Atomic Force Microscopy in Biomedical Research: Methods and Protocols (Methods in Molecular Biology Vol 736)

Methods

in

Molecular Biology™

Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield, Hertfordshire, AL10 9AB, UK



For further volumes: www.springer.com/series/7651

wwwwwwwwwwwwww

Atomic Force Microscopy in Biomedical Research Methods and Protocols

Edited by

Pier Carlo Braga Department of Pharmacology, School of Medicine, University of Milan, Milan, Italy

Davide Ricci Robotics, Brain and Cognitive Sciences Department, Italian Institute of Technology, Genoa, Italy and Department of Biophysical Electronic Engineering, University of Genoa, Genoa, Italy

Editors Pier Carlo Braga Department of Pharmacology School of Medicine University of Milan Milan, Italy [email protected]

Davide Ricci Robotics, Brain and Cognitive Sciences Department, Italian Institute of Technology Genoa, Italy and Department of Biophysical Electronic Engineering University of Genoa Genoa, Italy [email protected]

ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-61779-104-8 e-ISBN 978-1-61779-105-5 DOI 10.1007/978-1-61779-105-5 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011926794 © Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or ­dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. Printed on acid-free paper Humana Press is part of Springer Science+Business Media (www.springer.com)

Preface The invention and development of the optical microscope in the seventeenth century revealed the presence of a previously unseen and unimaginable world within and around us. Our lives would not be what they are today if optical microscopy had never existed or if it had not helped us to understand better what we are, how we function, and how we can improve our condition – first in the fields of biology and medicine, and then in many other fields. Another great step was made with the introduction of transmission and scanning electron microscopy in the 1930s, which was initially integrated with optical microscopy but subsequently developed its own identity and technology and opened up new horizons in human knowledge. Starting in 1986, further technological advances led to the development of atomic force microscopy (AFM), which is completely different from its predecessors: instead of being based on lenses, photons, and electrons, it directly explores the surface of the sample by means of a local scanning probe while the use of dedicated software allows the results to be visualized on a monitor. AFM has a number of special characteristics: very high magnification with very high resolution; minimal sample preparation (none of the dyes of optical microscopy, or the vacuum, critical point, or gold sputtering required by scanning electron microscopy); real three-dimensional topographical data that allow us to obtain different views of the samples from a single collected dataset; and the ability to work in a liquid in real time, thus making it possible to study the dynamic phenomena of living specimens in their biological environment and under near-physiological conditions. Over the years, an increasing number of researchers have started to use AFM and, in addition to a wide range of scientific articles, there are now also various books on the subject. In 2004, we edited a book published by Humana Press (Atomic Force Microscopy: Biomedical Methods and Applications) that described a series of practical AFM procedures in various applications with the aim of stimulating researchers to use the technique. We were therefore surprised when Humana Press proposed the publication of a second book on the subject so quickly after the first, and hesitated to accept the challenge. However, upon further reflection, we had to agree that the sheer breadth and originality of the new applications that have emerged since the first book was published more than justified this further review. The reason is quite simple: AFM is no longer simply just another form of microscopy, but has given rise to a completely new way of using microscopy that fulfils the dreams of all microscopists: being able to touch, move, and interact with the sample while it is being examined, thus making it possible to discover not only morphological, but also chemical and physical structural information. Optical microscopy made it possible to talk at the “micron” level (cells), and ­transmission and scanning electron microscopy introduced the idea of the “nano” level (sub-cellular), but still only in two dimensions; however, when speaking of AFM, it is not only usual to talk in three-dimensional “nano” terms, but it is also already possible to talk

v

vi

Preface

at the “pico” level (molecular). Together with continuous technical improvements, the reaching of this new dimensional range means that AFM can provide an opportunity to interact with individual molecules, observing them while we touch them and move them around in order to be able to discover their physical characteristics. All of this has also led to the development of a parallel “nano-technology” insofar as an AFM workstation has become a “nano-robot” that can dynamically interact with and manipulate samples on a “nano-scale”, and acquire information of sub-pico Newton “force spectroscopy” data on which to base the study of “nano-biology”. Functionalizing the AFM tip has made it possible to obtain “nano-biosensors” that can be used in the field of dynamic biomolecular processes in ways that could not even be imagined just a few years ago. Finally, combining AFM with other microscopic techniques, such as confocal or fluorescence microscopy is now being actively explored, and a number of interesting synergies have been discovered. This book brings together different types of applications in order to provide examples from different fields in the hope that this will stimulate researchers to apply their ingenuity in their own specialization and allow them to add significant originality to their studies. We gratefully acknowledge all of the contributions of our colleagues, each of whom donated their experience in order to cross-fertilize this new and fascinating technology. “GOD BLESS MICROSCOPY (ALL TYPES) …AND MICROSCOPISTS TOO ” because they show us what and how wonderful life is. Milan, Italy  Genoa, Italy 

Pier Carlo Braga Davide Ricci

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v xi

Part I The Basics of Atomic Force Microscopy   1 How the Atomic Force Microscope Works? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bruno Torre, Davide Ricci, and Pier Carlo Braga   2 Measurement Methods in Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . Bruno Torre, Claudio Canale, Davide Ricci, and Pier Carlo Braga   3 Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging . . . . . . Claudio Canale, Bruno Torre, Davide Ricci, and Pier Carlo Braga

3 19 31

Part ii Molecule Imaging   4 Imaging the Spatial Orientation of Subunits Within Membrane Receptors by Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stewart M. Carnally, J. Michael Edwardson, and Nelson P. Barrera   5 High Resolution Imaging of Immunoglobulin G Antibodies and Other Biomolecules Using Amplitude Modulation Atomic Force Microscopy in Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sergio Santos and Neil H. Thomson   6 Atomic Force Microscopy of Ex Vivo Amyloid Fibrils . . . . . . . . . . . . . . . . . . . . . Claudio Canale, Annalisa Relini, and Alessandra Gliozzi   7 Studying Collagen Self-Assembly by Time-Lapse High-Resolution Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clemens M. Franz and Daniel J. Muller   8 Atomic Force Microscopy Imaging of Human Metaphase Chromosomes in Liquid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Osamu Hoshi and Tatsuo Ushiki   9 Atomic Force Microscopy of Proteasome Assemblies . . . . . . . . . . . . . . . . . . . . . . Maria Gaczynska and Pawel A. Osmulski 10 Atomic Force Microscopy of Isolated Mitochondria . . . . . . . . . . . . . . . . . . . . . . . Bradley E. Layton and M. Brent Boyd 11 Imaging and Interrogating Native Membrane Proteins Using the Atomic Force Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andreas Engel

vii

47

61 81

97

109 117 133

153

viii

Contents

Part III Nanoscale Surface Analysis and Cell Imaging 12 Atomic Force Microscopy Investigation of Viruses . . . . . . . . . . . . . . . . . . . . . . . . Alexander McPherson and Yurii G. Kuznetsov 13 Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Christian Rankl, Linda Wildling, Isabel Neundlinger, Ferry Kienberger, Hermann Gruber, Dieter Blaas, and Peter Hinterdorfer 14 Atomic Force Microscopy as a Tool for the Study of the Ultrastructure of Trypanosomatid Parasites . . . . . . . . . . . . . . . . . . . . . . . . Wanderley de Souza, Gustavo M. Rocha, Kildare Miranda, Paulo M. Bisch, and Gilberto Weissmuller 15 Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Andreas Ebner, Hermann Schillers, and Peter Hinterdorfer 16 The Growth Cones of Living Neurons Probed by the Atomic Force Microscope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Davide Ricci, Massimo Grattarola, and Mariateresa Tedesco 17 Highlights on Ultrastructural Pathology of Human Sperm . . . . . . . . . . . . . . . . . . Narahari V. Joshi, Ibis Cruz, and Jesus A. Osuna 18 High-Speed Atomic Force Microscopy and Biomolecular Processes . . . . . . . . . . . Takayuki Uchihashi and Toshio Ando

171

197

211

223

243 259 285

Part IV Non-topographical Applications (Force-Spectroscopy) 19 Atomic Force Microscopy in Mechanobiology: Measuring Microelastic Heterogeneity of Living Cells . . . . . . . . . . . . . . . . . . . . . Evren U. Azeloglu and Kevin D. Costa 20 Force-Clamp Measurements of Receptor–Ligand Interactions . . . . . . . . . . . . . . . Félix Rico, Calvin Chu, and Vincent T. Moy 21 Measuring Cell Adhesion Forces: Theory and Principles . . . . . . . . . . . . . . . . . . . . Martin Benoit and Christine Selhuber-Unkel 22 Nanoscale Investigation on E. coli Adhesion to Modified Silicone Surfaces . . . . . . Ting Cao, Haiying Tang, Xuemei Liang, Anfeng Wang, Gregory W. Auner, Steven O. Salley, and K.Y. Simon Ng

303 331 355 379

Part V Investigating Drug Action 23 Imaging Bacterial Shape, Surface, and Appendages Before and After Treatment with Antibiotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 391 Pier Carlo Braga and Davide Ricci 24 Thymol-Induced Alterations in Candida albicans Imaged by Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 401 Pier Carlo Braga and Davide Ricci

Contents

ix

25 Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix Interactions in Vascular Smooth Muscle and Endothelial Cells . . . . . . . . . 411 Zhe Sun and Gerald A. Meininger 26 Atomic Force Microscopy Studies on Circular DNA Structural Changes by Vincristine and Aspirin . . . . . . . . . . . . . . . . . . . . . . . . . . . 425 Zhongdang Xiao, Lili Cao, Dan Zhu, and Zuhong Lu

Part VI Atomic Force Microscopy as a Nanotool 27 Combined Atomic Force Microscopy and Fluorescence Microscopy . . . . . . . . . . . Miklós S.Z. Kellermayer 28 Chemical Modifications of Atomic Force Microscopy Tips . . . . . . . . . . . . . . . . . . Régis Barattin and Normand Voyer 29 Atomic Force Microscopy as Nanorobot . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ning Xi, Carmen Kar Man Fung, Ruiguo Yang, King Wai Chiu Lai, Donna H. Wang, Kristina Seiffert-Sinha, Animesh A. Sinha, Guangyong Li, and Lianqing Liu Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

439 457 485

505

wwwwwwwwwwwwww

Contributors Toshio Ando  •  Department of Physics, Kanazawa University, Kanazawa, Japan; CREST, JST, Tokyo, Japan Gregory W. Auner  •  College of Engineering, Wayne State University, Detroit, MI, USA Evren U. Azeloglu  •  Department of Pharmacology and Systems Therapeutics, Mount Sinai School of Medicine, New York, NY, USA Régis Barattin  •  CEA-Grenoble, Grenoble, France Nelson P. Barrera  •  Department of Physiology, Pontificia Universidad Católica de Chile, Santiago, Chile Martin Benoit  •  Lehrstuhl für Angewandte Physik, LMU, Sektion Physik, München, Germany Paulo M. Bisch  •  Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, Brazil Dieter Blaas  •  Department of Medical Biochemistry, Max F. Perutz Laboratories, Vienna Biocenter, Medical University of Vienna, Vienna, Austria M. Brent Boyd  •  Department of Mechanical Engineering and Mechanics, Drexel University, Philadelphia, PA, USA Pier Carlo Braga  •  Department of Pharmacology, School of Medicine, University of Milan, Milan, Italy Claudio Canale  •  Nanophysics Unit, Italian Institute of Technology, Genoa, Italy Lili Cao  •  State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Ting Cao  •  College of Engineering, Wayne State University, Detroit, MI, USA Stewart M. Carnally  •  Department of Pharmacology, University of Cambridge, Cambridge, UK Calvin Chu  •  Miller School of Medicine, University of Miami, Miami, FL, USA Kevin D. Costa  •  Cardiovascular Research Center, Mount Sinai School of Medicine, New York, NY, USA Ibis Cruz  •  Department of Physiology, Laboratory of Andrology, University of Los Andes, Merida, Venezuela Wanderley de Souza  •  Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, Brazil; Diretoria de Programas, Instituto Nacional de Metrologia, Normalização e Qualidade Industrial – INMETRO, Rio Comprido, Rio de Janeiro, Brazil Andreas Ebner  •  Institute for Biophysics, University of Linz, Linz, Austria J. Michael Edwardson  •  Department of Pharmacology, University of Cambridge, Cambridge, UK

xi

xii

Contributors

Andreas Engel  •  Maurice E. Müller Institute for Structural Biology, Biozentrum, University of Basel, Basel, Switzerland; Department of Pharmacology, Case Western Reserve University, Cleveland, OH, USA Clemens M. Franz  •  DFG-Center for Functional Nanostructures, Karlsruhe Institute of Technology, Karlsruhe, Germany Carmen Kar Man Fung  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Maria Gaczynska  •  Department of Molecular Medicine, Institute of Biotechnology, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA Alessandra Gliozzi  •  Department of Physics, University of Genoa, Genoa, Italy Massimo Grattarola  •  Dipartimento di Ingegneria Biofisica ed Elettronica, University of Genoa, Genoa, Italy Hermann Gruber  •  Institute for Biophysics, University of Linz, Linz, Austria Peter Hinterdorfer  •  Institute for Biophysics, University of Linz, Linz, Austria Osamu Hoshi  •  Division of Microscopic Anatomy and Bio-Imaging, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan Narahari V. Joshi  •  Department of Physiology, University of Los Andes, Merida, Venezuela Miklós S.Z. Kellermayer  •  Department of Biophysics and Radiation Biology, Semmelweis University, Budapest, Hungary Ferry Kienberger  •  Institute for Biophysics, University of Linz, Linz, Austria; Agilent Technologies Austria GmbH, Linz, Austria Yurii G. Kuznetsov  •  Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA King Wai Chiu Lai  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Bradley E. Layton  •  Applied Computing and Electronics, The University of Montana College of Technology, Missoula, MT, USA Guangyong Li  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Xuemei Liang  •  College of Engineering, Wayne State University, Detroit, MI, USA Lianqing Liu  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Zuhong Lu  •  State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Alexander McPherson  •  Department of Molecular Biology and Biochemistry, University of California, Irvine, CA, USA Gerald A. Meininger  •  Department of Medical Pharmacology and Physiology, Dalton Cardiovascular Research Center, University of Missouri-Columbia, Columbia, MO, USA

Contributors

xiii

Kildare Miranda  •  Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, Brazil; Diretoria de Programas, Instituto Nacional de Metrologia, Normalização e Qualidade Industrial – INMETRO, Rio Comprido, Rio de Janeiro, Brazil Vincent T. Moy  •  Miller School of Medicine, University of Miami, Miami, FL, USA Daniel J. Muller  •  Department of Biosystems Science and Engineering, ETH Zurich, Basel, Switzerland Isabel Neundlinger  •  Institute for Biophysics, University of Linz, Linz, Austria K.Y. Simon Ng  •  College of Engineering, Wayne State University, Detroit, MI, USA Pawel A. Osmulski  •  Department of Molecular Medicine, Institute of Biotechnology, University of Texas Health Science Center at San Antonio, San Antonio, TX, USA Jesus A. Osuna  •  Department of Physiology, Laboratory of Andrology, University of Los Andes, Merida, Venezuela Christian Rankl  •  Institute for Biophysics, University of Linz, Linz, Austria; Agilent Technologies Austria GmbH, Linz, Austria Annalisa Relini  •  Department of Physics, University of Genoa, Genoa, Italy Davide Ricci  •  Robotics, Brain and Cognitive Sciences Department, Italian Institute of Technology, Genoa, Italy; Department of Biophysical Electronic Engineering, University of Genoa, Genoa, Italy Félix Rico  •  Centre de Recherche, Institut Curie, UMR168-CNRS, Paris, France Gustavo M. Rocha  •  Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, Brazil; Diretoria de Programas, Instituto Nacional de Metrologia, Normalização e Qualidade Industrial – INMETRO, Rio Comprido, Rio de Janeiro, Brazil Steven O. Salley  •  College of Engineering, Wayne State University, Detroit, MI, USA Sergio Santos  •  School of Physics and Astronomy, University of Leeds, Leeds, UK Hermann Schillers  •  Institut fur Physiologie II, University Munster, Munster, Germany Kristina Seiffert-Sinha  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Christine Selhuber-Unkel  •  Institute for Materials Science, University of Kiel, Kiel, Germany Animesh A. Sinha  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Zhe Sun  •  Dalton Cardiovascular Research Center, University of Missouri-Columbia, Columbia, MO, USA Haiying Tang  •  College of Engineering, Wayne State University, Detroit, MI, USA

xiv

Contributors

Mariateresa Tedesco  •  Dipartimento di Ingegneria Biofisica ed Elettronica, University of Genoa, Genoa, Italy Neil H. Thomson  •  School of Physics and Astronomy, University of Leeds, Leeds, UK Bruno Torre  •  Italian Institute of Technology, Genoa, Italy Takayuki Uchihashi  •  Department of Physics, Kanazawa University, Kanazawa, Japan; CREST, JST, Tokyo, Japan Tatsuo Ushiki  •  Division of Microscopic Anatomy and Bio-Imaging, Niigata University Graduate School of Medical and Dental Sciences, Niigata, Japan Normand Voyer  •  Département de chimie, Université Laval, Quebec, QC, Canada Anfeng Wang  •  College of Engineering, Wayne State University, Detroit, MI, USA Donna H. Wang  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Gilberto Weissmuller  •  Instituto de Biofísica Carlos Chagas Filho, Universidade Federal do Rio de Janeiro, Ilha do Fundão, Rio de Janeiro, Brazil Linda Wildling  •  Institute for Biophysics, University of Linz, Linz, Austria Ning Xi  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Zhongdang Xiao  •  State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China Ruiguo Yang  •  Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI, USA Dan Zhu  •  State Key Laboratory of Bioelectronics, School of Biological Science and Medical Engineering, Southeast University, Nanjing, China

Part I The Basics of Atomic Force Microscopy

wwwwwwwwwwwwww

Chapter 1 How the Atomic Force Microscope Works? Bruno Torre, Davide Ricci, and Pier Carlo Braga Abstract This chapter aims at giving a quick but precise introduction of the atomic force microscope from the working principle point of view. It is intended to provide a useful starting point to those who first approach the instrument giving a general sketch of the working principles and technical implementations as well as last improvements. Subheading 1 is introductory: it gives an overview of what the instrument does and why it has been developed. Subheading 2 is focused on measurement ranges and on the comparison with scanning electron microscope (SEM) and transmission electron microscope (TEM) which have similar ranges and resolutions but different sample interactions and applications. Subheading 3 gives an overview of the working principles and the most diffused technical implementations on which most of the commercial microscopes rely, as we think it gives the useful base knowledge to understand possible applications, instrument capabilities, and results. In particular, technical improvements taking place over the past few years are highlighted. Despite of the simple and not very technical approach, it has a key importance in understanding concepts at the base of Chapter 3, which is, on the other side, useful for beginners and experienced users as well. Subheading 4 compares different instrument architectures and can, therefore, be useful for those who are going to choose an instrument having clear final applications. Latest solutions are once more highlighted. Subheading 5 gives an overview and some suggestions to start working, both in air and in liquid. Following the general philosophy of the book, it follows more an “how to do” concept than a general theoretical approach. Subheading 6 contains the future developments of the techniques. Key words: Introduction to AFM, AFM working principles, AFM basics

1. Introduction Microscopes have always been one of the essential instruments for research in the biomedical field. The capability of optical ­microscopes to magnify and resolve details well below 1 mm has soon reached its intrinsic physical limit due to the well-known “diffraction limit”: when radiation hits obstacles of size comparable with its wavelength (visible light: 380–750 nm), diffraction

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_1, © Springer Science+Business Media, LLC 2011

3

4

Torre, Ricci, and Braga

and interference became important, and smaller details cannot be distinguished. Historically, two solutions have been found to image samples with few nanometers resolution or better: the first one is to shorten radiation wavelength, using ultraviolet, X-ray, or electronbased microscopes, to push the diffraction limit from hundreds to few nanometers scale. Radiation-based microscopes (such as the light microscope and the electron microscope) have become trustworthy companions in the laboratory and have contributed greatly to our scientific knowledge. However, short-wavelength radiation can induce sample damaging because high-energy interaction can be involved; moreover, measurements often require special sample preparation or controlled (vacuum) conditions, that can often be incompatible with physiological environment or in vivo measurements. A second strategy relies on a completely different system: a very sharp tip is set in (weak) interaction with the sample and rastered on it while interaction is measured and controlled. In this way, the tip tracks surface morphology while its XYZ position is registered by the electronics to compose a 3D map of the sample surface. Since the interaction can be controlled and limited to very low values, this kind of imaging is usually nondestructive. Depending on the type of interaction measured, scanning probe microscopes (SPMs) take different names, such as scanning tunneling (STM; current between tip and samples), scanning near field optical (SNOM; optical coupling), atomic force (AFM; force between last part of the tip and sample) (1), etc. Historically, AFM has been invented after STM to allow measurements on insulating samples: very soon it was clear that it could archive nanometer resolution working in different environments – air, liquid, or vacuum – regardless of conductive or optical properties of the sample. Moreover, it measures (and controls) tip–sample interaction forces and, therefore, it allows to probe (nano) mechanical properties of the specimen applying pressure or pulling the sample. Hereafter, we refer mainly to AFM for its wide range of applicability in biological field. After using it for the first time, three things can be noticed: ●●

●●

●●

Despite of a rather high-sounding name, imaging with the AFM can be quite simple: no special sample preparation is required, and images with unexpected resolution can be obtained on the very first time; Images are real 3D ones: height is measured with even higher resolution, and subnanometer steps are commonly resolved; questions on surface corrugation and feature height can be easily answered. At first use, imaging appears quite slow: one image can take some minutes to acquire and video rate measurements are not possible with commonly used instruments: this is a

How the Atomic Force Microscope Works?

5

c­ ommon feature for all the mechanical scanning techniques. Nevertheless, a few minutes (without preparation needed) time interval is compatible with most experiments even for biological applications. After some experience, one learns that in some cases it is ­ ossible to push resolution to the “atomic” level (2–4) and p that  images do contain details not observable with any other instrument. A noteworthy feature is that imaging is only one of the experiments that can be performed with the instrument: the tip can be pushed on the sample, pulled out, used to make scratches, functionalized to bind to specific chemical groups, electrically connected to detect currents or potentials, and used to induce catalytic reactions or for lithography purpose. The number of experiments that can be redesigned on the nanometer scale seems to be limited just by applicant imagination: this capability gave rise to a new definition for AFM applications, “lab on tip.” These are, anyway, advanced techniques and are not described in this work: readers interested in the topic are referred to specialized literature.

2. Performance Range of AFM AFM images show significant information about surface features with unprecedented clarity. The AFM can perform nondestructive examinations on any sufficiently “rigid” surface either in air or in liquid, regardless if the specimen is insulating, conductive, transparent, or opaque. Modern instruments can be endowed with temperature control stages and closed chamber for environmental control; some of them are especially designed to be coupled with an optical microscope for simultaneous imaging through advanced optical techniques so that a huge variety of complementary information can be archived. The field of view can vary from the atomic and molecular scale up to sizes larger than 100 mm so that data can be coupled with other information obtained with lower resolution – and wider field of view – techniques. The AFM can also examine rough surfaces with (sub)nanometer resolution on the vertical range up to more than 10 mm; large samples can be fitted directly in the microscope without cutting. With stand-alone instruments, any area on flat or nearly flat specimens can be investigated. Compared with the SEM, AFM provides topographic contrast of surface features with quantitative height information. Moreover, as the sample need not be electrically conductive, no metallic coating of the sample is required. Hence, no dehydration of the sample is necessary as with SEM, and samples may be imaged in their hydrated state. This eliminates the shrinkage of biofilm associated with SEM imaging, yielding a nondestructive

6

Torre, Ricci, and Braga

technique. The resolution of AFM is higher than that of environmental SEM, where hydrated images can also be obtained and extracellular polymeric substances may not be imaged. Compared with transmission electron microscopes, where the electron beam gives a planar projection of the sample by flowing through it, AFM images give information on 3D properties of the surface: in this sense, these two techniques can be regarded as the most complementary ones, since one (TEM) provides contrast on inner structures of the sample, but it is intrinsically 2D, while the other one (AFM) gives real 3D images with similar resolution, but it can only access to the exposed surface. Finally, it can be commented that with respect to TEM, no expensive and destructive (cross-sectioning) sample preparation is needed. Moreover, image contrast is quantitative and can be expressed in nanometer units by default and this is a pretty unique characteristic, allowing direct comparison between different samples. In the following subheadings, we give a brief outline of how the AFM works followed by a description of the parts that can be added to the basic instrument. Our overview has no pretense of completeness but aims at simplicity. For a more thorough description of the physical principles involved in the operation of these instruments, we refer you to the specialized literature.

3. The Microscope In Fig. 1, a schematic diagram of the AFM working principle is shown (1, 5). In principle, AFM can remind one of those old style record players, but it incorporates a number of technical solutions that allow to detect atomic-scale corrugation: very sharp tips at the end of flexible cantilevers and a sensitive deflection sensing system capable of controlling with high accuracy the tip–sample relative position are used. A basic configuration is made up as follows: ●●

●●

●●

●●

A 3D positioning system, called scanner, to adjust tip–sample relative position: if the tip is attached to the scanner, the configuration is called scanning probe; otherwise (as marked with 1 in Fig. 1a), if the tip is fixed and the sample is moved it is called scanning sample. A sample holder where the specimen can be placed in a stable configuration (Fig. 1b). A sharp tip at the end of a flexible cantilever (marker 3 in Fig. 1a and b). A deflection detecting system: in Fig. 1a and c (marker 4), the widely used optical beam deflection (OBD) configuration is

How the Atomic Force Microscope Works?

7

Fig. 1. Schematic diagram of a scanned-sample AFM, based on five quadrant piezo scanner configuration (see below). In the case of scanned probe, it is the tip that is scanned instead of the sample. (a) The piezoelectric scanner (1) is the (nano)positioning element allowing movement: it works by applying opposite voltages to ±X and ±Y sectors to move the sample in X and Y directions, respectively; an additional Z sector moves the sample in the vertical direction. The sample (2) is positioned on the scanner; (3) cantilever; (4) optical beam deflection system (OBD) to detect tip displacement; (5) position-sensitive photodetector (PSD) and preamplifier; (6) electronics. (b) A magnification of tip and sample. (c) A detail of the OBD system.

shown; in this configuration, a tiny tip displacement is detected by a laser beam, amplifying the deflection of the cantilever holding the tip. Laser light, reflected from the rear of the cantilever, is centered on a (usually four vector) photodiode by means of mirrors placed on the optical path. This method allows good signal amplification and it is of rather simple use ; therefore, it is employed on almost all commercial instruments. ●●

●●

Some signal conditioning and preamplifying stage: in case of OBD system, signals from sector A, B, C, and D are used to calculate overall power SUM = A + B + C + D, normal deflection N = (A + B − C − D)/SUM, and lateral deflection as L = (A + C − B − D)/SUM. A digital control system to control tip–sample position on the basis of collected signals.

The following sections contain further details on single ­components and on the working principle.

8

Torre, Ricci, and Braga

3.1. The Scanner

High-resolution (nanometer) positioning can be performed using piezoelectric ceramic materials. These materials undergo a reversible deformation when an external (high) bias voltage is applied across two opposite faces of it: in a first approximation, such deformation can be considered to depend linearly on the applied voltage. A widely used scanner configuration relies on piezoelectric tubes made up by four or five sectors (Fig. 2). A differential bias (with respect to the inner part of the tube, grounded) applied to opposite electrodes induces a bending of the tube in one of the two main directions, while common mode voltage induces a contraction or elongation in the vertical direction. The same happens for the other two electrodes so that differential signals can be used for X and Y movement and common bias for Z movement and four electrodes are sufficient to provide a complete 3D positioning. Anyway, for technical reasons, it is preferable to decouple the Z movement from the XY one, by adding a fifth dedicated electrode (see Figs. 1a and 2) so that a common voltage to side electrodes is no more needed. As shown in Fig. 2, this configuration gives an undesired parabolic component to the motion, therefore this type of scanners are usually endowed with embedded positioning sensors that allow distortion compensation and linearization. This kind of distortion (commonly referred as bow) is mostly relevant for high scan ranges (above some micrometers) and becomes less important for smaller regions, that is to say in case of high resolution: for this reason, some instruments allow to operate also in open-loop mode (i.e., without sensor compensation) for high-resolution imaging, to further reduce electrical noise of readout circuitry. Modern microscopes use a slightly different configuration: single linear piezoelectric elements are embedded in a metallic frame machined by electroerosion to be easily deformable in

Fig. 2. Working principle of a five quadrant piezo tube: right image shows how deflection occurs upon differential ­biasing of two opposite sectors, here −X and +X; the Z sector is visible just below the sample holder on the upper part of the tube. Figure is not in scale and deflection is intentionally exaggerated to highlight the effect.

How the Atomic Force Microscope Works?

9

­ redefined directions (flexure system); for each axis, a ­deformation p occurs easily in the parallel direction to the piezoelectric strain so that the three directions are efficiently decoupled on the three axes. The frames usually incorporate low noise, often capacitive or inductive positioning sensors, and therefore are good candidates for metrology purposes. In some configurations, one of the axes (vertical one) is also mechanically and physically decoupled from the other two. A few words can be spent on the topic of positioning sensors to detect displacements on the nanometer scale. Neglecting for the moment interferometric solutions, that are often used for metrological standards but are not very easy to be integrated, three different accurate positioning sensors are commercially available: ●●

●●

●●

Strain gauge (resistive) sensors can be easily integrated even on piezoelectric tubes by simply gluing them: upon deformation, sensors change their electrical resistance that can be directly read by the electronics. Anyway, since resistors are intrinsically thermal noise generators, this kind of sensors is noisy and resolution is usually limited to a few nanometers. Capacitive sensors: basically they are made up by a capacitor with one plate coupled with the moving part and one fixed to a standing position; a change in the relative position of the two plates implies a change in the relative capacitance that is electrically detected. This type of sensors has very low noise and commonly allows sensitivities of the order of tens of nanometers or better. Integration of these sensors in a scanner is more difficult than for the strain-based ones, and parallelism between faces is often an issue, so they are more often found on flexure scanners than on piezo tubes motors. Inductive or eddy current sensors: recently, some commercially available microscopes have successfully employed inductive sensors for embedded position detection. These have reported resolutions of few tens of nanometer.

Readers interested on this topic can find further details in ref.  6. For our purpose, it is sufficient to keep in mind that in some lower cost microscope, where this compensation is not implemented, a postprocessing software correction can always be performed. 3.2. Tip and Cantilever

The tip, which is mounted at the end of a small cantilever, is the heart of the instrument because it is brought in closest contact with the sample and gives rise to images though its force interactions with the surface. When the first AFM was made, a very small diamond fragment was carefully glued to one end of a tiny piece of gold foil. Today, the tip–cantilever assembly typically is

10

Torre, Ricci, and Braga

Fig. 3. The essential parameters in a tip are the radius of curvature (R ) and the aspect ratio (ratio of H to W ).

f­abricated from silicon or silicon nitride and, using technology similar to that applied to integrated circuit fabrication, allows a good uniformity of characteristics and reproducibility of results (7, 8). The essential parameters to consider are the sharpness of the apex, measured by the radius of curvature (spherical approximation), and the aspect ratio of the whole tip (Fig. 3). Nowadays, a variety of cantilevers are commercially available: in addition to standard pyramid tips, usually 3  mm tall with approximately 30-nm apex radius, also tetragonal, high aspect ratio and conical tips can be found. The tip can end with diamond-like carbon spikes, carbon nanotubes, or whiskers for low curvature radius, and they can also be further machined by means of focused electron beam (FEB) or focused ion beam (FIB) to obtain even higher aspect ratios. Commercial tips commonly end with curvature radius smaller than 10 nm, but ultrasharp tips with R < 2 nm or even <1 nm are commercially available. Moreover, tips can be coated with metal films to enhance conductivity in the contact area, or to obtain magnetic properties; low electrical resistance doped silicon tips are also available. Chemically functionalized tips can be also purchased for applications involving specific bindings with biochemical species on the sample surface. Although it would seem that sharper tips should yield more detailed images, this may not occur with all samples: in fact, quite often, the so-called atomic resolution on crystals is obtained best with standard silicon nitride tips. This is because reducing apex radius has the drawback of increasing tip fragility. Moreover, the measurement load rises quickly due to contact area reduction, and this can lead to quick tip erosion or to sample damage.

How the Atomic Force Microscope Works?

11

Fig. 4. SEM image of triangular (A) and single-beam (B) cantilevers (MLCT silicon nitride probe, Veeco) (courtesy S. Marras). The mechanical properties, such as the force constant and resonant frequency, depend on the values of width (W ), length (L), and thickness (T ). Bottom right image shows pyramidal tip.

The cantilever carrying the tip is attached to a small glass “chip” that allows easy handling and positioning of the instrument. There are essentially two designs for cantilevers, the “V” shaped and the single-arm kind (Fig. 4), which have different torsional properties. The length, width, and thickness of the beam(s) determine the mechanical properties of the cantilever and have to be chosen depending on mode of operation needed and on the sample to be investigated. Cantilevers are essentially classified by their force (or spring) constant and resonance frequency: soft and low-resonance frequency cantilevers are more suitable for imaging in contact and resonance mode in liquid, whereas stiff and high-resonance frequency cantilevers are more appropriate for resonance mode in air (9). 3.3. Deflection Sensor

AFMs can generally measure the vertical deflection of the cantilever with picometer resolution. To achieve this, most AFMs today use the optical lever or OBD method that achieves resolution comparable to an interferometer while remaining inexpensive and easy to use.

12

Torre, Ricci, and Braga

In this system, a laser beam is reflected on the backside of the cantilever (often coated by a thin metal layer to enhance reflectivity) onto a position-sensitive photodetector (PSD), consisting of two (more often four, as in Fig. 1) side-by-side mounted photodiodes. In this arrangement, a small deflection of the cantilever tilts the reflected beam and changes the position of the light spot on the photodetector. The signal difference between the different sections of the photodiode indicates the position of the laser spot on the detector, and thus the deflection of the cantilever. Because the distance between cantilever and detector is generally three orders of magnitude greater than the length of the cantilever (millimeters compared to micrometers), the optical lever greatly magnifies motions of the tip giving rise to an extremely high sensitivity. 3.4. Image Formation

Images are formed by recording the effects of the interaction forces between tip and surface as the cantilever is scanned over the sample. The scanner and the electronic feedback circuit, together with sample, cantilever, and optical lever form a feedback loop set up for the purpose. The presence of a feedback loop is a key difference between AFM and older stylus-based instruments so that AFM not only measures the force on the sample, but also controls it, allowing acquisition of images at very low tipto-sample forces (5, 10). The scanner is an extremely accurate positioning stage used to move the tip over the sample (or the sample under the tip) to form an image. The AFM electronics drives the scanner across the first line of the scan and back. It then steps in the perpendicular direction to the second scan line, moves across it and back, then to the third line, and so forth (Fig. 5). Usually, both forth and back traces (trace and retrace) are recorded, giving two images that ideally should overlap: if images differ at some point, this can

Fig. 5. Raster scan for image acquisition. The AFM electronics drive the scanner across the first line of the scan and back. The scanner then steps in the perpendicular direction to the second scan line, moves across it and back, then to the third line, and so forth.

How the Atomic Force Microscope Works?

13

be due to tip contamination, sample modification, or to an improper choice of measurement parameters, and therefore suggests that something needs to be adjusted. As the probe is scanned over the surface, a topographic image is obtained storing the vertical control signals sent by the feedback circuit to the scanner moving it up and down to follow the surface morphology while keeping the interaction forces constant. The image data are sampled digitally at equally spaced intervals, up to some thousands points per line. The number of lines is usually chosen to be equal to the number of data points per line, obtaining at the end a square grid of data points each corresponding to the relative X, Y, and Z coordinates in space of the sample surface (11). Usually, during scanning, data are represented by gray scale or RGB images, in which the brightness of points can range from black to white across 256 levels, corresponding to the information acquired by the microscope (that can be height, force, phase, and so on); anyway, data are usually collected with higher resolution, since they are digitalized as 16 bit data (65,536 ­levels) or better, therefore the available information is by far more than what is displayed on the screen. Usually, microscopes are endowed with software solutions allowing statistical analysis and quantitative mathematical parameterizations of collected data. A number of free software programs, compatible with the more widespread AFM file formats, can also be downloaded from the Web (12).

4. Instruments, Architectures, and Options

The first instruments introduced on the market had all very similar features and range of applications. They had scanners with small range, limited optical access, and could accommodate only small samples. Essentially, they were built to make very highresolution imaging on flat samples in a dry environment. With the development of new technical solutions, fields of  application grew very rapidly, and now it is possible to find ­instrument add-ons and architectures allowing to perform very sophisticated measurements in different fields: we can now find instruments that are specifically designed for large samples, such as silicon wafers, that have metrological capabilities, utilize closedloop scanners that are optimized for liquid and electrochemistry operation and can be mounted on an inverted microscope for biological investigations. Usually, one single instrument can have different options to extend its capabilities to a wide range of applications and a huge variety of experiments are easily software controllable so that even nonexperts can relatively quickly perform advanced experiments; anyway, it is still true that instruments are designed keeping in mind some particular application

14

Torre, Ricci, and Braga

so that they are optimized for a specific purpose. For this reason, it is ­necessary to have clearly in mind the main features that are desired in an instrument before its purchase, understanding at the same time that a loss of performance in other applications may be possible. One can still distinguish, to clarify this point, essentially between two main instrumentation categories, even if intermediate or hybrid configurations are possible: scanned-sample and scanned-tip microscopes. With reference to the widely used piezo tube configuration, we give a brief description of the advantages of one system with respect to the other and of their limitations; finally, we are going to clarify how these problems are minimized, thanks to the introduction of modern flexure scanners. 4.1. Scanned Sample

This scanned-sample AFM is the oldest design in which the sample is attached to the scanner and moved under the tip. Depending on how the cantilever holder, laser, and photodetector are assembled, it can easily accommodate an overhead microscope provided that long focal length objectives are used. A clear view of where the tip is landing is usually possible, speeding up the time it takes to get a meaningful image of the sample. Since in this configuration scanner and optical lever are physically independent, it is less packed, allowing an easier optical access to the tip and to the scanner. Scanners with wide X, Y, and Z range are usually available and closed-loop control feedback is more easily implemented in this scheme, and often a lower mechanical noise level can be obtained allowing higher ultimate resolution. There are quite a few drawbacks. First of all, the size and weight of the sample have to be limited because it is sitting on the scanner and may change its behavior. For the same reason, operation in liquid is impaired because liquid cells tend to be small and difficult to seal, and liquid flow or temperature control are more complicated to implement. Notwithstanding these difficulties, excellent results can be obtained on typical biomedical science specimens by ingeniously adapting them to the instrument’s characteristics.

4.2. Scanned Tip

In the scanned-tip method of operation, the sample stays still and it is the cantilever, attached to the scanner, which is moved across the surface. Although for scanning tunneling microscopes this was one of the first solutions applied, to build a scanned-tip AFM requires overcoming some difficulties, essentially related to adapting the beam bounce detection scheme to a moving cantilever. In fact, because of its weight, it is not possible to mount all the optical setup on the moving assembly and some position-related undesired effects can often occur. Some design couple correction lenses moving with the tip to partially compensate for these

How the Atomic Force Microscope Works?

15

effects. One advantage of this configuration is that no limitations in sample weight or size occur: since the scanner is loaded with always the same weight and the sample is still, its mechanical performance is independent from the sample changes. Instrument operation can be easily automated, by coarse positioning the sample below the scanner using a micrometric motor, and some models are endowed with three supports that enable them to scan the surface of any object under their probe. More recently, specialized instruments were developed, capable of being coupled or even integrated into inverted optical microscopes for biological applications. With respect to the scanned-sample models, scanned-tip instruments can be more easily equipped with temperaturecontrolled stages, open or closed liquid cells, liquid flow systems, electrochemistry cells, and controlled atmosphere chambers. Concerning limitations, one could say that what is gained on one side is lost on the other. For example, often the overall noise level is higher, limiting ultimate resolution. Large scan areas are more difficult to perform because tracking systems have to be used to keep the laser spot on the back of the cantilever. A top view of samples is obstructed by the scanner assembly: special hollow tubes have been developed recently, but even so on-axis microscopes, which are useful on nontransparent samples, still have limited resolution and lateral field of view. As anticipated, these limitations, occurring in both cases, have been greatly reduced by the introduction of flexure technology for scanners: the rigid metallic frame is by far less sensitive to loads so that heavier samples can be measured without affecting microscope capabilities (if the sample is scanned) or the whole optical lever can be moved with the tip (scanned-tip configuration), eliminating undesired tip-lased displacements. Some intermediate configurations scan the sample in the XY direction, while moving the tip (and the optical lever) in Z direction to physically decouple raster from retroactive movements. Moreover, they can be easily endowed with closed liquid cells, external circuitries, and coarse movements for advanced experiments or easy repositioning. The choice of configurations in most cases is more influenced by the possibility to couple the instrument with inverted or top-view optical microscopes, than from real technical limitations. Even if modern flexure stage microscopes prevents most of the architecture-related problems, it is still true that the greater the mechanical paths (big frame for large scanning, “open” architecture for easy optical access), the higher are the effects of mechanical vibrations or thermal expansions so that high (atomic) resolution instruments still rely on a very packed design and usually do not allow scanning frames larger then few micrometers.

16

Torre, Ricci, and Braga

5. Loading a Sample in the Microscope 5.1. Imaging Dry Samples

5.2. Imaging in Liquid

Samples to be viewed in atmospheric environment are often simply glued to a sample holder, usually a metal disk to be magnetically positioned or a microscope glass. An essential feature is that the sample has to adhere firmly adherent to the sample holder; otherwise, very poor imaging is achieved. For this reason, one has to be careful in the choice of the glue or sticky tape: slow drying glue or thick sticky tape should be avoided. A drawback is that after use in the AFM, the sample is difficult to take off without damage. Some systems, usually scanned-tip ones, can accept samples directly, securing them with a metal clip or springs. This method allows sample recovery without damage for further use in other experiments, but it can be less stable and needs special care for high-resolution work. Sometimes, because of the ease of use of the AFM, one forgets to be careful while handling the sample: fingerprints, dust, or scratches contaminate sample and affect all measurements, therefore one should avoid touching the surface in any case. To remove some dust on the surface, one can try to gently flush dry gas, e.g., nitrogen, on the sample, obtaining, in some cases, benefits. As a general rule, it is better to wear (powder free) gloves and to use tweezers and clean tools to handle the specimen (for these reason toolsets are provided with most microscopes). Also, it is best to keep a reserved area of the laboratory free from contaminants for the operations of sample and cantilever mounting. One of the main reasons for the success of AFM in biomedical investigations is its ability to scan samples in physiological condition, that is, immersed in liquid solutions (13, 14). Just to make an example, scanned-tip systems can often be directly used to image cells into a standard Petri dish. Each manufacturer has its own design of liquid cells, sometimes different ones depending on the application, and users may decide to make their own to fit specific needs. A few additional things that have to be taken care of when imaging in liquid are the temperature of the solution (eventually added during imaging; ref. 15), maintenance of the liquid cell, and cantilever holder assembly. Because the cantilever is extremely sensitive to temperature changes, it is important to let the system equilibrate before taking images. For example, in the case of contact-mode imaging with silicon nitride cantilevers and tips, a large variation in time of the signal on the photodetector corresponding to cantilever deflection can be observed in the presence of a temperature change (16). If temperature is not stable prior to approach of the tip to the sample and one starts taking images, after some time the applied force could be quite different than at the beginning of the imaging session.

How the Atomic Force Microscope Works?

17

If liquid has to be refilled or replaced, a good rule is to avoid abrupt temperature changes to the cantilever and to the sample: if the microscope operates in a closed acoustic box, we suggest keeping some small amount of spare liquid inside so that it stays at the same temperature of the system. Another problem can arise from bubble formations. Quite often, some air remains trapped below the tip or above the sample: in these cases, the cantilever is bent toward the surface and landing on the sample is not possible. This effect can be easily recognized because the deflection signal does not give reasonable values and sometimes the optical lever cannot be aligned. This effect occurs very often at first landing and during liquid refilling, mostly if the added liquid has a different temperature from the measurement bath. In these cases, the best solution is to remove the cantilever holder, dry the sensor with some gentle nitrogen flux, and remount the setup. Usually, wetting the cantilever with a droplet of buffer liquid before starting the experiment helps to prevent gas trapping. Once finished using the microscope for imaging in liquid, it is essential to immediately clean thoroughly all parts that have been in contact with the solution to avoid contamination of future experiments. Usually, it should be possible to disassemble and sonicate all vital parts of the liquid cell and the cantilever holder.

6. Future Developments The AFM is part of a family of SPMs that has a great growth potential. It is a fact that the majority of novel applications and techniques developed in SPMs in the last years are related to the life sciences. There is still much room for technical improvement: electronics, scanners, and tips are constantly improving. Scan speed limitations, sample accessibility, and ease of use have been addressed and can be still improved. As more and more biomedical researchers will be involved in the use of AFM, with their experience they will be able to contribute in developing an instrument less related to the physical sciences (its origin) and more tailored to our specific needs. References 1. Binnig, G., Quate, C. F., and Gerber, Ch. (1986) Atomic force microscope. Phys. Rev. Lett. 56, 930–933. 2. Binnig, G., Gerber, C., Stoll, E., Albrecht, T. R., and Quate, C. F. (1987) Atomic resolution with the atomic force microscope. Europhys. Lett. 3, 1281–1286.

3. Hug, H. J., Lantz, M. A., Abdurixit, A., et al. (2001) Subatomic features in atomic force microscopy images. Science 291, 2509. 4. Jarvis, M. R., Perez, R., and Payne, M. C. (2001) Can atomic force microscopy achieve atomic resolution in contact mode? Phys. Rev. Lett. 86, 1287–1290.

18

Torre, Ricci, and Braga

5. Alexander, S., Hellemans, L., Marti, O., et al. (1989) An atomic-resolution atomic-force microscope implemented using an optical lever. J. Appl. Phys. 65, 164–167. 6. Kwon, J., Hong, J., Kim,Y.S., Lee, D.Y., Lee K., Lee, S., Park,S.,(2003) Atomic force microscope with improved scan accuracy, scan speed,and optical vision Rev. Sci. Inst. V 74, N 10, 4378–4383. 7. Albrecht, T. R., Akamine, S., Carver, T.E., and Quate, C. F. (1990) Microfabrication of cantilever styli for the atomic force microscope. J. Vac. Sci. Technol. A 8, 3386–3396. 8. Tortonese, M. (1997). Cantilevers and tips for atomic force microscopy. IEEE Engl. Med. Biol. Mag. 16, 28–33. 9. Cleveland, J. P., Manne, S., Bocek, D., and Hansma, P. K. (1993) A non-destructive method for determining the spring constant of cantilevers for scanning force microscopy. Rev. Sci. Instrum. 64, 403–405. 10. Meyer, G. and Amer, N. M. (1988) Novel approach to atomic force microscopy. Appl. Plrys. Lett. 53, 1045–1047. 11. Baselt, D. R., Clark, S. M., Youngquist, M. G., Spence, C. F., and Baldeschwieler, mJ. D.

12.

13.

14.

15.

16.

(1993) Digital signal control of scanned probe microscopes. Rev. Sci. Instrum. 64, 1874–1882. I. Horcas, R. Fernandez, J.M. GomezRodriguez, J. Colchero, J. Gomez-Herrero, and A.M. Baro, Review of Scientific Instruments 78, 013705 (2007). Wade, T., Garst, J. F., and Stickney, J. L. (1999). A simple modification of a commercial atomic force microscopy liquid cell for in situ imaging in organic, reactive or air sensitive environments. Rev. Sci. Instr. 70, 121–124. Lehenkari, P. P., Charras, G. T., Nykanen, A., and Horton, M. A. (2000) Adapting atomic force microscopy for cell biology. Ultramicroscopy 82, 289–295. Workman, R. K. and Manne, S. (2000) Variable temperature fluid stage for atomic force microscopy. Rev. Sci. Instrum. 71, 431–436. Radmacher, M., Cleveland, J. P., and Hansma, P. K. (1995) Improvement of thermally induced bending of cantilevers used for atomic force microscopy. Scanning 17, 117–121.

Chapter 2 Measurement Methods in Atomic Force Microscopy Bruno Torre, Claudio Canale, Davide Ricci, and Pier Carlo Braga Abstract This chapter is introductory to the measurements: it explains different measurement techniques both for imaging and for force spectroscopy, on which most of the AFM experiments rely. It gives a general overview of the different techniques and of the output expected from the instrument; therefore it is, at a basic level, a good tool to properly start a new experiment. Concepts introduced in this chapter give the base for understanding the applications shown in the following chapters. Subheading 1 introduces the distinction between spectroscopy and imaging experiments and, within the last ones, between DC and AC mode. Subheading 2 is focused on DC mode (contact), explaining the topography and the lateral force channel. Subheading 3 introduces AC mode, both in noncontact and intermittent contact case. Phase imaging and force modulation are also discussed. Subheading 4 explains how the AFM can be used to measure local mechanical and adhesive properties of specimens by means of force spectroscopy technique. An overview on the state of the art and future trends in this field is also given. Key words: AFM imaging modes, Contact mode, Noncontact mode, Intermittent contact mode, Phase imaging, Force modulation, Force spectroscopy

1. Introduction Different kinds of measurements and advanced experiments can be grouped under two main categories: imaging and spectroscopy. In the first case, the tip is scanned over the surface to compose a topographic map of the sample and depending on the operation mode and on the parameters under control, the imaging mode can take different names (i.e., contact, noncontact, constant height, amplitude modulation, frequency modulation, etc.); in all these cases, data is organized to compose an image that is representative of surface morphology and/or related to some of its properties (e.g., composition).

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_2, © Springer Science+Business Media, LLC 2011

19

20

Torre et al.

In general, one can easily distinguish between two imaging modes depending on which tip–sample interaction is detected: usually, interaction can be detected by looking at static deflection of the cantilever (force measurements) or by forcing it into resonance and measuring the changes in its oscillation due to the presence of an interaction force. The first case is usually called static mode, or DC mode, because it records the static deflection of the cantilever, whereas the second takes a variety of names (some patented) among which we may point out the resonant or AC mode. In this case, the feedback loop tries to keep at a set value not the deflection but one of the oscillation parameters, usually the amplitude, of the cantilever while scanning the surface. To do this, more complicated electronics are necessary in the detection circuit, including a lock-in or a phase-locked loop amplifier, and also some actuator in the cantilever holder to induce the oscillatory excitation; anyway, almost all modern instruments have all these capabilities already implemented and software controllable as default so that inducing cantilever resonance and tuning measurements parameters can be easily done. From a physical point of view, one can make a distinction between the two imaging modes depending on the sign of the forces involved in the interaction between tip and sample, that is, by whether the forces there are purely repulsive or also account for attractive contributions (1). In Fig. 1, an idealized plot of the forces between tip and sample is shown, highlighting where typical imaging modes are operated.

intermittent

Fig. 1. Idealized plot of the forces between tip and sample, highlighting where typical imaging modes are operative.

Measurement Methods in Atomic Force Microscopy

21

An alternative method, the spectroscopy mode, is more related to the evaluation of the forces shown in Fig. 1 than to the reconstruct ion of image morphology. It consist of sweeping one of the measurement parameters S while the tip is not scanning the surface (point mode) and recording cantilever response R at the same time: in this way, a curve R = f (S) is obtained. A very wellestablished technique consists in monitoring cantilever deflection D (i.e., force, after spring constant calibration) while changing the tip–sample distance Z to bring the tip into contact with the sample, starting very far from the surface. This technique is referred as force spectroscopy, and D = f (Z) curves are called force–distance curves: data contain quantitative information on tip–sample interaction and on sample mechanical properties, such as elasticity and plasticity, therefore, this kind of spectroscopy is very popular. The spectroscopy mode is a general technique and other experiments can be performed to measure sample properties, such as conductivity, piezoelectric response, or dynamic mechanical response. Spectroscopy curves can be acquired on an N × M grid of points, obtaining a three-dimensional map in which each pixel is a spectroscopy curve containing information about the interaction. In the following chapter and subheadings, we briefly describe the DC and AC imaging modes and the basics of spectroscopy operation for relevant applications in the biomedical field.

2. DC Modes 2.1. Contact Mode

Also called constant force mode, the contact mode is the most direct AFM mode, where the tip is brought in contact with the surface and the cantilever deflection is kept constant during scanning by the feedback loop. Image contrast depends on the applied force which again depends on the cantilever spring constant (Fig. 2). Softer cantilevers are used for softer samples. It can be employed easily also in liquids, allowing to considerably reduce the capillary forces between tip and sample and, hence, limit the damage to the surface (Fig. 3; refs. 2, 3). Because the tip is permanently in contact with the surface while scanning, a considerable shear force can be generated, causing sample damage, especially on very soft specimens, such as biomolecules or living cells (4). Since share forces depend on load, recently very soft (spring constant 10–100 times smaller than usual contact-mode cantilever), small, low-noise silicon nitride cantilever has been specifically designed to work in contact mode on biological samples, particularly in liquid environment, allowing stable imaging: for these reasons, many authors prefer to work in contact mode even on very soft biological samples.

22

Torre et al.

Fig. 2. In contact mode, the tip follows directly the topography of the surface while it is scanned.

Fig. 3. In contact mode, capillary forces caused by a thin water layer can considerably increase the total force between sample and tip.

2.2. Deflection or Error Mode

In some cases, especially on rough and relatively rigid samples, the error signal (i.e., the difference between the set point and the effective deflection of the cantilever that occurs during scanning as a result of the finite time response of the feedback loop) is used to record images. By turning down on purpose the feedback gain, the cantilever will press harder on asperities and less on depressions, giving rise to images that contain high-frequency information otherwise not visible (5). This method has been extensively used to image submembrane features in living cells. The same method is also often used to record high-resolution images on crystals.

2.3. Lateral Force Microscopy

In this case (a variation of standard contact mode), not only the vertical deflection of the cantilever, but also the lateral deflection (torsion) is measured by the photodetector assembly, which in

Measurement Methods in Atomic Force Microscopy

23

Fig. 4. Using a four-section photodetector, it is possible to measure also the torsion of the cantilever during contact mode AFM scanning. The torsion of the cantilever reflects changes in the surface chemical composition.

this case has four photodiodes instead of two (Fig. 4). The degree of torsion of the cantilever supporting the probe is a relative measure of the surface friction caused by the lateral force exerted on the scanning probe (6). This method has been used to discriminate between areas of the sample that have the same height (i.e., that are on a same plane), but that present different frictional properties due to absorption of different chemical properties.

3. AC Modes All AC modes require setting the cantilever in oscillation using an additional driving signal. This can be accomplished by inducing oscillations in the cantilever with a piezoelectric motor (acoustic mode) or, as developed more recently, by directly driving using external coils a probe coated with a magnetic layer (magnetic mode). By using this second method interesting results have been obtained, especially in liquid, as it allows better control of the oscillation dynamics and has inherently less noise (7, 8). 3.1. Noncontact Mode

An oscillating probe is brought into proximity of (but without touching) the surface of the sample and senses the van der Waals attractive forces that induce a frequency shift in the resonant frequency of a stiff cantilever (Fig. 5; ref. 9). Images are taken by keeping a constant frequency shift during scanning, and usually this is performed by monitoring the amplitude of the cantilever oscillation at a fixed frequency and feeding the corresponding value to the feedback loop exactly as for the DC modes.

24

Torre et al.

Fig. 5. In the noncontact operation mode, a vibrating tip is brought near the sample surface, sensing the attractive forces. This induces a frequency shift in the resonance peak of the cantilever that is then used to operate the feedback.

The ­tip–sample interactions are very small in noncontact mode, and good vertical resolutions can be achieved, whereas lateral resolution is lower than in other operating modes. The greatest drawback is that it cannot be used in liquid environment, but only on dry samples. Also, even on dry samples, if a thick contamination or water layer is present, as the oscillation amplitude is small, the tip can sometimes get trapped as it does not have sufficient energy to detach from the sample. 3.2. Intermittent Contact Mode

The general scheme is similar to that of noncontact mode, but in this case, during oscillation, the tip is brought into contact with the sample surface causing a dampening of the oscillation amplitude by the same repulsive forces that are present in contact mode (Fig. 6). Usually, in the intermittent contact mode, the cantilever oscillation amplitude is larger than the one used for noncontact. There are several advantages that have made this mode of operation quite popular. The vertical resolution is very good together with the lateral resolution, there are less interactions with the sample compared to contact mode (especially lateral forces are greatly reduced), and it can be used in liquid environment (10–14). This mode of operation is the most generally used for imaging biological samples and is still under constant improvement, thanks to additional features, such as Q-control (12) or magnetically driven tips (7, 8).

Measurement Methods in Atomic Force Microscopy

25

Fig. 6. In intermittent contact mode, the free oscillation of a vibrating cantilever is dampened when the tip touches the sample surface at each cycle. The image is performed keeping constant the oscillation amplitude decrease while scanning.

3.3. Phase Imaging Mode

If the phase lag of the cantilever oscillation relative to the ­ riving signal is recorded in a second acquisition channel durd ing imaging in intermittent contact mode, noteworthy information on local properties that are not revealed by other AFM techniques (15), such as stiffness, viscosity, and adhesion, can be detected. In fact, it is good practice to always acquire both the amplitude and phase signals simultaneously during intermittent contact operation, as the physical information is entwined and all the data are necessary to interpret the images obtained (16–20).

3.4. Force Modulation

In this case, a low-frequency oscillation is induced (usually to the sample) and the corresponding cantilever deflection recorded while the tip is kept in contact with the sample (Fig.  6). The ­varying stiffness of surface features induces a corresponding dampening of the cantilever oscillation so that local relative viscoelastic properties can be imaged.

26

Torre et al.

4. Beyond Topography Using Force Curves

The AFM can provide much more information than simply taking images of the surface of the sample. The instrument can be used to record the amount of force felt by the cantilever as the probe tip is brought close to a sample surface, eventually indent the surface and then pulled away. By doing this, the long-range attractive or repulsive forces between the probe tip and the sample surface can be studied, local chemical and mechanical properties like adhesion and elasticity may be investigated, and even the bonding forces between molecules may be directly measured (21–23). By acquiring a series of force curves, one at each point of a square grid, it is possible to acquire the so-called force vs. volume map that allows the user to compute images representing local mechanical properties of the sample observed. Force curves typically show the deflection of the cantilever, as the probe is brought vertically toward and then away from the sample surface using the vertical motion of the scanner driven by a triangular wave (Fig. 7). By controlling the amplitude and frequency of the vertical movement of the scanner, it is possible to change the distance and speed that the AFM probe travels during the force measurement. Conceptually, what happens during a

Fig. 7. From positions A to B, the tip is approaching the surface, and at position B contact is made (if an attractive or repulsive force is active before contact, the portion of the force curve will reflect it). After position B, the cantilever bends until it reaches the specified force limit that is to be applied (S). Depending on the relative stiffness of the cantilever with respect to the sample, during this portion of the curve the tip can indent the surface. The tip is then withdrawn toward positions C and D. At position D under application of the retraction force, the tip detaches from the sample (often referred to as “snap off”). Between positions D and A, the cantilever returns to its resting position and is ready for another measurement.

Measurement Methods in Atomic Force Microscopy

27

force curve is not much different from what happens between tip and sample during intermittent contact imaging. The differences are in the frequency used, much lower for force curves, and the probe, much smaller in intermittent contact. In a force curve, many data points are acquired during the motion so that very small forces can be detected and interpreted by fitting the force curve according to theoretical models. In order to obtain quantitative data from force vs. distance curves, two technical details need special care. The positionsensitive photodetector signal has to be calibrated so to measure accurately the cantilever deflection, and after calibration it is essential that the laser alignment is left unchanged. Usually, the AFM software has a routine for such calibration, performed by taking a force curve on a hard sample and using the scanner’s vertical movement as reference (which means that the scanner also has to be accurately calibrated). At this point, the curve we are plotting is not yet a force curve but a calibrated deflection curve. The next step is to convert it to a force curve using the force constant of the cantilever we are using. Manufacturers usually specify this value, but for each cantilever there can be quite large variations so that for accurate work direct determination becomes necessary. There are different ways to measure the force constant, some requiring external equipment for measuring resonant frequency (such as spectrum analyzers) and others making use of reference cantilevers (24, 25). From the point of view of biomedical applications, interesting experiments can be performed by coating the tip with a ligand and approaching through a force curve a surface where receptor molecules can be found. In this case, the portion of the curve before snap-off has a different shape, reflecting the elongation of the bond between ligand and receptor before dissociation: from the shape the curve, it is possible to derive quantitative information on the binding forces (26–28). If a force curve is taken at each point of an N × N grid, it is possible derive images that are directly correlated to a physical property of the surface of the sample. For example, if the approach portion of each curve after contact is fitted using indentation theory, a map of the sample stiffness can be calculated. This data can be represented by an image in which the level of gray of each pixel, instead of representing the height of the sample, corresponds to the elasticity modulus. Similar images can be calculated for adhesion, binding, electrostatic forces, and so forth (29, 30). 4.1. State of the Art and Future Challenges: Dynamic Spectroscopy

If the same operation is done while dithering the cantilever close to its resonance frequency, tip–sample interaction is probed in dynamic mode (dynamic force spectroscopy), and several para­ meters can be measured as a function of distance (such as static deflection, amplitude, phase, higher harmonics, frequency, etc.)

28

Torre et al.

containing a larger amount of data: valuable information about local interactions can be extracted or reconstructed, revealing material properties such as short- and long-range forces (31), friction (32), plasticity (33), chemical composition (34), and so on. Quantitative interpretation of all the interactions at the base of spectroscopy data is still under development and involve the inversion of dynamic parameters to reconstruct interaction forces: historically, they have been first inverted to reconstruct tip– sample interaction forces in the case of FM–AFM by Durig using Hamilton–Jacobi perturbation theory in the large amplitude oscillation – or short range forces – case (35), and then generalized by Sader et  al. (36, 37) considering first resonance. Durig also investigated dynamic behavior by considering amplitude and phase of higher harmonics, using the Chebyshev polynomial expansion method (38). Under particular oscillating regimes, also subharmonic and chaotic cantilever dynamics, fingerprints of tip–sample interactions have been found (39). Each of these advanced spectroscopy methods imply a huge amount of data that require very high computational power to reconstruct physically valuable parameters from comparison with contact models (40); as a result, a fast and easy analysis relying on these dynamic methods is still far to be routinely implemented for spectroscopy maps or it is limited to a subset of information. References 1. Israelachvili, J. N. (1985) Intermolecular and Surface Forces. Academic Press, London. 2. Weisenhorn A. L., Maivald, P., Butt, H. J., and Hansma, P. K. (1992) Measuring adhesion, attraction, and repulsion between surfaces in liquids with an atomic-force microscope. Phys. Rev. B. 45, 11,226–11,232. 3. Weisenhorn A.L., Hansma, P. K., Albrecht T. R., and Quate, C. F. (1989) Forces in atomic force microscopy in air and water. Appl. Phys. Lett. 54, 2651–2653. 4. Butt, H.-J., Siedle, P., Seifert, K., et al. (1993) Scan speed limit in atomic force microscopy. J. Microsc. 169, 75–84. 5. Putman, C. A., van der Werf, K. O., de Grooth, B. G., van Hulst, N. F., and Greve, J. (1992) New imaging mode in atomic-force microscopy based on the error signal. SPIE Proceedings 1639, 198–204. 6. Gibson, C. T., Watson, G. S., and Myhra, S. (1997) Lateral force microscopy–a quantitative approach. Wear 213, 72–79. 7. Han, W. and Lindsay, S. M. (1998) Precision interfacial molecular force measurements with

8.

9. 10.

11.

12.

13.

a MAC mode atomic force microscope. Appl. Phys. Lett. 72, 1656–1658. Han, W., Lindsay, S. M., and Jing, T. (1996) A magnetically-driven oscillating probe microscope for operation in liquids. Appl. Phys. Lett. 69, 4111–4113. Garcia, R. and San Paulo, A. (2000) Amplitude curves and operating regimes in dynamic atomic force microscopy. Ultramicroscopy 82, 79–83. Hansma, P. K., Cleveland, J. P., Radmacher, M., et al. (1994) Tapping mode atomic force microscopy in liquids. Appl. Phys. Lett. 64, 1738–1740. Lantz, M., Liu, Y. Z., Cui, X. D., Tokumoto, H., and Lindsay, S. M. (1999) Dynamic force microscopy in fluid. Surface Interface Anal. 27, 354–360. Tamayo, J., Humphris, A. D., Owen, R. J., and Miles, M. J. (2001) High-Q dynamic force microscopy in liquid and its application to living cells. Biophys. J. 81, 526–537. Burnham, N. A., Behrend, O. P., Oulevey, F., et  al. (1997) How does a tip tap? Nano­ technology 8, 67–75.

Measurement Methods in Atomic Force Microscopy 14. Behrend, O. P., Oulevey, F., Gourdon, D., et al. (1998) Intermittent contact: Tapping or hammering? Appl. Phys. A66, S219–S221. 15. Magonov, S. N., Elings, V., and Whangbo, M.-H. (1997) Phase imaging and stiffness in tapping mode AFM. Surface Sci. 375, L385–L391. 16. Bar, G., Delineau, L., Brandsch, R., Bruch, M., and Whangbo, M.-H. (1999) Importance of the indentation depth in tapping-mode atomic force microscopy study of compliant materials. Appl. Phys. Lett. 75, 4198–4200. 17. Bar, G. and Brandsch, R. (1998) Effect of viscoelastic properties of polymers on the phase shift in tapping mode atomic force microscopy. Langmuir. 14, 7343–7347. 18. Cleveland, J. P., Anczykowski, B., Schmid, A. E., and Elings, V. B. (1998) Energy dissipation in tappingmode atomic force microscopy. Appl. Phys. Lett. 72, 2613–2615. 19. Chen, X., Davies, M. C., Roberts, C. J., Tendler, S. J. B., and Williams, P. M. (2000) Optimizing phase imaging via dynamic force curves. Surface Sci 460, 292–300. 20. Pang, G. K., Baba-Kishi, K. Z., and Patel, A. (2000) Topographic and phase-contrast imaging in atomic force microscopy. Ultramicroscopy 81(2), 35–40. 21. Butt, H-J. (1991) Measuring electrostatic, van der Waals, and hydration forces in electrolyte solutions with an atomic force microscope. Biophys. J. 60, 1438–1444. 22. Vinckier, A. and Semenza, G. (1998) Measuring elasticity of biological materials by atomic force microscopy. FEBS Lett. 430, 12–16. 23. Hutter Jeffrey L. and John Bechhoefer (1994) Measurement and manipulation of Van der Waals forces in atomic force microscopy. J. Vacuum Sci. Technol. B, 12, 2251–2253. 24. Cleveland, J. P., Manne, S., Bocek, D., and Hansma, P. K. (1993) A non-destructive method for determining the spring constant of cantilevers for scanning force microscopy. Rev. Sci. Instrum. 64, 403–405. 25. D’Costa, N. P. and Hoh, J. H. (1995) Calibration of optical lever sensitivity for atomic force microscopy. Rev. Sci. Instrum. 66, 5096–5097. 26. Hoh, I., Cleveland, J. P., Prater, C. B., Revel, J.-P., and Hansma, P. K. (1992) Quantized adhesion detected with the atomic force microscope. J. Am. Chem. Soc. 4917–4918. 27. Mckendry, R. A., Theoclitou, M., Rayment, T., and Abell, C. (1998) Chiral discrimination by chemical force microscopy. Nature 14, 2846–2849.

29

28. Okabe, Y., Furugori, M., Tani, Y., Akiba, U., and Fujihira, M. (2000) Chemical force microscopy of microcontact-printed selfassembled monolayers by pulsedforce-mode atomic force microscopy. Ultramicroscopy 82, 203–212. 29. Willemsen, O. H., Snel, M. M., van Noort, S. J., et  al. (1999) Optimization of adhesion mode atomic force microscopy resolves individual molecules in topography and adhesion. Ultramicroscopy 80, 133–144. 30. Thundat, T., Oden, P. I., and Warmack, R. J. (1997) Chemical, physical, and biological detection using microcantilevers. Electrochem. Society Proc. 97, 179–187. 31. F.J. Giessibl.1997.Forces and frequency shifts in atomic-resolution dynamic force microscopy” Phys. Rev. B 56: 16010–16015. 32. Holscher H, Schwarz U D, Zworner O and Wiesendanger R.1998.Consequences of the stick-slip movement for the scanning force microscopy imaging of graphite. Phys. Rev. B 57: 2477–81. 33. Butt HJ, Cappella B and Kappl M. 2005. Force measurements with the atomic force microscope: Technique, interpretation and applications. Surf. Sci. Rep. 59:1–152. 34. Magonov S  N, Elings V and Papkov V  S. 1997. AFM study of thermotropic structural transitions in poly(diethylsiloxane). Polymer 38: 297–307. 35. Durig U. 2000. Extracting interaction forces and complementary observables in dynamic probe microscopy. Appl. Phys. Lett. 76: 1203–1205. 36. Sader J E and Jarvis S P.2004. Accurate formulas for interaction force and energy in frequency modulation force spectroscopy. Appl. Phys. Lett. 84:1801–1803. 37. Sader J E, Uchihashi T, Higgins MJ, Farrell A, Nakayama Y and Jarvis S P. 2005. Quantitative force measurements using frequency modulation atomic force microscopy—theoretical foundations. Nanotechnology 16:S94–S101. 38. Durig U. 1999. Relations between interaction force and frequency shift in large-amplitude dynamic force microscopy. Appl. Phys. Lett. 75:433–435. 39. F Jamitzky, M Stark, W Bunk, WM Heckl and R W Stark, 2006. Chaos in dynamic atomic force microscopy.Nanotechnology 17: S213–S220. 40. W. N. Unertl “Implications of contact mechanics models for mechanical properties measurements using scanning force microscopy”, J. Vac. Sci. Technol. A 17, 1779 (Jul/ Aug 1999).

wwwwwwwwwwwwww

Chapter 3 Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging Claudio Canale, Bruno Torre, Davide Ricci, and Pier Carlo Braga Abstract Atomic force microscopy (AFM) measurements could be affected by different kinds of artifacts; some of them derive from the improper use of the instrument and can be avoided by setting the correct experimental parameters and conditions. In other cases, distortions of the images acquired by AFM are intrinsically related to the operating principle of the instrument itself and to the kind of interactions taken into account for the reconstruction of the sample topography. A perfect knowledge of all the artifacts that can perturb AFM measurements is fundamental to avoid misleading interpretations of the results. In this chapter, all the most common sources of artifact are presented, and strategies to avoid them are proposed. Subheading 1 is a brief introduction to the chapter. In Subheading 2, the artifacts due to the interactions between the sample and the AFM tip are presented. Subheading 3 is focused on the deformations due to the AFM scanner nonlinear movements. The interaction with the environment surrounding the instrument can affect the quality of the AFM results and the environmental instability are discussed in Subheading  4. Subheading  5 shows the effects of an incorrect setting of the feedback gains or other parameters. Subheading 6 aims on the artifacts that can be produced by the improper use of the image processing software. Subheading 7 is a short guide on the test that can be done to easily recognize some of the artifacts previously described. Key words: AFM, Tip artifacts, Nonlinearity, Instability, Creep

1. Introduction Images and other information obtained by using atomic force microscopy (AFM) are derived from the physical interaction between the AFM probe and the sample. The different working principles of the SPMs with respect to the conventional microscopes are responsible for a new series of artifacts that affect

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_3, © Springer Science+Business Media, LLC 2011

31

32

Canale et al.

images and are not easily recognizable by inexpert users. Since we are addressing novices in this field, we would like to give an idea of what can happen while taking images with the AFM, how one can recognize the source of the artifact, and then try to avoid it or minimize it. Sources of artifacts in AFM images are essentially the tip, the scanner, the environment, the control system electronics, and the image-processing software.

2. Tip Artifacts The AFM tip plays a fundamental role in the generation of the AFM image: it explores the sample surface while the cantilever bends under the action of the complex force field established between the sample and the tip itself. The geometrical shape of the tip always affects the AFM images acquired using it. The images result as the convolution between the sample and the tip shape; intuitively, as long as the tip is sharper than the feature under observation, the profile resembles closely the true shape of the sample. The choice of the optimal probe is important to minimize the artifacts due to tips: the smaller the size of the object, the sharper the tip. A notable exception arises in the case of high-resolution imaging on ordered crystals, where often better images are obtained with standard tips. This can be explained by realizing that at this dimensional scale the measurable radius of curvature of the tip is not in fact involved in the imaging process, but instead smaller local protrusions on the apex of the probe perform as the real tip (or tips) effectively taking the image. Further details on AFM tip properties and related artifacts can be gathered from the vast literature on the subject, together with a variety of methods for their correction (1–9). Specific artifacts, depending on the mode of operation, have been investigated and explanations have been proposed (10–14). Since we are now interested in showing a general overview of the subject for beginners in the field, we shall have a look at the main tip artifacts in a very simple way. 2.1. Tip Broadening Effect on Protruding Features

Different profiles were obtained using a dull or a sharp tip (Fig. 1). Depending on the lateral size and height of the feature to be imaged, both the sharpness of the apex and the sidewall angle of the tip become important. In general, when scanning rough samples, the tall features are displayed as a mirror image of the tip sidewall (Fig. 2). In addition to sharpness, the geometrical shape of the tip also is important, as the level of broadening in a particular direction depends on the tip geometrical symmetry. In particular, a conical tip affect the lateral size of the sample features

Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging

33

Fig. 1. Line profiles obtained using two tips with different aspect ratio. The shape of the object is better approximated using tip with a sharper profile. In spite of this, tip broadening ever affects AFM images.

Fig. 2. Three-dimensional view obtained from an AFM image showing part of a neuron. In particular, the soma (the taller part of the sample) is strongly affected by the tip and it appears as a mirror image of the tip sidewall itself, while the neuritis structure, although broadened by the tip, is clearly displayed.

symmetrically in all the directions while the level of broadening due to a pyramidal tip is dependent on the scan angle; the distance between two faces of the pyramid is significantly smaller with respect to the distance between two opposite edges. Very small features, such as nanoparticles, nanotubes, proteins, and DNA strands, ideally interact only with the tip apex; therefore, the images result as the convolution between the sample features and the hemisphere approximating the apex of the tip. Due to tip image broadening, the measured lateral size should be taken as an upper limit for the true size of the objects imaged by AFM. Note that in all these cases, the measured height of the sample is reported accurately.

34

Canale et al.

Fig. 3. Different artifacts arise from the use of deformed tips. Single amyloid fibrils are displayed as two separate adjacent structures due to the use of a “double tip” (a). Data scale 3 × 3 mm2; Z-range 12 nm. The asymmetrical geometry of a contaminated tip affected the shape of globular protein aggregates: all the features on the mica substrate displayed a similar elongated shape (b). Data scale 2 × 2 mm2, Z-range 13 nm.

2.2. Tip-Induced Deformations and “Double Image”

Multiple protrusions at the AFM tip apex can be present as a result of damage or contamination. Due to the interaction of the multiple tip apexes with the sample features, repetitive patterns may appear in an image (Fig. 3). Images affected by this artifact are often called “double image”; actually, sample particles can be replicated several times in the AFM image, depending on the number of apical protrusion interacting with the sample surface. Furthermore, spherical nanoparticles or small molecules may assume an elongated or triangular shape, reflecting an asymmetrical geometry of the apex of the tip (Fig. 3).

2.3. Flattening of Pits and Holes

The finite size of the tip has an effect also in the visualization of features that are below the surface mean level, such as a hole. The lateral size of small holes at the sample surface is underestimated. Furthermore, the tip may not be able to reach the bottom of a hole, resulting on a lack of physical depth in the AFM image.

3. Scanner Artifacts AFM scanners are made of piezoelectric ceramic, a material that undergoes a change of its shape under the effect of an applied voltage. Piezoelectric scanners can provide subnanometric positioning of a probe, and they have been one of the breakthroughs that made AFM possible. In spite of this, a number of artifacts arise from their physical and mechanical properties, even though their design has been constantly improved and some of the

Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging

35

a­ rtifacts have been removed, or at least minimized, in the newest instruments. One point that must not be neglected is that scanner properties change with time and use. In fact, the piezoelectric material changes its sensitivity to driving signals. If it is used often, it will become slightly more sensitive; if left idle, it will depolarize and become less sensitive. The best thing to do is to periodically calibrate the scanner according the manufacturer’s instructions. 3.1. Effects of Intrinsic Nonlinearity

Piezoelectric scanners are inherently nonlinear: if the extension of the scanner in any one direction is plotted as a function of the driving signal, the plot will not be a straight line but it will appear a curve similar to the one shown in Fig. 4. The nonlinear relationship between the applied voltage and the displacement of a piezoactuator contributes to positioning error (15). Nonlinear effects are more pronounced for large scans while they can be neglected for small scans. In this case, we refer to large scans when they are more than 70% of the full scale displacement of the piezoelectric scanner. The nonlinearity may be expressed as a percentage (describing the deviation from linear behavior), and it typically ranges from 2 to 25%, depending on the driving signal applied and the scanner construction. The effects are present both in the plane and in the vertical directions. An AFM image of a calibration grid with periodic structures, such as squares, appears severely distorted, with inhomogeneous spacing and anomalous curvature of features, typically appearing smaller on one side of the image than on the other (Fig.  5).

Fig. 4. Plot of the scanner extension vs. driving signal. Notice the large deviation from linearity.

36

Canale et al.

a

b

Fig. 5. Distortion of a test pattern caused by scanner nonlinearity. The square structures of a calibration grid (a) as may appear in a AFM image acquired using a noncalibrated scanner (b).

On a generic nonregular sample, such distortions may be hardly recognizable, but they will certainly be present. Once the scanner is properly linearized, the scanner calibration is also critical. It is possible for a scanner to be linear but not calibrated: in this case, the x and y distances measured from line profiles are incorrect. 3.1.1. In Plane Linearization

There are essentially two methods to linearize a scanner in the x and y directions: by software or by hardware. Software correction is performed by mathematically modeling the nonlinear behavior of the scanner: it is necessary to perform an image of a known calibration grid, to find the parameters for a correction algorithm, and then to apply this algorithm while scanning new samples. The limit of this method lies in the fact that unfortunately the correction strongly depends upon the scan speed, scan direction, and offset that have been used during the calibration procedure. When images are taken under conditions that are similar to the conditions used during the calibration, the correction is accurate; otherwise, nonlinearities are again present. More recently, hardware correction of AFM scanner nonlinearity has become popular (15). The true position of the scanner in the x and y directions is measured by a sensor during scanning and compared with the intended scanner position. A feedback circuit applies an appropriate driving signal to the scanner in order to attain the desired position (closed-loop feedback). Using closed-loop feedback systems, the deviation from the linear behavior is strongly reduced and can be generally neglected.

3.1.2. Out of Plane Linearization

Because the height range of scanners along the vertical direction is usually an order of magnitude smaller than the range in the scanning plane, effects of nonlinearity are less severe but still pre­ sent. Also, in this case, the linearization and calibration of the scanner can be obtained imaging a sample with well-defined step height. Often, the microscope is calibrated with only one reference step. This implies that the height measurements are not correct unless the feature being observed has a height close to the

Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging

37

Fig. 6. The out-of-plane movement of an AFM scanner is often calibrated in only one point, i.e., using a single step-height calibration standard. The plot represents the deviation from the true height value obtained measuring step heights that differ from the height at the calibration point.

one used in the calibration measurement (Fig. 6). It is also to be noted that although calibration gratings are reasonably easy to make by lithographic techniques, step–height calibration standards are more difficult to obtain, especially for very high-resolution work. Crystals with known height steps are often used for an accurate calibration on the nanometer/subnanometer range of vertical displacements. In the newest instruments, the use of closed-loop feedback strongly reduces the nonlinear effects also along the out of plane direction. 3.2. Effects of Hysteresis

All piezoelectric ceramics display hysteretic behavior; while scanning back and forth between two points, the same driving signal does not correspond to the same position in the two scanning directions. This can be easily observed by comparing the profiles taken from left to right and in the opposite direction on a feature on the surface of a sample. The result would be like in Fig.  7, where there is a lateral shift between the two profiles. Hysteresis is more pronounced over long-range operations while the effect can be neglected into small scans. Notice that hysteresis is also present in the vertical direction, giving rise to asymmetric step heights.

3.3. Effects of Creep

When a driving voltage is applied to an AFM scanner, the response of the piezo-actuator occurs in two steps. The first step takes place in less than a millisecond, providing approximately 95% of the full extent of the desired movement, and then the piezo-actuator slowly provides the remaining 5% on a timescale of several minutes.

38

Canale et al.

Fig. 7. Due to hysteresis, the profiles of a step obtained scanning in opposite directions (trace and retrace) appear laterally translated.

Fig. 8. Effect of creep on a scan performed zooming up onto a detail in a large image. This effect is often called drift, but must not be confused with thermal drift.

Fig. 9. Vertical overshooting at the edges of a step are due to the scanner creep in the out-of-plane direction.

This second slower movement is called creep. Due to creep, scans taken at different scan rates have slightly different magnification. Creep is particularly evident when a large variation of the drive voltage occurs, for example, when a zoom on a small area is performed just after the acquisition of a large scan area (Fig. 8). In the vertical direction, creep becomes apparent as an overshoot of the scanner position at the leading and trailing edge of features that have steep sides (Fig. 9). This can be often found as a lateral “shading” of protruding features on flat substrates in top view topographical images.

Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging

3.4. Effects of Cross Coupling Sample Tilting

39

Although alternative configurations have become available in the last years, several AFM instruments still use piezoelectric tube scanners to provide the movement along the x, y, and z axis. In this case, the scanners are assembled in the AFM having a free scanning end (to which either the cantilever or the sample is attached) while the other end is attached to the microscope body. For this reason, the motion of the scanner follows an arc (spherical or parabolic, depending on the type of scanner) and not a plane (Fig. 10). The affected images show a bow, which is especially evident in large scans. This artifact can easily be subtracted by image processing. Bow effect precludes the visualization of small features distributed on flat surfaces. AFMs often use tools to subtract the appropriate curve from each line during acquisition, allowing small features to become immediately evident during scan. Mechanical or electronic cross coupling between the x and y direction elements of the scanner can be present; in this case, the angles between features in the x and y plane are modified, and this becomes apparent in the image of test structures. Mechanical coupling between the piezoelectric ceramics that provide probe displacement along the different directions can cause substantial errors when measuring sidewall angles. Another source of cross coupling arises when the scan direction is not parallel to one of the piezoelectric elements that constitute the scanner. Rotated scans are obtained by sending appropriately mixed driving signals to both the x and y piezoelectric elements: if they both are not accurately calibrated, the image will be affected by a geometrical distortion.

x,y,z scanner

Fig. 10. Typical bow artifacts are due to the working mode of piezo-tube scanners. This effect is more evident on large scans of flat surfaces and it can be removed by image processing.

40

Canale et al.

It is useful to add that quite often (in fact, always) the sample has a plane tilt relative to the motion of the scanner. Although all acquisition softwares allow for subtracting the tilt during scanning, it is good practice to try and mount the sample as planar as possible so that the piezoelectric element responsible for the vertical movement operates across a smaller range, hence behaving linearly.

4. Environmental Effects All the devices that provide high-resolution measurements require a stable environment to provide optimal results. The AFM operates – thanks to its very high sensitivity to the small deflections of its cantilever; it is evident that if external vibrations affect the cantilever bending, these will create artifacts in the images. The two main sources of instability are the mechanical vibrations and the thermal drift. 4.1. Mechanical Vibrations

Low-frequency mechanical vibrations coming from the building and the acoustic vibrations propagating through air or from the floor are important sources of artifacts in AFM images. In particular, the mechanical vibrations, coming from the floor, can have an amplitude of several micrometers at frequencies below 15  Hz. The floor vibrations, if not properly filtered, can cause periodic artifact in an image. This type of artifact is most often noticed when imaging very flat samples. Sometimes, the vibrations can be started by an external event, such as an elevator in motion, a train going by, or even people walking in a hallway. Several devices to minimize the effect of vibrations are available: special air tables, bungee cords, and active vibration isolation platforms can be used to shield the AFM from these vibrations. If possible, it is better to install the instrument near a corner instead of at the center of a room, choosing the lowest floor in the building. Is it a good choice to use an acoustic hood or enclosure to isolate the AFM from external acoustic noise; several kinds are commercially available.

4.2. Thermal Drift

The AFM and its scanner are affected by external temperature changes or gradients. AFM manufacturers try to minimize this phenomenon by choosing materials and appropriate design, but nevertheless thermal drift can be present. In the case of very high-resolution imaging of atomic structures, it is often necessary to wait some time during scanning before the system stabilizes and stops drifting. Also, electronics and the laser spot on the cantilever can induce drift in measurement settings that need to stabilize in time. In the case of AFMs mounted onto inverted microscopes and eventually equipped with a heated stage, special care has to be taken. A good environmental conditioning system can be useful in reducing thermal effects.

Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging

5. Effects of Feedback and Other Parameter Settings

41

Depending on the mode of operation, several parameters have to be set by the user to obtain the best images. Among these, one can find deflection set point (in contact mode), oscillation amplitude and dampening (in AC modes), feedback gain (sometimes separated into a proportional gain setting and integral–derivative setting), low-pass filters, scan speed, and so on. The setting of these parameters is a trial-and-error process; each time a new sample is scanned, the best values must be searched. In spite of this, some general rules can help in finding the best working parameters. Soft samples generally must be imaged using low scan speeds and low interaction forces; otherwise, glitches in the scan direction or even sample deformation may occur. Rough samples again need to be imaged slowly, but larger amplitude or deflection might be needed to keep track of the surface. Special care must be taken in tuning the gain parameters of the feedback. If the feedback loop of a scanning probe microscope is not optimized, the image can be affected. When feedback gains are too low, the tip cannot track the surface, and features are distorted and smeared out. However, when feedback gains are too high, the system can oscillate, generating highfrequency periodic noise in the image. This may occur throughout the image or be localized to features with steep slopes. A good rule is to increase the gain until periodic noise appears in the image, and then image with a setting just a little below such value.

6. Image Processing Image processing is readily available in AFM as the data are stored digitally on a computer disk. One can easily access routines for flattening, polynomial line or surface subtraction, removal of bad data, matrix filtering, and three-dimensional representation with sophisticated rendering. Often some kind of processing is necessary to analyze data and compare them with other results, but care must be taken to avoid introducing artifacts. For example, as we have seen in Subheading 3.3, nearly all images are affected by a tilt and by a bow introduced by the scanner geometry. If the wrong curve fit is applied or if large features are not excluded from the surface subtraction parameter computation (all image analysis software allow the exclusion of surface area portions from the computation), distortions will be introduced. Low-pass filters, although capable of reducing noise in the data, introduce smoothing of sharp features and, in the worst cases, delete smaller details. Fourier transform and power spectrum filtering if misused can create periodic features that may look like topographical structures, whereas in reality they are mere noise.

42

Canale et al.

The number of commercially available or free software to process AFM data has remarkably increased. Modern software are generally provided with a user-friendly graphical interface, so that advanced algorithms for image treatment and physical parameter calculation are now widely used in all AFM laboratories. On the other hand, the off-line elaboration of the AFM data can induce substantial modification or deletion of physical data that are present in the raw data. Thus, it is fundamental to read with care the instruction manual of the image-processing software before using them in data elaboration.

7. Some Guidelines for Artifact Testing If during a measurement you get suspicious that an image may contain artifacts, here are some things you can do to be sure whether or not they are really present: ●●

●●

●●

●●

●●

Take more than one image of the same area or the same line to ensure that it looks the same. When looking at a single scan line profile during acquisition, look if the traces are identical and stable in time. Try changing the scan direction and take a new image. You can do this also on a single scan line looking at the profile and observing directly the difference between the trace and retrace plots. Change the scan size and take an image to ensure that the features scale properly. Rotate the sample and take an image to identify artifacts induced by the shape of the tip. Change the scan speed and take another image (especially when suspicious periodic or quasiperiodic features are present). If they scale, you are looking at periodical noise.

References 1. Keller, D., and Chih-Chung, C. (1991) Reconstruction of STM and AFM images distorted by finite-size tips. Surface Sci. 253, 353–364. 2. Hellemans, L., Waeyaert, K., Hennau, F., Stockman, L., Heyvaert, I., and Van Haesendonck, C. (1991) Can atomic force microscopy tips be inspected by atomic force microscopy? J. Vac. Sci. Technol. B. 9, 1309–1312. 3. Keller, D. and Chou, C. C. (1992) Imaging steep, high structures by scanning force microscopy with electron beam deposited tips. Surface Sci. 268, 333–339.

4. Keller, D., Deputy, D., Alduino, A., and Luo, K. (1992) Sharp, vertical-walled tips for SFM imaging of steep or soft samples. Ultramicroscopy 42–44, 1481–1489. 5. Wang, W. L. and Whitehouse, D. J. (1995) Application of neural networks to the reconstitution of scanning probe microscope images distorted by finite-size tips. Nanotechnology 6, 45–51. 6. Markiewicz, P. and Goh, M. C. (1995). Atomic force microscope tip deconvolution using calibration arrays. Rev. Sci. Instrum. 66, 1–4. 7. Villarrubia, J. S. (1996) Scanned probe microscope tip characterization without cantilever

Recognizing and Avoiding Artifacts in Atomic Force Microscopy Imaging

8. 9.

10.

11.

tip characterizers. J. Vac. Sci. Technol. B. 14, 1518–1521. Sheng, S., Czajkowsky, D. M., and Shao, Z. (1999) AFM tips: How sharp are they? J. Microsc. 196, 1–5. Taatjes, D. J., Quinn, A. S., Lewis, M. R., and Bovill, E. G. (1999) Quality assessment of atomic force microscopy probes by scanning electron microscopy: Correlation of tip structure with rendered images. Microsc. Res. Tech. 44, 312–326. Dinte, B. P., Watson, G. S., Dobson, J. F., and Myhra, S. (1996) Artefacts in noncontact mode force microscopy: The role of adsorbed moisture. Ultramicroscopy 63, 115–124. Yang, J., Mou, J., Yuan, J.-Y., and Shao, Z. (1996) The effect of deformation on the lateral

12.

13.

14. 15.

43

resolution of the atomic force microscopy. J. Microsc. 182, 106–113. van Noort, S. J., van der Werf, K. O., de Grooth, B. G., van Hulst, N. F., and Greve, J. (1997) Height anomalies in tapping mode atomic force microscopy in air caused by adhesion. Ultramicroscopy 69, 117–127. Kühle, A., Sorenson, A. H., Zandbergen, J. B., and Bohr, J. (1998) Contrast artifacts in tapping tip atomic force microscopy. Appl. Phys. A. 66, S329–S332. Paredes, J. I., Martinez-Alonso, A., and Tascon, J. M. (2000) Adhesion artefacts in atomic force microscopy imaging. J. Microsc. 200, 109–113. Cao, H., and Evans A.G. (1993). Nonlinear deformation of ferroelectric ceramics. J. Amer. Ceram. Soc. 76, 890–896.

wwwwwwwwwwwwww

Part II Molecule Imaging

wwwwwwwwwwwwww

Chapter 4 Imaging the Spatial Orientation of Subunits Within Membrane Receptors by Atomic Force Microscopy Stewart M. Carnally, J. Michael Edwardson, and Nelson P. Barrera Abstract Our experimental approach is based on the atomic force microscope (AFM) imaging of epitope-tagged subunits within membrane protein complexes purified in small amounts and decorated by anti-tag antibodies. Furthermore, we can produce simultaneous decoration of protein complexes using Fab fragments and IgG antibodies, which, combined with chemical modification of the substrate, allows us to determine the protein orientation across the cell membrane. Here, we describe a detailed protocol for membrane protein purification, AFM data collection, analysis, and interpretation of results. The protocol also covers basic AFM instrument settings and best practices for both observation of membrane protein complexes by AFM and automatic detection of the structures by an in-house algorithm. Once a sufficient number of membrane protein complexes have been visualized by AFM, data acquisition and processing can be completed in approximately 10 min using a scanning surface of 1 mm2. Key words: AFM, Molecular architecture, Membrane receptor, Structural biology, Subunit stoichiometry

1. Introduction Structural studies of membrane proteins have gained the interest of many research groups. Well-known methods such as X-ray crystallography and cryoelectron microscopy have provided valuable information about the molecular architecture of this family of proteins. However, solubilization conditions, heterogeneity, and size of the protein complexes and different subunit compositions frequently hinder their structural characterization (1). As a consequence, new experimental methods are required to tackle this exciting field. Atomic force microscope (AFM) imaging of

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_4, © Springer Science+Business Media, LLC 2011

47

48

Carnally, Edwardson, and Barrera

Fig.  1. Scheme showing the molecular architecture of ionotropic receptors and ion ­channels formed by identical (homomultimers) or different subunits (heteromultimers), identified by AFM imaging of membrane proteins decorated with anti-subunit antibodies.

membrane proteins at single-molecule level has provided valuable insights into receptor–ligand interaction force and structural analysis of subunits and domains (2, 3). However, many receptors expressed at the plasma membrane are functional only as heteromeric entities. In the case of ionotropic receptors in particular, the subunit orientation is key to understanding parameters such as ligand binding, stoichiometry, and ion conductance (4, 5). Furthermore, when subunits within a receptor have similar size and shape, AFM imaging alone cannot detect the subunit spatial orientation. Recently, we have designed a novel approach combining AFM imaging and expression of engineered receptor subunits. This method has led us to propose the molecular architecture of a number of ionotropic receptors and channels that have physiological and therapeutic relevance (Fig. 1) (6–12). This chapter focuses on the isolation and AFM imaging of intact membrane receptors. This AFM experimental approach is based on the imaging of epitope-tagged subunits within membrane receptors purified in small amounts and decorated by antitag antibodies. To reduce possible damage of the isolated proteins by frictional forces, the AFM tapping mode is selected to measure the topography of the samples. An in-house algorithm automatically identifies receptor–antibody complexes and analyzes the molecular volume of their components and the angles of doubly decorated receptor–antibody complexes. This information is then interpreted, in order to deduce the stoichiometry and spatial orientation of subunits within a receptor (Fig.  2). This chapter

Imaging the Spatial Orientation of Subunits Within Membrane Receptors

49

Fig. 2. Workflow of the AFM method to determine molecular architecture of receptors. The whole procedure consists of nine steps, as indicated in the illustration. Steps 1–3 are associated with the purification process of the receptors. Step 4 represents the AFM imaging of individual purified receptors in an adequate density for analysis. Usually, two controls are carried out for each experiment: (a) a mock transfection, and subsequent AFM imaging of the same elution fraction; (b) a co-incubation of receptors with nonspecific antibodies. Analysis of the AFM images is done automatically using an inhouse Matlab algorithm (step 5 ). The same imaging procedure is performed with commercial anti-IgG antibodies (steps 6 and 7 ). Once volume distributions of the receptors and antibodies have been calculated, AFM imaging of the complexes is performed on approximately 100–150 mm2 of scanning area (step 8 ). Note that volumes of antibodies are usually much smaller than the volumes of the receptors. Volumes of receptor–antibody complexes consistent with the presence of one receptor and one or more antibodies bound are selected (step 9 ). Doubly decorated receptors are further analyzed to calculate the angle between their particle peaks. The figure indicates the decoration of a tetrameric receptor, and so angles show the expected separation of antibodies at 90°.

describes a step-by-step protocol for the membrane protein ­purification process, adsorption onto mica, and AFM imaging of an adequate number of receptor–antibody complexes. It is expected that a moderately skilled person with basic knowledge of AFM operation and biochemistry will be able to reproduce the method faithfully.

2. Materials 2.1. Stock Buffers

1. Lysis stock solution: filtered 10 mM Tris–HCl, pH 7.6. 2. HBS 10× stock solution: filtered 0.5 M HEPES, pH 7.6, and 1 M NaCl.

50

Carnally, Edwardson, and Barrera

2.2. Working Solutions for Receptor Isolation

1. HBS/EDTA buffer: HBS 1× solution, 2  mM ethylenediaminetetraacetic acid disodium salt (EDTA). 2. Lysis buffer: lysis stock solution, 0.5 mM EDTA. For a 5 ml solution, add 1  mM phenylmethylsulfonyl fluoride (PMSF) and one-half protease inhibitor tablet. 3. Solubilization buffer: HBS 1× solution, 0.5 M NaCl, and 1% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate (CHAPS). For a 9 ml solution, add 1 mM PMSF and one protease inhibitor tablet. 4. Washing buffer: HBS 1× solution and 0.5% CHAPS. 5. Washing buffer/low imidazole: washing buffer, 100 mM imidazole, and 0.5% CHAPS. 6. Elution buffer: washing buffer, 500 mM imidazole, and 0.5% CHAPS. All working solutions should be prepared ~1 h before use and kept on ice.

2.3. AFM and Accessories

1. An AFM equipped with a 120-mm J-scanner and a dry ­imaging cell (Veeco, Santa Barbara, CA, USA). 2. Silicon cantilevers with a drive frequency of ~300 kHz and a specified spring constant of 40 N/m (MikroMasch, Madrid, Spain).

3. Methods 3.1. Receptor Isolation

All procedures are carried out at 4°C. 1. Collect flasks containing transfected cells. 2. Discard medium. 3. Wash once with HBS/EDTA buffer (add it gently, swirl a little, and then discard by pipetting – avoid losing live cells). 4. The cells are then removed from the flasks by agitation (hitting the sides and base of the flasks, and squirting the HBS directly onto the cells, pipetting up and down to collect more) with 9 ml of the same buffer. Transfer the cells into a 50-ml tube. 5. Rinse the flasks with another 10  ml of the same buffer (to collect remaining cells) and transfer the content into another 50-ml tube. 6. The cells are pelleted by centrifugation at ~900 × g for 7 min at 4°C. 7. Begin centrifuge’s cooling program (to 4°C). 8. The pellets are resuspended in 5 ml of lysis buffer. Transfer 5  ml of lysis buffer into the first tube and mix thoroughly. Transfer this to the second tube and mix thoroughly.

Imaging the Spatial Orientation of Subunits Within Membrane Receptors

51

9. Incubate on ice for 20 min. 10. Put the homogenizer on ice. 11. The cells are homogenized with 30 strokes of the homogenizer. 12. The homogenate is transferred to four tubes and centrifuged at 660 × g for 8 min at 4°C. 13. The supernatant is removed and divided equally between four tubes. The pellet can sometimes have a rather diffuse boundary with the supernatant, so to improve purity of the final product, leave a little supernatant behind to ensure that the pellet remains undisturbed. 14. Two further tubes are filled with 400 ml of the supernatant, taking 200 ml from each of the four tubes. 15. All six tubes are centrifuged again at 21,000 × g for 15 min at 4°C. 16. Remove the supernatant. 17. Freeze the pellet from one of the two tubes containing 400 ml – this will be the “membrane fraction” for the Western blot (Fig. 3). 18. The pellets of the remaining five tubes are solubilized in a total volume of 9  ml of solubilization buffer. Take 1  ml of solubilizing buffer and thoroughly resuspend a pellet, and add the suspension to a fresh tube. 19. Repeat the same for all the pellets and pool the suspensions. 20. Add the remaining solubilization buffer, mix by inversion, and incubate for 1 h at 4°C with gentle rotation. 21. The suspension is transferred to a fresh Beckman ultracentrifuge tube.

Fig. 3. Purification of membrane receptors containing His6-tagged subunits, using nickel columns. This procedure corresponds to the step 3 in the AFM method workflow shown in Fig. 2. (a) Silver staining of the membrane fraction showing a large number of solubilized proteins. After elution from the nickel column, only the first fraction with 500 mM imidazole gives a positive band for the His6-tagged subunit. (b) Western blot using anti-His6 antibody shows that the subunit is present in both the membrane fraction and the imidazole elution fraction.

52

Carnally, Edwardson, and Barrera

22. Spin at 170,000 × g for 1 h at 4°C. 23. Start washing the beads 30  min before the spin is due to finish. 24. Cut the very end of a blue P1000 tip. Mix the beads (which will have settled) by inversion and transfer 1 ml of slurry into a fresh 15- ml tube. 25. Spin down (1,000 × g for 3 min). 26. Aspirate the supernatant, add 8  ml washing buffer, mix by inversion until all beads are resuspended, and spin down (1,000 × g for 3 min). 27. Repeat the previous step (2× washes in total). 28. Add the supernatant from the 170,000 × g spin (leaving roughly 0.5 ml behind to avoid disturbing the pellet) to the washed beads. Bring to 100  mM imidazole and incubate overnight at 4°C with gentle inversion. 29. On the next day, spin down the beads (1,000 × g for 3 min). 30. The beads are washed four times with 8 ml of washing ­buffer/ low imidazole (as previously done – spin down, aspirate, and resuspend). 31. The beads are resuspended in an equal volume (~0.5 ml) of washing buffer and transferred to a purification column with a blue P1000 tip with the end cut off. 32. Wait a few minutes for the beads to settle, then break off the end of the column’s cap. 33. Washing buffer eluent is allowed to run through into a beaker. 34. Prewash with 5 ml washing buffer/low imidazole. 35. Protein is eluted with 3 × 1 ml of elution buffer. 36. Western blot and silver staining protocol are carried out with the different fractions in order to detect which contains the greatest quantity of protein at the greatest purity (Fig. 3). 3.2. Adsorption of Purified Receptors onto Mica Support

1. Mica is cleaved to reveal pristine basal plane, which is then incubated with 1% poly-l-lysine for 30  min, washed under ultra-pure water, and dried under nitrogen. 2. Isolated receptors from the selected elution fraction are adsorbed onto poly-l-lysine-coated mica for 10  min (see Note 1), which is then washed 10× with 1 ml ultra-pure water and dried gently under dry nitrogen.

3.3. AFM Operation for the Imaging of Individual Receptors and Anti-tag Antibodies

1. For an appropriate AFM scan of an area of 1 mm2 (see Note 2), not more than 50 isolated particles (receptors) should be observed, which in our experimental conditions corresponds to dilutions of the selected fraction in a range of 1:2 to 1:100 (see Note 3). Similarly, anti-tag antibodies previously absorbed

Imaging the Spatial Orientation of Subunits Within Membrane Receptors

53

Fig.  4. Effect of sample preparation on the membrane protein adsorption onto mica. (a) Optimal density of adsorbed receptors (R) is observed after 10-min incubation on mica. More concentrated samples can generate either higher density with the presence of dimers of receptors (R–R) (b) or micelle-like structures (M) interacting with isolated receptors (c).

onto mica should be of a particle density of no more than 50–70 per 1 mm2. The common AFM imaging parameters for air tapping mode are amplitude set point = 80–90% of free level, X/Y axis length = 1 mm, frequency of scanning = 3–4 Hz, integral gain = 0.25, proportional gain = 0.35, Z limit = 500– 1,000 nm (see Notes 4 and 5), Z scale = 5–10 nm, and engage set point = 0.98–0.99 (Fig. 4a).

54

Carnally, Edwardson, and Barrera

3.4. Co-incubation of Isolated Receptors and Anti-tag Antibodies In Vitro

1. Once optimal dilution conditions for the AFM imaging of individual receptors and antibodies are obtained, an overnight co-incubation of the components is carried out in vitro in a 1-ml Eppendorf tube at 4°C prior to deposition onto poly-l-lysine-coated mica. Depending upon the number and type of subunits within a receptor, it might be necessary to estimate the subunit orientation by decorating receptors with two antibodies at the same time. As IgG antibodies have similar molecular weights, one antibody should be cleaved by papain to generate Fab fragments, which are co-incubated with anti-tag IgG antibodies and purified receptors. Chemical modification of the mica surface with poly-glutamate or receptor ligands can also be used to force the adsorption of some domains. The application of Fab fragments and chemical modification is explained in detail in our paper on the characterization of the a4b3d GABAA receptor stoichiometry (6).

3.5. AFM Operation for the Imaging of Antibody–Receptor Complexes

1. Antibody–receptor complexes are imaged with AFM para­ meters similar to those used for the individual components (see Subheading  3.3 and Note 6). In order to perform an optimal identification of all the structures – free antibody and receptors, and complexes – the density of particles in an area of 1 mm2 should not be above 100 (Fig. 5a).

3.6. Automatic Analysis of the Receptors and Complexes Using an In-House Algorithm and Interpretation of the Results

1. All the AFM images are analyzed using an in-house algorithm currently designed for Matlab language. Briefly, multiple thresholding is performed on the images in a stepwise decreasing manner to reduce the influence of the uneven mica surface due to poly-l-lysine or protein spreading. 2. Once geometric conditions are satisfied for each individual receptor, individual antibody, and antibody–receptor complex (formed when one, two, or more antibodies bind), segmentation in a probabilistic way is carried out on each structure. For each particle, parameters such as particle height (h) and half-height radii (r) are included in the equation of molecular volume Vm = (p × h/6) (3r 2 + h 2) (13). As a result, distributions of molecular volumes for all the components are automatically calculated. Each time a new antibody is used, a volume analysis of the isolated antibody should be carried out, as different IgGs, despite close similarities in molecular weight, can have different apparent molecular volumes. 3. Doubly decorated receptor–antibody complexes are further considered to calculate the angle distribution between the highest peaks of each protein (Fig. 5b, c). This information will give us the subunit orientation within a receptor. For example, a trimeric receptor should have adjacent bound

Imaging the Spatial Orientation of Subunits Within Membrane Receptors

55

Fig. 5. Automated identification of tetrameric receptors, antibodies, and antibody–receptor complexes. (a) AFM images of  receptors co-incubated with anti-subunit antibodies. (b) Multiple thresholding and segmentation of receptors (R), antibodies (A), and the following complexes: single antibody-bound receptors (S) and double antibody-bound receptors (D). It is observed that the optimal procedure allows positive identification of the different structures. A sequential increase in the thresholding values decreases the number of structures detected and makes artifacts such as (1) the S-complex generated from the correctly identified D-complex, and (2) the R generated from the correctly identified S-complex. (c) Scheme of the resulting interpretation derived from AFM imaging of the tetrameric receptor.

56

Carnally, Edwardson, and Barrera

antibodies separated by an angle of 120°, and a tetramer by 90° and 180° (see Note 7). Depending upon the number of different subunits composing the receptor, a series of different anti-tag antibodies will be used. Note that a non-related anti-tag antibody should be tested to calculate unspecific binding profiles (Fig. 2). A full binding profile is then calculated for receptors co-incubated with several antibodies, which allows us to deduce the stoichiometry and orientation of subunits (details of the full automatic procedure are described in ref. 14). 4. Once a sufficient number of membrane protein complexes have been visualized by AFM, data acquisition and processing can be completed in approximately 10  min per 1  mm2 scan (see Note 8). 3.7. Conclusion

We have presented here a detailed protocol for the AFM imaging of purified membrane receptors. We have demonstrated that the AFM imaging of receptor–anti-subunit antibody complexes can provide structural information about the subunit stoichiometry and orientation within a receptor. To our knowledge, apart from high-resolution techniques such as X-ray crystallography and EM, this is the only method that has been consistently applied to study the molecular architecture of membrane proteins, including receptors and ion channels. Therefore, we envisage its application to other families of membrane proteins including transporters, multiprotein enzymes, and molecular machines in the near future.

4. Notes 1. Timing of AFM sample deposition: We have demonstrated that the density of protein adsorbed onto mica does not change during 5–20 min after deposition (data not shown); therefore, as stated in our protocol, a period of 10  min is adequate to obtain enough individual receptors without producing overlapping of sample layers (Figs. 4a and 5a). 2. Capture of AFM images: When capturing consecutive images, ensure that the separation between images is double the width of the images being taken. For example, if capturing 1 mm × 1 mm images, a 1 mm separation between consecutive images will be insufficient, as the lateral positioning for most open-loop atomic force microscopes is such that a portion of the current image may overlap with a portion of the previous image, resulting in duplication of data for the regions imaged twice.

Imaging the Spatial Orientation of Subunits Within Membrane Receptors

57

3. Adsorption of membrane proteins onto mica: Optimal ­density of adsorbed receptors is obtained by trial and error. If there is an excess of adsorbed receptors, unspecific aggregation of receptors can be observed (Fig. 4b). Similarly, a high concentration of detergent present on the mica can generate strong interactions with the isolated receptors (Fig. 4c). 4. Z limit: The Veeco Multimode AFM has 16-bit resolution on all three axes. This means that the data for each axis are expressed across ~66,000 divisions (216 = 65,536). If the Z limit is set to maximum (~5.2 mm for a J-scanner), each division represents ~0.1 nm, which, when imaging objects, particularly antibodies that are often below 1 nm in height, can lead to significant pixellation on the vertical axis (Fig.  6a). Lowering the Z limit to 500  nm, or 10% of maximum, will decrease the size of each vertical division by a factor of 10, so that each division now represents 0.01  nm. This will not increase the vertical resolution fully tenfold, as 0.01 nm is near or most probably below the working vertical noise limit of most AFMs, but the reduction will reduce vertical pixellation and lead to more precise vertical measurements (Fig. 6b). 5. Automatic scanning: If using automatic scanning, approach the surface and image for ~30 min with the Z limit set to maximum, to allow the Z-piezo element to “settle.” Once any

Fig. 6. Vertical pixellation effect on AFM imaging of antibodies. (a) Z limit set to 5.2 mm for the AFM imaging of an isolated IgG antibody. (b) Z limit set to 500 nm reduces vertical pixellation. Section analysis is indicated for both images.

58

Carnally, Edwardson, and Barrera

drift has ceased, lower the Z limit to 500–1,000  nm and c­ ommence automated imaging. This is to prevent the Z-piezo from drifting out of its range due to drift. 6. Resolution of antibody–receptor complexes: The amplitude set point is kept at or above 80% of the free level to avoid excessive vertical compression of the sample. Furthermore, the selection of high engage set points (0.98–0.99) helps to preserve the geometry of the probe apex by helping to minimize the forces applied to it (and as a result also to the ­sample). Maximum resolution for scanning of membrane proteins has been achieved using contact mode AFM over ordered monolayers of proteins; however, the resolution is significantly lower for isolated membrane proteins (15). Although we usually use probes with stiff cantilevers (see Subheading  2.3, item 2) to ensure stable scanning of single membrane receptors, the instrument parameters mentioned above allows, in rare cases, high-resolution AFM images of receptors decorated with antibodies (Fig. 7). 7. Automatic identification of antibody–receptor complexes: Multiple thresholding values should be kept as low as possible to identify all the structures present on the surface. Although poly-l-lysine-coated mica is very flat, very small fluctuations in the surface will affect the thresholding values needed for receptor identification. A general suggestion is to keep this at 0.02  nm over the averaged height of surface fluctuations. Under higher thresholding values, some complexes and individual receptors may be absent from particle segmentation (Fig. 5b). In addition, this can lead to the incorrect identification of complexes, including the disappearance of antibodies

Fig. 7. High-resolution AFM image of the decoration of trimeric receptor. Left panel indicates an AFM image of a purified receptor complexed with anti-His6 antibodies. Right panel shows the interpreted stoichiometry based on the image. Two IgG antibodies with the characteristic three domains (2 × Fab and Fc) are visible. Only one Fab domain interacts with the receptor. The angle between the highest peaks at every particle is around 120°, which suggests a trimeric stoichiometry for the receptor.

Imaging the Spatial Orientation of Subunits Within Membrane Receptors

59

correctly bound to receptors (Fig. 5b). Note that calculated dimensions for each structure do not change under different thresholding. A scheme showing the final interpretation of the receptor stoichiometry is shown in Fig. 5c. 8. AFM scanning area: An area of 1 mm2 is considered optimal for the automatic identification of receptor and antibody– receptor complexes. If an area bigger than 2 mm2 is chosen, the frequency of scanning should be decreased up to 2 Hz to maintain stable tip–sample interaction. In addition, at a resolution of 512 × 512 pixels, selection of bigger scanning areas will result in a decrease in the number of complexes identified automatically. This low efficiency is due to the uncertainty of geometrical parameters associated with the multiple thresholding, which in turn prevents segmentation of complexes.

Acknowledgments We thank Dr. Robert Henderson (Department of Pharmacology, University of Cambridge) and Drs. Haifang Ge and William Fitzgerald (Department of Engineering, University of Cambridge) for their support in AFM imaging and image processing. References 1. Lacapere, J. J., Pebay-Peyroula, E., Neumann, J. M., and Etchebest, C. (2007) Determining membrane protein structures: still a challenge! Trends Biochem Sci 32, 259–270. 2. Muller, D. J. (2008) AFM: a nanotool in membrane biology Biochemistry 47, 7986–7998. 3. Shahin, V., and Barrera, N. P. (2008) Providing unique insight into cell biology via atomic force microscopy Int Rev Cytol 265, 227–252. 4. Barrera, N. P., and Edwardson, J. M. (2008) The subunit arrangement and assembly of ionotropic receptors Trends Neurosci 31, 569–576. 5. Barrera, N. P., Henderson, R. M., and Edwardson, J. M. (2008) Determination of the architecture of ionotropic receptors using AFM imaging Pflugers Arch 456, 199–209. 6. Barrera, N. P., Betts, J., You, H., Henderson, R. M., Martin, I. L., Dunn, S. M., and Edwardson, J. M. (2008) Atomic force microscopy reveals the stoichiometry and subunit arrangement of the alpha4beta3delta GABA(A) receptor Mol Pharmacol 73, 960–967.

7. Barrera, N. P., Henderson, R. M., MurrellLagnado, R. D., and Edwardson, J. M. (2007) The stoichiometry of P2X2/6 receptor heteromers depends on relative subunit expression levels Biophys J 93, 505–512. 8. Barrera, N. P., Herbert, P., Henderson, R. M., Martin, I. L., and Edwardson, J. M. (2005) Atomic force microscopy reveals the stoichiometry and subunit arrangement of 5-HT3 receptors Proc Natl Acad Sci USA 102, 12595–12600. 9. Barrera, N. P., Ormond, S. J., Henderson, R. M., Murrell-Lagnado, R. D., and Edwardson, J. M. (2005) Atomic force microscopy imaging demonstrates that P2X2 receptors are trimers but that P2X6 receptor subunits do not oligomerize J Biol Chem 280, 10759–10765. 10. Barrera, N. P., Shaifta, Y., McFadzean, I., Ward, J. P., Henderson, R. M., and Edwardson, J. M. (2007) AFM imaging reveals the tetrameric structure of the TRPC1 channel Biochem Biophys Res Commun 358, 1086–1090. 11. Carnally, S. M., Dev, H. S., Stewart, A. P., Barrera, N. P., Van Bemmelen, M. X., Schild, L., Henderson, R. M., and Edwardson, J. M. (2008) Direct visualization of the trimeric

60

Carnally, Edwardson, and Barrera

structure of the ASIC1a channel, using AFM imaging Biochem Biophys Res Commun 372, 752–755. 12. Ormond, S. J., Barrera, N. P., Qureshi, O. S., Henderson, R. M., Edwardson, J. M., and Murrell-Lagnado, R. D. (2006) An uncharged region within the N terminus of the P2X6 receptor inhibits its assembly and exit from the endoplasmic reticulum Mol Pharmacol 69, 1692–1700. 13. Schneider, S. W., Lärmer, J., Henderson, R. M., and Oberleithner, H. (1998) Molecular

weights of individual proteins correlate with molecular volumes measured by atomic force microscopy Pflugers Arch 435, 362–367. 14. Barrera, N. P., Ge, H., Henderson, R. M., Fitzgerald, W. J., and Edwardson, J. M. (2008) Automated analysis of the architecture of receptors, imaged by atomic force microscopy Micron 39, 101–110. 15. Fotiadis, D., and Engel, A. (2004) Highresolution imaging of bacteriorhodopsin by atomic force microscopy Methods Mol Biol 242, 291–303.

Chapter 5 High Resolution Imaging of Immunoglobulin G Antibodies and Other Biomolecules Using Amplitude Modulation Atomic Force Microscopy in Air Sergio Santos and Neil H. Thomson Abstract The atomic force microscope (AFM) is a very versatile tool for studying biological samples at ­nanometre-scale resolution. The resolution one achieves depends on many factors, including the sample properties, the imaging environment, the AFM tip and cantilever probe characteristics, and the signal detection and feedback control mechanism, to name a few. This chapter describes how to routinely achieve the highest possible spatial resolution on isolated protein molecules on mica surfaces. This is illustrated with Immunoglobulin G antibodies but the methods apply equally well to any other globular multi-subunit protein, as well as other biomolecules. Double-stranded DNA is used as a model sample to illustrate the effects of the force regime in amplitude modulation atomic force microscopy (AM AFM) on the image resolution and contrast. AM control is a widely used technique in biological AFM for reasons which are discussed. Key words: Immunoglobulin G, DNA, Mica, Attractive, Repulsive, Forces, Protein subunit

1. Introduction Amplitude modulation atomic force microscopy (AM AFM) has been the main AFM technique applied to biological systems since the invention of the AFM in 1986 (1). The AFM can be operated in all environments, including air and liquid, allowing biomolecules and biological samples to remain hydrated and often active (2). Dynamic AFM, where the cantilever is vibrated so that the tip contacts the surface in a normal direction, has been essential for the progression of biological imaging, since these methods eliminate damaging shear forces that arise in contact mode (3, 4). Using the amplitude as the feedback signal for dynamic AFM

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_5, © Springer Science+Business Media, LLC 2011

61

62

Santos and Thomson

allows imaging in environments, where the damping of the ­cantilever is large, namely, ambient air and liquids. In general, the resolution in liquids is lower because of intrinsic sample movement and the influence of hydrodynamic forces smearing out image contrast (5). Obtaining the highest resolution images of single, isolated biomolecules is, therefore, usually best achieved on samples that have been dried. The molecules must be bound to a suitable support surface, which is normally freshly cleaved mica. Mica undergoes basal plane cleavage to leave a fresh uncontaminated surface which is atomically flat over large distances (microns to millimetres) on which molecules can be easily detected by the AFM tip (6). Importantly, mica is a hydrophilic surface where under ambient laboratory conditions, there is sufficient humidity that a thin, nanometre-thick water layer present on the surface maintains the biomolecules in a hydrated state (7, 8). Molecules are typically deposited from a buffer solution through incubation on the mica surface, before rinsing and drying with water. The binding of the proteins to the mica surface occurs through non-covalent interactions, such as Van der Waals and other electrostatic interactions. The weak interactions between the molecule and the surface means that dynamic AFM methods are essential (unless working in a very dry environment), and AM AFM is the ideal choice (9, 10) since frictional forces are minimised. The resolution achieved by AFM depends on a number of factors, including the magnitude of height variations across the sample and the local stiffness. These parameters influence the size of the tip–sample contact area, where rougher and softer samples give lower lateral resolution. Importantly, the distance at which the AFM tip scans over the sample also strongly influences resolution. For the highest resolution, it is preferable for the tip to be as close to the sample as possible without causing damage. In AM AFM, there can exist two stable cantilever amplitudes at a given tip–sample separation, such as the High and Low state amplitudes (11). This bistable behaviour arises from the harmonic oscillation of the cantilever interacting with the non-linear potential that exists between the tip and surface. At large distances from the surface, the overall net forces are attractive (e.g., van der Waals), whereas at short distances hard sphere repulsive forces take over. The AM AFM can, therefore, be operated stably in either force regime, and imaging is known as the attractive force regime or the repulsive force regime and these are usually correlated with the bistable solutions of Low and High, respectively, but not in all cases. In this article, we refer to the attractive regime where the average force is attractive when imaging is carried out in the Low state. When imaging is carried out in other than the Low state, the repulsive regime, where the net force is repulsive, is implied and this is generally reached through the High state. Generally,

High Resolution Imaging of Immunoglobulin G Antibodies

63

the tip is, on average, closer to the sample in the repulsive force regime, and therefore this is preferred for high-resolution imaging. Nevertheless, care must be taken to ensure that the interaction forces are not so high as to induce sample damage. The progress in high-resolution AM AFM was limited for a long time by lack of understanding of the bistable nature of the cantilever and its effect on image resolution and contrast. The community is now beginning to get a full understanding and control of the AFM to give sub-molecular resolution on isolated globular proteins, such as IgG and fibrinogen (12–14). The need for reducing the magnitude of the overall forces in the tip–sample interactions and the high pressures involved in high-resolution imaging in AFM have long been recognised since the invention of the instrument (15). The large tip–sample forces involved in the interaction are particularly relevant when imaging soft molecules which, when hydrated, can have a Young’s modulus as low as 105 Pa (16). This is a key limiting factor in terms of achieving high resolution in biological systems. Thus, in order to maximise the force sensitivity and enhance resolution some have used non-contact (NC) AFM (or frequency modulation AFM) (17) where, by using extremely small free amplitudes (0.2–0.3 nm) in liquid, true molecular resolution of polydiacetylene has been achieved. Non-contact AFM in UHV has also been used to achieve high-resolution images of single- and doublestranded DNA (dsDNA) molecules, where detailed structures are revealed with these methods that cannot, according to some authors (18), be revealed in standard AM AFM. Moreover, it should not be surprising that a long-sought goal, among those aiming at reproducibly obtaining high resolution consists of reliably constructing ultra sharp AFM tips and using them without causing sample or tip deformation. To that purpose, some authors have considered functionalising the tip with carbon nanotubes and have achieved high-resolution images of dsDNA molecules via AM AFM (19). Another recent technique consists of exciting the second mode of the cantilever where the increased tip–sample sensitivity has reportedly been used to obtain high-resolution images of individual protein molecules by reducing the average force from the nano-Newtons typically required in standard AM AFM to pico-Newtons (20); the amplitude of the fundamental mode is still used as the modulating parameter in these techniques. Moreover, in order to address the problems involving thermal motion, cryo-AFM techniques were developed not long after the invention of the AFM (21). The technique involves imaging at very low temperatures (e.g., 80–100°K), where the AFM is sometimes operated in nitrogen vapour, and has been reported to generally improve resolution compared to environmental AFM techniques. The current improvement in structural stability and reproducibility at cryogenic temperatures has led to

64

Santos and Thomson

advances in resolution, such as directly resolving the linker histone in a chromatosome (22). More recently, NC AFM has been used in UHV and extremely low temperatures (5°K) with an atomically functionalised tip and single atoms within a pentacene absorbed molecule have been resolved (23). From the above, it is clear that there are many fronts open in the AFM research field that may lead to routinely and reliably obtaining high-resolution images resolving the secondary structure of single biomolecules in ambient conditions. Still, to date and to the best of our knowledge, none of these methods surpasses the capabilities of standard AM AFM when it is used appropriately. Resolving the secondary structure of single biomolecules should lead to advances in important areas of molecular biology, such as single antibody–antigen interactions.

2. Materials 2.1. Preparation of Mica for Sample Immobilisation

1. Muscovite mica in 3″ × 1¢ sheets (Agar Scientific Ltd., Unit 7, Parsonage Lane, Stansted, Essex, CM24 8GF, UK). 2. Steel SPM specimen discs measuring 12 or 15  mm in dia­ meter (Agar Scientific Ltd. as above). 3. Five-minute epoxy glue from Devcon (ITW Devcon, Unit 3, Northampton Road, Rushden, NN10 6GL, UK). 4. Homemade “punch and die” set. However, a set can be purchased from Precision Brand Products Inc (Downers Grove, Il 60515, USA). 5. Scotch tape. 6. Toothpick or Gilson plastic pipette tip (Gilson, Inc. 3000 Parmenter Street P.O. Box  620027 Middleton, WI 535620027, USA).

2.2. Immunoglobulin G Antibodies and Sample Preparation Materials

1. Immunoglobulin G (IgG) antibodies. Stock solution 0.21 mg/ml in phosphate-buffered saline with 0.01% sodium azide (NaN3). The stock solution can be pipetted in small aliquots (1–5  ml) into eppendorf tubes and flash frozen in liquid nitrogen prior to storage at −80°C. If needed, store protein solution at 4°C but it is preferable to make a fresh solution frequently. 2. Milli Q water (Millipore (UK) Ltd., Watford, WD18 8YH, UK). 3. Gilson pipettes (Gilson, Inc. 3000 Parmenter Street P.O. Box 620027 Middleton, WI 53562-0027, USA). 4. Oxygen-free nitrogen gas (cylinder).

High Resolution Imaging of Immunoglobulin G Antibodies

2.3. AFM and Other Instrumentation

65

1. A commercial Multimode AFM equipped with a 14  mm ­scanner, E-scanner (Digital Instruments, Veeco Metrology Inc., Santa Barbara, CA, USA). Nanoscope IIIa controller or later version required. 2. Diving board-shaped silicon micro-cantilevers of 160 mm in length, nominal spring constant of k = 42 N/m, and resonant frequency 300 kHz (Olympus, Tokyo, Japan).

3. Methods The methods described below outline (1) the preparation of mica supports where the sample is to be immobilised, (2) a method to physisorb the antibodies onto the mica surface for AFM imaging, and (3) a method to operate and optimise the operational para­ meters of the AFM for high-resolution imaging of antibodies in ambient conditions. 3.1. Preparation of Mica Supports for Sample Immobilisation

1. Punch out suitable circular mica sheets to fit on the steel SPM specimen discs using the 10–15-mm holes on the “punch and die” set and a hammer. 2. Mix the two component epoxy glue and place 5–10 ml of the mixture in the centre of a specimen disc, using a toothpick or a Gilson plastic pipette tip. 3. Place the mica disc in the centre of the specimen disc and press gently with tweezers, so a continuous and thin layer of glue forms between the mica and the specimen (the glue should be evenly spread on the disc and form a thin and homogeneous layer). 4. Let the glue dry for 24–48 h.

3.2. Physisorption of Antibodies onto Mica

1. Dilute the stock solution of IgG 40-fold into Milli Q water (see Note 1). 2. Cleave the mica with the scotch tape. 3. Pipette 20 ml of the diluted solution onto the freshly cleaved mica support. 4. Incubate the sample for approximately 20 s before rinsing in excess Milli Q water and subsequently drying in a stream of dry N2 gas (1 bar pressure at a distance of several cm). Rinsing can be achieved by running up to 5 ml of Milli Q water across the mica sample while it is tilted at a 30–45° angle. The degree of rinsing may affect the density of molecules left on the surface after drying.

66

Santos and Thomson

5. Transport the completed specimen disc onto the piezo tube scanner. 6. Mount the cantilever chip into the air cantilever holder using fine, square-ended tweezers. 7. Mount the cantilever holder into AFM head and the head onto the tube scanner. 8. Keep both the AFM and any cables free from any source of vibrations where possible (see Note 2). 3.3. Operation of the AFM 3.3.1. Engaging and Setting the Regime of Operation

3.3.2. Optimisation of Resolution and Reduction of Sample Damage by Setting the Appropriate Operational Parameters

After mounting the cantilever and the AFM head, the tip can be engaged to the surface. It is advisable to set the scan size to zero during engagement; the sample might be contaminated at regions or might have topographically challenging regions, thus a scan size of zero while engaging minimises tip contamination and sometimes tip deformation. Once the imaging set-point and feedback gains have been optimised, and then one can set the scan size. It is also advisable to zero the phase immediately after setting the drive frequency prior to engaging. Then the working regime, that is Attractive and Repulsive force regimes, can be monitored in the Multimode system by checking the average phase shift value in each scan; negative and positive values correspond to the Attractive and Repulsive regime, respectively. Both z-piezo and phase-shift contrast should be acquired simultaneously in order to monitor the force regime and stability as the tip scans the sample (see Notes 2 and 3). Standard imaging is performed with free amplitudes (A) in the range of 15–40 nm, commonly with the cantilever driven below resonance in order to avoid bistability and work in the intermittent contact mode (9). This typically leads to imaging in the singlebranched region, where there is no bistability and a single branch exists, and the repulsive force regime. However, resolution might be maximised by working at or above resonance sometimes by using smaller free amplitudes. It is also customary to use values of amplitude set-point (Asp) as high as possible, for example Asp/A > 0.9. Generally, lower amplitudes and higher working frequencies (relative to the resonant frequency) increase the probability of the cantilever being in the Low state and being influenced by attractive forces only. In these cases, the tip is often operating in a non-contact imaging regime. In general, biomolecular specimens and, in particular, globular proteins such as antibodies might be easily deformed by using large forces and these are highly dependent on free amplitude. Nevertheless, imaging in the repulsive regime can lead to higher resolution because the tip is, on average, closer to the specimen, thus a compromise is needed. The repulsive regime can be stably reached by sufficiently decreasing the Asp for sufficiently high free amplitude.

High Resolution Imaging of Immunoglobulin G Antibodies

67

Fig. 1. Imaging using amplitude modulation at very high set-points or with the scan speed set too high results in the AFM tip skipping over the features and not resolving the molecules clearly. These are images of IgG antibodies physisorbed onto mica. (a) Height or z-piezo image; (b) phase signal image. The dotted white arrows indicate the fast scan direction during the simultaneous data acquisition of both images. One can see “shadows” behind the molecules due to the ­comparatively slow response time of the AFM system in the vertical direction compared to the scan plane. The scan speed is 4.1 Hz and the normalised set-point ratio is 0.89.

Thus, in order to minimise tip–sample interactions, the free amplitude should be kept as low as possible and slowly increased to reach the repulsive regime, that is, any increase in free amplitude should be performed smoothly from low values; for example, increasing from 5 to 6 nm. The Asp should also be reduced gradually, since abruptly reaching the repulsive regime can result in either tip or sample deformation (see Notes 5 and 6). The number of pixels per frame should be kept as high as possible (512 × 512 in a commercial Multimode). For high magnifications, the scan size should be kept in the range of 200 nm to 1 mm. Typical scan rates for these scan sizes go from 1 to 3 Hz (lines per second). If the scan rate is too high, then the features become blurred as the reaction time of the cantilever and the electronic feedback loop is not sufficiently fast to bring the tip back onto the surface when it goes over a molecule (see Fig. 1). The integral gain is commonly kept as high as possible, but not so high as to produce resonance of the z-piezo, which creates ripples in the scan. Typically, the integral gain needs to be around 0.5 for these cantilevers operating with AM AFM in air. The exact value may need to be adjusted, however, depending upon the sample and exact conditions. 3.3.3. Optimisation of Resolution by Controlling Relative Humidity and Regime of Operation

Relative humidity (RH) also plays a role in the operational parameters required to obtain high resolution. In particular, when ­operating in the non-contact regime, that is, when small values of free amplitude are used, the apparent height of the molecules is significantly reduced at high relative humidity (e.g., RH > 70%).

68

Santos and Thomson

This is due to the effects of water layers partially or totally ­covering the molecules and screening the surface van der Waals forces that would otherwise be used as feedback to track the topography. Thus, in order to scan in the non-contact regime with small free amplitudes, it is advisable to image at lower relative humidity (e.g., RH < 60%, see Note 4). 3.4. Conclusion

We have presented detailed information on the methods and ­sample preparation required to obtain high-resolution images of antibodies and other biomolecules that could be similar in size or even smaller. We show typical images of IgG and other molecules, such as DNA gyrase and dsDNA, obtained with the use of a commercial multimode AFM operated with Amplitude Modulation control in environmental conditions. It has been emphasised that an understanding of the operational parameters, force regimes, and origin of the surface forces is required in order to optimise tip–sample interactions. The optimisation procedure consistently leads to higher resolution and improved reproducibility. Establishment of a reliable technique allows interaction between molecules (e.g., antibody binding to antigens) to be observed with a high degree of precision. The AFM used in this way has increasing impact on the area of biomedical research.

4. Notes 1. Biomolecular concentration during deposition: the stock solution should be diluted to an appropriate concentration; too many molecules on the surface may impede visualisation of single molecules and sub-molecular structure, and induce unwanted molecular interactions. Typically, we find that concentrations in the range of pico-molar to nano-molar (nM) are usually sufficient, but depend on the propensity of binding of a given biomolecule to mica. The concentration should be high enough to guarantee some binding. Similar problems of unoptimised surface density can result from prolonged incubation time and excessive/insufficient rinsing and/or drying. 2. Damping of background vibrations. For high-resolution AFM, damping out of environmental vibrations is essential. Acoustic disturbances can be eliminated by covering the AFM head in an enclosure made of dense foam. An inexpensive way to eliminate building vibration is to suspend the AFM head on a heavy platform. We use a concrete slab connected to a single suspension point using bungee cords that are extended but not to their elastic limit. The platform can be raised from or lowered onto a stable laboratory bench by means of a mechanical winch.

High Resolution Imaging of Immunoglobulin G Antibodies

69

3. The AFM design: All the data presented here have been acquired with a Digital Instruments Multimode Nanoscope IIIa AFM (Veeco Metrology Inc.). This design of AFM is mechanically optimised for high-resolution imaging and performs very well. These experiments might be successful on other AFM platforms that are specifically optimised for highresolution work. 4. Effects of relative humidity in AM AFM in air. Even though the resolution might improve at low relative humidity, it is known that surface features tend to form on the surface of mica. These are clearly visible in the non-contact mode of operation and, more generally, in the Attractive force regime. It has been recently verified that potassium carbonate crystallites (24) are one of the structures that may spontaneously form on mica surfaces upon cleavage. The features might affect molecular apparent height and the overall interpretation of the image; the background might appear dirty and lacking flatness. We have found that these structures disappear at higher relative humidity (e.g., RH > 70%); potassium carbonate is soluble in water. In fact, similar studies carried out with the surface force apparatus show that some of the surface precipitates are water soluble and that the contamination layer can be removed by immersing the mica surface in water (25, 26). Unfortunately, as previously mentioned, at high relative humidity, the water layers have a screening effect on the intermolecular surface forces. This screening has a direct effect on the apparent molecular height, giving lower values in the Attractive regime at high RH, particularly if small free amplitudes (A < 10 nm) and low aspect ratio tips (R > 15–20 nm) are used. It is important to stress that this is only a tendency, since results might vary significantly with even small differences in tip radii, cantilever spring constant, drive frequency, free amplitude, Asp, and the thickness of the water layers. This is due to the nature of the screening effect of the water and its relationship to the tip–sample proximity and the tip radius. In the attractive force regime, and in particular, when there is no tip–sample intermittent contact, the dynamics of the cantilever are solely controlled by the longrange interactions (e.g., van der Waals interactions). When there is sufficient water around, the van der Waals forces due to the water molecules screen the van der Waals forces due to the surface. Thus, in order to “sense” the true solid surface, the tip must get as close as possible to the surface and ideally through the water layer. Since many parameters contribute to the final tip–sample separation in a non-linear fashion, the user must experimentally find the optimum separation in each situation. An example for dsDNA is shown in Fig.  2. Although dsDNA fragments have been used in this example,

70

Santos and Thomson

Fig. 2. Effects of relative humidity (RH) on the molecular width and height of dsDNA fragments on mica by AM AFM in the non-contact mode of operation. Height (z-piezo) images taken in the Low state. First, (a) and (b) together are an example for which the background impurities at low RH are of the order of 0.15–0.2 nm in apparent height. For the same molecules and cantilever-tip system, an increase in RH reduced the background impurities but the screening effect of the water molecules induces some loss in apparent height (b). A second sample and AFM probe have been used to produce the scans in (c) and (d) where the effects of the impurities are more dramatic at low RH (c). However, the same clearing effect of the background is observed at the high RH (d). Molecular average heights and widths are given on the bottomleft part for each scan in nm as H and W, respectively. In (a) and (c) the heights are given as two-part values and these correspond to the height of the molecules relative to the salt layers (lighter part of the background) and the true mica surface (darker part of the background). At high RH, these effects are minimised, so a single average value is given.

the effects are equally valid for antibody imaging or other single biomolecules. Figure  2 shows two different samples, one in (a and b), and another in (c and d). The four images have been obtained in the non-contact mode of operation (Low state) and in the attractive force regime with A = 3 nm and Asp/A = 0.9 driving at resonance. The differences in molecular and background contrast are due mainly to operating at different RH, namely, 30% in (a) and (c) and above 70% in (b) and (d). The apparent average height and width of the molecules are given on the bottom-left corner of each scan as H and W, respectively. The common outcome of these experiments as the humidity increases is twofold; first, the average apparent height of the DNA tends to decrease and second, the background becomes clearer as surface contamination is cleared and/or solubilises. In Fig. 2, two values of the height are given at low relative humidity, these are the

High Resolution Imaging of Immunoglobulin G Antibodies

71

heights of the molecules relative to the salt layer (lighter part of the ­background) and relative to the mica surface. The ­discrepancies in apparent height and width for a given RH can be attributed to slight differences in spring constant, tip radius, average tip–sample separation, the formation of molecular structures on the surface of the mica, and sample preparation and degeneration due to variations in RH. To sum up, the two examples in Fig. 2 below represent the two most common results when imaging biomolecules on a mica surface with small amplitudes (A < 10  nm), relatively sharp tips (R < 10–15  nm) and relatively stiff cantilevers (k ~ 40 N/m). Note that the effects of varying RH are generally irreversible, especially in terms of molecular height. As stated, sometimes it is possible to increase both lateral and topographic contrast by reaching intermittent contact (typically through the repulsive force regime). This can be achieved by either increasing the cantilever free amplitude or reducing the Asp. However, there exists a minimum value of free amplitude (depending on the particular system) for which the repulsive regime is reached. In Fig. 3, the differences in contrast when imaging in the attractive and the repulsive force regimes at both very low (Fig. 3a, c) and very high (Fig. 3b, d) relative humidity are shown. The four scans have been obtained at resonance with a free amplitude of 4–5 nm. The attractive regime has been reached by operating at high Asps (Asp/A > 0.85) and has been used to obtain (a) and (b). Conversely, the repulsive regime has been reached by lowering the Asp while keeping the free amplitude and drive frequency constant (Fig. 3b, d). The values of apparent DNA height and width are also given for each scan. Comparing Fig.  3a, b with Fig.  3c, d, it can be concluded that higher contrast can be obtained by reaching intermittent contact. Moreover, in the Repulsive regime, surface features are generally minimised as compared to the attractive regime even at low RH (compare Fig. 3a with b). Additionally, even when imaging in the attractive regime, the blurring effect caused by the background or surface features can sometimes be minimised by increasing the free amplitude. 5. Optimising operational parameters to image antibodies in the Repulsive regime. It should be clear now that small tip–sample forces are typically required to obtain high-resolution images of biomolecules, such as IgG. However, it is still debated whether purely non-contact interactions or non-contact imaging is the ultimate mode of operation. In particular, it is argued that if the tip is kept far enough from the sample and the interactions are always long range or van der Waals, then neither the tip nor the sample are ever submitted to the

72

Santos and Thomson

Fig. 3. Topography images of dsDNA fragments obtained at both (a) and (c) very low and (b) and (d) very high RH. The top scans (a and b) have been obtained in the attractive regime, whereas the bottom scans (c and d) have been obtained in the repulsive regime. It is observed that spatial resolution can be greatly increased by reaching the repulsive regime. Significantly, the repulsive regime can be sued to increase both lateral resolution and height at high RH by careful control of the operational parameters. Again, the screening effects of the water layers are noticeable in the attractive regime at high RH. The increase in apparent height at high RH in the repulsive regime as compared to low RH can be due to both molecular hydration and differences in average tip–sample separation induced by changes in dynamics due to the molecular interactions with the water layers. In fact, the average tip–sample separation is responsible for breaking through the water layer and reaching the surface, thus minimising the screening effect of the water layers. Note that some background features can still be observed in the repulsive regime at low RH.

repulsive forces responsible for permanently deforming surfaces. In AM AFM, the non-contact mode of operation is typically reached by keeping the free amplitude low (this has been demonstrated in Figs.  2 and 3 above). In fact, our experiments have also shown that the sub-molecular structure of IgG can be resolved in the attractive force regime and non-contact mode of operation (13). Examples are shown in Fig. 4, where a dessicated sample of IgG on mica has been imaged. In Fig. 4a, there are four IgG molecules lying next to each other showing the characteristic tri-nodular ­structure,

High Resolution Imaging of Immunoglobulin G Antibodies

73

Fig.  4. Attractive force regime imaging of IgG antibodies physisorbed on muscovite mica. (a) High-magnification z-piezo (topographic height) image of four IgG molecules which all show the characteristic tri-nodular appearance. The Fc domains are indicated by the white arrows. (b) Height (z-piezo) image of a field of IgG molecules with random physisorbed orientations. The molecules marked A, B, and C also give the tri-nodular appearance. ((b) Reproduced from Fig. 2, with permission from Elsevier (13)).

where the two Fab and Fc subunits of the ­antibody are clearly resolved. The white arrows highlight where the Fc domain is expected to be, in between the two Fab domains. Figure 4b shows a whole field of molecules at lower magnification where most molecules appear as projections of this tri-nodal shape, due to the stochastic nature of physisorption of the molecules to the mica does not preferentially orient them at the surface. The tri-nodular appearance is achieved when the IgG ­molecules are lying “flat” on the mica surface with all three subunits accessible to the tip (e.g., molecules marked A, B, and C). On the other hand, the versatility of the AFM easily allows direct comparison between force regimes and operation modes. In short, by sufficiently increasing the free amplitude and/or sufficiently decreasing the Asp, it is possible to intermittently contact the sample. This may or may not induce irreversible sample deformation depending on tip radii, free amplitude, Asp, spring constant and several other relatively less important parameters. In particular, from Fig. 2, it is clear that in certain occasions it might be beneficial to image in the Repulsive regime. In fact our results (Fig. 5) show that it is possible to image antibodies, in particular IgG, in the repulsive regime without causing permanent sample deformation and actually increasing contrast. Nevertheless, the outcome is highly sensitive on free amplitude and tip radii. The procedure should be as follows: the drive frequency should be set near resonance, in particular at or slightly above

74

Santos and Thomson

Fig. 5. Effect of reducing the amplitude set-point (Asp) on image contrast and resolution. The image was taken with the slow scan direction downwards. The normalised Asp was reduced step-wise from 1.00 at the top of the image to 0.96 at the bottom. One can see that the IgG molecules are not well resolved in the top of the image in the attractive force regime. In the middle of the image, with intermediate set-points, the cantilever is switching back and forth between the two oscillation states. In the repulsive force regime at the bottom, imaging becomes stable and sub-molecular resolution is achieved. The circle highlights an IgG in the “flat” orientation showing the tri-nodular structure. (Adapted from Fig. 4, with permission from Wiley-Blackwell (8)).

resonance in order to avoid sudden jumps to the repulsive regime. Then the free amplitude should be initially set relatively low, for example to 5 nm, while keeping the Asp high, for example Asp/A > 0.95. After engaging, the normalised Asp should be slightly reduced to a minimum of 0.9–0.85 in order to try and reach the repulsive regime. If the repulsive regime is not reached, then the tip should be disengaged and the free amplitude should be slightly increased (say 0.5 nm) and/or the drive frequency should be set slightly lower (Df~ −0.01%). The same procedure should then be followed until the repulsive regime is reached stably. The main aim is to reach the repulsive regime stably with a free amplitude as small as possible and with an Asp as high as possible. This is a direct consequence of the tip–sample forces ever increasing with free amplitude and decreasing Asp. Note that a region of bistability could be reached before stably reaching the Repulsive regime. As the set-point is decreased, constant switching between the two cantilever oscillation states can occur; an example of this is shown in Fig. 5. From top to bottom, the Attractive regime is first reached stably by setting the Asp high (Asp/A > 0.985). Then an area of bistability follows

High Resolution Imaging of Immunoglobulin G Antibodies

75

Fig. 6. (a) AFM height data showing sub-domain resolution on the Fab domains, within the tri-nodular structure. (b) Ribbon diagram of an IgG antibody structure rendered from X-ray coordinates from the protein data bank. The X-ray structure is rendered at ~2.3 times larger magnification than the AFM image. (Adapted from Fig. 5, with permission from Elsevier (13)).

as shown in the middle section of the image. Finally, further reducing the Asp to 0.956 allows stably reaching the Repulsive regime while increasing resolution. In particular, it is observed that the tri-nodular structure of the antibody IgG is clearly resolved. Care should be taken, however, since abruptly increasing the free amplitude or reducing the Asp might cause permanent damage to the sample or, in extreme cases, even the tip (12). In the best case scenario to date, IgG has been imaged with a resolution down to 25 kDa (13), that is, the two pairs of immunoglobulin folds that make up one fragment have been resolved (Fig. 6). 6. Using amplitude and phase distance (APD) curves to find optimal operational parameters to image in the Repulsive regime. Another method to try and reach the repulsive regime with a free amplitude as low as possible and an Asp as high as possible consists of systematically acquiring APD curves prior to scanning. In APD curves, the cantilever-sample separation (or z-piezo distance) is decreased and increased at a single point on the sample, while monitoring the amplitude reduction and the phase shift. First, a free amplitude and a drive frequency are set, then the z-distance is decreased through extension of the z-piezo and the amplitude goes down because of damping. It is not advisable to completely damp the amplitude while performing APD curves, since it might cause tip deformation (27). Usually, a trigger is set which limits the damping to a percentage of the free amplitude. Typically, the trigger is set so that only up to 70 or 80% of the cantilever free amplitude are damped. Once the amplitude reaches that minimum value, the z-distance is increased which is typically termed retraction. A method to establish the optimum free

76

Santos and Thomson

amplitude and set-point to image the antibodies is as follows; first, the drive frequency is set near resonance, typically at or above to avoid sudden jumps to the Repulsive regime. Then, a relatively low free amplitude is set, for example 5 nm. An APD curve is subsequently acquired by triggering the amplitude reduction to no more than 80%. If no switch to the repulsive regime is observed after several curves, the tip has to be disengaged and the free amplitude set slightly higher. Shifting the drive frequency slightly towards lower frequencies might also help reach the repulsive regime with a lower free amplitude. As soon as switching from the attractive to the repulsive regime is observed, the critical value of free amplitude (Ac) to reach the repulsive regime has been reached and scans can be performed with that value. It is advisable, however, to slightly increase the free amplitude, since scanning with a free amplitude of just Ac might cause constant switching (see the middle section of Fig. 5) or might require an Asp low enough to cause permanent damage to the molecules, for example Asp/A < 0.8. To sum up, the main positive effect of increasing the free amplitude slightly above Ac is that the repulsive regime can be reached stably at high enough values of Asp/A as shown in Fig. 5. Again, using very high free amplitudes (especially above Ac) should be avoided due to the fast escalation of average and peak forces with increasing A. What is required then is a compromise between resolution and deformation, and the user needs to gain experience in how to optimise the parameters to achieve sub-molecular resolution on multi-subunit proteins. An experimental example of an APD curve where a switch from the attractive (repulsive) to the repulsive (attractive) force regime has taken place during extension (retraction) is shown in Fig. 7. The behaviour of (a) the amplitude is shown on the top plot and (b) the phase behaviour is shown in the bottom plot. The free amplitude has been chosen to be 15 nm (A = 15 nm), and the cantilever has been driven at resonance. Since the phase has been zeroed there (at resonance) before engaging, the phase shift shows a value of zero degrees for the free cantilever. The stages through which the cantilever goes through during the APD curve are numbered in the top plot and correspond to the same stages for the phase behaviour. Briefly, the cantilever (1) first finds itself oscillating far above the surface. Thus, the phase shift is zero or close to zero and the amplitude remains constant (A = 15  nm). Then the z-piezo starts decreasing (extension). (2) When the zc separation is close to 15 nm, the amplitude starts decreasing almost linearly with zc separation and the phase shift starts becoming negative (attractive regime). (3) Eventually, there is a step-like discontinuity in amplitude and phase shift. This indicates a switch from the

High Resolution Imaging of Immunoglobulin G Antibodies

77

Fig. 7. (a) Amplitude and (b) phase distance curve obtained simultaneously with a free amplitude of approximately 15 nm while driving at resonance. The phase shift is given in degrees and, as shown, a value of zero is typically set in the Multimode AFM for the free cantilever at resonance. This allows distinguishing between force regimes by monitoring phase shifts where values above zero (positive) and below zero (negative) correspond to the repulsive and attractive regime, receptively. A switch between force regimes (or more thoroughly oscillation states) is observed in the curves during both extension and retraction.

attractive to the repulsive regime, or more thoroughly, from the Low to the High state. The phase becomes positive at this point. (4) As the z-piezo continues decreasing, the amplitude keeps decreasing linearly but now in the repulsive regime. The phase shift starts becoming negative again but since there is no step-like discontinuity in amplitude and phase shift, this indicates that there is much energy dissipation due to the cantilever being so close to the sample (28, 29) rather than a switch between states. Since the amplitude reduction had been triggered to about 80% prior to acquiring the curve, when the amplitude is close to 3  nm the z-piezo distance stops decreasing (e.g., approaching the surface) and the retraction stage starts. (5) Now, during retraction the cantilever oscillates in the repulsive regime (High state) and the phase shift is positive. (6) Eventually, there is a switch from the repulsive to the attractive regime and there is a step-like discontinuity in both amplitude and phase shift. This ­corresponds to a reverse switch between states (high to low).

78

Santos and Thomson

(7) The cantilever now oscillates in the attractive force regime where the phase shift is negative. Other APD curves taken with A = 12 nm showed no switch between states, that is, the attractive regime had dominated throughout via the Low state, meaning that the critical value of free amplitude to reach the repulsive regime for this system is of approximately 15  nm. Thus, the tip could now be engaged at resonance with a free amplitude of 15 nm and the Asp could be slowly decreased while scanning in order to reach the repulsive regime. Note, however, that in this case the Asp might have to be dramatically decreased below a normalised value of 0.8 (Asp/A << 0), since no switch is observed in the curve until Asp = 6  nm (Asp/A = 0.4) during extension. Thus, in this case, it would be advisable to either shift the driving frequency closer to resonance (e.g., Df = −10 Hz) or slightly increase the free amplitude, for example to 17  nm. Any of the two approaches (or both) could be used to reach the repulsive regime with higher normalised set-points.

Acknowledgements We thank Anthony Maxwell for kindly providing us with monoclonal IgG antibodies against the A-subunit of DNA gyrase. SS is funded through a Doctoral Training Grant of the BBSRC. We acknowledge the University of Leeds for strategic investment in AFM infrastructure. References 1. Binnig, G. et  al. (1986) Atomic Force Microscope. Physical Review Letters 56, 930–933 2. Alessandrini, A. and Facci, P. (2005) AFM: a versatile tool in biophysics. Measurement Science and Technology 16, 65–92 3. Tamayo, J. and Garcia, R. (1996) Deformation, Contact Time, and Phase Contrast in Tapping Mode Scanning Force Microscopy. Langmuir 12 (18), 4430–4435 4. Hansma, H.G. and Hoh, J.H. (1994) Biomolecular Imaging with the Atomic Force Microscope. Annual Review of Biophysics and Biomolecular Structure 23, 115–140 5. Thomson, N.H. et al. (1996) Protein tracking and detection of protein motion using atomic force microscopy. Biophysical Journal 70 (5), 2421–2431

6. Ostendorf, F. et  al. (2008) How flat is an ­air-cleaved mica surface? Nanotechnology 19, 305705–305710 7. Moreno-Herrero, F. et al. (2003) DNA height in atomic force microscopy. Ultramicroscopy 96, 167–174 8. Thomson, N.H. (2005) Imaging the substructure of antibodies with tapping-mode AFM in air: the importance of a water layer on mica. Journal of Microscopy 217, 193–199 9. Garcia, R. and Perez, R. (2002) Dynamic Atomic Force Microscopy Methods. Surface Science Reports 47, 197–301 10. Hansma, P.K. et  al. (1994) Tapping mode atomic force microscopy in liquids. Applied Physics Letters 64, 1738–1740 11. Garcia, R. and San Paulo, A. (2000) Dynamics of a vibrating tip near or in intermittent ­contact

High Resolution Imaging of Immunoglobulin G Antibodies with a surface. Physical Review B 61, 13381–13384 12. San Paulo, A. and Garcia, R. (2000) HighResolution Imaging of Antibodies by TappingMode Atomic Force Microscopy: Attractive and Repulsive Tip-Sample Interaction Regimes. Biophysical Journal 78, 1599–1605 13. Thomson, N.H. (2005) The substructure of immunoglobulin G resolved to 25 kDA using amplitude modulation in air. Ultramicroscopy 105, 1003–1110 14. Abou-Saleh, R.H. et  al. (2009) Nanoscale Probing Reveals that Reduced Stiffness of Clots from Fibrinogen Lacking 42 N-Terminal Bb-Chain Residues Is Due to the Formation of Abnormal Oligomers. Biophysical Journal 96 (6), 2415–2427 15. Quate, C.F. (1994) The AFM as a tool for surface imaging. Surface Science 299/300, 980–995 16. Urry, D.W. (1988) Entropic elastic processes in protein mechanisms. I. Elastic structure due to an inverse temperature transition and elasticity due to internal chain dynamics Journal of Protein Chemistry 7 (1), 1–34 17. Fukuma, T. et al. (2005) True molecular resolution in liquid by frequency modulation atomic force microscopy. Applied Physics Letters 86, 193108–193110 18. Y. Maeda et al. (1999) Observation of singleand double-stranded DNA using non-contact atomic force microscopy. Applied Surface Science 140, 400–405 19. J. Martinez et al. (2005) Length control and sharpening of atomic force microscope carbon nanotube tips assisted by an electron beam. Nanotechnology 16, 2493–2496

79

20. Patil, S. et  al. (2007) Force microscopy ­imaging of individual protein molecules with sub-pico Newton force sensitivity. Journal of Molecular Recognition 20, 516–523 21. Zhang, Y. et  al. (1996) Imaging Biological Structures with the Cryo Atomic Force Microscope. Biophysical Journal 71, 2168–2176 22. Sitong Sheng et  al. (2006) Localization of Linker Histone in Chromatosomes by CryoAtomic Force Microscopy. Biophysical Journal 91 (4), L35–L37 23. Gross, L. et al. (2009) The Chemical Structure of a Molecule Resolved by Atomic Force Microscopy. Science 325, 1110–1114 24. Ostendorf, F. et  al. (2009) Evidence for Potassium Carbonate Crystallites on AirCleaved Mica Surfaces. Langmuir 25 (18), 10764–10767 25. Christenson, H.K. and Israelachvili, J.N. (1987) Growth of Ionic Crystallites on Exposed Surfaces. Journal of colloid and interface science 17, 576–577 26. Christenson, H.K. (1993) Adhesion and surface energy of mica in air and water. Journal of Physical Chemistry 97, 12034–12041 27. Zitzler, L. et al. (2002) Capillary forces in tapping mode atomic force microscopy. Physical Review B 66, 155436–155443 28. Martinez, N. and Garcia, R. (2006) Measuring phase shifts and energy dissipation with amplitude modulation atomic force microscopy. Nanotechnology 17, 167–172 29. Cleveland, J.P. et  al. (1998) Energy ­dissipation in tapping-mode atomic force ­microscopy. Applied Physics Letters 72 (20), 2613–2615

wwwwwwwwwwwwww

Chapter 6 Atomic Force Microscopy of Ex Vivo Amyloid Fibrils Claudio Canale, Annalisa Relini, and Alessandra Gliozzi Abstract Here, we report a study of ex vivo amyloid fibrils formed, respectively, by the Leu174Ser Apolipoprotein A-I (ApoA-I-LS) variant and by b2-microglobulin (b2-m) (Relini et  al., J. Biol. Chem. 281:16521– 16529, 2006; Relini et al., Biochim. Biophys. Acta 1690:33–41, 2004). In the work on ApoA-I-LS, the AFM has been used to characterize and compare the morphologies of amyloid fibrils isolated from two different patients, while in the study on b2-m our investigation provided important information about the factors that can promote the aggregation in vivo. Key words: AFM, Amyloid, Aggregation, Fibrillogenesis

1. Introduction An entire class of pathologies, including Alzheimer’s disease, Parkinson’s disease, spongiform encephalopathy, Huntington Corea, and amyloidoses (1), is associated with the deposition of intracellular or extracellular fibrillar protein aggregates, characterized by a high b-sheet content. The mechanism responsible for the conversion of soluble polypeptide chains into stable, insoluble fibrillar structures termed amyloid fibrils is not completely known. Fibrillogenesis is considered as an alternative pathway to normal protein folding, in which a destabilized polypeptide chain is sequestered by a competing path of aggregation (2). At present, at least 40 different diseases related to the deposition of amyloid or amyloid-like aggregates are known (3); representative examples are listed in Table 1. The AFM has been widely used in the last decade in the study of fibrillogenesis; the very high spatial resolution of this technique and its capability to obtain a three-dimensional image of the sample

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_6, © Springer Science+Business Media, LLC 2011

81

82

Canale, Relini, and Gliozzi

Table 1 A summary of the main diseases related to amyloid/amyloid-like deposition and the proteins or peptides involved therein. For a comprehensive list, see ref. 3 Clinical syndrome

Fibril component

Alzheimer’s disease

Ab peptides (1–40, 1–41, 1–42, 1–43); Tau

Parkinson’s disease

a-synuclein (wild type or mutant)

Spongiform encephalopathies

Prion protein (full-length or fragments)

Hereditary cerebral hemorrhage with amyloidosis

Cystatin C (minus a 10-residue fragment); Ab peptides

Fronto-temporal dementias

Tau (wild type or mutant)

Familial British dementia

ABri peptide

Familial Danish dementia

ADan peptide

Amyotrophic lateral sclerosis

Superoxide dismutase (wild type or mutant)

Cerebellar ataxias

Ataxins (polyQ expansion)

Spinocerebellar ataxia 17

TATA box-binding protein (polyQ expansion)

Dentatorubral-pallidoluysian atrophy

Atrophin 1 (polyQ expansion)

Huntington disease

Huntingtin (polyQ expansion)

Kennedy disease

Androgen receptor (polyQ expansion)

Primary systemic amyloidosis

Ig light chains (full-length or fragments)

Secondary systemic amyloidosis

Serum amyloid A (fragments)

Hemodialysis-related amyloidosis

b2-microglobulin

Familial Mediterranean fever

Serum amyloid A (fragments)

Senile systemic amyloidosis

Transthyretin (wild-type or fragments thereof )

Familial amyloidotic polyneuropathy I

Transthyretin (over 45 variants or fragments thereof)

Familial amyloid polyneuropathy III

Apolipoprotein A-1 (fragments)

Finnish hereditary systemic amyloidosis

Gelsolin (fragments of the mutant protein)

Type II diabetes

Pro-islet amyloid polypeptide (fragments)

Lysozyme systemic amyloidosis

Lysozyme (full-length, mutant)

Insulin-related amyloidosis

Insulin (full-length)

Pulmonary alveolar proteinosis

Lung surfactant protein C

Cataract

g-Crystallins

Pituitary prolactinoma

Prolactin

Medullary carcinoma of the thyroid

Calcitonin

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils

83

surface at the nanometric scale had a fundamental role in the morphological characterization of amyloid fibrils as well as prefibrillar aggregates (4–8). Moreover, the ability of the AFM to work in different environments without needing any particular sample preparation procedures, such as staining, fixation, or metallization, enables the observation of the fibrillization process in different solution conditions and/or in the presence of different cofactors (6, 7, 9). In vitro experiments can provide important information on the mechanisms involved in the aggregation process. On the other hand, the analysis of fibrils extracted from ex vivo pathological deposits is fundamental for characterizing the structural and morphological features of amyloid aggregates and their possible interplay with other biological factors. In spite of this, the number of studies involving in  vitro fibrils is highly dominant over the analysis of ex vivo specimens, likely due to difficulties related to the availability of natural samples. 1.1. Morphological Characterization of Ex Vivo Amyloid Fibrils Formed by the Apolipoprotein A-I Leu174Ser Variant

Apolipoprotein A-I (ApoA-I) is the major protein component of high-density lipoproteins (HDL). ApoA-I is responsible for cholesterol extraction and transport from body tissues. Variants of ApoA-I are involved in some forms of hereditary systemic amyloidosis. In particular, ApoA-I-LS is related to an amyloidosis form with predominant heart involvement; in this study, we examined fibrils isolated from the heart of two patients. Amyloid fibrils of ApoA-I had not yet been characterized by AFM when we started this work, the only data available on their structure had been obtained by small angle X-ray diffraction on the ApoA-I-LS mutant (10) and by electron microscopy crosssectional images on the Leu60Arg ApoA-I mutant (11). Characterization of the ultrastructural assembly of ApoA-I-LS in natural amyloid fibrils provided complementary information on the three-dimensional organization of this type of fibrillar deposits. We compared the morphologies of ApoA-I-LS amyloid material obtained from the two patients and we determined the structural parameters of the aggregates observed both in air and under liquid working in tapping mode. Figure 1 shows two representative areas acquired in air (Fig.  1a) or in liquid (Fig.  1b). The fibrils extracted from the two patients showed common morphologies, the same structural features and compatible sizes (Fig. 2). In addition, we discussed the role of the nonfibrillar aggregates, found to coexist with the fibrils, as potential fibril precursors (12).

1.2. The Role of Collagen in the Aggregation of b2-Microglobulin

The deposition of insoluble fibrils of b2-m in the musculo-skeletal system is characteristic of dialysis-related amyloidosis. The aggregation process of b2-m has been the object of extensive investigation since many years. Several experiments have been designed to reproduce b2-m amyloid deposition in  vitro and understand the

84

Canale, Relini, and Gliozzi

Fig. 1. Tapping mode AFM images of ApoA-I-LS aggregated material obtained in air (a) and in liquid (b). In (b) only short fibrillar aggregates are present, while the elongated fibrils clearly displayed in (a) are absent. This indicates a weak adhesion between elongated fibrils and the mica substrate. Scan size: 3 mm (a), 4 mm (b), Z-range 25 nm (a), and 15 nm (b).

Fig. 2. Height distributions for ApoA-I-LS fibrils extracted from two different patients (black and gray bars). Data were obtained measuring the height of fibril cross sections in the AFM images acquired in air (a) or in liquid (b). The height of the dehydrated fibrils is significantly reduced with respect to fully hydrated conditions, due to water loss and flattening of the fibrils on the mica substrate.

­ olecular conformational changes involved in this process. b2-m m was shown to aggregate in  vitro under acidic conditions hardly compatible with the physiologic ones (13). At neutral pH, b2-m was able to form fibrils in the presence of high concentrations of free metal ions (14) and to elongate preformed fibrils in the pre­ sence of TFE (15), SDS (16), or when the protein populates an intermediate state of the folding pathway (17). Although the information resulting from these in  vitro experiments can contribute to get insight into the aggregation process of b2-m, it cannot be directly related to in vivo fibrillogenesis. Our aim was the understanding of the conditions that promote the in vivo fibrillization of b2-m. In the presence of kidney failure, the concentration of free circulating b2-m can increase by up to 50-fold; the persistent

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils

85

increase in b2-m concentration results in amyloid deposition, preferentially localized in the musculo-skeletal system. The accumulation of b2-m deposits has been shown to cause arthralgias, destructive osteoarthropathies, and carpal tunnel syndrome. Although a high concentration of b2-m is a typical condition at the onset of the disease, there is not a strict correlation between the disease severity and b2-m levels (18), suggesting that other factors might be involved in b2-m amyloid deposition. b2-m is ubiquitously present in the human body, since the peptide is released from every cell expressing the major histocompatibility complex class I (MHCI). On the contrary, one of the most peculiar properties of DRA is its strict specificity for tissues of the skeletal system; in fact, the tissues that are primarily involved are bones and ligaments. Therefore, it might be reasonable to associate the localized protein aggregation and fibrillization to the molecular environment that is present in the skeletal system. The AFM inspection of ex vivo material has been fundamental to reveal the strict association between b2-m fibrils and collagen fibers (Fig. 3). We performed in vitro experiments using conditions as close as possible to the conditions occurring in the periarticular tissues of patients subject to long-term hemodialysis. In the presence of a flogistic process, the local pH can decrease locally to 6.4 (18). The temperature was set to 37°C, the physiological temperature, and to 40°C, in order to mimic the extreme temperature that can be induced by an inflammatory state. We studied the aggregation

Fig. 3. Surface plots of a tapping mode AFM image (height data) of b2-m ex vivo amyloid fibrils surrounding a collagen fiber. The collagen is easily detectable from the characteristic band pattern, with a periodicity of 67 nm. The strict association between amyloid fibrils and collagen suggested an active role of the collagen in the aggregation process of b2-m. This research was originally published by Relini et al. (6).

86

Canale, Relini, and Gliozzi

Fig. 4. Tapping mode AFM images of b2-m amyloid fibrils obtained in vitro at 40°C and pH 6.4 in the presence of collagen. (a) and (c) are topographical images, while (b) and (d) are amplitude images. In (a) and (b), the arrows show a collagen fibril embedded in the surrounding material partially composed of amyloid fibrils. Scan size: 880 nm (a, b), 1.5 mm (c, d), Z-range 12  nm (a), and 10  nm (c). In the absence of collagen, only a low number of globular aggregates and rare ­filamentous structures are found (data not shown). This research was originally published by Relini et al. (6).

process in the absence or in the presence of type I collagen. Aggregation experiments were performed at a protein concentration ranging between 0.2 and 0.6 mg/ml. Fibrillization occurred only in the presence of collagen (Fig. 4). In this work, the analysis of ex vivo materials provided important information, not only on the morphology of amyloid aggregates, but also on the relation between amyloid deposition and collagen. In vitro experiments demonstrated that the latter has an active role in the fibrillogenesis of b2-m.

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils

87

2. Materials 2.1. Materials for ApoA-I Leu174Ser Variant 2.1.1. Preparation of the Substrates for the AFM Analysis

1. Metallic magnetic disks of 12 mm in diameter (provided by the mechanical workshop of the Physics Department at the University of Genoa). 2. Mica sheets (Agar Scientific, purchased from Assing, Monterotondo, Rome). 3. Tape and double-side adhesive tape. 4. Two-component epoxy glue.

2.1.2. Chemicals

1. Amyloid material purification buffer: 10  mM Tris, 1  mM EDTA, 140 mM NaCl, 1.5 mM phenylmethylsulfonyl fluoride (PhMeSO2F), and 0.1% NaN3, pH 8.0. 2. AFM imaging and sample preparation buffer: Tris 50  mM, pH 7.0.

2.1.3. AFM

1. A Dimension 3000 microscope (Digital Instruments-Veeco, Santa Barbara, USA), equipped with a “G” scanning head (maximum scan size 100 mm) and driven by a Nanoscope IIIa controller, was used in all the experiments. 2. V-shaped gold-coated Si3N4 cantilevers (DNP, Veeco, Santa Barbara, USA) with a nominal spring constant of 0.06 N/m and pyramidal tips having a nominal curvature radius of about 40 nm were used for imaging in liquid. 3. Single beam silicon cantilevers (type OMCL-AC160TS, Olympus, Japan) with a nominal spring constant of 40 N/m and typical tip curvature radius around 7 nm were used for imaging in air.

2.2. Materials for Aggregation of b2-Microglobulin

See Subheading 2.1.1.

2.2.1. Preparation of the Substrates for the AFM Analysis 2.2.2. Chemicals

1. 10 mM Tris/EDTA, containing 1.5 M phenylmethylsulfonyl fluoride. 2. Ammonium acetate 50 mM, pH 7.4.

2.2.3. AFM

1. AFM images were acquired using a Multimode Scanning Probe microscope (Digital Instruments-Veeco, Santa Barbara, USA) equipped with an “E” scanning head (maximum scan size 10 mm) and driven by a Nanoscope IV controller. For larger scan sizes,

88

Canale, Relini, and Gliozzi

a Dimension 3000 microscope (Digital Instruments-Veeco), equipped as described in Subheading 2.1.3 was employed. 2. The same AFM cantilevers Subheading 2.1.3 were used.

described

previously

in

3. Methods 3.1. Methods for ApoA-I Leu174Ser Variant 3.1.1. Purification of the Ex Vivo Material

3.1.2. AFM Sample Preparation

The ex vivo material was purified following the methods developed by Pras et  al. (19). Fibrils were extracted from the explanted hearts of two patients affected by cardiac amyloidosis. Approximately 800  mg of tissue was homogenized and centrifuged at 60,000 × g in 2  ml of buffer containing 10  mM Tris, 1  mM EDTA, 140  mM NaCl, 1.5  mM phenylmethylsulfonyl fluoride (PhMeSO2F), and 0.1% NaN3, pH 8.0. The procedure that allowed the clearing of the non-amyloid material was stopped when the absorbance at 280  nm of the supernatant dropped under 0.1 units. The buffer was replaced by distilled water, and homogenization and centrifugation were carried out six times. Amyloid material was collected from the supernatant and the protein concentration monitored by Congo red staining and thioflavin assay. See Obici et al. (11) for further details. 1. The Dimension 3000 sample holder is equipped with a magnet, for this reason it is useful to use magnetic metallic disks to fix the sample. 2. Cut a piece of mica of about 1 × 1 cm2 using a pair of scissors. 3. Mica substrates must be attached to metallic disks. To perform imaging in air of dried samples, mica substrates can be attached before or after sample deposition and dehydration by means of double-side adhesive tape. This procedure is not suitable for liquid imaging, as usually the tape tends to detach from the metallic disk if in contact with an aqueous solution. The most convenient method to attach mica onto the metallic disks is to use a two-component epoxy glue. This is probably the best choice also for the preparation of dried samples; the only inconvenient is the long time required to dry the glue, generally more than 1  h, depending on the kind of glue. However, it is possible to prepare several substrates in advance and then to cleave mica just before sample deposition. 4. Sample deposition: the sample was diluted to a concentration of 50 mg/ml, a small aliquot of 20 ml was deposited on freshly cleaved mica. For imaging in air, the sample was dried under mild vacuum for 30 min (Note 1); for imaging in liquid, the sample was incubated on mica for 30  min and then rinsed with the same buffer to eliminate the material which had not adhered to the substrate (Note 2).

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils 3.1.3. AFM Operation

89

1. Mount the cantilever in the AFM holder as suggested in the user manual of the AFM. 2. Mount the holder in the AFM head and start the laser alignment. This operation is generally described in the AFM user manuals as well. The basic concepts are the same for all AFMs, but setups produced by different companies may adopt different procedures for laser alignment. When the laser is reflecting on the back of the cantilever and the laser spot is centered on the photodiode, the sum signal increases to a value that depends on the kind of cantilevers used. All the measurements described here were obtained using cantilevers that were metal-coated on the back side, in order to increase reflectivity; with our setup the sum was around 5–6 V. 3. When working in liquid with very soft metal-coated cantilevers (0.03–0.4 N/m), a drift of the laser spot on the photodetector can be clearly displayed; the vertical deflection changes with time due to cantilever bending. This pheno­ menon is due to the different thermal expansion coefficients of the two materials that are in contact. Wait until the cantilever reaches the thermal equilibrium with the environment or, alternatively, use uncoated silicon cantilevers. 4. Before engagement, the cantilever resonance frequency (fo) must be determined. Some of the AFM systems, as for example the Veeco systems, have an automatic procedure to determine fo and the working amplitude of oscillation (amplitude setpoint, A). Otherwise, fo and A have to be chosen manually, following the AFM manual instructions. 5. When working in tapping mode, a fundamental parameter is the amplitude setpoint A; lower setpoints correspond to a stronger interaction between the tip and the sample. Generally A is chosen equal to the 70–80% of the free oscillation amplitude Ao. A high interaction force between the AFM probe and the sample may damage the AFM tip or the sample itself, especially if working on delicate biological materials. To engage softly on the sample, it is recommended to use high amplitude setpoints, more than 80% of Ao. In this case, it might happen that the tip is not really engaged to the surface. There are essentially two ways to optimize the tip engagement: –– Looking at the scope mode, decrease gradually the amplitude setpoint during imaging until the surface profile can be clearly seen. –– In liquid, amplitude vs. z-piezo displacement curve can help in determining the setpoint (Fig.  5); to know more about this procedure, see Lantz et al. (20). In the case of stiff silicon cantilevers used for tapping mode in air, this operation is not recommended as the strong interaction between tip and surface can easily damage the tip.

90

Canale, Relini, and Gliozzi

Fig. 5. An amplitude vs. z-piezo displacement curve can be helpful in the determination of the correct setpoint amplitude A. When working on soft biological samples, A must be chosen as close as possible to the free amplitude of oscillation of the cantilever Ao. Soft tapping conditions correspond to a setpoint ratio A /Ao close to unit; A /Ao has a value around 0.85 in this example. The setpoint ratio can be reduced to 0.5 or less (hard tapping), but the strong interaction between tip and sample can damage the sample or the tip itself.

6. Set the scan size to zero before the engagement in order to reduce the risk of contamination while the tip is approaching the surface. 7. After the tip engagement start to scan large areas (10–15 mm) to have a large scale view of the sample and check the fibrils spatial distribution. The scan rate must be chosen depending on the scan area, larger scans require a slower scan rate to have a tip velocity between 0.5 and 5 mm/s. 8. The gain must be maintained as high as possible in order to have good quality images. 9. When studying fibril morphology, several factors contribute to define the apparent fibril size and shape (Note 3). In particular, the measured fibril width (Note 4) is much larger than the real fibril size as a result of broadening effects due to the AFM tip size. Assuming that the cross section is circular, the apparent width (w) is related to the height (h) according to the equation:

w = 2 2Rh ,

(1)

where R is the radius of curvature of the tip. The radii of curvature of the tips were determined by imaging latex calibration spheres (94 nm in diameter) dried on to a freshly cleaved mica substrate (21) (Note 5).

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils

3.2. Methods for Aggregation of b2-Microglobulin

3.2.1. Purification of the Ex Vivo Material and In Vitro Fibrillization

91

1. Amyloid fibrils were isolated from the femoral head of a patient. The isolation procedure was very similar to that used for the extraction of ApoA-I-LS described in Subheading 3.1.1, details are reported in Relini et al. (6). 2. Recombinant b2-m was expressed and purified according to the procedure reported by Esposito et al. (22). The protein was dehydrated after purification. 3. When lyophilized b2-m was dissolved in a buffer at pH 6.4 at room temperature, some tangles of fibrils were observed. To avoid the presence of aggregated material, the protein was dissolved in ammonium acetate 50 mM at pH 7.4 and then the pH was adjusted to 6.4 just at the beginning of the experiment. The protein concentration was 2  mg/ml, and the solution was centrifuged at 16,500 × g for 1 h at the temperature of 4°C in order to remove large aggregates. The supernatant was collected and filtered with 20-nm pore size filters (Anotop 10, cat. no. 6809-1002, Whatman, USA). Aggregation experiments were performed at a protein concentration of 0.2–0.6 mg/ml, after acidification to pH 6.4. 4. Type I collagen was purified from calf skin (23). SDS–PAGE was used to check the collagen purity. Fibrillar collagen was prepared by solubilizing the purified collagen in 5 mM acetic acid. The solution was incubated at 37°C for 30 min after a dilution 1:1 with 2× phosphate-buffered saline. 5. Fibrillar collagen was sonicated for 10 min in a bath sonicator (ACAD, Genoa, Italy) and washed with the ammonium acetate buffer. Then, it was put on a microscope slide, it was cut into several small pieces that were washed again and added to the protein solution (Note 6).

3.2.2. AFM Sample Preparation

3.2.3. AFM Operation

1. Subheadings 1–3 of paragraph 3.1 describe general procedures that have been used also in the preparation of b2-m samples. A different method has been used to deposit the sample in the presence of fibrillar collagen. In this case, a piece of collagen was collected from the solution and deposited onto a freshly cleaved mica substrate together with its surrounding protein solution (Note 7). The sample was dried under mild vacuum and imaged in tapping mode in air. As the AFM settings have been already described in Part I, the present paragraph must be intended as complementary to Subheading 3.1.3. 1. Special attention must be taken when engaging on the b2-m/ collagen samples. The piece of collagen dried on the mica substrate forms a relatively thick layer of very rough material. It is not possible to obtain nanometric resolution if the AFM

92

Canale, Relini, and Gliozzi

probe is engaged in the center of the collagen network. To achieve a resolution allowing amyloid fibrils detection, the tip should be engaged close to the collagen edge, but in a flat area. This operation can be performed easily using the optics that is integrated in the Dimension 3000 system or using the top view optics of the Multimode system (Note 8). 3.3. Conclusions

The AFM technique has been successfully applied to the study of ex vivo amyloid material. We had the rare opportunity to analyze amyloid material isolated by two patients affected by hereditary cardiac amyloidosis. The morphological characterization of two unrelated samples of ApoA-I amyloid aggregates demonstrated that aggregate sizes and shapes are quite comparable. Our AFM inspections show that not only mature fibrils, but also the prefibrillar species display the same morphological features. The analysis of ex vivo material extracted from the femoral head of a patient revealed the strict association between b2-m fibrils and collagen fibers. The interesting correlation between amyloid fibrils and collagen has been a starting point for a series of in  vitro aggregation experiments. Following the hypothesis, suggested by the ex vivo inspection, of an active role of collagen in the aggregation process of b2-m, we have been able to obtain b2-m fibrillation in vitro under conditions which are very close to the physiopathological ones.

4. Notes 1. Amyloid proteins usually have a high tendency to aggregate. In this case, the drying procedure could induce an increase of the aggregation rate due to the increased protein concentration as consequence of solvent evaporation. Therefore, it is important to dry the sample as quickly as possible, either using a low vacuum chamber or a dry nitrogen flow. 2. One of the most important steps for the morphological and structural characterization of a heterogeneous sample by AFM is the sample deposition procedure. During this operation, part of the material composing the sample may be lost or artifacts could be generated, precluding the visualization of important components of the specimen. When working in liquid, the materials that are adsorbed on the substrate are the only ones that can be observed by AFM, the part of the sample weakly adhering to the substrate is removed by the AFM tip during scan. In the case of ApoAI-LS, the globular aggregates that were present in the solution adsorbed onto mica and a homogeneous distribution of

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils

93

this material was clearly displayed (Fig. 1b), while only short fibrillar aggregates (200–250  nm in length) were observed (Fig. 1b); longer fibrils, such as those imaged in air (Fig. 1a), did not adhere to the substrate. In principle, using substrates with different characteristics (electric charge, hydrophilicity, etc.), should allow observing different species that are present in the sample, thus obtaining important information on the chemical-physical characteristics of these species. The problem of poor sample adhesion to the substrate can be overcome working in air onto dehydrated samples, therefore avoiding the loss of material. Moreover, when using a buffer the dehydration procedure gives rise to salt crystals that can completely cover the sample masking its features. In this case, it is preferable to incubate the sample on the mica substrate for some minutes and then rinse it gently with Milli-Q water. 3. Shape and size deformations induced by the dehydration process must be taken into account when measuring the size of a dried biological sample. The evaporation of water can induce a significant shrinking in a material with high water content. In addition, the drying procedure can induce a flattening of the sample on the mica surface. The decrease of fibril thickness is clearly shown in Fig.  2 for ApoA-I-LS amyloid aggregates. This result was in agreement with previous data obtained on amyloid aggregates formed by another peptide (24). 4. Fibril height and width must be measured in individual fibrils from the image sections taken perpendicularly to the fibril longitudinal axis (Fig. 6). 5. Any material with a well-known size and geometrical shape can be used to determine the AFM tip shape and size. In our work, latex spheres have been used. It is possible to use other kinds of calibration standards, for those commercially available see the brief overview provided by Peter Markiewicz in 1999 (25), showing all the most common techniques that are used for AFM tip characterization. 6. Fibrillar collagen is very sticky; all the operations described at point 5 of Subheading 3.2.1 must be performed with particular care. In our case, we used Teflon tweezers to manipulate the sample. We prepared at the same time different vials containing b2-m in the conditions described in Subheading 3.2.1 and we added a single piece of collagen per vial. 7. For the AFM inspection of the sample, the collagen must be transferred from the solution to the mica surface. This operation was also performed by using Teflon tweezers. 8. A great attention to sample size must be paid using a Multimode AFM head. Due to its particular head design, this instrument can be used only for the study of small samples.

94

Canale, Relini, and Gliozzi

Fig. 6. An example of fibril cross section. The cross section was performed along the line in the AFM image. Fibril height is measured as the vertical distance between the two points indicated by the arrows in the cross section, corresponding in this case to a value of 2.8 nm.

Acknowledgment We thank Vittorio Bellotti and Giampaolo Merlini (Department of Biochemistry, University of Pavia -Laboratori di Biotecnologie IRCCS Policlinico San Matteo, 27100 Pavia, Italy) for extraction and purification of Apo-A-I and beta2-m ex-vivo amyloid fibrils and expression and purification of recombinant beta2-m. References 1. Dobson, C.M. (2001) Protein folding and its links with human disease. Biochem. Soc. Symp. 68, 1–26. 2. Booth, D. R., Sunde, M., Bellotti, V., Robinson, C. V., Hutchinson, W. L., Fraser, P. E. (1997) Instability, unfolding and aggregation of human lysozyme variants underlying amyloid fibrillogenesis. Nature 385, 787–793. 3. Chiti, F. and Dobson, C.M. (2006) Protein misfolding, functional amyloid, and human disease. Annu Rev Biochem 75, 333–366.

4. Blackley, H. K. L., Sanders, G. H. W., Davies, M. C., Roberts, C. J., Tendler, S. J. B. and Wilkinson, M. J. (2000) In-situ Atomic Force Microscopy study of b-Amyloid fibrillization. J.Mol.Biol. 298, 833–840. 5. Chamberlain, A.K., MacPhee, C.E., Zurdo, J., Morozova-Roche, L., Hill, H.A.O., Dobson, J.J. and Davis, J.J. (2000) Ultrastructural organization of amyloid fibrils by atomic force microscopy. Biophys. J. 79, 3282–3293.

Atomic Force Microscopy of Ex Vivo Amyloid Fibrils 6. Relini, A., Canale, C., De Stefano, S., Rolandi, R., Giorgetti, S., Stoppini, M., Rossi, A., Fogolari, F., Corazza, A., Esposito, G., Gliozzi, A. and Bellotti, V. (2006) Collagen plays an active role in the aggregation of b2microglobulin under physiopathological conditions of dialysis-related amyloidosis. J. Biol. Chem. 281, 16521–16529. 7. Arimon, M., Dıez-Perez, I., Kogan, M. J., Durany, N., Giralt, E., Sanz, F. and FernandezBusquets, X. (2005) Fine structure study of Ab1-42 fibrillogenesis with atomic force microscopy. FASEB J. 19, 1344–1346. 8. Natalello, A., Prokorov, V. V., Tagliavini, F., Morbin, M., Forloni, G., Beeg, M., Manzoni, C., Colombo, L., Gobbi, M., Salmona, M. and Doglia, S. M. (2008) Conformational plasticity of the Gerstmann–Sträussler– Scheinker disease peptide as indicated by its multiple aggregation pathways. J. Mol. Biol. 381, 1349–1361. 9. Ha, C., Ryu, J. and Park, C.B. (2007) Metal ions differentially influence the aggregation and deposition of Alzheimer’s b-amyloid on a solid template. Biochemistry 46, 6118–6125. 10. Mangione, P., Sunde, M., Giorgetti, S., Stoppini, M., Esposito, G., Gianelli, L., Obici, L., Asti, L., Andreola, A., Viglino, P., Merlini, G., Bellotti, V. (2001) Amyloid fibrils derived from the apolipoprotein A-I Leu174Ser variant contain elements of ordered helical structure. Protein Sci. 10, 187–199. 11. Obici, L., Bellotti, V., Mangione, P., Stoppini, M., Arbustini, E., Verga, L., Zorzoli, I., Anesi, E., Zanotti, G., Campana, C., Vigano, M., Merlini, G. (1999) The new apolipoprotein A-I variant Leu174Ser causes hereditary cardiac amyloidosis, and the amyloid fibrils are constituted by the 93-residue N-terminal polypeptide. Am. J. Pathol. 155, 695–702. 12. Relini, A., Rolandi, R., Bolognesi, M., Aboudan, M., Merlini, G., Bellotti, V., Gliozzi, A. (2004) Ultrastructural organization of ex vivo amyloid fibrils formed by the apolipoprotein A-I Leu174Ser variant: an atomic force microscopy study. Biochim. Biophys. Acta 1690, 33–41. 13. Kad, N.M., Myers, S.L., Smith, D.P., Smith, D.A., Radford, S.E. and Thomson, N.H. (2003) Hierarchical assembly of b2-microglobulin amyloid in vitro revealed by atomic force microscopy. J. Mol. Biol. 330, 785–797. 14. Morgan, C. J., Gelfand, M., Atreya, C., and Miranker, A. D. (2001) Kidney dialysis-associated amyloidosis: a molecular role for copper in fiber formation. J. Mol. Biol. 309, 339–345.

95

15. Yamamoto, S, Gejyo, F. (2005) Historical background and clinical treatment of dialysisrelated amyloidosis. Biochim. Biophys. Acta 1753, 4–10. 16. Kihara, M., Chatani, E., Sakai, M., Hasegawa, K., Naiki, H., and Goto, Y. (2005) Seedingdependent maturation of beta2-microglobulin amyloid fibrils at neutral pH. J. Biol. Chem. 280, 12012–12018. 17. Chiti, F., De Lorenzi, E., Grossi, S., Mangione, P., Giorgetti, S., Caccialanza, G., Dobson, C. M., Merlini, G., Ramponi, G., and Bellotti, V. (2001) A partially structured species of beta 2-microglobulin is significantly populated under physiological conditions and involved in fibrillogenesis, J. Biol. Chem. 276, 46714–46721. 18. Floege,J, and Ehlerding, G. (1996) Beta-2microglobulin-associated amyloidosis. Nephron. 72, 9–26. 19. Pras, M., Schubert, D., Zucker-Franklin, D., Rimon, A., Franklin, A.C. (1968) The characterization of soluble amyloid prepared in water. J. Clin. Invest. 47, 924–933. 20. Lantz, M., Liu, Y. Z., Cui, X. D., Tokumoto, H., and Lindsay, S. M., (1999) Dynamic force miscroscopy in fluid. Surf. Interface Anal. 27, 354–360. 21. Ramirez-Aguilar, K.A., Rowlen, K.L. (1998) Tip characterization from AFM images of nanometric spherical particles, Langmuir 14, 2562–2566. 22. Esposito, G., Michelutti, R., Verdone, G., Viglino, P., Hernandez, H., Robinson, C.V., Amoresano, A., Dal Piaz, F., Monti, M., Pucci, P., Mangione, P., Stoppini, M., Merlini, G., Ferri, G., and Bellotti, V. (2000) Removal of the N-terminal hexapeptide from human b2-microglobulin facilitates protein aggregation an fibril formation. Protein Sci 9, 831–845. 23. Giorgetti, S., Rossi, A., Mangione ,P., Raimondi, S., Marini, S., Stoppini, M., Corazza, A., Viglino, P., Esposito, G., Cetta, G., Merlini, G., Bellotti, V. (2005) b2-Microglobulin isoforms display an heterogeneous affinity for type I collagen. Protein Sci. 14, 696–702. 24. Grégoire, C., Marco, S., Thimonier, J., Duplan, L., Laurine, E., Chauvin, J.-P., Michel, B., Peyrot, V., Verdier, J.-M. (2001) Three-dimensional structure of lithostathine protofibril, a protein involved in Alzheimer’s disease. EMBO J. 20, 3313–3321. 25. Markiewicz, P. http:// www.weizmann.ac.il/ Chemical_Research_Support/surflab/peter/ standard/index.htm.

wwwwwwwwwwwwww

Chapter 7 Studying Collagen Self-Assembly by Time-Lapse High-Resolution Atomic Force Microscopy Clemens M. Franz and Daniel J. Muller Abstract Fibrillar collagens constitute a main component of many tissues, where they form a scaffold for cell attachment and provide mechanical strength. Gaining insight into molecular mechanisms of collagen self-assembly from in vitro experiments is important for better understanding the complex hierarchical processes involved in collagen fibril formation in vivo. In addition, such insight can be used to assemble collagen into desirable structures for the biofunctionalization of surfaces in different biotechnological and medical applications. Here, we describe a method to direct the assembly of type I collagen into welldefined nanoscopic matrices of different patterns. Within these matrices, the self-assembly of collagen molecules into fibrils can be directly observed by time-lapse atomic force microscopy (AFM). Highresolution AFM topographs reveal substructural details of the collagen fibril architecture and provide information about mechanisms and dynamics of fibril formation. Key words: Collagen, Self-assembly, Atomic force microscopy (AFM), Biofunctionalization, Fibrillogenesis, Topography

1. Introduction We have recently developed a method to create ultrathin (»3 nm) collagen matrices composed of highly ordered fibrils assembled from collagen type I molecules (1–3). Fine-tuning of the preparation conditions provides precise control over the self-assembly of type I collagen molecules into fibrils displaying different spacing, width, and structural properties. As a result, the structure of the assembled matrix is highly reproducible and defined down to the nanometer scale (3). From the information contained in ­high-resolution atomic force microscopy (AFM) topographs, the ultrastructure of self-assembled collagen fibrils can be correlated with that of collagen fibrils assembled in vivo. Because with AFM Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_7, © Springer Science+Business Media, LLC 2011

97

98

Franz and Muller

biological samples can be imaged in a fully hydrated state and under physiological conditions (4, 5), collagen fibril formation can be directly monitored using continuous imaging. For instance, time-lapse and high-resolution AFM topographs reveal that ­collagen fibrils grow in lateral steps by incorporating molecular building blocks corresponding to the size of four to five collagen molecules (6). Furthermore, using the AFM stylus as a nanoscopic stylus, it is possible to laterally realign the collagen matrix fibrils into freely designable structures ranging from the nanometer to the micrometer scale (1). Recently, we have also used these ordered collagen matrices as structurally and biochemically defined cellculture substrates to study cell adhesion and migration (7–12).

2. Materials 2.1. Collagen Matrices 2.1.1. Mica Supports

1. Muscovite mica supports: highest-quality (grade V1) ruby muscovite mica sheets 3  cm × 5  cm or larger (http://www. grmica.com). The polygon-shaped sheets should be clear, transparent and flat with a thickness of less than 0.3 mm. 2. TruPunch punch-and-die set (http://www.precisoinbrand. com) to produce mica disks ranging from 3 to 10  mm in diameter. Rubber hammer for punching and a solid wooden board as a support for the die set. 3. Adhesive tape (width 12 mm) for mica cleaving (http://www. timemed.com). 4. Two-component epoxy resin (http://www.r-g.de) or Dymax OP-29 optical adhesive (http://www.dymax.com) to attach mica disks to a support. 5. UV illumination system for curing the optical adhesive, for instance a UV-A dual diode (http://www.roithner-laser. com). Alternatively, a suitable UV light source is frequently provided by agarose gel documentation stations with UV-capability.

2.1.2. Collagen Assembly

1. All chemical are of analytical grade and from Sigma–Aldrich (http://www.sigmaaldrich.com) unless stated otherwise (see Note 1). 2. PureCol bovine dermal collagen type I solution at 3 mg/ml in 0.01  N HCl (pH 2) (http://www.purecol.nu/purecol. html). Store at 4°C (see Note 2). 3. Phosphate-buffered saline (PBS) without Ca2+ and Mg2+ (http://www.invitrogen.com). 4. Extracellular buffer: 109.2 mM NaCl, 4.1 mM KCl, 1.7 mM CaCl2, 0.65  mM MgCl2, 7.9  mM monosodium glutamate, 0.4  mM NaH2PO4, 0.3  mM Na2HPO4, 27  mM NaHCO3,

Studying Collagen Self-Assembly

99

20  mM Hepes, pH 7.4, adjusted with NaOH. This buffer resembles the extracellular environment in which collagen fibrils form in vivo. 5. Standard collagen buffer: 50  mM l-glycine, and 200  mM KCl. Adjust to pH 9.2 with 1 M NaOH. The pH of this buffer is close to 9.3, the isoelectric point (pI) of collagen type I (13). 6. Glass cover slips, slides, or standard tissue culture dishes (∅ 35 mm) as supports for the mica disks. 7. A humidified chamber to prevent drying of the collagen supports, for instance a plastic tissue culture dish (∅ 150 mm), wet 3MM Whatman blotting paper (http:// www.whatman. com) and Parafilm “M” laboratory film for air-tight sealing of the chamber. 2.1.3. Collagen Matrix Cross-Linking

1. Glutaraldehyde solution (2%) in PBS, store aliquots at −20°C. 2. Glutaraldehyde quenching solution: 1 mg/ml sodium borohydride solution, prepare fresh immediately before use. 3. Photosensitizer: 0.5% riboflavin solution (http:// www. jenapharm.de). 4. UV irradiation: UV-A dual diode (http:// www.roithner-laser. com) and UV-A meter (LaserMate-Q, http:// www.coherent. com).

2.2. Atomic Force Microscopy

1. An AFM equipped with a fluid cell for contact and tapping mode imaging in solution and piezoelectric scanners with a scan range of 100  mm2 × 100  mm2, for instance a Veeco MultiMode (http:// www.veeco.com) or a JPK NanoWizard (http:// www.jpk.com) (see Note 3). 2. Active vibration isolation table, such as a TS 300-LT, rested on a stable support frame (http:// www.hwlscientific.com). 3. AFM cantilevers: oxide-sharpened Si3N4 cantilevers with nominal force constants of 0.09  N/m (OMCL TR400PS, http:// www.olympus.com) or 0.06  N/m (NP-S, http:// www.veeco.com). 4. The effect of temperature on collagen assembly can be investigated using a temperature-controlled sample holder, such as a JPK BioCell (http:// www.jpk.com).

3. Methods On muscovite mica surfaces, the collagen self-assembly process can be directed to provide structurally well-defined two-dimensional (2D) matrices of highly aligned type I collagen molecules (Fig. 1a) (2). In this procedure, a drop of an appropriate buffer solution is

100

Franz and Muller

Fig. 1. High-resolution AFM topographs showing the directed self-assembly of nanoscopic collagen matrices on muscovite mica surfaces. Depending on the buffer conditions used during collagen assembly, collagen fibrils are either nonperiodic (b) or display the characteristic D-periodicity of about 67  nm (a). The vertical scale of the AFM topographs corresponds to 4 nm.

first placed onto a freshly cleaved mica disk (14). Monomeric collagen molecules solubilized in an ice-cold, low pH solution are then injected into the buffer covering the mica ­surface at RT. The raise in temperature and pH initiates the adsorption and selfassembly of fibrillar collagen structures on the supporting mica surface. The self-assembled collagen layer exhibits a thickness of »3 nm and can span up to several square centimeters. Depending on the collagen concentration and the temperature of the buffer solution, the assembly time of the collagen matrix ranges from minutes to a few hours (3, 6). The composition of the buffer solution plays a crucial role in guiding the self-assembly process of the collagen molecules. By varying pH, electrolyte composition and/or concentration, the collagen self-assembly process (2, 3) can be directed. For instance, the matrix density, governed by both interfibrillar distance and the fibril width, can be adjusted by varying the pH, while inclusion of K+-ions in the buffer leads to the establishment of the 67  nm D-band periodicity (Fig.  1b), a characteristic structural feature observed in native collagen fibrils (11). Within the first 5 h after assembling the collagen matrix on a mica surface, single fibrils can be realigned in a controlled manner using the AFM tip as a nanostylus (Fig.  2) (1, 3). In contrast, mature collagen matrices (>5  h) are structurally and biochemically stable, and rescanning these matrices after months of storage in buffer at room temperature (RT) reveals no changes in fibril

Studying Collagen Self-Assembly

101

Fig. 2. Reshaping the collagen matrix using the AFM tip as a nanoscribe. (a) Schematic depiction showing how the AFM stylus can be applied to realign collagen fibrils within a rectangular region of the matrix. (b) The nano-manipulated ­collagen fibrils are aligned perpendicular to the surrounding fibrils. (c) Several fibrils realigned in a horizontal direction. Adapted from ref. 1.

orientation or mechanical stability. However, when used as ­cell-culture substrates for adhesion and migration studies, cellular contraction forces frequently cause matrix remodeling, sometimes leading to complete matrix destruction (12). In this case, matrix stability can be increased by chemical cross-linking the collagen fibrils. Cross-linked matrices retain their biological function, preserve structural features, such as the D-banding, and can be used as stable cell-culture substrates. Considerable structural information about collagen fibril architecture can be obtained from AFM topographs of fully matured unfixed or fixed matrices. However, to gain insight into dynamic aspects of the collagen self-assembly process, time-lapse AFM imaging has to be employed (Fig. 3) (6). Using standard buffer conditions, the self-assembly of collagen molecules into an

102

Franz and Muller

Fig. 3. Self-assembly of collagen fibrils observed by high-resolution time-lapse AFM. Topographs were recorded continuously at a rate of 5.1 min per frame in standard buffer solution (50 mM glycine, 200 mM KCl, pH 9.2) containing a final collagen concentration of 6 mg/ml. Individual images recorded at the indicated times from the moment of collagen seeding were extracted from a time-lapse series. AFM topographs show laterally and longitudinally growing fibrils. The characteristic D-band is clearly resolved with molecular detail. The full color scale of the image corresponds to a height of 5 nm. AFM scanning performed by Dr. David Cisneros.

ordered 2D matrix is usually completed within a few minutes. As commercially available AFMs typically require several minutes (<2–6) to obtain one high-resolution topograph of a fragile biological sample, these fast processes cannot easily be resolved by time-lapse AFM. Thus, observation of the molecular growth steps

Studying Collagen Self-Assembly

103

of collagen self-assembly requires a reduced collagen ­concentration to slow down the speed of matrix assembly. 3.1. Assembling Collagen Matrices 3.1.1. Preparing Mica Supports

1. Check mica sheets for bends, cracks, or impurities. Punch disks out of clear and structurally intact regions of the mica sheet using a rubber hammer and a punch-and-die set sitting on top of a solid wooden support. Depending on the desired size of the collagen support, the diameter of the mica disk can be varied between about 3 and 10  mm. Mica disks with a diameter of about 6.5 mm have proven useful for a number of applications. Recheck the mica disks and only keep intact, unsplit, and nonflaking disks with smooth edges. 2. Glue mica disks onto dust-free glass cover slips, glass microscope slides, or into plastic tissue culture dishes using a twocomponent epoxy resin. Ensure that the mica disks are level with the support before the resin hardens. 3. Alternatively, when the optical properties of the collagen substrates matter, for instance when they are used in cell migration studies using light microscopy, an optical adhesive (Dymax OP-27) can be used. In this case, the adhesive is cured for a minimum of 5 min using a UV light source. 4. After the glue has solidified and immediately prior to collagen coating, cleave the mica disk by first firmly attaching and then quickly removing an adhesive tape. Check the cleaved mica surface on the support as well as on the adhesive tape for steps or faults in the crystal structure. Only continue if perfectly flat, homogeneous mica surfaces have been exposed.

3.1.2. Collagen Self-Assembly

1. Immediately after mica cleaving, add standard collagen buffer to the mica surface at RT. Use 30 ml of the collagen buffer for a 6.5-mm disk. For other disk diameters, adjust the buffer volume accordingly. The buffer must cover the disk completely without spilling over and running off the disk. 2. If collagen matrices with different fibrillar arrangement are desired, exchange standard collagen buffer with others, such as extracellular buffer. 3. To produce collagen fibrils lacking the characteristic D-periodicity (see Fig.  1b), use a buffer containing no K+ions, such as PBS. 4. Use a micropipette to inject type I collagen stock solution to a final collagen concentration of 100 mg/ml. After collagen addition, briefly mix the solution by carefully pipetting up and down several times without disturbing the overall buffer droplet. 5. Incubate the disks overnight (16  h) at RT in a humidified chamber. A 150-mm plastic tissue culture dish containing

104

Franz and Muller

small pieces of wet filter paper provides a simple chamber accommodating several collagen substrates. The rim of the dish should be sealed using Parafilm to prevent the samples from drying out. 6. On the next day, remove unadsorbed collagen by washing the mica disks at least five times with 50  ml PBS each using a micropipette. The collagen substrate should not run dry during the wash steps, and the micropipette should not touch the collagen surface. 7. After washing, overlay the collagen-coated disks with PBS and use them either immediately or within 1 week. 8. For time-lapse imaging of collagen fibrillogenesis, prepare the mica support as described in standard collagen buffer but inject collagen at a reduced final concentration of 6–20 mg/ml and insert the support into the AFM sample chamber for immediate imaging (see Subheading 3.2.2). 3.2. AFM Imaging 3.2.1. Imaging Mature Collagen Matrices

1. Insert the matured collagen matrix into an AFM equipped with a fluid cell. After thermal equilibration, imaging can be performed in contact or tapping mode at RT. For contactmode imaging of unfixed collagen samples, the scanning force should be below <500  pN to avoid sample distortion. For reproducible high-resolution scan results, use low scan speeds (0.3–1 Hz) and constantly adjust the applied force to compensate for thermal drift (see Note 4). 2. Tapping mode imaging is performed with the drive amplitude of the cantilever set to a root mean square (RMS) of approximately 25 nm and a drive frequency close to the resonance frequency (8–9.5 kHz) of the cantilevers mounted to the fluid cell immersed in buffer solution.

3.2.2. High-Resolution Time-Lapse Imaging of Nascent Collagen Fibrils

1. For time-lapse imaging, the sample is inserted into the fluid cell of the AFM immediately after addition of collagen monomers (see Note 5). 2. Continuous tapping mode imaging is started immediately using the same settings outlined in Subheading 3.2.1. Depending on the collagen concentration, fibrils appear in the AFM topographs instantaneously or within 20–30 min (Fig. 3). 3. To resolve fast processes during fibril formation, the perframe scan time can be shortened by increasing the scan speed and by decreasing the scan size. However, scan speed must be sufficiently low to reliably trace the fibril contours. Fast lateral fibril growth can be monitored at comparatively low scan speeds by limiting the scan size to few or even a single line perpendicular to the fibril direction and extending the line pixel number to the maximum setting offered by the AFM software (2,048 or 4,096 pixel/line).

Studying Collagen Self-Assembly

105

4. Complete surface coverage with collagen may take between 1 and 5 h. Uninterrupted scanning for several hours may place a considerable demand on the AFM operator and on equipment access. 3.3. Reshaping the Collagen Matrix Using the AFM Tip as a Nanoscribe

1. Matrices are first imaged in tapping mode. The maximum scanning force in this mode (<100 pN) is sufficiently small to prevent reversible or irreversible structural changes to the fibril arrangement. 2. Subsequently, predefined regions of the matrix are manipulated using AFM contact-mode scanning at elevated scanning forces of around 300–500 pN. The nanoscale radius of the AFM tip ensures that only individual fibrils are rearranged at a given time. Subsequent high-resolution reimaging of the manipulated region in tapping mode reveals nanoscale matrix restructuring (Fig.  2b, c). Individual matrix fibrils can be realigned within the first 5 h after the self-assembly process. After this setting period, fibrils cannot be reoriented even at scanning forces in excess of 500 pN. Higher scanning forces may lead to overall matrix destruction rather than controlled fibril reorientation.

3.4. Chemically Cross-Linking the Collagen Matrix to Increase Its Stability

1. Add 30 ml glutaraldehyde solution to the collagen matrix for 2 min at RT. 2. Quench unreacted aldehyde groups by two washes with freshly prepared 1 mg/ml sodium borohydride solution. 3. The use of glutaraldehyde may be undesirable if the matrix is used as a cell-culture substrate. In this case, the matrices can be cross-linked using simultaneous UV-A/riboflavin treatment that is known to induce singlet oxygen. After the riboflavin solution is pipetted onto the collagen matrices, they are irradiated with UV-A at 370 nm using a UV-A double diode at 3 mW/cm2 for 45 min. Surface irradiance can be controlled with a calibrated UV-A meter. Maximum matrix stability is usually reached after 45 min of UV irradiation.

4. Notes 1. All buffers are made with ultrapure water (18.4 MW/cm) using an ELGA Purelab Ultra purifier (http://www.elgalabwater. com) or equivalent. 2. Different collagen type I preparations may contain low levels of additional proteins or collagens of other types. According to the manufacturer, PureCol collagen contains 3% collagen type III. While this low concentration of collagen type III

106

Franz and Muller

does not inhibit collagen type I fibrillogenesis, the precise effect of the residual type III collagen in the solution is unknown. 3. An optional housing can be erected around the AFM to minimize acoustic noise and thermal drift. 4. Because of a well-characterized nonlinear tip-convolution effect, collagen fibrils appear laterally expanded in AFM topographs recorded with pyramid-shaped AFM tips. This should be considered when interpreting the information contained in AFM images. 5. AFM high-resolution time-lapse imaging requires constant imaging conditions throughout the entire scan period. Since scanning forces are affected by changes in temperature (thermal drift), it may be advisable to equilibrate the buffer solution for at least 1 h inside the AFM sample chamber to obtain complete thermal relaxation. Microscope design permitting, the collagen solution can be directly added while the probe is inserted into the microscope.

Acknowledgments We thank Dr. David Cisneros and Dr. Fengzhi Jiang for providing AFM images and Dr. Pierre-Henri Puech for helpful comments. References 1. Jiang, F., Khairy, K., Poole, K., Howard, J. & Muller, D. J. (2004) Creating nanoscopic collagen matrices using atomic force microscopy. Microsc Res Tech 64, 435–40. 2. Cisneros, D. A., Friedrichs, J., Taubenberger, A., Franz, C. M. & Muller, D. J. (2007) Creating ultrathin nanoscopic collagen matrices for biological and biotechnological applications. Small 3, 956–63. 3. Jiang, F., Horber, H., Howard, J. & Muller, D. J. (2004) Assembly of collagen into microribbons: effects of pH and electrolytes. J Struct Biol 148, 268–78. 4. Ludwig, T., Kirmse, R., Poole, K. & Schwarz, U. S. (2008) Probing cellular microenvironments and tissue remodeling by atomic force microscopy. Pflugers Arch 456, 29–49. 5. Franz, C. M. & Puech, P. H. (2008) Atomic force microscopy - a versatile tool for studying cell morphology, adhesion and mechanics. Cell Mol Bioeng 1, 289–300.

6. Cisneros, D. A., Hung, C., Franz, C. M. & Muller, D. J. (2006) Observing growth steps of collagen self-assembly by time-lapse highresolution atomic force microscopy. J Struct Biol 154, 232–45. 7. Friedrichs, J., Manninen, A., Muller, D. J. & Helenius, J. (2008) Galectin-3 regulates ­integrin alpha2beta1-mediated adhesion to collagen-I and -IV. J Biol Chem 283, 32264–72. 8. Friedrichs, J., Torkko, J. M., Helenius, J., Teravainen, T. P., Fullekrug, J., Muller, D. J., Simons, K. & Manninen, A. (2007) Contributions of galectin-3 and -9 to epithelial cell adhesion analyzed by single cell force spectroscopy. J Biol Chem 282, 29375–83. 9. Taubenberger, A., Cisneros, D. A., Friedrichs, J., Puech, P. H., Muller, D. J. & Franz, C. M. (2007) Revealing early steps of alpha2beta1 integrin-mediated adhesion to collagen type I by using single-cell force spectroscopy. Mol Biol Cell 18, 1634–44.

Studying Collagen Self-Assembly 10. Franz, C. M., Taubenberger, A., Puech, P. H. & Muller, D. J. (2007) Studying integrinmediated cell adhesion at the single-molecule level using AFM force spectroscopy. Sci STKE 2007, pl5. 11. Poole, K., Khairy, K., Friedrichs, J., Franz, C., Cisneros, D. A., Howard, J. & Mueller, D. (2005) Molecular-scale topographic cues induce the orientation and directional movement of fibroblasts on two-dimensional collagen surfaces. J Mol Biol 349, 380–6. 12. Friedrichs, J., Taubenberger, A., Franz, C. M. & Muller, D. J. (2007) Cellular remodelling

107

of individual collagen fibrils visualized by  time-lapse AFM. J Mol Biol 372, 594–607. 13. Hattori, S., Adachi, E., Ebihara, T., Shirai, T., Someki, I. & Irie, S. (1999) Alkali-treated collagen retained the triple helical conformation and the ligand activity for the cell adhesion via alpha2beta1 integrin. J Biochem 125, 676–84. 14. Muller, D. J., Amrein, M. & Engel, A. (1997) Adsorption of biological molecules to a solid support for scanning probe microscopy. J Struct Biol 119, 172–88.

wwwwwwwwwwwwww

Chapter 8 Atomic Force Microscopy Imaging of Human Metaphase Chromosomes in Liquid Osamu Hoshi and Tatsuo Ushiki Abstract Methods for atomic force microscopy (AFM) imaging of human metaphase chromosomes were ­introduced in the present study. Chromosomes from the lymphocytes were fixed and prepared onto glass slides as the chromosome spread, and observed in phosphate-buffered saline by dynamic mode AFM. On the contrary, chromosomes from the human cell line BALL-1 were isolated using the hexylene glycol method, absorbed onto a silane-coated glass slide, and observed in a hexylene glycol buffer solution by dynamic mode AFM. AFM provides three-dimensional topographic images of both fixed and unfixed human chromosomes with height information. The ultrastructural image of a pair of chromatids was also obtained by AFM in a liquid condition. The combined use of the AFM and light microscopy of cytochemically and/or immunocytochemically stained chromosomes is also expected to be useful for studies on the localization of chemical components in relation to the higher-order structure of the chromosomes. Key words: Human metaphase chromosome, AFM

1. Introduction Chromosomes are thread-like structures that appear during cell division in eukaryotic cells. They consist of DNA, histones, and nonhistone proteins, and act as a vehicle for equal distribution of genetic materials into the two daughter cells. It is well known that the DNA strand wraps histone octamers to form a beadson-a-string fiber, which further condenses into a 30-nm chromatin fiber. However, the mechanism by which the 30-nm chromatin fiber condenses into the chromosome is still unelucidated because of the technical difficulties in observing the ­ultrastructure of the chromosome by conventional electron microscopy.

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_8, © Springer Science+Business Media, LLC 2011

109

110

Hoshi and Ushiki

Atomic force microscopy (AFM), which was invented in 1986, has recently attracted biologists because it can obtain the three-dimensional image of the sample surface at the resolution comparable to that of electron microscopy. Thus, many studies have been conducted on the use of AFM to observe chromosomes since Degrooth and Putman (1992) first applied AFM to chromosomal studies (1–5). Imaging chromosomes by AFM has several advantages: (1) AFM provides the three-dimensional image of chromosomes at a high resolution in vacuum, air, and liquid environments. (2) It enables the direct observation of chromosomes, in contrast with scanning electron microscopy that requires metal coating and/or conductive staining. (3) After AFM imaging, the samples can be further treated with various biological methods such as chromosome banding, immunocytochemistry, and fluorescence in situ hybridization. In this chapter, we will introduce the techniques for ­observing human chromosomes by AFM. We will describe the preparation methods for fixed and unfixed human chromosomes and discuss the suitable techniques for imaging chromosomes by AFM, ­especially in a liquid condition.

2. Materials 2.1. For Preparation of Fixed Human Metaphase Chromosomes

1. Heparinized peripheral human blood. 2. Ficoll-Paque PLUS. 3. Karyotyping culture medium. 4. Colcemid. 5. 75 mM KCl. 6. Carnoy’s solution (methanol/acetic acid, 3:1, v/v). 7. Glass slide.

2.2. For Preparation of Isolating Native Chromosomes

1. Human cell line BALL-1 (RCB0256; RIKEN Cell Bank, Tsukuba, Japan). 2. Culture medium. 3. Fetal calf serum. 4. Colcemid. 5. 75 mM KCl. 6. Hexylene glycol buffer: 1.0  M hexylene glycol (2-methyl-2, 4,-pentanediol), 0.5 mM CaCl2, and 0.1 mM PIPES, pH 6.5. 7. Dounce homogenizer (15 ml capacity). 8. Silane-coated glass slide. 9. Glass cutter.

Atomic Force Microscopy Imaging of Human Metaphase Chromosomes in Liquid

2.3. For AFM

111

1. Atomic force microscope (an SPA-400 microscope unit ­controlled by an SPI 4000 probe station, SII NanoTechnology, Chiba, Japan). 2. phosphate-buffer saline (PBS). 3. Open-typed glass chamber. 4. V-shaped silicon nitride cantilever with a nominal spring constant of 0.32 N/m (DNP-S20, Veeco Instruments Inc., NY, USA). 5. 4¢,6-Diamino-2-phenylindole dihydrochroride (DAPI).

3. Methods The methods described below outline (1) the fixation of human metaphase chromosomes, (2) the isolation of native human chromosomes, (3) the preparation of fixed human metaphase ­chromosomes for observation by AFM in liquid, (4) the preparation for isolating native human metaphase chromosomes for observation by AFM in liquid, and (5) the observation by AFM in liquid. 3.1. Fixed Human Metaphase Chromosomes

Human lymphocytes were isolated from heparinized peripheral blood by Ficoll-Paque gradient density centrifugation (see Note 1). The lymphocytes were cultivated in a karyotyping culture medium for 72 h at 37°C under an atmosphere containing 5% CO2 and 95% air. Lymphocytes were then arrested in metaphase by adding colcemid to the culture medium at a final concentration of 0.05 mg/ml for 1 h (see Note 2). The cell suspension was then exposed to 10 ml of 75  mM KCl for 30  min at room temperature for hypotonic treatment. At the end of this treatment, 5 ml of chilled Carnoy’s solution was added to the cell suspension, followed by very gentle mixing using a pipette. Three changes of the fixative were made by centrifugation, and the cells were finally resuspended in fresh Carnoy’s solution and stored in a freezer at −20°C.

3.2. Isolating Native Chromosomes

Cells from the human cell line BALL-1 were cultivated in the culture medium (RPMI 1640) supplemented with 10% fetal calf serum at 37°C under an atmosphere containing 5% CO2 and 95% air. After arrest of cell division in metaphase with 0.06 mg/ml colcemid for 12 h, the cells were suspended and centrifuged at 190 × g for 10 min at 4°C. The cells were then resuspended in 10 ml of the culture medium, kept for 20  min at 4°C, centrifuged again at 190 × g for 10 min, and exposed to 75 mM KCl for 15 min at 4°C. After the hypotonic treatment, the cells were collected by centrifugation and isolated by the hexylene glycol method according to Wray and Stubblefield (6). Briefly, the cells were collected by

112

Hoshi and Ushiki

c­ entrifugation at 190 × g, resuspended in the hexylene glycol ­buffer (pH 6.5), and incubated in the same buffer for 10 min at 37°C. The cells were then broken with a Dounce homogenizer (15 ml capacity) by five gentle strokes (7) (see Note 3) and centrifuged at 400 × g at 4°C for 3 min in order to remove nuclei. The supernatant was transferred to a clean tube and centrifuged at 1,680 × g for 15 min at 4°C to concentrate the chromosomes. 3.3. Preparation of Fixed Human Metaphase Chromosomes for Observation by AFM in Liquid

Fixed human metaphase chromosomes were observed after spreading the samples onto glass slides (S2644). The chromosome spreads were formed simply by dropping the cell suspension onto glass slides. These spreads were briefly dried in air to affix them onto the glass slides and then immediately immersed in PBS. These specimens were observed by phase contrast microscopy, and chromosome spreads with little cytoplasmic debris were selected for AFM imaging (see Note 4). After the chromosome spreads were photographed with a light microscope, the glass slides were cut into small pieces (about 7 mm2) with a glass cutter, and the pieces bearing the chromosomes were mounted onto an open-typed glass chamber (about 2 cm in diameter and 5 mm in depth). During these processes, the samples were kept in PBS to avoid air-drying of the chromosomes (see Note 5).

3.4. Preparation for Isolating Native Chromosomes for Observation by AFM in Liquid

Isolated native chromosomes were adsorbed on a silane-coated glass slide (S9443) by dropping the chromosome suspension onto the slide and observed with a phase contrast microscope to determine the portions to be studied by AFM (see Note 6). After the determined portion was photographed with a light microscope, the glass slides were cut into small pieces (about 7 mm2), and the pieces bearing the chromosomes were mounted on a glass chamber. During these processes, the samples were kept in hexylene glycol buffer to avoid air-drying of the chromosomes.

3.5. AFM in Liquid

AFM observations were made using a commercial atomic force microscope equipped with a light microscope (see Note 7). Under the light microscope, the tip of the cantilevers was carefully set above the chromosomes to be observed with AFM. The AFM images were obtained in a dynamic force mode (i.e., intermittent contact mode) either in PBS for fixed chromosomes, or in hexylene glycol buffer for isolating native chromosomes. Reduction in the oscillation amplitude was used as the feedback parameter by a slope detection technique (see Note 8). When V-shaped silicon nitride cantilevers with a nominal spring constant of 0.32  N/m (DNP-S20) were used for imaging, the typical resonance frequency of the cantilevers was about 12.5 kHz in liquid and the Q-value was usually 1–2 in this condition. We sometimes used the Q-control technique, by which the Q-value is increased three to four times (e.g., from Q = 1 to 3 or 4) (see Note 9). The AFM images were

Atomic Force Microscopy Imaging of Human Metaphase Chromosomes in Liquid

113

usually obtained in three modes: height mode, amplitude mode, and phase mode. The size of the images were 512 × 512 pixels. AFM images showed that fixed metaphase chromosomes in the chromosome spread are 200–400 nm in height, while unfixed isolated chromosomes are 400–800  nm in height, as shown in Figs. 1 and 2 (see Note 10). Metaphase chromosomes have a pair of chromatids, and alternating ridges and grooves were often observed in the chromatid arms by AFM. At high magnification, chromatids are observed as an aggregation of globular or fibrous structures about 50–60 nm thick. After the AFM images are obtained, samples can be stained with DAPI and observed by fluorescence microscopy. The samples can also be prepared for immunocytochemistry in order to compare the structures with various chemical components. Thus, AFM imaging of chromosomes is expected to be useful not only for the investigation of the three-dimensional surface structure of the chromosomes with height information, but also for the studies on the localization of chemical components in relation to the higher-order structure of the chromosomes, by the combined use of the AFM and light microscopy of cytochemically and/or immunocytochemically stained chromosomes.

Fig. 1. (a) Phase contrast microscopic images of the spread of fixed chromosomes in phosphate-buffered saline (PBS). The square line indicates the portion for AFM imaging in (b). (b) AFM images of the fixed human chromosomes in PBS. After observation of the chromosome spread with S phase contrast microscope, chromosomes to be studied were imaged by dynamic mode AFM in PBS. The sister chromatids of each chromosome have ridges and grooves on their surface. (c) The profile of the chromosome arm (lower right) shows that the height of this chromosome is about 250 nm.

114

Hoshi and Ushiki

Fig. 2. Isolated unfixed chromosomes in a hexylene glycol buffer solution. (a) Phase contrast microscopic image of isolated unfixed chromosomes mounted on a silane-coated glass slide. The chromosome indicated by an arrow was imaged by AFM in the same buffer solution. (b) AFM image of the isolated unfixed chromosome indicated by the arrow in (a). (c) Closer view of a part of (b). Chromatids are observed as an aggregation of globular or fibrous structures about 50 nm thick. (d) The profile of the chromosome arm shows that the height is 400 nm. (e) The fluorescent microscopic image of isolated unfixed chromosomes after AFM observation. These chromosomes were stained positively with DAPI. The chromosome found in the lower portion of this micrograph is the same as that observed in (b). (Reproduced with permission from ref. 4).

4. Notes 1. The mitotic index of lymphocytes might be different among individual donors. 2. The time span of colcemid treatment affects the number of the mitotic cells and the degree of compaction of chromosomes. If the time of colcemid treatment is long, the number of mitotic cells becomes high and the compaction of chromosomes is tight. 3. This homogenizer treatment destroys the cell membrane and releases the chromosomes into the buffer. An excess of homogenization should be avoided. 4. The chromosome spread with few cytoplasmic debris should be carefully selected by phase contrast microscopy prior to the AFM examination. Otherwise, a blanket-like layer of cytoplasmic debris covers the chromosome.

Atomic Force Microscopy Imaging of Human Metaphase Chromosomes in Liquid

115

5. Attention should be paid to drying of chromosomes during the AFM operation, because the buffer evaporates in the open-typed chamber. 6. Phase contrast microscopy is useful in observing chromosome samples without any staining. Because isolated chromosomes are scattered sparsely on the glass slide, the position of chromosomes should be marked with a pigmented pen in order to find them out easily for AFM imaging. 7. The light microscopic observation of samples in the AFM is important for the adjustment of the probing tip of the cantilever over the chromosomes to be studied. 8. The interaction force between the tip and the sample should be carefully adjusted to the minimum; otherwise chromosomes are easily deformed during scanning. 9. When the original Q-value is very small (around 1), the Q-control technique is sometimes useful for obtaining stable AFM images of chromosomes by dynamic mode AFM in a liquid condition (8, 9). 10. Tip artifacts might be serious when the conventional V-shaped cantilever is used for AFM imaging of chromosomes with height of more than 400 nm. This is because the cantilevers have a pyramidal tip with a large aspect ratio, which affects the lateral shape of the chromosomes in AFM images. A cantilever with a conical tip at an angle around 10° and the tip height of 9 mm (Au-ISC, 225C3.0, Team Nanotec GmbH, Villingen-Schwenningen, Germany) might be useful to improve the imaging quality (8). References 1. De Grooth, BG. and Putman, CA. (1992) High-resolution imaging of chromosomerelated structures by atomic force microscopy. J. Microsc. 168, 239–247. 2. Putman, CAJ., van er Werf, KO., de Grooth, BG., van Hulst, NF., Greve, J. and Hunsma, PK. (1992) A new imaging mode in Atomic Force Microscopy based on the error signal. SPIE. 1639, 198–2047. 3. Hoshi, O., Owen, R., Miles, M. and Ushiki, T. (2004) Imaging of human metaphase ­chromosomes by atomic force microscopy in liquid. Cytogenet. Genome. Res. 107, 28–31. 4. Hoshi, O., Shigeno, M. and Ushiki, T. (2006) Atomic force microscopy of native human metaphase chromosomes in a liquid. Arch. Histol. Cytol. 69, 73–78. 5. Ushiki, T and Hoshi, O. (2008) Atomic force  microscopy for imaging human meta-

phase chromosomes. Chromosome Res. 16, 383–396. 6. Wray, W. and Stubblefield, E. (1970) A new method for the rapid isolation of chromosomes, mitotic apparatus, or nuclei from mammalian fibroblasts at near neutral pH. Exp. Cell. Res. 59, 469–478. 7. Marsden, MPF. and Laemmli, UK. (1979) Metaphase chromosome structure: Evidence for radial loop model. Cell 17, 849–868. 8. Ushiki, T., Shigeno, M. and Hoshi, O. (2008) Techniques for imaging human metaphase chromosomes in liquid conditions by atomic force microscopy. Nanotechnology 19, 384022 (5 pp). 9. Humphris, ADL. and Round, AN. and Miles, MJ. (2001) Enhanced imaging of DNA via active quality factor control. Surf. Sci. 49, 468–472.

wwwwwwwwwwwwww

Chapter 9 Atomic Force Microscopy of Proteasome Assemblies Maria Gaczynska and Pawel A. Osmulski Abstract The proteasome is the essential prime protease in all eukaryotes. The large, multisubunit, modular, and multifunctional enzyme is responsible for the majority of regulated intracellular protein degradation. It constitutes a part of the multienzyme ubiquitin–proteasome pathway, which is broadly implicated in recognition, tagging, and cleavage of proteins. The name “proteasome” refers to several types of protein assemblies sharing a common catalytic core particle. Additional protein modules attach to the core, regulate its activities, and broaden its functional capabilities. The structure of proteasomes has been studied extensively with multiple methods. The crystal structure of the core particle was solved for several species. However, only a single structure of the core particle decorated with PA26 activator has been determined. NMR spectroscopy was successfully applied to probe a much ­simpler, archaebacterial type of the core particle. In turn, electron microscopy was very effective in exploring the spatial arrangement of many classes of assemblies. Still, the makeup of higher-order ­complexes is not well established. Besides, the crystal structure provided very limited information on proteasome molecular dynamics. Atomic force microscopy (AFM) is an ideal technique to address questions that are unanswered by other approaches. For example, AFM is perfectly suited to study allosteric regulation of proteasome, the role of protein dynamics in enzymatic catalysis, and the spatial organization of modules and subunits in assemblies. Here, we present a method that probes the conformational diversity and dynamics of yeast core particle using the oscillating mode AFM in liquid. We are taking advantage of the observation that the tubeshaped core particle is equipped with a swinging gate leading to the catalytic chamber. We demonstrate how to identify distinct gate conformations in AFM images and how to characterize the gate dynamics controlled with ligands and disturbed by mutations. Key words: Atomic force microscopy, Single molecule analysis, Proteasome, Allostery, Protein dynamics, Enzyme catalysis, Conformational selectivity

1. Introduction All proteins display a certain level of local and global dynamics (1). It is accepted that globular proteins exist in more than one ligandindependent conformational state, the phenomenon called

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_9, © Springer Science+Business Media, LLC 2011

117

118

Gaczynska and Osmulski

c­ onformational diversity or conformational instability (2). Even more, the intrinsic dynamics is postulated to be the basis of catalytic functions and allosteric regulation of enzymes (3, 4). Allostery is now broadly understood as a coupling of conformational states between widely separated sites. So far, the conformational diversity has been studied mostly with NMR spectroscopy, hydrogen– deuterium exchange, and molecular modeling (1, 5, 6). These methods, especially NMR spectroscopy, provide a comprehensive analysis of molecular dynamics; however, they are well suited for much smaller proteins. Paradoxically, the robust allosteric transitions are well pronounced in large multisubunit enzymes, where the traditional methods of studying protein dynamics meet overly complicated systems. Allosteric transitions induce nanometer-scale movements within a molecule occurring on a timescale of milliseconds to seconds. Both the spatial and temporal ranges are within the limits of atomic force microscopy (AFM). The gentle and noninvasive tapping mode AFM in liquid allows for real-time analysis of single protein molecules without disturbing their natural dynamics. Here, we show that analysis of images obtained with tapping mode AFM in liquid provides a valuable information about conformational diversity, allosteric transitions, and their coupling with enzymatic activity of the exemplary large and ­complex eukaryotic enzyme, the proteasome. The name “proteasome” in eukaryota encompasses several assemblies sharing a common catalytic core (7). The core particle (CP), 20S proteasome, is a heteromultimeric assembly of 28 ­subunits arranged in four stacked heptameric rings forming a ­hollow cylinder. According to the crystal structure models of yeast and mammalian 20S proteasomes, the diameter of the ­cylinder is nearly 12 nm, and the length approaches 16 nm (8, 9). The inner b rings form a concealed catalytic chamber with three pairs of active sites. The outer a rings are equipped with a swinging gate guarding the entrance to the central channel leading to the catalytic chamber (Fig.  1). The gate is built from short N-terminal sequences of a subunits. Opening of the gate clears the way for substrates to reach the catalytic centers; however, crystal structures of eukaryotic core particles consistently showed the a rings with tightly closed gates (8–11). The “tops” of proteasomal cylinder, called a faces, are nearly flat and are used as platforms for the attachment of several kinds of additional proteins or protein complexes, modulating the activity of core particle and forming higherorder proteasomal assemblies (7). The CP can degrade peptides and unfolded ­proteins. Attachment of 19S ­heteromultimeric regulatory particle (RP; cap) bestows the resulting assembly with the ability to recognize and process polyubiquitinated proteins (7, 12). The core particle with two RPs attached forms 26S proteasome, the most structurally and functionally advanced form of the proteasome. The attachment of 19S cap is believed to open the gate in a ring of the core particle (13). The increased catalytic activity

Atomic Force Microscopy of Proteasome Assemblies

119

Fig. 1. A diagram of eukaryotic core proteasome particle (20S proteasome). Left : the drum-shaped molecule is built from 28 subunits arranged in four heteroheptameric rings. Right : the top surface of external rings is equipped with a gate formed by swinging N-terminal fragments of a subunits.

is treated as a ­biochemical evidence of the gate opening in most of the higher-order proteasomal assemblies (14–16). In turn, solving the crystal structure of the proteasome with attached heptameric activator complex (17) brought the structural evidence for the gate opening. The 3D reconstruction of cryo-electron micrographs showed that another activator, PA200, opens the gate as well (18). The low but ­consistently detectable catalytic activity of free core particle was explained by gate opening due to external factors, such as binding of hydrophobic peptides or unfolded protein domains hypothetically interacting with the gate, or by destabilization of the gate area in a small population of naturally inactive CP (7, 19, 20). The ­crystal structure models of free 20S proteasome or proteasome liganded with competitive inhibitors did not support allosteric movements in the gate area or elsewhere, leading to a perception of proteasome as a hardly allosteric enzyme (7). Meantime, biochemical studies provided an evidence of allostery between catalytic centers and putative regulatory sites (21), as well as between the catalytic centers and a face. The latter route included modulation of catalytic activity by attachment of distinct activator complexes (22) and modulation of RP stability by signals from active sites engaged by substrates or selected inhibitors (23, 24). A sparse structural ­evidence of dynamics in proteasomal complexes includes “wagging” of the 19S cap and a ring dilation observed in cryo-electron micrographs (25, 26). An evidence of increased dynamics around the central opening was provided by NMR spectroscopy of the archaebacterial proteasome, which is considered as a simpler version of the core particle built with a homoheptameric a ring but without a fully developed gate (27). Besides our studies, the AFM imaging of the CP was directed at micropatterning (28) or ­obtaining ordered and highly immobilized molecules (29–31), without attempts to address the molecular dynamic behavior.

120

Gaczynska and Osmulski

Below we present an AFM method that allows to distinguish between an eukaryotic CP with closed gate and that with an open gate. We discuss methods to approximate the timescales of openclosed transitions. The described procedures can serve as a guide for AFM imaging-based detection of conformational changes in other large protein molecules.

2. Materials 1. High-purity eukaryotic 20S proteasome. The particles for AFM imaging should be as pure as reasonably possible. Commercially available 20S proteasomes (Enzo, Calbiochem) perform well. However, it is always worth to check the purity of each new batch using 12% SDS-PAGE. The enzyme is stable for months when stored frozen in 20% glycerol-containing buffer (pH 7–8, −80°C) (see Note 1). Thawed stock solution of the enzyme should be kept on ice and diluted only immediately before use. 2. Buffer for proteasome dilution and imaging: 50  mM Tris– HCl, pH 7.0. Lower ionic strength buffer, for example, 10  mM, will work as well. Sterile-filtered buffer is usually stored at 4°C, so bring a few milliliters to the room temperature before use (see Note 2). 3. A ligand that interacts with proteasome active center, for example, a fluorogenic model peptide substrate, succinylLeu-Leu-Val-Tyr-4-methylcoumarin-7-amide (Suc-LLVYMCA, for example, from Bachem), or a substrate-mimicking transition-state analog inhibitor, carbobenzoxy-LeuLeuLeuB(OH)2 (MG262, Calbiochem). Stock solutions of these ligands in dimethyl sulfoxide (DMSO) are kept frozen and need to be thawed and kept at room temperature for the duration of experiments. 4. DMSO, the solvent of peptide substrates (molecular biology grade). DMSO is highly hygroscopic; therefore, for the best results, the stock solvent should be kept under inert atmosphere. For uncompromised accuracy, the dissolved compounds need not be exposed to air for longer than a brief sample taking. 5. Automatic pipettors of up to 10–20 and 50–100 mL range, with matching tips. 6. Eppendorf tubes. 7. Magnetic stainless steel specimen disks (Ted Pella, Inc.). 8. Muscovite mica in sheets (economy choice) or precut to the size of specimen disks (for example, SPI Suppliers). 9. Tweezers to handle the disks and the probes.

Atomic Force Microscopy of Proteasome Assemblies

121

10. Pressure-sensitive or scotch-type tape. 11. An atomic force microscope with the tapping mode in liquid capability (here, Nanoscope IIIa, Veeco Instruments; ­software version 5.1) equipped with a scanner suitable for nanometer/ micrometer-scale work (here, scanner E for Nanoscope IIIa) (see Note 3). 12. An AFM sample chamber for in-liquid work. 13. Probes (cantilever tips) for contact mode/tapping mode in liquid (here, silicon nitride oxide-sharpened NP-ST or NP-STT probes, Veeco Instruments).

3. Methods 3.1. Preparation of Proteasome Specimens for Imaging

There are several ways to immobilize proteasomes and other ­proteins for imaging. The methods used so far included specific binding to a lipid bilayer containing phosphoinisitol (30), or taking advantage of His-tagged particles and nickel-containing coating of support (28, 29). However, the simplest and least invasive, ­preferred here, is to let the particles electrostatically attach to the mica ­surface (see Note 4). The mica supports should be prepared in bulk in advance. For this purpose, cut the mica sheet into pieces roughly the size of metal disks. Clean the metal disks with acetone. Mix the two-component epoxy glue. Use a toothpick to put a small dab of the glue on the disk and press the mica onto it (see Note 5). Leave for at least 24 h to dry and then store in a dustproof container. When ready for imaging, dilute the proteasome with the buffer. Prepare several microliters of a sample diluted to a concentration of about 5 nM (see Note 6). Proteasomes in such concentration should produce an image of loosely dispersed particles on mica (see Note 7), a condition preferred for studying molecular dynamics (see Subheading 3.3.3). Adjust the concentration accordingly if more or less coverage is desired (Fig.  2). Expose the fresh layer of mica by attaching a piece of a sticky tape to the surface and pulling out a thin layer of the material (see Note 6). Immediately pipette a few microliters (3 mL in this study) on the mica. Leave for about 2 min (see Note 8). Cover the small droplet, which should not dry-out by that time, with a larger droplet of 30 mL of buffer. Using forceps, position the disk in the microscope head, taking care not to spill the droplet. Mount the probe in the chamber for in-liquid work and position the chamber in the AFM, at an extended distance above the sample. Carefully lower the probe manually toward the sample just to snap it into the buffer droplet (see Note 9). Align the laser with the tip. With multicantilever probes such as those used here, choose the one with the nominal spring constant of 0.32 N/m. The s­ ystem is ready for probe engagement and imaging.

122

Gaczynska and Osmulski

Fig. 2. AFM topographs of 600 nm × 600 nm fields of yeast 20S proteasomes electrostatically immobilized on mica and imaged in tapping mode in liquid. Left : the particles are loosely and randomly dispersed. To achieve such effect, 3 mL of 5 nM proteasome preparation was deposited on mica. Right : after depositing 3 mL of 100 nM proteasome preparation on mica, the particles cover the mica in a dense layer. The bar on the far right shows the range of gray scale – coded heights of the particles. The identical gray scale is applied through all figures in this chapter.

3.2. AFM Imaging

Set the field size at 1 mm × 1 mm in the Nanoscope software. Tune the microscope manually and engage the probe. With the microscope and the probes used, the excitation frequency is usually set between 9 and 10  kHz, slightly below the peak resonant frequency. The drive voltage is set at 200–500 mV and the set point ranges from 1.4 to 1.8 V. The higher the set point introduced, the lower the force of tapping applied. Integral gain is usually set in the 0.2–0.4 range, with the highest value not inducing the “ringing” distortion of the image. A typical speed of scanning is set in the 2–3 Hz range; however, analysis of images generated with slower or faster scan rates can provide valuable information (see Subheading  3.3.3). It is useful to readjust the parameters (amplitude and set point) right after successful engagement. After the first scan or two, it should be apparent if the quality of the image and the imaged sample is acceptable or if further adjustments or changes are needed (scan conditions, probe, protein density in the sample, or protein/buffer purity) (see Note 10). As soon as the image is of acceptable quality, the scanning session can proceed even for an hour with the same sample and tip (see Note 11). It is advisable to collect both trace (the probe moving left to right) and retrace images (right to left, along the same scan line) in height mode, as sometimes objects distorted in one scan direction will be acceptable in the other image, and artifacts may be easier to spot when there are two sets of surface topography images created back to back to compare. For the purpose of ­single-molecule analysis of conformational changes, the same group of particles (the same field) should be scanned consecutively at least ten times. The image capture should be set at the automatic mode, and usually no condition adjustment is ­necessary during a dozen or more scans (see Note 12). The scanned field

Atomic Force Microscopy of Proteasome Assemblies

123

changes its position slightly with every scan; however, the thermal drift does not exceed 10 nm per scan, which allows for repeated monitoring of most of the originally set region. After moving the scanning field to a different area of the mica, the sequence of consecutive scanning can be repeated to accumulate enough images of single particles for analysis. Imaging fields smaller than 1 mm2, for example, 700 nm × 700 nm, may offer more detailed information. Care should be taken, however, not to compromise the speed of scanning and resolution in time domain (see Subheading 3.3.3). To test the conformational response of proteasome to the presence of a ligand, for example, the SucLLVY-MCA substrate, the ligand dissolved in buffer can be gently injected directly into the chamber with a pipette tip, even without stopping the scan (see Note 13). Ten microliters of the liquid can be safely injected and will replenish the buffer lost due to evaporation during preceding scans. A final concentration of 100  mM of the substrate approximates the steady-state conditions of an enzymatic reaction well (32). The concentration of DMSO, the solvent for the model substrate, should not exceed 3% (v/v) (33). In a separate control experiment, add buffer with solvent only to the sample and proceed with imaging and analysis as with the ligand. Alternatively, DMSO can be added to the sample after initial scans, several images recorded, the solvent washed-out (see Note 14), and the substrate added as before. 3.3. Analysis of AFM Images of Proteasome Particles 3.3.1. Determination of Orientation of Particles

The typical image of a low-concentration 20S proteasome sample contains a few dozen of particles (Fig. 2). The particles are attached to mica support stably enough so as not to change ­position after multiple rounds of scanning. They appear somewhat larger than the crystal structure – set dimensions due to a tip broadening (tip dilation) phenomenon, straightforward to remove (see Note 15). Before searching for the open and closed conformations, it is mandatory to determine the way particles are positioned on a mica surface. Two major exclusive alignments of CP are possible. In the first case, the cylinder-shaped molecule is “standing” upright on its a ring (top view), allowing for direct observation of the gate status. In the second orientation, the particle is lying on its side (side view) and observation of the gate is not possible. The topography of top-view particles shows rounded, taller objects. In contrast, side-view particles are rectangular and shorter (Fig. 3a). To distinguish between these two orientations, it is necessary to measure the length (l) to width (w) ratio (l/w) (34). The classification described here was performed with the ­section analysis of the Nanoscope software, but any other image processing software (for example, SPIP) capable of morphometric analysis can be employed. The l and w parameters are measured at a fixed depth; in this case, 1 nm from the top to avoid interference from sloping sidewalls.

124

Gaczynska and Osmulski

Fig. 3. The 20S proteasomes attach to mica either in “standing” or in “lying” orientations. The upright position prevails in baker’s yeast particles under the applied conditions. (a) Zoomed-in AFM topographs of single typical particles in top-view (“standing”; left ) and side-view (“lying”; right ) orientations. Sections of the topmost parts of the particles accompany the topographs. The dimensions of the presented particles, measured at 1 nm from the top (see arrows), are as follows: topview length l = 11.3 nm, width w = 9.6 nm, and l/w = 1.18; side-view l = 17.9, w = 10.1, and l/w = 1.77. (b) Histogram analysis of dimensions (length/width ratios) of proteasome particles collected in a single field. The lengths and widths of the particles were measured as shown in (a). Among the total of 37 particles, 35 clustered in the 1–1.2 range of l/w, and only two had the l/w above 1.6. We consider the rounded-top particles in the former cluster as standing on their a rings (top view ) and the rectangular-top particles in the latter cluster as lying on their sides (side view).

The representative histogram analysis of an l/w ratio distribution of particles in a single field is shown in Fig.  3b. There are two clearly separated groups of particles: the first type (majority of the molecules) showing the l/w values close to one, and the second with a much larger l/w ratio. In all our experiments with yeast proteasomes, the top-view particles are prevalent (>90%) (see Note 16). On the basis of analysis of more than 100 particles from multiple fields, the l/w for the first group ranges from 1.00 to 1.16, and for the second group, from 1.26 to 1.68 (35). Accordingly, all particles with the l/w ratio below 1.2 are classified as top-view (standing) particles and can be subjected to the analysis described below. The presented analysis is possible if an AFM tip is regularly shaped (such as a pyramid); otherwise, the routine that removes tip broadening has to be used first. 3.3.2. Classification of Particles into “Open” and “Closed”

Analysis of topography of the zoomed-in images of wild-type, not liganded, standing proteasomes reveals that beyond expectable variations in diameter or height, the particles differ significantly in

Atomic Force Microscopy of Proteasome Assemblies

125

the shape of the surface of a ring. The topmost part of the majority of particles is cone shaped and smooth; however, others are crater shaped, with a well-pronounced dip, or depression, in or near the center of a face. An obvious hypothesis would consider the smooth particles as those with a closed gate, and the crater-like as those with an open gate. We attempted to set the criteria for unambiguous distinguishing between “closed” and “open” conformations of proteasomes. The section tool in AFM-image processing software aids the analysis. In short, if sections of a molecule performed in four different directions were all cone shaped, the particle should be considered closed. If all four sections are crater shaped (with a dip or hole surrounded by a higher-positioned rim), the particle is classified as open (Fig. 4). The gate is positioned in the center of a face in the crystal structure models of 20S proteasome (8, 9, 17); however, the dip in AFM images quite often appears off-center (33–35). This observation is fully explainable by the scan conditions. The diameter of a single top-view particle is nearly 12 nm based on the crystal structure models (8, 9). The expected maximal size of the central channel opening is about 2  nm (17, 36). Under typical conditions of 512 × 512 pixel resolution and 1 mm2 scanned field, the width of a single scan line in digital pixels is approximately 2 nm, about the expected size of the gate. If the scan line happens to be positioned perfectly in the middle of the scanned particle, then the potential opening will appear symmetrically positioned in the middle of the particle image. In a common case of less than-perfect alignment of a scan

Fig. 4. The AFM images of yeast core proteasomes show two stable conformations distinguishable by the shapes of sections through the topmost parts of their topographs. The top row presents the zoomed-in AFM topographs of representative particles, with their respective sections cut in four directions (as shown) below. Left: the particle with a smooth and concave a face is classified as closed. Middle : the particle with a deep central opening in a face, prominent in all four sections, is labeled as open. Right : when the AFM probe ran too slow (see Table 1) through the a face, the open and closed forms were indistinguishable.

126

Gaczynska and Osmulski

line with a center of a particle, the opening registered by an offcenter scan line will appear off-center in the image of a face. Generally, the method of analysis of four sections can be used not only for proteasomes, but also for images of objects capable of opening and closing, for example, membrane channels. The potential ambiguity in the classification may arise from the question about the minimal depth of the depression. In the case of yeast proteasomes, the average depth is 0.56 ± 0.14 nm (n = 151 particles), and the data range is 0.31–0.84  nm. Since random background fluctuations in our images, approximated from the root mean square (RMS) of background power spectral density, are below 0.2 nm, the depth of a dip in the a face is safely above the noise level. Comprehensive analysis of other topographical features of CP will certainly enrich the knowledge about proteasome conformations (see Note 17); however, simplicity of the open-closed distinction is unsurpassed (see Note 18). 3.3.3. Capturing Elusive Conformational Transitions

Analysis of the topography of the same particles that were identified by their unique Cartesian coordinates and recognizable pattern on the mica in successive scans revealed that each and every top-view molecule can reversibly switch between open and closed conformations. The observation strongly suggests that the proteasome is capable of robust conformational transitions. Moreover, the two strikingly distinct shapes of a face are likely manifestations of conformational diversity (conformational instability): the presence of more than one ligand-independent conformational state (2). Interestingly, the partition of open and closed conformers is always about 1:3, both when all particles from a single scan of a field are analyzed and when results of multiple scans of a single particle are computed (see Note 19) (33). The partition changes significantly in favor of the closed form (more than 90%) when the molecules are attached to mica as a dense layer (see Fig. 2 – right panel) (35). We hypothesize that packing molecules in a dense layer and also in a crystal lattice results in molecular crowding and restricts conformational fluctuations. Such effect of molecular crowding was observed for other proteins by combining X-ray scattering and computational modeling (37). The results underline the importance of imaging and analyzing loosely dispersed molecules (see Fig.  2 – left), not packed in clusters, dense layers, or 2D crystals. The high participation of the less-preferred open conformation suggests relatively small free energy difference between the two states: the more stable closed and the less stable open conformation. Addition of certain ligands changes the partition of the conformations in a striking way. For example, introduction of protein or peptide substrates results in the flip in the partition of the conformations to three parts open and one part closed (33). If the substrates are in molar concentrations comparable to the enzyme and thus well below steady state (32), the 3:1 partition will gradually change back to the control 1:3 (open:closed), a

Atomic Force Microscopy of Proteasome Assemblies

127

phenomenon we interpret as the enzyme using up all the ­available substrate (35). A shift toward almost all open particles is observed with inhibitors mimicking tetrahedral transition stage of the catalytic cycle or engaging certain residues in the active site (35). The open control proteasome molecules are undistinguishable from the ones open in the presence of a substrate or an appropriate inhibitor. The same is true about the closed conformers, alike with or without a ligand added. Therefore, we hypothesize that proteasome harnesses its natural conformational transitions to the needs of catalytic activity by allowing the substrates to enter the catalytic chamber through a periodically open gate (35). The characteristics of AFM scanning in time domain limit the range of structural transitions detectable with the method. On the contrary, modulating the scan conditions allows for temporal characteristics of the transitions already within detection limits. The probe set for a typical scan rate of 2.44 Hz needs 210 s to complete a scan of 1 mm2 field (Table 1, see Note 20). Adjacent scan lines that form the full image of the top surface of a single particle are separated in a timescale of a couple of seconds, a very long time even for relatively slow allosteric transitions. However, under such conditions, the probe moves across the 12-nm alpha ring within in about 5 ms, and across the gate area in less than a millisecond (Table 1), an acceptable timescale to study allosteric movements. Interestingly, the open and closed forms are blurred, but still distinguishable when the tip moves across the gate area in

Table 1 The open and closed forms of core particles were distinguishable following the four sections criterion only when the probe scanned the gate area with proper speed Scan rate (Hz)

Scan size (mm2)

Scan length (s)

Tip velocity (nm/ms)

Scan across the a face (ms)

Scan across the gate (ms)

Detectable open–closed forms

1.00

0.25

512

0.5

24

4.0

No

2.44

0.25

210

1.2

10

1.7

Blurred open

2.44

1.00

210

2.4

  5.0

0.8

Yes

3.05

1.00

168

3.1

  3.9

0.6

Yes

2.44

4.00

210

4.9

  2.4

0.4

Yes

4.36

4.00

117

8.7

  1.4

0.2

Yes

The scan time of the gate should match the average lifetime of at least one of the conformations closely. Examples of different scanning conditions with comments about detection of the conformers are presented. Apparently, when the AFM probe traversed the gate area in more than about 2 ms, the images of particles were blurred and the identification of the distinct conformers was not possible

128

Gaczynska and Osmulski

about 2 ms. The four-section method fails to distinguish the two forms when the time to scan the gate increases to 4 ms (Table 1, Fig. 4). Therefore, we hypothesize that the average lifetime of the open form in control proteasomes is close to 2 ms. Consequently, we predict that the average lifetime of a more stable closed form should be three times the average ­lifetime of the open form, a notion currently under investigation.

4. Notes 1. Yeast Saccharomyces cerevisiae 20S proteasome from Enzo Life Sciences Inc. or proteasome purified from MHY501 strain (“wild type”) (35) was used here. More than two freeze–thaw cycles should be avoided, as they “activate” the core proteasome particle. 2. The buffers should be prepared from ultrapure water and the purest reagents available on the market. They have to be filtered through a 0.2-mm bottle filter before storage. It is not useful, however, to filter a small sample of the buffer before use through a spin filter, since such filters often shed nanosized debris into the solution. 3. A good system isolating from vibrations and a high-quality external battery with current smoothing capabilities can tremendously improve the stability of imaging conditions and the image quality. We positioned the microscope on a benchtop vibration isolation platform (BM-4, Minus K Technology), supported on a custom-made sturdy wooden table. An antistatic mat covers the floor in front of the table, where an operator stands. The controller and computer are connected to the Cyber Power PP1100 battery. 4. The freshly exposed mica surface has a negative charge. Most of the proteins, proteasome including, in a neutral-pH buffer will bear a mild positive charge. Therefore, the molecules will freely attach to the mica with presumed most stable attachment positions related to availability of relatively large, flat, and nonnegative surface. 5. It is not necessary to apply the pressure to glued mica for ­longer than a brief finger pressing. Stacking the disks and weighting the stack down or pressing it with a small clamp may result in freshly glued mica sliding off the disk or even breaking. 6. Use the highly diluted protein sample immediately. Prepare a freshly diluted sample if more than about 10 min passed after dilution and before pipetting on the mica. Similarly, use a mica surface within a few minutes (preferably in less than

Atomic Force Microscopy of Proteasome Assemblies

129

5  min) after its exposing. The mica support can be peeled again and reused several times, as long as the fresh unbroken surface can be exposed. 7. If the proteasome preparation is of very low concentration, it can be loaded on mica directly, without dilution in the ­glycerol-free buffer. The glycerol-free buffer should be used to overlay the 3-mL droplet, as usual. 8. The use of a clean cabinet, for example, a PCR enclosure, greatly helps to keep the sample and all small equipment dust free. 9. Instead of “free-standing” buffer droplet on mica, the liquid can be fully enclosed and sealed in the chamber with o-rings. Sealing the chamber prevents sample drying and helps with buffer exchange, if needed. However, setting-up the system with a droplet is much simpler. Drying can be prevented by occasional replenishment of buffer: for example, about 20 mL can be injected with a pipettor through one of the openings in the sample chamber after about an hour of imaging of the same preparation. Replacement of buffer by pipetting buffer in and out slowly and carefully is possible. The sealed system is advantageous when a flow-through type of liquid exchange is desired. 10. Sometimes the probe repositioning in a spring holder of the chamber dramatically improves the image quality. However, if no adjustment of conditions or repositioning works, the tip is likely blunt and needs to be replaced. 11. The out-of-the-box tip will not usually wear down after a few hours of scanning. It may, however, be blunted by collected debris or start to produce a “double-image” artifact (all ­features on the image appear in pairs) resulting from a contaminating particle attached to the tip and acting as a second tip. The “second tip” can sometimes be shaken-off by several seconds of fast scanning (4–5  Hz) without engagement. Several hours of exposure to UV irradiation may help to clean the tip covered with debris. 12. The most often needed adjustment is a slight increase in the amplitude or decrease in the set point. 13. The disturbance introduced by injection is usually brief and the area of scan lost to resulting noise is 20% or less. The acceptable image quality usually comes back without intervention or after a slight increase in the amplitude. 14. To wash out a reversible proteasome ligand or any other compound added to the buffer, wash the sample by pipetting several changes of 50 mL of fresh buffer in and out of the chamber. Ten changes are sufficient to remove the ligand/compound (34, 38).

130

Gaczynska and Osmulski

15. The broadening can be removed with a simple formula

(

)

D = 2 R 2 + d 2 /4 − R , where D = correct diameter of the molecule, d = the apparent diameter measured at half-height, and R = spherical radius of the tip (39). The R is approximated by scanning objects of well-known dimensions, for example, AFM standard grids, with the tip. The radii of NP-ST or NP-STT tips are usually in the 7–9 nm range. 16. Among proteasomes isolated from several eukaryotic sources of CP we analyzed (38), proteasomes from S. cerevisiae were the most efficiently attached in the top-view position (>90%). In contrast, particles from fission yeast (Schizosaccharomyces pombe) tend to partition evenly between the standing and “lying” particles (34). 17. Analysis of length and width of the “lying” (side-view) particles allowed for the distinction of two populations: “barrels” and “drums” (“cylinders”) (34). The diameter of a ring has been found to cluster into two groups as well, the smaller diameters correlated with closed gate and the larger, with the open gate (35). The latter result is fully consistent with the electron microscopy-based observation of a dilation of a ring in 26S proteasome assembly, in which the gate is presumed open (25). 18. Averaging of images. It is possible to apply image processing software to attempt to average topographs of the proteasome molecules. Such approach works fine with proteasomes arranged in a dense lattice of relatively stiff molecules. However, such molecules do lose the ability to switch to the open conformation. Although the disperse arrangement of molecules on mica preserves the conformational diversity of proteasomes, it makes successful averaging of particles representing open and closed conformations separately unexpectedly very difficult. Since the proteasomes are working and constantly change details of their topography, they seldom look as smooth regular assemblies as one would expect from the crystal structure models. Instead, the rims around the crater are most often of uneven height, slanted, and sometimes discontinuous, leading to errors and misclassifications in automatic procedures. 19. Obviously, the more cases (particles in a field or scans of a particle) collected for statistical analysis, the more reliable the results. A single field with about 100 particles gives a good initial approximation of the partition. It is prudent to process images of at least ten fields with at least 20 particles in each field. For analysis of single particles successively scanned ­multiple times in a movie-like fashion, images of at least ten particles with at least ten scans each should be collected.

Atomic Force Microscopy of Proteasome Assemblies

131

20. Scan rates between 2 and 3 Hz and scan sizes between 0.25 and 1  mm2 are most convenient to use; however, images of acceptable quality can be obtained with rates up to 5.55 Hz. Fields larger than about 4 mm2 will not be practical since the size of a single pixel will approach the size of a face, obviously disabling observations of conformational differences. References 1. Henzler-Wildman, K., and Kern, D. (2007) Dynamic personalities of proteins, Nature 450, 964–972. 2. Bahar, I., Chennubhotla, C., and Tobi, D. (2007) Intrinsic dynamics of enzymes in the unbound state and relation to allosteric regulation, Curr. Opin. Struct. Biol. 17, 633–640. 3. Gunasekaran, K., Ma, B., and Nussinov, R. (2004) Is allostery an intrinsic property of all dynamic proteins? Proteins 57, 433–443. 4. Henzler-Wildman, K. A., Thai, V., Lei, M., Ott, M., Wolf-Watz, M., Fenn, T., Pozharski, E., Wilson, M. A., Petsko, G. A., Karplus, M., Hubner, C. G., and Kern, D. (2007) Intrinsic motions along an enzymatic reaction trajectory, Nature 450, 838–844. 5. Dodson, G. G., Lane, D. P., and Verma, C. S. (2008) Molecular simulations of protein dynamics: New windows on mechanisms in biology, EMBO Rep. 9, 144–150. 6. Liu, Y. H., and Konermann, L. (2008) Conformational dynamics of free and catalytically active thermolysin are indistinguishable by hydrogen/deuterium exchange mass spectrometry, Biochemistry 47, 6342–6351. 7. Groll, M., Bochtler, M., Brandstetter, H., Clausen, T., and Huber, R. (2005) Molecular machines for protein degradation, Chembiochem 6, 222–256. 8. Groll, M., Ditzel, L., Lowe, J., Stock, D., Bochtler, M., Bartunik, H. D., and Huber, R. (1997) Structure of 20S proteasome from yeast at 2.4 A resolution, Nature. 386, 463–471. 9. Unno, M., Mizushima, T., Morimoto, Y., Tomisugi, Y., Tanaka, K., Yasuoka, N., and Tsukihara, T. (2002) Structure determination of the constitutive 20S proteasome from bovine liver at 2.75 A resolution, J. Biochem. 131, 171–173. 10. Groll, M., Berkers, C. R., Ploegh, H. L., and Ovaa, H. (2006) Crystal structure of the boronic acid-based proteasome inhibitor bortezomib in complex with the yeast 20S proteasome, Structure 14, 451–456. 11. Groll, M., Kim, K. B., Kairies, N., Huber, R., and Crews, C. M. (2000) Crystal structure of

12.

13.

14.

15.

16.

17.

18.

19.

20.

epoxomicin:20S proteasome reveals a molecular basis for selectivity of alpha¢, beta¢ – epoxyketone proteasome inhibitors, J. Am. Chem. Soc. 122, 1237–1238. Osmulski, P. A., Tokmina-Lukaszewska, M., Endel, L., Sosnowska, R., and Gaczynska, M. (2008) Chemistry of proteasome, in Wiley Encyclopedia of Chemical Biology, John Wiley and Sons, Inc. Bajorek, M., and Glickman, M. H. (2004) Keepers at the final gates: regulatory complexes and gating of the proteasome channel, Cell. Mol. Life Sci. 61, 1579–1588. Rechsteiner, M., and Hill, C. P. (2005) Mobilizing the proteolytic machine: Cell biological roles of proteasome activators and inhibitors, Trends Cell Biol. 15, 27–33. Rosenzweig, R., Osmulski, P. A., Gaczynska, M., and Glickman, M. (2008) The central unit within the 19S regulatory particle of the proteasome, Nat. Struct. Mol. Biol. 15, 573–580. Smith, D. M., Chang, S. C., Park, S., Finley, D., Cheng, Y., and Goldberg, A. L. (2007) Docking of the Proteasomal ATPases’ Carboxyl Termini in the 20S Proteasome’s ? Ring Opens the Gate for Substrate Entry, Mol. Cell 27, 731–744. Whitby, F. G., Masters, E. I., Kramer, L., Knowlton, J. R., Yao, Y., Wang, C. C., and Hill, C. P. (2000) Structural basis for the activation of 20S proteasomes by 11S regulators, Nature 408, 115–120. Ortega, J., Heymann, J. B., Kajava, A. V., Ustrell, V., Rechsteiner, M., and Steven, A. C. (2005) The axial channel of the 20S proteasome opens upon binding of the PA200 activator, J. Mol. Biol. 346, 1221–1227. Kisselev, A. F., Kaganovich, D., and Goldberg, A. L. (2002) Binding of hydrophobic peptides to several non-catalytic sites promotes peptide hydrolysis by all active sites of 20S proteasomes – Evidence for peptide-induced channel opening in the alpha-rings, J. Biol. Chem. 277, 22260–22270. Liu, C. W., Corboy, M. J., DeMartino, G. N., and Thomas, P. J. (2003) Endoproteolytic activity of the proteasome, Science 299, 408–411.

132

Gaczynska and Osmulski

21. Schmidtke, G., Emch, S., Groettrup, M., and Holzhutter, H. G. (2000) Evidence for the existence of a non-catalytic modifier site of peptide hydrolysis by the 20S proteasome, J. Biol. Chem. 275, 22056–22063. 22. Li, J., and Rechsteiner, M. (2001) Molecular dissection of the 11S REG (PA28) proteasome activators, Biochimie. 83, 373–383. 23. Babbitt, S. E., Kiss, A., Deffenbaugh, A. E., Chang, Y. H., Bailly, E., Erdjument-Bromage, H., Tempst, P., Buranda, T., Sklar, L. A., Baumler, J., Gogol, E., and Skowyra, D. (2005) ATP hydrolysis-dependent disassembly of the 26S proteasome is part of the catalytic cycle, Cell 121, 553–565. 24. Kleijnen, M. F., Roelofs, J., Park, S., Hathaway, N. A., Glickman, M., King, R. W., and Finley, D. (2007) Stability of the proteasome can be regulated allosterically through engagement of its proteolytic active sites, Nat. Struct. Mol. Biol. 14, 1180–1188. 25. da Fonseca, P. C. A., and Morris, E. P. (2008) Structure of the human 26S proteasome: subunit radial displacements open the gate into the proteolytic core, J. Biol. Chem. 283, 23305–23314. 26. Walz, J., Erdmann, A., Kania, M., Typke, D., Koster, A. J., and Baumeister, W. (1998) 26S proteasome structure revealed by threedimensional electron microscopy, J. Struct. Biol. 121, 19–29. 27. Sprangers, R., and Kay, L. E. (2007) Quantitative dynamics and binding studies of the 20S proteasome by NMR, Nature 445, 618. 28. Tinazli, A., Piehler, J., Beuttler, M., Guckenberger, R., and Tampe, R. (2007) Native protein nanolithography that can write, read and erase, Nat. Nanotechnol. 2, 220–225. 29. Dorn, I. T., Eschrich, R., Seemuller, E., Guckenberger, R., and Tampe, R. (1999) High-resolution AFM-imaging and mechanistic analysis of the 20S proteasome, J. Mol. Biol. 288, 1027–1036. 30. Furuike, S., Hirokawa, J., Yamada, S., and Yamazaki, M. (2003) Atomic force microscopy studies of interaction of the 20S protea-

31.

32. 33.

34.

35.

36.

37.

38.

39.

some with supported lipid bilayers, Biochim. Biophys. Acta – Biomembranes 1615, 1–6. Thess, A., Hutschenreiter, S., Hofmann, M., Tampe, R., Baumeister, W., and Guckenberger, R. (2002) Specific orientation and twodimensional crystallization of the proteasome at metal-chelating lipid interfaces, J. Biol. Chem. 277, 36321–36328. Gaczynska, M., and Osmulski, P. A. (2005) Small-molecule inhibitors of proteasome activity, Meth. Mol. Biol. 301, 3–22. Osmulski, P. A., and Gaczynska, M. (2000) Atomic force microscopy reveals two conformations of the 20 S proteasome from fission yeast, J. Biol. Chem. 275, 13171–13174. Osmulski, P. A., and Gaczynska, M. (2002) Nanoenzymology of the 20S proteasome: proteasomal actions are controlled by the  allosteric transition, Biochemistry. 41, 7047–7053. Osmulski, P. A., Hochstrasser, M., and Gaczynska, M. (2009) A Tetrahedral Transition State at the Active Sites of the 20S Proteasome Is Coupled to Opening of the  a -Ring Channel, Structure 17, 1137–1147. Forster, A., Whitby, F. G., and Hill, C. P. (2003) The pore of activated 20S proteasomes has an ordered 7-fold symmetric conformation, EMBO J. 22, 4356–4364. Makowski, L., Rodi, D. J., Mandava, S., Minh, D. D. L., Gore, D. B., and Fischetti, R.  F. (2008) Molecular Crowding Inhibits Intramolecular Breathing Motions in Proteins, J. Mol. Biol. 375, 529–546. Gaczynska, M., Osmulski, P. A., Gao, Y., Post, M. J., and Simons, M. (2003) Proline- and arginine-rich peptides constitute a novel class of allosteric inhibitors of proteasome activity, Biochemistry. 42, 8663–8670. Kiselyova, O. I., and Yaminsky, I. V. (2004) Atomic force microscopy of protein complexes, in Atomic force microscopy: biomedical methods and applications (Braga, P. C., and Ricci, D., Eds.), pp 217–230, Humana Press, Totowa, NJ.

Chapter 10 Atomic Force Microscopy of Isolated Mitochondria Bradley E. Layton and M. Brent Boyd Abstract This chapter describes methods for isolating and imaging metabolically and toxicologically challenged mitochondria with atomic force microscopy. Mitochondria were isolated from rat dorsal root ganglia or brain and exposed to glucose or dinitrobenzene (DNB) to simulate the cellular environment of a diabetic animal that has been exposed to excess glucose or to DNB. It is one of only a few articles to present images of membrane structures, such as voltage-dependent, anion-selective channel pores, on intact organelles. The purpose of the chapter is not to report on the metabolic or toxic effects, but to communicate in more detail than a typical journal paper allows the methods used to image isolated organelles. We also provide a series images revealing the outer membrane and outer membrane pores. An image of an isolated nucleus as well as a set of notes written to avoid common pitfalls in isolation, labeling, and imaging is also included. Key words: AFM, Mitochondria, VDAC, Organelle, In situ, Membrane imaging, Apoptosis

1. Introduction In the original work (1) that provided the impetus for this chapter, we explored the hypothesis that in the presence of excess glucose, one or more mitochondrial membrane proteins may become altered, in either morphology or prevalence, leading to a metabolic cascade causing the production of oxidizing agents that eventually destroy mitochondria and subsequently the neurons they support. For review, see ref. 2. A primary challenge in the preparation and subsequent atomic force microscopy (AFM) imaging of single mitochondria was to determine their optimal concentration after isolation from both embryonic rat dorsal root ganglia (DRG) and SY5Y cells. Fluorescence imaging was performed prior to AFM after using a mitochondria isolation

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_10, © Springer Science+Business Media, LLC 2011

133

134

Layton and Boyd

technique derived from ref. 3. This chapter outlines in detail isolation methods, imaging techniques, and provides suggestions for streamlining the process. The research challenge and primary contribution were to establish whether or not mitochondrial membrane proteins and the outer membrane itself become altered in situ as quantified by morphological features at the surface of glucose-challenged or dinitrobenzene (DNB)-challenged mitochondria. Imaging membrane-bound proteins and other membrane morphological features in whole, isolated mitochondria had not been accomplished prior to our work. In addition to a description of the imaging techniques, we provide example images obtained from the technique developed. The atomic force microscope is an ideal tool for imaging in situ proteins, and our investigation was one of the first to attempt to observe proteins in intact organelles. Other AFMbased approaches include the imaging of mitochondria with immunogold labeling (4), isolated outer mitochondrial membranes with intact voltage-dependent, anion-selective channel (VDAC) pores (5, 6), isolated nuclear membranes (7), and imaging single cells, such as spermatozoa (8) or neurons (9). AFM has also been used to perturb whole cells in an effort to track the resulting movement of groups of mitochondria near the point of tip contact (10). With its subnanometer resolution, and the ability to image in aqueous environments, it is also well suited to image organelle membrane proteins in situ (4) or in monolayers (11–14). In this regard, AFM is thus frequently used as a complementary technique to SEM and is often invaluable for imaging subcellular structures since less sample preparation is generally needed, and sputtering artifacts are eliminated (15). However, two other relevant papers (16, 17) obtained images of membrane proteins in air contact mode (see Note 1). There has been a relatively recent revolution in the description of the morphology of the mitochondrial inner membrane. The textbook “radiator-fin” shape presented decades ago is no longer valid. Recent three-dimensional reconstruction techniques of thick-sectioned TEM preparations of single mitochondria have revealed that the cristae form tubular structures 30–40 nm in diameter and several hundred nanometers in length (18). Other evidence indicates that under altered chemical environments, mitochondria themselves may become dramatically reticulated (19). There is strong evidence that mitochondrial size increases in the presence of excessive glucose. This was found by Russell et  al. (20), where mitochondria from rat primary DRGs challenged with 45 mM glucose for 6 h sustained an increase in

Atomic Force Microscopy of Isolated Mitochondria

135

cross-sectional area of ~50% (0.43 ± 0.03 mm2 vs. 0.66 ± 0.08 mm2). This increase in size was also found to be dose- and time-dependent, with increased glucose molarities resulting in greater swelling rates, and longer exposure resulting in greater size increases. A plateau appears to be reached somewhere between 24 and 48 h of excessive glucose exposure. This time course appears to coincide with the complete membrane depolarization at 24 h. In addition to being key components in metabolic disorders, mitochondria may also play an important role in infectious disease pathways. For example in hepatitis, the presence of the hepatitis B protein HBx has been implicated in the opening of the so-called mitochondrial permeability transition pore (21–23). One of the most thoroughly studied mitochondrial proteins is VDAC (24), so named because it is selectively permeable to anions. Expressed in the nucleus, but found in the outer membranes of mitochondria, its molecular mass is ~30  kDa and occurs at a mean density of 103–104 per mm2 on the outer membrane surface (25). Approximating a single mitochondrion as a sphere with a radius of 0.25  mm yields an approximate mitochondrial surface area of 4pr2 = 0.8 mm2, and thus between 800 and 8,000 VDAC pores per mitochondrion. The VDAC diameter reported by Manella (25) is 4.3–5.3 nm with an inner pore diameter of 2–4  nm, and a 0.5  nm projection from the outer membrane surface. A general three-dimensional shape of the VDAC pore has also been obtained by electron crystallography (25), confirming pore geometry. The functional implications of this is that VDAC is easily permeable to many respiration-related and metabolism-related molecules, such as AMP, ADP, ATP, inorganic phosphate, acylcarnitine (26), or other molecules, with diameters <1 nm (27). Thus, the ability to quantify number, spatial distribution, and potentially the state of individual porins may yield valuable insight into the molecular mechanisms of apoptosis. Despite the progress made in imaging isolated mitochondria and isolated mitochondrial membrane proteins, however, the direct observation of a dynamic pore formation event is yet to be achieved. It is commonly thought to be comprised of VDAC, adenine nucleotide translocase (ANT), and cyclophilin D. However, VDAC and ANT can be deleted and the mitochondrial permeability transition still occurs implying that may facilitate pore formation, but are not necessarily required for it to occur. Some investigators believe that its opening is required to release calcium, whereas others argue that a closed pore actually releases more calcium from the mitochondria. Additionally, some claim that a closed MPTP facilitates apoptosis whereas others argue that an open MPTP facilitates apoptosis. Thus, there is likely a prominent role to play for AFM in preparations of living and metabolically active mitochondria. For further reading see ref. 28.

136

Layton and Boyd

2. Materials 2.1. Chemicals

1. Sodium pentobarbital (100 mg/kg) for rat euthanization. 2. Trypsin (0.25 M) for cell dissociation. 3. Ice-cold isotonic (pH = 7.4) phosphate-buffered saline for suspension of cells subsequent to DRG removal. 4. Protease inhibitor, such as leupeptin, aprotinin, or phenylmethanesulphonylfluoride (PMSF), for arresting trypsin activity. 5. Glucose or other metabolite for providing a metabolic challenge to mitochondria. 6. Mannitol for use as a glucose control. 7. Paraformaldehyde (4%) for mitochondria fixation. 8. DNB for mitochondrial toxicity challenge.

2.2. Atomic Force Microscopy

1. Digital Instruments Bioscope AFM model BS3-N2 with a Nikon eclipse TE2000-U optical microscope, using Nanoscope Software v 5.12r5 also from Digital Instruments (Santa Barbara, CA). 2. AFM tips. These may be either tapping or contact with spring constants k = 0.1–1.0  N/m, or natural frequencies of wn ~ 300 kHz (air) or ~10 kHz (fluid). For these experiments, the “short skinny” tips DNP-S (sharp, silicon nitride probes, Veeco Probes), were used. 3. Poly-l-lysine glass slides, e.g., (Polysciences 22247-1 or polyl-lysine Biocoat® from six-well plastic plates from Becton– Dickinson, 354413 or BD two-well poly-l-lysine slides, 354629) for making mitochondria adherent to imaging substrate.

2.3. Animals, Cells, and Organelles

1. Adult female Sprague-Dawley rats. 2. We used mitochondrial samples extracted from embryonic rat DRG, adult rat liver, and from SY5Y cells, an immortalized neural cell line originally expanded from a bone tumor in 1970. The SY5Y cells were obtained for the purpose of isolation procedure verification, since extraction from embryonic DRG mitochondria is more challenging and expensive. 3. Three groups of embryonic rat DRG mitochondria samples were assigned notation (Table 1).

Atomic Force Microscopy of Isolated Mitochondria

137

Table 1 Embryonic rat DRG, rat liver, and SY5Y mitochondria samples collected on 4-28-03, 5-06-03, and 5-20-03 as well as for rat liver mitochondria from previous successful imaging Sample

Notation

Neural control

C

Neural mannitol

M

Neural glucose

Gn

SY5Y

SY5Y

Liver control unfixed

N

Liver control fixed

Nf

Liver glucose unfixed

G

Liver glucose fixed

Gf

Liver DNB-challenged

DNB

3. Methods 3.1. Isolation

1. Using the National Research Council’s “Guide for the Care and Use of Laboratory Animals,” one E15 rat, with 17 embryos, was euthanized using sodium pentobarbital (100  mg/kg) injected intraperitoneally. Dissection begins once no involuntary response results from a strong pinch to the paw pads. Other euthanization techniques, such as decapitation or asphyxiation, are thought to produce rapid alterations in the metabolic state of the animal and may lead to altered mitochondrial states. From each embryo, as many DRGs as possible were extracted under a dissecting microscope with Roboz 0.20 mm × 0.12 mm tweezers. In this case, approximately 500 DRG clusters were extracted in approximately 2  h. Extraction of mitochondria from liver cells is identical with the exception that the liver tissue must be dissected and homogenized cold prior to trypsinization. SY5Y cell prep is identical to that of both DRG and liver from this point forward. For a more in-depth description of dissection procedure, see Note 2 (29). 2. Cell clusters were then exposed to ice-cold trypsin (0.25 M) in phosphate-buffered saline for 10 min.

138

Layton and Boyd

3. Suspension was then spun for 5 min at 2,200 × g in 1.8-mL tubes. 4. The pellet, which contained noncellular membrane material, was resuspended in 1 mL ice-cold isotonic buffer with a protease inhibitor, such as leupeptin, aprotinin, or PMSF. 5. Solution was then homogenized with a glass Dounce homogenizer, e.g., MS851 Acris Antibodies (Germany) on ice for 20 strokes to further gently disrupt cell membranes. 6. Subsequent to homogenization, 1 mL ice-cold isotonic buffer and protease inhibitor were added to arrest damage to exposed mitochondria. 7. Solution was then spun at 900 × g for 5  min at −4°C in 1.8-mL tubes. 8. A 1-mL pellet of cells, nuclei, and other matter was removed using a 10-mL pipette. Note: for AFM, 1 mL is an enormous sample. The volume of a single mitochondrion is approximately 100 aL (100 attoliters = 10−16 L). Thus, a 1-mL sample could contain as many as ten billion mitochondria. 9. Supernatant containing mitochondria was carefully transferred into an ultracentrifuge tube with a 1-mL pipettor. Care was taken to not remove the uppermost layer of the supernatant in order to avoid lipid debris. The lower most layer was also avoided. 10. Solution was then spun at ~12,000 × g for 8 min in 1.8-mL tubes. 11. A small visible pellet with a footprint of less than a square millimeter is now resolvable. This pellet contains intact mitochondria. 12. Supernatant is then poured off or pipetted off and the pellet resuspended in 1  mL ice-cold isotonic buffer without protease inhibitors. 13. Respin solution at 12,000 × g for 8 min. 14. Pull off supernatant and resuspend pellet in 10 mL of ice-cold isotonic buffer. This step is critical for maintaining a high concentration of mitochondria. Resuspending in 100  mL results in a concentration that is too sparse for AFM imaging. Using a volume smaller than 10 mL, if feasible, may be desirable if the pipettor is skilled and the sample can be kept hydrated prior to and during imaging. 3.2. Treatment

1. At this point glucose, mannitol or DNB challenge may be added. In the present study, treatments were started midday after dissections were complete, and metabolic or toxic challenge was added to suspensions at 4°C on agitator for approximately 2  h. Care must be taken to avoid sample from interacting with the container lid to avoid mitochondria

Atomic Force Microscopy of Isolated Mitochondria

139

coming out of solution. Samples may be directly imaged or stored at 4°C overnight. 2. At this point, prior to storage, 2–4% paraformaldehyde may be added to fix cells prior to AFM imaging. 3.3. Atomic Force Microscopy

1. Pipette cells onto polylysine slides and keep them moist if fluid-tapping mode is desired. The time required to align the AFM laser and position the sample under the tip is typically sufficient to allow the mitochondria to become adherent. This occurs through surface charge reactions (Fig.  1). A typical view of the mitochondria or mitochondrial clusters looks similar to the image seen in Fig. 2. At this point in the process, patience must be practiced. Obtaining a few quality AFM images can take a full day in the lab, and obtaining them consistently can take weeks or months of practice. In placing the tip near the target, begin with the “best guess” as to where the tip is actually located on the cantilever and aim directly for the sample (see Note 2) (see Fig. 3). Distinguishing stationary from suspended particles is best accomplished by creating

Fig. 1.  After attaching to poly-l-lysine-coated slide, mitochondria are likely to flatten as increasing numbers of adhesion sites become bound to the lysine. Additionally, if samples are allowed to dry as is necessary for air contact mode atomic force microscopy, mitochondrial volume is likely to decrease due to water loss from the matrix.

Fig. 2. Light microscopy Images of clusters of isolated mitochondria on mica at 400×. (a) control (b) mannitol-control, (c) glucose-challenged. Scale bars = 100 mm (see Note 4).

140

Layton and Boyd

Fig. 3.  AFM tip being brought into proximity of the apparent cluster of control mitochondria. Scale bar = 100 mm.

small waves with a pipette while searching for imageable clusters. Clusters that are not moving as a result of wave motion are likely stuck to the lysine and are the best candidates for imaging. Ideally, a rinse should be done at least once to remove unattached debris from the imaging field that might interfere with the laser or cantilever. It is typically good practice to allow the microscope to complete a 1 mm × 1 mm scan before attempting to fine tune gain, target force, scan size, and scan rate settings (see Note 3). Gradually begin increasing scan size until sample becomes visible, then begin optimizing settings for minimal force by moving the deflection set point away from 0 V and toward −2 V until image disappears, and then return to surface by slowing moving back toward 0  V until image reappears. In tapping mode, begin moving amplitude set point toward the free amplitude until image disappears, then gradually bring amplitude set point back until image reappears. Typical TMR values should be ~1 nm, once tip force was minimized. All imaging for this work was performed at an RMS amplitude of approximately 1 V. 2. Alternatively, if the AFM is equipped with fluorescence imaging, individual mitochondria or mitochondrial clusters may be labeled with MitoTracker® (e.g., Invitrogen formerly M7514 Molecular Probes) and applied according to the manufacturer’s specifications, typically at a concentration of 1–10 nM (Fig. 4). MitoTracker green is excited at 490 nm and emits at 516  nm. Images were taken with a Nikon Microphot microscope and captured on a SPOT RT digital camera. 3. In our original study, fluorescence was not available on the AFM optical microscope and we relied on a probabilistic

Atomic Force Microscopy of Isolated Mitochondria

141

Fig. 4. Fluorescence imaging of mitochondrial preparations reveals that the mitochondrial yield is approximately 150 clusters per mm2 in a 1-mL sample on the left, and fewer than thirty in a 10 mL sample on the right. A maximum AFM scan and a typical AFM scan (10 mm × 10 mm) is shown for comparison in the left image. Scale bars = 100 mm.

formula for finding individual mitochondria. For example, with random placement of the tip in a sparse population of mitochondria and 10 mm × 10 mm scans, each of which takes approximately 10 min to acquire, the probability of success is approximately 5–10% after an hour and a half of imaging. 4. If imaging cannot occur immediately after dissection and dissociation, fixation and storage may be used. In this study, several mitochondria were imageable after nearly a year in storage at 4°C. To do so, respin for 5  min at 12,000 × g, remove supernatant, and resuspend pellet in 1 – 10 mL of 10–100 nM MitoTracker Green. Gently spin and resuspend, and pull off 1 mL for fluorescence imaging as above. Another alternative is to add an anti-fade reagent (e.g., ProLong® Invitrogen). After a 20 mm × 10 mm scan, a typical cluster may appear as in Fig. 5. Heights of some of the mitochondria were as large as 500 nm trailing down to less than 50 nm. Several spherical-like structures with diameters on the order of 0.5 mm were found in all three of the samples G, M, and C, indicating that mitochondria were in fact present in these preparations as evidenced in the fluorescence imaging (G shown in Fig. 6). Under a coverslip, the entire 1 mL droplet spreads to a diameter of approximately 1 cm (~80 mm2), yielding a volumetric density of approximately 106 mitochondria (or clusters) per microliter. Using an order-of-magnitude calculation, with an approximate mitochondrial radius of 0.5  mm and approximating a mitochondrion as a sphere, a typical mitochondrion volume is

V Mt =

4 3 4 p rMt = 3.14(0.25µm)3 = 0.065µm3 = 65 aL. 3 3

142

Layton and Boyd

Fig. 5. (a) Two-dimensional images of SY5Y mitochondria imaged in air contact mode. Scale = 20 mm.

Fig. 6. 2 mm × 2 mm scans of glucose-challenged mitochondria imaged in air-contact atomic force microscopy.

3.4. Fortuitous Impurities

After several instances where apparent mitochondrial clusters stuck to the tip upon attempting to engage, a series of images were obtained of an apparent mitochondrial cluster. As mentioned previously, the method that appeared to be most successful was that of lowering the tip directly onto the sample as opposed to approaching it laterally, thus reducing the chance of releasing it from the poly-l-lysine substrate. The cluster appeared to be only

Atomic Force Microscopy of Isolated Mitochondria

143

Fig. 7. Fluid-tapping atomic force microscopy image of 3hr-glucose challenged unfixed mitochondria preparation from embryonic rat DRGs. (a) 2 mm × 2 mm Scan of apparent nuclear (b) 250 nm close-up of region. (c) Cross-sectional analysis and (d) top projection of apparent pore showing pore diameter of ~10 nm. As an alternative, we could write Red (inner arrowheads) horizontal distance = 40 nm, green (outer arrowheads) horizontal distance = 158 nm.

loosely attached, however, since rescanning the same area often yielded a different image from the previous one. Most notable in these scans are the presence of structures with an approximate density of 50 per mm2 (Fig. 7a) diameters of 100 nm (Fig. 7b) and an approximate height of 10 nm (Fig. 7c), and an apparent inner pore diameter of 10 nm (Fig. 7c and d). See Note 5.

144

Layton and Boyd



The structures seen in Fig. 7 are almost certainly proteinaceous membrane pores. Upon attempting to obtain a larger scan size of the regions surrounding those seen in Fig. 7, to determine if the AFM tip was in fact on the cluster under it, or imaging something stuck to the lysine substrate, the cluster broke free and became unimageable. There exists the possibility that the image in Fig. 7 is the inner or outer surface of an outer mitochondrial membrane stuck to the lysine. Such a large (2 mm × 2 mm) region of membrane could be that of a single mitochondrion, since those shown in Fig. 7 are that of unfixed, glucose-challenged. There also exists the possibility that this is a fragment of nuclear membrane. As a very rough estimate of the molecular mass, Mo, of the protein being imaged, the following equation may be used: NA Mo = V prot , V1 + dV 2 where NA is the Avogadro constant (6.022 × 1023 mol−1), V1 is the partial specific volume of the protein (0.74 cm3/g), V2 is the specific volume of water (1 cm3/g), and d is a factor describing the extent of hydration. Shillers et al. (30) used d = (0.4 mol H2O/ mol protein) for their air-dried method. Substituting in values for our membrane protein with its height of ~4  nm, and d = 1 for fluid, we obtain a molecular mass of ~10  kDa. Note that the protein radius of ~50 nm is likely to be overestimated by as much as 20% since its radius is approximately 10 nm (31). This protein size, however, is greater than that of the VDAC membrane protein reported in the literature (25).

3.5. Single Mitochondria at Close Range

The following sections (Figs. 8–13) consist of a series of images taken at scan sizes of approximately 500 nm × 500 nm of a single mitochondrion. Common to most is the indication that the inner membranous structure is partially visible, since it likely supports the outer membrane as the tip passes over.

3.6. Future Work

Future work may include attempting a membrane protein imaging method similar to that used by Schillers et  al. (17) who obtained excellent results of protein density from frog oocyte membranes. The technical difficulties involved in “rolling out” a mitochondrial membrane onto a lysine slide may be much greater, since the frog oocyte is in fact macroscopic (~1 mm diameter) as compared to the <1 mm mitochondrial diameter. Another protein that may be imageable with the protocol described herein is the nuclear pore complex, or other nuclear membrane protein. In fact, a nucleus was imaged with the protocol described herein. The resulting image is shown elsewhere (32).

Atomic Force Microscopy of Isolated Mitochondria

145

Fig.  8. Fluid-tapping atomic force microscopy image of unfixed control mitochondria from embryonic rat DRGs. Ridge seen running horizontally across image may be a result of this mitochondrion changing shape during imaging.

Fig. 9. Fluid-tapping atomic force microscopy image of fixed control mitochondria from embryonic rat DRGs. Overall morphology is more stable than seen in the unfixed specimens.

146

Layton and Boyd

Fig. 10. Fluid-tapping atomic force microscopy image of fixed control mitochondria from embryonic rat DRGs. Overall morphology is more stable than seen in the unfixed specimens.

Fig. 11. Fluid-tapping atomic force microscopy image of fixed control mitochondria from embryonic rat DRGs. While more convoluted, than those seen above, overall morphology is more stable than seen in the unfixed specimens.

Fig. 12. Fluid-tapping atomic force microscopy image of fixed control mitochondria from embryonic rat DRGs depicting what appears to be a breach in the outer membrane. Whether caused by glucose challenge or sample preparation is yet unknown.

Fig.  13. 20  mm Scan of control DRG mitochondrial preparation in air contact mode atomic force microscopy showing crystalline structures presumably from dried salts in isotonic buffer. It is unclear whether or not the punctuating circular features along crystalline lines are mitochondria or not. Including only the circular features with diameters 0.5 mm or larger, results in ~10 in this 20 mm × 20 mm scan, or the equivalent of 250 per 100 mm × 100 mm area, 2.5× more than was seen with fluorescence imaging of the same sample preparation.

148

Layton and Boyd

4. Notes 1. Two of the most relevant papers for this research, (16, 17), imaged in air. For example, Boujrad et al. (16) obtained AFM images of apparent clusters of an 18 kDa mitochondrial membrane protein, peripheral-type benzodiazepine receptor PBR, associated with the VDAC. Schillers et al. (14) obtained AFM images of at least two membrane proteins of Xenopus laevis oocytes. The heights of these membrane proteins were 10 and 14 nm, corresponding to molecular weights of 275 and 750 kDa, respectively. 2. Currently, the two most time consuming portions of the protocol are DRG removal and mitochondrial location. A large advantage would be gained if the location mitochondria could be visually verified immediately prior to AFM probe engagement. Schillers et al. (14) overcame this obstacle with their oocyte membrane preparation by using a fluorescent marker from Molecular Probes FM1-43 that apparently fluoresces in air, eliminating the need to keep the sample moist for imaging. It could be however, that their samples were still in the process of drying, and thus not completely free of water when fluorescence imaging was done. Hydration is critical to maintain the molecular structure for fluorophore activity. 3. Additional key things learned were that typically a voltage sum above the required 0.5 V is not obtainable until fluid has been added between the fluid cell and the sample. This requires that the bottom of the fluid cell be within ~1 mm of the sample slide to allow for a meniscus to form between cell and sample. Two other potential problems with obtaining a sufficient sum are (1) not having the fluid cell squarely seated on the end of the piezo tube resulting in loss of signal into the photoarray, or (2) using a previously used tip that may have become corroded and thus less reflective. Finally, if care is not taken to do a final rinse with deionized water prior to air imaging, crystallized salts can result (Fig. 13). 4. It became clear from light microscopy, fluorescence imaging, air contact mode AFM, and fluid-tapping AFM that mitochondria are present in all SY5Y, liver, and DRG samples. From air contact mode and fluorescence imaging results, mitochondria are present in a surface area fraction between 2,000 and 10,000 per mm2 and a volume fraction of ~0.01%, when resuspended in 10  mL. In several fluid-tapping mode images, the presence of mitochondria appears to be sparser. More than likely the scarcity is caused by rinsing to remove suspensions from the fluid-tapping arena. Resuspension in no more than 10 mL of fluid appears was sufficient to increase the

Atomic Force Microscopy of Isolated Mitochondria

149

density of mitochondria to a point, where AFM is feasible with a random, nonfluorescent search. Imaging unfixed mitochondria in fluid-tapping mode is considerably more challenging because of mitochondria sticking to the probe, but is essential if in situ pore morphology is to be resolved. Timing is critical in imaging the unfixed mitochondria since osmotic swelling occurs. It does appear that imaging the full boundary of the mitochondria is significantly more challenging since the attachment to the poly-l-lysine substrate in fluid-tapping mode is much more tenuous than that in air contact mode. 5. The images obtained with this technique represent the configuration of the membrane and its proteins in situ. While it is desirable to obtain a map of where individual membrane proteins are located on the scale of seconds or fractions of a second, this is not currently feasible due to the scan rate of AFM which must necessarily be on the order of a few micrometers per second with a path width of a few nanometers. It is also unlikely that the molecular scale images attainable in planar membranes with intact proteins, e.g., (6, 33) or indeed the mitochondrial inner membrane (34) will be feasible with current AFM technology with either fixed or fresh mitochondria. However, three-dimensional reconstruction of a single cell with AFM images of serial sections has recently been achieved (35). The method described in this paper is also distinct from that used in both refs. (4) and (36), whereby immuno-labeled gold nanoparticles are used to detect the presence, prevalence, and spatial distribution of specific proteins.

Acknowledgments The authors thank Michael Bouchard for helpful discussions on mitochondrial structure and function, Chia-Wei Wang and Hui Wang for AFM assistance, Carrie Backus and Terry Miller for mitochondrial isolation assistance, Eva Feldman for funding through the Michigan Diabetes Research and Training Center, Martin Filbert for donation of animal specimens, Ann Marie Sastry for atomic force microscope access, and the Keck Foundation for support. References 1. Layton, B.E., et al. (2004) In situ imaging of mitochondrial outer membrane pores using atomic force microscopy. BioTechniques 37: 564–573. 2. Feldman, E.L. (2003) Oxidative stress and diabetic neuropathy: a new understanding of an old problem. Journal of Clinical Investigation 111(4): 431–3.

3. Verweij, B.H., et  al. (1997) Mitochondrial dysfunction after experimental and human brain injury and its possible reversal with a selective N-type calcium channel antagonist (SNX-111). Neurol Res 19(3): 334–9. 4. Papadopoulos, V., et  al. (1994) Topography of the Leydig cell mitochondrial peripheral-type

150

5.

6. 7.

8. 9.

10.

11.

12.

13. 14.

15.

16.

17.

18.

Layton and Boyd benzodiazepine receptor. Mol Cell Endocrinol 104(1): R5–9. Goncalves, R.P., N. Buzhysnskyy, and S. Scheuring (2008) Mini review on the structure and supramolecular assembly of VDAC. J Bioenerg Biomembr 40(3): 133–8. Goncalves, R.P., et al. (2007) Supramolecular assembly of VDAC in native mitochondrial outer membranes. J Mol Biol 369(2): 413–8. Kramer, A., et al. (2008) Apoptosis leads to a degradation of vital components of active nuclear transport and a dissociation of the nuclear lamina. Proceedings of the National Academy of Sciences of the United States of America 105(32): 11236–11241. Joshi, N.V., et al. (2000) Ultrastructural investigation of human sperm using atomic force microscopy. Archives of Andrology 44(1): 51–57. Parpura, V., P.G. Haydon, and E. Henderson (1993) Three-dimensional imaging of living neurons and glia with the atomic force microscope. J Cell Sci 104 ( Pt 2): 427–32. Silberberg, Y.R., et al. (2008) Mitochondrial displacements in response to nanomechanical forces. Journal of Molecular Recognition 21(1): 30–36. Arechaga, I. and D. Fotiadis (2007) Reconstitution of mitochondrial ATP synthase into lipid bilayers for structural analysis. Journal of Structural Biology 160(3): 287–294. Hoogenboom, B.W., et al. (2007) The supramolecular assemblies of voltage-dependent anion channels in the native membrane. Journal of Molecular Biology 370(2): 246–255. Neff, D., et al. (1997) Chloroplast F0F1 ATP synthase imaged by atomic force microscopy. Journal of Structural Biology 119(2): 139–148. Schillers, H., et al. (2000) Plasma membrane plasticity of Xenopus laevis oocyte imaged with atomic force microscopy. Cellular Physiology and Biochemistry 10(1–2): 99–107. De Stasio, G., et al. (2004) Imaging the cell surface: Argon sputtering to expose inner cell structures. Microscopy Research and Technique 63(2): 115–121. Boujrad, N., B. Vidic, and V. Papadopoulos (1996) Acute action of choriogonadotropin on Leydig tumor cells: changes in the topography of the mitochondrial peripheral-type benzodiazepine receptor. Endocrinology 137(12): 5727–30. Schillers, H., et al. (2000) Plasma membrane plasticity of Xenopus laevis oocyte imaged with atomic force microscopy. Cell Physiol Biochem 10(1–2): 99–107. Mannella, C.A., et al. (1998) Electron microscopic tomography of rat-liver mitochondria

19.

20. 21.

22.

23.

24. 25.

26.

27.

28. 29.

30.

31.

and their interaction with the endoplasmic reticulum. Biofactors 8(3–4): 225–8. Handran, S.D., et  al. (1997) Mitochondrial morphology and intracellular calcium homeostasis in cytochrome oxidase-deficient human fibroblasts. Neurobiology of Disease 3(4): 287–298. Russell, J.W., et  al. (2002) High glucoseinduced oxidative stress and mitochondrial dysfunction in neurons. Faseb J 16(13): 1738–48. Clippinger, A.J. and M.J. Bouchard (2008) Hepatitis B virus HBx protein localizes to mitochondria in primary rat hepatocytes and modulates mitochondrial membrane potential. J Virol 82(14): 6798–811. Clippinger, A.J., T.L. Gearhart, and M.J. Bouchard (2009) Hepatitis B virus X protein modulates apoptosis in primary rat hepatocytes by regulating both NF-kappaB and the mitochondrial permeability transition pore. J Virol 83(10): 4718–31. McClain, S.L., et al. (2007) Hepatitis B virus replication is associated with an HBx-dependent mitochondrion-regulated increase in cytosolic calcium levels. J Virol 81(21): 12061–5. Saccone, C., et  al. (2003) Molecular clock and gene function. J Mol Evol 57 Suppl 1: S277–85. Mannella, C.A. (1998) Conformational changes in the mitochondrial channel protein, VDAC, and their functional implications. J Struct Biol 121(2): 207–18. Holmuhamedov, E. and J.J. Lemasters (2009) Ethanol exposure decreases mitochondrial outer membrane permeability in cultured rat hepatocytes. Arch Biochem Biophys 481(2): 226–33. Nitzan, Y., K. Orlovsky, and I. Pechatnikov (1999) Characterization of porins isolated from the outer membrane of Serratia liquefaciens. Curr Microbiol 38(2): 71–9. Kroemer, G., L. Galluzzi, and C. Brenner (2007) Mitochondrial membrane permeabilization in cell death. Physiol Rev 87(1): 99–163. Malin, S.A., B.M. Davis, and D.C. Molliver (2007) Production of dissociated sensory neuron cultures and considerations for their use in studying neuronal function and plasticity. Nat Protoc 2(1): 152–60. Schillers, H., et al. (2001) Plasma membrane protein clusters appear in CFTB-expressing Xenopus laevis oocytes after cAMP stimulation. Journal of Membrane Biology 180(3): 205–212. Wang, H., B.E. Layton, and A.M. Sastry (2003) Nerve collagens from diabetic and nondiabetic Sprague-Dawley and biobreeding rats: an atomic

Atomic Force Microscopy of Isolated Mitochondria force microscopy study. Diabetes Metabolism Research and Reviews 19(4): 288–298. 32. Layton, B.E. (2008) Recent patents in bionanotechnologies: nanolithography, bionanocomposites, cell-based computing and entropy production. Recent Patents in Nanotechnology 2(1): 1–12. 33. Scheuring, S., T. Boudier, and J.N. Sturgis (2007) From high-resolution AFM topographs to atomic models of supramolecular assemblies. J Struct Biol 159(2): 268–76. 34. Domenech, O., et al. (2006) Thermodynamic and structural study of the main phospholipid

151

components comprising the mitochondrial inner membrane. Biochim Biophys Acta 1758(2): 213–21. 35. Chen, Y., et al. (2005) Atomic force microscopy imaging and 3-D reconstructions of serial thin sections of a single cell and its interior structures. Ultramicroscopy 103(3): 173–82. 36. Layton, B.E., et al. (2008) Collagen’s triglycine repeat length may help to explain an interdomain transfer event from a eukaryote into Trichodesmium erythraeum. Journal of Molecular Evolution 66(6): 539–554.

wwwwwwwwwwwwww

Chapter 11 Imaging and Interrogating Native Membrane Proteins Using the Atomic Force Microscope Andreas Engel Abstract Membrane proteins exist in a lipid bilayer and provide for cell–cell communication, transport solutes, and convert energies. Detergents are used to extract membrane proteins and keep them in solution for purification and subsequent analyses. The atomic force microscope (AFM) is a powerful tool for imaging and manipulating membrane proteins in their native state without the necessity to solubilize them. It allows membranes that are adsorbed to flat solid supports to be raster-scanned in physiological solutions with an atomically sharp tip. Therefore, AFM is capable of observing biological molecular machines at work. Superb images of native membranes have been recorded, and a quantitative interpretation of the data acquired using the AFM tip has become possible. In addition, multifunctional probes to simultaneously acquire information on the topography and electrical properties of membrane proteins have been produced. This progress is discussed here and fosters expectations for future developments and applications of AFM and single-molecule force spectroscopy. Key words: Atomic force microscope, High-resolution imaging, Single-molecule force spectroscopy, Bacterial porin, pH gating, Voltage gating, Voltage-dependent anion channel

1. Introduction In the early 1980s, only a few would have believed that a single atom could be visualized simply by touching it. The groundbreaking paper by Binnig, Gerber, and Quate in 1986 laid the foundation for such an almost unbelievable possibility (1). Another milestone came from the Hansma group that demonstrated the possibility to image proteins in their aqueous environment, thus opening an exciting avenue for structural biologists (2). This progress stimulated early work on a native biological membrane, the gap junction (3). The resolution of the method is determined by the size of the probe that touches Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_11, © Springer Science+Business Media, LLC 2011

153

154

Engel

the surface – ultimately, it is the atom at the tip apex. Since the tip cannot contour deep and narrow crevasses properly, rather flat objects such as biological membranes packed with proteins that protrude only by a few nanometers are quite amenable to highresolution imaging by atomic force microscope (AFM). Native and reconstituted membranes densely packed with proteins embedded in the lipid bilayer have given the most significant results, depicting the surface structure of single membrane proteins with a lateral resolution of 0.5 nm, and a vertical one that is even better (4–9). However, imaging is not the only possibility the AFM offers: single molecules can also be addressed and interrogated by single-molecule force spectroscopy (SMFS) (10, 11). This opens an avenue to probe single membrane proteins in their native environment, the lipid bilayer. This chapter focuses on the preparation methods for membrane samples and the way to acquire high-resolution images. We describe some early work on bacterial porins and summarize recent results from native outer mitochondrial membranes.

2. Materials 2.1. Preparation of Mica Supports for Sample Immobilization

1. Inoxydable and magnetic steel disks of 11  mm diameter (internal services of the Biozentrum, Basel, Switzerland). 2. Teflon sheets of 0.25-mm thickness (Maag Technic AG, Birsfelden, Switzerland). 3. Mica sheets with a thickness between 0.3 and 0.6 mm (Mica House, 2A Pretoria Street, Calcutta 700071, India). 4. “Punch and die” set from Precision Brand Products Inc. (2250 Curtiss Street, Downers Grove, IL 60515). 5. Ethanol (concentration >96% (v/v)). 6. Loctite 406 superglue from KVT König, Dietikon, Switzerland. 7. Araldite Rapid: Two-component epoxy glue from Ciba-Geigy, Basel, Switzerland. 8. Scotch tape.

2.2. Membranes and Buffers

1. All membranes used were prepared as described. Stock solution: 0.25 mg/ml in physiological buffer as indicated. Samples are stored at 4°C and protected from unnecessary light irradiation. 2. Adsorption buffer: typically 20  mM Tris–HCl (pH 7.8), 150 mM KCl, else as specified in the examples. 3. Ionic strength and pH of imaging buffers were adjusted to the sample as indicated in the examples.

Imaging and Interrogating Native Membrane Proteins Using the AFM

155

4. Nanopure water (E18 mOhm/cm). 5. Analytical grade buffers (Tris–HCl, MES). 6. Analytical grade electrolytes. 2.3. AFM and Accessories

1. For contact mode AFM, a commercial multimode AFM was used, equipped with a 120-mm scanner (j-scanner) and a liquid cell (Digital Instruments, Veeco Metrology Group, Santa Barbara, CA). 2. For frequency modulated (FM) AFM, a home-built instrument was used (12). 3. Oxide-sharpened Si3N4 microcantilevers of 100-mm length and a nominal spring constant of k = 0.08  N/m (OMCLTR400PSA, Olympus Optical Co., Ltd, Tokyo, Japan) were employed for contact mode, and for FM-AFM commercial silicon cantilevers (NCH, Nanosensors; nominal force constant k = 30  N/m) were coated with a 30-nm gold layer to enhance their reflectivity.

3. Methods 3.1. Preparation of Mica Supports for Sample Immobilization

1. Punch mica disks of 6 mm and Teflon disks of 13 mm diameter using the “punch and die” set and a hammer. 2. Clean the Teflon and steel disks with ethanol and paper wipes. 3. Glue a Teflon disk on a steel disk using Loctite 406. 4. Glue a mica disk on the Teflon surface of the Teflon-steel disk with the two-component epoxy glue. 5. Let the supports dry for at least 1 day.

3.2. Adsorption of Membranes to Mica

1. Dilute and mix 3 ml of membrane stock solution with 30 ml of adsorption buffer in an Eppendorf tube. 2. Cleave mica with Scotch tape. 3. Pipet the diluted purple membranes on the freshly cleaved mica support. 4. Adsorb membranes for 15–30 min. 5. Wash away the membranes that are not firmly attached to the mica by removing approximately two-thirds of the fluid volume from the mica surface and re-adding the same amount of the corresponding imaging buffer. Repeat this washing procedure at least three times. 6. Mount the support onto the piezo scanner.

156

Engel

7. Install the AFM head containing the fluid cell (without o-ring seal) and cantilever on the microscope. 8. Fill the space between the mica surface and the fluid cell with the corresponding imaging buffer to avoid drying of the protein. 3.3. Operation of the AFM

Clean the AFM fluid cell using detergent and filtered/nanopure water. Rinse the fluid cell with ultrapure ethanol and nanopure water three times. Then, dry the fluid cell using clean nitrogen gas. After thermal relaxation of the instrument, initial engagement of the tip is performed. Specimen deformation and contamination of the tip are minimized during the engagement process by setting the scan size to 0. Prior to scanning the surface, the operating point of the instrument is set to forces below 1 nN. During scanning in the contact mode, the forces are kept as small as possible (<1  nN) and corrected manually to compensate for thermal drift. Such adjustments are not required with the FM-AFM as it tracks the forces and compensates drift. Two frames are simultaneously recorded showing either topography or deflection signal in trace or retrace direction. Usually, deflection and height signals are recorded at low magnification (frame size >1 mm) to find the flat membranes quickly, whereas height signals are acquired in both, trace and retrace direction at high magnification (frame size <1 mm). Typically, the scan speed is set to 4.7–5.5 Hz (lines per second). At high magnification, the scan range of the z-piezo is reduced to avoid limitation of the axial z-resolution by the digitalization of the signal (AD conversion). All measurements are carried out under ambient pressure and at room temperature.

3.4. Imaging Reconstituted Escherichia coli Outer Membrane Porin OmpF

An Escherichia coli outer membrane contains approximately 105 porins to allow the passage of nutrients <600 Da (13). Conductance measurements have shown that porin OmpF trimers exist in either open or closed states, depending on the transmembrane potential (14). The critical voltage above which channels close is Vc > 90 mV for OmpF, and depends on the ionic strength (15). As this voltage is larger than that expected for the outer membrane, the physiological relevance of voltage gating has been questioned (16). However, evidence that Vc is affected by pH has been reported (17), and that membrane-derived oligosaccharides, polycations, low ionic strength buffer, and pressure lower Vc (18). Although the structures of several porins have been solved (19, 20), the mechanism of channel closure is not understood. As suggested by Schulz (21), the narrowest part of the channel (referred to as eyelet) with its particular distribution of positively and negatively charged residues is the favorite candidate for a voltage switch. A model of voltage gating based on oppositely charged moving domains has been discussed, indicating that the loop

Imaging and Interrogating Native Membrane Proteins Using the AFM

157

lining the eyelet is the moving part of the gate (15). However, experiments with engineered mutants where the loop forming the eyelet was tethered by a disulfide bond suggested that this hypothesis needs to be reconsidered (22). An early study of 2D porin OmpF crystals with the AFM revealed two conformations of the extracellular surface (23), but the reason for the conformational switch was not identified. Purple membranes of Halobacterium halobium have been used to optimize the imaging conditions, and to demonstrate that the AFM allows a lateral resolution of »0.5 nm and a vertical resolution of »0.1 nm to be achieved (24, 25). The exceptionally high signal-to-noise ratio of the AFM allows structural details of single-membrane proteins to be observed, and the flexibility of protrusions, comprising a small number of amino acid residues, to be visualized. This progress has been exploited to characterize both the periplasmic and the extracellular surface of porin OmpF (9). As shown in Fig. 1, 2D OmpF crystals adsorbed to mica and imaged under optimized conditions exhibited the periplasmic

Fig. 1. Comparison of the atomic model of a rectangular porin crystal with the raw topographs recorded in buffer solution, by operating the AFM in contact mode at minimal tiploading force. Such crystals often consisted of two layers with the extracellular surfaces facing each other (23). (a) Topographs of the periplasmic OmpF surface displayed features that correlated directly with the periplasmic aspect of the atomic model rendered at 3 Å. (b) To image the extracellular surface, the upper crystalline layer was removed with the stylus. Extracellular domains protruded 13(±2) Å (n = 78) from the lipid bilayer and revealed a substructure consistent with the extracellular aspect of the atomic model, notably a distinct cleft that was visible in the protruding domains.

158

Engel

and extracellular surface topography at high resolution, revealing features such as protruding b-turns confirmed by the atomic structure. Porin membranes adsorbed to highly oriented pyrolytic graphite (HOPG), however, showed an amazing conformational change of the extracellular surface that has not been predicted by the atomic structure when the electric field across the 2D crystal was changed by applying a voltage to a platinum wire closely positioned above the membrane. At zero voltage, topographs of extracellular porin surfaces revealed the protruding domain triplets of 13 Å height (Fig. 2a), but this aspect changed dramatically into doughnut-like structures of about 6 Å height when the voltage was increased to 4,500  mV (HOPG negative; Fig.  2b). At intermediate voltages, overviews showed the transition from the triplet protrusions to the doughnut structures, the porin trimer being unambiguously identified by correlating the topograph with the extracellular surface derived from the atomic model (ellipses in Fig. 2c).

Fig. 2. Voltage-induced conformational change of the extracellular surface of porin OmpF. Crystals were adsorbed to HOPG and topographs taken in 10 mM Tris–HCl (pH 7.8) and 50 mM KCl. The brightness range corresponds to 15 Å. (a) Trigonally packed porin trimers (marked by circles) exhibit 13 Å long protrusions surrounding a triangular vestibule. (b) Upon application of a membrane potential, a dramatic conformational change occurs: the extracellular domains convert into doughnuts of about 6 Å height, marked by circles. (c) Sometimes both conformations were observed simultaneously, allowing the center of the trimer in its extended conformation to be correlated with the center of the doughnuts, as indicated by rows of circles. To observe voltage-induced conformational changes, the voltage (4,500 mV, HOPG negative) was applied with a platinum wire positioned above the cantilever. The effective membrane potential could not be measured, because a significant ion current flowed between the electrode and the HOPG surface that was not fully covered with porin membranes.

Imaging and Interrogating Native Membrane Proteins Using the AFM

159

The transition between the two conformations was fully reversible. To demonstrate that this voltage-induced conformational change represented channel closure, 2D porin crystals were adsorbed to mica in 0.3  M KCl (pH 7.8). Shielding the negative surface charges of porin and mica with a buffer containing >0.1  M KCl was required for adsorption of the sample through van der Waals attraction (26–28). The latter led to an increased K+ concentration between the porin membrane and the mica. Flushing the liquid cell in the AFM with a low ionic strength buffer (5 mM KCl, pH 7.8) created a >60-fold cation gradient. This corresponded to a Nernst potential >100  mV, which was expected to induce channel closure. If the channels remained open, ions between porin sheets and mica would rapidly equilibrate with the bulk solution, and membranes were expected to desorb as a result of electrostatic repulsion (28). While most of the double-layered porin crystals desorbed immediately, most of the single-layered porin sheets remained firmly attached for more than 1  h, consistent with channel closure. Their extracellular surface revealed domains of 13 Å height in the high ionic strength buffer, but after exchanging it with a low ionic strength buffer (Fig. 3a, right), doughnut-like structures were observed that protruded 6(±2) Å (n = 87) from the bilayer. These structures appeared similar to those seen at an applied membrane potential (Fig.  2b, c). Again, this conformational change was fully reversible. After flushing the sample with 0.3 M KCl and imaging the same patch, the extracellular domains were fully extended. As the porin channels are known to be sensitive to pH (17), we used the AFM to probe for any conformation change of the OmpF surface when exposing the 2D crystals to low pH. Between pH 4 and 3, the extracellular protrusions collapsed, converting from a single protrusion (Fig. 3a, left) into a bilobed domain (Fig. 3a, right), concomitant with a height change from 13 Å at pH 7.8 to 6(±1) Å at pH 3 (n = 120). Three such domains formed doughnut-like structures similar to those induced by a membrane potential or by an ion gradient. The reversibility of this structural rearrangement was demonstrated by imaging the same membrane patch after increasing the pH to 4 (Fig. 3a): the extracellular domains snapped back to protrude from the membrane by 12(±2) Å (n = 85). Between pH 4 and 9, and in 0.1–0.5 M KCl, loops forming the extracellular domains of the OmpF assume the conformation determined by x-ray analysis (19). Using AFM, we found them to collapse upon applying a membrane potential >Vc, or changing the pH to 3. In the latter case, the loops are likely to assume a different conformation, because their charge distribution changes profoundly upon lowering the pH (29). For steric reasons, and since the b-barrel trimer is a rigid structure stabilized by intermolecular

160

Engel

Fig. 3. (a) The conformational change of porin OmpF induced by either a low pH or a membrane potential is summarized by the morphed averages from extracellular surfaces of 2D porin crystals recorded at different voltages (adjusted by the ion concentration gradient) or at different pH. (b) This conformational change can be modeled as a rotation of the extracellular domain about a hinge at the rim of the b-barrel. A cross-section of the monomer demonstrates the open conformation (left porin monomer) and the closed conformation (right monomer). The black arrow indicates the putative rotation of the extracellular domain, while gray lines indicate the membrane surface.

interactions (19), our observations suggest that the protrusions probably move toward and collapse into the extracellular vestibule about the threefold axis as indicated in Fig. 3b. Such a conformational change would reduce the height of extracellular domains from 13 to 6 Å and close the channel entrance. Because under physiological conditions a voltage beyond Vc is unlikely to occur (16), we speculated that it is the acidic pH, which leads to channel closure for the bacterial outer membrane porin OmpF (9). Later, the pH-driven closure of maltoporin was demonstrated by black-lipid membrane experiments (30), while the structure of an open and a closed conformation of porin OmpG showed the involvement of extracellular loops (31), and this conformational change was demonstrated most recently under physiological conditions (32).

Imaging and Interrogating Native Membrane Proteins Using the AFM

3.5. Imaging the Voltage-Dependent Anion Channel in Native Mitochondrial Outer Membranes

161

The outer mitochondrial membrane houses specific proteins, which facilitate metabolic coupling and signaling between the cytosol and mitochondria. This membrane must be tight, because stress-induced release of cytochrome c leads to apoptosis. The voltage-dependent anion channel (VDAC), a general diffusion pore exhibiting a diameter of 2–3 nm, mediates a major part of the molecular traffic. These porins have a molecular mass of around 30 kDa per functional channel, and have been found in all eukaryotic organisms. For membrane potentials >|20  mV|, VDACs are in a state that has a reduced conductance. Electron microscopy has shown that these pores form lattices whose unit cells comprise six VDACs when native mitochondrial outer membranes are treated by phospholipase A2 (33). NMR and x-ray analyses have produced several atomic structures of the single channel (34). However, only recently it has been possible to reveal these channels directly in the native outer membranes of mitochondria from potato (35) and from yeast (36). In some membrane domains, VDACs were found to be packed at high density like bacterial outer membrane porins (36), whereas in other domains, VDACs were loosely packed, exhibiting single pores and oligomeric clusters comprising two, three, four, and six channels (35) (Fig. 4). The strength of FM-AFM compared to that of contact mode AFM is demonstrated by the images of the VDAC in native potato mitochondrial membranes (35). Whereas in contact mode small oligomers were hard to observe and single VDAC channels were not found, monomers, dimers, trimers, tetramers, and hexamers

Fig. 4. AFM topographs of the voltage-dependent anion channel in outer mitochondrial membranes (OMMO). (a) Topography of OMM patches. Two different types of surface are evident: (1) the OMM surface and (2 ) the mica surface. (b) VDAC proteins arranged in hexagons are visualized by contact mode AFM. The selected VDAC hexagons marked by broken squares are displayed in the gallery at the bottom. Scale bar represents 75 nm and the frame size of the magnified particles in the gallery is 18 nm. (c) The surface structure and organization of single VDAC proteins in the native OMM are revealed by frequency modulation AFM. Various oligomeric states of the VDAC are displayed in the gallery at the bottom: monomers (marked 1 ), dimers (2 ), trimers (3 ), tetramers (4 ) and hexamers (6 ). The scale bar represents 75 nm and the frame size of the boxes in the gallery is 21 nm.

162

Engel

were unambiguously identified in FM-AFM. This observation suggests that FM-AFM induces smaller lateral forces, thereby making the observation of single channels embedded in the bilayer possible.

4. Conclusion and Perspectives Since its invention more than two decades ago the AFM has become an important tool for structural biologists. It is the only instrument that allows surfaces of cells, supramolecular assemblies, and single molecules to be imaged in the native aqueous environment at nanometer-scale resolution. In addition, it makes manipulation of such structures at this scale possible. Recent developments of cantilevers, deflections sensors, imaging modes, and fast-scan systems demonstrate the potential possibilities to improve the AFM. Such progress will enhance the applicability of AFM to a wider range of biological questions, and will allow data to be acquired more efficiently than hitherto possible. AFM images of native membranes at submolecular resolution have provided a wealth of novel insights, and it is likely that this particular application of AFM will yield further important results. Measurements of forces between cells, within supramolecular aggregates, or forces dictating the fold of proteins can now be executed with great efficiency, allowing large datasets to be acquired, delivering quantitative information that were hitherto not accessible. As instrumentation development progresses, sample preparation methods have to be improved as well. There is room for improving the immobilization of native biological membranes to make them accessible to high-resolution imaging and manipulation with the AFM. The rapid progress over the past few years promises the AFM to deliver substantial new information about the structure, dynamics, and function of diverse native biological membranes.

5. Notes 5.1. Preparing the Sample Holders

Glue should be completely hardened and be uniformly distributed between mica and Teflon (or support), and be devoid of air bubbles. In AFMs that displace the supporting surface, air bubbles between mica and support can cause vibration or drift on the nanometer scale. Before mounting the sample support in the AFM, clean all contact surfaces using propanol and/or ethanol. Even small particles between the support and AFM can cause vibrations and drift.

Imaging and Interrogating Native Membrane Proteins Using the AFM

163

5.2. Damping of Vibrations

For high-resolution AFM imaging, an acoustic and vibration-isolated setup of the microscope is crucial. Isolate AFM from sources that may cause electronic and mechanic noise and thermal drift. Noise may be detected by vibration detectors or by scanning the mica surface in buffer solution at minimal forces of 50 pN or less. Sources of electronic noises may be found by switching off the devices individually. The AFM should be placed on an actively or passively damped table. For acoustic isolation, a glass bell may be used.

5.3. Choosing Suitable AFM Cantilevers

For achieving the best possible results, the cantilever properties must be adapted to the experiment. For high-resolution contact mode imaging AFM, cantilevers should be soft (0.1 N/m) and exhibit resonance frequencies in buffer solution, which allow tracking the surface features at the scanning speed applied (see (37)). For oscillating mode imaging, the cantilevers can be up to 30 times stiffer since the amplitude changes of the oscillating cantilever can be detected with sufficient accuracy to sense even very subtle force differences, which is required to prevent the deformation of membrane proteins (12). For specific applications, even stiffer cantilevers may be needed, but the sensitivity of the deflection detection will impose limits. In all cases, the cantilever stylus should have a nominal radius of less than 10 nm. SMFS requires soft cantilevers having a high resonance frequency. Otherwise, small forces may not be detected and the maximum sampling rate of the cantilever limits capturing fast unfolding events.

5.4. Sample Preparation

For AFM experiments, membranes are usually adsorbed to a chemically inert hydrophilic and flat solid support by properly adjusting pH and ionic strength. Such solid-supported membranes have allowed, and will still allow, important insights to be gained into the structure–function relationship of native membrane proteins. Solid supports that have proven suitable for highresolution imaging of membrane proteins include mica (28), HOPG (9, 38, 39), molybdenum disulfide (38), template-stripped gold (38, 40), and template-stripped platinum (38). Templatestripped metal surfaces appear to be particularly useful for combined topographical and electronic measurements (38). Detailed, step-by-step protocols for preparing biological membranes and for AFM imaging have been provided (7). Although high-resolution imaging is possible exclusively on solid-supported membranes, certain question may not be addressed by this preparation method. For example, membrane proteins in membranes directly attached to the support often exhibit impaired mobility (41–43), because the gap between membrane and support is only 0.5–2 nm. Moreover, adsorption forces may influence the conformation of membrane proteins, and it is known that lipids of a solid-supported lipid bilayer can show different structural features than the lipids of a vesicle or

164

Engel

a freestanding lipid bilayer (42). Various schemes have been proposed to circumvent this problem by using spacers that warrant a larger gap, or polymer cushions (44). Freestanding bacterial S-layers spanned over small wells have been imaged at high resolution in the AFM (45), representing an ideal situation. Nevertheless, there is room for further progress in sample preparation strategies to study structure and function of native membrane protein assemblies by AFM. 5.5. Imaging Membrane Channels at High Resolution Using the AFM

Tip, cantilever, deflection detector, piezo elements, fluid cell, and the electronic control system dictate the performance of an AFM. Since the early days, the tip was subject to attempts making it sharper, more reproducible, and more robust. However, even the best tip will inevitably change upon interaction with the sample. Therefore, it is up to the experience and skill of the operator to judge the performance of the tip to collect data when it is devoid of contaminants. Whereas the tip apex dominates the lateral resolution of an AFM, the vertical detection limit is ultimately given by thermal fluctuations of the cantilever, whose properties have been analyzed in great detail (for a recent summary see (37)). The AFM can be operated in several imaging modes that seek to minimize variations in the tip–sample interaction. In contact mode, the vertical deflection signal is the servo loop input for maintaining a constant deflection by a vertical displacement of sample or cantilever. Both the servo in- and output can produce AFM images. The deflection signal (i.e., the small deviation from a set cantilever deflection that the servo attempts to maintain) is mainly used to visualize topographical features in images with large height differences, such as overview images. Because it is a differential signal, edges are enhanced, allowing membrane patches to be easily identified. Lateral forces acting between the tip and the sample in contact mode often leads to a displacement of the sample and can limit the obtainable resolution. In oscillation type AFM, the cantilever is only intermittently in contact with the sample, thus reducing lateral forces. The servo loop in oscillation mode AFM uses the amplitude, phase, or frequency change for feedback. In FM-AFM a phase-locked loop measures the frequency shift, which is used as input for the servo loop operating the z-piezo. An advantage of FM-AFM is that the frequency shift gives quantitative information on the force acting between the tip and the sample (12, 46). When running the FM-AFM with small amplitudes (£1  nm), a most stable and quantitative operation yielding high-resolution images was achieved (12, 47). While these advances in instrumentation and novel imaging modes indeed make the acquisition of high-resolution images more reproducible, it is the tip–sample interactions that ultimately dictate the quality of topographical information obtained. Ionic strength and pH have a critical influence on the tip–sample interactions, and thus provide a handle to optimize them. Ideally, the

Imaging and Interrogating Native Membrane Proteins Using the AFM

165

Fig. 5. Selecting buffer conditions for high-resolution AFM imaging. (a) Small asperities of the AFM stylus contour the sample topography at high resolution. Interactions between AFM stylus and protein membrane can be divided into long- and short-range interaction forces. Long-range electrostatic repulsion forms a cushion that compensates forces applied to the tip. (b) Force–distance curves recorded reveal electrostatic repulsion most clearly from (1) to (2 ). Increasing ion concentration and valency screens electrostatic interactions, revealing the van der Waals attraction at high ionic strength (lowest curve). Balancing van der Waals attraction and electrostatic repulsion is used to promote adsorption of the sample onto the support, and to minimize the forces between asperities of the AFM stylus and the protein.

force applied to the tip (i.e., the set cantilever deflection) to obtain stable operation in the contact mode should be distributed over a significant surface area, while the tip apex barely touches the protein surface. In low ionic strength buffers, electrostatic forces can have a decay length of 100 nm or more, which is observed by the deflection of the cantilever long before the tip contacts the specimen surface. In contrast, van der Waals attraction has a range of about 1 nm and depends neither on pH nor on ionic strength. Therefore, the buffer can be used to adjust the electrostatic contribution to counteract the van der Waals attraction and to distribute the force applied to the tip, thereby preventing severe tip-induced sample deformations (25) (Fig. 5). The strategy for optimizing the recording buffer has been detailed recently (7).

Acknowledgments The author thanks Daniel J. Müller, Dimitrios Fotiadis, and Bart Hoogenboom for providing beautiful topographs and constructive discussions. This work was supported by the Maurice E. Müller Foundation of Switzerland and by the Swiss National Foundation. The used AFM facility was built with contributions from the Swiss University Conference and JPK-Instruments AG, Berlin, Germany.

166

Engel

References 1. Binnig, G., C. F. Quate, and C. Gerber. 1986. Atomic force microscopy. Phys. Rev. Lett. 56:930–933. 2. Drake, B., C. B. Prater, A. L. Weisenhorn, S. A. Gould, T. R. Albrecht, C. F. Quate, D. S. Cannell, H. G. Hansma, and P. K. Hansma. 1989. Imaging crystals, polymers and processes in water with the atomic force microscope. Science 243:1586–1589. 3. Hoh, J. H., R. Lal, S. A. John, J. P. Revel, and M. F. Arnsdorf. 1991. Atomic force microscopy and dissection of gap junctions. Science 253:1405–1408. 4. Buzhynskyy, N., J. F. Girmens, W. Faigle, and S. Scheuring. 2007. Human cataract lens membrane at subnanometer resolution. J Mol Biol 374:162–169. 5. Czajkowsky, D. M., E. M. Hotze, Z. Shao, and R. K. Tweten. 2004. Vertical collapse of a cytolysin prepore moves its transmembrane beta-hairpins to the membrane. Embo J 23:3206–3215. 6. Kienberger, F., H. Mueller, V. Pastushenko, and P. Hinterdorfer. 2004. Following single antibody binding to purple membranes in real time. EMBO Rep 5:579–583. 7. Müller, D. J., and A. Engel. 2007. Atomic force microscopy and spectroscopy of native membrane proteins. Nat Protoc 2:2191–2197. 8. Müller, D. J., G. M. Hand, A. Engel, and G. E. Sosinsky. 2002. Conformational changes in surface structures of isolated connexin 26 gap junctions. EMBO Journal 21:3598–3607. 9. Müller, D. J., and A. Engel. 1999. Voltage and pH-induced channel closure of porin OmpF visualized by atomic force microscopy. J Mol Biol 285:1347–1351. 10. Engel, A., and H. E. Gaub. 2008. Structure and mechanics of membrane proteins. Annu Rev Biochem 77:127–148. 11. Kedrov, A., H. Janovjak, K. T. Sapra, and D. J. Müller. 2007. Deciphering molecular interactions of native membrane proteins by single-molecule force spectroscopy. Annu Rev Biophys Biomol Struct 36:233–260. 12. Hoogenboom, B. W., H. J. Hug, Y. Pellmont, S. Martin, P. L. T. M. Frederix, D. Fotiadis, and A. Engel. 2006. Quantitative dynamicmode scanning force microscopy in liquid. Appl. Phys. Lett. 88:193109. 13. Nikaido, H., and M. Vaara. 1985. Molecular basis of bacterial outer membrane permeability. Microbiol. Rev. 49:1–32.

14. Schindler, H., and J. P. Rosenbusch. 1978. Matrix protein from Escherichia coli outer membranes forms voltage-controlled channels in lipid bilayers. Proc. Natl. Acad. Sci. USA 75:3751–3755. 15. Brunen, M., and H. Engelhardt. 1993. Asymmetry of orientation and voltage gating of the Acidovorax-delafieldii porin Omp34 in lipid bilayers. Eur. J. Biochem. 212:129–135. 16. Sen, K., J. Hellman, and H. Nikaido. 1988. Porin channels in intact cells of Escherichia coli are not affected by donnan potentials across the outer membrane. J. Biol. Chem. 263:1182–1187. 17. Todt, J. C., W. J. Rocque, and E. J. McGroarty. 1992. Effects of pH on bacterial porin function. Biochem. 31:10471–10478. 18. Delcour, A. H. 1997. Function and modulation of bacterial porins: insights from electrophysiology. FEMS Microbiol Lett 151:115–123. 19. Cowan, S. W., T. Schirmer, G. Rummel, M. Steiert, R. Ghosh, R. A. Pauptit, J. N. Jansonius, and J. P. Rosenbusch. 1992. Crystal structures explain functional properties of two E. coli porins. Nature 358:727–733. 20. Schirmer, T. 1998. General and specific porins from bacterial outer membranes. J. Stuct. Biol. 121:101–109. 21. Schulz, G. 1993. Bacterial porins: structure and function. Curr. Opin. Cell Biol. 5:701–707. 22. Phale, R. S., T. Schirmer, A. Prilipov, K.-L. Lou, A. Hardmeyer, and J. Rosenbusch. 1997. Voltage gating of Escherichia coli porin channels: role of the constriction loop. Proc. Natl. Acad. Sci. USA 94:6741–6745. 23. Schabert, F. A., C. Henn, and A. Engel. 1995. Native Escherichia coli OmpF porin surfaces probed by atomic force microscopy. Science 268:92–94. 24. Fotiadis, D., S. Scheuring, S. A. Müller, A. Engel, and D. J. Müller. 2002. Imaging and manipulation of biological structures with the AFM. Micron 33:385–397. 25. Müller, D. J., D. Fotiadis, S. Scheuring, S. A. Müller, and A. Engel. 1999. Electrostatically balanced subnanometer imaging of biological specimens by atomic force microscopy. Biophys J 76:1101–1111. 26. Müller, D. J., and A. Engel. 1997. The height of biomolecules measured with the atomic force microscope depends on electrostatic interactions. Biophys J 73:1633–1644.

Imaging and Interrogating Native Membrane Proteins Using the AFM 27. Israelachvili, J. 1991. Intermolecular & surface forces. Academic Press Limited, London. 28. Müller, D. J., M. Amrein, and A. Engel. 1997. Adsorption of biological molecules to a solid support for scanning probe microscopy. J Struct Biol 119:172–188. 2 9. Karshikoff, A., V. Spassov, S. W. Cowan, R. Ladenstein, and T. Schirmer. 1994. Electrostatic properties of two porin channels from Escherichia coli. J. Mol. Biol. 240:372–384. 30. Andersen, C., B. Schiffler, A. Charbit, and R. Benz. 2002. PH-induced collapse of the extracellular loops closes Escherichia coli maltoporin and allows the study of asymmetric sugar binding. J Biol Chem 277:41318–41325. 31. Yildiz, O., K. R. Vinothkumar, P. Goswami, and W. Kuhlbrandt. 2006. Structure of the monomeric outer-membrane porin OmpG in the open and closed conformation. Embo J 25:3702–3713. 32. Mari, S. A., S. Koster, C. A. Bippes, O. Yildiz, W. Kuhlbrandt, and D. J. Müller. 2010. pHinduced conformational change of the betabarrel-forming protein OmpG reconstituted into native E. coli lipids. J Mol Biol 396:610–616. 33. Mannella, C. A. 1986. Mitochondrial outer membrane channel (VDAC, porin) twodimensional crystals from Neurospora. Methods Enzymol 125:595–610. 34. Hiller, S., and G. Wagner. 2009. The role of solution NMR in the structure determinations of VDAC-1 and other membrane proteins. Curr Opin Struct Biol 19:396–401. 35. Hoogenboom, B. W., K. Suda, A. Engel, and D. Fotiadis. 2007. The supramolecular assemblies of voltage-dependent anion channels in the native membrane. J Mol Biol 370:246–255. 36. Goncalves, R. P., N. Buzhynskyy, V. Prima, J. N. Sturgis, and S. Scheuring. 2007. Supramolecular assembly of VDAC in native mitochondrial outer membranes. J Mol Biol 369:413–418. 37. Frederix, P. L., P. D. Bosshart, and A. Engel. 2009. Atomic force microscopy of biological membranes. Biophys J 96:329–338.

167

38. Frederix, P. L. T. M., P. D. Bosshart, T. Akiyama, M. Chami, M. R. Gullo, J. J. Blackstock, K. Dooleweerdt, N. F. de Rooij, U. Staufer, and A. Engel. 2008. Conductive supports for combined AFM-SECM on biological membranes. Nanotechnology 19. 39. Scheuring, S., D. J. Müller, P. Ringler, J. B. Heymann, and A. Engel. 1999. Imaging streptavidin 2D crystals on biotinylated lipid monolayers at high resolution with the atomic force microscope. J Microsc 193:28–35. 40. Cisneros, D. A., D. J. Müller, S. M. Daud, and J. H. Lakey. 2006. An approach to prepare membrane proteins for single-molecule imaging. Angew Chem Int Ed Engl 45:3252–3256. 41. Müller, D. J., A. Engel, U. Matthey, T. Meier, P. Dimroth, and K. Suda. 2003. Observing membrane protein diffusion at subnanometer resolution. J Mol Biol 327:925–930. 42. Tanaka, M., and E. Sackmann. 2005. Polymersupported membranes as models of the cell surface. Nature 437:656–663. 43. Wagner, M. L., and L. K. Tamm. 2000. Tethered polymer-supported planar lipid bilayers for reconstitution of integral membrane proteins: silane-polyethyleneglycol-lipid as a cushion and covalent linker. Biophys J 79:1400–1414. 44. Müller, D., and A. Engel. 2008. Strategies to prepare and characterize native membrane proteins and protein membranes by AFM. Curr Opin in Colloids & Interface Science 13:338–350. 4 5. Goncalves, R. P., G. Agnus, P. Sens, C. Houssin, B. Bartenlian, and S. Scheuring. 2006. Two-chamber AFM: probing membrane proteins separating two aqueous compartments. Nat Methods 3:1007–1012. 46. Uchihashi, T., M. J. Higgins, S. Yasuda, S. P. Jarvis, S. Akita, Y. Nakayama, and J. E. Sader. 2004. Quantitative force measurements in liquid using frequency modulation atomic force microscopy. Applied Physics Letters 85:3575–3577. 47. Fukuma, T., K. Kobayashi, K. Matsushige, and H. Yamada. 2005. True atomic resolution in liquid by frequency-modulation atomic force microscopy. Appl. Phys. Lett. 87:034101.

wwwwwwwwwwwwww

Part III Nanoscale Surface Analysis and Cell Imaging

wwwwwwwwwwwwww

Chapter 12 Atomic Force Microscopy Investigation of Viruses Alexander McPherson and Yurii G. Kuznetsov Abstract Atomic force microscopy (AFM) has proven to be a valuable approach to delineate the architectures and detailed structural features of a wide variety of viruses. These have ranged from small plant satellite viruses of only 17 nm to the giant mimivirus of 750 nm diameter, and they have included diverse morphologies such as those represented by HIV, icosahedral particles, vaccinia, and bacteriophages. Because it is a surface technique, it provides images and information that are distinct from those obtained by electron microscopy, and in some cases, at even higher resolution. By enzymatic and chemical dissection of virions, internal structures can be revealed, as well as DNA and RNA. The method is relatively rapid and can be carried out on both fixed and unfixed samples in either air or fluids, including culture media. It is nondestructive and even non-perturbing. It can be applied to individual isolated virus, as well as to infected cells. AFM is still in its early development and holds great promise for further investigation of biological systems at the nanometer scale. Key words: Imaging, Nanoscale, Structure, Infection, Nucleic acids, Icosahedra

1. Introduction A direct imaging technology that promises to have a significant impact on structural biology, and which is, in most ways, complementary to X-ray diffraction and electron microscopy, the classical approaches, is atomic force microscopy (AFM) (1–3). An immediate advantage of AFM is that it is based on relatively simple physical principles, unlike X-ray crystallography, and the instruments are mechanically and electronically rather straightforward, unlike electron microscopy. Unlike both of the other technologies, AFM is fairly inexpensive to institute and apply, even to biological specimens. The acuity and investigative size range of the AFM have proven to be quite remarkable and it is now permitting researchers new access to virus structure and the effects of viruses on organisms. Indeed, it has allowed us to Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_12, © Springer Science+Business Media, LLC 2011

171

172

McPherson and Kuznetsov

visualize the surface features, internal structures, and the nucleic acid cores of many viruses. At the same time, it has proven to be an effective instrument for observing viruses emerging from animal cells, and the perturbations they produce to the cells.

2. Materials 1. Glutaraldehyde for sample fixation was made by diluting a commercial solution to a concentration of 5% w/v with distilled water, and kept in a light-free container. Glutaraldehyde should be freshly prepared at least every 2 weeks and maintained at 4°C. 2. Poly-l-lysine for coating cover slips was prepared by dissolving lyophilized polymer in distilled water to a final concentration of 1 mg/ml. This stock solution, maintained at 4°C, was further diluted to 0.1  mg/ml before application to substrates. 3. Magnesium and nickel chloride for treating substrates were made by dissolving reagent-grade salts in distilled water to concentrations of 50 mM. 4. Substrates for AFM were acid-washed glass or plastic, 1-cm diameter cover slips, which were extensively rinsed with distilled water. Alternately, the substrates were freshly cleaved mica. 5. Virus samples were diluted from their stocks into distilled water when possible, and into phosphate-buffered saline otherwise. Substrates with samples, unless air-dried directly, were rinsed with distilled water before air drying. 6. For scanning in fluids, the fluid cell was filled with either distilled water or phosphate-buffered saline. 7. Any other reagents, such as enzymes, detergents, or reducing agents, were prepared from the highest grade materials available and dissolved in distilled water or phosphate-buffered saline. These were maintained at −20°C in small aliquots and thawed for use as needed.

3. Methods 3.1. AFM Instrument Setup

AFM instruments can be operated in either contact mode, or what is referred to as tapping mode. In contact mode, a probe made of silicon or silicon nitride is placed in near contact with the surface of interest, say the capsid of a virus, and then translated in a systematic raster mode over the surface. The AFM probe is a

Atomic Force Microscopy Investigation of Viruses

173

sharp stylus similar in function to a minute phonograph needle. The tip ideally has a single point, with a very small radius of curvature. The probe is mounted at the end of a short cantilever, typically 100–250 mm in length, which has a low spring constant to minimize the force between the tip and the sample. Scanning is achieved by translating the sample beneath the probe, using a piezoelectric positioned x–y stage, along a continuous sequence of raster lines. As the probe tip passes over the surface, it interacts through “aggregate atomic forces,” which remain somewhat mysterious, with structural features on the surface. Encounters with these substructures cause the probe to be displaced vertically as the tip rides across. Exceedingly small displacements of the tip are amplified by deflection of a laser beam that is reflected from the upper surface of the cantilever, and these deflections are detected and tracked by a split photodiode. Photoelectric circuitry converts the deflections into height information. The resulting scan data, recorded as a digital topographical image, can then be presented in a number of visual formats. Sample perturbation and other problems arising from unfavorable probe–surface interactions have been obviated to a great extent by the development of “tapping” mode instruments (4). With tapping mode, the probe tip is not in continuous contact with the sample surface, but rapidly oscillates up and down as it is scanned over the surface, essentially “tapping” its way and gently sensing the heights of obstacles it encounters. In tapping mode, the vertical position of the sample is continually adjusted by a feedback mechanism to maintain the amplitude of the freely oscillating probe constant (see Note 1). The “tapping mode” approach has proven to be a significant boon in biological investigations as it has allowed the characterization of samples that would otherwise be too soft or too fragile to withstand contact mode examination. Operating with tapping mode in a liquid environment presents some complications due to fluid dynamics, but these are not severe. A constraint that sometimes presents obstacles during analysis in a liquid medium is that the specimen under study must be fixed to, or made to adhere firmly to the substrate surface of the fluid cell, which may be glass, cleaved mica, plastic, or any other hard material (see Note 2). One particular feature of AFM must be borne in mind whenever one is interpreting images. The one- or two-dimensional profile obtained of any object, or surface substructure, is the convolution of the tip shape with that of the feature being scanned. This is illustrated in Fig. 1. An image of an object scanned with a broad, dull tip is not the same as that acquired with a sharper tip. In particular, while the height of the object will be the same regardless of the tip shape (because the maximum vertical deflection of the cantilever tip would be the same), the lateral dimensions will not. A broader tip yields a broader object, and a sharper tip produces the more accurate size (see Note 3). Whereas height information is almost

174

McPherson and Kuznetsov

Fig. 1. Schematic illustration of the convolution of the shape of the AFM tip with the shape of the feature, or particle being scanned. The side of the cantilever tip contacts the object and begins to produce a deflection of the cantilever before the tip apex actually reaches the object. Similarly, the opposite side of the tip is still in contact with the object even after the apex itself has passed. Thus the total deflection implies a virtual lateral dimension for the object greater than its actual dimension. The difference between the virtual and actual dimensions is a function of the width of the cantilever tip. The sharper the tip, the more accurate the observed dimensions, and the greater the resolution attainable.

always trustworthy, lateral measurements are frequently suspect. The reliability of lateral measurements can, however, be increased if some standard having defined spatial features is first scanned and its known spacings or cell dimensions compared with those in the image. Such standards may be etched grids on silicon, or the surfaces of protein crystals (5). Height resolution for all samples is typically better than 1 nm (see Notes 4–6). Specimens, however, are not always best visualized under physiological conditions, particularly when high resolution is desired.

Atomic Force Microscopy Investigation of Viruses

175

Because cantilever tip pressure, even in “tapping mode,” may produce deformation, for example, of a cell membrane, in some cases fixation is the better option. This, as with light microscopy histological procedures, usually relies on glutaraldehyde, paraformaldehyde, or osmium tetroxide fixation, followed by dehydration and imaging in water–alcohol mixtures, or in air. These methods have been developed by microscopists for more than a century to preserve the natural morphology of a sample but still allow high-resolution imaging. While not as ideal as in situ observation, the cells are no longer alive or viruses infective, fine details of their structures can be visualized that would otherwise be obscured by membrane flexion. 3.2. Virus Imaging

The resolution of AFM, in the best of cases, is roughly that of current cryo-EM models (6). It is applied to individual particles and does not yield an average structure over an entire population as do many EM reconstructions. It does not require that a virus have symmetrical or uniform architecture, or even that all particles be the same in structure. Thus, it is equally applicable to small icosahedral viruses such as tomato bushy stunt virus, helical viruses such as tobacco mosaic virus, and completely irregular, complex viruses such as vaccinia or the retroviruses. There is no size restriction. It has been used to analyze small plant viruses such as turnip yellow mosaic virus (TYMV) (7) to massive icosahedral viruses such as PBCV-1, an algal virus (8), to mimivirus (9), the largest virus known. Viruses were first visualized by AFM in their crystalline form (7, 10, 11), as illustrated by Fig. 2, rather than as single isolated particles, in investigations of the growth of crystals of satellite tobacco mosaic virus (STMV) and TYMV (12–14). Because they were immobilized on the surfaces of crystals, conditions were suitable for direct imaging of even the small 17–30-nm diameter

Fig. 2. In (a), a low magnification AFM image shows the two-dimensional growth islands that characterize the surfaces of orthorhombic T = 3 Brome Mosaic Virus (BMV) crystals. In (b) is a high magnification AFM image of the surface of the same crystal. As with most virus crystals, vacancies frequently occur in clusters to produce large defects. In (c) is a low magnification AFM image of the reassembled T = 1 particles of BMV in a tetragonal lattice. The scan areas are (a) 2 mm × 2 mm, (b) 542 nm × 542 nm, (c) 272 nm × 272 nm.

176

McPherson and Kuznetsov

virions. Larger icosahedral plant viruses in crystalline form were studied subsequently (15–19). The first AFM studies of noncrystalline viruses were retroviruses on cell surfaces (20–22), again, principally because they were immobilized by their association with cell surfaces. Single particles of larger viruses, and helical viruses, were eventually visualized by AFM, and these included tobacco mosaic virus, cauliflower mosaic virus, Tipula iridescent virus (15, 23), herpes simplex virus (24), vaccinia virus (25, 26), and mimivirus (9). Although virus crystals were investigated using both contact and tapping mode, noncrystalline specimens were imaged exclusively with tapping mode, in both air and buffer. 3.3. Internal Structure Imaging

Because AFM images the surfaces of specimens, it might be thought that AFM would be of little use in visualizing the interior features of viruses or cells. This, however, is not the case. As has been shown in AFM investigations of a number of viruses, it is, in fact, an invaluable tool for deducing the interior architecture of virions, regardless of their external form or size. This is because it is possible to strip away layers of structure systematically by chemical, physical, and enzymatic means (23, 26, 27) and to accompany this process of dissection by AFM visualization. Using the same strategy as that used by conventional anatomists, it has been proven possible to disassemble viral specimens, see what is inside, and ascertain how the components are linked.

3.4. Rapid Shape Classification

A valuable qualitative result that emerges almost immediately from AFM images is what the virus looks like, what is its overall architecture, and how similar are particles to one another. Are they uniformly the same in appearance, or are there a variety of forms? Thus even a cursory investigation may quickly reveal certain general features that allow rapid classification. This is illustrated by the various structural classes of viruses shown in Fig. 3. The virions may be spherical, cylindrical, or filamentous. They may have symmetrically arranged capsomeres or other surface units, fibers, protruding vertices, prolate or icosahedral shapes, unusual morphologies, pleiomorphic character, etc. Tail assemblies may be observed directly, as on phages for example. AFM is, therefore, a useful tool for simply deducing the kind of virus one is dealing with, whether more than one kind of virus is present in a population, and the general level of contamination that may accompany the virus as a consequence – cellular material, degraded virions, and macromolecular impurities of all sorts.

3.5. Quantitative Dimensional Measurements

A fundamental parameter for virus particles is their diameter if they are spherical viruses, or their diameter and length if they are helical. AFM can provide measures of these in both the hydrated and dried states, which also gives an estimate of the degree of shrinkage they undergo as a result of dehydration. Because of the finite tip size, and tip-to-tip variation in radius of curvature, it is

Atomic Force Microscopy Investigation of Viruses

177

Fig. 3. A variety of viruses seen at relatively low magnifications by AFM. In (a) vaccinia virus, in (b) the retrotransposon Ty3, in (c) Tobacco Mosaic Virus, in (d) T4 bacteriophage, in (e) PBCV-1 algal virus, an irridovirus, and in (f) Moloney Mouse Leukemic Virus. The scan areas are (a) 850  nm × 850  nm, (b) 500  nm × 500  nm, (c) 200  nm × 200  nm, (d) 250 nm × 250 nm, (e) 185 nm × 185 nm, and (f) 200 nm × 200 nm.

risky to measure linear dimensions directly by AFM (see Note 3). It is, however, safe to measure the heights of objects above the substrate plane, and the distances between the points of maximum elevation (e.g., capsomere to capsomere) on particles, or centerto-center distances (e.g., particles in a crystal or in a cluster). As has been emphasized already, for spherical and cylindrically symmetric particles, measurements of particle heights above the substrate plane yield reasonably accurate values for their dia­ meters, and individual measurements are usually accompanied by rather modest error, generally of the order of 5% or less. By repeating measurements for a number of particles in the field, and using different scan directions, good statistics can be obtained, and histograms of size distributions compiled. Precision of a few angstroms is possible. Histograms of particle sizes, as illustrated in Fig. 4, are often informative (8, 22) (see Note 7). 3.6. Topography as a Function of Composition and Architecture

The surfaces of virus particles vary topographically as a function of their composition and architectures. Plant viruses, for example, generally exhibit protein capsids with few embellishments, and this is true of many animal viruses and bacteriophages as well. These capsids are generally based on icosahedral architectures,

178

McPherson and Kuznetsov

Fig. 4. The heights above substrate level were measured for about 200 isolated MuLV virions and plotted as a histogram. The spread of sizes is not due to error in the measurements, which is only a few nanometers at most, but represents the real variation in size of particles produced in infection. The very large and very small particles are aberrant virions.

Fig. 5. A phase-contrast AFM image of a herpes simplex virion absorbed onto mica. The large sheet of white material around the capsid is the membrane envelope of the virus that has been partially discarded. The capsid lattice of the virus is clearly evident. The scan area is 770 nm × 770 nm.

and clusters of coat protein subunits, or capsomeres, are symmetrically distributed (28, 29). Many animal viruses, on the contrary, though they may contain an icosahedral capsid in their interior, often have either a lipid membrane over their surface, as does the herpes virus in Fig. 5, a covering of protein clusters, or even a hair-like coating of fibers. These various surfaces are readily apparent by AFM, and can be identified and delineated with a high degree of precision with the aid of some histological procedures, such as osmium tetroxide fixation, or protease treatment.

Atomic Force Microscopy Investigation of Viruses

179

Fig. 6. At higher magnification, the surface features of many viruses emerge. Seen here in (a) is the surface lattice of PBCV-1, a giant algal virus belonging to the irridovirus ­family, in (b) in vitro reassembled particles of the Gag protein from Mason–Pfizer Monkey Virus, in (c) Ty3 retrotransposons, and in (d) the capsid of the giant Mimivirus. The scan areas are (a) 42 nm × 42 nm, (b) 63 nm × 63 nm, (c) 117 × 117 nm, and (d) 200 nm × 200 nm.

Icosahedral capsids, or bullet-shaped or elongated capsids based on that symmetry, can be characterized in terms of the structure of the fundamental capsomere, along with the icosahedral triangulation number, T (29). Some examples are shown in Fig. 6. This will vary from small integral numbers like T = 1 for satellite viruses, or the T = 1 reassembly particles of Brome Mosaic Virus (30) to T = 3 and higher for more conventional, small icosahedral viruses such as poliovirus or TYMV, to very large numbers for complex viruses such as the irridoviruses such as PBCV-1 (T = 169) and mimivirus (T one of nine possibilities lying between 972 and 1,200). In many cases, the exterior shell of a virus may not be icosahedral, but it might possess an inner capsid which is. For example, though membrane covered and of pleiomorphic external shape, herpes simplex virus possesses a nucleic acid containing capsid of icosahedral form T = 16. Mimivirus exhibits a complex outer surface coated with a forest of fibers, but it too contains an

180

McPherson and Kuznetsov

Fig. 7. On the left is a capsid of the Ty3 retrotransposon. If the pattern of five- and sixfold capsomeres are plotted on its surface, as on the right, then it can be deduced that the virion has T = 4 icosahedral symmetry as shown in the center drawing. Other Ty3 virions were shown to have T = 3 and T = 7l icosahedral architectures.

icosahedral core (9). The T number, then, provides much of the information one needs to describe an icosahedral capsid. The triangulation numbers of icosahedral viruses can frequently be deduced from AFM images, as for the retrotransposon Ty3 in Fig. 7. With very large icosahedral capsids, which include PBCV-1 and mimivirus, one determines the two indices h and k (29), which define T (T = h2 + hk + k2). This is done by following a row of hexagonal capsomeres from one pentagonal vertex to the next icosahedral edge, and by simply counting the number of capsomeres along one edge h and the other k (the h and k coordinates of the intersection point on the icosahedral edge) that one needs to traverse (31) (see Note 8). In the cases of Ty3 retrotransposon and of Mason–Pfizer monkey virus (MPMV) in Fig. 8, for example, particles of their truncated Gag proteins were reassembled in vitro and imaged at high magnification by AFM. From the images, individual protein subunits were visible, and this allowed the discrimination of two possible models for the capsomeres (5). A similar analysis was used in the case of the large algal virus PBCV-1 (8). Knowing the diameters of capsomeres is often of considerable importance, even when individual subunits cannot be resolved. In mimivirus, for example, capsomere diameter provided a crucial clue in delineating the capsid architecture and permitting subsequent detailed analysis and reconstruction by cryo-EM (9). Although capsids of native HIV have yet to be visualized by AFM, helical tubes of capsid protein reassembled in  vitro have (5). In these tubes, a hexagonal arrangement of coat proteins could be clearly seen, and this provided support for a capsid model based on modified icosahedral architecture (32). The tubes reassembled from HIV Gag protein should remind us that helical and rod-shaped structures having periodic substructure are also excellent specimens

Atomic Force Microscopy Investigation of Viruses

181

Fig.  8. Virus-like particles (VLP) reassembled in  vitro or in recombinant bacteria often appear to have icosahedral ­symmetry, but upon closer inspection, they do not, the pentameric vertices are replaced by random defects or overall disorder. This is true of the Ty3 Gag particles in (a) and the mutant Gag of MPMV seen in (b). VLPs often take on entirely different shapes than the native capsids, as seen in the tubular forms of reassembled HIV Gag in (c). The scan areas are (a) 250 nm × 250 nm, (b) 125 nm × 125 nm, and (c) 200 nm × 200 nm.

for AFM analysis. These can appear in investigations of intact viruses and even in studies of spherical viruses when their interiors are explored. 3.7. Virus Sample Preparation Techniques 3.7.1. Isolated Particles

With AFM, it is not essential that highly purified virus particles be used as samples (15), although that might be ideal. Because individual particles can be investigated whenever a good specimen is spatially distinct from the surrounding rubble of proteins, cellular debris, and biological detritus, it may still yield excellent images. A problem, however, is that biological debris often adhere to and foul the AFM tip, severely degrading the quality of images. Contaminated tips are one of the most frustrating and annoying accompaniments of biological AFM.

3.7.2. Viruses on the Surface of Host Cells

Viruses on the surfaces of host cells may be visualized as well as free particles, and sometimes with better results because they are better immobilized (8, 20, 22, 33). Moloney mouse leukemia virus (MuLV) emerging from an infected 3T3 cell are clearly delineated in Fig. 9. They may be seen entering cells upon infection, or budding from cells after replication and assembly. This often provides valuable insights into which cells in a population are producing virus, the distribution of virus particles on the surface of the cells (are there preferred sites for budding?), and some details of the budding process itself.

3.7.3. Mutant Viruses: Anomalous Features

Mutant viruses, naturally occurring or produced in the laboratory, can be imaged as well as native virions and virus-like particles (VLP) created in vitro from capsid proteins. In some cases, the phenotype of the mutant can be revealed by observing infected host cells for unique or anomalous features. This was done, as shown in Fig. 10, in a study of MuLV-infected 3T3 cells, where a

182

McPherson and Kuznetsov

Fig. 9. Viruses can also be observed while emerging from, or stall attached to the host cell plasma membrane. In (a) is a low magnification AFM image of a 3T3 cell in culture infected with MuLV. The virus is clearly seen as white spots over the surface of the cell. In (b) is a higher magnification image showing four MuLV particles budding from a host cell in culture. In (c) is a mass of HIV virus bursting from the surface of a human lymphocyte in culture, and in (d), a higher magnification image of four HIV budding from a lymphocyte surface. The scan areas are (a) 10  mm × 10  mm, (b) 2  mm × 2  mm, (c) 5 mm × 5 mm, and (d) 460 nm × 460 nm.

mutant lacked glycosylated Gag protein (20, 34). Prior evidence suggested that such mutants failed in some stage of viral budding. This was confirmed by AFM visualization of infected, virus-producing cells. As seen in Fig.  10, instead of normal, spherical virus emerging from the cell surface, bullet- and ­comet-shaped protrusions were found distributed all over the plasma membrane of the host cells. The comets were viruses that were apparently trying to escape, but were unable to pull away and terminate association with the host cell. From this, it was concluded that the failure of glycosylation produced a defect in late stages of the budding process. Other mutations in virus genomes may produce alterations in external features of virus particles that are readily observable by AFM.

Atomic Force Microscopy Investigation of Viruses

183

Fig. 10. Two mutant forms of Moloney Mouse Leukemic Virus (MuLV) are shown in (a) and (b). In (a), the mutant lacks the ability to make the envelope protein, and the lipid membrane of the virus is observed directly. Its wavy pattern is due to its fluid nature and its motion when subjected to the pressure of the AFM cantilever. In (b) is another mutant which does not properly glycosylate its Gag protein, resulting in an inability to bud from the host cell properly, resulting in the comet-shaped structures on the surface of the host 3T3 cell in culture. The scan areas are (a) 250 nm × 250 nm. In (b) are composites of serial, positionally incremented AFM images representing total scan areas of 2 mm × 8.44 mm and 2 mm × 4.72 mm.

MuLV particles that failed to make an envelope protein (gp120 protein), one of which is seen in Fig. 10, were examined in another study (22). While normal particles are characterized by a coating of protein tufts, about 100–150 in number, mutant particles were “bald” virions lacking any such protein clusters. Instead, only an outer lipid membrane was visible. Some viruses exhibit special external structures, or deviations from their general architectures. For example, MuLV particles generally have a single small bump or a brief protrusion somewhere on their otherwise uniformly crenulated surfaces. These are likely to be a “budding scar” resulting from breaking away from the host cell (22). Other MuLV particles, perhaps defectives, exhibited small sectors on their surfaces where protein was absent and a channel into the interior appeared (20). Other, more prominent features are the surface fibers on the surfaces of mimivirus and the lateral bodies of vaccinia (seen in Fig. 11) (25, 26). PBCV-1 Chlorella Virus, an irridovirus, exhibited a unique pentagonal assembly of proteins at every fivefold vertex of its icosahedral capsid (8), shown in Fig. 12. The assembly had a single protein in the center that could “push in” and “pull out” as demonstrated by the application of AFM tip pressure. Its exact function is speculative. Many bacteriophages have tail assemblies of one sort or

184

McPherson and Kuznetsov

Fig. 11. AFM can capture a variety of specialized structures produced by viruses. In (a) is a fragment of the anchor protein, at the center, which connects a corona of glycosylated surface fibers to the capsid. In (b) is the contractile DNA injection assembly, or tail, of bacteriophage T4 showing helical architecture. In (c) is the “stargate” apparatus found at a unique vertex of Mimivirus which, upon opening, allows escape of the DNA.

Fig.  12. Fivefold vertices on large viruses which have icosahedral capsids often have unique clusters of proteins or unusual structures. In (a) and (b) are fivefold vertices of the large algal virus PBCV-1. In (c) and (d) is seen the stargate apparatus of Mimivirus opening to allow expulsion of the DNA inside the capsid. The scan areas are (a) 96 nm × 96 nm, (b) 100 nm × 100 nm, (c) 500 nm × 500 nm, and (d) 800 nm × 800 nm.

Atomic Force Microscopy Investigation of Viruses

185

another for packaging and injecting their DNA. Mimivirus is, in a sense, similar to these phages and has an assembly, seen in Fig. 12c, d, of presumably similar function at a single, unique, fivefold vertex. This star shaped structure is undoubtedly analogous to the tail assemblies of phages and is, as is evident in the figures, a distinctive feature of the virions. It is a complex structure likely composed of many proteins, and AFM reveals much of that complexity. 3.8. Population Analysis

An important point that deserves particular emphasis is that all of the particles within a population of virus are not absolutely identical, and often there are very significant differences in the detailed features of individual virus particles. This is a point often obscured by the results of X-ray crystallography (35, 36) or cryo-EM reconstructions (37, 38). These techniques rely totally on an assumption of structural conformity and produce models that represent the average in time and space of the individuals that make up the population. AFM, on the contrary, allows the revelation of the eccentricities and unique features of the individuals, and these are instructive. They often define the extremes of what is possible among a large population of viruses having, presumably, the same genome and the same environment for replication and assembly. What we see with AFM is that anomalous and aberrant individuals are not only present, but are also common.

3.9. Internal Structure Determination

One might think that because AFM provides images of the surfaces of objects and does not peer into their interiors, as do X-ray diffraction and electron microscopy, they would be of little value in delineating the interior structure of viruses, the layers beneath the external surface. This is not true, however, as we can apply the same technique that has been used by anatomists for centuries – dissection. With the aid of chemical, enzymatic, and physical tools, we can systematically pare a complex entity, including a virus, down to its core, layer by layer (23). At each stage, AFM may then be used to visualize what remains and what has been removed as well (see Note 9). Among the most useful agents for chemical dissection have been detergents, usually 0.5–2% of some nonionic detergents such as NP40, and reducing agents such as DTT or DTE. The former causes protein structure to unravel gradually and detergents strip away the lipid membrane. The latter reduces disulfide bonds and liberates polypeptides otherwise bonded to one another. Disulfide bond reduction appears to be particularly important in large, complex viruses where such covalent linkages cross-link coat proteins and stabilize capsids (25, 39) (see Note 10). The most effective enzymatic tools have been proteases that degrade polypeptides. These are particularly useful because they have a range of activities and a spectrum of specificities. As a consequence, a whole variety of proteases have been employed, including trypsin, bromelin, proteinase K, subtilisin, and mixtures

186

McPherson and Kuznetsov

of pancreatic proteases. Viruses are usually exposed to the ­proteases for anywhere from 15  min to several hours, or even overnight, at concentrations of 0.5 mg/ml to as high as 5 mg/ml (see Note 11). Physical forces have also been used to disrupt viruses, and often fortuitous perturbations, resulting simply from preparation and handling, have proven to be structurally illuminating. Heat, for example, was used to open TYMV (7) to release its encapsidated RNA, and direct physical pressure was used on mimivirus sandwiched between two layers of atomically smooth cleaved mica, as well. There are also instances where “hammering” of individual particles with the AFM tip has been utilized, taking advantage of the fact that AFM can serve as a tool as well as an imaging device. In carrying out the dissection of a virus, or even in simply visualizing particles spread on a glass, plastic, or mica substrate, it is necessary to ensure that the virus particles adhere firmly to the substrate. Failure to do so allows the particles to move beneath the AFM tip, rendering imaging impossible. Occasionally, altering the charge on the substrate is sufficient (see Notes 12 and 13). Altering charge is, however, frequently insufficient for virions. To fix most viruses to the substrate, as well as a wide variety of other biological entities and materials, an effective procedure is to coat the substrate with poly-l-lysine before depositing the virus. Presumably, salt bridges between the e amino groups of the lysines and the glutamic and aspartic acid carboxyl groups on the particles lock them in place. After such substrate–particle attachment, the substrate can be rinsed with water several times without loss of sample (see Note 14). It is occasionally unnecessary to actually treat viruses with any chemical or biochemical agent to view the interior, as the physical stress of preparation and purification may result in damaged or partially degraded particles. These may expose interior structural features that are otherwise not apparent. Retroviruses, in particular, are physically fragile. Some MuLV, as shown in Fig. 13, when subjected to the shear forces of centrifugation, lose portions of the shell surrounding the capsid. This permits direct visualization of the virus core still embedded within the layers of envelope and matrix protein (22). HIV is another example where even the mildest procedures produce some damaged virions. Although the cores of HIV have not yet been visualized by AFM, likely due to their fragility, the remainder of the virus without the cores has been (22). Some examples can be seen in Fig.  13c, d. Such partially disrobed particles, both MuLV and HIV, provide specimens that can be subjected to quantitative examination and thereby yield the dimensions, the thicknesses of internal structural layers, and they give some clues as to their components as well. The best example of a complete dissection of a complex virus using AFM is that of vaccinia virus, a pox virus of about 300-nm

Atomic Force Microscopy Investigation of Viruses

187

Fig. 13. Virions are sometimes damaged in their preparation, especially fragile viruses such as retroviruses. These occasions often provide images of interior structure that would otherwise not be accessible to the AFM tip. In (a) and (b) are MuLV whose outer shell has been partially sheared away to reveal the nucleic acid containing capsid inside. In (c) and (d) are similarly sheared HIV particles. The interior capsids have been lost, leaving the empty outer shells behind. The scan areas are (a) 200 nm × 200 nm, (b) 200 nm × 200 nm, (c) 500 nm × 500 nm, and (d) 250 nm × 250 nm.

diameter that is delimited by a lipid membrane (25, 26). It ­contains a double-stranded DNA genome bounded by several protein shells. It also has two unusual protein assemblies of still unknown function, known as lateral bodies, associated with its inner core. Vaccinia was sequentially degraded with a 0.5% NP40 nonionic detergent combined with 0.05  M DTT, followed by exposure to this same mixture but containing either trypsin or proteinase K, or to the proteases alone. Six stages in this process are presented in Fig.  14. At the end, the innermost core was breached and the DNA was exposed. 3.10. Imaging of Nucleic Acids of Viruses

The nucleic acids of viruses, some of which are seen in Fig. 15, from a structural standpoint, are of considerable interest, and in particular, how they are condensed and packaged inside capsids and cores. Clearly, packaging is accomplished differently by ­specific families of viruses. It is unlikely, for example, that

188

McPherson and Kuznetsov

Fig. 14. A series of AFM images presenting the dissection of vaccinia virus using a variety of enzymatic and chemical reagents to remove successive layers of structure. In (a) is the intact virus in buffer before any treatment. In (b), the lipid membrane is pronounced as a corona surrounding a virion dried upon the substrate. The lateral body at the center is also more pronounced on the dried virions. In (c), the outer protein shell has been etched away to reveal the inner protein capsid which is perforated and still has the lateral bodies connected to it. In (d) is a mass of capsids having lost both their lateral bodies and their DNA. In (e) are higher magnification AFM images of the lateral bodies that decorate the capsids, and in (f) is a mass of vaccinia DNA released onto the AFM substrate. The scan areas are (a) 400 nm × 400 nm, (b) 350 nm × 350 nm, (c) 350 nm × 350 nm, (d) 2 mm × 2mm, (e) 500 nm × 500 nm, and (f) 2 mm × 2 mm.

bacteriophage and pox virus package their genomic doublestranded DNA the same way. The packing densities of the nucleic acid differ by more than tenfold (26). Also, is it unlikely that large, single-stranded RNA-containing viruses, such as retroviruses, package their genomes the same way as do T = 1 or T = 3 icosahedral viruses (40). Certainly, helical and filamentous viruses use entirely different mechanisms. AFM investigations have been conducted on RNA extracted by phenol from a series of small icosahedral viruses, and from tobacco mosaic virus, the classical rod-shaped, helical virus (41). The spherical viruses included poliovirus, STMV, TYMV, and brome mosaic virus. In this study, the gradual unraveling of the tertiary structure of the RNA, and ultimately the secondary structure as well, could be produced in stages simply by heating. A counter example was provided by the rod-shaped, helical tobacco mosaic virus RNA which appeared initially as a thread, a completely extended molecule lacking any secondary structure. With time, it began forming local secondary structural elements

Atomic Force Microscopy Investigation of Viruses

189

Fig. 15. Expulsions or emissions of nucleic acid by a variety of viruses. In (a), a shattered Ty3 retrotransposon disgorges a mass of nucleic acid. Although it is known that the genome of the retrotransposon has two single strands of RNA as its genome, the nucleic acid seen here has the characteristics of double-stranded DNA. In (b), the DNA core of the algal virus PBCV-1 throws out a splash of double-stranded DNA. In (c) is a mass of DNA released by vaccinia virus upon degradation with proteases. In (d), a shattered virion of STMV spreads its single-stranded RNA genome of 1,058 nucleotides around itself. In (e), virions of TYMV are losing their single-stranded RNA genomes after loss of a capsomere. In (f), damaged T4 bacteriophages release their DNA on the AFM substrate. The scan areas are (a) and (b) 1 mm × 1 mm, (c) 5 mm × 5 mm, (d) 200 nm × 200 nm, (e) 500 nm × 500 nm, and (f) 2 mm × 2 mm.

and eventually condensed into forms similar to those seen for the RNA from the icosahedral viruses (41). One conclusion of the study was that the single strands of RNA spontaneously condensed as linear arrangements of stem-loop substructures following synthesis, the condensed RNA bound coat protein to it, and the two cooperatively coalesced into the completed particle. In studies such as these, AFM proved itself as an able technique for directly visualizing nucleic acid structure, demonstrating its fluidity, and suggesting the mechanisms by which it is encapsidated. DNA and RNA appear quite different in AFM images, and this is evident in Fig.  16, which presents both kinds of nucleic acids. The former looks like strands and coils of stiff rope lacking any higher levels of structure, while the latter appears as complicated, linear sequences of self-involved secondary structure. Sometimes, however, the distinction is not entirely clear and further evidence may be needed to show whether a filament, strand, or complex is DNA or RNA.

190

McPherson and Kuznetsov

Fig. 16. Viral nucleic acid can be readily visualized by AFM after expulsion from the capsid. In (a) is a DNA plasmid used in recombinant DNA research to transform bacteria. In (b) is a tangle of double-stranded DNA from vaccinia virus. In (c) is the genomic single-stranded RNA of STMV, still condensed due to secondary structure, and in (d), RNA from poliovirus escaping from a condensed core obtained by phenol extraction of the virus. The scan areas are (a) 587 nm × 587 nm, (b) 500 nm × 500 nm, (c) 220 nm × 220 nm, and (d) 500 nm × 500 nm.

A method was devised for additional identification based on exposure of the nucleic acid to high concentrations of bovine RNase A (42). RNA, naturally, was hydrolyzed to small pieces by RNase A and left only fragments on the substrate which corresponded to protected stem-loops. DNA, on the contrary, became coated with the protein and the resulting strands exhibited diameters two to three times that of naked double-stranded DNA. Thus it is possible to practice a kind of crude histology with AFM. A second example of histological AFM is the immunolabeling of viruses with antibodies specific for certain proteins. Although individual IgG are not clearly identifiable by AFM when bound to a virion, IgG conjugated with gold particles generally are. In a sense, these are used in the same way as they are used in transmission electron microscopy immunolabeling, except that instead of visualizing points of high electron density, one images with AFM objects having the size and shape of the immuno-gold particles.

Atomic Force Microscopy Investigation of Viruses

191

Fig. 17. Virus can be labeled with gold particles conjugated with antibody against specific viral proteins. These can be visualized using AFM as seen here. In (a) and (b) are two different MuLV virions that have been exposed to antibody/gold particles directed against the envelope protein. Arrows indicate the gold particles. The scan areas are (a) 500 nm × 500 nm and (b) 600 nm × 600 nm.

Using gold–IgG conjugate particles against the envelope ­protein, as shown in Fig. 17, it was possible to show that protein tufts on the surfaces of MuLV were indeed envelope proteins (20, 22). The major problem with IgG–gold conjugates at this point is that their physical size limits the resolution of the method. Conjugated gold particles can bind only as close as their diameters allow. The answer to the question, what can AFM visualize that is of value to the structural biology of viruses, is that it can visualize virtually every part of a virus, and to resolutions that approach, and in some cases surpass, those of electron microscopy. At this time, lipid membranes have been identified, both RNA and DNA have been visualized, and large protein assemblies resolved. The capsids of icosahedral viruses, and the icosahedral capsids of nonicosahedral viruses have been seen at high resolution, in some cases sufficiently high to deduce the arrangement of coat protein units in the capsomeres, or to determine the triangulation number T. In addition, viruses have been recorded budding from infected cells and suffering the consequences of a variety of stresses. Mutant viruses have been examined and phenotypes described. Unusual structural features have appeared, and very importantly, the unexpectedly great amount of structural nonconformity within populations of virus particles has been well documented. It has, furthermore, been shown that the structures of viruses observed by AFM are entirely consistent with models derived by X-ray crystallography and cryo-EM (16). Although there are currently no examples, there is certainly no reason why structural information derived from X-ray crystallography and/or electron microscopy cannot be combined with AFM images, just as it has been for the latter two technologies.

192

McPherson and Kuznetsov

4. Notes 1. Tapping mode minimizes contact between the probe tip and the sample surface and greatly reduces lateral forces. An even more sensitive means of scanning in tapping mode is called phase modulation scanning. Here, phase changes are introduced into the tip oscillations due not only to height differences, but also to changes in the nature of the aggregate atomic interactions, caused in turn by variations in the physical or chemical properties of the sample surface. This approach has been shown to be useful for imaging very thin and delicate materials such as biological membranes (24). 2. To achieve this, it may be necessary to treat the substrate with various reagents, such as poly-l-lysine, to induce better adhesion of samples. If this condition is not met, then the specimen will move due to interaction with the probe, and no useful information will be gathered. 3. Because one does not, in general, know the tip shape one is working with at the time, the image cannot be easily deconvoluted to provide true lateral dimensions. 4. On large soft samples, such as living animal cells (33), lateral resolution may be more limited by the motion and deformation of the cell surface in response to tip pressure rather than tip structure. 5. Because visualization can be carried out in a fluid environment, specimens may suffer no dehydration as is generally the case with electron microscopy, and they usually require no fixing or staining. 6. Indeed, specimens can be observed over long periods, so long as they stay relatively unchanged and immobilized during a single frame interval. For the most part, even living cells seem oblivious to the presence of the probe tip (33). 7. If the distribution is a simple Gaussian, then it can be presumed that particles of only one general morphology, or icosahedra of only one triangulation number are present, but that their diameters vary to some degree about the mean, perhaps due to physiological state or degree of maturation. On the contrary, if a more complex distribution is observed, one having multiple peaks and shoulders, then particles of separate classes may be present. 8. While the T number describes the overall distribution of capsomeres on the surface of an icosahedral capsid, the more complete description of a virus structure would require the distribution of protein units in the individual capsomeres to be defined, and ultimately coordinates of the atoms

Atomic Force Microscopy Investigation of Viruses

193

comprising the virus coat proteins. The last can only be obtained by X-ray crystallography, but the distribution of subunits within capsomeres can sometimes be determined or deduced by AFM analysis. 9. This approach is particularly effective with large, complex viruses such as vaccinia virus (25, 26) or mimivirus (9). With these large assemblies, ordered and disordered protein shells, lipid membranes, and the nucleic acid within can be revealed and analyzed. By deconstruction, the architecture of particles is revealed, and, at the same time, the kinds of biochemical interactions that maintain each level of structure are delineated as well. 10. In some cases, nonionic detergents are insufficient to disrupt structure and more vigorous ionic detergents such as SDS must be used. There is difficulty with SDS, however. It tends to have an all-or-none effect, so that upon reaching a concentration sufficient to disrupt viruses, it completely degrades them uncontrollably. SDS can also produce artifacts due to drying on the substrate. 11. The proteases must be washed from the virions with buffer or water before imaging as they otherwise produce a dense, irregular background that makes imaging problematic, and they foul the cantilever tip. 12. Mica is negatively charged on its surface, but exposure to nickel or magnesium salt such as MgCl2 coats it with divalent ions and leaves it positively charged. 13. Some viruses or macromolecules, such as nucleic acids, may be firmly held by a positive surface if they are repelled by a negative surface, or vice versa. 14. The only serious disadvantage of coating with poly-l-lysine is that it produces a rather rough and irregular background. As a consequence, molecular objects, such as lipid membranes or nucleic acids, which rise only about a nanometer or two above the substrate plane, become difficult to identify and visualize. The method is excellent, however, for imaging cells and intact or partially degraded virions. References 1. Binning, G., and Quate, C.F. (1986). Atomic force microscope. Phys. Rev. Lett. 56: 930–933. 2. Bustamante, C., and Keller, D. (1995). Scanning force microscopy in biology. Phys. Today 48: 32–38. 3. Allen, S., Davies, M.C., Roberts, C.J., Tendler, S.J.B., and Williams, P.M. (1997). Atomic

force microscopy in analytical biotechnology. Trends in Biotech. 15: 101–105. 4. Hansma, P.K. (1994). Tapping mode atomic force microscopy in liquids. Appl. Phys. Lett. 64: 1738–1740. 5. Kuznetsov, Y.G., Ulbrich, P., Haubova, S., Ruml, T., and McPherson, A. (2007).

194

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

McPherson and Kuznetsov Atomic force microscopy investigation of Mason-Pfizer monkey virus and human immunodeficiency virus type 1 reassembled particles. Virology 360: 434–446. Baker, T.S., Olson, N.H., and Fuller, S.D. (1999). Adding the third dimension to virus life cycles: Three-dimensional reconstruction of icosahedral viruses from cryo-electron micrographs. Microbiology and Molecular Biology Reviews 63: 862–922. Kuznetsov, Y.G., and McPherson, A. (2006). Atomic force microscopy investigation of Turnip Yellow Mosaic Virus capsid disruption and RNA extrusion. Virology 352: 329–337. Kuznetsov, Y.G., Gurnon, J.R., Van Etten, J.L., and McPherson, A. (2005). Atomic force microscopy investigation of a chlorella virus, PBCV-1. J Struct Biol 149: 256–263. Xiao, C., Kuznetsov, Y.G., Sun, S., Hafenstein, S.L., Kostyuchenko, V.A., Chipman, P.R., Suzan-Monti, M., Raoult, D., McPherson, A., and Rossmann, M.G. (2009). Structural studies of the giant mimivirus. PLoS Biology 7: 958–966. Malkin, A.J., Land, T.A., Kuznetsov, Y.G., McPherson, A., and DeYoreo, J.J. (1995). Investigation of virus crystal growth mechanisms by in situ atomic force microscopy. Physical Review Letters 75: 2778–2781. Land, T.A., Malkin, A.J., Kuznetsov, Y.G., McPherson, A., and DeYoreo, J.J. (1996). Mechanisms of protein and virus crystal growth: An atomic force microscopy study of Canavalin and STMV crystallization. J. Cryst. Growth 166: 893–899. Malkin, A.J., Kuznetsov, Y.G., Lucas, R.W., and McPherson, A. (1999). Surface processes in the crystallization of turnip yellow mosaic virus visualized by atomic force microscopy. J Struct Biol 127: 35–43. Malkin, A.J., Kuznetsov, Y.G., and McPherson, A. (2001). Viral Capsomere Structure, Surface processes and growth kinetics in the crystallization of macromolecular crystals visualized by in situ atomic force microscopy. J. Cryst. Growth 232: 173–183. Malkin, A.J., Plomp, M., and McPherson, A. (2002). Application of atomic force microscopy to studies of surface processes in virus crystallization and structural biology. Acta crystallographica 58: 1617–1621. Kuznetsov, Y.G., Malkin, A.J., Lucas, R.W., Plomp, M., and McPherson, A. (2001). Imaging of viruses by atomic force microscopy. The Journal of general virology 82: 2025–2034. McPherson, A., Malkin, A.J., Kuznetsov, Y.G., and Plomp, M. (2001). Atomic force microscopy

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

applications in macromolecular crystallography. Acta Cryst. D57: 1053–1060. McPherson, A., Kuznetsov, Y.G., Malkin, A.J., and Plomp, M. (2004). Macromolecular crystal growth investigations using atomic force microscopy. Journal of synchrotron radiation 11: 21–23. Makino, D.L., Larson, S.B., and McPherson, A. (2005). Preliminary analysis of crystals of panicum mosaic virus (PMV) by X-ray diffraction and atomic force microscopy. Acta crystallographica 61: 173–179. Kuznetsov, Y.G., Daijogo, S., Zhou, J., Semler, B.L., and McPherson, A. (2005b). Atomic force microscopy analysis of icosahedral virus RNA. J Mol Biol 347: 41–52. Kuznetsov, Y.G., Datta, S., Kothari, N.H., Greenwood, A., Fan, H., and McPherson, A. (2002). Atomic force microscopy investigation of fibroblasts infected with wild-type and mutant murine leukemia virus (MuLV). Biophys J 83: 3665–3674. Kuznetsov, Y.G., Victoria, J.G., Robinson, W.E., Jr., and McPherson, A. (2003). Atomic force microscopy investigation of human immunodeficiency virus (HIV) and HIV-infected lymphocytes. J Virol 77: 11896–11909. Kuznetsov, Y.G., Low, A., Fan, H., and McPherson, A. (2004). Atomic force microscopy investigation of wild-type Moloney murine leukemia virus particles and virus particles lacking the envelope protein. Virology 323: 189–196. Malkin, A.J., Plomp, M., Leighton, T.J., McPherson, A., and Wheeler, K.E. (2006). Unraveling the Architecture and Structural Dynamics of Pathogens by High Resolution In vitro Atomic Force Microscopy. In Microscopy and Microanalysis, pp. 32–85. Cambridge Univ. Press, Cambridge. Plomp, M., Rice, M.K., Wagner, E.K., McPherson, A., and Malkin, A.J. (2002). Rapid visualization at high resolution of pathogens by atomic force microscopy: structural studies of herpes simplex virus-1. Am J Pathol 160: 1959–1966. Malkin, A.J., McPherson, A., and Gershon, P.D. (2003). Structure of intracellular mature vaccinia virus visualized by in situ AFM. J. Viol. 77: 6332–6340. Kuznetsov, Y., Gershon, P.D., and McPherson, A. (2008). Atomic Force Microscopy Investigation of Vaccinia Virus Structure. Journal of Virology 85: 7551–7566. Kaper, J.M. (1975). The Chemical Basis of Virus Structure, Dissociation and Reassembly. In The Frontiers of Biology Series.

Atomic Force Microscopy Investigation of Viruses

28. 29.

30.

31.

32.

33.

34.

(eds. A. Neuberger, and E.L. Tatum). NorthHolland, Amsterdam. Horne, R.W., and Wildy, P. (1961). Symmetry in virus architecture. Virology 15: 348–373. Caspar, D.L.D., and Klug, A. (1962). Physical principles in the construction of regular viruses. In Cold Spring Harbor Symposium on Quant. Biol., pp. 1–24. Lucas, R.W., Kuznetsov, Y.G., Larson, S.B., and McPherson, A. (2001). Crystallization of Brome Mosaic Virus (BMV) and T = 1 Brome Mosaic Virus Particles following a structural transition. Virology 286: 290–303 Reddy, V.S., Natarajan, P., Okerberg, B., Li, K., Damodaran, K.V., Morton, R.T., Brooks, C.L.I., and Johnson, J.E. (2001). Virus Particle Explorer (VIPER), a website for virus capsid structures and their computational analyses. J. Virol. 75: 11943–11947. Ganser, B.K., Li, S., Klishko, V.Y., Finch, J.T., and Sundquist, W.I. (1999). Assembly and analysis of conical models for the HIV-1 core. Science 283: 80–82. Kuznetsov, Y.G., Malkin, A.J., and McPherson, A. (1997). Atomic force microscopy studies of living cells: visualization of motility, division, aggregation, transformation, and apoptosis. J Struct Biol 120: 180–191. Low, A., Datta, S., Kuznetsov, Y., Jahid, S., Kothari, N., McPherson, A., and Fan, H. (2007). Mutation in the glycosylated gag protein of murine leukemia virus results in reduced in vivo infectivity and a novel defect

35. 36.

37.

38. 39.

40. 41.

42.

195

in viral budding or release. J Virol 81: 3685–3692. Rossmann, M.G. (2000). Fitting atomic models into electron-microscopy maps. Acta crystallographica 56: 1341–1349. Rossmann, M.G., Bernal, R., and Pletnev, S.V. (2001). Combining electron microscopic with x-ray crystallographic structures. J Struct Biol 136: 190–200. Adrian, M., Dubochet, J., Lepault, J., and McDowall, A.W. (1984). Cryoelectron microscopy of viruses. Nature (London) 308: 32–36. Carragher, B., Potter, C.S., and Sigworth, F.J. (2007). Software tools for macromolecular microscopy. J. Struct. Biol. 157: 1–288. Wikoff, W.R., Liljas, L., Duda, R.L., Tsuruta, H., Hendrix, R.W., and Johnson, J.E. (2000). Topologically linked protein rings in the bacteriophage HK97 capsid. Science 289: 2129–2133. McPherson, A. (2005). Micelle formation and crystallization as paradigms for virus assembly. Bioessays 27: 447–458. Kuznetsov, Y.G., Zhang, M., Menees, T.M., McPherson, A., and Sandmeyer, S. (2005c). Investigation by atomic force microscopy of the structure of Ty3 retrotransposon particles. J Virol 79: 8032–8045. Kuznetsov, Y.G., and McPherson, A. (2006b). Identification of DNA and RNA from retroviruses using ribonuclease A. Scanning 28: 278–281.

wwwwwwwwwwwwww

Chapter 13 Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions Christian Rankl, Linda Wildling, Isabel Neundlinger, Ferry Kienberger, Hermann Gruber, Dieter Blaas, and Peter Hinterdorfer Abstract Human rhinoviruses are the causative agents of the common cold. The serotypes belonging to the minor receptor group attach to members of the low-density lipoprotein receptor family and enter the host cell via receptor-mediated endocytosis. Receptor binding, the very first step in infection, was characterized by force spectroscopy measurements at the single molecule level. We demonstrate how kinetic on- and offrate constants can be derived from such experiments carried out with the atomic force microscope. Key words: Force spectroscopy, Molecular recognition, Viral receptor, Very low density lipoprotein receptor, Minor receptor group rhinovirus, PicoRNAvirus, Atomic force microscope

1. Introduction The first step of viral infection is the attachment of the virus to a cell receptor. This interaction initiates a cascade of processes resulting in the delivery of the viral genome into the host cell. The route can be receptor-mediated endocytosis (e.g., many naked and enveloped viruses (1)) or direct translocation of nucleoproteins via fusion of viral envelope and plasma membrane (2, 3). Once inside the cell, proteins encoded by the viral genome are translated and, ultimately, viral progeny is produced. Thus, understanding cell attachment of viruses is important. Here, we show how single molecule force spectroscopy is used to gain insight into this elementary process. The atomic force microscope was invented in the late 1980s (4). Its ability to measure in liquids and at room temperature made it a tool to investigate biological interactions at the single Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_13, © Springer Science+Business Media, LLC 2011

197

198

Rankl et al.

molecule level, such as binding of receptor–ligand pairs (5), antibody–antigen interaction (6), and the interaction of complementary DNA strands (7). Recently, this technique was applied to investigate the interaction of single viruses with live cells (8). In order to perform force measurements of interacting components (molecules or molecular assemblies), one binding partner is immobilized on the cantilever tip and the other one on a sample surface. Approaching the tip until it touches the surface eventually leads to interaction between the molecules on the tip and on the surface. Subsequent retraction loads the bond with an increasing force due to cantilever bending. At a certain force, the bond ruptures and the bent cantilever jumps off the surface into its equilibrium position. This unbinding is caused by thermal fluctuations rather than by purely mechanical dissociation. If the thermal lifetime of the interaction is short compared to the time it takes to apply the bending force of the cantilever, no unbinding event will be observed. Faster loading results in measurable unbinding forces. Therefore, unbinding forces depend on the applied loading rate and on the details of the functional relation of bond lifetime and applied force. A detailed theoretical consideration yields the direct link between such single molecule pulling experiments and bulk experiments, where thermodynamic data are experimentally acquired. The single molecule approach gives access to the full spectrum of a property instead of an averaged value gained by bulk experiments. This chapter describes the protocol to measure kinetic onand off-rates of a human rhinovirus type 2 (HRV2) interacting with the human low-density lipoprotein receptor, overexpressed in mouse fibroblast cells. In addition, we show how the number of interacting receptors can be determined. We used a PicoPlus AFM (Agilent Technologies, Chandler, AZ, USA) with commercially available cantilevers (Microlevers, Veeco). The virus was bound to the tip using a homemade amine reactive cross-linker.

2. Materials 2.1. AFM

1. 5500 AFM (Agilent Technologies, Chandler, AZ, USA). 2. National Instruments data acquisition card (PCI-6121). 3. Cantilever: MSCT Microlever (Veeco, Santa Barbara, CA).

2.2. Linking Viruses to the Tip

1. HRV2 produced in suspension culture and purified by differential and sucrose density gradient centrifugation (9). 2. DMSO (Sigma, Taufkirchen, Germany): highest available purity. 3. Ethanolamine hydrochloride (Sigma, Taufkirchen, Germany): highest available purity.

Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions

199

4. Chloroform and ethanol (Fisher Scientific): analytical grade. 5. Triethanolamine (Sigma, Taufkirchen, Germany): highest available purity. 6. 4-(2-Hydroxyethyl)piperazine-1-ethanesulfonic acid, N-(2hydroxyethyl)piperazine-N ¢-(2-ethanesulfonic acid) (HEPES) (Sigma, Taufkirchen, Germany): highest available purity. 7. NaCl (Sigma, Taufkirchen, Germany): highest available purity. 8. NaOH (Sigma, Taufkirchen, Germany): highest available purity. 9. NaCNBH3 (Sigma, Taufkirchen, Germany): highest available purity. 10. HBS buffer solution: 10 mM HEPES, 150 mM NaCl, pH 7.4. 11. HBS-Ca buffer solution: HBS with 2 mM CaCl2. 2.3. Mouse Fibroblast Cells

1. M4-LDLR cells (10), a SV40 large T-antigen-immortalized mouse fibroblast cell line deficient in endogenous LDLR and LRP (LDLR-related protein) overexpressing human lowdensity lipoprotein receptor (11, 12). 2. DMEM (Invitrogen, Cat. No. 21969-035). 3. l-Glutamin (Invitrogen, Cat. No. 25030-024). 4. Penicillin–streptomycin (Invitrogen, Cat. No. 151140-122). 5. Fetal bovine serum (Invitrogen, Cat. No. 10270-106). 6. Growth medium: DMEM with 1% l-glutamin, 1% penicillin– streptomycin, and 10% fetal bovine serum.

3. Methods 3.1. Immobilizing the Virus to the Tip

Attachment of ligand molecules to the measuring tip of an AFM converts it into a biospecific sensor by which cognate receptor molecules can be detected on a sample surface. Attachment of ligands to AFM tips via PEG chains is done in three steps: (1) amino groups are generated on the tip surface, (2) PEG chains are attached to the amino groups on the tip, and (3) a ligand molecule is coupled to the free-tangling end of PEG: 1. For amino functionalization of the cantilever chips, 3.3  g ethanolamine hydrochloride was dissolved in 6  mL DMSO by heating to 60°C in a glass beaker. 2. When all solid was dissolved, the beaker was removed from the heater and molecular sieve beads (3 Å) were added to the mixture and a small thin glass plate (prewashed in ethanol and dried with nitrogen gas) was placed on top of the beads, taking care to avoid inclusion of air bubbles. 3. The cantilevers were washed three times in chloroform and dried with nitrogen gas.

200

Rankl et al.

4. Cantilevers were immersed in the ethanolamine hydrochloride/ DMSO solution and incubated overnight. 5. The tips were washed with DMSO (3×), ethanol (3×), and dried with nitrogen gas. 6. For linking of aldehyde–PEG–NHS, 3.3 mg of this compound were dissolved in 0.5 mL chloroform and transferred into a small glass reaction chamber. 7. Triethylamine (30  ml) was added and the cantilevers were immersed into the solution at RT for 2 h. 8. After the reaction, the cantilevers were extensively washed in chloroform and dried with nitrogen gas. 9. For covalent binding of viral particles, a 5-mL aliquot of the virus stock solution (HRV2, ~1  mg/mL in HBS, freshly thawed from −25°C) was mixed with 45 ml of HBS. 10. The bottom of a disposable Petri dish was covered with parafilm and the dried cantilevers were moderately pressed onto the parafilm. The cantilevers were arranged in a circular manner so that the ends with the tips pointed to the center of the circle. Then, the 50-mL virus suspension was pipetted into the center of the circle such that all tips were covered with liquid. 11. Immediately, a 2-mL aliquot from a 1  M NaCNBH3 stock solution (freshly prepared by dissolving 32 mg solid NaCNBH3 in 500 mL 10 mM NaOH) was added and the cantilever was incubated for 60 min. 12. Free aldehyde functions on the tips were inactivated by addition of 5 mL 1 M ethanolamine hydrochloride (pre-adjusted to pH 9.6 with NaOH and stored in aliquots at −25°C). 13. After another 10 min reaction time at room temperature, the cantilevers were washed with HBS and stored in HBS-Ca at 4°C for less than 24 h. 3.2. Cell Preparation

1. For AFM measurements, cells were grown on a 25 mm cover slip. Best results were achieved when the cells were nearly confluent (~80%). 2. The cover slip was mounted in the AFM. It was taken care that the cells never dried. 3. Growth medium was exchanged against HBS-Ca buffer solution.

3.3. Calibration of the Optical Detection System of the AFM

Prior to interaction measurements, the optical lever detection (OLS) system was calibrated allowing to correlate photodiode voltage with the deflection of the cantilever. This calibration was done for every cantilever used. The result was a calibration factor

Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions

201

fols, which describes the relationship between physical cantilever deflection and measured photodiode voltages: 1. The AFM was set up according to the manual. 2. A cantilever containing virus immobilized to its tip was inserted into the AFM. It was taken care that the cantilever never dried, as this might denature the virus on the tip. 3. A freshly cleaned glass slide was mounted in the AFM. The liquid cell was filled with HBS buffer. 4. For sensitivity determination approx. 50 force distance cycles with 100 nm length and 1 s sweep time were acquired. In order to obtain most accurate results, the z-piezo was freshly calibrated and the force distance cycles were performed around 0 nm (i.e., start at 50 nm and end at −50 nm z-piezo position). 5. The slope of the contact part was determined in every force curve individually (see Note 1, Fig. 1a). 6. The calibration factor fols represents the negative average of the above determined inverse slopes. The obtained value depends on the cantilever and AFM system, typical values for fols are between 10 and 100 nm/V.

Fig. 1. (a) A force–distance cycle (dash dotted line) on mica was acquired so as to calibrate the optical lever detection system. The tip was aligned away from the surface and continuously approached toward it. As long as the tip does not contact the surface, no deflection is seen (right part of the graph). As long as the tip is in contact with the surface, further approaching results in an upward bending (left part of the graph). A linear fit (solid line) through the contact region yielded a calibration factor fols = 18.3 nm/V. (b) A force–distance cycle showing an unbinding event in the retrace; note the nonlinear stretching indicating specific binding. The rupture force F can be directly calculated from the height of the rupture event. The uncertainty of the rupture force s is determined by the standard deviations left (sl) and right (sr) of the rupture: s 2 = s 2l + s 2r . The slope at rupture keff is needed to determine the loading rate r = keffv.

202

Rankl et al.

3.4. Dynamic Force Spectroscopy Experiments

The interaction force is measured by performing force distance cycles; i.e., the tip is approached to the surface until it touches it and subsequently retracted. During retraction the linker stretches until the bond breaks, which results in a characteristic nonlinear force signal. The dissociation of the complex is mainly thermally driven and of stochastic nature. Therefore, many such force signals must be acquired to obtain statistically relevant distributions of the rupture forces. It is another consequence of the stochastic nature of this process that the rupture force distribution depends on the pulling speed. From this dependence, the kinetic off-rate is extracted: 1. A cell sample was mounted in the AFM and the growth medium was exchanged against HBS-Ca buffer solution. 2. A virus-carrying tip was mounted and placed over a cell making use of the optical access. During approach, it was taken care that the tip did not crash into the cell, so as to avoid unspecific adhesion of proteins to the AFM tip. Such a “dirty” tip can result in false force signals or in no force signals at all. 3. At least 1,000 force–distance cycles of typically 1,000  nm sweep length and 1  s sweep duration time were recorded. Again, the maximum force applied by the cantilever was kept as low as possible to prevent contamination of the tip. 4. Step 3 was repeated using different sweep duration times (0.1, 0.2, 0.5, 2, and 5 s). For very short duration times (0.1 and 0.2 s), at least 2,000 force curves were recorded so as to compensate for the low binding probabilities due to the short contact times. 5. The steps 3 and 4 were repeated using cantilevers of different spring constants (10, 20, and 30 pN/nm). 6. The specificity of the interaction was proven by blocking the interaction. Usually, one binding partner is added at high concentration, leading to a saturation of available binding sites. Here, we choose reversible inactivation. Activity of the LDL receptor depends on Ca2+ ions. In order to block the interaction with the virus, the buffer was exchanged against HBS buffer, containing 2 mM EGTA, which does not contain Ca2+. Again, 1,000 force curves with 1,000 nm sweep length and 1 s sweep duration were acquired. The binding probability was reduced to less than 3% (compared to 15% in HBS-Ca buffer). Next, the buffer was changed back to HBS-Ca, resulting in restoration of the binding probability to a value comparable before Ca2+ removal. 7. Step 6 was repeated for each cantilever used. Only data from cantilevers that showed the above-described reversible blocking behavior were used.

Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions

3.5. Determining the Spring Constant of the Cantilever

203

The cantilever acts as a small spring and, accordingly, the deflection can be converted into a force F acting on the cantilever, using Hooke’s law F = k × z. Here k is the cantilever spring constant and z the deflection of the cantilever. The spring constant can be measured in several ways, either using a calibrated reference (13), the added mass method (14), the thermal noise method (15), or the Sader method (16). Each of these methods has advantages and drawbacks, a good overview can be found in ref. (17). The thermal noise method is most widely used and commonly accepted. It is based on modeling the cantilever as a spring, making use of the equipartition theorem: 1. The washed and dried cantilever was mounted in the AFM with a freshly calibrated scanner (see Note 2). A cleaned mica sheet was used as sample. 2. The free (at least 100 mm away from the surface) cantilever movement was recorded using a data acquisition card. 3. The optical lever sensitivity was determined (see above). 4. Multiplying the recorded cantilever movement with the optical lever sensitivity converts the photodiode output into the deflection (nanometer) of the cantilever. 5. The power spectrum density of the data gained in step 4 was estimated using a fast Fourier transform (see Note 3). The resulting squared amplitudes of the Fourier transform are proportional to the power spectrum. However, there are several different conventions for the normalization of the power spectrum and many opportunities for making it wrong. In order to achieve the time-integral squared amplitude, the proper normalization factor is given by the sampling interval. A detailed discussion can be found in ref. 18. 6. This power spectrum estimate was fitted with a simple harmonic oscillator model:



A = Awhite +

4 A0w 0 2 2 2 æ ww 0 ö w - w0 +ç è Q ÷ø

(

)

2

,



(1)

where Awhite is a white noise floor, A0 is the zero frequency amplitude, w0 is the radial resonance frequency, and Q is the quality factor (Fig. 2).



7. The spring constant of the cantilever using the thermal noise method is finally given by: 4 kb T k =α , (2) A0ω 0 Q where kb is the Boltzmann constant and T the absolute temperature, a = 0.817 for rectangular cantilevers (19) and a = 0.764 for triangular cantilevers (20).

204

Rankl et al.

Fig. 2. The power spectral density (dots) of a freely oscillating cantilever used to calculate the spring constant is shown. It was calculated by applying a Fourier transform to the photo diode output shown in the inset and subsequent proper normalization. The harmonic oscillator fit (solid line) yielded a radial resonance frequency of ~92 krad/s, a quality factor of 25.6 and a zero frequency amplitude of 2.4 nm2/Hz. These parameters were used to estimate the spring constant of the cantilever (23.4 pN/nm).

3.6. Extraction of the Kinetic Off-Rate from Force Spectroscopy Experiments

A simple two-state binding model with a separating energy barrier can be used to describe dynamic force spectroscopy experiments (21, 22). It links dynamic force spectroscopy experiments to the kinetic off-rate constant and to the thermally averaged distance between the binding state and the barrier along the projection of the applied force: 1. A program (see Note 1) was written, which converts the force curves into proper units (pN) by multiplying the photo diode output with fols × k and allowed toggling through the proper converted force curves. Curves showing a nonlinear force signal, terminated by an abrupt return to the baseline were identified as an unbinding event. The program allowed extracting 2 2 2 the height F i and the uncertainty ( s i = s l,i + s r,i ) of a rupture event, and the slope at the time of rupture keff of proper unbinding events (Fig. 1b).



2. The probability density estimate of the rupture forces was estimated using: æ (F - Fi )2 ö 1 n 1 (3) pdf (F ) = å exp ç ÷. 2 n i =1 2ps i2 è 2s i ø 3. The most probable rupture force F * was determined from this pdf.

Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions

205

Fig. 3. The dependence of the applied loading rate on the distribution of rupture forces is shown. Theory predicts an increase of observed rupture forces with higher loading rate, which was indeed found.

4. The loading rate of a dataset was determined to be: r = keff

2 ´ scansize . sweeptime

5. Steps 1–4 were repeated for different pulling speeds resulting in a pdf and a loading rate estimate for each dataset (see Fig. 3). 6. The most probable unbinding force F * was plotted against the logarithm of the loading rate r (Fig. 4). 7. A straight line ( y = ax + b ) with slope a and y axis intersection b was fitted into the above plot. 8. Parameters were extracted according to -

3.7. Measuring the Kinetic On-Rate Constant



b

kT e a , x b = B , c.f. Notes 4 and 5 (21, 22). koff = a a The virus–receptor interaction was approximated with pseudo first-order kinetics. Estimation of the kinetic on-rate constant kon from single molecule unbinding force measurements requires the determination of the interaction time t and the effective concentration ceff: kon = (tc eff ) . -1

The interaction time t was determined from the binding probability at different encounter times. An effective concentration

206

Rankl et al.

Fig. 4. The most probable rupture force as a function of the loading rate is shown. As predicted (21, 22), the unbinding force depends linearly on the logarithm of the loading rate. The straight line shows the linear fit used to calculate the model parameters, koff = 0.6/s and x b = 0.4 nm.

was estimated from the number of binding partners nb (typically 1) within a free volume accessible by the virus tethered to the tip. The free volume was approximated by a sphere, whereby the radius reff of the sphere is the sum of the equilibrium length of the cross-linker (3  nm) plus the diameter of the virus (30  nm). Therefore, the kinetic on-rate constant can be calculated using the following formula (8, 23): kon

4preff3 N A = , 3nbt

(4)

where NA is the Avogadro constant = 6.022 × 1023/mol. 1. Virus tip and cell samples were mounted as described above. 2. The size and duration for force distance cycle were set to 1,000 nm sweep length and 1 s sweep time. 3. First, a dwell time of about 10 ms on the surface was used. As our AFM software does not directly support dwell times on the surface, we set a relative maximum deflection limit of 0.2  V. The dwell time was then determined from the force curve, i.e., the time needed to travel from the contact point to the maximum force limit and back again. 4. At least 1,000 force distance cycles were acquired. 5. Step 4 was repeated using different dwell times (10, 20, 50, 100, 200, 500, 1,000 ms, etc.), c.f. Note 6.

Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions

207

Fig. 5. Binding probability as a function of the contact time. A typical first-order reaction kinetics was found allowing for the determination of the kinetic on-rate constant kon.

6. The binding probability, i.e., the fraction of force distance cycles that showed an unbinding event was determined for every dwell time. 7. The unbinding probability as a function of the contact time was plotted. 8. This graph was fitted using p = A (1 - exp((t - t 0 ) / t )) , where t0 is the time lag, A the maximum observable binding probability, and t the interaction time (Fig. 5). 9. The suchlike gained t was used to determine the kinetic on-rate constant: 4preff 3N A . kon = 3nb t 3.8. Counting the Number of Bound Receptors

HRV2 has 12 receptor-binding sites (i.e., each receptor molecule can wind around each of the 12 vertices of the viral icosahedron simultaneously attaching via up to five symmetry-related sites; (24)). In order to determine how many receptors are bound to the virus, the distribution of rupture forces were further analyzed. Multiple receptor binding stabilizes the interaction and leads to higher rupture forces. If an individual receptor–virus bond breaks, the applied load is distributed over the remaining bonds and reduces the bond lifetime to nanosecond levels (25). Therefore, the remaining bonds rupture very shortly after the first one is broken and the finite bandwidth of the AFM causes the multiple bond breakages to be registered as a single rupture event. A direct consequence of this fact is that the governed rupture force distribution shows multiple force peaks. This distribution was fitted with æ F -m 2 ö 1 ( i ) ÷ , including N A exp ç a sum of Gaussians: p (F ) =

å

i =1

i

si 2p

çè

2si2

÷ø

208

Rankl et al.

Fig. 6. The rupture force distribution of virus interacting with a cell is shown at two different encounter times. (a) 330 ms and (b) 670 ms. Shorter encounter times show more single receptor–virus interaction breakings. With longer encounter times, the number of single receptor–virus binding decreases, concomitantly double receptor–virus bindings increase. N

åA

=1 the boundary condition i=1 i , where Ai is the fraction of rupture events corresponding to this peak, m i is the position of the peak, and si is the width of the peak. With increasing encounter time, the positions mi and widths of the peaks si were constant. In contrast, the fractions Ai changed and more multiple receptor bonds were found (Fig. 6). The force needed to unbind two receptors from a single virus is not necessarily twice the force needed to rupture a single receptor–virus bond. For uncorrelated binding, it is smaller than the double value. In addition, the most probable unbinding force for multiple receptors–virus interactions does not depend linearly on the logarithm of the loading rate anymore (8). Therefore, the kinetic off-rate constant must be determined from the loading rate behavior of the force peak of a single receptor–virus interaction.

4. Notes 1. MATLAB (The Mathworks, Natick, MA, USA) was used for all programs and calculations. 2. The thermal noise method to calibrate the spring constant depends as 1 / f ols 2 on the sensitivity fols; therefore, a proper calibration of the z-scanner is necessary to achieve highest accuracy prior to every spring constant measurements. A sensitivity-free method to determine the spring constant is the Sader method, which requires knowing the plain view geo­ metry of the cantilever. In our lab, a program was written that

Determination of the Kinetic On- and Off-Rate of Single Virus–Cell Interactions

209

is capable to determine the spring constant using the thermal noise and Sader method within the same measurement. 3. Another important detail to note is that the standard deviation of the Fourier transform is 100%, independent of the number of sample points. But the mean value can be estimated very accurately by averaging several measurements. For this, the recorded free cantilever movement was split into five to ten parts of equal length; a power density estimation of every part was performed. The final power density was calculated by averaging over all parts (18). 4. Accuracy estimates of koff and x b can be derived using bootstrapping (26). 5. In the last years, additional models describing dynamic force spectroscopy have been published (27–29), which could also be used to extract information about the binding energy landscape. 6. Prior to kinetic on-rate measurements, an expected t value should be estimated by solving Eq. 3 for t, using an estimate for kon. The dwell times on the surface should be adjusted to this estimate. References 1. Marsh, M. and A. Helenius (2006) Virus entry: Open sesame. Cell 124, 729–740. 2. Harrison, S.C. (2008) Viral membrane fusion. Nat Struct Mol Biol 15, 690–8. 3. Falanga, A., et  al. (2009) Membrane fusion and fission: enveloped viruses. Protein Pept Lett 16, 751–9. 4. Binnig, G., C.F. Quate, and C. Gerber (1986) Atomic Force Microscope. Phys. Rev. Lett. 56, 930–933. 5. Moy, V.T., E.L. Florin, and H.E. Gaub (1994) Intermolecular forces and energies between ligands and receptors. Science 266, 257–259. 6. Hinterdorfer, P., et  al. (1996) Detection and localization of individual antibody-antigen recognition events by atomic force microscopy. Proc. Natl. Acad. Sci. USA. 93, 3477–3481. 7. Lee, G.U., L.A. Chrisey, and R.J. Colton (1994) Direct measurement of the forces between complementary strands of DNA. Science 266, 771–773. 8. Rankl, C., et  al. (2008) Multiple receptors involved in human rhinovirus attachment to live cells. Proceedings of the National Academy of Sciences 105, 17778–17783. 9. Hewat, E.A., et al. (2000) The cellular receptor to human rhinovirus 2 binds around the

5-fold axis and not in the canyon: a structural view. EMBO J. 19, 6317–6325. 10. Herdy, B., et al. (2004) Identification of the human rhinovirus serotype 1A binding site on the murine low-density lipoprotein receptor by using human-mouse receptor chimeras. J Virol 78, 6766–74. 11. Ishibashi, S., et al. (1993) Hypercholesterolemia in low density lipoprotein receptor knockout mice and its reversal by adenovirus-mediated gene delivery. J Clin Invest 92, 883–893. 12. Herz, J., D.E. Clouthier, and R.E. Hammer (1992) LDL receptor-related protein internalizes and degrades uPA-PAI-1 complexes and is essential for embryo implantation. Cell 71, 411–421. 13. Gibson, C.T., G.S. Watson, and S. Myhra (1996) Determination of the spring constants of probes for force microscopy/spectroscopy. Nanotechnology 7, 259–262. 14. Cleveland, J.P., et al. (1993) A nondestructive method for determining the spring constant of cantilevers for scanning force microscopy. Review of Scientific Instruments 64, 403–405. 15. Hutter, J.L. and J. Bechhoefer (1993) Calibration of atomic-force microscope tips. Rev. Sci. Instrum. 64, 1868–1873.

210

Rankl et al.

16. Sader, J.E., et al. (2005) General scaling law for stiffness measurement of small bodies with applications to the atomic force microscope. J. Appl. Phys. 97, 124903–7. 17. Ohler, B. Practical Advice on the Determination of Cantilever Spring Constants. 2007; Available from: http://www.veeco.com/pdfs/appnotes/AN94%20Spring%20Constant%20 Final_304.pdf. 18. Press, W.H., (2007) Numerical Recipes: The Art of Scientific Computing. 3 ed. 2007: Cambridge University Press. 19. Butt, H.J. and M. Jaschke (1995) Calculation of thermal noise in atomic force microscopy. Nanotechnology 6, 1–7. 20. Stark, R.W., T. Drobek, and W.M. Heckl (2001) Thermomechanical noise of a free v-shaped cantilever for atomic-force microscopy. Ultramicroscopy 86, 207–215. 21. Evans, E. and K. Ritchie (1997) Dynamic strength of molecular adhesion bonds. Biophysical Journal 72, 1541–1555. 22. Izrailev, S., et al. (1997) Molecular dynamics study of unbinding of the avidin-biotin complex. Biophys. J. 72, 1568–1581.

23. Baumgartner, W., et  al. (2000) Affinity of Trans-interacting VE-cadherin Determined by Atomic Force Microscopy. Single Molecules 1, 119–122. 24. Verdaguer, N., et al. (2004) X-ray structure of a minor group human rhinovirus bound to a fragment of its cellular receptor protein. Nature Structural Molecular Biology 11, 429–434. 25. Sulchek, T., R.W. Friddle, and A. Noy (2006) Strength of Multiple Parallel Biological Bonds. Biophys J 90, 4686–4691. 26. Rankl, C., et al. (2007) Accuracy Estimation in Force Spectroscopy Experiments. Japanese Journal of Applied Physics 46, 5536. 27. Dudko, O.K., et al. (2003) Beyond the conventional description of dynamic force spectroscopy of adhesion bonds. Proc. Natl. Acad. Sci. USA. 100, 11378–11381. 28. Hummer, G. and A. Szabo (2003) Kinetics from nonequilibrium single-molecule pulling experiments. Biophys. J. 85, 5–15. 29. Raible, M., et al. (2006) Theoretical analysis of single-molecule force spectroscopy experiments: heterogeneity of chemical bonds. Biophys. J. 90, 3851–3864.

Chapter 14 Atomic Force Microscopy as a Tool for the Study of the Ultrastructure of Trypanosomatid Parasites Wanderley de Souza, Gustavo M. Rocha, Kildare Miranda, Paulo M. Bisch, and Gilberto Weissmuller Abstract Here, we describe the methodology currently used to analyze the structural organization of protozoa of the Trypanosomatidae family by atomic force microscopy. The results are compared with those obtained using light, scanning, and transmission electron microscopy. Key words: Parasitic protozoa, Trypanosomatids, Atomic force microscopy, Scanning electron microscopy

1. Introduction Prior to the 1980s, analysis of the structural organization of ­biological samples was carried out primarily by light, scanning, and transmission electron microscopy. In 1981, Gerd Binnig and Heinrich Rohrer, working at the IBM laboratory in Zurich, developed the scanning tunneling microscope, which was first microscope able to generate atomic-resolution three-dimensional images, and used it to analyze the surfaces of materials with high conductivity (1). A few years later, the same group developed the atomic force microscope (AFM). The AFM did not require a highly conductive surface, and thus could be applied to analyze the surfaces of a large number of materials, including biological materials (2). More importantly, especially from a biological point of view, AFM can provide nanometer-resolution images of living cells in gaseous and liquid environments. As a result, even cells in culture can be examined by AFM, which opens new avenues for  the application of this technique in the biological sciences. Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_14, © Springer Science+Business Media, LLC 2011

211

212

de Souza et al.

Since its inception, AFM has been used extensively to analyze diverse biological samples, including parasitic protozoa (3–10). It is now well established that AFM has an enormous potential for imaging and analyzing cells. This potential has been increasingly realized in the last few years, especially with the release of commercially available equipments, some even dedicated to biological applications. AFM has also been applied to study the structural organization of trypanosomatid parasites, which was first attempted by Dvorak and collaborators almost 10 years ago (4). While this study drew the attention of the parasitology community to the potential of this technique, the published images did not add significant new information regarding the structure of these cells. However, a subsequent study that used a similar methodology to prepare the samples but with the addition of a pretreatment of the protozoa with detergent produced images wherein both the cell surface and the intracellular structures of epimastigote forms of Trypanosoma cruzi could be well recognized (10). In addition, new structures were described, including (a) a flagellar furrow separating the axoneme from the paraflagellar rod (PFR) and running from the point of emergence of the flagellum from the flagellar pocket to the flagellar tip, (b) a row of periodically organized structures localized in the flagellar furrow, and (c) the structural organization of the flagellar necklace, which appears as nine protrusions positioned about 400  nm from the flagellar base and distributed around the circumference of the flagellum. This chapter describes the methods we have used to analyze the structure of trypanosomatids by atomic force microscopy. The methodology used is explained in great detail so that it can be easily repeated and adapted for similar studies with other trypanosome species and other protozoa. There are many AFM scan modes. The intermittent contact mode produces topographical information. Phase imaging is a complementary imaging mode that, in addition to providing topographical information, allows for the analysis of adhesive and elasticity properties. Herein, we describe such approaches for acquiring images of parasitic protozoa of the Trypanosomatidae family.

2. Materials 2.1. Protozoan Culture Medium

1. Liver infusion tryptose (LIT) medium: contains 0.4% sodium chloride, 0.2% glucose, 0.65% anhydrous disodium hydrogen phosphate, 0.5% tryptose, 0.5% liver infusion broth, 0.04% potassium chloride, 0.05% hemin, and 0.02% folic acid. 2. LIT medium is supplemented with 10% fetal bovine serum.

Atomic Force Microscopy as a Tool for the Study of the Ultrastructure

213

3. Sodium phosphate buffer (to obtain 200 mL of 0.1 M sodium phosphate buffer pH 7.4, combine 19 mL of 0.2 M monobasic sodium phosphate, NaH2PO4, 81  mL of 0.2  M dibasic sodium phosphate, Na2HPO4, and 100  mL of distilled water). 4. Glass pipettes (5, 1 mL) and rubber bulb. 5. Calibrated pH meter. 6. Conventional centrifuge and a high-speed microcentrifuge. 7. Centrifuge tubes (conical, 15 mL) or microcentrifuge tubes. 8. Top loading balance for measuring buffer salts and resin components. 9. Hot plate with magnetic stirrer and stirring bar. 10. Magnetic stir bars. 11. Culture tubes or plastic flasks. 12. Dissecting needle. 2.2. Whole Intact Cell Analysis

1. 2.5% Glutaraldehyde (EM grade). 2. Sodium cacodylate buffer (to obtain 100 mL of 0.1 M sodium cacodylate buffer pH 7.4, combine 25 mL of 0.2 M sodium cacodylate, Na(CH3)AsO2 · 3H2O, 1.4  mL of 0.2  N hydrochloric acid, HCl, and 23.6 mL of distilled water). 3. Ethanol dehydration series (Merck – product number 100983). 4. Fume cupboard (Fume hood). 5. Critical point dryer (Baltec – CPD 030). 6. 0.01% Poly-l-lysine (70,000–150,000 MW; Sigma–Aldrich). 7. Ethanol dehydration series (Merck – product number 100983). 8. Fume cupboard (Fume hood). 9. Critical Point Dryer (Baltec – CPD 030).

2.3. Atomic Force Microscopy

1. PHEM buffer containing 60 mM pipes (1,4-piperazine diethylsulfonic acid), 20 mM Hepes (N-2-hydroxyethylpiperazine N-1-2-ethanesulfonic acid), 10  mM EGTA, and 2  mM MgCl2, pH 7.2. 2. 1% Nonidet NP-40. 3. V-shaped standard narrow cantilevers, model NP-S (Veeco Probes, Camarillo, CA, USA). 4. Tetrahedral-shaped cantilevers, model AC240TS (Olympus, Tokyo, Japan).

214

de Souza et al.

2.4. Field Emission Scanning Electron Microscopy

1. 2.5% Glutaraldehyde (EM grade). 2. Sodium cacodylate buffer (to obtain 100 mL of 0.1 M sodium cacodylate buffer pH 7.4, combine 25 mL of 0.2 M sodium cacodylate, Na(CH3)AsO2 · 3H2O, 1.4  mL of 0.2  N hydrochloric acid, HCl, and 23.6 mL of distilled water). 3. 0.01% Poly-l-lysine (70,000–150,000 MW; Sigma–Aldrich). 4. Osmium tetroxide (Electron Microscopy Sciences, stock aqueous solution). 5. Potassium ferrocyanide (to obtain 50 mL of 1.6% potassium ferrocyanide stock solution, combine 0.8 g potassium ferrocyanide, K4[Fe(CN)6] · 3H2O, 10  mM calcium chloride, CaCl2, and 50 mL of 0.2 M sodium cacodylate buffer). 6. Ethanol dehydration series (Merck – product number 100983). 7. Fume cupboard (Fume hood). 8. Critical Point Dryer (Baltec – CPD 030). 9. High-resolution ion beam coater (GATAN – model 681). 10. SEM specimen stubs. 11. Desiccated or dry compartment to store SEM specimen stubs.

2.5. Plasma Membrane Extraction

1. PHEM buffer containing 60 mM pipes (1,4-piperazine diethylsulfonic acid), 20 mM Hepes (N-2-hydroxyethylpiperazine N-1-2-ethanesulfonic acid), 10  mM EGTA, and 2  mM MgCl2, pH 7.2. 2. 1% Nonidet NP-40. 3. 2.5% Glutaraldehyde (EM Grade). 4. 0.01% Poly-l-lysine (70,000–150,000 MW; Sigma–Aldrich). 5. Ethanol dehydration series (Merck – product number 100983). 6. Fume cupboard (Fume hood). 7. Critical Point Dryer (Baltec – CPD 030).

3. Methods 3.1. Cultivation of Trypanosomatids

T. cruzi epimastigotes were cultivated in LIT medium supplemented with 10% fetal bovine serum and 1% hemin for 3–5 days at 28°C. Only parasites collected in the exponential phase of growth were used for experiments.

Atomic Force Microscopy as a Tool for the Study of the Ultrastructure

3.2. Preparation of Samples for Scanning Electron Microscopy

215

1. Fixation. Cells were collected by centrifugation (1600 × g for 10 min at 4°C) from 3- to 5-day-old cultures, washed with 0.1  M phosphate buffer, pH 7.2, and fixed for 60  min at room temperature in a solution containing 2.5% glutaraldehyde in 0.1 M phosphate or PHEM buffer, pH 7.2. For scanning electron microscopy following glutaraldehyde fixation, the cells were washed in buffer, postfixed for 1 h at 4°C in a solution containing 1% osmium tetroxide and 0.8% potassium ferrocyanide in 0.1 M sodium cacodylate buffer, pH 7.2. 2. Adhesion of protozoa. As trypanosomatids are single isolated cells, they were adhered for 10 min at room temperature into glass coverslips previously coated with 0.1% poly-l-lysine. Before use, the coverslips were cleaned with soap and water, rinsed in distilled water, and dried in a dust-free environment. For more details, see Note 1. 3. Plasma membrane extraction. For some experiments, the adhered cells were incubated with 1% Nonidet NP-40 diluted in PHEM buffer for 5–7 min in order to remove the plasma membrane. They were then fixed in glutaraldehyde as described above. For more details, see Note 2. 4. Dehydration. The cells were dehydrated using an ethanol series (30, 50, 70, 90, and 100%). The ethanol was of good quality. The cells were incubated for 10 min in each ethanol solution. 5. Critical point drying. To dry the samples and avoid surface tension, critical point drying (BALTEC – CPD030) was used. After dehydration, the cells still adhered to the glass coverslips were transferred to the equipment. The compartment containing the samples was closed and the temperature and pressure were changed to about 8°C and 50 atm. Subsequently, the ethanol was gradually replaced by liquid carbon dioxide. When the substitution was complete, the temperature and pressure of the compartment were elevated to over 31°C and 73 atm to pass to the critical point of the working fluid, thus avoiding the direct liquid–gas transition seen in ordinary drying. For more details, see Note 3. 6. Metal coating. After critical point drying, small pieces of the glass coverslips were adhered to special stubs that were then introduced into the chamber of the high-resolution ion beam coater (GATAN – model 681). The samples were coated for 5 min with a 2–3-nm thick gold or chromium layer using the appropriate voltages and currents indicated by the manufacturer (400 mA). 7. Scanning electron microscopy. Observation was carried out using secondary electron imaging in a Jeol JSM 6340F field emission scanning electron microscope operating at 5.0 kV,

216

de Souza et al.

Fig. 1. Light and electron micrographs of epimastigote forms of Trypanosoma cruzi. (a) Differential interference contrast micrograph of T. cruzi showing the whole parasite. It is possible to visualize the whole parasite using this technique. The arrow points to the anterior region where a free flagellum is seen; (b) Transmission electron micrograph of a thin section of T. cruzi. Subcellular structures, such as the nucleus (N), kinetoplast (K), Golgi complex (G), mitochondrion (asterisk), and the flagellum (arrow ), are seen. Microtubules of axonem are also observed (arrowheads ). (c) Field emission scanning electron micrograph of T. cruzi showing its elongated form and its smooth surface. Arrow indicates the flagellum. Bars – (a) 5 mm; (b) 1 mm; and (c) 2 mm.

12  A emission current, and a working distance of 8  mm. Figure  1 shows a general view of the epimastigote form of T. cruzi as seen by light and electron microscopy. These images show the general shape of the protozoan from the posterior to anterior end of the parasite. The flagellum of this trypanosomatid emerges from the flagellar pocket (FP) and remains tightly attached to the cell body along its length (Fig. 1b, c). Scanning electron microscopy shows the flagellar tip as a single and uniform cylindrical shaped structure (Fig. 2). Figure 3 shows a schematic view of the protozoan, based mainly on observations made both by scanning and transmission electron microscopy of thin sections in order to facilitate the interpretation of the images shown. For more details, see Note 4. 3.3. AFM Imaging Techniques

1. Contact mode analysis. The glass slide containing the sample was mounted onto the XY scanner of the AFM and a CCD camera was used to locate the parasites. For contact mode analysis, a soft cantilever was used. V-shaped standard narrow cantilevers, model NP-S (Veeco Probes, Camarillo, CA), were used. This model of cantilever includes four cantilevers differing in length and width. For the experiments, we used the softer cantilever with a nominal spring constant of 0.12 N/m and nominal frequency of 18 kHz. Cantilever elastic ­constants

Atomic Force Microscopy as a Tool for the Study of the Ultrastructure

217

Fig. 2. FESEM image of the end of the flagellum of an epimastigote form of T. cruzi. The flagellum appears as a cylindrical structure and its end is seen as a unique and uniform structure. Bar – 1 mm.

Fig. 3. Illustration showing some structures and organelles found in the epimastigote form of T. cruzi based on images obtained by transmission electron microscopy.

were obtained by the thermal noise method. Samples were scanned at a constant force with a low scan rate (0.6 Hz) to reduce noise and minimize sample damage. Force and integral gain was constantly monitored to use minimum force with good response of the feedback system to obtain better images. To get high-resolution images, samples were acquired with 512 points × 512 lines of resolution. In the contact mode, the cytoskeleton could be detected after plasma membrane extraction (Fig. 4). 2. Intermittent contact mode analysis. For the intermittent contact mode, tetrahedral-shaped cantilevers (AC240TS, Olympus, Tokyo, Japan – nominal spring constant 2  N/m and 70  kHz of nominal frequency) were used. These are

218

de Souza et al.

Fig. 4. Three-dimensional view of a height image of the flagellum (f) of a T. cruzi epimastigote previously extracted with detergent and scanned in contact mode. Microtubules (arrow ) of the flagellar axoneme can be observed in this image. The subpellicular microtubules (asterisk) and the basal body (bb) are also seen. Bar – 2 mm.

interesting because their tips are located on the very end of the cantilever, and they offer a very good lateral resolution. Cantilever elastic constants were obtained by the thermal noise method. Samples were scanned at a constant force with a low scan rate (0.5 Hz) to reduce noise and minimize sample damage. Force and integral gain were constantly monitored to use minimum force with good response of the feedback system to obtain better images. To get images with good resolution, samples were acquired with 512 points × 512 lines of resolution. Figure 5a shows the topographic signal obtained by scanning the flagellum region of a T. cruzi epimastigote using the intermittent contact mode. This image allows the observation of a furrow along the main axis of the flagellum. This structure has not been seen by any other kind of light or electron microscopy techniques. 3. Phase mode analysis. Phase mode occurs when the tip is in intermittent contact with sample. This mode determines the topographic, viscoelastic, and adhesive properties of the samples. In this mode, the phase shift of the oscillating cantilever is measured relative to the driving signal. Images were obtained at the same moment of intermittent contact mode. Figure 5b shows phase signal image of the same region detected by intermittent contact mode shown in Fig. 5a. The images obtained by phase mode have more contrast, allowing for better

Atomic Force Microscopy as a Tool for the Study of the Ultrastructure

219

Fig. 5. AFM images of the flagellum of T. cruzi previously extracted with detergent. The use of intermittent contact mode allowed a better view of the topography of the flagellum and the furrow that exists in its major axis (arrows) (a). Phase image (b) complements the information from the flagellum height signal (a). Moreover, phase mode is more evident when the phase signal is overlaid with the topographical image (c). Three-dimensional visualization reveals higher levels of detail in the analyzed structures.

v­ isualization of some flagellar structures, such as the PFR and the periodic organization of protrusions in the furrow. 4. Image analysis. AFM image processing (line-wise flattening only) was performed in IGOR-PRO (Wavemetrics, Portland, OR) using a MFP-3D template developed by Asylum Research. For better topographic visualization, topographic, phase, and an overlay of both images are displayed as three-dimensional views. Figure 5c shows an overlay of the phase image and the topographic image. For more details, see Note 5.

220

de Souza et al.

4. Notes 1. 106 cells/mL is a good cell density starting point to have the trypanosomatids well distributed in a 100 mm2 region. 2. The plasma membrane extraction could be done with many different detergents, or solvents, such as methanol. We found that 1% Nonidet NP-40 diluted in PHEM buffer worked well. However, it is important to note that this concentration, as well as the type of detergent and period of membrane extraction, can be changed. We suggest starting with 1% Nonidet NP-40 in PHEM buffer and then analyzing the refractive power of the cell by light microscopy. Transmission electron microscopy must also be used to confirm plasma membrane extraction. 3. To observe biological samples by SEM or in air conditions by AFM, cells must be critical point-dried because of the high vacuum in the SEM chamber, which could change the integrity of the specimen surface due to the boiling of reminiscent hydrated specimens. Air-dried specimens could cause deformations and the collapse of structures. Because of this, the use of a machine that substitutes the liquid embedding the samples for one with a lower surface tension could reduce damage to the sample. Using critical point drying, it is possible to pass from liquid to gas without any abrupt change in state. Critical point drying is the safest method for drying samples. Additionally, after plasma membrane extraction, critical point drying can be used to maintain the architecture of cells. 4. Field emission scanning electron microscopy allows for the acquisition of high-resolution images due to the field emission cathode in the electron gun of the scanning electron microscope, which produces narrower probing beams at both low and high electron energies, resulting in improved spatial resolution. 5. The image processing techniques used in all types of microscopy are intended to most accurately reproduce the information obtained. The simplest processing technique is the adjustment of the brightness and contrast of the image. Other techniques could include operations, such as image rotation, warping, color balancing, etc. These may change the results of the image and increase the fluorescence signal. For example, algorithmic programs like blur and sharpen can be used to reduce and enhance image details, respectively, by changing pixel values based on a subjective combination with surrounding pixels.

Atomic Force Microscopy as a Tool for the Study of the Ultrastructure

221

Acknowledgments The authors are grateful for support from Conselho Nacional de Desenvolvimento Cinetífico e Tecnológico (CNPq), Financiadora de Estudos e Projetos (FINEP), Fundação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES), and Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro (FAPERJ). References 1. Binnig, G. and Rohrer, H. (1982) Scanning tunneling microscopy. Helv. Phys. Acta. 55, 726–735. 2. Binnig, G., Quate, C. F. and Gerber, C. (1986) Atomic force microscopy. Phys. Rev. Lett. 56, 930–933. 3. Bustamante, C. and Dunlap, D. (1991) Application of scanning tunneling microscopy to structural biology. Semin. Cell Biol. 2, 179–185. 4. Dvorak, J. A., Kobayashi, S., Abe, K., Fujiwara, T., Takeuchi, T. and Nagao, E. (2000) The application of the atomic force microscope to studies of medically important protozoan parasites. J. Elec. Microsc. 49, 429–435. 5. Akaki, M., Nakano, Y., Nagayasu, E., Nagakura, K., Kawai, S. and Aikawa, M. (2001) Invasive forms of Toxoplasma gondii, Leishmania amazonensis and Trypanosoma cruzi have a positive charge at their contact site with host cells. Parasitol. Res. 87, 193–197.

6. Dufrêne, Y. F. (2002) Atomic force microscopy, a powerful tool in microbiology. J. Bacteriol. 184, 5205–5213. 7. Dufrêne, Y. F. (2008) Towards nanomicrobiology using atomic force microscopy. Nat. Ver. Microbiol. 6, 674–680. 8. Wright, C. J. and Armstrong, I. (2006) The application of atomic force microscopy force measurements to the characterization of microbial surfaces. Surf. Interface Anal. 38, 1419–1428. 9. Garcia, C. R., Takeuschi, M., Yoshioka, K. and Miyamoto, H. (1997) Imaging Plasmodium falciparum-infected ghost and parasite by atomic force microscopy. J. Struct. Biol. 119, 92–98. 10. Rocha, G. M., Miranda, K., Weissmüller, G., Bisch, P. M., de Souza, W. (2008) Ultrastructure of Trypanosoma cruzi revisited by atomic force microscopy. Microsc. Res. Tech. 71, 133–139.

wwwwwwwwwwwwww

Chapter 15 Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy Andreas Ebner, Hermann Schillers, and Peter Hinterdorfer Abstract Erythrocytes (red blood cells, RBCs) are the most common type of blood cells in vertebrates. Many diseases and dysfunctions directly affect their structure and function. Employing the atomic force microscope (AFM) physical, chemical, and biological/physiological properties of RBCs can be studied even under near-physiological conditions. In this chapter, we present the application of different AFM techniques to investigate and compare normal and pathological RBCs. We give a detailed description for nondestructive immobilization of whole intact RBCs and explain preparation techniques for isolated native RBC membranes. High-resolution imaging of morphological details and pathological differences are demonstrated with healthy and systemic lupus erythematosus (SLE) erythrocytes revealing substructural changes due to SLE. We also present the technique of simultaneous topography and recognition imaging, which was used to map the distribution of cystic fibrosis transmembrane conductance regulator sites on erythrocyte membranes in healthy and cystic fibrosis-positive RBCs. Key words: Erythrocytes, TREC, Membrane preparation, Molecular recognition, Recognition imaging, Cystic fibrosis, AFM tip chemistry

1. Introduction Erythrocytes (red blood cells, RBCs) are the most common type of blood cells in vertebrates. They are well known for their ability to transport oxygen. Many diseases and dysfunctions directly affect the structure and function of erythrocytes. Employing the atomic force microscope (AFM) physical, chemical, and ­biological/physiological properties of RBCs can be studied even under near ­physiological conditions. A number of AFM studies have focused on disease-related changes (for a review thereof, see ref. 1). In 1992, Zacheé already investigated changes in the shape of RBCs of patients after splenectomy (2). Investigations on RBC ­deformations Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_15, © Springer Science+Business Media, LLC 2011

223

224

Ebner, Schillers, and Hinterdorfer

(3–5), nano-rheological properties (6), as well as viscoelastic s­ tudies (7) brought new insights into the behavior of RBCs. Hypercholesterolemia decreases the deformability of RBCs, which impairs their hemorheological behavior and ­promotes atherosclerosis (8). The hemolytic activity of the amphipathic peptides has been correlated to the phosphocholine-to-sphingomyelin ratio (PC/SM) (9). Hyperhomocysteinemia and lipid abnormalities are commonly found in patients with chronic renal failure; both are recognized as risk factors for atherosclerosis (10). Many diseases of the heart and circulatory system have been linked with insufficient deformability and increased aggregability of RBCs (11). Further­ more, in essential hypertension, the deformability of RBCs is reduced (12). Excessive amounts of NO leads to damage of erythrocyte plasticity. The loss of deformability is accompanied by increased fragility of erythrocyte membranes as measured by enhanced release of free hemoglobin (13). Changes in erythrocyte membrane parameters, especially in phospholipids, can produce serious metabolic disorders and influences the rheologic properties of erythrocytes in patients with Binswanger’s disease (14). The observed changes in erythrocyte membrane fluidity among Cyclosporin A-treated patients correlate with more frequent prevalence of hemolytic anemia among Cyclosporin A (15). Therefore, RBCs have been used as a prototypical cellular system to study drug-mediated plasma bilayer effects (16). Here, we describe the application of different AFM techniques to investigate and compare normal and pathological RBCs. Healthy and cystic fibrosis (CF)-positive RBCs, as well as systemic lupus erythematosus (SLE)-positive RBCs have been probed. Since AFM allows measurements under physiological conditions, it offers a well-suited technique to explore morphological details and follow functional changes in RBC. For imaging of morphological details and pathological differences, a whole intact RBC has to be immobilized in a nondestructive way (e.g., on a wheat germ agglutinin matrix) and should be scanned using a gentle imaging technique like tapping-mode AFM. In this chapter, such an experiment is demonstrated with healthy and SLE erythrocytes revealing substructural changes due to SLE. For the investigation of membrane proteins at the single protein level and for investigation of the inner side of the cellular membrane, further preparation techniques are needed. Hence, a second focus of this chapter is a preparation technique to produce isolated but native inside-out orientations of RBC membranes. RBCs are tightly attached to a support and, e.g., exposed to a fluid flow-imposed shear stress which opens the cells. By imaging the flat RBC membrane, more topographical details can be resolved, than when imaging the whole cell. A third main aspect of this chapter is the introduction of a recently developed AFM technique, which facilitates the simultaneous acquisition of topographical images and corresponding

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

225

recognition maps. This technique is exemplified in a comparative study of RBC membranes from healthy individuals and cystic fibrosis patients. The cystic fibrosis transmembrane conductance regulator (CFTR) is a cAMP activatable membrane protein that acts not only as an ion channel, but also as a regulator of several membrane conductances (17). A mutation in the gene encoding for CFTR results in the severe disease of cystic fibrosis. The most common CF-associated mutation is the deletion of phenylalanine at residue 508, ∆F508 CFTR. CFTR bearing the ∆F508 mutation fails to progress through the normal biosynthetic pathway and fails to traffic to the plasma membrane. As a result, CFTR ∆F508 is mislocalized and is not present in the apical membrane of epithelia cells (18). Consequently, the apical membrane of CF cells is Cl− -impermeable resulting in an impaired electrolyte transport and fluid secretion by several epithelia, including the sweat duct, exocrine pancreas, and the pulmonary airways (19). However, it was shown that the membrane distribution of ∆F508 CFTR is tissue-specific and exhibits variation of expression from null to apparently normal amounts (20, 21). CFTR is not only found in epithelia, but also in human erythrocytes as shown by using Western blot techniques (22), in studies of Plasmodium falciparum-induced channel activation (23), by deformationinduced CFTR-dependent ATP-release (24) and functional studies (25, 26). With recognition imaging, it was possible to prove unequivocally that CF-positive erythrocytes have a significant lower number of CFTR proteins compared to healthy RBCs.

2. Materials 2.1. Tight Attachment of Whole RBCs

1. Mica (Muscovit) sheets (E. Groepl, Austria).

2.1.1. Aminofunctio­ nalization

3. Molecular sieve 0.4 nm (Merck, Germany).

2. Ethanolamine hydrochloride (Sigma–Aldrich, Austria). 4. Ethyleneglycol-bis(Succinimidyl.succinate) = EGS (Sigma–Aldrich, Austria). 5. Desiccator 5 L with O-ring.

2.1.2. Covalent Coupling of Capturing Biomolecules (e.g., Wheat Germ Agglutinin) 2.1.3. Red Blood Cell Preparation and Attachment

1. Phosphate buffered saline (PBS): 5 mM Na2HPO4, 150 mM NaCl, pH 7.4. 2. Capturig biomolecules (e.g. wheat germ agglutinin, WGA). 3. Small glass chamber, ~30 mm diameter. 1. Purified RBCs.

226

Ebner, Schillers, and Hinterdorfer

2.2. Isolated RBC Membranes 2.2.1. Membrane Preparation

1. PBS: 5 mM Na2HPO4, 150 mM NaCl, pH 7.4. 2. Ethylene glycol tetra-acetic (EGTA) (Sigma, Deisenhofen, Germany). 3. Very low salt buffer (vlsb): 0.3  mM Na2HPO4/NaH2PO4, 0.2 mM EGTA, pH 7.4. 4. Poly-l-lysine (PLL) (Sigma, Deisenhofen, Germany). 5. Paraformaldehyde (PFA) (Sigma, Deisenhofen, Germany). 6. Bovine serum Germany).

albumin

(BSA)

(Sigma,

Deisenhofen,

7. Glutaraldehyde (Sigma, Deisenhofen, Germany). 8. Masterflex PTFE diaphragm pump 7090-62 (Cole-Parmer Instruments, Chicago, USA). 9. Microcentrifuge 5417R (Eppendorf, Germany). 10. Shaker MTS 4 electronic (IKA-Works, Wilmington, USA). 2.2.2. Imaging of Isolated Membranes

1. Multimode AFM equipped with Nanoscope IIIa controller (Veeco, Santa Barbara, USA) or PicoPlus AFM equipped with MACMode controller (Agilent, Santa Clara, USA). 2. MSCT SiN3 cantilever k = 0.01 N/m (Veeco, Santa Barbara, USA).

2.2.3. Recognition Imaging of Isolated Membranes

1. Magnetically coated AFM cantilever, spring constant ~0.1 N/m. 2. Aminopropyl-triethoxysilane (APTES). 3. Triethylamine (TEA). 4. Heterobifunctional polyethylene(glycol) cross-linker. 5. Specific target molecule (e.g., antibody). 6. Dessicator (5 L) with an O-ring. 7. AFM setup equipped with a topography and recognition imaging (TREC) box (Agilent, Santa Clara, USA).

3. Methods The investigation of RBCs with the AFM can be performed on the outer or inner cell membrane. In the case of the outer membrane, it is possible to image whole intact cells, which should be gently fixed to increase the lateral resolution. This allows the exploration of morphological details and pathological differences of whole RBCs, as was observed in a study of Kamruzzahan et al. There, the morphology of erythrocytes from healthy humans with erythrocytes from SLE-positive patients was compared. To perform such measurements, the RBCs have to be mounted firmly on a flat solid

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

227

support. For anchoring cells to a substrate, flexible linkers are appropriate tools to achieve immobilization avoiding direct physical contact to the sample surface (27). Thus, the cells are anchored at a small distance from the surface, which may facilitate unaffected cell physiology. Major tasks for the development of such binding protocols are the realization of defined and suitable coupling chemistry protocols and the adjustment of the surface density of the linkers. In Subheading 3.1 a tight and noninvasive immobilization technique to mount intact RBCs to a flat surface using a short flexible linker is presented. This anchoring protocol is a prerequisite to resolve structural details of the outer cellular membrane using a gentle AFM imaging technique like the tapping mode. To probe the inner side of a red blood cell, a completely different approach has to be used, whereby the RBCs need to be opened. This is described in detail in Subheading  3.2. Here, a technique called “shear opening” allows the generation of tightly attached erythrocyte membranes on a glass coverslip surface facing the inner side out. As a result, the membrane appears significantly stiffer and high-resolution AFM techniques like TREC are possible (28). Subheading 3.3 focuses on TREC, a recent development in dynamic force microscopy. By oscillating a functionalized tip close to its resonance frequency during the lateral scan across the surface, both the sample topography and the corresponding map of recognition sites can be simultaneously obtained. This highly sophisticated technique has proven to be very powerful in biophysical research (29–31) (e.g., in the investigation of RBC membranes (28)). 3.1. Tight Attachment of Whole RBCs

The coupling process of intact RBCs to a solid support can be performed in different ways. One possibility is the use of PLLcoated glass slides (1, 32). The attachment of RBCs is achieved by simple electrostatic interactions between the negatively charged cellular membrane and the positively charged PLL-coated glass surface. In this section, a tight but noninvasive and physiological way to anchor RBCs is shown (3), in which the cell is immobilized via small spacers. The protocol consists of several coupling steps. In the first step, the nonreactive flat surface (e.g., mica) has to be converted into a chemically addressable surface. The ethanolamine aminofunctionalization (33) turned out to be best suited for this purpose (Fig. 1(1)) (see Note 1). In the next step, a short homobifunctional cross-linker allows conversion of the outer amino groups into an amino-reactive surface. EGS has two reactive N-hydroxy-succinimide ester groups, highly reactive toward amino groups, whereas the ethyleneglycol in between acts as a spacer molecule. This space is short enough to avoid loop formation on the aminofunctionalized surface. The first EGS coupling step (to surface amino groups) is performed in a solvent

228

Ebner, Schillers, and Hinterdorfer

Fig. 1. Coupling scheme of erythrocytes via surface-bound lectin. (1) Aminofunction­ alization using ethanolamine hydrochloride. (2) EGS is used to invert the aminofunctionalized surface into a lysine reactive one. (3) Finally, the outer NHS-ester of EGS is used to couple WGA (a lectin with high affinity to membrane glycoproteins). Taken from ref. 3.

(Fig. 1(2)) while the second reactive group of EGS, which is used to covalently couple proteins via their lysine residue, is performed in aqueous buffer (Fig. 1(3)). WGA is used as a coupling protein for tight but noninvasive attachment of RBCs. WGA is a plant lectin that was found to bind carbohydrate moieties of the glycocalix of RBCs (34). In Fig. 3, the coupling scheme is shown. 3.1.1. Aminofunc­ tionalization of the Solid Support

1. Clean the surface by washing 3 × 5 min with ethanol (an ultrasonic bath is ideal). When using mica sheets, freshly cleave both sides directly before performing the following steps. 2. Dissolve 3.3 mg ethanolamine hydrochloride in 6 mL DMSO by gently heating to ~70°C. 3. Add 10% v/v molecular sieves (0.4 nm). 4. Cool down to room temperature. 5. Degas in a desiccator (at aspirator vacuum) for 30 min. 6. Place the freshly cleaved mica or the cleaned glass slides into the solution. 7. Incubate overnight.

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

229

8. Wash 3× 5 min with DMSO and 3× 5 min with ethanol, dry in a gentle nitrogen or inert gas stream (see Note 2). 3.1.2. Coupling of WGA Via EGS

1. Dissolve 10  mg EGS in 10  mL chloroform in a small glass chamber. 2. Place the aminofunctionalized mica sheets or glass slides in the solution. 3. Add 0.5% (v/v) TEA and mix carefully. 4. React for 60 min. Afterward, wash extensively with chloroform and dry the slides in an inert gas or nitrogen stream. 5. Dissolve 0.5 mg/mL wheat germ agglutinin in PBS. 6. Incubate this solution on the EGS functionalized surface and react for 120  min. Afterwards remove unbound WGA by extensive rinsing with PBS. 7. Store in PBS at 4°C until further use.

3.1.3. Red Blood Cell Preparation and Attachment

1. RBCs should be freshly prepared from blood by a suitable isolation method (e.g., by discontinuous density gradient centrifugation). 2. Incubate the suspension of purified RBCs (less than 0.05% white blood cells) in PBS buffer on the WGA-coated surface for 20 min, and wash afterward extensively with PBS buffer. 3. Incubate the RBCs in 1% glutaraldehyde for 1 min and wash with PBS.

3.1.4. AFM Imaging

AFM imaging of RBCs at high lateral resolution is difficult because of the softness and the spherical shape of the cells. In very early studies, Häberle (35) presented an elegant solution to achieve a resolution of ~10 nm by using a special micropipette technique. However, the technique involved excessive technical effort and did not present a generally applicable approach. When using a commercial AFM, RBCs should be imaged with very soft AFM cantilevers to avoid any damage. Furthermore, intermittent contact imaging modes like acoustic or magnetic AC mode AFM are best suited. The AFM probe thus oscillates above the sample and touches the surface only at the end of the downward movement. Thereby, only a very low force is applied to the biological sample. In Fig. 2, AFM images of healthy (left A–D) and SLE-positive (right A–D) RBCs are shown, where the pathological SLE erythrocytes clearly exhibit deformations when compared to the healthy cells.

3.2. Isolated Membranes

The plasma membrane of a cell accommodates diverse membrane proteins, including integral membrane proteins such as receptors, ion channels, and transporters, as well as certain antigens that are peripherally associated with the membrane. The plasma membrane of eukaryotic cells is generally anything but flat. The curvature of

230

Ebner, Schillers, and Hinterdorfer

Fig. 2. Tapping-mode images of RBCs from healthy (left ) and SLE-positive (right ) patients at different scan sizes. Taken from ref. 3.

a cell is formed by lamellopodia, cell body, and cell nucleus. The membrane shows major structures like membrane ruffles (36), microvilli, and cilia, and submembranous structures like the cytoskeleton (37). A closer view reveals unevenness of the plasma membrane like humps and pits, representing endo- and exocytotic activity (38). A huge variety of proteins are heterogenically distributed within the membrane, and many membrane proteins are equipped with highly branched sugars forming the glycocalyx. The glycocalyx, a network of polysaccharides that protrudes up to 100 nm from cellular surfaces, limits the tip access to the membrane surface, thus reducing the resolution (39). These factors impede molecular resolution on whole cells, which is indispensable for dissolving the structure of membrane proteins with AFM. In imaging whole cells, the generated topographical information is limited to the outer cell surface, but several membrane proteins are located on the cytoplasmic side of the cell membrane. To gain more information about the structure of membrane proteins by AFM imaging, it is necessary to isolate cell membranes on a solid support in such a way that the intracellular face of the membrane is accessible for the AFM tip (“inside-out” orientation). 3.2.1. Plasma Membrane Preparation

The membrane of a RBC can be isolated according to a modified method of Swihart et al. (40). By this preparation, the RBC membrane is spread in patches on glass with an inside-out orientation, and gently shaken in a vlsb at 37°C for 20 min in order to remove hemoglobin and remnant cytoskeletal proteins. This approach

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

231

offers several advantages: (a) the membrane is flat due to the fact that no curvature is imposed by underlying structures (e.g., cytoskeleton), (b) the membrane is stiff because it lies on a hard support instead of soft cytosol, and (c) the intracellular face of the membrane can be imaged by AFM at high resolution, since none of it stretches out over the surface. In general, the main principle in isolating cellular membranes is to “glue” the cell on a functionalized surface and subsequently remove the cell body, whereby membrane patches remain in an inside-out orientation (41, 42) on the support (Fig. 3). Preparation steps: 1. Wash glass coverslips with acetone and use a lens cleaning wipe. 2. Coat glass coverslips with 150 mL of 0.01% Poly-L-Lysine (PLL) in H2O for 30 min followed by two washing steps with H2O. 3. Collect ~5 mL of venous blood from donors with an EDTA blood draw container. Dilute four drops of blood with 1,000 mL isotonic PBS containing 0.2 mM EGTA. Spin down cells with 1,450 × g for 4  min and remove the supernatant. Repeat this washing procedure three times. Dilute packed RBC 1:600 with PBS. Transfer 150  mL of this suspension onto a PLL-coated glass coverslip allowing the RBC to adhere for 15 min. Then, remove the droplet. Attention: do not let the sample dry during all following steps!

Fig. 3. Schematic overview of shear opening procedure of red blood cells. Erythrocytes are exposed to fluid flow-imposed shear stress and, as a result, the cells are opened, exposing their cytoplasmic side of the membrane.

232

Ebner, Schillers, and Hinterdorfer

Fig. 4. Detailed scheme of shear opening procedure.

4. By applying shear stress to the immobilized RBC you get “inside-out” oriented RBC membranes. Use a stream of PBS buffer to remove the upper part of the cells by pressing PBS with a flow rate of 36 mL/min through a needle (21 G, distance and angle to the glass coverslip: 5  cm and 20°). The shear stress is applied for 12 s (Fig. 4). It is crucial to use the correct parameters. Changes in flow rate, time, or angle result in a deficient membrane preparation. 5. Wash the membranes, which are firmly attached to the glass surface, two times with a vlsb to clean membranes from hemoglobin and membrane residues. For this, add 150 mL vlsb to the glass coverslips and remove the droplet after 2 min. 6. To detach cytoskeletal proteins, add 150  mL of vlsb to the samples and shake the glass coverslips at 37°C for 20  min (1,000/m in a shaker: e.g., IKA-shaker MTS 4 electronic). 7. Wash the samples three times using 1,000 mL of PBS. This can be done in a 12-well-plate with one glass coverslip per well. 8. Fix the RBC membranes with 500 mL of 4% PFA in PBS for 40 min. Wash three times with 1,000 mL of PBS containing 3% of BSA. 9. To block against unspecific binding, incubate the samples in 1,000 mL of 3% BSA in PBS at room temperature (RT) for 60 min. 10. Wash the samples three times using 1,000 mL of PBS. 3.2.2. Imaging of Isolated Membranes

Soft cantilevers (e.g., with spring constants of 0.01 N/m) should be used for imaging of isolated RBC membranes to avoid damage to the membrane. Imaging can either be done in contact or in tapping mode. An optimal imaging speed is 1–20  mm/s. The above-described membrane preparation method results in a certain configuration of the sample, which is depicted in Fig. 5. Only region C has the required inside-out orientation of the RBC membrane. Thus, use an AFM equipped with a light microscope, so that the cantilever tip should be positioned over this region.

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

233

Fig. 5. Scheme of the sample after preparation seen from the topview.

Fig. 6. AFM images demonstrating the membrane isolation procedure. (a) Erythrocytes attached to poly-l-lysine-coated glass, before exposure to shear stress and (b) after application of the jet stream, resulting in flat inside-out oriented membranes with circular shapes.

In Fig. 6, AFM images are presented showing intact RBCs on the left (Fig. 6a) and RBC membrane patches after application of shear stress on the right (Fig. 6b). The membrane patches attached on glass (Fig.  6b) exhibit flat surfaces with circular shapes and diameters of about 8 mm. Observed heights of the isolated plasma membranes, including the protruding membrane proteins, ranged between 10 and 15 nm. Overlapping membrane edges appeared frequently with heights of about 25  nm (yellow structures in Fig. 6b).

234

Ebner, Schillers, and Hinterdorfer

Fig. 7. High-resolution scan (5 mm × 5 mm) of the cytosolic face of a human erythrocyte. The high protein density together with the tip geometry (tip convolution) almost prevents imaging the lipid bilayer between the proteins. However, at some points the lipid bilayer is visible defining the zero height in this image (taken from ref. 44).

The fluid mosaic model describes a cell membrane as a twodimensional oriented solution of integral proteins in a viscous phospholipid bilayer. Current textbooks estimate a protein content of 50% for typical membranes (43). With this, not only ­proteins, but also a lipid bilayer should be visible in high-resolution­ AFM images of RBC membranes. Figure 7 shows a 5 mm × 5 mm scan of the cytosolic face of a human erythrocyte. A dense package of proteins with different shapes and sizes can be observed, yet the lipid bilayer is not visible. These observations are in good agreement to other AFM studies of human erythrocytes (45) where high-resolution images of the RBC membranes did not reveal ­lipids. The tip convolution contributes to this result, but not exclusively. Either globular shapes of huge intracellular polypeptide chains or a protein content much higher than expected could cause the observed appearance of the human red cell membrane without a visible lipid bilayer. 3.3. Recognition Imaging of Isolated Membranes

TREC allows mapping of topographical details of a specimen and simultaneous investigation of the distribution of receptors on the surface (see Note 3). A prerequisite for such measurements is the use of a sensing, ligand-functionalized AFM tip. The upgrade of a bare tip (usually silicon or silicon nitride) into a monomolecular sensor requires a number of coupling steps, which are listed in Subheading  3.3.1. Subsequently, optimal conditions for the

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

235

TREC experiments are explained and an example for successful TREC imaging is presented with the sensing of CFTR in RBCs of healthy and CF diseased patients. 3.3.1. Tip Chemistry

The vapor phase deposition of APTES is the best suited method for aminofunctionalization of a given surface (see Note 1). By APTES coating, a nondense monomolecular silane layer with optimal physical and chemical behavior is introduced on the AFM tip. In addition, choosing the right parameters allow an accurate and easy adjustment of reactive sites per tip apex (33). A high number of distensible homo- and heterobifunctional cross-linkers have been developed. In the following, one of the most commonly used linkers, the NHS– PEG–aldehyde (46) linker, is exemplary presented to demonstrate ligand coupling to an AFM tip (see Note 4 and Fig. 8). 1. Wash the cantilevers 3× 5 min with chloroform and dry in a gentle nitrogen gas stream. Then, wash 3× 5 min with ethanol and dry again in a gentle nitrogen gas stream. 2. Flood a 5-L desiccator with argon. Open the desiccator and place a vial with 30 mL freshly distilled APTES and a vial with 10 mL TEA inside. 3. Place the cleaned cantilevers in the desiccator, rinse with argon for 60 s, and let it react for 120 min.

Fig. 8. Aminofunctionalized AFM tip gets upgraded into a monomolecular sensor. NHS– PEG–aldehyde linker is coupled via amide-bond formation. Subsequently, the aldehyde group is coupled to the amine residue of a lysine of a protein (e.g., of an antibody). Taken from ref. 28.

236

Ebner, Schillers, and Hinterdorfer

4. Remove the vials from the desiccator and rinse the desiccator with argon for 5–10  min. Store the APTES-coated tips for 2 days under argon before further use. 5. Dissolve the heterobifunctional cross-linker NHS–PEG– aldehyde in chloroform (3.3 mg/mL). Transfer the solution into a small reaction chamber, add the aminofunctionalized cantilevers, and start the reaction by adding 10 mL of TEA. 6. After 60 min, wash the tips with chloroform and dry them in a gentle stream of inert gas. 7. Place the aldehyde terminated tips on a clean surface (e.g., Parafilm™) and add 100 mL of the protein solution (typically 0.1–3 mg/mL protein in PBS buffer) on the cantilevers. 8. Add 2 mL 1 M NaCNBH3 and allow to react for 110 min. 9. Add 5 mL 1 M ethanolamine and allow to react for 10 min. 10. Wash the tips with PBS and store in PBS at 4°C. 3.3.2. Parameters for Recognition Imaging

When performing TREC experiments, ligand-functionalized tips have to be approached to the surface followed by the oscillation of the tip close to the sample surface. For successful measurements, the working amplitude has to be adjusted in a way that the upper part of the oscillation amplitude is damped (Fig. 7). Details for setting up the optimal imaging conditions can be found in Preiner et al. (47). Finally, the specificity of the interaction has to be proven. This can be done in two ways. By adding free ligands to the sample, the surface-embedded receptors form a receptor–ligand complex resulting in a hindrance of the tip–ligand surface–receptor ­complex  formation (“surface block”). Adding free receptors, on the other hand, passivates the AFM tip to complexation of free receptors with the tip-bound ligand (“tip block”) (see Note 5). For a more detailed explanation of TREC, other literature is recommended (29, 31, 48, 49). In the following, the necessary steps for successful TREC imaging are explained (see Fig. 9).

Fig.  9. TREC working principle: A ligand-functionalized AFM tip is oscillated over the sample surface. The lower part of the amplitude is used for driving the AFM feedback loop, resulting in the topography image, whereas the upper part is affected by molecular recognition, yielding a simultaneously acquired recognition image.

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

237

1. Approach the functionalized cantilever to the sample surface (see Note 6). 2. Adjust the amplitude for TREC imaging. When using the typical PEG-18 cross-linker, start with a free amplitude of ~20 nm and adjust the set point amplitude to a value close to the free amplitude. Decrease the free amplitude until recognition signals are detected. If no recognition signals are detected, even when using various amplitudes, replace the cantilever. 3. Image with a scanning velocity less than 3 mm/s (in the fast scan axis). 4. Specificity proof experiment: use either a tip or a surface “block experiment” (see Note 5). All specifity proof experiments should performed at the same conditions as the unblocked recognition imaging experiments (see Fig. 10) (1) Tip block. After performing a successful TREC experiment, incubate the ligand-functionalized tip in a solution containing the corresponding receptor. Remove the excess of unbound receptor molecules by extensive washing with measuring buffer (2) Surface block. Alternatively, add a solution of free ligands to the immobilized RBC surface and incubate for 30–60 min. Wash the membrane with measuring buffer to remove unbound ligands. The following steps are optional to estimate the CFTR ­density using nanocrystalline fluorophores quantum dots: 5. To localize CFTR, incubate the isolated membranes at 4°C overnight with a mouse monoclonal antibody against a C-terminal epitope of CFTR (Mab25031, R&D Systems Inc., Minneapolis, MN) diluted to 8 mg/mL in PBS containing 3% BSA. 6. Wash the samples five times with 1,000 mL of 3% BSA in PBS for 5 min. 7. Incubate at RT for 1 h with anti-mouse antibody, labeled with the nanocrystalline fluorophores Quantum Dots (1102-1, Quantum Dot Corp., Hayward, CA) diluted to 10 nM in PBS with 3% BSA). 8. Wash the sample three times with 1,000 mL of PBS with 3% BSA for 5 min. 9. Wash the sample two times with 1,000 mL of PBS (without BSA) for 5 min. 10. Fix the membrane patches with 0.5% glutaraldehyde in PBS for 40 min.

238

Ebner, Schillers, and Hinterdorfer

Fig. 10. Topography and recognition images of isolated erythrocyte membranes. TREC imaging topography of a non-CF (a) and of a CF (d) erythrocyte membrane. Dark spots in the recognition images (b) and (e) represent the specific interaction sites between the modified tip (i.e., anti-CFTR antibody tip) and CFTR, corresponding to the same areas as shown in (a) and (d). The CF membrane (e) clearly reveals fewer recognition events compared to the non-CF membrane (b). Blocking the membrane of non-CF (c) and CF (f) erythrocytes with free anti-CFTR antibody results in the disappearance of the recognition signals (block efficiency >90%), confirming the specificity of recognition. Scale bar is 200 nm, z scale 80 nm. Taken from ref. 28.

11. Wash the samples with H2O three times to remove buffer salts 12. Dry the samples in air.

4. Notes 1. Aminofunctionalization: Two aminofunctionalization techniques are presented here. The vapor phase deposition of APTES is preferable for the tip coating. Although this coating requires accurate handling, it prevents possible damage to the magnetic coating of cantilevers used for TREC. In contrast, for the first step in the immobilization protocol of whole RBCs, the simpler ethanolamine procedure is sufficient but can also be substituted with the APTES vapor phase deposition.

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

239

2. The stability of aminofunctionalized glass surfaces significantly depends on the storage conditions. When stored under argon, the critical oxidation of the amino groups can be avoided and the slides (or sheets) can be stored over weeks. 3. TREC is a sophisticated method allowing investigation of receptor distributions under near physiological conditions, which can only be briefly explained within this book chapter. For more details, please find literature focused on this (29, 31, 48, 49). 4. Tip chemistry. The shown aldehyde–PEG–NHS cross-linker is well suited to couple ligands like proteins via their lysines (using the primary amine group). In addition, an ever increasing number of other heterobifunctional cross-linker exists, using different coupling strategies. The most prominent ones are the site-directed coupling via a His6-tag (50) or the covalent coupling of a thiolated ligand via PDP–PEG–NHS (46). More possibilities and detailed information are given in ref. 51. 5. Specificity proof experiments. In general, the interaction between a tip-bound ligand and a surface-bound receptor can be hindered (blocked) in two ways. Either the tip-bound ligand gets inactivated due to binding of free receptor molecules forming a tip–ligand–receptor complex on the tip (=“tip block”), or the surface-bound receptor gets blocked by addition of free ligand molecules resulting in a surface–receptor– ligand complex. The terminus block in this context means the inactivation of ligands or receptor as specificity proof. 6. The initial approach with a functionalized tip is very critical, since hard contact at the onset can destroy the functional unit on the tip. To overcome this, the approaching velocity should be set very low and the feedback parameters high. References 1. Dulinska, I., Targosz, M., Strojny, W., Lekka, M., Czuba, P., Balwierz, W. & Szymonski, M. (2006). Stiffness of normal and pathological erythrocytes studied by means of atomic force microscopy. Journal Of Biochemical And Biophysical Methods 66, 1–11. 2. Zachee, P., Boogaerts, M. A., Hellemans, L. & Snauwaert, J. (1992). Adverse Role Of The Spleen In Hereditary Spherocytosis - Evidence By The Use Of The Atomic Force Microscope. British Journal Of Haematology 80, 264–265. 3. Kamruzzahan, A. S. M., Kienberger, F., Stroh, C. M., Berg, J., Huss, R., Ebner, A., Zhu, R., Rankl, C., Gruber, H. J. & Hinterdorfer et, a. (2004). Imaging morphological details and pathological differences of red blood cells

using tapping-mode AFM. Biological Chemistry 385, 955–960. 4. Wu, Y. Z., Hu, Y., Cai, J. Y., Ma, S. Y., Wang, X. P., Chen, Y. & Pan, Y. L. (2009). Timedependent surface adhesive force and morphology of RBC measured by AFM. Micron 40, 359–364. 5. Ho, M. S., Kuo, F. J., Lee, Y. S. & Cheng, C. M. (2007). Atomic force microscopic observation of surface-supported human erythrocytes. Applied Physics Letters 91. 6. Bremmell, K. E., Evans, A. & Prestidge, C. A. (2006). Deformation and nano-rheology of red blood cells: An AFM investigation. Colloids And Surfaces B-Biointerfaces 50, 43–48.

240

Ebner, Schillers, and Hinterdorfer

7. Strasser, S., Zink, A., Kada, G., Hinterdorfer, P., Peschel, O., Heckl, W. M., Nerlich, A. G. & Thalhammer, S. (2007). Age determination of blood spots in forensic medicine by force spectroscopy. Forensic Science International 170, 8–14. 8. Koter, M., Franiak, I., Strychalska, K., Broncel, M. & Chojnowska-Jezierska, J. (2004). Damage to the structure of erythrocyte plasma membranes in patients with type-2 hypercholesterolemia. International Journal of Biochemistry and Cell Biology 36, 205–215. 9. Belokoneva, O., Villegas, E., Corzo, G., Dai, L. & Nakajima, T. (2003). The hemolytic activity of six arachnid cationic peptides is affected by the phosphatidylcholine-tosphingomyelin ratio in lipid bilayers. BBABiomembranes 1617, 22–30. 10. de Gómez Dumm, N., Giammona, A. & Touceda, L. (2003). Variations in the lipid profile of patients with chronic renal failure treated with pyridoxine. Lipids in Health and Disease 2, 7. 11. Starzyk, D., Korbut, R. & Gryglewski, R. J. (1999). Effects of nitric oxide and prostacyclin on deformability and aggregability of red blood cells of rats ex vivo and in vitro. J Physiol Pharmacol 50, 629–37. 12. Sandhagen, B. (1999). Red cell fluidity in hypertension. Clin Hemorheol Microcirc 21, 179–81. 13. Starzyk, D., Korbut, R. & Gryglewski, R. (1997). The role of nitric oxide in regulation of deformability of red blood cells in acute phase of endotoxaemia in rats. Journal of physiology and pharmacology: an official journal of the Polish Physiological Society 48, 731. 14. Chen, C., Jia, H., Ma, H., Wang, D., Guo, S. & Qu, S. (1999). Rheologic determinant changes of erythrocytes in Binswanger’s disease. Zhonghua yi xue za zhi = Chinese medical journal; Free China ed 62, 76. 15. Wrobel, A., Kaminska, D. & Klinger, M. (2003). 16. Li, A., Seipelt, H., Müller, C., Shi, Y. & Artmann, G. (1999). Effects of salicylic acid derivatives on red blood cell membranes. Pharmacology & toxicology 85, 206–211. 17. Schwiebert, E., Benos, D., Egan, M., Stutts, M. & Guggino, W. (1999). CFTR is a conductance regulator as well as a chloride channel. Physiological reviews 79, 145–166. 18. Welsh, M., Denning, G., Ostedgaard, L. & Anderson, M. (1993). Dysfunction of CFTR bearing the delta F508 mutation. Journal of cell science. Supplement 17, 235. 19. Fuller, C. & Benos, D. (1992). Cftr! American Journal of Physiology- Cell Physiology 263, 267–286.

20. Dupuit, F., Kälin, N., Brezillon, S., Hinnrasky, J., Tümmler, B. & Puchelle, E. (1995). CFTR and differentiation markers expression in nonCF and delta F 508 homozygous CF nasal epithelium. Journal Of Clinical Investigation 96, 1601. 21. Kälin, N., Claaß, A., Sommer, M., Puchelle, E. & Tümmler, B. (1999). F508 CFTR protein expression in tissues from patients with cystic fibrosis. Journal Of Clinical Investigation 103, 1379–1389. 22. Sterling Jr., K., Shah, S., Kim, R., Johnston, N., Salikhova, A. & Abraham, E. (2004). Cystic fibrosis transmembrane conductance regulator in human and mouse red blood cell membranes and its interaction with ecto-apyrase. Journal Of Cellular Biochemistry 91. 23. Verloo, P., Kocken, C., Van der Wel, A., Tilly, B., Hogema, B., Sinaasappel, M., Thomas, A. & De Jonge, H. (2004). Plasmodium falciparum-activated chloride channels are defective in erythrocytes from cystic fibrosis patients. Journal Of Biological Chemistry 279, 10316. 24. Sprague, R., Ellsworth, M., Stephenson, A., Kleinhenz, M. & Lonigro, A. (1998). Deformation-induced ATP release from red blood cells requires CFTR activity. American Journal of Physiology- Heart and Circulatory Physiology 275, 1726–1732. 25. Stumpf, A., Almaca, J., Kunzelmann, K., Wenners-Epping, K., Huber, S., Haberle, J., Falk, S., Duebbers, A., Walte, M. & Oberleithner, H. (2006). IADS, a decomposition product of DIDS activates a cation conductance in Xenopus oocytes and human erythrocytes: new compound for the diagnosis of cystic fibrosis. Cell Physiol Biochem 18, 243–252. 26. Stumpf, A., Wenners-Epping, K., Wälte, M., Lange, T., Koch, H., Häberle, J., Dübbers, A., Falk, S., Kiesel, L. & Nikova, D. (2006). Physiological concept for a blood based CFTR test. Cellular Physiology And Biochemistry 17, 29–36. 27. Schilcher, K., Hinterdorfer, P., Gruber, H. J., Schindler, H. (1997). A non-invasive method for the tight anchoring of cells for scanning force microscopy. Cell Biology International 21, 769–778. 28. Ebner, A., Nikova, D., Lange, T., Haberle, J., Falk, S., Dubbers, A., Bruns, R., Hinterdorfer, P., Oberleithner, H. & Schillers, H. (2008). Determination of CFTR densities in erythrocyte plasma membranes using recognition imaging. Nanotechnology 19. 29. Ebner, A., Kienberger, F., Kada, G., Stroh, C. M., Geretschlager, M., Kamruzzahan, A. S. M., Wildling, L., Johnson, W. T., Ashcroft, B.,

Normal and Pathological Erythrocytes Studied by Atomic Force Microscopy

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

Nelson, J., Lindsay, S. M., Gruber, H. J. & Hinterdorfer, P. (2005). Localization of single avidin-biotin interactions using simultaneous topography and molecular recognition imaging. Chemphyschem 6, 897–900. Stroh, C., Wang, H., Bash, R., Ashcroft, B., Nelson, J., Gruber, H., Lohr, D., Lindsay, S. M. & Hinterdorfer, P. (2004). Single-molecule recognition imaging microscopy. Proceedings Of The National Academy Of Sciences Of The United States Of America 101, 12503–12507. Stroh, C. M., Ebner, A., Geretschlager, M., Freudenthaler, G., Kienberger, F., Kamruzzahan, A. S. M., Smith-Gill, S. J., Gruber, H. J. & Hinterdorfer, P. (2004). Simultaneous topography and recognition imaging using force microscopy. Biophysical Journal 87, 1981–1990. Nowakowski, R., Luckham, P. & Winlove, P. (2001). Imaging erythrocytes under physiological conditions by atomic force microscopy. Biochimica Et Biophysica Acta-Biomembranes 1514, 170–176. Ebner, A., Hinterdorfer, P. & Gruber, H. J. (2007). Comparison of different aminofunctionalization strategies for attachment of single antibodies to AFM cantilevers. Ultramicroscopy 107, 922–927. Salzer, U., Hinterdorfer, P., Hunger, U., Borken, C. & Prohaska, R. (2002). Ca(++)-dependent vesicle release from erythrocytes involves stomatin-specific lipid rafts, synexin (annexin VII), and sorcin. Blood 99, 2569–2577. Haberle, W., Horber, J. K. H. & Binnig, G. (1991). Force Microscopy On Living Cells. Journal Of Vacuum Science & Technology B 9, 1210–1213. Braet, F., Seynaeve, C., De Zanger, R. & Wisse, E. (1998). Imaging surface and submembranous structures with the atomic force microscope: a study on living cancer cells, fibroblasts and macrophages. Journal Of Microscopy 190, 328–338. Rotsch, C. & Radmacher, M. (2000). Druginduced changes of cytoskeletal structure and mechanics in fibroblasts: an atomic force microscopy study. Biophysical Journal 78, 520–535. Schneider, S., Sritharan, K., Geibel, J., Oberleithner, H. & Jena, B. (1997). Surface dynamics in living acinar cells imaged by atomic force microscopy: identification of plasma membrane structures involved in exocytosis, Vol. 94, pp. 316–321. National Acad Sciences. Le Grimellec, C., Lesniewska, E., Cachia, C., Schreiber, J., De Fornel, F. & Goudonnet, J. (1994). Imaging of the membrane surface of MDCK cells by atomic force microscopy. Biophysical Journal 67, 36–41.

241

40. Swihart, A., Mikrut, J., Ketterson, J. & Macdonald, R. (2001). Atomic force microscopy­ of the erythrocyte membrane ­skeleton. Journal Of Microscopy 204, 212–225. 41. Nikova, D., Lange, T., Oberleithner, H., Schillers, H., Ebner, A. & Hinterdorfer, P. (2006). Atomic force microscopy in nanomedicine. Applied scanning probe methods III. Springer, Berlin, 1–27. 42. Oberleithner, H., Schillers, H., Schneider, S. & Henderson, R. (2001). Nanoarchitecture of Plasma membrane visualized with atomic force microscopy. Ion channel localization methods and protocols methods in pharmacology and toxicology. Humana, Totowa, NJ, 405–424. 43. Cooper, G. & Hausman, R. (2000). The cell: a molecular approach, ASM Press Washington, DC. 44. Schillers, H. (2008). Imaging CFTR in its native environment. Pflugers Archiv-European Journal Of Physiology 456, 163–177. 45. Yamashina, S. & Katsumata, O. (2000). Structural analysis of red blood cell membrane with an atomic force microscope. Journal of Electron Microscopy 49, 445–451. 46. Ebner, A., Wildling, L., Kamruzzahan, A. S. M., Rankl, C., Wruss, J., Hahn, C. D., Holzl, M., Zhu, R., Kienberger, F., Blaas, D., Hinterdorfer, P. & Gruber, H. J. (2007). A  New, Simple Method for Linking of Antibodies to Atomic Force Microscopy Tips. Bioconjugate Chem. 18, 1176–1184. 47. Preiner, J., Ebner, A., Chtcheglova, L., Zhu, R. & Hinterdorfer, P. (2009). Simultaneous topography and recognition imaging: physical aspects and optimal imaging conditions. Nanotechnology 20. 48. Hinterdorfer, P. & Reich, Z. (2008). Molecular Recognition Force Microscopy: From Simple Bonds to Complex Energy Landscapes. Nanotribology and Nanomechanics: An Introduction, 279. 49. Preiner, J., Ebner, A., Chtcheglova, L., Zhu, R. & Hinterdorfer, P. (2009). Simultaneous topography and recognition imaging: physical aspects and optimal imaging conditions. Nanotechnology 20, 215103. 50. Kienberger, F., Kada, G., Gruber, H. J., Pastushenko, V., Riener, C., Trieb, M., Knaus, H.-G., Schindler, H. & Hinterdorfer, P. (2000). Recognition force spectroscopy studies of the NTA-His6 bond. Single Mol. 1, 59–65. 51. Ebner, A., Wildling, L., Zhu, R., Rankl, C., Haselgrubler, T., Hinterdorfer, P. & Gruber, H. J. (2008). Functionalization of probe tips and supports for single-molecule recognition force Microscopy. Topics in Current Chemistry 285, 29–76.

wwwwwwwwwwwwww

Chapter 16 The Growth Cones of Living Neurons Probed by the Atomic Force Microscope Davide Ricci, Massimo Grattarola, and Mariateresa Tedesco Abstract A detailed report of experimental findings concerning the use of atomic force microscopy to probe growth cones of chick embryo spinal cord neurons under vital conditions is given. The role played by indentation in the making of images and force-versus-distance curves is critically discussed. As a result, the thickness of growth cone regions is quantitatively estimated. By comparing the obtained images with descriptions given in the literature on the basis of other microscopy techniques, a central (C) region and a peripheral (P) region are identified, characterized by a different thickness and a different structural organization. Moreover, clusters of adhesion molecules are tentatively identified in regions where neuron arborizations were challenged by the atomic force microscope (AFM) tip. Key words: AFM, Indentation, Force maps, Living cells, Arborizations

1. Introduction There is a large body of recent literature describing the use of atomic force microscopy (1) for the study of living cells. These experimental findings clearly indicate that atomic force microscope (AFM) is a very valuable tool for the three-dimensional imaging of flat biological samples strongly adhering to a substrate, with a lateral resolution in between the resolutions of optical and electron microscopy. Moreover, a very relevant feature of AFM is the capability of analyzing local mechanical properties of living cells. The expression “flat biological samples” includes layers of cells, such as epithelia (2, 3), and single cells such as fibroblasts and glial cells (4, 5). The technique, in its present state, seems to be less appropriate for globular structures, such as neuron bodies (6), and for string-like structures, such as neuron arborizations (7, 8).

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_16, © Springer Science+Business Media, LLC 2011

243

244

Ricci, Grattarola, and Tedesco

On the contrary, neuronal growth cones are subcellular structures that seem to be very appropriate for AFM analysis: they are flat, highly specialized regions that very strongly adhere to the substrate. Moreover, the mechanical properties of these structures (i.e., the cytoskeleton local organization) are of great relevance for understanding the development of neural architectures and therefore, the potential of micromechanical information of the AFM is of particular value. On the basis of these premises, this chapter will be devoted to a detailed report of experimental findings concerning the use of AFM to probe growth cones of chick embryo spinal cord neurons under vital conditions.

2. Materials 2.1. Animals

1. Chick embryos 7–8 days old.

2.2. Chemicals and Reagents

1. Chick embryo extract (Gibco 16460-016). 2. Hanks’ balanced solution (HBSS, Gibco 24020-091). 3. Bovine serum albumin (BSA, Sigma A-7030). 4. Trypsin solution 0.25% (Gibco 25050-014). 5. Trypsin inhibitor (Sigma T-6522). 6. Deoxyribonuclease (DNAse), type I (Sigma D-5025). 7. DMEM-F12 (Gibco 31331-028). 8. Fetal bovine serum (heat inactivated) (Gibco 10108-157). 9. Horse serum (heat inactivated) (Gibco 26050-070). 10. Poly-d-lysine (Sigma P-7280) or poly-l-lysine (Sigma P-9155). 11. Stock supplement solution N-2 (Gibco 17502-048). 12. 5-Fluoro-2¢-deoxyuridine antimitotic agent (Sigma F-0503). 13. Phosphate-buffered saline (PBS, Gibco 14287-080). 14. Glutaraldehyde solution (Sigma G-6257).

2.3. Equipment

1. Atomic force microscope: Park Scientific Instrument Autoprobe CP (Thermomicroscopes, Sunnyvale, CA). 2. Silicon nitride pyramidal tips on cantilevers with 0.01  N/m nominal spring constant (Thermomicroscopes, Sunnyvale, CA). 3. Dissection microscope (Wild-Leitz). 4. Microdissection forceps (FTS Dumon #5 biologie). 5. Forceps (large, small) (FST). 6. Scissors (fine) (FST). 7. Disposable conical tubes (Falcon 2170, 2195 or equivalent).

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope

245

8. Disposable cell culture dishes (100 mm ø Falcon, 35 mm ø Falcon). 9. Phase contrast microscope (Diavert-Leitz). 10. Thermo-controlled water bath (37.5°C). 11. Centrifuge. 12. Glass slides.

3. Methods The methods described below outline (1) the neuron cell culture and sample preparation, (2) the atomic force microscope setup for imaging, (3) the acquisition of force-versus-distance and indentation curves, (4) the results obtained and their interpretation, and (5) a comparison with other techniques. 3.1. Neural Cell Culture and Sample Preparation 3.1.1. Chick Embryo Spinal Cord Neuron Extraction

Spinal cord neurons were obtained through dissection of spinal cords from 8-day chick embryos and plated on treated coverslips. The dissected cords were minced in HBSS and enzymatically dissociated in 0.05% trypsin at 37°C for 25  min, then washed in CMF-HBSS containing 0.3% BSA, 0.005% DNAse (deoxyribonuclease, type I), and 0.025% trypsin inhibitor. After mechanical dissociation, the resulting single cells were suspended in MEMF12 (1:1) supplemented with 5% fetal bovine serum (FBS), 5% inactivated horse serum, and 5% chick embryo extract for plating on culture substrata.

3.1.2. Neural Cell Culture and Sample Preparation for AFM Investigations

To prepare the culture substrata, first glass slides were cut to 20 × 40 mm pieces and then cleaned and sterilized (see Note 1). They were then incubated overnight in a poly-d-lysine solution (5  mg in 50  ml distilled water), rinsed three times in distilled water, and then dried in a sterile hood (see Note 2). Plating was made on the glass slides kept in plastic Petri dishes, and the cultures were incubated at 37°C and 5% CO2 (see Note 3). Two days after plating, the medium was replaced with 96% MEM-F12, 3% horse serum, and 1% stock supplement solution N2. To free cultures from nonneural cells, an antimitotic agent (5-fluoro-2 deoxyuridine, 10−6  M) was added to the culture medium 72 h after plating.

3.1.3. Cell Fixation

For the purpose of comparing results obtained on living cells, fixated cells were also prepared. In this case, after keeping the cells in culture for 4 or 5 days, the medium was removed and cultures were briefly rinsed with PBS (Gibco BRL). The cells were fixed for 20–30 min using 0.8% glutaraldehyde in PBS. Finally, the slides were rinsed twice with PBS and dried.

246

Ricci, Grattarola, and Tedesco

3.2. Atomic Force Microscopy Setup for Imaging 3.2.1. AFM Setup

A Park Scientific Instruments Autoprobe CP (Sunnyvale, CA) atomic force microscope was used, equipped with a scanner tube allowing 100 mm (x, y) maximum scan size and 6 mm (z) excursion (see Note 4). All experiments were performed using cantilevers with 0.01 N/m nominal spring constant (see Note 5) and silicon nitride pyramidal tips (see Note 6). Special care was taken in order to avoid contact between liquids and scanner, which would cause permanent damage to the piezoelectric element and eventually to the high voltage electronics. For this purpose, the top half of the microscope containing the scanner was enclosed in a polyethylene film sheet. This allows the scanner to move freely and does not interfere with the magnetic coupling of the sample holder to the scanner. The cantilever chip was mounted on a chip holder that has a glass window behind the cantilever chip. In order to avoid air bubble formation, before mounting the chip holder into the microscope, the glass and cantilever chip were wetted with buffer solution using a syringe, and a droplet of water was trapped (kept in place by surface tension) between the chip and the glass window. The sample was then taken out of the Petri dish, taking care to keep a film of buffer solution onto the surface. To overcome the difficulties of gluing a wet glass slide to the sample holder metal disk and also to overcome the limitations of the x–y table that has only a 12 × 12 mm range, we used the following method. First, we fixed a whole glass slide with cyanoacrylate glue to the metal sample holder disk, which was then placed on the scanner as usual. Second, we placed Vaseline onto this glass slide and pressed the cell-covered glass slide firmly onto it. This allows to move the sample easily in search of a good area for imaging and also to change it quickly (see Note 7).

3.2.2. Tip to Sample Approach Procedure

The first step is to approach the tip to the sample as usual with the stepper motor until the drop hanging from the cantilever holder assembly meets the liquid covering the sample glass slide. A meniscus is then formed and from this moment, the surface of the sample can be seen through the on-axis optical microscope (see Note 8). The tip-to-sample approach was always performed on a glass area next to the cell to be imaged, and before scanning, the force set point was lowered to a small value (0.5 nN) in order to avoid cell damage.

3.2.3. AFM Settings for Imaging

Force-versus-distance curves prior to and after imaging were recorded routinely for cantilever deflection calibration purposes and for sample stiffness estimation. These curves were transformed into force-versus-indentation plots, using as reference a force-versus-distance curve obtained on glass during the same session. Images were taken by recording two acquisition channels

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope

247

in the AFM simultaneously: the Z-piezo driving voltage and the error signal from the feedback loop. The first signal is proportional to the Z-piezo displacement necessary to maintain the cantilever deflection (force) at the set point during scanning, while the second one records deviations of the cantilever deflections (hence from the set force) from the set point value. In order to obtain imaging with higher spatial frequency resolution, we tuned the feedback loop parameters so that only the average cantilever deflection was kept near the set point value, allowing the system to generate a meaningful image from the error signal channel, which has a wider frequency band (9). Typical scanning speeds were between 13 and 41 mm/s (see Note 9). 3.3. Acquisition of Force-VersusDistance and Indentation Curves 3.3.1. Force-VersusDistance Curves

3.3.2. Force-VersusIndentation Curves

Force-versus-distance curves were obtained by using the standard PSI software, which records the cantilever deflection, while driving the piezo in the z-direction following a triangular wave. The software allowed us to set the wave frequency and to average the force-versus-distance curves obtained consecutively at the same point. The curves corresponding to a given image were stored in a digital file (1,024 points for each force curve) for further processing. The force scale for these curves was calibrated by using, as a reference substrate, the glass the cells adhered to. As the glass did not appreciably indent under the loads applied, from the slope of the linear portion (after tip contact) of the force-versusdistance curve, we derived the conversion factor from the error signal (in mV) to the cantilever deflection (in nm) and hence to the applied force (in nN), through the spring constant K of the cantilever (Force = K∙cantilever deflection, nominal K = 0.01 N/m). This conversion factor depended on the intensity of the laser beam reflected from the backside of the cantilever and on the area of the spot on the photodiode. Therefore, for each series of curves taken in the same session, we left the laser alignment unchanged, and began and finished the experiment by performing a calibration curve on the glass. When pushed against a soft sample, the tip of the AFM will indent the surface, and the shape of the indentation curve (i.e., the relationship between the load applied and the tip penetration) will give information on the stiffness of the sample. The force-versusindentation curves were calculated by using the approach portion of the force-versus-distance curves. The first step was to take a force-versus-distance curve on a naked glass portion of the sample as reference. From this curve, the coefficient of linear relationship between the Z-piezo displacement and cantilever deflection was derived. From each of the force-versus-distance curves taken on the cells, the calibration line was subtracted, thus obtaining the force-versus-indentation curve (see Note 10).

248

Ricci, Grattarola, and Tedesco

3.4. Results Obtained and Their Interpretation 3.4.1. Imaging

Figure 1 is a collage of various images (acquired in error mode) taken on the same growth cone of a spinal cord neuron, adhering to a treated slide just taken out of the incubator. Figure 1a shows a topview rendering of the growth cone. Filamentous cytoskeletal structures are evident in the thick region (arrows). Figure  1b shows the top region of the cone (partially missing in Fig. 1a). Small “dot-like” structures are visible (arrows). A further zoom of the top right corner of the cone is shown in Fig. 1c, showing a meshwork of cytoplasmic structures (arrows). Finally, Fig. 1d shows the image of the growth cone after about 10 min of continuous scanning. The background globular structure on the left (arrow), present in both images, can be used to align the two images. Most of the periphery of the growth cone has clearly retracted.

Fig. 1. Growth cone of a living spinal cord neuron adhering to a polylysine-coated glass slide. (a) Topview rendering of a error mode image. Filamentous cytoskeletal structures are evident in the thick region (arrows). (b) Scan of the top region of the cone (partially missing in a). Small “dot-like” structures can be seen in the thick domain. (c) Zoom of the top right corner of the cone. A meshwork of cytoplasmic structures appears (arrows). (d) Image of the growth cone after about 10 min of continuous scanning. Most of the periphery of the growth cone has retracted.

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope

249

Fig. 2. 3D-shaded rendering of the z-piezo signal image acquired simultaneously with the image in Fig. 1a, with a pictorial representation of the possible “real” versus measured profile on the growth cone.

An increase in the relief of the filamentous structures projecting toward the neurite can be noticed. Figure 2 shows a 3D rendering of the growth cone, as derived from the z-piezo (topographic) image (not shown), taken simultaneously with the image in Fig. 1a. It should be noted that the cone thickness shown in the figure is affected by the indentation of the tip on the neuron. Nevertheless, “true” thickness can be estimated, as described in Subheading 3.4.2. In the 3D image, a thick and a flat region can be tentatively identified, separated by a continuous relief. Figure 3a, b show the growth cone of another neuron belonging to another slide analyzed immediately after taking it out of the incubator. A thick tubular zone is again evident toward the neurite. Careful inspection allows one to detect a surrounding lowcontrast region with flat protrusions (arrows). For comparison, Fig.  3c shows a similar growth cone after fixation. Similar to Fig. 3a, d shows a growth cone from another living neuron where one can identify a thick tubular region surrounded by spiky structures (arrows). Figure 4a shows a small whole neuron with several arborizations. Toward the apical end, most of them seem to be disrupted. Interestingly enough, a “trace” of the borders of the arborizations is evident (Fig.  4b, c). The trace is made of small (150-nm diameter) dot-like structures, which could be identified as clusters of adhesion molecules.

250

Ricci, Grattarola, and Tedesco

Fig. 3. Series of three images of different growth cones where the peripheral region has been detected by the AFM. (a) Growth cone of a living neuron analyzed immediately after taking it out of the incubator. A thick tubular zone is again evident. Careful inspection allows one to detect a surrounding low-contrast region with flat protrusions (arrowheads). Image obtained recording the z-piezo signal. (b) Image of a growth cone after fixation, shown for comparison purposes. Z-piezo signal image. (c) Growth cone from another living neuron where one can identify a thick tubular region, surrounded by spiky structures (arrows). Image obtained by recording the error signal.

3.4.2. Indentation, Topography, and Mechanical Properties

Figure 5 shows a series of representative force-versus-indentation curves acquired upon a growth cone of a living neuron. Curves 1–3 in Fig. 5 were taken moving at steps of 3 mm away from the growth cone edge toward the neurite and represent the typical behavior of a “soft” portion of a living cell (4, 7). The indentation at first shows a parabola-like trend followed by a quasi-vertical slope, at higher applied forces. The first part can be explained with the classical indentation theory of a solid punch into a half space, progressively deviating from such behavior as the glass substrate contribution becomes dominant (10). The quasi-vertical trend with increasing force indicates that the maximum compression of the cell material has been reached, and the indentation limit value attained will give an indication of the cell thickness (7). Let us now compare curves 3 and 4: curve 3 was taken onto an apparent depression of the surface, at a point like the one identified by A in Fig. 2, while curve 4 was recorded onto a stiff portion of the cell surface like the one identified by B. It is evident how in the

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope

251

Fig. 4. (a) A small whole living neuron showing arborizations, image acquired using the error channel signal. The apical end of most of them seem to be disrupted. (b) Higher magnification error signal image of the apical end of an arborization. (c) Error signal and z-piezo signal simultaneous image of the same arborization apical end. The trace is made of small (approximately 150  nm in diameter) dot-like structures that could be identified as clusters of adhesion molecules.

case of curve 3 we are progressively indenting a thick portion of the cell until we reach the glass substrate, while in the second case, after a parabolic behavior for the first 100 nm of indentation, we reach a constant slope corresponding to an elastic spring constant of 0.0045 N/m. This means that after indenting the most external “soft” cell surface, the tip interacts with submembrane structures that exhibit an elastic response (4). At the force value of 0.5 nN that corresponds to the nominal set

252

Ricci, Grattarola, and Tedesco

Fig. 5. Series of representative force-versus-indentation curves acquired upon a growth cone of a living neuron. Curves 1–3 are taken moving away from the growth cone edge toward the neurite at steps of 3 mm. Each curve shows an increase in indentation with the applied force with a parabola-like behavior, until a quasi-vertical trend is reached. Curve 4 was taken next to the point where curve 3 was acquired, but on a protrusion. A totally different trend can be observed, where after an initial parabolic indentation, a linear dependence on the force increase is established.

point used during imaging in the case of curve 3, we have an indentation of 950 nm and for curve 4, 200 nm. We need to keep in mind these features and figures in order to understand the contrast mechanism of both the z-piezo signal and error signal images on such specimens. This means that the deep shallows next to high peaks found in the z-piezo image cannot purely be due to morphological features. In fact, on thick and “soft” locations, indentation can reach the micron range and on stiffer ones, it lowers down to the 100-nm range; the shallows and peaks in the “topographical” image must be essentially due to differences in stiffness of the sub-membrane growth cone structure encountered by the cantilever during scanning. A pictorial representation of the possible “real” versus measured profile on the growth cone is shown in the second half of Fig. 2. Similar effects are found in the error images, where the gray scale levels represent the deviations from the feedback force (cantilever deflection) set point. As the feedback loop has been tuned to give a high contrast in the error image, thus allowing temporary and relatively large deviations from the force set point, on the left hand side (scanning is from left to right) of an upward slope or an increase in stiffness, the pixels will be darker, while on the left hand side, it will be lighter. This can be clearly observed in Fig. 1a. Generally speaking, it is not possible to discriminate between a change in topography and a variation in stifness, unless one has independent knowledge of the properties of the surface. By use of force-versus-indentation curves, it is possible to discriminate the effects and estimate the thickness of the undeformed surface, at least at the point where the curve is taken. Extrapolation to similar areas can be made to have an estimate of numerical values.

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope

253

A feature common to Figs. 1 and 3 is a thick tubular region extending toward the neurite. The thickness of this region as read on the z-piezo image is in the order of 1 mm, to which at least 200 nm must be added to take the indentation into account. 3.4.3. Identification of Growth Cone Regions 3.4.3.1. Lamellipodia

The filaments in evidence in Fig. 1a, d can be easily identified as microtubules. By comparing these with images generated by the other techniques and described in the literature, the thick region can be easily identified as the so-called C-domain. Note that submicrometer-size structures are visible in this region. In Figs. 1–3, this domain is surrounded by a flat area, with thickness in the order of a few hundred nanometers. This could be identified as a lamellipodia-rich P-domain. The flat protrusions shown in Fig. 3a, b can be identified as lamellipodia structures. Their thickness is in the order of 30–60 nm. Irregularities, distortion along the scanning direction, and “islands” of biological material underline the extent of the tip–sample interaction. For comparison, Fig. 3c shows a similar growth cone after fixation. Lamellipodia structures with a smooth profile are now evident. The thickness is now in the 100–200-nm range: the fixation process has affected the membrane stiffness so that only negligible indentation occurs during scanning.

3.4.3.2. Filopodia

The spiny protrusions surrounding the C-domain in Fig. 3d can be identified as filopodia structures. Interestingly enough, these protrusions appear to be made of globular subunits, often arranged in a discontinuous way. These subunits have a diameter of 120– 180 nm and a thickness ranging from 5 nm to 30 nm. They could be identified as clusters of proteins or patches of membrane adhering to the substrate, left after the tip–sample interaction. The distribution of proteins in the filopodia of growth cones is a subject of active research. Filopodia are known to be filled with bundles of actin filaments (11), and the presence of spots of tyrosinephosphorylated proteins has been recently demonstrated at the tips of growth cones by immunofluorescence techniques (12). The formation of focal contacts by the tip of filopodia with the substrate remains an open question, which could be addressed by further investigation of the described structures.

3.4.3.3. Arborizations

Finally, traces of discontinuous biological material are evident in the terminal regions of the arborizations of a whole neuron (Fig. 4). A fixed similar neuron is shown for comparison (Fig. 6). Here, the arborizations are smooth and continuous. The morphology of the biological details is better preserved, but no indications about the adhesion to the substrate can be inferred. On the contrary, we can conclude that the tip–living material interaction does somehow affect the morphology but, at the same time, it gives hints about the organization of the biological structure at the nanometer scale and about the way contact is made with the substrate.

254

Ricci, Grattarola, and Tedesco

Fig. 6. A small whole fixated neuron showing arborization, to compare with the living one imaged in Fig. 4.

3.5. Comparison with Other Techniques

Detailed AFM images of living flat cells, such as glia cells (5, 6), fibroblasts (4, 7) and epithelial cells (13) have been already analyzed in the literature. Low resolution images of whole neurons have also been produced (6). The other available techniques for studying growth cones are as follows: 1. Whole-mount electron microscopy that gives images with detailed information down to the nanometer scale (14), but on dead materials and without thickness quantification. 2. Fluorescence microscopy that allows one to identify cytoskeleton components by immunofluorescence staining. Timelapse analysis of stained (lipid probe Dioc6) living growth cones is also described (15). 3. Video-enhanced DIC imaging that is widely used for generating detailed images of unstained living growth cones. This technique has allowed the identification of two distinct domains: a central, relatively thick, organelle-rich region (C-domain) and a peripheral, thin region devoid of organelles (P-domain) (16, 17). AFM shares with the last technique the capability of imaging unstained samples. Moreover, compared to all the mentioned methods, it is the only one to have the potentiality of giving quantitative information on thickness. On the contrary, to put the last statement in the correct perspective, it should be underlined that any AFM-originated image of a soft sample is the result of a mechanical interaction between the tip and the sample. This implies

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope

255

indentation, and force-versus-distance curves must be utilized to correct the data to obtain “real” thickness. At the same time, a challenge to the adhesion of the living cone to the substrate is exerted. Finally, it is worth mentioning that, in principle, information about viscosity could be also obtained, similar to that in the laser tweezers technique (18), by carefully comparing the forward and retraction portions of the force-versus-distance curves.

4. Notes 1. Cut glass slides are used as the AFM employed for these investigations is a “scanned sample” model that can accommodate only flat samples having a maximum width of 25 mm (width) and a length of about 50  mm. When using “scanned tip” instruments, it is possible to use the AFM directly onto the Petri dish. One advantage of using glass slides is the surface flatness with respect to Petri dishes. 2. Even if it is possible to grow cultures for some cell lines on bare substrates, for AFM investigation, a good adhesion with the substrate is essential, allowing the cells to withstand the lateral forces induced by the tip during scanning. 3. An advantage of using small glass slides is the possibility of preparing several samples in the Petri dish at one time that can be kept in the incubator until just before use. This allows to increase the throughput of one single primary cell line culture both in time and number of observable cells, as usually it is not possible to keep temperature and CO2 control during AFM measurements. Recently research groups have been developing systems that allow to control physiological environmental conditions during AFM imaging. 4. An important feature when using the AFM on living cells is the range available for the Z direction in the scanner, as variations in height of several microns can be found during scanning. Also X and Y range should be several tens of microns. 5. The use of small spring constants avoids damages to the cell surface or even detachment of the cell from the substrate. We have also tried using 0.003 N/m spring constant cantilevers, but the adhesion forces between the tip and sample did not allow achieving good imaging. 6. Standard silicon nitride tips are quite sufficient for good imaging in contact mode on living cells as the visco-elasticity of the cell is the main limiting factor in resolution: in fact, a sharper tip does not improve resolution but can produce damage. Moreover, silicon nitride seems to behave better than silicon or polysilicon with respect to tip-to-sample adhesion.

256

Ricci, Grattarola, and Tedesco

7. A fresh supply of the buffer solution used for the culture, possibly held at 37°C, should be kept at hand, in order to able to add it when required. 8. An on-axis microscope is essential for positioning the tip on the sample in the required position. Either a scanned-sample AFM equipped with a high magnification long focal length microscope, or a scanned tip AFM mounted onto an inverted microscope can be used. The latter obviously opens the possibility to use a wider range of optical microscopy techniques in conjunction with AFM. 9. One of the serious limitations of AFM is the low scanning speeds that have to be used on soft surfaces. A 10 mm × 10 mm image with 512 points per line is typically acquired in 8 min. 10. On the thinner portions of the cells, one observes at first the indentation process and then a constant relationship between load and Z-piezo travel is found. This means that all the cells have been compressed and the glass surface is “reached.” One can use this last linear portion of the force-versus-distance curve to derive the coefficient and obtain the corresponding force-versus-indentation curve. References 1. Binning, G., Quate C. F., and Gerber C. (1986) Atomic force microscope. Phys. Rev. Letters. 56, 930–933. 2. Schoenenberger C.A., and Hoh J. H. (1994) Slow cellular dynamics in MDCK and R5 cells monitored by time-lapse atomic force microscopy. Biophys. J. 67, 929–936. 3. Hoh, J. H., and Schoenenberger C.-A. (1994) Surface morphology and mechanical properties of MDCK monolayers by atomic force microscopy. J. of Cell Science. 107, 1105–1114. 4. Ricci, D., Tedesco M., and Grattarola M. (1997) Mechanical and morphological properties of living 3T6 cells probed via scanning force microscopy. Micros. Res. Tech. 36(3), 165–171. 5. Henderson, E., Haydon P. G., and Sakaguchi D. S. 1992. Actin filaments dynamics in living glial cells imaged by atomic force microscopy. Science. 257, 1944–1946. 6. Parpura, V., Haydon P., and Henderson E. (1993) Three - dimensional imaging of living neurons and glia with the atomic force microscope. J. of Cell Science. 104, 427–432. 7. Ricci, D., and Grattarola M. (1994) Scanning force microscopy on live cultured cells: Imaging and force-versus-distance investigations. J. of Microsc. 176, 254–261. 8. Butt, H.-J., Siedle P., Seifert K., Fendler K., Seeger T., Bamberg E., Weisenhorn A.L.,

Goldie K., and Engel A. (1993) Scan speed limit in atomic force microscopy. J. of Microsc. 169, 75–84. 9. Putman, C.A.J., van der Werf. K.O., de Grooth B.G., van Hulst N.F., Greve J., and Hansma P.K. (1992) A new imaging mode in Atomic Force Microscopy based on the error signal. Proc. SPIE. 1639, 198–204. 10. Sneddon, J.N. (1965) The relation between load and penetration in the axisymmetric Boussinesq problem for a punch of arbitrary profile. Int. J. Eng. Sci. 3, 47–57. 11. Lewis, A.K., and Bridgam P.C. (1992) Nerve growth cone lamellipodia contain two populations of actin filaments that differ in organization and polarity. J. Cell. Biol. 119, 1219–1243. 12. Da-Yu Wu and Golberg D.J. (1993) Regulated tyrosine phosphorylation at the tips of growth cone filopodia. The J. of Cell Biology. 123, 653–664. 13. Hoh, J.H., Sosinsky G.E., Revel J.-P., and Hansma P.K. (1993) Structure of the extracellular surface of the gap junction by atomic force microscopy. Biophys. J. 65, 149–163. 14. Bridgman, P.C., and Dailey M.E. (1989) The organization of myosin and actin in rapid frozen nerve growth cones. J. Cell. Biol. 108, 95–109.

The Growth Cones of Living Neurons Probed by the Atomic Force Microscope 15. Bridgman, P.C. (1991) Functional anatomy of the growth cone in relation to its role in locomotion and neurite assembly. In The nerve growth cone. Letourneau P.C., Kater S.B., and Macagno E.R., editors. Raven Press, New York. 39–53. 16. Gordon-Weeks, P.R., and Mansfield G.S. (1991) Assembly of microtubules in growth cones: the role of microtubule-associated proteins. In The nerve growth cone. Letourneau P.C., Kater S.B., and Macagno E.R., editors. Raven Press, New York. 55–64.

257

17. Goldberg, D.J., Burmeister D.W., and Rivas R.J. (1991) Video microscopic analysis of events in the growth cone underlying axon growth and the regulation of these events by substrate-bound proteins. In The nerve growth cone. Letourneau P.C., Kater S.B., and Macagno E.R., editors. Raven Press, New York. 79–95. 18. Dai, J., and Sheetz M.P. (1995) Mechanical properties of neuronal growth cone membranes studied by tether formation with laser optical tweezers. Biophys. J. 68, 988–996.

wwwwwwwwwwwwww

Chapter 17 Highlights on Ultrastructural Pathology of Human Sperm Narahari V. Joshi, Ibis Cruz, and Jesus A. Osuna Abstract Applications of atomic force microscopy to ultrastructural investigation of human spermatozoa are ­discussed, with particular emphasis to their most common pathological alterations, which are recognized to be associated with male infertility. Morphological alterations can be located in the head, neck piece, and/or in the flagellum. The consequences of these defects on infertility-related topics are examined in the light of aberrations caused in varicocele and in other spermatozoa morphological alterations like globozoospermia, oligoasthenospermia, and in semen from patients with HIV syndrome. Special attention is given to the temperature effects on sperm abnormalities. The application of the present approach to pharmacology, namely, the development of male contraceptive methods, is also referred. Key words: Atomic force microscope, Human sperm, Infertility, Pathological alterations

1. Introduction A conventional technique to investigate ultrastructure is ­transmission electron microscopy (TEM) (1) which permits to examine the materials, biological as well as nonbiological, with a resolution of the order of a few nanometers, even though some instruments have been recently developed with increased power of resolution pushing the limit further up to a fraction of nanometer. This widely used practice undergoes several serious limitations, namely, that TEM cannot be used for living biological systems. Moreover, in the sample preparation and the measurement, processes such as fixating, vacuum creation, staining, osmotic damage, and other deteriorating factors to the biological material are present. In a way, the measurement process itself seriously affects the pathological aspects in which we are mainly interested. Special methods, such as the use of betaine (2), is recommended to preserve spermatozoa for ultrastructural morphology, Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_17, © Springer Science+Business Media, LLC 2011

259

260

Joshi, Cruz, and Osuna

however, its effectiveness is very limited. Applications of artifacts, which are indispensable, create additional defects and also damage or alter the surface completely or partially. These types of defects and irregularities are not only unavoidable but they change, up to a certain extent, the morphology and create doubts about the authenticity of the original data. In addition to these aspects, TEM is not able to provide topographical information which is really needed to examine the pathological aspects of living cells. Besides, TEM cannot provide three-dimensional images at ease and with precision, as measurements are not carried out by a scanning procedure and hence there is a lack of topographical information. These difficulties are overcome by a recently developed technique, namely, atomic force microscopy (AFM) (3, 4). AFM is a scanning probe microscope which consists of a microscale cantilever with a sharp tip (probe) at its end, and it is used to scan a specimen surface with high resolution, more than 1,000× the limit of an optical microscope. The spatial resolution of AFM depends upon the dimensions of the tip, and generally it has atomic resolution of the order of a few nanometer. Recent technology permits to fabricate the tip as small as 1 nm radius, and hence a slight improvement in spatial resolution is observed. As it is a scanning probe microscope, it provides information with precision in three dimensions. Obviously, it is a breakthrough in nanotechnology, and it has an extensive and sizable impact in biological and medical sciences. In biological systems, one encounters very flexible micromolecules and conventional AFM fails to yield useful images or topographical information. For this purpose, Cryo-AFM has been designed which permits to acquire images in liquid nitrogen vapor, and hence it is possible to control the temperature from 77 to 220 K. Cry-AFM is found to be very suitable for bio materials like immunoglobulins, examination of DNA, red blood cells (4), and other biological materials. Recently, temperature control unit is incorporated to AFM (5). Such achievements guarantee that new and useful information is available to examine the temperature dependence on defect formation in human sperm. This issue is addressed in the coming section. On the other hand, measurements at low temperature are not required for several biological entities, where ultrastructural images and topographical information at room temperature are essential for understanding pathological alterations in a living system at nanoscale level. With this view, an AFM is applied to Entamoeba histolytica (6, 7) in their natural environment and useful structural information has been reported. Andrology is not an exception and the applications of AFM to ultrastructural (in the range of nanometer) features and analysis of human sperm, both normal and pathological, have become a reality (8–12). In order to look at alterations caused in human sperm by a disorder of any kind, it is necessary to have complete and reliable

Highlights on Ultrastructural Pathology of Human Sperm

261

information of healthy human sperm in its physiological environment­, and a detailed topographical data have been reported earlier by Joshi et al. (8). This has been used to compare the alterations caused in the pathology of human sperm and to evaluate alterations caused by different deleterious conditions. Morphological alterations in human sperm are directly related with infertility, which is the inability of a couple to conceive after 1 year of regular unprotected intercourse. The prevalence of infertility has been estimated in about 15% in all couples of reproductive age (13). Many couples that come to an infertility clinic do not present apparently the causes related to infertility, and disturbances in fertility may remain occult for years, until a couple are interested in having a child. It is estimated that about 40–50% of the infertility cases is attributed to “male factor” (14–16). According to etiology, male infertility has been categorized in three groups: pretesticular causes, testicular causes (spermatogenesis factors), and post-testicular causes (17). Some aspects of infertility (treatable and untreatable) are certainly related with spermatozoon morphology. Normal testicular function is expressed in two different ways: sex hormone synthesis and spermatozoa production. Sperma­ togenesis is a very complex process in which different ana­ tomic and functional structures are involved: the central nervous system (the hypothalamo-pituitary axis) and, at the periphery, the testes (17). Spermatogenesis disruption is reflected in several pathological conditions. Primary testicular alterations are the most common causes of spermatogenesis absence or derangement, in which a variety of factors are involved: mainly genetic causes and chromosome disorders. Scrotal temperature is a crucial factor for testicular heat regulation, maintaining its temperature in the physiological range required for normal spermatogenesis. High testicular temperatures have been related to male infertility, probably altering sperm morphology (18). Some of these sperm morphological changes can be examined by TEM or by AFM, and could be an important source of information for the study of male infertility.

2. Experimental 2.1. Sperm Preparation

The initial evaluation of a man from an infertile couple should include a thorough clinical evaluation, with particular emphasis in his reproductive history, and at the same time semen analysis should be performed. It is a generally accepted criterion that semen analysis is the cornerstone of the laboratory evaluation of the infertile male; it provides information on semen volume as well as sperm concentration, motility, and morphology. Subjects from

262

Joshi, Cruz, and Osuna

our studies received standard written instructions for semen ­collection. Semen was obtained by masturbation, after 3–5 days of sexual abstinence, and the sample was delivered to the laboratory within 1 h of collection, avoiding exposure to extremes temperatures. Semen analysis was performed according to the manual published by the World Health Organization (WHO) (19). Samples were prepared for routine semen analysis, including estimation of sperm concentration, assessment of sperm morphology, hypoosmotic swelling test, and leukocytes quantification by the peroxidase technique. Spermatozoa separation was carried out by the “swim-up” technique. The conventional “culture medium” used in our experiments was Hanks 199 obtained from GIBCO laboratory, USA. It is observed that spermatozoa were maintained in healthy conditions (or without alteration) for 6–8  h, time enough to carry out AFM measurements. In some special cases, like sperm from HIV patients, a written permission was obtained and confidentiality about the identity was guaranteed. Selected sperms are placed on conventional microscopic slides. AFM images are recorded in normal atmosphere (not in vacuum) and saline environments without damaging the cell membrane. The present technique, without doubt, preserves the small cellular structure and also retains cytoplasmatic structure which is generally damaged in TEM measurement procedure. 2.2. Examination by Optical Microscopy

Sperms were examined by conventional optical microscopy (400×) for gross defects, such as bending, coiling, and relative length of flagellum. Very often, the form and the shape of the head can be viewed. Therefore, a preliminary study is generally carried out with an optical microscope. It is possible to detect some likely abnormalities by optical microscopy, even though the quantification can be carried out only by AFM.

2.3. AFM Operations and Precautions in the Measurement Techniques

AFM is a high-resolution instrument which permits to obtain images and examine ultrastructural details with precision in three dimensions in living and nonliving specimens. However, certain precautions are required to obtain good and reliable results: 1. Nano structure grown and/or grooved on semiconductor wafers like Si or GaAs is a good example of nonliving specimen where three-dimensional ultrastructure is found to be very useful. In this case, the sample is hard enough and the use of contact mode is very functional. However, in case of biological objects (living or nonliving) the use of contact mode is totally inadequate as the outer membrane is delicate and soft. Moreover, dragging the tip on the sample just damages the membrane and the sample. Therefore, the most important recommendation for non biological objects is measurements in contact mode are just not possible and its use should be avoided completely.

Highlights on Ultrastructural Pathology of Human Sperm

263

2. The scanning rate should be adjusted to match the software and hardware of a given equipment. It is always recommended to try a couple of scanning rates before the final imaging. In most cases, 1 line/s was found to be an adequate rate and the images reveal all the details. The speed higher than 3 lines/s is not recommended, unless the software is specially designed for this purpose. 3. In biological samples, noncontact mode, commonly known as tapping mode, should be essentially employed. In this mode, three parameters for measurements procedure should be considered. The first is the stiffness of the cantilever. If the cantilever is not stiff enough, it will stick to the water molecules or liquid environment. The recommended value is about 1 nN/m. A value lower than this might be unsuitable for reliable measurements if the environment contains lipids or sticky solutions as it is in the present case. 4. The magnitude of the applied force is a critical parameter and it approximately varies from 0.2 to 0.8  nN/m depending upon the type of membrane and its elasticity. For example, for E. histolytica (6, 7) the optimum force is 0.8 nN/m, meanwhile for a healthy spermatozoon it is 0.2  nN/m. A slight variation is expected from one sperm to another and it is generally determined by experiment. It is worth to mention that when the sperm is deteriorated (e.g., incubated at 40°C), a lower force is recommended. 5. The resonating frequency of the vibrating tip is also a key parameter for high-resolution imaging, particularly when the sample is soft and the structure below the membrane is not compact. This is the present situation, particularly near the neck pieces. The vibrating frequency used for the spermatozoon lies between 170 and 190 kHz. For soft parts, sometimes, a separate scanning is required with lower frequency. High-resolution good images can be obtained with the help of a delicate balance between the force and the vibrating frequency (8). 2.4. Data Collection System and Imaging

Three-dimensional images were obtained by image processing software which was provided with the equipment. These images offer information about alterations caused in spermatozoon’s morphology of a specific disorder, and could help to correlate the cause for infertility; however, they do not permit to quantify the alterations, and the only way is to make them available with the help of longitudinal and transverse profiles of their structures (e.g., the head, the neck, or the tail). Therefore, we have supplied these profiles for healthy and pathological sperms. A typical spermatozoon is a stripped-down cell, which has a strong flagellum to propel it through the seminal fluid (see Fig. 1). The mature spermatozoon is devoid of cytoplasmic organelles

264

Joshi, Cruz, and Osuna

Fig. 1. Image obtained by AFM of a healthy spermatozoon. The height scale shown on the right hand side provides only a rough idea and hence the topography is a must. Reproduced from Archives of Andrology, Joshi et al. (8).

such as ribosomes, endoplasma reticulum, and Golgi apparatus. Meanwhile, it contains numerous mitochondria strategically located, particularly where they are needed to efficiently propel the flagellum (20). Spermatozoa consist of two functional distinct regions: the head, and the tail or flagellum. The neck region is considered as the beginning of the tail. However, morphologically the neck has distinct features, and therefore in the present work, we have classified it into three different sections, namely, the head, neck piece, and the flagellum. Each part has specific functions and pathological alterations or abnormalities, and each of them has a definite modified response. For example, if the tail has some defect, its motility will have an irregular response. We, therefore, believe that it is more useful if pathological irregularities are presented by subsection head, neck piece, and tail region.

3. Sperm Anatomy 3.1. Head

The head of the spermatozoon is occupied mostly by the nucleus and the acrosome, with lesser amounts of cytoskeletal structures and cytoplasm. The nucleus and the acrosome are usually symmetrical structures. Human spermatozoa are uniform as far as the size and the shape properties are concerned. Variations in them may be less than 5%. Therefore, it is just appropriate to investigate morphological details of healthy sperms and examine the alterations caused in them. Deviation, if it is observed, is due to external factors, which may change their morphology, as in some diseases like varicocele. The temperature at which the sperms are formed and maintained is also important. The spermatozoa nuclei are haploid, containing only one member of each chromosome pair. The chromatin, which is the male DNA contribution in the fertilization process, becomes highly condensed at the latter stages

Highlights on Ultrastructural Pathology of Human Sperm

265

of the spermatogenesis; at these stages, the histones, high in lysine content compared to arginine, are replaced by protamines, highly basic proteins rich in arginine and cysteine. The m-RNA encoding for mouse protamines are synthesized in spermatids, indicating that protamines are products of the haploid genome. The highly condensed protamine–DNA complex is stabilized by disulfide bonds between the protamines (21, 22). Therefore, the major nuclear proteins associated with sperm DNA are protamines. An increase in the height of the sperm indicates that chromatin is not in a compact form, and protamines and arginine are not in the required balance proportion. Defects in the region of head or neck pieces are very common, and they lack of specificity for any pathological condition of the testes. They are usually found in patients with severe low sperm counts (oligozoospermia), but also in patients with normal sperm count. The causes for oligozoospermia remain elusive, and they are the expression of damage of germinal epithelium. Defects of the head and neck pieces are included in the teratozoospermia category (23). Systemic illness, chronic fever, frequent exposure to high temperatures, cytotoxic chemotherapy and radiation therapy, drugs and toxins, and a large list of other factors may be related to aberrations in the spermatozoa head and neck pieces. For an evaluation of any kind of anomalous morphological structure, it is necessary to know the standard features. We are, therefore, presenting here an image with transverse and longitudinal topographical details of healthy spermatozoon in its natural environment. The pathological features are compared with these details. Figure  1 shows the image of a healthy spermatozoon. Figure 2a–c are head image, the longitudinal, and the transversal topography, respectively. A table of the magnitudes of lengths and heights appear at the bottom of each figure (Fig.  2b, c). These values are given with a precision of 0.1 nm. This is a measurement accuracy of the system used. However, as mentioned earlier, these values might vary by about 5% from person to person, and sperm to sperm. This aspect should be considered while comparing normal values with the pathological ones. The height measurements, sometimes, are provided with the help of a histogram, however, they are not suitable to quantify and hence their use is avoided in the present work. Figure 2b reveals that the longitudinal topography discloses some structural information like acrosome cap and a well-marked boundary between head and neck which could be an important feature. The length of the acrosome cap is about 1,880  nm and approximately the same value is reported by Kumar et al. (24, 25). The transverse profile, in this case, does not throw much light on the structural details and generally its form is symmetric if measured perpendicular to the length. For a pathological study, the longitudinal profile is

266

Joshi, Cruz, and Osuna

Fig.  2. (a) shows the image of the head of the spermatozoon. (b) and (c) show longitudinal and transverse profiles. Reproduced from Archives of Andrology, Joshi et al. (8).

more significant for the analysis of the head, and we take it into account for further discussion. A morphological examination of the head necessarily should focus the attention on the acrosomal region. It is a large secretory granule or vesicle that closely surrounds and overlies the anterior end of the nucleus. The acrosome originates from Golgi complex in the spermatids and contains hydrolytic enzymes necessary for the sperm to fuse and penetrate the egg’s outer coat to achieve fertilization. When a sperm comes in contact with an egg, the contents of the acrosome vesicle are released by exocytosis in the so-called acrosomal reaction (21, 23). This is a critical moment in the fertilization process. Specific proteins are also released which are essential to a normal sperm–ovum interaction. This membranous structure sits as a cap over the nucleus in the anterior region of the sperm head. The morphology of this cover cap and the alterations caused on it can be viewed with AMF (see Fig. 2b), and this is one of our concerns in the present investigation as it is directly related with infertility in male.

Highlights on Ultrastructural Pathology of Human Sperm

267

The other important component of the head of the sperm is chromatin which can also be viewed with the help of topographical configurations, as shown in Fig. 2b, c. These figures permit to calculate the volume of the desired area. The height of the head of a healthy sperm is about 750 nm, and Fig. 2b reveals a clear distinction between the acrosome and the chromatin regions. By this or by using an equivalent method, it is possible to measure the volume of the desired area. Sperm chromatin contains DNA material and it is located at the central part of the head of the sperm. Considering its importance, Lee et al. (26) have investigated chromatin volumes in human sperm nuclei for seven of nine classes of head shape abnormalities. On the basis of experimental data, it is concluded that the nuclear volume is essentially identical even though the shape and the projected area have substantially different forms. This study indicates that different head forms in fertile men are not originated from sperm chromatin or DNA content of the sperm nucleus, and the way in which chromatin is organized. Such information is valuable to compare the results obtained by TEM and with other observational tools. 3.1.1. Varicocele

Varicocele is the most commonly identified and correctable cause of male infertility. It is about three times more common in infertile men than in men of proven fertility. Varicocele is found predominantly on the left scrotal sac. The seminal profile seen in infertile men with varicocele was first described by MacLeod in 1965 (27). Oligozoospermia of varying degrees was observed, but the striking findings were marked impairment of sperm motility as well as a noticeable increased of immature and tapered forms in the seminal fluid (stress pattern). Nevertheless, the semen quality in men with varicocele varies from normal to azoospermia. A study from the WHO showed on 9.043 men that the incidence of varicocele was 25% on men with abnormal semen, and 11.7% on men with normal semen (23). Zucchi et al. (28) found that 18 out of 43 patients had oligoasthenospermia, 20 had asthenospermia, and 5 had normal values. Several theories have been postulated for the effect of varicocele on the testicular function, including vascular stasis, back pressure, reflux of renal or adrenal substances into the pampiniform plexus, and interference with the heat exchange mechanism of the pampiniform plexus; nevertheless, the precise mechanisms have not been elucidated (29). Recently, Mieusset et al. (30) studying a group of 13 fertile men have reached to the conclusion that clothing and seated with legs crossed position has a persisting thermogenic effect, stressing the fact that the scrotum regulatory system plays a fundamental role to keep the testes at a temperature lower than the rest of the body. The temperature effect seems to be very critical, and for this purpose a separate investigation was carried out to examine the effect of temperature on sperm morphology and it is discussed separately.

268

Joshi, Cruz, and Osuna

The above discussion confirms that varicocele is one of the predominant causes for male infertility and it has several implications, since it causes different types of abnormalities not only in head, but also in the neck piece and flagellum region. Therefore, we have given importance to the morphological investigation of sperm of a patient suffering from varicocele. By using AFM technique, we have demonstrated earlier that head abnormalities were predominant in sperm obtained from varicocele patients (9); probably, the difference is not remarkable between varicocele grade I and grade III. (11). In semen from patients with varicocele, the morphological and topographical variations in the head form of the spermatozoon can be appreciated with the help of Fig. 3. In normal sperm in the longitudinal topography, the acrosome cap is clearly visible (see Fig. 2b), and it is about 1,800 nm length while it is altered for sperm obtained from varicocele patients grade I which is about 1,450  nm (not well marked), and it is nearly absent for the sperms obtained from varicocele grade III. This is a clear indication that some spermatozoa of varicocele grade III have practically no possibility in the fertilizing process. The variation in height is also noticeable. In a normal cell it is about 750 nm; meanwhile, in varicocele patients (both grade I and III) the height is above 1,050 nm. In case of varicocele grade I, the increase in height is abrupt; meanwhile, in varicocele III it is rather continuous. This also indicates that the damage is not localized but spread over the entire region. An increase in height indicates that chromatin (or DNA) material is not sufficiently condensed or compact, and it may affect the fertilizing process. There is no significant variation in length and it is

Fig. 3. Images of the head along with the longitudinal and the transverse profiles of the sperm of a patient with varicocele I and varicocele III. Reproduced from Archives of Andrology, Joshi et al. (11).

Highlights on Ultrastructural Pathology of Human Sperm

269

of the order of 5,000  nm for varicocele grade I. In case of ­varicocele grade III, the length is difficult to determine because the depression corresponding to the neck is not detected, and hence it is not possible to pinpoint the exact location where the neck region starts. In this case, much information is obtained from the longitudinal topography. 3.1.2. Oligoasthenoter­ato­ zoospermia

Oligoasthenoteratozoospermia (OAT) signifies the alterations of three variables of the seminal fluid; characterized by abnormal low sperm concentration, low motility, and also abnormal sperm morphology. Different factors may be responsible for the OAT condition, including hypothalamic pituitary dysfunction, causing low luteinizing hormone (LH), follicle-stimulating hormone (FSH) syntheses, and inadequate stimulation of spermatogenesis; in cases of Leydig cell and germinal cell dysfunction, as may be the case of varicocele; and in the focal Sertoly-Cell-Only (SCO) syndrome (31). Because of the above reasons, understanding the morphological defects in sperms of OAT patients is very important. The morphological details of the sperm obtained from OAT patients were examined carefully as alterations are expected for the reasons mentioned above (see Fig. 4). The general trend is similar to that obtained from varicocele patients. The height of the head is increased significantly up to approximately 1,425 nm instead of about 750 nm. Meanwhile the length is not altered, it has the value approximately 5,000–5,200 nm (in this case also the beginning of the neck is not located and hence the value is not precise). Therefore, the increase in height appears more noticeable. In globozoospermic sperm, which is a severe form of teratozoospermia, probably genetic nature is characterized by round nuclei and the absence of acrosome, as a consequence the Golgi apparatus does not evolve into the acrosome (32). In globozoospermic spermatozoon, the situation is a little different (9) as compared with OAT. The height of the sperm lies between two extreme values, and it is about 1,300 nm, but the common factor is that the layer of acrosome is absent in both types. In this case also, the neck structure is not visible, which is expected to be located at 5,200 nm, but the most curious aspects is that the variation in the surface is smooth and continuous up to approximately 7,000 nm. This means that in the region of the head, neck piece, and the flagellum are merged without showing any structural variation, indicating that there are serious alterations in the sperm of patients with globozoospermia.

3.1.3. HIV Syndrome

Defects in the region of the head or the neck pieces are very common, and they lack specificity for any pathological condition of the testes. They are usually found not only in patients with severe oligozoospermia, but also in patients with normal sperm count.

270

Joshi, Cruz, and Osuna

Fig. 4. The image and the topography of the sperm of a patient having OAT. Reproduced from Fertility and Sterility, Joshi et al. (9).

Systemic illness, fever, and other factors previously mentioned may be related to changes in the spermatozoa head and neck pieces. AFM is an adequate research tool which may contribute to a precise knowledge of teratozoospermia and its implications in male infertility. Defects in the region of head or neck pieces are very common in sperms from other conditions, for example patients affected by HIV and diseases characterized by high and prolonged fever. In case of HIV patients, the situation is more interesting and complex as some of the patients are going through antiretroviral therapy (HAART) treatment for survival, and therefore AFM investigation had been carried out in both cases, patients with and without HAART treatment, and the results are very valuable (see Figs. 5 and 6).

Highlights on Ultrastructural Pathology of Human Sperm

271

Fig. 5. The image and the topography of the head of the sperm of a patient obtained from HIV patient. Reproduced from Archives of Andrology, Barboza et al. (12).

In the case of HIV patients, the height of the head of the sperm is slightly reduced (approximately 600 nm) but the length of the head is increased. The acrosome layer is not well marked or it might be very thin. After going through HAART treatment, these values try to normalize. The height of the spermatozoon approaches to the normal value. However, in this case the evaluation and conclusions are more difficult because the situation varies from patient to patient as they are undergoing different combination of drugs with different doses. The observation reveals only a tendency. Here, it is worth to mention that AFM does show the presence of HI-virus on the head and neck regions (12). This is the direct evidence of the active role of the spermatozoon in HI-virus transportation process. 3.1.4. Temperature Effects

From the aspects discussed in varicocele patients, it is clear that the thermogenic effect is very significant, not only from the point of view of sperm concentration, but also for the temperature at which the sperm is formed and maintained. High temperatures affect motility as well as sperm morphology. Healthiness of the

272

Joshi, Cruz, and Osuna

Fig. 6. The image and the topography of the head of the sperm of a HIV patient who is going through HAART. Reproduced from Archives of Andrology, Barboza et al. (12).

sperm is very sensitive to temperature. Moreover, from the ­clinical practice, it is well known that patients suffering from fever for a long time (chronic high temperature for days or weeks) show problems related with infertility, even though they are reversible. It is also known that the effect of prolonged sauna exposure on the testes is negative (33). In this case, the environment of the testes is inadequate, and hence spermatozoa are expected to have modification and malfunctioning (34). With the above consideration, a morphological investigation has been carried out by incubating sperms at 36.5, 38, and 40°C for 24 h (unpublished work), and the results are shown in Figs. 7 and 8. Obviously, the sperm morphology at 36.5°C is normal and coincides with the image and topography as given in Figs. 1 and 2. The difference is visible for the sperm incubated at 38°C, however, it is not drastic. The acrosome layer is visible but its form is slightly

Highlights on Ultrastructural Pathology of Human Sperm

273

Fig. 7. The image and the longitudanl and transeverse the topography of the head and the neck of a sperm incubated at 38°C.

distorted (Fig. 7). The height of the head does not really change but the structure near the neck pieces acquires an imperfect form. This means that DNA is still in a compact form. However, the sperm incubated at 40°C shows dramatic alterations (Fig. 8). The height of the head is increased from the expected value 750– 1,150 nm. The surface is rough and the head has a bulky form. In this case, DNA is not in a compact form, and the bulkiness of the head may cause impediments in the swimming process also. These features, namely, irregularities in the head form, increase in the height, and the absence of the acrosome structure, have been used positively by pharmacologists to control and examine the quality of a given drug toward the male contraceptive (24, 25). By injecting a polyelectrolytic compound (a special drug) in a

274

Joshi, Cruz, and Osuna

Fig. 8. The image and the topography of the head and neck of a sperm incubated at 40°C.

healthy person, the defect in the sperms and the ­alterations caused due to the drug are studied with the help of AFM, and its effectiveness toward a male contraceptive is obtained. The faulty head form revealed by the topography together with the absence of the acrosome layer is the fundamental feature of the pathology of the head of the human sperm. The abrupt and asymmetrical increase in height, which is also a common feature in pathological sperm, has also a remarkable negative effect on the floating and the swimming capacity of the sperm due to the principle of buoyancy, and this becomes as an additional ­impediment in the fertility process.

Highlights on Ultrastructural Pathology of Human Sperm

275

In this study, we have given special importance to the a­ bnormalities in the head region with reference to the sperm chromatin form; if it is intact, compact or not. These features can be viewed with the help of AFM. This is a key aspect in the artificial reproduction process and is referred again. New methods to evaluate the sperm chromatin and particularly sperm DNA fragmentation have been developed in recent years (35). The combination of one of these methods with AFM could be of great value to increase the knowledge about spermatozoa functional capability for fertilization. 3.2. Neck Structure

The neck of the spermatozoon is very often called “the connecting piece” as it is a bridge between the head and the upper part of the flagellum. The main structural components of the connecting piece are the basal plate or capitulum and the segmented columns. Interior in the connecting piece, and distally to the basal plate, a transversely oriented proximal centriole lies between the longitudinal oriented distal centriole and a depression in the capitulum. The connecting piece joins the middle piece distally at the beginning of the mitochondrial sheath (20). The tapping mode of AFM allows imaging of light and dense materials equally well. It is known that the mammalian spermatozoa have significant importance in mitochondrial activity as it deals with ATP, and hence the energy for the motility and its healthiness. Abnormalities in the neck piece are a direct indication that the motility is affected in one or in other form. Neck pieces along with the flagellum are related with the motility, one of the key qualities, of the sperm in the fertility process. Pathological modifications are not only limited to the head structure, but they are also extended to the neck pieces and the flagella regions. Because of the presence of the “neck piece” and the centriole structure, a depression or a discontinuity is created around the neck and this can be visualized with the help of longitudinal and transverse topography (Figs. 9 and 10). In this case, transverse topography reveals significant information like the width and the depth of the depression between the neck pieces. It is observed that the depth is of the order of 30 nm and the width lies between 670 and 690 nm. In case of varicocele, these details are more easily noticeable in the transverse topography than in the longitudinal. Therefore, only the transverse profile is provided. In normal sperm, the depression is clearly observed (Fig. 9) meanwhile in varicocele I, a slight tendency for the depression is observed and in varicocele III, the depression corresponding to the neck structure is totally absent and instead of it a peak is observed (see Fig.  10). This indicates that the deterioration and/damage of the neck pieces are progressive from varicocele grade I to varicocele grade III. The absence of the central structure and the disorganization

276

Joshi, Cruz, and Osuna

Fig. 9. The image of the neck structure along with the longitudinal and transverse topographies, of a healthy sperm. Reproduced from Archives of Andrology, Joshi et al. (8).

around the neck are two noteworthy features. The image of the neck structure also confirms similar conclusions. These defects are not only observed in varicocele, but are also observed in globozoospermia and hence these particular features cannot be used for identification purpose, even though they are useful when varicocele is the outstanding clinical feature accompanied with other clinical findings. In the case of OAT patients, the membrane near the “neck pieces” region is completely damaged (Fig. 11) and an unidentified granular structure is detected. This creates obstructions both in the motility and the energy-related features giving rise to slow, sluggish, and incoherent movements, affecting the sperm fertilizing capability, and therefore this sperm is not properly functional. The effects on the neck pieces are more drastic in case of HIV patients. Figure  6 shows that the neck structure is completely damaged and it is not recovered even after HAART. On the contrary, the region is distorted and bulky patches, probably the agglomeration of protein, are extended. A close look shows that the height of the neck is slightly reduced, but the more remarkable feature is that the depression corresponding to the neck structure is absent.

Highlights on Ultrastructural Pathology of Human Sperm

277

Fig. 10. The transverse profile of the neck of the sperm of a patient with varicocele grade I and grade III. Reproduced from Archives of Andrology, Joshi et al. (11).

Fig. 11. Image of the neck piece of a sperm suffering OAT. Reproduced from Fertility and Sterility, Joshi et al. (9).

The effect of the temperature on the neck pieces is still ­ ramatic and can be seen by Figs. 7 and 8 with the corresponding d longitudinal and transverse profiles. Images of the sperms incubated at 38 and 40°C show very strong damage, an uneven surface and irregular patches. At 38°C, the longitudinal profile reveals the region of the neck pieces, which is defective and such

278

Joshi, Cruz, and Osuna

structure disappears totally for the sperm incubated at 40°C. Moreover, there is a steep fall in the height without indication of the presence of any structure. In short, there is a complete damage of the centriole structure. 3.3. Flagellum

The structural components of the flagellum are the axoneme, the mitochondrial sheath, the outer dense fibers, and the fibrous sheath. The axoneme is composed of a “9 + 2” complex of microtubules which extends up to the end of the flagellum. The main structural components of the axoneme are a circle of nine microtubule doublets, interlinked by nexin and connected to the central sheath of the central pair of microtubules by radial spokes. Protruding from each doublet, there is a row of outer and inner dynein arms, which are the motors for the flagellum to move further in a special fashion (20). These structural details are covered under the surface and they are not possible to be detected by AFM unless the membrane is damaged. In stained human spermatozoon, the tail should be straight, uniform, and thinner than the mid piece, uncoiled and approximately 45  mm long. (WHO Manual). In Fig.  12, a complete

Fig. 12. The image and topographical data of the flagellum of a healthy sperm. Reproduced from Archives of Andrology, Joshi et al. (8).

Highlights on Ultrastructural Pathology of Human Sperm

279

length of the flagellum is not shown. Spermatozoa with abnormal flagella are relatively common in human semen. Fertile men have been reported to have 18% or more sperm with abnormal tails (36). In cases of total immotile spermatozoon (severe astenozoospermia), ovum fertilization can be achieved by the new assisted reproductive technologies like ICSI (36, 37). Defects in sperm motility in infertile men might be originated from axonemal defects, which could be related to the absence of dynein arms; such investigation has been carried out by TEM (37). In healthy sperm, the central channel or nanogroove is a wellorganized part from the end of the neck pieces to the end of the tail. This nanogroove is supposed to run throughout its length in a uniform manner (see Fig. 12) without obstruction or conglomeration of protein patches. The presence and details of the channel are appreciated more in the transverse section than in the longitudinal one. It is obvious that in varicocele grade I, the central canals deep is getting covered with proteins and the plateau is recorded. This process is continued even in varicocele III where instead of a dip, a peak is observed. This confirms that the damage is progressive from varicocele I to varicocele III (see Fig. 13). Considering abnormality in the head and also the flagellum (means motility), the sperm of the patient of varicocele III has serious limitations in the fertilizing process. The pathological alterations are more severe in HIV patients (12). The defective nature of the flagellum is all over the length. Not only the channel is absent but the tail is twisted and there are random patches and irregularity over the entire surface of the tail. These types of aberrations are typical in sperms of HIV patients (12).

Fig. 13. The image and topographical data of the flagellum of sperms from varicocele grade I and grade III. Reproduced from Archives of Andrology, Joshi et al. (11).

280

Joshi, Cruz, and Osuna

The flagellum of the sperm incubated at temperatures 38 and 40°C also show considerable alterations which affect their motility. In this particular case, both the transverse and the longitudinal profiles are useful. The last one provides data about the variation in thickness of the sperm. In normal sperm, the thickness varies in a smooth and undulatory manner meanwhile in the case of the sperm incubated at 38°C, the variation in the thickness is not smooth instead it is uneven. However, the overall structure remains the same with little alterations. The situation becomes worst in the case of the sperm kept warm at 40°C. Where the thickness of the tail varies dramatically and randomly, at certain points it reaches up to 70 nm and suddenly drops to 20 nm. This might be due to the conglomeration of some proteins which are unevenly deposited on the flagellum (see Fig. 14). The sperm incubated at 38°C shows the tendency of the ­presence of the central canal. This is observed by the transverse

Fig. 14. The image and longitudinal topographical data of the flagellum of a sperm incubated at 40°C.

Highlights on Ultrastructural Pathology of Human Sperm

281

profile and the images themselves. In healthy sperm, the groove is well marked from the neck to the end of the tail. This is not so for the incubated sperm. When it is incubated at 40°C, the central groove disappears completely and a plateau is observed (Fig. 14). The width of the present plateau is considerable (approximately 540 nm). All these changes cause impediments in the floating and swimming motion, and hence they fail the “swim-up test.” As mentioned earlier, this view, namely, the pathological alteration for the evaluation of the sperm quality has been applied to male contraceptives and the tail has been examined after the treatment. Certainly, the tail has notable alterations not only topographical, but also in its form. The coiling of the tail is clearly visible: obviously, the sperm fertilization capability is out of question (25). It is known that at a high temperature of the scrotal (18), the sperm production is reduced. The present investigation shows that at high temperature, the quality of the sperm is also considerably lowered and abnormalities, which affect functioning, are spread over the entire part of the sperm.

4. Supplementary Information AFM is the only technique which provides information at subnanostructure of the living organisms in their natural environment. It is, therefore, expected that this technique could reveal some facts or specific data which were not revealed before. This has been precisely observed in the case of HIV patients. The role of the sperms in HIV is an interesting aspect in the last decade. It was thought that HI-virus cannot enter in the sperm. However, AFM clearly reveals that the images where virus is on the head of the surface (12) and they are merging or penetrating. This is because HIV surface glycoprotein, gp 120, binds specifically to the galactosyl–acylglycerol (GALAAG) and its sulfate form. The seminolipid (SGALAAG) acts as an HIV receptor. Such valuable information was expected but never detected experimentally. This is the first time that the presence of the virus and its merging process is clearly perceived with the help of AFM technique. Thus, it seems that semen is a safe haven for HI-virus. This awareness is very significant for pharmacological research in HIV treatment.

5. Applications and Prospects The outcome of the ultrastructural investigation of human sperm has direct application in identifying certain infertility-related diseases and also in the developing of certain drugs for contraceptive purposes.

282

Joshi, Cruz, and Osuna

The morphology of the sperm is a significant prognostic f­actor for fertility-related issues, and potential targets are in the development of male contraceptive drugs. The investigation shows that there are certain drugs which produce charge ­imbalance on the sperm membrane, and consequently the sperm surface is destabilized and the acrosomal contents are dispersed (24, 25). Obviously, the morphology and the topology are altered and the effect of the drugs can be visualized and evaluated with the help of AFM. Such work has been recently carried out by Kumar et al. (24, 25), and it is expected that male contraceptives are available in a short time. It is necessary to examine the causes of infertility in men who are trying to surpass an infertility condition by assisted reproductive techniques for artificial reproduction purpose. The patients having globozoospermia are the most suitable candidates for artificial reproduction as the only defect is in the acrosome cap, which can be viewed by AFM. If the height of the head is normal and the acrosome is missing then independent of the morphology of the tail, the sperm is just fit for artificial reproduction. This suggests that AFM technique should be in the hands of the andrologists for routine procedures. Investigation on the temperature dependence of the alterations in the morphology of the sperm is now possible to be detected with very recent developments in AFM technology (5). It has been reported that atomic force microscopic measurements can be carried out with the precision of 0.025°C with long-term reproducibility. This clearly suggests that now it is time to examine pathological alterations in human sperm very precisely, and to examine what type of alterations and how they are taking place in human sperm to evaluate the impact of fever and related diseases on the fertility of the patients. Such information is truly needed for research in genetics-related fields and for reproductive medicine. It is our great concern to get more information about the environmental effects on the healthiness of the sperm. Morphological alterations also help to assess the chromatin structure as well as intact and reacted acrosome in human sperm. These features are decisive in fertility-related diseases. In the last couple of years, substantial improvements have been made in the technology of AFM both in speed and resolution and now it is emerging as a powerful tool in the hands of scientists who are working in reproductive medicine. The pathology of human sperm has impact not only in pharmacology, but also in biotechnology, microbiology, histology, and genetics. These features will be exploited in the near future. Thanks to AFM.

Highlights on Ultrastructural Pathology of Human Sperm

283

Acknowledgments We are thankful to CONICIT (Venezuela) for financial support. We also thank to Honorio Medina and J.M. Barboza for their valuable help and collaborations in this project. References 1. Egerton R.F. (ed.) (2008) Principles of electron microscopy: An introduction to TEM, SEM& AEM, Springer. 2. Swam M.A.(1997) Improved preservation of ultrastructural morphology in human spermatozoa using betaine in the primary fixation International journal of andrology 20, 45–54 3. Cohen S.H. and Lightbody M.L. (1999) Eds Atomic Force Microscopy, Kluwer academic, New York. 4. Han W., Mou J., Sheng J., Yang J. and Shao Z. Cryo (1995) Atomic Force Microscopy: a new approach for biological imaging at high resolution. Biochemistry 34 (26), 8215–20. 5. Agilent Technologies. www.agilent.com/find. 5600LS AFM, data sheet. 6. Joshi N.V., Medina H, Urdanedta H, Barboza J (2000) Atomic force microscopic study of ultrastructure of entamoeba histolytica in Scanning and force microscopies for biomedical applications II. Edt. Nie S, Tamiya E. and Yeng E. published in SPIE, 3922, 222–227. 7. Joshi N.V., Medina H., Urdaneta H. and Berrueta L. (1999) In vivo nanoimaging and ultrastructure of entamoeba histolica by using atomic force microscopy Experimental Parasitology 93, 95–100. 8. Joshi N.V., Medina. H, Colasante C. and Osuna A (2000) Ultrastructural investigation of human sperm using atomic force microscopy, Archives of Andrology 44, 51–57. 9. Joshi N., Medina H., Cruz I. and Osuna J. (2001) Determination of the ultrastructural pathology of human sperm by atomic force microscopy Fertility and Sterility, 75, 961–965. 10. Allen M.J, Bradbury E.M and Balhorn R. (1996) The chromatin structure of well spread demembranated human sperm nuclei revealed by atomic force microscopy. Scanning Microscope, 10, 989–94. 11. Joshi N.V., Medina H. and Osuna J.A. (2001) Ultrastructural pathology of varicocele spermatozoa by using atomic force microscopy Archives of Andrology 47, 143–152.

12. Barboza J.M., Medina H., Doria M. Rivero L. Hernandez L. and Joshi N. V. (2004) Use of atomic force microscopy to reveal sperm ultrastructure in HIV-patients on highly active antiretroviral therapy. 50, 121–129. 13. Irvine D.S. (1998). Epidemiology and etiology of male infertility. Hum. Reprod. 13 (Suppl 1):33–44. 14. Bhasin S, de Krester D.M and Baker H. W. (1994) Pathophysiology and natural history of male infertility. J Clin Endocrinol Metab. 79:1525–9. 15. World Health Organization. (1991) Infertility: a tabulation of available data on prevalence of primary and secondary infertility. Geneva., WHO Programme on Maternal and Child Health and Family Planning, Division of Family Health. 16. Gordon Baker H.W. Clinical Management of Male Infertility (2006). In De Groot L.J, Jameson J.L (eds) Endocrinology. Fifth ed. Vol.3 Elsevier Saunders, Philadelphia, PA, USA. 3199–3225. 17. Kerr J.B, de Krester D. Functional Morphology of the Testis.In: De Groot L.J, Jameson J.L. (2006) (eds) Endocrinology. Fifth ed. Vol.3 Elsevier Saunders, Philadelphia, PA, USA. 3089–3138. 18. Mieusset R, Bujan L. (1995) Testicular heating and its possible contributions to male infertility: a review. Int. J. Androl. 18, 169–184. 19. WHO laboratory manual for the examination of human semen and sperm-cervical mucus interaction (1999). Fourth edition. World Health Organization. Cambridge University Press. 20. Eddy E.M. (1998) The spermatozoon in “The physiology of reproduction”. Eds. Knobil E. and Neil J., Raven press, New York, 27–88. 21. Davis J.R. and Langfordt G.A. Testicular proteins. In: Johson A.D, Gomes W.R, and Vandemark N.L. (1970) The Testis. Vol.II. Academic Press New York and London. 259–306.

284

Joshi, Cruz, and Osuna

22. Hecht N.B, Bower P.A, Waters S.H, Yelick PC, Distel R.J. (1986) Evidence for haploid expression of mouse testicular genes. Exp. Cell. Res. 164, 183–90. 23. WHO: The influence of varicocele on parameters of fertility in a large group of men presenting to infertility clinics. World Health Organization. Fertil Steril. (1992) 57: 1289–93. 24. Kumar S. Chaudhary K. Sen P. and Guha S.K. (2006) Topological alterations in human spermatozoa associated with polyelectrolytic effect of RISUG Micron 37, 526–32. 25. Kumar S. Chaudhary K. Sen P. and Guha S.K. (2005) Atomic Force Microscopy: a powerful tool for high-resolution imaging of spermatozoa J. nanobiology. 3, 9–16. 26. Lee I.D.. Allen M-J- and Balhorn (1997) Atomic force microscope analysis of chromatin volumes in human sperm with head shape abnormalities. 56, 42–49. 27. MacLeod J. (1965). Seminal cytology in the presence of varicocele. Fertil Steril. 16, 735–57. 28. Zucchi A, Mearini L, Mearini E, Fioretti F, Bini V, Porena M. (2006) Varicocele and Fertility: Relatioship Between Testicular Volume and Seminal Parameters Before and After Treatment. J Androl. 27, 548–51. 29. Sandlow J. (2004) Pathogenesis and treatment of varicocele. Br Med. 328, 967–68. 30. Mieusset R, Bengoudifa B, and Bujan L.(2007) Effect of Posture and Clothing on Scrotal Temperature in Fertile men. J. Androl. 28:170–75.

31. Bergmann M, Behre HM, Nieschlag E. (1994) Serum FSH and testicular morphology in male infertility. Clin. Endocrinol. 40, 133–36. 32. Holstein A.F, Schirren C.G, Schirren C. Human spermatids and spermatozoa lacking acrosome. J Reprod. Fertil. (1973), 35, 489–91. 33. Brown-Woodman P.D., Post E.J., Gass C. and White I.G. (1984). The effect of a single sauna exposure on spermatozoa Arch. Andrology 12, 9–15. 34. Wang C. McDonald V., Leung A. Superlano L.; Berman N., Hull L. and Swerdloff R.S. (1997) Effect of increased scrotal temperature on sperm production in normal men. Fertility and sterility 68, 334–339. 35. Chohan K.R, Griffin J.T, Lafrombroise M, De Jonge C.J, Carrel D.T. (2006) Comparison of Chromatin Assays for DNA Fragmentation Evaluation in Human Sperm. J. Androl. 27, 53–59. 36. Ord T, Patrizio P, Marello E, Balmaceda J.P, ASCH RH. Mini-Percoll ( 1990): A new method of semen preparation for IVF in severe male factor infertility. Hum Reprod.; 5 :987–989. 37. Palermo G, Joris H, Derde M.P, Camus M, Devroey P, Van Steirghem A.C. ( 1993) Sperm characteristics and outcome of human assisted fertilization by sub-zonalinsemination and intracytoplasmatic sperm injection. Fertil.& Steril. 59, 826-835.

Chapter 18 High-Speed Atomic Force Microscopy and Biomolecular Processes Takayuki Uchihashi and Toshio Ando Abstract Atomic force microscope (AFM) is unique in its capability to capture high-resolution images of biological samples in liquids. This capability will become more versatile to biological sciences if AFM additionally acquires an ability of high-speed imaging, because “direct and real-time visualization” is a straightforward and powerful means to understand biomolecular processes. However, the imaging speed of conventional AFM is too slow to capture moving protein molecules at high resolution. In order to fill this large gap, various efforts have been carried out in the past decade. In this chapter, the past efforts for increasing the scan rate and reduction of tip–sample interaction force of AFM and demonstration of direct visualization of biomolecular processes are described. Key words: AFM, Protein, Bio-imaging, Dynamics, High-speed AFM

1. Introduction Atomic force microscopy (AFM) allows us to visualize individual protein molecules in aqueous solutions directly at submolecular resolution (1). However, its application range in biological research is limited because of its low imaging rate. With conventional AFM instruments, we cannot study dynamic molecular processes carried out by proteins which occur in the subsecond time scale. Over the past decade, significant efforts have been carried out to increase the scan speed of AFM and to achieve low invasiveness under high-speed imaging conditions (for recent reviews, see refs. 2–4). The most advanced high-speed AFM can now capture successive images at 30–60  ms/frame under the condition of a scan range of ~250 nm and ~100 scan lines (5, 6).

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_18, © Springer Science+Business Media, LLC 2011

285

286

Uchihashi and Ando

The tip–sample interaction force is greatly reduced without sacrificing the imaging rate significantly. As a result, dynamic biomolecular processes including molecular interactions, conformational changes, and diffusion of proteins can be captured on video (7–9). This chapter focuses on key techniques that realize fast and low-invasive imaging and gives examples of fast imaging of biomolecular processes.

2. Materials 1. AFM cantilever (AC-10, Olympus, Tokyo, Japan). 2. Mica (natural muscovite or synthetic fluorophlogopite). 3. Epoxy adhesive. 4. Highly oriented Pyrolytic Graphite (HOPG). 5. Phosphatidyl choline (PC, Avanti Polar Lipids, Inc, Alabama, USA). 6. Phosphatidyl ethanolamine (PE, Avanti Polar Lipids, Inc, Alabama, USA). 7. His-tag. 8. Dioleoyl-phosphatidyl-choline (DOPC, Avanti Polar Lipids, Inc, Alabama, USA). 9. Streptavidin (Sigma-Aldrich). 10. Phenol (Wako, Osaka, Japan). 11. Plasma etcher (South Bay Technology, CA, USA). 12. GroEL and GroES (Sigma-Aldrich). 13. Caged adenosine triphosphate (ATP, DOJINDO, Kumamoto, Japan). 14. The imaging buffer I: 50  mM HEPES–NaOH, pH 7.4, 50 mM KCl, 10 mM MgCl2, 400 mM caged ATP and 70nM GroES. 15. UV laser light. 16. Cytoplasmic dynein (gifted by Prof. Y. Toyoshima, Tokyo Univ.). 17. Microtubule (purified from chick brains). 18. Imaging buffer II: 80 mM PIPES–NaOH, pH 6.8, 25 mM potassium acetate, 0.1  mM ATP, 10% dimethyl sulfoxide (DMSO), and 2  mM taxol. Taxol was used to stabilize microtubules.

High-Speed Atomic Force Microscopy and Biomolecular Processes

287

3. Methods 3.1. Theoretical Background for Imaging Speed

Our high-speed AFM is based on the tapping mode (10, 11) in which the AFM tip is vertically oscillated and periodically brought into contact to a sample surface during scanning. The tip oscillation reduces the lateral force between tip and sample, and thus minimizes damage and/or deformation of biological molecules. The vertical tip force acting on the sample is controlled by a Proportional-Integral-Derivative (PID) feedback control so that the oscillation amplitude of the cantilever is kept constant. Precise and fast feedback control is the mandatory requirement for fast and low-invasive imaging (see Note 1). Here we consider factors that limit the scanning speed of AFM. Supposing that an image is taken in a time T for a scan range W × W with scan lines N, the scan velocity Vs in the x-direction is then given by Vs = 2WN/T. For W = 240  nm, N = 100, and T = 30 ms, Vs becomes 1.6 mm/s. Assuming that the sample surface has a sinusoidal shape with a periodicity l in the x-direction, the sample stage is moved in the z-direction with a frequency of f = Vs/l for the tip–sample distance to be maintained constant. When l = 10 nm and Vs = 1.6 mm/s, f becomes 160  kHz. The feedback bandwidth fB should be equal to f or higher and thus can be expressed as fB ³ 2WN/lT. Because of the chasing-after nature of feedback control, the sample topo­ graphy is always traced with a phase delay. The phase delay q is given by ~2 × 2pfDt, where Dt is the open-loop time delay (the sum of time delays of devices contained in the feedback loop). The main delays in tapping-mode AFM are the reading time of the cantilever’s oscillation amplitude, the cantilever’s response time, the z-scanner’s response time, the integral time of error signals in the feedback controller, and the parachuting time (tp). “Parachuting” means that the cantilever tip completely detaches from the sample surface at a steep down-hill region of the sample and thereafter takes time until it lands on the surface again. The feedback bandwidth is usually defined by the feedback frequency that results in a phase delay of p/4. With this definition, we obtain fB = 1/(16Dt). Tip parachuting significantly deteriorates obtained images under high-speed scanning and hence its avoidance or minimization is the most important subject in developing high-speed AFM. The parachuting time tp is a function of various parameters such as the sample height h0, the ratio r of the cantilever amplitude set point As to the free oscillation amplitude A0 (see Note 2), the phase delay q, and the cantilever’s resonant frequency fc (for details, see ref. 2).

288

Uchihashi and Ando

3.2. Devices of High-Speed AFM

The basic structure of our high-speed AFM instrument is similar to that of conventional AFM setups. However, various devices are optimized for high-speed scanning (2, 12). The key devices are given below.

3.2.1. Cantilevers

Cantilevers for fast and low-invasive imaging should have a highresonant frequency and a small spring constant. As a result, cantilevers should have small dimensions (see Fig. 1 and Note 3). Most advanced cantilevers for high-speed AFM are made of silicon nitride (13). They are ~6  mm long, ~2  mm wide, and ~90  nm thick, and are coated with a layer of ~20-nm-thick gold, which results in a resonant frequency fc of ~3.5 MHz in air and ~1.2 MHz in water, spring constant kc of ~0.2 N/m, and quality factor Q c of ~2.5 in water. Therefore, their response time tc(=Q c/pfc) is 0.66 ms in water. This type of cantilever is not yet commercially available, but cantilevers with a resonant frequency of 600  kHz in water and a spring constant of ~0.1 N/m are commercially available.

3.2.2. High-Speed Scanner

Several conditions are required to establish a high-speed scanner: (a) high resonant frequencies, (b) a small number of resonant peaks in a narrow frequency range, (c) sufficient maximum displacements, (d) small crosstalk between the three-dimensional (3D) axes, and (e) low quality factors. We employ flexure stages made of blade springs for the x- and y-scanners (see Fig. 2 and Note 4). The flexure stages are made by monolithic processing to minimize the number of resonant peaks (12). The y-scanner displaces the x-scanner, and the x-scanner displaces the z-scanner over which a sample stage is placed. The maximum displacements of the x- and y-scanners at 100 V are 1 and 3 mm, respectively. The x-piezoactuator is held at both ends with flexures, so that its center of mass is hardly displaced and, consequently, no large mechanical excitation is produced. The x-scanner has resonant

Fig. 1. Electron micrograph of a small cantilever developed by Olympus. Scale bar, 1 mm.

High-Speed Atomic Force Microscopy and Biomolecular Processes

289

Fig. 2. Sketch of the high-speed scanner currently used for imaging studies. A sample stage is attached on the top of the upper z-piezoelectric actuator (the lower z-piezoelectric actuator used for counterbalancing is hidden). The dimensions (W × L × H  ) of the z-actuators are 3 × 3 × 2 mm3.

peaks at 45 and 65 kHz, and higher frequencies; however, these peaks are not large. The z-piezoactuator (maximum displacement, 2 mm at 100 V; self-resonant frequency, 360 kHz) is held only at the four side rims parallel to the displacement direction. The z-piezoactuator can be displaced almost freely in both counter directions, and consequently, impulsive forces are barely exerted on the holder. This holding method has an additional advantage in that the resonant frequency is not lowered by holding, although the maximum displacement decreases by half. The x-scanner is actively damped either by the previously developed Q-control technique (14) or by feedforward control using inverse compensation (15, 16). The z-scanner is also actively damped either by the Q-control technique or by inverse compensation. An electronic circuit that automatically produces an inverse transfer function for a given transfer function was developed (6, 17). By using this compensation, the z-scanner bandwidth fs is extended to ~500 kHz, and the quality factor Q  s is reduced to ~0.5. Therefore, its response time ts (=Q  s/pfs) is ~0.32 ms. 3.2.3. Feedback Controller

The tip–sample interaction force has to be minimized for lowinvasive imaging. This is practically carried out by using an amplitude set point r close to 1. However, under this condition, the oscillating cantilever tip easily detaches from the sample surface at steep down-hill regions of the sample. Once detached completely, like “parachuting,” it takes time for the tip to touch the surface again. Parachuting results in a loss of sample topography information and a low feedback bandwidth. During parachuting, the error signal is saturated at (2A0 − As) = 2A0(1 − r), where 2A0 is the free oscillation peak-to-peak amplitude. A shallower set point results in a smaller saturated error signal and hence prolongs the parachuting time tp. The feedback gain cannot be increased to shorten tp, as a larger gain induces overshoot at up-hill regions of the sample, resulting in the instability of the feedback operation.

290

Uchihashi and Ando

Fig. 3. Schematic showing principle of dynamic PID control. Solid line : an amplitudedistance curve; gray line : an error signal used in conventional PID control; and broken line : an error signal used in dynamic PID control.

To solve this problem, a novel PID controller named “dynamic PID controller” was developed. It can automatically change the feedback gain depending on the oscillation amplitude (18). Namely, the feedback gain is increased when the error signal exceeds a threshold level (see Fig. 3). By using this technique, the parachuting time tp is significantly reduced, or parachuting is avoided completely, resulting in a significant increase in the feedback bandwidth and a decrease in the tapping force. 3.2.4. Other Devices

To drive the z-scanner at a high frequency, a piezodriver with a high slew rate and low electronic noise is required. The piezodriver is custom-made and characterized with a maximum output voltage of 50 V, rms noise <2 mV, high slew rate of 1,000 V/ms, and bandwidth of 3 MHz for capacitive loads up to 2 nF (Mess-tek Co., Ltd., Saitama, Japan). Our small cantilevers have a ~1,000 times larger value of fc/kc, when compared with conventional cantilevers. Therefore, a large phase shift occurs by tip–sample interaction even with a small Q, indicating the potential for fast phase-contrast imaging. We developed a fast phase detector (19, 20). This device can detect phase shifts within a single oscillation cycle and, more importantly, at any timing within a cycle. Because of this feature, we can choose the detection timing at the point where the largest phase shift occurs. In addition, it can be inferred whether the tip–sample interactions conserve or dissipate energy. This is because the phase shift due to an energy-conservative interaction decreases very

High-Speed Atomic Force Microscopy and Biomolecular Processes

291

quickly with time, while the phase shift caused by an energy-dissipative interaction is maintained over the oscillation cycle (20). 3.3. Preparations of Substrate and AFM Tip

The choice of the substrate to support sample is very important for AFM observation of biomolecular processes. Visualization of biomolecular processes requires the substrate to have various properties: for example, (a) the surface must have appropriate binding affinity for the sample so that the sample does not diffuse too quickly on it, (b) the functional activity of the sample should be retained on the surface, (c) the surface should selectively bind a specific component in a multicomponent sample, and (d) the sample should attach to the surface in a desired orientation. For high-speed AFM observation, we use various substrates such as mica, HOPG, lipid bilayers, and streptavidin crystals on biotinylated lipid bilayers (21). Here, we describe only frequently used substrates: bare mica and mica surface-supported planar lipid bilayers. High spatial resolution is also important in high-speed AFM. However, small cantilevers with a sharp tip are not commercially available at present. We describe the preparation of sharp tips as well.

3.3.1. Bare Mica

Mica has often been used as a substrate for AFM observations owing to its surface flatness at the atomic level over a large area. It has net negative charges and is, therefore, quite hydrophilic. A bare mica surface adsorbs various proteins through electrostatic interactions. Except in some cases (e.g., GroEL attachment in an end-up orientation), the orientation of adsorbed proteins is not unique, and the selective attachment of a specific species is not expected. However, when the dynamic processes of a single species of protein are to be observed without other proteins, a mica surface is useful. We can control the affinity for a specific protein by varying the ionic strength or pH, or by adding divalent cations such as Mg2+. 1. Punch mica disks 1–2 mm in diameter and <0.1-mm thick. Avoid serrated edge formation which often accompanies partial cleavage of interlayer contacts in the disk (see Note 5). 2. Glue a mica disk to the sample stage using an epoxy and dry it for at least 1 h. 3. Press Scotch tape to the mica surface and remove the tape from the mica. Check whether the mica surface is smooth and does not have some burrs. 4. Place a sample solution of 2–4 ml on the freshly cleaved surface and leave it for 1–3 min, and then rinse with an appropriate buffer solution (see Note 6). 5. Find an appropriate solution condition by repeated imaging using different solutions (see Note 7).

292

Uchihashi and Ando

3.3.2. Mica-Supported Planar Lipid Bilayers

Mica surface-supported planar lipid bilayers are useful for adsorbing proteins to their surfaces with a controlled affinity or selectivity. A membrane surface with zwitterionic polar head groups such as PC and PE is known to resist protein adsorption (22, 23). Various lipids with functional groups that are attached to polar groups (e.g., biotin attached to PE and Ni-NTA attached to GS) are available. They enable specific attachment of proteins labeled with biotin or His-tag onto planar lipid bilayers (21). DOPC is useful for the preparation of 2D streptavidin crystals when it is used together with biotinylated lipids because streptavidin allows the specific attachment of a biotinylated protein without adsorbing nonbiotinylated proteins (see Fig. 4 and Note 8). Here, we describe a simple vesicle fusion method for preparing uniform and smooth surfaces of lipid bilayers (see Note 9). 1. Dissolve each lipid compound in chloroform or in a mixture of chloroform, methanol, and water (follow the manufacturer’s instructions). 2. Mix each lipid solutions at a desired ratio in a glass test tube. 3. Dry the organic solvent under a stream of N2 (or Ar) gas. 4. Place the test tube in a desiccator evacuated by an aspirator for more than 30 min. 5. Add a buffer solution to the test tube (typical final concentration of lipids, 0.125 mM) and vortex it. At this stage, multilamellar lipid vesicles are formed. 6. To obtain small unilamellar vesicles, sonicate the multilamellar vesicle suspension (typically 100  mL) with a tip sonicator at

Fig. 4. Two-dimensional streptavidin crystals formed on bilayers of DOPC + DOPE-biotin supported on a mica surface. Scan range is 150 nm.

High-Speed Atomic Force Microscopy and Biomolecular Processes

293

intervals of 1 s with a duty ratio of ~0.5 until the suspension becomes transparent (typically, no more than 30 cycles). 7. Place a drop of the small unilamellar vesicle solution on a freshly cleaved mica surface and leave it for 30 min in a sealed container, while maintaining high humidity in the container to avoid sample drying. Subsequently, rinse the sample with an appropriate buffer solution. 3.3.3. EBD Tip

We use electron beam deposition (EBD) to grow an amorphous carbon tip on the original tip. The tip fabrication process is given below (see Note 10). 1. Prepare a small container with small holes (~0.1 mm diameter) in the lid. 2. Place a piece of phenol crystal (sublimate) in the container and put small cantilevers above the holes on the lid. Then place the container in the SEM chamber. 3. Wait until thermal drift of the SEM stage ceases. It usually takes at least 1 h. 4. Irradiate an electron beam on the cantilever tip with the spot mode for 1–2 min. The growth rate of an amorphous carbon tip is ~50 nm/s. The newly formed tip has an apex radius of 15–25 nm. 5. Sharpen the amorphous carbon tip using a plasma etcher (PE-2000, South Bay Technology, California, USA) for 5–10 min at 15 W in argon or oxygen gas. The tip radius can be decreased to 4–5 nm (see Fig. 5).

Fig. 5. (a) SEM image of an EBD tip grown from phenol as a vapor source. The deposition time was 90 s. The electron energy, working distance, and aperture size of the FE-SEM during the deposition are 20 kV, 4 mm, and 30 mm, respectively. (b) The apex radius of the grown tip is about 17 nm. (c) After etching in Ar gas for 8 min using a plasma etcher with the power of 16 W, the apex radius is reduced to about 5 nm.

294

Uchihashi and Ando

3.4. Imaging of Biomolecular Processes

High-speed AFM is not completely established yet as a tool for routinely observing biomolecular processes. At present, it is important to examine whether we can really image biological processes that have been expected or known to occur. It is also important to survey adequate substrates and experimental conditions for different samples. These conditions are often different from those for still imaging. Along with gradual accumulation of successful exemplifications of known molecular processes, high-speed AFM will be considered a reliable tool, while newly filmed data on unexplored biological processes will be widely accepted. Here we show high-speed AFM imaging of biomolecular processes. It is important to observe how protein conformation changes dynamically upon binding to its ligand. Protein molecules always exhibit Brownian motion. Therefore, thermal and ligand-induced conformational changes have to be distinguished. To facilitate this distinction, we introduced a technique of UV photolysis of caged compounds to the AFM (see Note 11). As a demonstration of this technique, we observed the binding events between an AAA ATPase chaperonin GroEL and a co-chaperonin GroES upon uncaging ATP from caged ATP. The AFM observation was carried out under the imaging buffer I. GroEL consists of two homoheptameric rings stacked back to back and assists its substrate polypeptides to fold into their functional 3D structures, whereas GroES is a single homoheptameric ring. GroES binds to GroEL only when GroEL is in the nucleotide-bound state. GroEL was attached to the mica surface with an end-up orientation, forming a highly packed single layer (see Fig. 6 and Note 12). Before UV flash, GroES floats in the solution containing caged ATP and is, therefore, not visible. High-frequency pulses of a UV laser light (355 nm) were applied within 5 ms. Immediately after that, GroES attached to the top of GroEL, forming protrusions on the GroEL layer. Since the concentration of ATP uncaged in a small UV-irradiated volume quickly decreased due to diffusion, the second application of UV pulses resulted in more protrusions.

Fig. 6. GroES binding to GroEL immediately after ATP release by flash photolysis of caged ATP. The number attached to each image indicates the time (s) elapsed after imaging began. UV flash light was applied shortly before the frames marked with “F.” Scan range, 500 nm; imaging rate, 0.96 s/frame.

High-Speed Atomic Force Microscopy and Biomolecular Processes

295

Fig. 7. Processive movement of yeast cytoplasmic dynein along microtubules. The same dynein molecule is indicated by arrows. The number attached to each image indicates the time. Scan range, 300 nm; imaging rate, 0.56 s/frame.

We have applied high-speed AFM to observe processive movement of motor proteins such as myosin V, dynein, and kinesin along fibrous protein tracks. Here, we show the processive movement of a microtubule-based cytoplasmic dynein which mediates various biological processes, including nuclear migration and organelle transport. The AFM observation was carried out under the imaging buffer II. Microtubules were adsorbed onto a bare mica surface (see Note 13). Figure 7 shows successive images of yeast cytoplasmic dynein on the microtubules in the presence of 0.1 mM ATP. The dynein molecule (arrow) moves from left to right on the images, often changing from one microtubule to the other. The average translocation velocity was 68 nm/s, which is in good agreement with that measured using fluorescent microscopy (see Note 14). Thus, the tip–sample interaction does not influence the physiological function of dynein. 3.5. Conclusion

Here, we presented key techniques of high-speed AFM to observe the dynamic behavior of biomolecules in action. We demonstrated the ability of high-speed AFM to visualize biomolecular processes by imaging GroEL–GroES binding events synchronized with uncaging of ATP and processive movement of yeast cytoplasmic dynein along microtubules. Direct observation of biomolecules in action is more straightforward than optical and electron microscopic observations. This new microscopy, therefore, has a great potential of solving various questions related to how biological molecules function.

4. Notes 1. The force acting in protein–protein interactions approximately ranges from 1 to 100  pN. Even the single “rigor” complex of a muscle-myosin head and an actin filament, which hardly dissociates in equilibrium, is ruptured quickly by a pulling force of ~15  pN (24). The force produced by motor proteins during ATP hydrolysis is generally a few pN

296

Uchihashi and Ando

(see ref. 25). We should note that the mechanical quantity which affects the sample is not the force itself but the force impulse, i.e., the product of force and the time over which the force acts. In tapping-mode high-speed AFM, the time of force action is short, and therefore, a relatively large peak force (<20 pN) would not affect the sample significantly. 2. The amplitude set point r is usually determined by compromising two factors: (1) increase in tapping force with decreasing r and (2) decrease in the feedback bandwidth with increasing r owing to parachuting. 3. The resonant frequency fc and the spring constant kc of a rectangular cantilever with thickness d, width w, and length L are expressed as

d E f c = 0.56 2 L 12r

(1)

and

wd 3 kc = E 4L3

(2)

where E and r are Young’s modulus and the density of the material used, respectively. Young’s modulus and the density of silicon nitride (Si3N4), which is often used as a material for soft cantilevers, are E = 1.46 × 1011 N/m2 and r = 3.087 kg/ m3, respectively. To attain a high resonant frequency and a small spring constant simultaneously, cantilevers with small dimensions must be fabricated. 4. The scanner is made of stainless alloy (SUS 304). The gaps in the piezoactuator are filled with an elastomer to damp the vibrations passively. This passive damping is effective in suppressing low-frequency vibrations. Aluminum or duralumin is often used as a material for the scanner. Magnesium and magnesium alloys appear to be more suitable materials because of their larger mechanical damping coefficients and larger ratios of Young’s modulus to density. However, from our experience, only a slight improvement is attained using these materials. 5. Hydrodynamic pressure produced by rapid scanning of the sample stage induces vibrations of the mica disk through movement of the cleaved sites. For high-speed imaging, the disk should be small (1–2 mm in diameter) to avoid generation of too large a hydrodynamic pressure (26). Serrated edges of the mica surface also induce unwanted vibrations which decrease the feedback bandwidth.

High-Speed Atomic Force Microscopy and Biomolecular Processes

297

6. The sample volume that can be placed on the mica surface is small and therefore rinsing should be carefully carried out so as not to dry the mica surface. One should also carefully apply a sample solution to the mica surface so that the solution does not spill over. Otherwise, the mica surface is sometimes smeared as the solution touches parts other than the mica surface. 7. When the affinity is too weak, reduce the ionic strength or pH, or increase the concentration of divalent cations. 8. DPPC contains no unsaturated hydrocarbons in the alkyl chains, and therefore, its phase-transition temperature is high (~41°C) and it is appropriate for preparing planar bilayers with low fluidity. For example, when planar bilayers are formed with DPPC at a high temperature (~60°C) together with a certain fraction of DPPE-biotin, streptavidin, which is sparsely attached to the surface, hardly diffuses at room temperature (Fig. 8a). When DOPE-biotin is used together with DPPC, the sparsely attached streptavidin diffuses at a moderate rate (Fig. 8b). 9. Examples for lipid compositions [buffer conditions] are given below. The mixing ratios of lipids should be changed depending on the samples and dynamic events to be visualized.

Fig. 8. (a) Streptavidin on bilayers of DPPC + DPPE-biotin. Scan range, 200 nm; imaging rate, 0.18 s/frame. (b) Streptavidin on bilayers of DPPC + DOPE-biotin. Scan range, 200 nm; imaging rate, 0.18 s/frame. The number attached to each image in (a) and (b) indicates the time (s) elapsed after capturing the first frame. The apparent difference between (a) and (b) in the size of streptavidin molecules is due to the difference in the sharpness of the cantilever tips used.

298

Uchihashi and Ando

(a) Highly fluidic bilayers with biotin; DOPC:DOPS:biotincap-DOPE = 7:2:1 (w/w) [10  mM HEPES–NaOH (pH 7.4), 150 mM NaCl, and 2 mM CaCl2]. (b) Low fluidic bilayers with biotin; DPPC:biotin-capDPPE = 9:1 (w/w) [10  mM HEPES–NaOH (pH 7.4), 150 mM NaCl, and 2 mM CaCl2]. (c) Slightly fluid bilayers with biotin; DPPC:biotin-capDOPE = 9:1 (w/w) [10 mM HEPES–NaOH (pH 7.4), 150 mM NaCl, and 2 mM CaCl2]. (d) Highly fluidic bilayers with Ni-NTA; DOPC:DOPS: DOGS-NTA(Ni) = 7:2:1 (w/w) [10 mM HEPES–NaOH (pH 7.4), 150 mM NaCl, and 2 mM CaCl2]. (e) Low fluidic bilayers with Ni-NTA; DPPC:DOGSNTA(Ni) = 9:1 (w/w) [50  mM Tris–HCl (pH 8.0), 50 mM KCl, and 3 mM MgCl2]. (f) Low fluidic bilayers with positively charged head groups; DPPC:DPTAP = 7:3 (w/w) or only DPTAP [distilled water]. (g) Low fluidic bilayers with negatively charged head groups; DPPG 100% [10 mM HEPES–NaOH (pH 7.4), 150 mM NaCl, and 2 mM CaCl2]. (h) Low fluidic bilayers with negatively charged head groups formed on positively charged lipid bilayers; DPPA 100% [distilled water]. 10. We tested various materials as a source gas. The growth rate of an amorphous carbon tip is the highest with naphthalene and phenol among various materials tested. However, the tip made from naphthalene is fragile and frequently broken by just contact with the mica surface even without scanning. The tip made from phenol has higher durability. In fact, it takes 1 day to remove the tip completely by plasma etching. 11. An intense flash illumination by a 355-nm laser causes large bending of a cantilever due to a photothermal expansion effect, which causes hitting of the tip against the substrate, leading to tip damage. To solve this problem, attenuated highfrequency laser pulses (~50 kHz) should be applied while the y-scanner is being scanned toward the starting point after the completion of one frame acquisition. In addition, during this period, the sample stage is withdrawn from the cantilever tip. 12. GroEL solution (0.03 mg/ml) was put on a mica surface, incubated for 3 min, and the unbound GroEL was carefully washed away. The GroEL solution was applied again on the substrate surface for 3 min and then the sample was rinsed. By this procedure, the whole mica surface was covered by GroEL.

High-Speed Atomic Force Microscopy and Biomolecular Processes

299

13. A solution containing 1 mg/ml microtubules was applied on a bare mica surface for 3  min and then unbound microtubules were washed away. A solution containing yeast cytoplasmic dynein was applied on the surface for 3 min and then the sample was washed again to remove unbound dynein. 14. In Fig. 7, the dynein molecule (arrow) moved on the top of the microtubules. We also observed movement of a dynein molecule along the side of the microtubule. In this case, translocation velocity was about 7 nm/s, much slower than that shown in Fig.  7. This is probably because the dynein bound at the side wall of a microtubule is influenced by the mica surface.

Acknowledgments We thank N. Kodera, D. Yamamoto, M. Shibata, A. Miyagi, M. Taniguchi, H. Yamashita, and all previous students for their dedicated studies for developing high-speed AFM. This work was supported by the Japan Science and Technology Agency (JST; the CREST program and a Grant-in-Aid for Development of Systems, Technology for Advanced Measurement and Analysis) and the Japan Society for the Promotion of Science (JSPS; a Grant-in-Aid for Basic Research (S)). References 1. Müller, D. J., Fotiadis, D., Scheuring, S., Müller, S. A., and Engel, A. (1999) Electrostatically balanced subnanometer imaging of biological specimens by atomic force microscope. Biophys. J. 76, 1101–1111. 2. Ando, T., Uchihashi, T., and Fukuma, T. (2008) High-speed atomic force microscopy for nano-visualization of dynamic biomolecular processes. Prog. Surf. Sci. 83, 337–437. 3. Ando, T., Uchihashi, T., Kodera, N., Yamamoto, D., Miyagi, A., Taniguchi, M., and Yamashita, H. (2008) High-speed AFM for nano-visualization of biomolecular processes. Pflugers Arch. – Eur. J. Physiol. 456, 221–225. 4. Yamamoto, D., Uchihashi, T., Kodera, N., Yamashita, H., Nishikori, S., Ogura, T., Shibata, M., and Ando, T. (in press) Highspeed Atomic Force Microscopy Techniques for Observing Dynamic Biomolecular Processes. Methods in Enzymology 5. Ando, T., Uchihashi, T., Kodera N., Miyagi, A., Nakakita R., Yamashita H., and Sakashita M. (2006) High-speed atomic force microscopy

6.

7.

8.

9.

for studying the dynamic behavior of protein molecules at work. Jpn J Appl Phys 45, 1897–1903. Yamashita, H., Kodera, N., Miyagi, A., Uchihashi, T., Yamamoto, D., and Ando, T. (2007) Tip-sample distance control using photo-thermal actuation of a small cantilever for high-speed atomic force microscopy. Rev Sci Instrum. 78, 083702 (5 pp) Miyagi, A., Tsunaka, T., Uchihashi, T., Mayanagi, K., Hirose, S., Morikawa, K., and Ando, T. (2008) Visualization of intrinsically disordered regions of proteins by high-speed atomic force microscopy. Chem. Phys. Chem. 9, 1859–1866. Yamamoto, D., Uchihashi, T., Kodera, N., and Ando, T. (2008) Anisotropic diffusion of point defects in two-dimensional crystal of streptavidin observed by high-speed atomic force microscopy. Nanotechnology 19, 384009 (9 pp). Yamashita, H., Voïtchovsky, K., Uchihashi, T., Contera, S. A., Ryan, J. F., and Ando, T.

300

10.

11.

12.

13. 14.

15.

16.

17. 18.

Uchihashi and Ando (2009) Dynamics of bacteriorhodopsin 2D crystal observed by high-speed atomic force microscopy. J. Struct. Biol. 167, 153–158. Zhong, Q., Inniss, D., Kjoller, K., and Elings, V. B. (1993) Fractured polymer/silica fiber surface studied by tapping mode atomic force microscopy. Surf. Sci. Lett. 290, L688–L692. Hansma, P. K., Cleveland, J. P., Radmacher, M., Walters, D. A., Hillner, P. E., Bezanilla, M., et al. (1994) Tapping mode atomic force microscopy in liquids. Appl. Phys. Lett. 64, 1738–1740. Ando, T., Kodera, N., Takai, E., Maruyama, D., Saito, K., and Toda, A. (2001) A highspeed atomic force microscope for studying biological macromolecules. Proc. Natl. Acad. Sci. USA 98, 12468–12472. Kitazawa, M., Shiotani, K., and Toda, A. (2003). Batch fabrication of sharpened silicon nitride tips. Jpn. J. Appl. Phys. 42, 4844 – 4847. Kodera, N., Yamashita, H., Ando, T. (2005) Active damping of the scanner for high-speed atomic force microscopy. Rev Sci Instrum. 76, 053708, (5 pp). Schitter, G., and Stemmer, A. (2004) Identification and open-loop tracking control of a piezoelectric tube scanner for high-speed scanning-probe microscopy. IEEE Trans Control Systems Technol. 12, 449–454. Zou, Q., Leang, K. K., Sadoun, E., Reed, M. J., and Devasia, S. (2004) Control issues in highspeed AFM for biological applications: collagen imaging example. Asian J Control 6, 164–178. Morita, S., Yamada, H., and Ando, T. (2007) Japan AFM roadmap 2006. Nanotechnol 18, 084001 (10 pp). Kodera, N., Sakashita, M., and Ando, T. (2006) Dynamic proportional-integral-differential

19.

20.

21.

22.

23. 24.

25. 26.

controller for high-speed atomic force microscopy. Rev Sci Instrum. 77, 083704, (7 pp). Stark, M., and Guckenberger, R. (1999) Fast low-cost phase detection setup for tappingmode atomic force microscopy. Rev Sci Instrum. 70, 3614–3619. Uchihashi, T., Ando, T., Yamashita, H. (2006) Fast phase imaging in liquids using a rapid scan atomic force microscope. Appl. Phys. Lett. 89, 213112, (3 pp). Yamaoto, D., Nagura, N., Omote, S., Taniguchi, M., and Ando, T., (in press) Streptavidin 2D crystal substrates for visualizing biomolecular processes by atomic force microscopy. Biophys. J. 97. Zhang, S. F., Rolfe, P., Wright, G., Lian, W., Milling, A. J., Tanaka, S., and Ishihara, K. (1998). Physical and biological properties of compound membranes incorporating a copolymer with a phosphorylcholine head group. Biomater. 19, 691–700. Vadgama, P. (2005). Surface biocompatibility. Annu. Rep. Prog. Chem., Sect. C: Phys. Chem. 101, 14–52. Nakajima, H., Kunioka, K., Nakano, K., Shimizu, M., Seto, M., and Ando, T. (1997). Scanning force microscopy of the interaction events between a single molecule of heavy meromyosin and actin. Biochem. Biophys. Res. Commun. 234, 178–182. Schmidt, J. J., and Montemagno, C. D. (2004). Bionanomechanical systems, Annu. Rev. Mater. Res. 34, 315–337. Ando, T., Kodera, N., Maruyama, D., Takai, E., Saito, K., and Toda, A. (2002). A high-speed atomic force microscope for studying biological macromolecules in action, Jpn. J. Appl. Phys. 41, 4851–4856.

Part IV Non-topographical Applications (Force-Spectroscopy)

wwwwwwwwwwwwww

Chapter 19 Atomic Force Microscopy in Mechanobiology: Measuring Microelastic Heterogeneity of Living Cells Evren U. Azeloglu and Kevin D. Costa Abstract Recent findings clearly demonstrate that cells feel mechanical forces, and respond by altering their ­phenotype and modulating their mechanical environment. Atomic force microscope (AFM) indentation can be used to mechanically stimulate cells and quantitatively characterize their elastic properties, providing critical information for understanding their mechanobiological behavior. This review focuses on the experimental and computational aspects of AFM indentation in relation to cell biomechanics and pathophysiology. Key aspects of the indentation protocol (including preparation of substrates, selection of indentation parameters, methods for contact point detection, and further post-processing of data) are covered. Historical perspectives on AFM as a mechanical testing tool as well as studies of cell mechanics and physiology are also highlighted. Key words: Nanoindentation, Elasticity, Young’s modulus, Biomechanics, Elastography

1. Introduction Living cells and tissues constantly interact with the physical world and as a result, their mechanical properties are intimately linked with their biological functions. Even as an organism stands idle, the gravitational field of the planet is ever present, exerting force on all of its cells and organelles. The effect of these forces becomes more apparent when one looks at the field of aerospace medicine, which deals with the clinical conditions associated with space explorers. During extended lower-orbit spaceflights, most habitation conditions remain unchanged except for the gravitational acceleration. Under microgravity environments, astronauts and cosmonauts may experience significant bone loss, alterations to blood and cardiac ventricular chamber volumes, increased

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_19, © Springer Science+Business Media, LLC 2011

303

304

Azeloglu and Costa

vascular resistance, and severe skeletal muscular atrophy (1). Such changes underscore the strength of mechanobiological feedback in physiology and remodeling, even in the absence of alterations in biochemical signaling. The link between physiological function and mechanical stimuli was established over a century ago by the pioneering orthopedic surgeon Julius Wolff, who observed that growth plates in bones were “following” the path of maximum compressive stress (2). He hypothesized that by performing a structural engineering analysis on the force equilibrium of trabecular bone, one can estimate the location and direction of maximal growth, which was a remarkably accurate prediction. Decades later, it is widely accepted that the concept of Wolff’s Law for bone remodeling extends to almost all biological tissues, and that growth of tissues is regulated by externally applied loads (3). Mechanical cues are used in a broad spectrum of applications from evaluation of cell function (4) to stimulating growth of fetal (5), injured (6), or engineered (7) tissues. Clearly, physical forces have a critical effect in growth and remodeling of biological tissues at the cellular level, such that they can dramatically alter physiological function (8). External mechanical forces play a substantial role in cell physiology because cells feel and respond to their mechanical environment (9). Vogel and Sheetz define the three pillars of organism growth as mechanosensing, mechanotransduction, and mechanoresponse (10). Accordingly, the mechanical sensory machinery of cells first evaluate the material properties of their surrounding environment (11, 12), which along with mechanosensitive ion channels (13) and/or stretch-activated opening of cryptic domains (14, 15) alter the intracellular biochemical environment and transcription (16). Finally, cells respond to such changes in mechanical environment by locally altering force modulation (17), cytoskeletal organization (18), and resultant cell stiffness (19). This process may take from sub-seconds to days and even weeks (10). Knowing the exact mechanobiological pathways that affect cell function, we can design cell-based therapies that specifically regulate the desired outcome; treatment of osteoporosis with the use of micro-amplitude oscillations is a prominent clinical example in which mechanical modulation of cellular and molecular function has been successfully carried out (20). Environmental stiffness-guided stem cell differentiation is another promising application (21). 1.1. Historical Perspective

The psychologist Julian Jaynes once suggested “history does not move by leaps into unrelated novelty, but rather by the selective emphasis of aspects of its own immediate past” (22). This idea is no less applicable to cell biomechanics than any other field; indeed, first mention of micromechanics dates back three

Atomic Force Microscopy in Mechanobiology

305

centuries to Alexander Stuart’s Royal Society lectures. While it was mentioned by many since then, it wasn’t until the early twentieth century when pioneer biochemists and physiologists such as George Kite, Robert Chambers, and William Seifriz started performing specialized mechanics-based experiments on isolated cells (23). Mechanical heterogeneity of the cell was recognized about a century ago using a microdissection technique that is similar to modern atomic force microscope (AFM) experiments. Kite and Chambers observed that the nucleus was stiffer than the surrounding cytoplasm in grasshopper spermatocytes (24). A micropipette technique developed by Marshall Barber in 1904 (25) was also used by Seifriz to study the cytoplasmic viscosity of a number of fungi and plant cells; similar to Kite, Seifriz also reported regional variations in intracellular viscosity (26). In an earlier work on oogonia, he also pointed out an increase in apparent viscosity with increasing age (27). Carlson noted spatiotemporal differences in mechanical properties of dividing invertebrate neuroblasts (28) and Lillie hypothesized that intracellular distribution of material within striated muscles was the cause of a cyclic heterogeneity in myofibrillar stiffness (29). While these groundbreaking studies were well ahead of their time, most of them lacked the quantitative rigor required for state of the art characterization of biomaterials and tissues. Many involved qualitative observations such that even the terminology used in these studies varied from one work to another (30). A proponent of standardization of biomaterial properties, Seifriz rated cytoplasmic viscosity using a “grading scheme” based on relative concentration of gelatin solutions (26); however, even in his methodical approach there was minimal quantification. In fact, previously mentioned observations by Carlson and Lillie regarding mechanical heterogeneity were quantitatively confirmed by AFM only within the past decade (31, 32). Throughout the century, as microscopy techniques and methods of micromanipulation advanced, the accuracy of observations and diligent quantification improved as well. In addition, biological significance of micromechanics was more deeply appreciated. Northen was one of the first investigators to report the effects of a cytoskeletal fiber on cellular viscosity, namely the reduction in viscous modulus due to microtubule disruption (33). Rosenberg, using a novel cell compression device, showed that extensibility of embryonic heart and liver cells changed during development (34); he was one of the first to suggest a functional relationship between cell stiffness and biological development. Similarly, Weiss and Garber were among the first to suggest a relationship between the mechanics of the extracellular environment and cell function (35). By the mid-twentieth century, as physicists and engineers started focusing on cell biology problems, studies of cell mechanics

306

Azeloglu and Costa

transformed from a descriptive exercise into a more quantitative discipline. Specialized instruments for applying known deformations, such as precision micropipette aspiration devices, were developed (36) and utilized along with increasingly complex structural models (37). In fact, the modern understanding of cell mechanics shares its roots with the field of bioengineering, as they were both in large part created by a group of visionaries such as Y. C. Fung, Richard Skalak, and Shu Chien. These investigators were some of the first to apply modern engineering analysis to the study of cell mechanics, revolutionizing the experimental (38), theoretical (39, 40), and computational (41) aspects of the field. In addition to their voluminous contributions to the literature, they have mentored scores of scientists who continue to advance and transform this exciting and critical area of research. 1.2. Cell Biomechanics and Pathophysiology

Mechanical properties of tissues are intertwined with their structural properties; thus, they have always played a significant role in clinical diagnostics. Physicians have relied heavily on their tactile senses to characterize, what are sometimes very subtle, changes in the “texture” or “hardness” of skin lesions or internal organs. It has been known for some time that the altered mechanical characteristics of the tissues were correlated with pathophysiology in macroscale – in particular, the rate and density of collagen and calcium turnover were known as important regulators of local mechanical properties in soft and hard tissues, respectively. Applicability of the same phenomenon to individual cells had been considered as early as 1920s by the American physiologist Lewis Heilbrunn, who suggested that pharmacological interventions significantly affect cytoplasmic mechanics in sea urchin eggs (42). Some of the first quantitative biomechanical studies on diseased cells were performed on isolated red blood cells, where erythrocytes from sickle cell anemia patients were shown to have a significantly higher viscosity and membrane rigidity compared to normal cells (43). Since the early 1970s, changes in cell mechanics have been associated with all of the leading human health problems, including cancer, heart disease, and diabetes. Thus, quantifiable cell mechanical properties offer the potential to serve as unique disease markers (44). In fact, several automated high-throughput methods have been recently introduced to test for changes in cell mechanics as a clinical diagnostic tool (45, 46). Some of the conditions that have been shown to affect cell mechanics are summarized in Table 1, along with the measurement technique that was used to probe the related properties.

1.3. Subcellular Mechanics and Mechanotransduction

Cells are composed of organelles and other structural components such as cytoskeletal proteins. Traditionally, these components are studied from a strictly biological perspective where transmission of biochemical signals via messenger molecules is thought to

Method

Optical tweezers, micropipette aspiration, and laser viscometer

AFM

Optical tweezers, AFM

AFM

Micropipette aspiration

Optical tweezers

AFM

Gel stretch, nanoindentation

AFM

Condition

Sickle cell

Leukemia

Cancer (prostate, breast, bladder, lung, and pancreas)

Diabetes

Osteoarthritis

Malaria

HIV

Hypertrophy

Aging

Table 1 Cell mechanics as a correlate of pathophysiology

Cardiomyocytes and epithelial cells stiffen with age

Hypertrophic cardiac myocytes exhibit significantly higher viscous moduli than controls

As it matures, HIV alters its stiffness to increase its infectivity

Parasitized erythrocytes get stiffer as the developmental stage of P. falciparum advances

Osteoarthiritis increases both the equilibrium modulus and viscosity of chondrocytes

Erythrocytes from diabetic patients are stiffer than normal

Deformability of cancerous cells is higher than normal cells; cells get even softer when they are malignant or metastatic. See (136) for a detailed review

Myeloids are significantly stiffer than lymphoids, which are stiffer than normal neutrophils Chemotherapy exposure further increases stiffness

Erythrocytes from sickle cell anemia patients are more viscous with higher membrane rigidity Hydroxyurea treatment restores red blood cell deformability

Result

(149, 150)

(72, 148)

(147)

(146)

(144, 145)

(143)

(137–142)

(131, 135)

(132–134)

References

Atomic Force Microscopy in Mechanobiology 307

308

Azeloglu and Costa

govern cell behavior. Recently, interaction of mechanical forces and biochemical signaling molecules has been realized as an important component of cellular physiology (47). Such mechanobiological pathways have been identified as principal driving forces behind critical biological phenomena such as cell morphology (48), motility (49), and differentiation (50). These interactions can be studied from three perspectives: mechanosensing, mechanotransduction, and mechanoresponse (10). Interestingly, subcellular mechanical properties play a critical role in all three aspects since they govern the relationship between local forces and deformation. Hence, characterization of local cell biomechanics – such as effects of specific organelles and cytoskeletal fibers on cytoplasmic material properties – has garnered a special interest in biomedical research. Accurate representation of the cell’s constitutive properties is critical for proper quantification of physical signals transmitted through the intracellular space (51). Intuitively, different components of an inhomogeneous material under mechanical stress are expected to behave differently (52); given a uniform boundary condition for a heterogeneous cell, the distribution of intracellular deformation will depend on local mechanics. Compared to a softer region, a stiffer organelle may have lower strains and higher stresses (53); therefore, inherent cellular heterogeneity plays a critical role in determining how physical forces are transferred to the mechanosensitive protein machinery of the cell (54). Effects of various local intracellular heterogeneities on cell mechanobio­ logy are summarized in Fig. 1. There are numerous mechanically active cellular components that have been shown to play critical roles in regulating material properties of cells, including but not limited to cytoskeletal fibers such as actin (63, 70, 76, 77), microtubules (72, 73, 78) or intermediate filaments (75, 79) as well as lipid granules (61), proteinsecreting vesicles (60), motor proteins (80, 81), and nuclei (53, 56). While the exact mechanical role of some of the individual components remains a controversial topic (74, 82, 83), it is widely accepted that spatial distribution of these molecules plays a significant role in regulating local forces (62, 64, 84–86). 1.4. AFM as a Mechanical Testing Device

Mechanics of cells have been extensively studied with micropipette aspiration (87), twisting bead rheology (88), uniaxial stretch via micromanipulators (72) or optical tweezers (89), unconfined compression of isolated (90) and cultured (91) cells, various custom-built test rigs (72, 92–96), and AFM (97). The AFM is described by its Nobel laureate inventor, Gerd Binnig, as a hybrid between a “force-encoding stylus” and a “scanning tunneling microscope” (98). Its working principles are similar to those of a profilometer; the sample is pushed against a soft elastic cantilever and the resultant probe deflection is measured (Fig. 2a). Knowing

Atomic Force Microscopy in Mechanobiology

Vesicles: Protein secreting vesicles and lipid granules were shown to have distinct material properties (60, 61).

309

Nucleus: Conflicting findings regarding nuclear stiffness. Most direct studies focusing on isolated nuclei find increased stiffness compared to the cytoplasm. Opposite findings in intact cells may be influenced by the surrounding cytoskeletal elements that are known to be stiff (53, 55-59)

Other Organelles and Structures: The effective modulus of a given locality within a cell depends on the underlying intracellular contents such as gap junctions (62).

Cytoskeleton: Actin stress fibers are the stiffest known intracellular component (18, 63-71). They are thought to be responsible for overall cellular rigidity. Microtubules (72-74) and intermediate filaments (64, 75) play structural roles as well, however their exact mechanical role is not entirely known.

Fig. 1. Schematic representation of intracellular biomechanical heterogeneity. Distinct material properties of spatially complex subcomponents create an environment where local mechanics may differ significantly from perceived aggregate behavior. Knowledge of constitutive properties and volume ratios of constituents can be used to simulate inhomogeneous deformation fields and determine mechanically relevant regions. Biomechanical assays that test for local changes in cell material properties may be able to recognize pathological changes related to alterations in specific cell components.

the spring constant of the cantilever, one can then determine the contact force between the probe and the sample, which is related to the mechanical properties of the sample. A laser beam is reflected off the back of the AFM cantilever and onto a strategically arranged four-quadrant photodetector, such that even angstrom-level vertical and horizontal deflections of the cantilever can be detected (99). Precise real-time three-dimensional (3D) positioning of the cantilever is achieved using high-voltage piezoelectric ceramic blocks or cylinders with software or hardware feedback control. This “haptic” system gives AFM the unique ability to directly interact with objects, observe structural features with sub-optical resolution, and physically manipulate them with quantifiable forces and molecular precision (100).

310

Azeloglu and Costa

b Deflection (nm)

a

k(N/m)

600 450 300 150 0

∆z

0

Force (nN)

c

∆h

1000

2000

3000

4000

5000

Extension (nm) 5.0

2.5

0.0 0

500

1000

1500

2000

Indentation Depth (nm)

Fig. 2. Principles of atomic force microscope indentation. (a) Schematic representation of an indentation where the AFM cantilever of a known spring constant, k (N/m), is extended a distance Dz towards a substrate. As the cantilever contacts the substrate, it deflects a measurable amount Dh, which is proportional to the amount of force required to indent the substrate to that given depth. (b) This process leads to extension–deflection curves for loading (dotted black) and unloading (dotted gray) phases of the indentation cycle, where the contact point (open square) can be determined through a number of algorithms (see Note 17). Force curve on a rigid surface (solid black) is superimposed for reference. (c) Mechanical properties of the underlying substrate can be extracted from the post-contact portion of the indentation, where force can be plotted (gray circles) as a function of indentation depth. In traditional Hertzian contact analysis, a single elastic modulus value for the whole substrate can be obtained by fitting the entire force–depth relationship (black dashed line).

Devised as a high-resolution imaging device (101), the AFM was quickly adapted as a mechanical testing platform for soft biological samples (85) and cells (102). Originally, the AFM stage was the actuator itself with a stationary probe sensor. Later, this design was modified so that the AFM scanner could be placed on top of an inverted microscope for easy adaptability to cell biology. Classical AFM contact mode imaging involves raster scanning the probe over the sample in the plane orthogonal to the indentation axis (z-axis) while applying a constant force onto the sample by keeping the cantilever deflection constant (103). Contact mode imaging provides exceptional detail and is still used extensively in cell biology for various applications (104–108). Other modalities such as tapping mode (109), jumping mode (110), or phase imaging (111) are also used in investigation of cellular processes (104, 112). Even though the working principles of these techniques may differ, an inherent attribute of the AFM that translates

Atomic Force Microscopy in Mechanobiology

311

through all these imaging modalities is the fact that none of the images are pure topographical information; instead they are convolutions of material height and the underlying mechanical properties. The challenge lies with the quantitative deconstruction of that information into its height and elasticity components (44). One of the most critical long-term objectives of the biomechanics field is to develop a rapid, noninvasive cell mechanics-based biomedical test to detect and diagnose early disease conditions, and AFM offers an ideal platform to achieve this lofty goal. In this manuscript, we will outline some of the techniques that have been used to optimize the processing of mechanical information using AFM indentation for biomedical applications.

2. Materials Numerous AFM systems (such as Asylum Research’s MFP-3DBIO) are available commercially with specialized biological capabilities such as simultaneous fluorescent microscopy and force measurements. Most of these AFM systems can be coupled with an inverted fluorescent microscope (such as Olympus’ IX-81), and almost all come with a biological chamber option that provides environmental controls for cell biology experiments. The reader is encouraged to investigate different capabilities and advantages/disadvantages of individual systems for particular applications. 1. Standard culture materials including culture media, serum supplement, and phosphate-buffered saline. 2. High-molecular weight poly-l-lysine and/or other adhesion molecules may be necessary to culture certain adhesiondependent cells such as neurons. 3. Standard polystyrene tissue culture dishes. Alternatively, open-top cover glass-bottom culture dishes are suitable for culture, indentation, and high-resolution immunofluorescent labeling of cells since they allow oil-immersion lenses. 4. CO2-independent medium is a non-HEPES pH buffer that is necessary to keep cells in standard atmospheric conditions for extended periods. 5. If necessary, formaldehyde or other fixative is needed to prepare cells for immunofluorescence. Accordingly, bovine serum albumin or other blocking agents, triton X-100 or other permeabilization agents, and primary/secondary antibodies along with Hoechst/DAPI stains are also necessary for this procedure. 6. Standard blunt-pyramidal silicon nitride probes with gold reflective coating (Veeco, Camarillo, CA; MLCT) are adequate

312

Azeloglu and Costa

for most local cell mechanics measurements. For aggregate modulus measurements, custom large sphere-mounted probes (e.g. Novascan, Ames, IA; PT.PS.15) are more suitable. See Note 1 for details on cantilever selection. In addition to the essential materials listed above, certain instruments and tools may prove useful in probing biomechanical properties of cells with AFM. While this list is not exhaustive, the user is encouraged to inquire into the slew of options that accompany current commercial AFM systems. Biological environment chambers (e.g. Asylum; BioHeater), vibration isolation cabinets (e.g. Asylum; BCH-45), and high-frequency actuators (e.g. Asylum; iDrive) for calibration or viscoelasticity measurements are some of the modules that may be practical upgrades for a biological AFM system. In addition, consider the following items as optional: ●●

●●

Low-power dissection stereoscope (e.g. Olympus; SZ-60) for examining AFM probes before and after experiments. Compact tabletop plasma cleaner with vacuum (e.g. Sigma; Z561673) allows removal of organic waste from expensive custom AFM cantilevers, and can also be used for surface treatment of culture substrates.

3. Methods While the methods outlined in this chapter are focused on utilizing AFM indentation for the study of cell mechanics, they are generally adaptable for whole tissues as well. The indentation techniques outlined in this manuscript can be obtained with most commercially available AFM without additional equipment. However, it is highly advantageous to combine the AFM with an inverted fluorescence microscope system for optical visualization of cells and specifically labeled constituents, and a heated stage enables physiologically relevant measurements on living cells. Most commercial AFMs are equipped with magnetic or vacuumbased rigs to ensure that the substrate, on which the cells or sample are adhered, is secured to the working platform. 3.1. Probing Cell Mechanics with AFM

Below is the step-by-step outline of a standard AFM indentation experiment for an adherent monolayer of cells. 1. Mount a calibrated (see Note 1) cantilever on the AFM fluid holder and place a clean glass slide (or culture dish; see Note 2) on the stage for use in calibrating the photodetector deflection sensitivity; place the scanner on the stage (see Note 3) and aim the laser at the end of the cantilever such that adequate return voltage is obtained on the position sensor. At this point, allow the instruments (mercury lamp, AFM controller and sensor) to

Atomic Force Microscopy in Mechanobiology

313

equilibrate for 30–60 min prior to measurements in order to minimize thermal fluctuations (see Note 4). 2. Once the equipment reaches a steady-state temperature, ­pinpoint a reasonable laser location on the cantilever such that fluid drag and motion artifacts are minimized. To achieve this, perform continuous noncontact indentations away from the substrate surface while slowly adjusting the location of the laser over the cantilever, until the shape of the indentation curve is as flat as possible (97). 3. Determine cantilever deflection sensitivity, i.e. the conversion coefficient (nm/V) between the position of the laser on the AFM sensor (i.e., voltage signal on the photodetector) and the cantilever deflection (see Note 5). Engage on glass and indent once. Since glass is infinitely stiff compared to the cantilever, 100% of the piezo extension should be converted to deflection; therefore, adjust deflection sensitivity such that the post-contact slope of indentation is 1 nm/nm (meaning for every nanometer of extension, the cantilever deflects a nanometer). Repeat calibration on another distant spot on glass and confirm the calibration value. 4. Prepare cells for indentation (see Note 6). 5. At this point, the protocol can vary depending on the goal of the investigator. Below, two separate protocol suggestions are outlined focusing on two biomedical indentation tests quantifying either aggregate or local cell mechanical properties. 3.1.1. Measuring Aggregate Elastic Properties

One of the most straightforward and accurate ways to utilize cell mechanical properties in biomedical applications involves rapid characterization of aggregate properties of cells isolated from patients. Such a method would be easy to replicate and standardize, and it would involve little technological challenge to implement as a clinical diagnostic method. It should be noted that a global mechanical test would involve compression, not indentation because the latter would mostly probe local properties. Keep in mind that for compression testing of aggregate cell mechanical properties, a large surface area AFM probe is required. Ideally a flat surface is needed, however, a large spherical particle (>10–20 mm diameter) may be used with some additional computational burden during post-processing. 1. If working with cultured cells, gently trypsinize cells, rinse and replate on a culture dish (or other work substrate). Place cells back in the incubator for preliminary adhesion to the dish. Ensure that cells are adhered but not spread because their spherical shape will be critical for proper analysis (see Note 7). If working with nonadherent blood cells, use polyl-lysine coated glass or microfabricated wells to secure cells (see Note 8).

314

Azeloglu and Costa

2. Place cells on the AFM stage and target an adhered spherical cell (see Note 9). 3. Approach within ~100–200  mm of the cell using manual control. 4. Once appropriate indentation parameters are set (e.g. indentation frequency, piezo range, etc.), engage onto the cell and approach manually until the probe-cell contact point occurs at about 30–50% of the extension range (which will vary from one AFM to another); adequate pre-contact range is critical to determine the contact point with an objective algorithm (Fig. 2b). Perform an indentation protocol (see Note 10). 5. Repeat measurements noting the contact point (see Note 11) in order to visually assess whether the cell was permanently deformed during the first indentation, and to ensure the results appear repeatable. 6. Withdraw cantilever, remove AFM head, and fix cells for subsequent fluorescent staining (see Note 12). 3.1.2. Dynamic AFM Elastography for Subcellular Biomechanics

Instead of a single global stiffness value, another approach for assessing cell mechanics involves characterization of subcellular properties. Under this scheme, mechanical properties can be mapped out with high resolution, and results can be correlated with the underlying cytoskeleton or organelles, either live or during post-processing. This method can also have tremendous diagnostic value as a biomedical test, since it involves spatial mapping that can give mechanistic information regarding cell behavior, such as local changes in mechanics during motility (67). It should be noted that the probe to be used for local cell mechanical ­measurements must be relatively small, preferably a blunted-­ pyramidal indenter or a small spherical particle (£5 mm diameter). 1. Place cells on the AFM stage and select an appropriate cell (see Note 13) for indentation. 2. Approach within ~100–200  mm of the cell using manual control. 3. Once appropriate indentation parameters are set, engage onto the cell and approach manually until initial contact with the cell occurs at about 75% of the extension range, yielding a deflection–extension curve that shows clear pre- and post-contact regions. Make sure to obtain a phasecontrast (and/or fluorescence) image before the first indentation such that the physical location of the array can be correlated with the underlying cellular structures during post-processing (see Note 14). 4. Perform an array of indentations with a set spacing between indentations. The density and breadth of this indentation

Atomic Force Microscopy in Mechanobiology

315

array could vary depending on the goal of the study. For example, if the mechanical contribution of a relatively small organelle (~200 nm) is being probed, a high-density array of 21 × 21 indentations with 100-nm spacing could be used. Conversely, a larger 11 × 11 array with 2-mm spacing would be more suitable to compare nuclear and cytoplasmic mechanics. Other critical indentation parameters should also be empirically determined (see Note 15). 5. Withdraw cantilever, remove AFM head, and fix cells. 6. At this point, cells can be permeabilized and stained for markers of interest for spatial correlation. 3.2. Data Analysis

The equations that govern contact between two smooth elastic bodies were derived in the late nineteenth century by Heinrich Hertz (113). Therefore, even though the contact equations for a rigid conical indenter (a simplified approximation of the most commonly used AFM probe geometry) were derived by Love (114), the general elastic contact solutions are collectively referred to as “Hertzian theory.” For AFM applications, theoretical assumptions of Hertzian theory typically include a rigid indenter with ideal geometry and a smooth, flat, nonadhesive, and infinitely thick substrate. Also, the substrate is assumed to have homogeneous, isotropic, and linear elastic mechanical properties. Despite such theoretical limitations and over 60  years of additional improvements, simple “Hertzian theory” is still used by numerous investigators to characterize cell mechanical properties. In this manuscript, some of the non-Hertzian approaches will be also reviewed as successful alternatives to simplified contact equations.

3.2.1. Quantitative Analysis of AFM Indentation

In AFM indentation, the indentation depth is proportional to the force applied to the rigid indenter, and inversely proportional to the “stiffness” of the substrate. The elasticity of the cell is the desired unknown variable in the AFM indentation problem, while force and depth are measured experimentally (Fig. 2c). Indentation force (F ) is linearly proportional to the deflection of the cantilever (h); if we assume perfect bending, the relationship is analogous to that of an elastic spring, and is simply governed by the spring constant of the cantilever (k):



F = k(h − h0 )

(1)

where the contact point (z0,h0) must be accurately identified to define the origin of the deflection–extension curve, as described below. The depth of indentation (D) is then easily found from the difference between the net z-extension and the probe deflection relative to this origin:

D = (z − z0 ) − (h − h0 )

(2)

316

Azeloglu and Costa



A number of methods have been designed to estimate the contact point. While some of these methods rely on visual inspection (115), other more quantitative methods may use the derivative of the force curve (116), alterations in thermal fluctuations (117) or search algorithms that include contact point as an unknown parameter (118). Another method that can successfully detect the contact point even under low signal-to-noise conditions involves fitting a bi-domain polynomial function to the raw force curve (97). Regardless of the chosen method, repeatable and objective characterization of contact point is critical in accurate determination of total contact area, thus the apparent modulus (see Note 16). Therefore, it is highly recommended to develop or adopt a quantitative user-independent contact point estimation method prior to initiating any AFM indentation experiments. The details of our bi-domain fitting method are provided in Note 17. According to Hertzian contact theory, knowing force and depth, one can compute the stiffness of the underlying elastic substrate using the idealized geometry of the contact area (119). This geometric relationship can conveniently be generalized in the following form: E (3) F (D) = 2π · · f (D) 2 1 − υ2



where f(D) is a geometric function, and E and n are Young’s modulus of elasticity and Poisson’s ratio of the underlying substrate, respectively. Notice how both Poisson’s ratio and the elastic modulus come into the indentation problem. This is due to an inherently complex deformation field that causes both compressive (axially) and tensile (laterally) stresses within the substrate. As a result, the force–depth relationship is often written in a shorter form: (4) F (D) = 2π · Eˆ · f (D )

(

)

using an aggregate term named apparent elastic modulus, which subsumes both the Young’s modulus of elasticity and Poisson’s ratio. The geometric function, f(D), is the mathematical relationship that defines the contact between the rigid indenter and the substrate. A list of different force equations for various indenter shapes is given in Table 2. The two most commonly used contact geometries in biomechanical indentation studies are the sphere and blunted pyramid. The geometry of the standard blunted pyramid is approximated with its blunted cone equivalent (120), which has a complex contact relationship (Table 2). The governing equation for its contact radius is given as the following:

D+

a (D ) R

·

(

)

a (D )  π  b  2 2 a (D ) − b − a (D ) − ·  − arcsin  = 0.     tan θ  2 a (D ) 

(5)

317

Atomic Force Microscopy in Mechanobiology

Table 2 Indentation force for some of the common indentation geometries Indenter geometry Cone

Indentation force versus indentation depth F (D) =

4 ˆ 2 ED tan θ π

Sphere

F (D ) =

8 ˆ E D 3R 3

Blunted cone

For D ≤

θ R

b2 b2 use the spherical form. For D > use: R R

 a2  π  b  F (D ) = 4Eˆ aD −  2 − arcsin  a   2 tan q  − Cylinder

Cross section

θ R

b

  b a a −b  + + a2 − b2  3R  2 tan q 3R   3

2

2

ˆ F (D) = 4ERD R

The required parameters for characterization of the force–depth relationship are: F indentation force, D indentation depth, R indenter radius for sphere and cylinder, defect radius of curvature for the blunted cone, q cone semi-angle, b size of the defect for the blunted cone, a contact radius, and Eˆ apparent modulus of elasticity. Note that the blunted cone indenter follows the spherical contact regime until indentation depth overcomes the cone defect region, which has a radius of curvature of R

Note that contact radius, a(D), in this expression cannot be solved analytically, but must be computed iteratively (121). Once a depth-dependent expression for the contact radius is obtained, the apparent modulus in Eq.  4 can be solved in a least-squares fashion as a linear coefficient relating the contact radius (times depth) to indentation force, with a constrained zero-intercept. Such is the classical approach for solving Hertzian elastic contact, which assumes homogeneous, isotropic, and linear elastic properties of the underlying semi-infinite material. While this traditional method may be applicable for panoply of materials, it is not generally suitable for complex or heterogeneous substrates such as cells (Fig.  1). Instead, a non-Hertzian modification of Eq. 4 may be used to deduce depth-dependent elastic properties, such that each instantaneous depth and force datum can be used to compute a “pointwise apparent modulus” value for that given indentation depth:

Eˆ pw (D) =

F (D) 2π · f (D)

(6)

Azeloglu and Costa

a

b

2

Force (nN)

318

1

Apparent Modulus (kPa)

0 4 3 2 1 0 0

200

400

600

800

1000

Indentation Depth (nm)

Fig. 3. Changes in depth-dependent material properties can be quite dramatic for mechanically heterogeneous objects such as cells. (a) An alveolar type-II epithelial cell has a number of intracellular structures that can influence its apparent local mechanical properties, such as surfactant-filled lamellar bodies (green) or the nucleus (blue). (b) AFM indentations on such structures (e.g. arrow in panel (a)) result in complex patterns of force and pointwise modulus versus indentation depth. Traditional fitting algorithms using Hertzian contact mechanics (black dashed lines) neglect such depth-dependent heterogeneities, limiting the information obtained from AFM indentation tests.

This pointwise method does not assume homogeneous material properties (120), and it is capable of revealing depth-dependent changes in apparent modulus (Fig. 3). It may also offer the potential to fit the resultant modulus profile for characterization of nonlinear material properties with a phenomenological model. Under such a scheme, the pointwise apparent modulus would be computed as a function of indentation depth, and an inverse finite element model with appropriate material configuration would be used to recreate the measured modulus response. This was recently demonstrated to be an accurate method to quantify viscoelastic properties of isolated cells with micropipette aspiration (122). 3.2.2. Analyzing Whole-Cell Unconfined Compression

Aggregate elastic properties of a cell can be analyzed using the above indentation analysis scheme with little modification. For a flat-punch indenter that can fully compress the cell, analysis is straightforward since this scenario involves simple spherical contact as noted in Table 2 (where the sphere in this case is the cell, not the indenter). Contact conditions should be modeled with care, especially if a spherical indenter is being used (123, 124).

Atomic Force Microscopy in Mechanobiology

319

4. Notes 1. While AFM cantilevers arrive from the manufacturer with a nominal range of stiffness values, there can be significant lotto-lot variation; therefore, it is critical to determine the exact stiffness of a given batch (125). A number of well-established methods can be utilized (126–128). Most cantilever manufacturers and retailers can supply pre-calibrated cantilevers as well, albeit for an extra fee. If there are no available calibration equipment or experts on-site, it is recommended to send a few of the cantilevers for calibration to determine a median stiffness value of a given batch. Since batches of cantilevers are manufactured in bulk, large deviations within a lot are not expected (129). For most cell mechanical measurements, the reader is encouraged to start experiments with softer cantilevers at first. While ultrasoft cantilevers (k = 0.01–0.1  N/m) will be more suitable for most squamous cells and neurons, moderately stiff cantilevers (k = 0.1–1 N/m) may be preferred for stiffer cells such as cardiac myocytes. Please note that the deflection of the cantilever (i.e. force) will be proportional to contact surface area; therefore, customized probes with large spherical particles may need to be stiffer than their unmodified counterparts. 2. Most commercial AFM systems are equipped to handle glass slides and various culture dishes as work substrates. 3. Prior to forming a meniscus between the cantilever and the calibration surface, add a drop of water over both surfaces to prevent trapping of air bubbles behind the cantilever. 4. If working with live cells at 37°C, make sure that all the ­calibration steps are performed at 37°C as well, since most commercial AFM cantilevers are temperature-sensitive (reflective top coating usually has a higher thermal expansion coefficient than the rest of the cantilever). It is critical to have a positive voltage on the optical sensor during the warm-up of the AFM head such that these circuits equilibrate as well. 5. Most modern AFM systems have closed-loop controllers that correct for piezo hysteresis; if an open-loop system is being utilized, it is highly recommended to stay within the middle of the piezo extension limits to minimize nonlinearities in the calibration curve. 6. Cells should be cultured on a stiff, flat substrate such as glass or polystyrene. Surface treatments (such as coating with extracellular matrix proteins) may be used to improve cell adhesion, but may also affect apparent elastic properties (19). If subcellular properties are being interrogated and spatial

320

Azeloglu and Costa

correlation of indentation sites with specific proteins is required, commercially available glass coverslip-bottomed culture dishes allow compatibility with fluorescent labeling and high-magnification oil-immersion objectives. Before starting an indentation experiment, ensure that critical calibration parameters such as the location of the laser on the cantilever, noncontact slope, and the return photovoltaic voltage on the AFM sensor are unaltered. Record these parameters, including an image of the laser location on the cantilever, for future reference. If cells are being probed under regular atmospheric conditions, culture medium could be replaced with CO2-independent media to control the pH. Check for temperature fluctuations; if performing experiments at room temperature, ensure that sample temperature is equilibrated at least 20–30 min prior to the experiment. 7. Exact timing for “preliminary plating” depends on cell type. While some cells, such as fibroblasts, adhere and spread rapidly, others may take longer time. A multi-well assay with 5-min increments (15–60 min) can be performed to determine the optimal time for the specific cell type of interest. Twenty to thirty minutes of preliminary incubation for primary fibroblasts, chondrocytes, and smooth muscle cells results in consistent adhesion with a well-preserved spherical geometry. 8. Theoretically, compression of cells for quantification of aggregate modulus can also be carried out on spread cells. However, these measurements may require additional post-processing to correct for the substrate effects due to the thin cellular layer relative to the probe tip diameter (130). This may further require finite element modeling for quantitative analysis of mechanical properties (91). 9. Microfabricated surfaces, such as nanowells, can be utilized to immobilize nonadherent cells during the indentation protocol (131). 10. The exact parameters for a successful elastic test would depend on the cell type, thus it should be iteratively determined by the user. It is suggested that overall compression of the ­spherical cell be limited to 10% of the cell diameter to ensure applicability of the Hertzian contact equations. For other parameters see Note 15. 11. Determination of the exact contact point is a challenging task. It is recommended that an automated, objective algorithm be used for this purpose; such a protocol is outlined in Note 17. 12. While it may seem trivial, it can be difficult to relocate the exact position of an indentation experiment on a glass slide (or culture dish) once it has been removed from the microscope stage.

Atomic Force Microscopy in Mechanobiology

321

Make sure that the location is marked by some natural pointer (such as a pattern of cells that are recognized at low magnification), or with a permanent marker or scratch mark on the bottom of the substrate (i.e. culture dish). Acquisition of an image of the cell with the AFM tip in view will help relocate the target cell on the microscope after it has been fixed and stained. 13. Appropriate cell selection is crucial. While functional expression of key molecules may seem to be the most important para­ meter, it is critical to consider others such as size, morphology­, and state of health. When selecting a healthy control sample for example, the user must ensure that the cell size is within a reasonable range, and that there are no obvious signs of functional problems such as excessive membrane blebbing, misshapen nuclei, or cellular detachment, which may suggest apoptotic behavior. 14. Most modern biological AFM systems are now equipped with fluorescent microscopes and control software that are capable of matching the location of an optical image with AFM data; for older models, an optical image with the engaged AFM tip is necessary to manually locate the array in relation to the cell. 15. Other indentation parameters may include the rate of indentation, size of the extension range, and whether to include automatic trigger (and if so, the magnitude of such a trigger). These parameters should be determined empirically for each sample; here are some guidelines for parameter selection. While faster indentations may improve experimental time, it may be detrimental to cells. It is not recommended to deform cells at rates higher than 20–30 mm/s. Optimal magnitude of extension also depends on the application; smaller extensions may be more suitable for fine structures (<1  mm), whereas large inclusions or aggregate properties may be easier to resolve with deeper indentations (3–6  mm). Regardless of extension size, indentation depth should not surpass 20% of overall sample size to avoid substrate-related stiffening artifacts. Trigger option could simplify array construction by changing the start-up location of piezo extension; however, it should be noted that automatic triggers in commercial AFM scanners perform extra indentations to determine an appropriate location. These extra indentations may not be suitable for certain applications, such as timed indentations or for cells that are particularly sensitive to mechanical stimulation. 16. Inaccuracies of the selected contact coordinates may lead to substantial errors in the calculated modulus. In order to demonstrate, five different points are selected as the “contact point” on a representative force curve obtained on a primary smooth muscle cell. The resultant depth-dependent pointwise modulus patterns were substantially different (Fig. 4).

322

Azeloglu and Costa

a

500

Deflection (nm)

400 300 200 100 0 0

1000

2000

3000

4000

5000

b

35

Apparent Modulus (kPa)

Extension (nm)

30

Point-1 Point-2 Point-3

25

Point-4 Point-5

20 15 10 5 0

0

200

400

600

800

1000

Indentation Depth (nm)

Fig. 4. The effect of contact point selection on computed modulus of a primary rat aortic smooth muscle cell. (a) Five different points, separated by 100 nm of extension, are demarcated on a force curve within the vicinity of contact (identified as the abrupt increase in deflection). (b) The resultant pointwise apparent elastic modulus distributions are distinctly different, particularly at small indentation depths.

17. Contact point can be determined objectively using a bi-domain polynomial fit of the raw force curve with a linear-quadratic function (types of the two fits were determined empirically). Linear pre-contact (f1) and quadratic post-contact (f2) functions are defined as:

y 1 = a + bx1

(7)



y 2 = k + cx 2 + dx 22

(8)

where y- and x-coordinates signify the extension and deflection axes, respectively. Since a continuity constraint is enforced at the point of contact (fits for pre- and post-contact zones must ­connect at xc, with x2 ≡ x1−xc), Eq. 8 becomes:

y 2 = a + bx c + cx 2 + dx 22

(9)

Atomic Force Microscopy in Mechanobiology

323

This overdetermined problem (5 equations, 4 unknowns) can be solved with the following system of equations:

 ∑ y1   ∑ y 1x1  ∑ y2  ∑ y 2x 2  ∑ y 2 x 22

  n1     ∑ x1  =  n2    ∑ x2   ∑ x 22

∑ x1 ∑ x12 xc ∑ n 2 xc ∑ x 2 xc ∑ x 22

0 0 ∑ x2 ∑ x 22 ∑ x 23

0  0  ∑ x 22   ∑ x 23  ∑ x 24 

a  b    c    d 

(10)

where n1 and n2 are the sample sizes of the first and second ­polynomial within which all the summations of x- and y-­coordinates are performed. The error between the prescribed bi-domain function and the raw data is minimized by solving Eq. 10 in a leastsquares fashion, using the Moore-Penrose pseudo-inverse method in MatLab or similar numerical methods.

Acknowledgments The authors wish to acknowledge a NSF CAREER Award (KDC; BES-0239138) and a fellowship from the Stony Wold-Herbert Fund (EUA) for funding. References 1. Barratt, M.R. and Pool, S.L., Principles of Clinical Medicine for Space Flight. (2008), Springer: New York. p. xiv, 596 p. 2. Wolff, J. (1870) Ueber die innere Architectur der Knochen und ihre Bedeutung für die Frage vom Knochenwachsthum. Virchows Archiv, 50(3): p. 389–450. 3. Fung, Y.C. (1990) Biomechanics: Motion, Flow, Stress, and Growth. New York: Springer-Verlag. xii, 569. 4. Trepat, X., Puig, F., Gavara, N., Fredberg, J.J., Farre, R., and Navajas, D. (2006) Effect of stretch on structural integrity and micromechanics of human alveolar epithelial cell monolayers exposed to thrombin. Am J Physiol Lung Cell Mol Physiol, 290(6): p. L1104–10. 5. Liu, M., Xu, J., Tanswell, A.K., and Post, M. (1993) Stretch-induced growth-promoting activities stimulate fetal rat lung epithelial cell proliferation. Exp Lung Res, 19(4): p. 505–17. 6. Ilizarov, G.A. (1989) The tension-stress effect on the genesis and growth of tissues. Part I. The influence of stability of fixation and soft-tissue preservation. Clin Orthop Relat Res, (238): p. 249–81. 7. Isenberg, B.C. and Tranquillo, R.T. (2003) Long-term cyclic distention enhances the mechanical

properties of collagen-based media-equivalents. Ann Biomed Eng, 31(8): p. 937–49. 8. Lim, C.T., Zhou, E.H., and Quek, S.T. (2006) Mechanical models for living cells – a review. J Biomech, 39(2): p. 195–216. 9. Discher, D.E., Janmey, P., and Wang, Y.L. (2005) Tissue cells feel and respond to the stiffness of their substrate. Science, 310(5751): p. 1139–43. 10. Vogel, V. and Sheetz, M. (2006) Local force and geometry sensing regulate cell functions. Nat Rev Mol Cell Biol, 7(4): p. 265–75. 11. Jiang, G., Huang, A.H., Cai, Y., Tanase, M., and Sheetz, M.P. (2006) Rigidity sensing at the leading edge through alphavbeta3 integrins and RPTPalpha. Biophys J, 90(5): p. 1804–9. 12. Bischofs, I.B. and Schwarz, U.S. (2003) Cell organization in soft media due to active mechanosensing. Proc Natl Acad Sci U S A, 100(16): p. 9274–9. 13. Patel, A.J., Lazdunski, M., and Honore, E. (2001) Lipid and mechano-gated 2P domain K(+) channels. Curr Opin Cell Biol, 13(4): p. 422–8. 14. Baneyx, G., Baugh, L., and Vogel, V. (2002) Fibronectin extension and unfolding within cell matrix fibrils controlled by cytoskeletal tension. Proc Natl Acad Sci USA, 99(8): p. 5139–43.

324

Azeloglu and Costa

15. Smith, M.L., Gourdon, D., Little, W.C., Kubow, K.E., Eguiluz, R.A., Luna-Morris, S., and Vogel, V. (2007) Force-induced unfolding of fibronectin in the extracellular matrix of living cells. PLoS Biol, 5(10): p. e268. 16. Hall, A. (1998) Rho GTPases and the actin cytoskeleton. Science, 279(5350): p. 509–14. 17. Choquet, D., Felsenfeld, D.P., and Sheetz, M.P. (1997) Extracellular matrix rigidity causes strengthening of integrin-cytoskeleton linkages. Cell, 88(1): p. 39–48. 18. Icard-Arcizet, D., Cardoso, O., Richert, A., and Henon, S. (2008) Cell stiffening in response to external stress is correlated to actin recruitment. Biophys J. 19. Takai, E., Costa, K.D., Shaheen, A., Hung, C.T., and Guo, X.E. (2005) Osteoblast elastic modulus measured by atomic force microscopy is substrate dependent. Ann Biomed Eng, 33(7): p. 963–71. 20. Judex, S., Zhong, N., Squire, M.E., Ye, K., Donahue, L.R., Hadjiargyrou, M., and Rubin, C.T. (2005) Mechanical modulation of molecular signals which regulate anabolic and catabolic activity in bone tissue. J Cell Biochem, 94(5): p. 982–94. 21. Engler, A.J., Sen, S., Sweeney, H.L., and Discher, D.E. (2006) Matrix elasticity directs stem cell lineage specification. Cell, 126(4): p. 677-89. 22. Jaynes, J. (1990) The origin of consciousness in the breakdown of the bicameral mind. Boston: Houghton Mifflin. 491 p. 23. Pelling, A.E. and Horton, M.A. (2008) An historical perspective on cell mechanics. Pflugers Arch, 456(1): p. 3–12. 24. Kite, G.L. and Chambers, R., Jr. (1912) Vital staining of chromosomes and the function and structure of the nucleus. Science, 36(932): p. 639–641. 25. Terreros, D.A. and Grantham, J.J. (1982) Marshall Barber and the origins of micropipette methods. Am J Physiol, 242(3): p. F293–6. 26. Seifriz, W. (1920) Viscosity values of protoplasm as determined by microdissection. Botanical Gazette, 70(5): p. 360–386. 27. Seifriz, W. (1918) Observations on the structure of protoplasm by aid of microdissection. Biological Bulletin, 34(5): p. 307–324. 28. Carlson, J.G. (1946) Protoplasmic viscosity changes in different regions of the grasshopper neuroblast during mitosis. Biological Bulletin, 90(2): p. 109–121. 29. Lillie, R.S. (1912) The physiological significance of the segmented structure of the striated muscle fiber. Science, 36(921): p. 247–255. 30. Kite, G.L. (1913) Studies on the physical properties of protoplasm I: The physical properties of

the protoplasm of certain animal and plant cells. Am J Physiol, 32(2): p. 146–164. 31. Matzke, R., Jacobson, K., and Radmacher, M. (2001) Direct, high-resolution measurement of furrow stiffening during division of adherent cells. Nat Cell Biol, 3(6): p. 607–10. 32. Akiyama, N., Ohnuki, Y., Kunioka, Y., Saeki, Y., and Yamada, T. (2006) Transverse stiffness of myofibrils of skeletal and cardiac muscles studied by atomic force microscopy. J Physiol Sci, 56(2): p. 145–51. 33. Northen, H.T. (1950) Alterations in the structural viscosity of protoplasm by colchicine and their relationship to c-mitosis and c-tumor formation. Am J Botany, 37(9): p. 705–711. 34. Rosenberg, M.D. (1963) The relative extensi­ bility of cell surfaces. J Cell Biol, 17: p. 289–97. 35. Weiss, P. and Garber, B. (1952) Shape and movement of mesenchyme cells as functions of the physical structure of the medium: contributions to a quantitative morphology. Proc Natl Acad Sci U S A, 38(3): p. 264–80. 36. Mitchison, J.M. and Swann, M.M. (1954) The mechanical properties of the cell surface .1. The cell elastimeter. J Exp Biol, 31(3): p. 443–&. 37. Cave, I.D. (1968) The anisotropic elasticity of the plant cell wall. Wood Sci Technol, 2(4): p. 268–278. 38. Chien, S., Usami, S., Dellenback, R.J., and Gregersen, M.I. (1967) Blood Viscosity: Influence of Erythrocyte Deformation. Science, 157(3790): p. 827–829. 39. Fung, Y.C. and Tong, P. (1968) Theory of the sphering of red blood cells. Biophys J, 8(2): p. 175–98. 40. Skalak, R., Tozeren, A., Zarda, R.P., and Chien, S. (1973) Strain energy function of red blood cell membranes. Biophys J, 13(3): p. 245–64. 41. Dong, C., Skalak, R., Sung, K.L., SchmidSchonbein, G.W., and Chien, S. (1988) Passive deformation analysis of human leukocytes. J Biomech Eng, 110(1): p. 27–36. 42. Heilbrunn, L.V. (1920) The physical effect of anesthetics upon living protoplasm. Biological Bulletin, 39(6): p. 307–315. 43. Chien, S., Usami, S., and Bertles, J.F. (1970) Abnormal rheology of oxygenated blood in sickle cell anemia. J Clin Invest, 49(4): p. 623–34. 44. Costa, K.D. (2003) Single-cell elastography: probing for disease with the atomic force microscope. Dis Markers, 19(2-3): p. 139–54. 45. Rosenbluth, M.J., Lam, W.A., and Fletcher, D.A. (2008) Analyzing cell mechanics in hematologic diseases with microfluidic biophysical flow cytometry. Lab Chip, 8(7): p. 1062–70.

Atomic Force Microscopy in Mechanobiology 46. Bao, N., Zhan, Y., and Lu, C. (2008) Microfluidic electroporative flow cytometry for studying single-cell biomechanics. Anal Chem, 80(20): p. 7714–9. 47. Discher, D.E., Mooney, D.J., and Zandstra, P.W. (2009) Growth factors, matrices, and forces combine and control stem cells. Science, 324(5935): p. 1673–7. 48. Janmey, P.A., Winer, J.P., Murray, M.E., and Wen, Q. (2009) The hard life of soft cells. Cell Motil Cytoskeleton, 66(8): p. 597–605. 49. Zaman, M.H., Trapani, L.M., Sieminski, A.L., Mackellar, D., Gong, H., Kamm, R.D., Wells, A., Lauffenburger, D.A., and Matsudaira, P. (2006) Migration of tumor cells in 3D matrices is governed by matrix stiffness along with cell-matrix adhesion and proteolysis. Proc Natl Acad Sci U S A, 103(29): p. 10889–94. 50. Engler, A.J., Griffin, M.A., Sen, S., Bonnemann, C.G., Sweeney, H.L., and Discher, D.E. (2004) Myotubes differentiate optimally on substrates with tissue-like stiffness: pathological implications for soft or stiff microenvironments. J Cell Biol, 166(6): p. 877–87. 51. Na, S., Sun, Z., Meininger, G.A., and Humphrey, J.D. (2004) On atomic force microscopy and the constitutive behavior of living cells. Biomech Model Mechanobiol, 3(2): p. 75–84. 52. Ni, Y. and Chiang, M.Y.M. (2007) Prediction of elastic properties of heterogeneous materials with complex microstructures. J Mech Phys Solids, 55(3): p. 517–532. 53. Guilak, F., Tedrow, J.R., and Burgkart, R. (2000) Viscoelastic properties of the cell nucleus. Biochem Biophys Res Commun, 269(3): p. 781–6. 54. Lammerding, J., Schulze, P.C., Takahashi, T., Kozlov, S., Sullivan, T., Kamm, R.D., Stewart, C.L., and Lee, R.T. (2004) Lamin A/C deficiency causes defective nuclear mechanics and mechanotransduction. J Clin Invest, 113(3): p. 370–8. 55. Berdyyeva, T.K., Woodworth, C.D., and Sokolov, I. (2005) Human epithelial cells increase their rigidity with ageing in  vitro: direct measurements. Phys Med Biol, 50(1): p. 81–92. 56. Broers, J.L.V., Peeters, E.A.G., Kuijpers, H.J.H., Endert, J., Bouten, C.V.C., Oomens, C.W.J., Baaijens, F.P.T., and Ramaekers, F.C.S. (2004) Decreased mechanical stiffness in LMNA-/- cells is caused by defective nucleocytoskeletal integrity: implications for the development of laminopathies. Hum. Mol. Genet., 13(21): p. 2567–2580.

325

57. Mathur, A.B., Collinsworth, A.M., Reichert, W.M., Kraus, W.E., and Truskey, G.A. (2001) Endothelial, cardiac muscle and skeletal ­muscle exhibit different viscous and elastic properties as determined by atomic force microscopy. J Biomech, 34(12): p. 1545–53. 58. Ricci, D., Tedesco, M., and Grattarola, M. (1997) Mechanical and morphological ­properties of living 3  T6 cells probed via ­scanning force microscopy. Microsc Res Tech, 36(3): p. 165–71. 59. Roca-Cusachs, P., Alcaraz, J., Sunyer, R., Samitier, J., Farre, R., and Navajas, D. (2008) Micropatterning of single endothelial cell shape reveals a tight coupling between nuclear ­volume in G1 and proliferation. Biophys J, 94(12): p. 4984–95. 60. Azeloglu, E.U., Bhattacharya, J., and Costa, K.D. (2008) Atomic force microscope ­elastography reveals phenotypic differences in ­alveolar cell stiffness. J Appl Physiol, 105(2): p. 652–61. 61. Darling, E.M., Topel, M., Zauscher, S., Vail, T.P., and Guilak, F. (2007) Viscoelastic properties of human mesenchymally-derived stem cells and primary osteoblasts, chondrocytes, and adipocytes. J Biomech. 62. A-Hassan, E., Heinz, W.F., Antonik, M.D., D’Costa, N.P., Nageswaran, S., Schoenenberger, C.A., and Hoh, J.H. (1998) Relative microelastic mapping of living cells by atomic force microscopy. Biophys J, 74(3): p. 1564–78. 63. Costa, K.D., Sim, A.J., and Yin, F.C. (2006) Non-Hertzian approach to analyzing mechanical properties of endothelial cells probed by atomic force microscopy. J Biomech Eng, 128(2): p. 176–84. 64. Haga, H., Sasaki, S., Kawabata, K., Ito, E., Ushiki, T., and Sambongi, T. (2000) Elasticity mapping of living fibroblasts by AFM and immunofluorescence observation of the cytoskeleton. Ultramicroscopy, 82(1-4): p. 253–8. 65. Laurent, V.M., Kasas, S., Yersin, A., Schaffer, T.E., Catsicas, S., Dietler, G., Verkhovsky, A.B., and Meister, J.J. (2005) Gradient of rigidity in the lamellipodia of migrating cells revealed by atomic force microscopy. Biophys J, 89(1): p. 667–75. 66. Nagayama, K., Nagano, Y., Sato, M., and Matsumoto, T. (2006) Effect of actin filament distribution on tensile properties of smooth muscle cells obtained from rat thoracic aortas. J Biomech, 39(2): p. 293-301. 67. Park, S., Koch, D., Cardenas, R., Kas, J., and Shih, C.K. (2005) Cell motility and local ­viscoelasticity of fibroblasts. Biophys J, 89(6): p. 4330–42.

326

Azeloglu and Costa

68. Smith, B.A., Tolloczko, B., Martin, J.G., and Grutter, P. (2005) Probing the viscoelastic behavior of cultured airway smooth muscle cells with atomic force microscopy: stiffening induced by contractile agonist. Biophys J, 88(4): p. 2994–3007. 69. Trepat, X., Grabulosa, M., Buscemi, L., Rico, F., Farre, R., and Navajas, D. (2005) Thrombin and histamine induce stiffening of alveolar epithelial cells. J Appl Physiol, 98(4): p. 1567–74. 70. Wakatsuki, T., Schwab, B., Thompson, N.C., and Elson, E.L. (2001) Effects of cytochalasin D and latrunculin B on mechanical properties of cells. J Cell Sci, 114(Pt 5): p. 1025–36. 71. Wu, H.W., Kuhn, T., and Moy, V.T. (1998) Mechanical properties of L929 cells measured by atomic force microscopy: effects of anticytoskeletal drugs and membrane crosslinking. Scanning, 20(5): p. 389–97. 72. Nishimura, S., Nagai, S., Katoh, M., Yamashita, H., Saeki, Y., Okada, J., Hisada, T., Nagai, R., and Sugiura, S. (2006) Microtubules modulate the stiffness of cardiomyocytes against shear stress. Circ Res, 98(1): p. 81–7. 73. Tsutsui, H., Ishihara, K., and Cooper, G.t. (1993) Cytoskeletal role in the contractile dysfunction of hypertrophied myocardium. Science, 260(5108): p. 682–7. 74. Hu, S., Chen, J., and Wang, N. (2004) Cell spreading controls balance of prestress by microtubules and extracellular matrix. Front Biosci, 9: p. 2177–82. 75. Shah, S.B., Davis, J., Weisleder, N., Kostavassili, I., McCulloch, A.D., Ralston, E., Capetanaki, Y., and Lieber, R.L. (2004) Structural and functional roles of desmin in mouse skeletal muscle during passive deformation. Biophys J, 86(5): p. 2993–3008. 76. Deng, L., Fairbank, N.J., Fabry, B., Smith, P.G., and Maksym, G.N. (2004) Localized mechanical stress induces time-dependent actin cytoskeletal remodeling and stiffening in cultured airway smooth muscle cells. Am J Physiol Cell Physiol, 287(2): p. C440–8. 77. Yourek, G., Hussain, M.A., and Mao, J.J. (2007) Cytoskeletal changes of mesenchymal stem cells during differentiation. Asaio J, 53(2): p. 219–28. 78. Stamenovic, D., Mijailovich, S.M., TolicNorrelykke, I.M., Chen, J., and Wang, N. (2002) Cell prestress. II. Contribution of microtubules. Am J Physiol Cell Physiol, 282(3): p. C617–24. 79. Helmke, B.P., Goldman, R.D., and Davies, P.F. (2000) Rapid displacement of vimentin intermediate filaments in living endothelial cells exposed to flow. Circ Res, 86(7): p. 745–52.

80. Martens, J.C. and Radmacher, M. (2008) Softening of the actin cytoskeleton by inhibition of myosin II. Pflugers Arch, 456(1): p. 95–100. 81. Schafer, A. and Radmacher, M. (2005) Influence of myosin II activity on stiffness of fibroblast cells. Acta Biomater, 1(3): p. 273–80. 82. Heidemann, S.R., Lamoureaux, P., and Buxbaum, R.E. (2000) Opposing views on tensegrity as a structural framework for understanding cell mechanics. J Appl Physiol, 89(4): p. 1670–1678. 83. Ingber, D.E. (2000) Opposing views on ­tensegrity as a structural framework for understanding­cell mechanics. J Appl Physiol, 89(4): p. 1663–70. 84. Hu, S., Chen, J., Fabry, B., Numaguchi, Y., Gouldstone, A., Ingber, D.E., Fredberg, J.J., Butler, J.P., and Wang, N. (2003) Intracellular stress tomography reveals stress focusing and structural anisotropy in cytoskeleton of living cells. Am J Physiol Cell Physiol, 285(5): p. C1082–90. 85. Radmacher, M., Tillamnn, R.W., Fritz, M., and Gaub, H.E. (1992) From molecules to cells: imaging soft samples with the atomic force microscope. Science, 257(5078): p. 1900–5. 86. Solon, J., Levental, I., Sengupta, K., Georges, P.C., and Janmey, P.A. (2007) Fibroblast adaptation and stiffness matching to soft elastic substrates. Biophys J, 93(12): p. 4453–61. 87. Guilak, F., Ting-Beall, H.P., Baer, A.E., Trickey, W.R., Erickson, G.R., and Setton, L.A. (1999) Viscoelastic properties of intervertebral disc cells. Identification of two biomechanically distinct cell populations. Spine, 24(23): p. 2475–83. 88. Hubmayr, R.D., Shore, S.A., Fredberg, J.J., Planus, E., Panettieri, R.A., Jr., Moller, W., Heyder, J., and Wang, N. (1996) Pharmacological activation changes stiffness of cultured human airway smooth muscle cells. Am J Physiol, 271(5 Pt 1): p. C1660–8. 89. Guck, J., Ananthakrishnan, R., Mahmood, H., Moon, T.J., Cunningham, C.C., and Kas, J. (2001) The optical stretcher: a novel laser tool to micromanipulate cells. Biophys J, 81(2): p. 767–84. 90. Shieh, A.C. and Athanasiou, K.A. (2006) Biomechanics of single zonal chondrocytes. J Biomech, 39(9): p. 1595–602. 91. Peeters, E.A., Oomens, C.W., Bouten, C.V., Bader, D.L., and Baaijens, F.P. (2005) Mechanical and failure properties of single attached cells under compression. J Biomech, 38(8): p. 1685–93. 92. Goldmann, W.H., Galneder, R., Ludwig, M., Kromm, A., and Ezzell, R.M. (1998) Differences in F9 and 5.51 cell elasticity determined by cell

Atomic Force Microscopy in Mechanobiology poking and atomic force microscopy. FEBS Lett, 424(3): p. 139–42. 93. Yamada, S., Wirtz, D., and Kuo, S.C. (2000) Mechanics of living cells measured by laser tracking microrheology. Biophys J, 78(4): p. 1736–47. 94. Puig-De-Morales, M., Grabulosa, M., Alcaraz, J., Mullol, J., Maksym, G.N., Fredberg, J.J., and Navajas, D. (2001) Measurement of cell microrheology by magnetic twisting cytometry with frequency domain demodulation. J Appl Physiol, 91(3): p. 1152–9. 95. Canetta, E., Duperray, A., Leyrat, A., and Verdier, C. (2005) Measuring cell viscoelastic properties using a force-spectrometer: influence of protein-cytoplasm interactions. Biorheology, 42(5): p. 321–33. 96. Thoumine, O. and Ott, A. (1997) Time scale dependent viscoelastic and contractile regimes in fibroblasts probed by microplate manipulation. J Cell Sci, 110 (Pt 17): p. 2109–16. 97. Costa, K.D. (2006) Imaging and probing cell mechanical properties with the atomic force microscope. Methods Mol Biol, 319: p. 331–61. 98. Binnig, G., Quate, C.F., and Gerber, C. (1986) Atomic force microscope. Phys Rev Lett, 56(9): p. 930–933. 99. Stark, R.W., Drobek, T., and Heckl, W.M. (2001) Thermomechanical noise of a free v-shaped cantilever for atomic-force microscopy. Ultramicroscopy, 86(1-2): p. 207–15. 100. Butt, H.J., Cappella, B., and Kappl, M. (2005) Force measurements with the atomic force microscope: Technique, interpretation and applications. Surf Sci Rep, 59(1-6): p. 1–152. 101. Gratz, A.J., Manne, S., and Hansma, P.K. (1991) Atomic force microscopy of atomicscale ledges and etch pits formed during ­dissolution of quartz. Science, 251(4999): p. 1343–1346. 102. Hoh, J.H. and Schoenenberger, C.A. (1994) Surface morphology and mechanical properties of MDCK monolayers by atomic force microscopy. J Cell Sci, 107 (Pt 5): p. 1105–14. 103. Moreno-Herrero, F., Colchero, J., GomezHerrero, J., and Baro, A.M. (2004) Atomic force microscopy contact, tapping, and jumping modes for imaging biological samples in liquids. Phys Rev E Stat Nonlin Soft Matter Phys, 69(3 Pt 1): p. 031915. 104. Nag, K., Munro, J.G., Hearn, S.A., Rasmusson, J., Petersen, N.O., and Possmayer, F. (1999) Correlated atomic force and transmission electron microscopy of nanotubular structures in pulmonary surfactant. J Struct Biol, 126(1): p. 1–15. 105. Davis, J.J., Hill, H.A.O., and Powell, T. (2001) High resolution scanning force microscopy of

327

cardiac myocytes. Cell Biology International, 25(12): p. 1271–1277. 106. Moloney, M., McDonnell, L., and O’Shea, H. (2004) Atomic force microscopy analysis of enveloped and non-enveloped viral entry into, and egress from, cultured cells. Ultramicroscopy, 100(3-4): p. 163–9. 107. Park, S., Costa, K.D., and Ateshian, G.A. (2004) Microscale frictional response of bovine articular cartilage from atomic force microscopy. J Biomech, 37(11): p. 1679–87. 108. Soboyejo, W.O., Nemetski, B., Allameh, S., Marcantonio, N., Mercer, C., and Ricci, J. (2002) Interactions between MC3T3-E1 cells and textured Ti6Al4V surfaces. J Biomed Mater Res, 62(1): p. 56–72. 109. Vie, V., Giocondi, M.C., Lesniewska, E., Finot, E., Goudonnet, J.P., and Le Grimellec, C. (2000) Tapping-mode atomic force microscopy on intact cells: optimal adjustment of tapping conditions by using the deflection signal. Ultramicroscopy, 82(1-4): p. 279-88. 110. de Pablo, P.J., Colchero, J., Gomez-Herrero, J., and Baro, A.M. (1998) Jumping mode scanning force microscopy. Applied Physics Letters, 73(22): p. 3300–3302. 111. Nagao, E. and Dvorak, J.A. (1999) Phase imaging by atomic force microscopy: analysis of living homoiothermic vertebrate cells. Biophys J, 76(6): p. 3289–97. 112. Schaus, S.S. and Henderson, E.R. (1997) Cell viability and probe-cell membrane interactions of XR1 glial cells imaged by atomic force microscopy. Biophys J, 73(3): p. 1205–14. 113. Hertz, H. (1881) Über die Berührung fester elastischer Körper (On the contact of elastic solids). J. Reine Angew. Mathematik, 92: p. 156–171. 114. Love, A.E.H. (1939) Boussinesq’s problem for a rigid cone. Q. J. Math., 10: p. 161–175. 115. Dimitriadis, E.K., Horkay, F., Maresca, J., Kachar, B., and Chadwick, R.S. (2002) Determination of elastic moduli of thin layers of soft material using the atomic force microscope. Biophys J, 82(5): p. 2798–810. 116. Radmacher, M. (2002) Measuring the elastic properties of living cells by the atomic force microscope. Methods Cell Biol, 68: p. 67–90. 117. Heinz, W.F., Antonik, M.D., and Hoh, J.H. (2000) Reconstructing local interaction potentials from perturbations to the thermally driven motion of an atomic force microscope cantilever. J Phys Chem B, 104(3): p. 622–626. 118. Rotsch, C., Jacobson, K., and Radmacher, M. (1999) Dimensional and mechanical dynamics of active and stable edges in motile fibroblasts investigated by using atomic force microscopy. Proc Natl Acad Sci U S A, 96(3): p. 921–6.

328

Azeloglu and Costa

119. Johnson, K.L. (1970) The correlation of indentation experiments. J Mech Phys Solids, 18(2): p. 115–126. 120. Costa, K.D. and Yin, F.C. (1999) Analysis of indentation: implications for measuring mechanical properties with atomic force microscopy. J Biomech Eng, 121(5): p. 462–71. 121. Briscoe, B.J., Sebastian, K.S., and Adams, M.J. (1994) The effect of indenter geometry on the elastic response to indentation. J Phys D: Appl Phys, 27(6): p. 1156–1162. 122. Zhao, R., Wyss, K., and Simmons, C.A. (2009) Comparison of analytical and inverse finite element approaches to estimate cell viscoelastic properties by micropipette aspiration. J Biomech, 42(16): p. 2768–2773. 123. Darling, E.M., Zauscher, S., and Guilak, F. (2006) Viscoelastic properties of zonal articular chondrocytes measured by atomic force microscopy. Osteoarthritis Cartilage, 14(6): p. 571–9. 124. Lulevich, V., Zink, T., Chen, H.Y., Liu, F.T., and Liu, G.Y. (2006) Cell mechanics using atomic force microscopy-based single-cell compression. Langmuir, 22(19): p. 8151–5. 125. Emerson, R.J.t. and Camesano, T.A. (2006) On the importance of precise calibration techniques for an atomic force microscope. Ultramicroscopy, 106(4-5): p. 413–22. 126. Butt, H.J. and Jaschke, M. (1995) Calculation of thermal noise in atomic force microscopy. Nanotechnology, 6(1): p. 1–7. 127. Cleveland, J.P., Manne, S., Bocek, D., and Hansma, P.K. (1993) A nondestructive method for determining the spring constant of cantilevers for scanning force microscopy. Rev Sci Instrum, 64(2): p. 403–405. 128. Sader, J.E. (1995) Parallel beam approxi­ mation for V-shaped atomic force microscope ­cantilevers. Rev Sci Instrum, 66(9): p. 4583–4587. 129. Senden, T.J. and Ducker, W.A. (1994) Experimental determination of spring constants in atomic force microscopy. Langmuir, 10(4): p. 1003-1004. 130. Mahaffy, R.E., Park, S., Gerde, E., Kas, J., and Shih, C.K. (2004) Quantitative analysis of the viscoelastic properties of thin regions of fibroblasts using atomic force microscopy. Biophys J, 86(3): p. 1777–93. 131. Rosenbluth, M.J., Lam, W.A., and Fletcher, D.A. (2006) Force microscopy of nonadherent cells: a comparison of leukemia cell deformability. Biophys J, 90(8): p. 2994–3003. 132. Brandão, M.M., Fontes, A., Barjas-Castro, M.L., Barbosa, L.C., Costa, F.F., Cesar, C.L., and Saad, S.T. (2003) Optical tweezers for

measuring red blood cell elasticity: application to the study of drug response in sickle cell disease. Eur J Haematol, 70(4): p. 207–11. 133. Nash, G.B., Johnson, C.S., and Meiselman, H.J. (1984) Mechanical properties of oxygenated red blood cells in sickle cell (HbSS) disease. Blood, 63(1): p. 73–82. 134. Ballas, S.K., Dover, G.J., and Charache, S. (1989) Effect of hydroxyurea on the rheological properties of sickle erythrocytes in vivo. Am J Hematol, 32(2): p. 104–11. 135. Lam, W.A., Rosenbluth, M.J., and Fletcher, D.A. (2007) Chemotherapy exposure increases leukemia cell stiffness. Blood, 109(8): p. 3505–8. 136. Suresh, S. (2007) Biomechanics and biophysics of cancer cells. Acta Biomater, 3(4): p. 413–38. 137. Guck, J., Schinkinger, S., Lincoln, B., Wottawah, F., Ebert, S., Romeyke, M., Lenz, D., Erickson, H.M., Ananthakrishnan, R., Mitchell, D., Kas, J., Ulvick, S., and Bilby, C. (2005) Optical deformability as an inherent cell marker for testing malignant transformation and metastatic competence. Biophys J, 88(5): p. 3689–98. 138. Lekka, M., Laidler, P., Gil, D., Lekki, J., Stachura, Z., and Hrynkiewicz, A.Z. (1999) Elasticity of normal and cancerous human bladder cells studied by scanning force microscopy. Eur Biophys J, 28(4): p. 312–6. 139. Li, Q.S., Lee, G.Y., Ong, C.N., and Lim, C.T. (2008) AFM indentation study of breast cancer cells. Biochem Biophys Res Commun, 374(4): p. 609–13. 140. Iyer, S., Gaikwad, R.M., Subba-Rao, V., Woodworth, C.D., and Sokolov, I. (2009) Atomic force microscopy detects differences in the surface brush of normal and cancerous cells. Nat Nanotechnol, 4(6): p. 389–93. 141. Faria, E.C., Ma, N., Gazi, E., Gardner, P., Brown, M., Clarke, N.W., and Snook, R.D. (2008) Measurement of elastic properties of prostate cancer cells using AFM. Analyst, 133(11): p. 1498–500. 142. Cross, S.E., Jin, Y.S., Rao, J., and Gimzewski, J.K. (2007) Nanomechanical analysis of cells from cancer patients. Nat Nanotechnol, 2(12): p. 780–3. 143. Lekka, M., Fornal, M., Pyka-Fosciak, G., Lebed, K., Wizner, B., Grodzicki, T., and Styczen, J. (2005) Erythrocyte stiffness probed using atomic force microscope. Biorheology, 42(4): p. 307–17. 144. Guilak, F. (2000) The deformation behavior and viscoelastic properties of chondrocytes in articular cartilage. Biorheology, 37(1-2): p. 27–44.

Atomic Force Microscopy in Mechanobiology 145. Zhang, Q.Y., Wang, X.H., Wei, X.C., and Chen, W.Y. (2008) Characterization of ­viscoelastic properties of normal and osteoarthritic chondrocytes in experimental rabbit model. Osteoarthritis Cartilage, 16(7): p. 837–40. 146. Suresh, S., Spatz, J., Mills, J.P., Micoulet, A., Dao, M., Lim, C.T., Beil, M., and Seufferlein, T. (2005) Connections between single-cell biomechanics and human disease states: gastrointestinal cancer and malaria. Acta Biomater, 1(1): p. 15–30. 147. Kol, N., Shi, Y., Tsvitov, M., Barlam, D., Shneck, R.Z., Kay, M.S., and Rousso, I. (2007) A stiffness switch in human immunodeficiency virus. Biophys J, 92(5): p. 1777-83.

329

148. Zile, M.R., Richardson, K., Cowles, M.K., Buckley, J.M., Koide, M., Cowles, B.A., Gharpuray, V., and Cooper, G., IV (1998) Constitutive properties of adult mammalian cardiac muscle cells. Circulation, 98(6): p. 567–579. 149. Lieber, S.C., Aubry, N., Pain, J., Diaz, G., Kim, S.-J., and Vatner, S.F. (2004) Aging increases stiffness of cardiac myocytes measured by atomic force microscopy nanoindentation. Am J Physiol Heart Circ Physiol, 287(2): p. H645–651. 150. Sokolov, I., Iyer, S., and Woodworth, C.D. (2006) Recovery of elasticity of aged human epithelial cells in vitro. Nanomedicine, 2(1): p. 31–6.

wwwwwwwwwwwwww

Chapter 20 Force-Clamp Measurements of Receptor–Ligand Interactions Félix Rico, Calvin Chu, and Vincent T. Moy Abstract Protein–protein interactions are the basis of both biochemical and biophysical signaling of living cells. In many cases, the receptor is present on the cell surface while the ligand is in solution or linked to another support (extracellular matrix or another cell). In the case of cellular adhesion, forces are continuously applied to receptor–ligand complexes and, as a consequence, the dissociation kinetics of the bonds may change. It is, thus, relevant to study the kinetics of protein–protein interactions in response to applied forces, as this is the most physiologically relevant situation. The atomic force microscope (AFM) was one of the first nanotools to be applied to this end. However, new approaches need to be developed to better understand the complex energy landscape of molecular interactions under applied stress. In this chapter, we described the use of the AFM to carry out force-clamp measurements on receptor–ligand bonds. Forceclamp measurements on bonds consist of applying a constant and controlled force to a receptor–ligand bond and measure the resulting dissociation lifetime. The described methods include the required materials, functionalization of tips and substrates, force-clamping measurements, and processing and interpretation of the results. An illustrative example is given with the well-studied streptavidin–biotin complex. Key words: Single molecule, Dynamic force spectroscopy, Bell model, Molecular interactions, Adhesion, Affinity, Binding strength, Force-clamp, Atomic force microscope

1. Introduction Ligand–receptor interactions play a central role in many ­biological functions, including cell adhesion, cell motility, and cell differentiation. Most of these bonds, such as integrin–ligand bonds, are subjected to mechanical forces. It is, thus, important to understand the effect of force on the bond kinetics. After the classic work by Bell suggesting how force would affect the lifetime of a biological bond, it was necessary to provide experimental evidence of bond rupture by mechanical stress (1). The first atomic force microscope (AFM) measurements to determine rupture Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_20, © Springer Science+Business Media, LLC 2011

331

332

Rico, Chu, and Moy

forces of receptor–ligand complexes were carried out more than a decade ago (2–4). Since then, single molecule force spectroscopy (SMFS) has evolved into a relatively well-established technique to study molecular interactions, one molecule at a time. Various methods have been developed to measure and interpret receptor– ligand interactions under applied force (5–10). The most used approach consists of measuring bond rupture forces by pulling the bound complexes at varying retraction speeds, which leads to varying loading rates (11, 12). The averaged rupture forces at each loading rate leads to the dynamic force spectrum of the interaction. The seminal works by Evans and coworkers allowed us to analyze force spectroscopy data in terms of transition state theory by defining a free energy potential of certain width and intrinsic dissociation rate (5, 11), which allows us to determine the effect of force on the dissociation rate (or lifetime) of the bond. In addition, recent refinement of the theory enabled us to assess the shape and depth of the free energy landscape (13–17). An alternative and elegant approach to study the effect of force on lifetime is the use of the force-clamp technique developed to study protein unfolding (18), and later applied to unbinding of receptors and ligands (19). In force-clamp measurements, the bond (protein) is held at a constant force until it unbinds (unfolds). The characteristic lifetimes are determined at different levels of clamping forces. The resulting data gives a direct measure of the force dependence of the lifetime of the bond, and it is particularly useful to detect catch-to-slip bonds (19, 20). The (strept) avidin–biotin complex was chosen as a model system of a strong molecular interaction (2–4). Even if the energy landscape of the interaction has been observed to be rather complex (21, 22), it is still a good example to illustrate the mechanisms involved in molecular interactions, being also adequate because of its easy handling and wide availability. In this chapter, we describe the methods and procedures required to probe the streptavidin–biotin interaction at the single molecule level using the force-clamp approach. We first list the necessary materials including the reagents and chemicals, the atomic force microscopy system, and the cantilevers. In the reported example, we used a commercial AFM, however, we provide some recommendations that can be useful for home built systems. Subheading 3 includes the AFM tip functionalization with streptavidin, the substrate coating with biotin, the force-clamp measurements, processing and analysis of the data, and interpretation of the results in terms of the Bell model. We have included recommendations and steps in Subheading 4 which require special care. Even if the specific protocol describes the methods to probe the streptavidin–biotin interaction in force-clamp mode, it can also be applied to different receptor–ligand systems and easily adapted for dynamic force spectroscopy measurements at varying loading rates.

Force-Clamp Measurements of Receptor–Ligand Interactions

333

2. Materials 2.1. Chemicals and Reagents

1. Gemologist loupe, magnifying glass, or surgical microscope for looking at the cantilevers.

2.1.1. General Materials

2. Fine stainless steel tweezers/forceps for handling cantilevers (see Note 2). 3. Pyrex Petri dish/similar inert vessel for treating cantilevers with acetone, acids, and other reactive reagents. 4. UV light for cleaning cantilevers. 5. Parafilm provides a clean, hydrophobic surface for coating cantilevers. 6. Alconox powder: 5% w/v Alconox solution for cleaning tools and certain substrates. 7. 24-well tissue culture plate for rinsing and treating cantilevers.

2.1.2. General Reagents

1. Blocking solution: 1% bovine serum albumin (BSA). 2. Phosphate Buffered Saline (PBS). 3. Nanopure MilliQ water. 4. HPLC grade or >95% purity acetone. 5. HPLC grade or >95% purity ethanol. 6. 3-aminopropyltriethoxysilane (3-APTES) (see Note 1).

2.1.3. Functionalization of Gold-Coated AFM Tips

1. 11-Mercaptoundecanoic acid. Store in acid cabinet. 2. 1-ethyl-3-(3-dimethlaminopropyl)carbodiimide (EDC). 3. N-hydroxysuccinimide (NHS). 4. HPLC grade ethanol.

2.1.4. Mica Substrate

1. 9.5 mm diameter V1-Mica disk or 15 mm2 or 25 mm2 square. 2. Die and punch set: for cutting round mica disks from a larger sample. This is necessary only if your mica disks/squares are too large to fit onto a glass slide. 3. Optically clear epoxy. 4. Scotch tape. 5. Glass slide/hard substrate. 6. Biotinylated BSA.

2.2. Atomic Force Microscope

An AFM consists of four major components: a cantilever, ­low-noise photodiode, laser source, and a piezoelectric element pictured in Fig. 1. A thermoelectric module or a similar temperature regulation device can also be used to regulate the temperature of the sample. In the configuration pictured in Fig. 1, the alignment of the laser with a beam splitter allows for a free optical path so that

334

Rico, Chu, and Moy beamsplitter piezo

light source

photdiode

dish

objective

cantilever

thermoelectric module

Fig.  1. Atomic force microscope. Schematic diagram of an atomic force microscope coupled to an inverted optical microscope with the relevant components: laser source, photodiode, piezoelectric element, and cantilever. The sample can be placed on a thermoelectric module for temperature control.

the AFM can be used on top of an inverted optical ­microscope, mechanically and acoustically isolated. Alternatively, the AFM can also be used as a stand-alone system on top of a mechanically ­isolated table. 2.2.1. Optical Detection Method

The cantilever is described in detail in the next section (see Note 3). Essentially, the cantilever behaves like a spring with spring constant k, which follows the Hooke’s law F = kd for small deflections, where F is force and d is the cantilever deflection. The deflection is measured by monitoring the position of the reflected laser with a segmented photodiode. Traditionally, laser diodes have been used to measure cantilever deflection, however, optical interference due to the high spatial coherence of lasers results in a sinusoidal-like wave in the deflection signal of force–distance curves (23, 24). The amplitude of interference is sometimes two to three times greater than single molecule rupture forces, thus compromising single molecule detection. To reduce optical interference, superluminescent diodes (SLDs) with a low coherence can be used in place of laser diodes. The light beam is focused on a spot of ~10 mm diameter using a microfocusing lens. The spot is positioned on the end of the cantilever and the reflected light is then collected by the segmented photodiode (Fig. 1). The difference in the photocurrents between the vertical quadrants is then transformed into the deflection of the cantilever. Thus, calibration of the photodiode sensitivity (sometimes referred as InvOLS, inverse optical lever sensitivity (25)) is required when carrying out force spectroscopy.

Force-Clamp Measurements of Receptor–Ligand Interactions

335

2.2.2. Piezoelectric Element

The piezoelectric element pictured in Fig. 1 moves the ­cantilever in the z-direction allowing the functionalized cantilever to make contact with the stationary protein-coated substrate. Another common configuration is to use the piezoelectric ­element to move the sample stage while keeping the cantilever stationary. The position of the piezoelectric element is monitored using either strain gauge (resistive) or capacitive elements. In the strain gauge method, as the piezoelectric element expands and contracts the resistance changes and the position can be monitored. The most common capacitive sensors are linear variable differential transformers (LVDTs), which measure the capacitance change between a fixed electrode and another electrode that moves along with the piezoelectric element. Since capacitive elements move with the moving piezoelectric element, capacitive elements ­generally provide more accurate position monitoring than strain gauge methods (Physik Instrumente has an excellent online piezo tutorial at http://www.physikinstrumente.com).

2.2.3. Piezoelectric Feedback Control

Piezoelectric elements are inherently nonlinear and present hysteresis and creep. This reduces the repeatability of the measurements and the accuracy in repositioning the AFM tip relative to the sample. Creep results in further displacement of the piezoelectric element, when the applied voltage signal is constant. Hysteresis is a complex process where the current displacement behavior is dependent on the previous history. A proportional-integrative-derivative (PID) controller can be used to correct for these nonlinear effects by computing a feedback signal from the piezoelectric position sensors mentioned in the last section to correct for these effects (26). AFM operation without nonlinear piezoelectric correction is referred to as open-loop while closed-loop refers to feedback correction. The advantages of closed-loop operation are a high reproducibility in the positioning of the tip and a dramatic reduction of nonlinearity, creep, and hysteresis. In force spectroscopy measurements, these corrections have an important effect in the applied velocity and the calibration of the photodiode sensitivity. The disadvantage of closed-loop operation is that the deflection signal is noisier as both electronic noise and mechanical/frictional noise, arising from feedback correction, are present. On the other hand, open-loop operation reduces these noise sources; hence, smaller forces can be resolved, but at the tradeoff of nonlinear cantilever displacement. The advantages of closed-loop operation are very useful in AFM imaging and lateral positioning, and most of the new commercial AFM systems include closed-loop operation.

2.3. Optimal Cantilevers for Single Molecule Rupture Force Measurements

AFM cantilevers are force transducers that can measure and apply pico to nano-Newton forces to the sample of interest. They come in many different sizes, shapes, and can be coated with different reflective metal coatings (Au, Al…). Each cantilever is optimized

336

Rico, Chu, and Moy

for a different function; hence, it is crucial for force spectroscopy measurements to select the cantilever optimal for the specific system. The tip at the end of each cantilever also comes in different sizes and shapes and may influence force spectroscopy measurements, as it is the actual part of the AFM to come into contact with the sample. A partial selection of cantilevers used for SMFS is provided in Table 1. 2.3.1. Composition and Dimensions of Cantilevers/Tips

Cantilevers commonly used for SMFS are mostly made of silicon nitride (Si3N4), and are mainly V-shaped or rectangular (see Fig. 2). These cantilevers vary in length from 60 mm for the small Biolevers (Olympus, Japan) to 320  mm for the Microlevers (Veeco, Santa Barbara, CA). At the free end of the cantilever, there is a tip that can be pyramidal, half pyramidal, or conical. Since the tip is the part of the AFM probe that comes into ­contact with the substrate, the tip shape affects the area of contact, which ultimately influences the probability of bond formation between the cantilever tip and the substrate. Generally, unsharpened cantilever tips have been mainly used for SMFS as they provide a smoother and slightly larger area of contact between the tip and the substrate while also reducing the local strains on biological systems, such as cells (2, 27–29). However, sharpened cantilevers have also been used effectively for SMFS (30). A possible drawback of sharpened tips is that they wear relatively faster than unsharpened ones after successive force scans, thus affecting the protein coating as well.

2.3.2. Reflective Coating of Cantilevers/Tips

The top or reflective surface of the cantilever is commonly coated with a thin gold layer to enhance the maximum amount of reflected light (2, 29, 30). Aluminum is also used as a reflective coating. Sometimes, both the reflective and tip surfaces of the cantilever can be coated with gold, in which case thiol chemistry is required for tip functionalization (this is the case of Olympus Biolevers, see Subheading 3.1) (27, 28). Lastly, it is also possible to obtain cantilevers without a reflective coating. Removal of the reflective coating allows for a clear optical light path through the cantilever and minimizes thermal drifts. The disadvantage of using cantilevers without a reflective coating is that the reflected laser signal is much less intense, leading to possibly lower sensitivity.

2.3.3. Ideal Range of Spring Constants

The cantilever spring constant determines the amount of deflection or displacement in the cantilever in response to an applied force. When a molecule tethered to the cantilever tip interacts with an apposing molecule attached to the substrate, the amount of cantilever deflection (d) upon retraction of the functionalized cantilever from the surface depends on the stiffness of the cantilever (kc), the stiffness of the linker and the protein (kl), and lastly the molecular displacement (x) upon separation of the cantilever from the substrate which are related by (31)

Si3N4

Olympus

Veeco

Veeco

Olympus

Biolever BL150VB-C1 (A/B)

NP series (B/D)

MSCT (B/C/D) sharpened MLCT

OMCLTR400PSA

Si3N4

Si3N4

Si3N4

Si3N4

Composition

Veeco

Manufacturer

MLCT-AUHW (B/C/D)

Catalog number

V-shaped

Rect./ V-shaped

V-shaped

Rectangular

Rect./ V-shaped

Shape

200

200/320/ 220

196

60/100

200/320/ 220

Length (mm)

Table 1 Cantilevers for single molecule force spectroscopy measurements

13.4/0.4

22/0.6

(33/18)/ 0.6

30/0.18

22/0.6

Width/ thickness (mm)

4-sided pyramid/ 2.9/20

4-sided pyramid/ 2.5-3.5/20

4-sided pyramid/ 2.5-3.5/20

Semi 4-sided pyramid/ 7/30

4-sided pyramid/ 2.5-3.5/50

Tip shape/ height (mm)/ radius(nm)

0.02

0.02/0.01/ 0.03

0.06/0.03

0.03/0.006

0.02/0.01/ 0.03

Nominal spring constant (N/m)

11

15/7/15

14/12

37/13

15/7/15

Resonant frequency in air (kHz)

(44)

(30)

(29)

(27, 28)

(2, 3)

References

338

Rico, Chu, and Moy

a

b

V-shaped cantilever

Bottom view

length

thickness

width

Rectangular cantilever Bottom view Reflective/top side Tip/bottom side

Side view

Fig. 2. AFM cantilevers. (a) Chip of AFM cantilevers with a regular size paperclip. Inset, magnification of one end of the chip showing four V-shaped cantilevers and one rectangular cantilever (MLCT-AUHW, Veeco). (b) Schematics showing the bottom and lateral views of V-shaped and rectangular cantilevers.



d=

kl x kl + kc

(1)

From Eq. 1, if kc » kl, the deflection of the cantilever is rather small and may be smaller than the resolution of the system. On the other hand, if kc « kl, it may require a larger amount of cantilever deflection (d) to stretch the molecule, which means a larger displacement of the piezoelectric element. Hence, it is important to keep these factors in mind when choosing an optimal cantilever for SMFS. In addition, it has been suggested that the spring constant of the cantilever may affect the energy landscape of the interaction (32–34). Cantilevers with spring constants ranging from 0.006 to 0.06 N/m have been successfully used previously (2, 27–30). The most commonly used cantilevers for SMFS generally have a nominal spring constant of 0.01 N/m (2, 30). 2.3.4. Noise Limit of Cantilevers



Assuming that the system is mechanically and electrically well ­isolated, the noise in the AFM detection system is mainly limited by the thermally driven motion of the cantilever. If the thermal noise in the cantilever is on the same scale or larger than the measured force of interest, then these cannot be well resolved. The thermal noise affecting the cantilever deflection is approximated by ∆Frms =

4g kBTf s

(2)

where kB is Boltzman’s constant, T is the absolute temperature, fs is the cutoff frequency, which should be well above the frequencies of interest over which SMFS measurements are acquired (31), and g is the viscous drag coefficient of the cantilever, which strongly depends on the separation between the cantilever and

Force-Clamp Measurements of Receptor–Ligand Interactions

339

the substrate (35, 36). Since the viscous drag coefficient is ­dependent on the size of the cantilever, it is thus possible to reduce thermal noise by reducing cantilever size (37). One of the smallest cantilevers commercially available to date is the Olympus Biolever (BLRC150V, Table 1). In addition, a higher tip increases the distance between the cantilever and the substrate, reducing the effective viscous drag. Low-pass filtering is also an effective way to reduce the noise in the deflection signal. In the case of the small Biolevers, a theoretical thermal noise limited detection of 5 pN is reasonably expected (38).

3. Methods 3.1. AFM Tip Functionalization

3.1.1. Protein Functionalization of Gold-Coated Tips via Thiol Self-Assembled Monolayers

The most common method of adsorbing proteins to AFM tips is either by coupling via amino (2, 30, 39–46) or carboxy termini (27, 28) (Notes 13 and 15). In addition, it is also possible to link proteins to AFM tips through specific tags if the protein of interest contains a tag and the tip is coated to specifically recognize the tag (2, 42, 46, 47) (see Note 14). Certain cantilevers manufactured with goldcoated tips (i.e., Olympus BL-R150VB and OMCL-TR400PB-1) can be functionalized with thiol-­reactive groups (-SH) to immobilize cross-linkers on the gold layer. Even more, the protein of interest can be linked directly to the gold-coated AFM tips if the protein of interest is labeled with a reactive thiol group (48). A more detailed description of common methods to coat AFM cantilevers with proteins is given in the chapter by Voyer and coworkers. 1. Wash tweezers in Alconox solution and rinse extensively (see Notes 6 and 7) 2. Shake cantilevers in a cleaned glass dish containing nanopure water for 1 min 3. Irradiate cantilevers with UV light for 5 min 4. Wash cantilevers with ethanol by completely submerging each cantilever once in each of five wells of a 24-well tissue culture plate containing 1 mL of ethanol 5. Completely submerge cantilevers in glass dish containing Nanothinks™ ACID11 for 24 h at room temperature. Nanothinks™ ACID11 contains a thiol group, a relatively short linker and a free carboxy terminus. A self-assembled monolayer then forms on the gold surface of the AFM tip, in which thiol groups interact with the gold surface while carboxyl groups remain free for cross-linking proteins to the tip (47, 49). 6. Rinse cantilevers five times with ethanol 7. Prepare activation buffer: 20 mg each of EDC and NHS in 1 mL of PBS pH 6.5–7.2. Equilibrate both EDC and NHS to

340

Rico, Chu, and Moy

room temperature before use. Mix thoroughly and then ­vortex for 1–2 min (see Note 8) 8. Completely submerge cantilevers in activation buffer EDC is a bifunctional cross-linker that reacts with carboxyl groups of the formed monolayer. Once reacted with a carboxyl group, EDC reacts with amine groups in the protein of interest. NHS prevents the rapid hydrolyzation of the EDC intermediate formed with ACID11 (see Pierce product info on NHS and EDC for more details). 9. Rinse cantilevers five times with PBS (see Note 9) 10. Dry the cantilever by gently touching it to the bottom surface of the well of a clean tissue culture plate (see Note 10) 11. Place a small square piece of parafilm (50 × 50  mm2) in the center of a clean Petri dish 12. Place the dry cantilever chips on top of the parafilm forming a circle with the cantilevers facing each other (see Fig.  3). Usually, around four cantilevers per protein drop 13. Gently press the center of each cantilever chip against the parafilm to prevent the cantilever from moving 14. Gently pipette 30–50 mL of the protein solution diluted to the desired protein concentration (see Note 11) 15. Check to see that all of the cantilevers are completely submerged in the solution, and that there are no air bubbles around the cantilevers 16. Transfer the entire tissue culture dish and protein-coated cantilevers into a humidified chamber (see Note 12)

a

b

Fig. 3. Handling of AFM cantilevers (see Note 4). (a) Stainless steel tweezers to manipulate c­ antilevers. (b) Four cantilever chips immersed in a drop of protein solution on top of parafilm forming a circle with the cantilevers facing each other (step 14 of Subheading 3.1.1).

Force-Clamp Measurements of Receptor–Ligand Interactions

341

17. Allow the proteins to coat the cantilevers either overnight at 4°C or for 3 h at room temperature 18. Rinse the excess protein from the protein-coated cantilevers five times in PBS or the desired experimental buffer 19. Store protein-coated cantilevers in PBS at 4°C for up to 5 days 20. Prior to using the cantilever in an experiment, block nonspecific binding sites by incubating for 30–60 min in 1% (w/v) BSA at room temperature 21. Rinse cantilevers five times in experimental buffer In our particular case, we used 50 mg/mL streptavidin as the protein solution. 3.1.2. Removing Thiols from Gold-Coated Cantilevers

Since the gold-coated cantilevers can be relatively expensive, and also coating cantilevers with gold is expensive and time consuming, it might be desired to reuse the cantilevers or any gold-coated surface. The protocol below shows how to remove thiols from gold-coated cantilevers by photooxidation (40). 1. Irradiate cantilever with UV light for 10  min (wavelength 254 nm @ 8 mW/cm2 or equivalently 60 min @ 1.25 mW/cm2) 2. Wash five times with nanopure water 3. Wash five times with ethanol 4. Store cantilever in air or argon

3.2. Substrate Coating

Hydrophilic surfaces are the most commonly used substrates to immobilize proteins for force spectroscopy. Hydrophilic surfaces include tissue culture dishes, glass coverslips/slides, beads, and mica. These are generally readily available and allow for relatively easy protein adsorption. Hydrophobic surfaces can also be used. The drawback of immobilizing proteins on hydrophilic surfaces is the increase in nonspecific binding compared to hydrophobic surfaces.

3.2.1. Mica

Mica is a mineral that exists as densely compact thin sheets. Commercially, it comes in different grades V1–V10, varying also in shape and pricing. Some of the nice properties of V1 mica are that each sheet is atomically flat, after cleavage each sheet is tremendously clean and optically clear, and many sheets can be cleaved from one sample. In most buffers, mica is negatively charged, making it relatively easy to functionalize mica surfaces. One of the most common methods to immobilize proteins to mica surfaces is through ionic interactions with a salt solution, namely NaCl (44). Mica can also be silanized to covalently crosslink adhesion molecules to mica (50, 51). The preparation of

342

Rico, Chu, and Moy

mica substrates is explained below in a two part protocol ­consisting of, firstly, fabrication and, secondly, protein coating of mica substrates (Notes 17 and 18 on alternative substrates). 3.2.1.1. Fabricating Mica Substrates

1. If your mica substrate fits onto a glass slide then proceed to step 2, otherwise use a dye and punch set and cut out a mica disk smaller than a glass slide. Avoid using scissors to cut mica. Some manufacturers sell precut mica disks/squares close to the desired size 2. Wash the glass slide or desired support with acetone 3. Glue the mica disk onto the dry clean glass slide with optically clear epoxy (EPOTEK 377). Follow manufacturer’s instructions for handling epoxy and curing time.

3.2.1.2. Functionalizing Mica Using a Salt Solution

1. Cleave the mica by taking a piece of scotch tape and pressing it against the mica disk. In one smooth, complete motion, rip the tape off (see Note 16) 2. Dilute the protein of interest in NaCl to a final molar concentration of 1 mM NaCl 3. Coat the mica disk with protein in 1 mM NaCl for 10 min 4. Rinse the protein-coated mica surface with 1 mM NaCl five times In our case, we use passive adsorption of biotinylated bovine serum albumin (bBSA) by incubating 1% bBSA overnight at 4°C.

3.3. AFM Force-Clamp Measurements

Force-clamp measurements require a fast enough feedback circuit to keep the deflection (force) constant, together with low-noise conditions to detect minimal forces. Low-noise conditions can be effectively achieved by vibration and acoustic isolation of the AFM system and by the reducing the size of the cantilever and the ­bandwidth of data acquisition. Most commercial AFM systems use  digital feedback to keep the force constant during contact mode imaging. However, its application to measurements in which the  force is held constant during pulling is not such a common feature of commercial software. We used a commercial AFM ­system MFP-3D-BIO (Asylum Research, Santa Barbara, CA), ­coupled to the stage of an inverted optical microscope (Nikon, Japan), which already provided force-clamp control procedures. The cantilevers used, biolevers, had a nominal spring constant of 6 pN/nm. The InvOLS was determined by calculating the slope of deflection (in V) versus distance (nm) from five consecutive force curves on the hard substrate, resulting in ~40  nm/V (this value depends on the specific electronics, on the geometrical configu­ ration of the AFM system, and on the specific cantilever used, Fig. 4). The spring constant of the cantilever was calibrated in ­liquid prior the measurements using a method provided by the AFM

Force-Clamp Measurements of Receptor–Ligand Interactions

343

0.8

Deflection (V)

0.6

slope=OLS

0.4

0.2

0

−20

−40

−60

−80

−100

Displacement (nm) Fig. 4. Sensitivity calibration. Representative example of a calibration curve obtained on a mica surface coated with BSA. The slope of the linear part of the curve was used to determine the optical lever sensitivity, OLS = 1/InvOLS.

manufacturer­ based on the thermal fluctuations method (52, 53), which lead to spring constants ranging from 5 to 7 pN/nm. The detailed force-clamp measurement procedure is described below. 1. Mount the cantilever on the cantilever holder (see Note 19) 2. Immediately immerse the cantilever in the measurement buffer to minimize any contact with air 3. Allow the system to equilibrate (15–30 min) 4. Be sure that the tip is far away from the surface (at least 20 mm) and acquire the thermal fluctuations of the cantilever to calibrate the spring constant using the thermal fluctuations method (52, 54) 5. Approach the tip to the substrate and make contact 6. Acquire five force-distance curves with a deflection ~1 V (this value depends on the particular AFM system) to calibrate the photodiode signal and determine the InvOLS by the inverse of the resulting slope (Fig. 4) 7. Position the tip on the protein-coated surface and make slight contact 8. Acquire force-clamp curves consisting in the following steps (Fig. 5): (a) Approach the tip to apply a force of few tens of pN (region I)

344

Rico, Chu, and Moy

(b) Maintain contact at this compression force for a certain time so as to assure that only 10–30% of the curves lead to bond formation (see Note 20) (region II) (c) Retract the cantilever at ~2 mm/s until it reaches the set clamping force (e.g., 10 pN) (region III) (d) If a bond is formed, hold the cantilever at this force until the bond breaks (region IV). If no bond is formed, the cantilever will reach its initial position (e) Retract the cantilever until the initial, resting position (regions V and VI) (f) Repeat the cycle (see Notes 20–22) 3.4. Control Measurements

Control measurements are crucial in adhesion measurements. In general, hundreds of force curves are acquired under determined conditions (compression force and contact time) and the adhesion frequency is calculated. Control measurements consist of acquiring force curves under the exact experimental conditions only modifying­ the ligand- or receptor-adhesive conditions and determining the resulting adhesion frequency. There are different methods to prove the specificity of the interaction that depend on each experimental system, such as blocking the receptor or ligand with antibodies, chelating ions required for the interaction, using cells not expressing the protein of interest, etc. An alternative method is using AFM tips (or substrates) in which the last step of protein coating was omitted. In our particular case, we blocked the biotinylated surface using free streptavidin as the most straightforward method, finding a dramatic reduction of the adhesion frequency. The relative adhesion frequency averaged over three positions of the sample surface dropped from 100% to 2 ± 1% after adding 100 mL of 100 mg/ml free streptavidin to the measurement buffer to block the biotin-coated substrate (final concentration ~50 mg/mL).

3.5. Data Processing and Analysis

Only force curves presenting a single adhesion event are considered. A representative example of a force-clamp curve is shown in Fig.  5, which represents the detected force and the piezo ­displacement as a function of time. Only the force-clamping region (region IV) is used to determine the two relevant parameters. 1. The actual clamping force, which depends on the resting force level, determined as the average level of force during clamping relative to the resting force level at zero velocity (see Note 23) 2. The bond lifetime measured from the beginning of the clamping regime until the force drops. The resulting list of forces and lifetimes is then pooled according to levels of force (in our case 10, 20, and 25 pN). The mean

Force-Clamp Measurements of Receptor–Ligand Interactions

345

(IV)

a 500 contact (II)

Displacement (nm)

400

(III)

300 200

approach (I)

100 0

Force (pN)

40

(V) (VI)

20

0

-20 0.2

b

Time (s)

0.6

0.8

60

40

Force (pN)

0.4

retraction clamping (III) (IV)

20

resting (VI)

repositioning (V)

0 clamping force

-20 lifetime

-40 0.80

0.85

0.90 Time (s)

0.95

1.00

Fig. 5. Force-clamp measurements. (a) Distance (top) and force (bottom) curves showing an example of a force-clamp measurement. The relevant steps of the curves are shown: (I) approaching regime in which force increased from zero (noncontact) to ~50 pN (trigger force); (II) in the contact regime, the force was maintained at the trigger value during a dwell time of ~0.5 s; (III) during retraction, the cantilever was withdrawn until the clamping force was reached (~20 pN, negative); (IV) during the clamping process, the force was maintained constant until the bond broke; (V) the cantilever was withdrawn 500 nm during repositioning; and (VI) the cantilever remained immobile during the resting regime. (b) Magnification of the four last steps of the force trace showing how to determine the lifetime of the bond and the actual clamping force.

346

Rico, Chu, and Moy

force of each group is thus calculated together with the standard deviation. For each set of clamping force, we generate the cumulative probability distribution of lifetimes by plotting the number of events with a lifetime of t or more against t (Fig. 6a), which can be modeled as an exponential decay (Aexp[-t/t (F) ]). The intrinsic lifetime of the interaction at the applied force, t(F), is then estimated by fitting the exponential to the data. More accurate methods can be applied to determine intrinsic dissociation lifetimes (55, 56). The lifetime obtained for each set is then represented against the average clamping force (Fig. 6b). The Bell model is the more generalized approach to interpret the effect of force on the lifetime of a bond (1, 5) and can be summarized by the following expression (see Notes 24 and 25)

3.6. Interpretation of the Results

t (F ) = t 0e



x b F / kBT



where xb is the effective range of the interaction, kB is the Boltzmann constant, and T, the absolute temperature. As observed, the lifetime of the bond decreases exponentially with the applied force, as is reflected from our experimental results (Fig. 6b). By fitting the above equation to our lifetime versus force data, we obtain the parameters of the bond, its intrinsic lifetime at no force (t0), and the width of the potential energy. In our case, we found a width of ~1 nm and a lifetime of ~9 s, which is in good agreement with previous results, given the low number of data points a

b 100

1

8 6

8

10 pN 20 pN 25 pN

6 4

2

Lifetime (s)

Number of events

4

10

8 6 4

2

0.1

8 6 4

2

2

1

8 6

0.01 0.0

0.5

1.0 1.5 Lifetime,t (s)

2.0

2.5

5

10

15 20 Force (pN)

25

30

Fig. 6. Dissociation kinetics. (a) Lifetime distributions representing the number of events with a lifetime of t or more against t at three different levels of clamping force (10, 20 and 25 pN, circles, crosses and squares, respectively). Solid lines are best fits of an exponential decay to determine the mean lifetime at each force. (b) Fitted lifetimes against actual forces. Error bars represent one standard deviation. The Bell model (solid line) was fitted to the curve obtaining a potential width xb = 1.0 ± 0.1 nm and an intrinsic lifetime at zero force t0 = 9 ± 1 s.

Force-Clamp Measurements of Receptor–Ligand Interactions

347

(11, 22). Given the virtually zero loading rate and the relatively low forces applied, it is possible that we were probing the outermost barrier of the streptavidin–biotin interaction, which has been described before at 1.3 nm (57). From previous studies of forced dissociation of the streptavidin–biotin complex, it is expected that an inner barrier would dominate the dissociation pathway at higher forces. This effect would be characterized by a weaker dependence of lifetime on the applied force (11, 21, 22).

4. Notes 1. APTES is a common cross-linker used to derivatize siliconbased substrates (i.e., glass or many types of cantilevers) and may also be used for quartz surfaces. The free amino group allows the cross-linking to proteins with a free carboxy terminus. This solution is very hygroscopic and can hydrolyze very quickly, so it is best to purchase in small volumes. 2. Handling AFM cantilevers is a difficult task when it is done for the first time. Handling broken cantilevers beforehand helps to acquire some practice without damaging new cantilevers. 3. Generally, cantilevers are mounted in an angle of 8–10° from the axis parallel to the substrate surface. 4. Cantilevers are flexible but fragile, and it is common to lose some of them during the various steps of the coating protocol. Thus, it is recommended to functionalize more than one cantilever, i.e., more than one chip, for each experiment (5–10 chips). Coated cantilever can be stored in PBS at 4°C for about 1 week. 5. The AFM cantilever holder can be made out of plexiglass or quartz. Quartz is a nice alternative to plexiglass as it is relatively more inert than plexiglass and is resistant to scratching. The disadvantage of quartz holders is the difficulty in machining. 6. Before starting the functionalization protocol, check to see whether all of the cantilevers on the chip are intact and undamaged to prevent wasting valuable protein and also your own time. 7. Long cantilevers (~320 mm) have the tendency to bend back onto themselves due to the surface tension of the fluid giving the appearance of a missing cantilever. To ensure that the cantilever is intact, gently tap the side of the vessel or use a pair of forceps and pick up the cantilever chip and tap it against the bottom of the vessel.

348

Rico, Chu, and Moy

8. Prepare fresh before using, and must be used immediately after preparing as the amine group on EDC and NHS can hydrolyze rapidly. Store both NHS and EDC in desiccant sieves at 4°C. 9. Notice that Tris–HCl buffer contains free amine groups. Avoid the use of this buffer for protein storage and during the functionalization process, as it reduces the coupling efficiency. PBS or HEPES buffers are recommended. 10. Residual fluid remaining on the cantilever dilutes the final protein concentration, but if the cantilever is too dry then it may be difficult to coat the cantilevers with protein solution. Just leave a small trace amount of fluid remaining on the chip, probably something less than 2 mL. 11. Depending on the system, the protein solution concentration for SMFS can range from 5 to 100  mg/mL, so generally a good starting protein concentration is 50 mg/mL. Usually a 30–50 mL is enough to coat four cantilever chips at once. 12. A humidified chamber can be made easily from a plastic box, e.g., a small sandwich lunchbox or an empty pipette tip box. Fill the bottom 10  mm of the box with water and place a stable platform protruding above the water so the Petri dish can sit inside the box without touching the water. 13. Another common method for coupling proteins to silicone nitride tips is through heterobifunctional cross-linker attached to the silanized silicone nitride surface (7, 27). 14. Most of the used methods to functionalize the tips, including those explained here, link the protein via any amine group, with no defined orientation. Alternatively, precise orientation of the protein can be achieved by making use of a specific tag present in it, such as His- or Avi-tags. Coating the tip with a receptor recognizing such tags or with an antibody binding to a known epitope of the protein would allow us to control the protein orientation. This method has been successfully used on different systems: His-tag (46, 47) and biotin (2, 42). The two major advantages of attaching proteins to AFM tips with site-specific tags are specificity and controlled protein orientation. However, it is important to first characterize the binding strength of such linking bonds because it may not be strong enough and may break before rupture of the complex under study (46). 15. Attachment of proteins to cantilevers with long linkers diminishes nonspecific adhesion of the tip with the functionalized surface. The long linker adds an elastic component to the single molecule rupture force profile. It also shifts the initial loading regime .The actual elastic response to force, usually described as a worm-like chain (58, 59), and the shift of the start of the loading regime depend on the length of the linker

Force-Clamp Measurements of Receptor–Ligand Interactions

349

and have been used as a signature of specific interaction (60). Two types of commonly used long linkers include polyethylene glycol (PEG) linkers (27, 28, 41, 60) and carboxymethyl amylose chains (49). Both of these moieties can be modified to obtain tethers of different length. 16. A nice cleavage should yield a very smooth surface devoid of any smaller pieces of mica. 17. Other convenient substrates on which proteins can be passively adsorbed are polystyrene dishes. As mentioned before, control measurements are required to confirm the specificity of the interaction. In general, most nonspecific binding in hydrophilic surfaces can be blocked with 1% BSA, but a small amount of nonspecific binding can still occur. The major advantage of hydrophobic surfaces is the reduction in nonspecific binding compared to hydrophilic surfaces. Ideal hydrophobic surface are Petri dishes manufactured to culture bacteria. Unlike tissue culture dishes, Petri dishes have not been chemically modified and are thus hydrophobic. One drawback of conducting adhesion measurements on hydrophobic surfaces is that for an equivalent protein concentration, less protein adsorbs to the hydrophobic than to the hydrophilic surface, hence, more concentrated protein solutions are generally required to functionalize hydrophobic surfaces. However, this is a small price to pay because there is either very little or, in some cases, absolutely no nonspecific binding to hydrophobic surfaces blocked with 1% w/v Pluronic (BASF), instead of BSA. 18. An alternative way of immobilizing proteins on a substrate is the use of commercially available beads. Beads come in many different materials and can vary greatly in size. One of the major advantages of beads is that many of them are commercially available with a wide range of conjugated proteins designed for protein purification, protein labeling, and other techniques requiring molecular protein specificity. Hence, if the adhesion molecule of interest is labeled with a tag specific (e.g., biotin, histidine, NTA, etc.) to the protein conjugated to the bead, then the molecule of interest can specifically adhere to the bead. Depending on the material composition, beads can also vary in stiffness, which may affect the contact area with the tip. In addition, the material composition may also affect the surface roughness of the beads. In order to work with beads in force spectroscopy, they must be immobilized to the substrate at least to the point where force curves do not affect bead position. Bead immobilization can be accomplished by using the same principle used for detection, i.e., using their own specific tag (61).

350

Rico, Chu, and Moy

19. Avoid exposure of all functionalized cantilevers to air, as ­proteins may be damaged. Thus, mount the cantilever on the AFM holder and immediately immerse it on the experimental buffer. 20. The adhesion frequency (the frequency of binding events) has to be kept <30%, preferably ~10%, to assure that the majority of rupture events are due to single bonds (62, 63). To control the adhesion frequency, various experimental conditions can be varied: (a) Contact time between adhering surfaces (b) Ligand and/or receptor densities (c) Contact area between AFM tip and sample surface (tip and sample geometry and elasticity) (d) Compression force, i.e., indentation, on elastic samples, such as cells or agarose beads 21. To prevent from possible artifacts due to wearing of the tip and/or substrate and/or damaging of the protein, it is recommended to probe different sample regions during the force measurements, e.g., change cantilever or sample lateral position every 100 curves. 22. Measurements under different conditions (e.g., various clamping force levels or different loading rates) should be carried out in random order. 23. As can be observed in Fig. 5b, the clamping force is measured relative to the force level when the cantilever is in its resting position, i.e., not in movement. Depending on the cantilever shape, the viscosity of the buffer and the speed at which it is moving, a viscous drag affects the deflection of the cantilever, and thus the force measured. This drag force depends also on the position of the cantilever relative to the sample surface. Therefore, it is important to measure the actual clamping force relative to the force level at zero velocity or to characterize the viscous drag and correct for it (64, 65). 24. The development of single molecule measurement techniques has been accompanied by a refinement of the theoretical background. The initial theoretical framework of dynamic force spectroscopy was established by the seminal paper by Evans and Ritchie (5) based on the classic works by Bell and Kramers (1, 66). More recent refinements have been ­developed to take into account the shape of the potential (13, 15, 67) and to derive more general expressions (16, 68, 69). 25. Recent studies suggest that the spring constant of the cantilever may affect the energy landscape of the bond and has to be taken into account when analyzing the data (33, 34, 70, 71).

Force-Clamp Measurements of Receptor–Ligand Interactions

351

References 1. Bell, G. I. (1978) Models for Specific Adhesion of Cells to Cells, Science 200, 618–627. 2. Moy, V. T., Florin, E. L., and Gaub, H. E. (1994) Intermolecular Forces and Energies Between Ligands and Receptors, Science 266, 257–259. 3. Florin, E. L., Moy, V. T., and Gaub, H. E. (1994) Adhesion Forces Between Individual Ligand-Receptor Pairs, Science 264, 415–417. 4. Lee, G. U., Kidwell, D. A., and Colton, R. J. (1994) Sensing Discrete Streptavidin Biotin Interactions with Atomic-Force Microscopy, Langmuir 10, 354–357. 5. Evans, E., and Ritchie, K. (1997) Dynamic strength of molecular adhesion bonds, Biophys. J. 72, 1541–1555. 6. Benoit, M., Gabriel, D., Gerisch, G., and Gaub, H. E. (2000) Discrete interactions in cell adhesion measured by single-molecule force spectroscopy, Nature Cell Biology 2, 313–317. 7. Zhang, X. H., Wojcikiewicz, E., and Moy, V. T. (2002) Force spectroscopy of the leukocyte function-associated antigen-1/intercellular adhesion molecule-1 interaction, Biophysical Journal 83, 2270–2279. 8. Hinterdorfer, P., Baumgartner, W., Gruber, H. J., Schilcher, K., and Schindler, H. (1996) Detection and localization of individual antibody-antigen recognition events by atomic force microscopy, Proc Natl Acad Sci U S A 93, 3477–3481. 9. Dufrene, Y. F., and Hinterdorfer, P. (2008) Recent progress in AFM molecular recognition studies, Pflugers Arch 456, 237–245. 10. Muller, D. J., Helenius, J., Alsteens, D., and Dufrene, Y. F. (2009) Force probing surfaces of living cells to molecular resolution, Nat Chem Biol 5, 383–390. 11. Merkel, R., Nassoy, P., Leung, A., Ritchie, K., and Evans, E. (1999) Energy landscapes of receptor-ligand bonds explored with dynamic force spectroscopy, Nature 397, 50–53. 12. Li, F. Y., Redick, S. D., Erickson, H. P., and Moy, V. T. (2003) Force measurements of the alpha(5)beta(1) integrin-fibronectin interaction, Biophys.J. 84, 1252–1262. 13. Dudko, O. K., Hummer, G., and Szabo, A. (2006) Intrinsic rates and activation free energies from single-molecule pulling experiments, Phys.Rev.Lett. 96. 14. Hummer, G., and Szabo, A. (2001) Free energy reconstruction from nonequilibrium single-molecule pulling experiments, Proc Natl Acad Sci U S A 98, 3658–3661. 15. Dudko, O. K., Hummer, G., and Szabo, A. (2008) Theory, analysis, and interpretation of

single-molecule force spectroscopy experiments, Proc Natl Acad Sci U S A 105, 15755–15760. 16. Freund, L. B. (2009) Characterizing the resistance generated by a molecular bond as it is forcibly separated, Proceedings of the National Academy of Sciences 106, 8818–8823. 17. Sulchek, T. A., Friddle, R. W., Langry, K., Lau, E. Y., Albrecht, H., Ratto, T. V., DeNardo, S. J., Colvin, M. E., and Noy, A. (2005) Dynamic force spectroscopy of parallel individual Mucin1-antibody bonds, Proc.Natl. Acad.Sci.U.S.A. 102, 16638–16643. 18. Oberhauser, A. F., Hansma, P. K., CarrionVazquez, M., and Fernandez, J. M. (2001) Stepwise unfolding of titin under force-clamp atomic force microscopy, Proc. Natl. Acad. Sci. USA 98, 468–472. 19. Marshall, B. T., Long, M., Piper, J. W., Yago, T., McEver, R. P., and Zhu, C. (2003) Direct observation of catch bonds involving celladhesion molecules, Nature 423, 190–193. 20. Marshall, B. T., Sarangapani, K. K., Lou, J. H., McEver, R. P., and Zhu, C. (2005) Force history dependence of receptor-ligand dissociation, Biophys. J. 88, 1458–1466. 21. Pincet, F., and Husson, J. (2005) The solution to the streptavidin-biotin paradox: The influence of history on the strength of single molecular bonds, Biophys.J. 89, 4374–4381. 22. Rico, F., and Moy, V. T. (2007) Energy landscape roughness of the streptavidin-biotin interaction, J Mol Recognit 20, 495–501. 23. Butt, H. J., Cappella, B., and Kappl, M. (2005) Force measurements with the atomic force microscope: Technique, interpretation and applications, Surface Science Reports 59, 1–152. 24. Weisenhorn, A. L., Maivald, P., Butt, H. J., and Hansma, P. K. (1992) Measuring Adhesion, Attraction, And Repulsion Between Surfaces In Liquids With An Atomic-Force Microscope, Phys. Rev. B 45, 11226–11232. 25. Proksch, R., Schaffer, T. E., Cleveland, J. P., Callahan, R. C., and Viani, M. B. (2004) Finite optical spot size and position corrections in thermal spring constant calibration, Nanotechnology 15, 1344–1350. 26. Devasia, S., Eleftheriou, E., and Moheimani, S. O. R. (2007) A survey of control issues in nanopositioning, Ieee Transactions on Control Systems Technology 15, 802–823. 27. Neuert, G., Albrecht, C., Pamir, E., and Gaub, H. E. (2006) Dynamic force spectroscopy of the digoxigenin-antibody complex, Febs Letters 580, 505–509. 28. Odorico, M., Teulon, J. M., Bessou, T., Vidaud, C., Bellanger, L., Chen, S. W. W.,

352

Rico, Chu, and Moy

Quemeneur, E., Parot, P., and Pellequer, J. L. (2007) Energy landscape of chelated uranyl: Antibody interactions by dynamic force spectroscopy, Biophysical Journal 93, 645–654. 29. Ray, C., Brown, J. R., and Akhremitchev, B. B. (2007) Correction of systematic errors in singlemolecule force spectroscopy with polymeric tethers by atomic force microscopy, Journal of Physical Chemistry B 111, 1963–1974. 30. Francius, G., Alsteens, D., Dupres, V., Lebeer, S., De Keersmaecker, S., Vanderleyden, J., Gruber, H. J., and Dufrene, Y. F. (2009) Stretching polysaccharides on live cells using single molecule force spectroscopy, Nature Protocols 4, 939–946. 31. Gittes, F., and Schmidt, C. F. (1998) Thermal noise limitations on micromechanical experiments, European Biophysics Journal with Biophysics Letters 27, 75–81. 32. Noy, A. (2008) Strength in numbers: Probing and understanding intermolecular bonding with chemical force microscopy, Scanning 30, 96–105. 33. Walton, E. B., Lee, S., and Van Vliet, K. J. (2008) Extending Bell’s model: How force transducer stiffness alters measured unbinding forces and kinetics of molecular complexes, Biophys. J. 94, 2621–2630. 34. Tshiprut, Z., Klafter, J., and Urbakh, M. (2008) Single-molecule pulling experiments: when the stiffness of the pulling device matters, Biophys J 95, L42–44. 35. Roters, A., and Johannsmann, D. (1996) Distance-dependent noise measurements in scanning force microscopy, Journal of PhysicsCondensed Matter 8, 7561–7577. 36. Alcaraz, J., Buscemi, L., Puig-de-Morales, M., Colchero, J., Baro, A., and Navajas, D. (2002) Correction of microrheological measurements of soft samples with atomic force microscopy for the hydrodynamic drag on the cantilever, Langmuir 18, 716–721. 37. Viani, M. B., Schaffer, T. E., Chand, A., Rief, M., Gaub, H. E., and Hansma, P. K. (1999) Small cantilevers for force spectroscopy of single molecules, Journal of Applied Physics 86, 2258–2262. 38. Junker, J. P., Ziegler, F., and Rief, M. (2009) Ligand-Dependent Equilibrium Fluctuations of Single Calmodulin Molecules, Science 323, 633–637. 39. Berquand, A., Xia, N., Castner, D. G., Clare, B. H., Abbott, N. L., Dupres, V., Adriaensen, Y., and Dufrene, Y. F. (2005) Antigen binding forces of single antilysozyme Fv fragments explored by atomic force microscopy, Langmuir 21, 5517–5523.

40. Brewer, N. J., Janusz, S., Critchley, K., Evans, S. D., and Leggett, G. J. (2005) Photooxidation of self-assembled monolayers by exposure to light of wavelength 254  nm: A static SIMS study, Journal of Physical Chemistry B 109, 11247–11256. 41. Ebner, A., Wildling, L., Kamruzzahan, A. S. M., Rankl, C., Wruss, J., Hahn, C. D., Holzl, M., Zhu, R., Kienberger, F., Blaas, D., Hinterdorfer, P., and Gruber, H. J. (2007) A new, simple method for linking of antibodies to atomic force microscopy tips, Bioconjugate Chemistry 18, 1176–1184. 42. Florin, E. L., Moy, V. T., and Gaub, H. E. (1994) Adhesion forces between individual ligand-receptor pairs, Science 264, 415–417. 43. Kamruzzahan, A. S. M., Ebner, A., Wildling, L., Kienberger, F., Riener, C. K., Hahn, C. D., Pollheimer, P. D., Winklehner, P., Holzl, M., Lackner, B., Schorkl, D. M., Hinterdorfer, P., and Gruber, H. J. (2006) Antibody linking to atomic force microscope tips via disulfide bond formation, Bioconjugate Chemistry 17, 1473–1481. 44. Raab, A., Han, W., Badt, D., Smith-Gill, S. J., Lindsay, S. M., Schindler, H., and Hinterdorfer, P. (1999) Antibody recognition imaging by force microscopy, Nat Biotechnol 17, 901–905. 45. Riener, C. K., Stroh, C. M., Ebner, A., Klampfl, C., Gall, A. A., Romanin, C., Lyubchenko, Y. L., Hinterdorfer, P., and Gruber, H. J. (2003) Simple test system for single molecule recognition force microscopy, Analytica Chimica Acta 479, 59–75. 46. Verbelen, C., Gruber, H. J., and Dufrene, Y. F. (2007) The NTA-HiS(6) bond is strong enough for AFM single-molecular recognition studies, in 1st International AFM BioMed Conference on Life Sciences and Medicine (AFM BioMed 2007), pp 490–494, John Wiley & Sons Ltd, Barcelona, SPAIN. 47. Dupres, V., Menozzi, F. D., Locht, C., Clare, B. H., Abbott, N. L., Cuenot, S., Bompard, C., Raze, D., and Dufrene, Y. F. (2005) Nanoscale mapping and functional analysis of individual adhesins on living ­bacteria, Nature Methods 2, 515–520. 48. Gad, M., Itoh, A., and Ikai, A. (1997) Mapping cell wall polysaccharides of living microbial cells using atomic force microscopy, Cell Biology International 21, 697–706. 49. Grandbois, M., Dettmann, W., Benoit, M., and Gaub, H. E. (2000) Affinity imaging of red blood cells using an atomic force microscope, Journal of Histochemistry & Cytochemistry 48, 719–724. 50. Favis, B. D., Leonard, B. J., and Prudhomme, R. E. (1983) The interaction of a cationic

Force-Clamp Measurements of Receptor–Ligand Interactions silane coupling agent with mica, Journal of Applied Polymer Science 28, 1235–1244. 51. Wang, H. D., Bash, R., Yodh, J. G., Hager, G. L., Lohr, D., and Lindsay, S. M. (2002) Glutaraldehyde modified mica: A new surface for atomic force microscopy of chromatin, Biophys. J. 83, 3619–3625. 52. Hutter, J. L., and Bechhoefer, J. (1993) Calibration of atomic-force microscope tips, Rev.Sci.Instr. 64, 1868–1873. 53. Walters, D. A., Cleveland, J. P., Thomson, N. H., Hansma, P. K., Wendman, M. A., Gurley, G., and Elings, V. (1996) Short Cantilevers for Atomic-Force Microscopy, Rev.Sci.Instr. Vol 67, 3583–3590. 54. Viani, M. B., Schaffer, T. E., Chand, A., Rief, M., Gaub, H. E., and Hansma, P. K. (1999) Small cantilevers for force spectroscopy of single molecules, J Appl Phys 86, 2258–2262. 55. McManus, O. B., Blatz, A. L., and Magleby, K. L. (1987) Sampling, Binning, Fitting, And Plotting Durations Of Open-And-Shut Intervals From Single Channels, Biophys. J. 51, A48–A48. 56. Sigworth, F. J., and Sine, S. M. (1987) Data Transformations For Improved Display And Fitting Of Single-Channel Dwell Time Histograms, Biophys. J. 52, 1047–1054. 57. Pierres, A., Touchard, D., Benoliel, A. M., and Bongrand, P. (2002) Dissecting streptavidinbiotin interaction with a Laminar flow chamber, Biophys. J. 82, 3214–3223. 58. Bustamante, C., Marko, J. F., Siggia, E. D., and Smith, S. (1994) Entropic Elasticity of LambdaPhage DNA, Science 265, 1599-1600. 59. Rief, M., Gautel, M., Oesterhelt, F., Fernandez, J. M., and Gaub, H. E. (1997) Reversible unfolding of individual titin immunoglobulin domains by AFM, Science 276, 1109–1112. 60. Hinterdorfer, P., Baumgartner, W., Gruber, H. J., Schilcher, K., and Schindler, H. (1996) Detection and localization of individual antibody-antigen recognition events by atomic force microscopy, Proceedings of the National

353

Academy of Sciences of the United States of America 93, 3477–3481. 61. Rico, F., and Moy, V. T. (2007) Energy landscape roughness of the streptavidin-biotin interaction, Journal of Molecular Recognition 20, 495–501. 62. Tees, D. F. J., Waugh, R. E., and Hammer, D. A. (2001) A microcantilever device to assess the effect of force on the lifetime of selectin-carbohydrate bonds, Biophys.J. 80, 668–682. 63. Tees, D. F. J., Woodward, J. T., and Hammer, D. A. (2001) Reliability theory for receptor-ligand bond dissociation, J.Chem.Phys. 114, 7483–7496. 64. Alcaraz, J. (2001) Micromechanics of Cultured Human Bronchial Epithelial Cells Measured with Atomic Force Microscopy, Universitat de Barcelona. 65. Janovjak, H. J., Struckmeier, J., and Muller, D. J. (2005) Hydrodynamic effects in fast AFM single-molecule force measurements, Eur Biophys J Biophy 34, 91–96. 66. Kramers, H. A. (1940) Brownian motion in a field of force and the diffusion model of chemical reactions, Physica 7, 304. 67. Hummer, G., and Szabo, A. (2005) Free energy surfaces from single-molecule force spectroscopy, Acc Chem Res 38, 504–513. 68. Friddle, R. W. (2008) Unified model of dynamic forced barrier crossing in single molecules, Physical Review Letters 100. 69. Evans, E. H., K; Kinoshita, K; Wong, WP. (2009) A new approach analysis of single-­ molecule force measurements, in Handbook of single molecule biophysics (Hinterdorfer, P. v. O., A, Ed.), pp 571–589, Springer, New York. 70. Friddle, R. W., Podsiadlo, P., Artyukhin, A. B., and Noy, A. (2008) Near-equilibrium chemical force microscopy, Journal of Physical Chemistry C 112, 4986–4990. 71. Zhang, Y., Sun, G. Y., Lu, S. Q., Li, N., and Long, M. A. (2008) Low Spring Constant Regulates P-Selectin-PSGL-1 Bond Rupture, Biophys. J. 95, 5439–5448.

wwwwwwwwwwwwww

Chapter 21 Measuring Cell Adhesion Forces: Theory and Principles Martin Benoit and Christine Selhuber-Unkel Abstract Cell adhesion is an essential prerequisite for survival, communication, and navigation of cells in organisms. It is maintained by the organized binding of molecules from the cell membrane to the extracellular space. This chapter focuses on direct measurements of cellular binding strength at the level of single adhesion molecules. Using atomic force microscopy-based force measurements, adhesion strength can be monitored as a function of adhesion time and environmental conditions. In this way, cellular adhesion strategies like changes in affinity and avidity of adhesion molecules (e.g., integrins) are characterized as well as the molecular arrangement of adhesion molecules in the cell membrane (e.g., molecular clusters, focal adhesion spots, and linkage to the cytoskeleton or tether). Some prominent values for the data evaluation are presented as well as constraints and preparative techniques for successful cell adhesion force experiments. Key words: Cell adhesion, Affinity, Anchoring, Cell membrane, Cytoskeleton, Force measurement, Focal adhesion, Living cells, Force sensor modifications, Clustering

1. Introduction Cells are complex microfactories that maintain the basic function of all living organisms. In their lifespan, cells produce huge amounts of distinct biomolecules that they organize within their membrane as well as in the intra- and extracellular space with extreme accuracy. A number of specific tasks are performed using this machinery, such as transducing signals from the extracellular environment and forming responses to these signals. The cell membrane is central to the interaction of a cell with its environment in that it provides a direct interface between the intraand extracellular space. Its lipid bilayer incorporates ­protecting molecules, receptors, ion channels, and adhesion molecules that mediate specific cellular interactions. Adhesion molecules may be

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_21, © Springer Science+Business Media, LLC 2011

355

356

Benoit and Selhuber-Unkel

distributed stochastically in the membrane, but they can also appear in the form of clusters at distinct spots. Typically, such clusters are centers for the communication between the intra- and extracellular space which generate special tasks, such as sensing the mechanical properties of the extracellular environment (1–3). An important function of cells is not only to sense forces, but also to actively exert forces (4). This ability is essential to processes, such as cell division, locomotion, and spreading (5, 6). Cellular forces are generated intracellularly by the concerted action of filamentous cytoskeletal elements and molecular motors. In most cell types, the cytoskeleton consists of the actin network, microtubule rods, and other filamentous cell-stiffening molecules. These cytoskeletal elements are not only critical to their role for force generation, but also responsible for the stability of cell shape, the visco-elastic properties of a cell, and stress propagation within a cell and its tissue (7, 8). Although electron microscopy, NMR, and X-ray diffraction analyses are steadily increasing our knowledge about the structure of extra- and intracellular units and biomolecules, the dynamic properties of biological processes are still unknown to a large extent. Such dynamic processes are ubiquitous. For example, biomolecules act as catalysts, signal amplifiers, gene regulators, energy suppliers, signal transducers, information stores, or transportation units, just to mention a few. With laser spectroscopy and molecular dynamics simulation, dynamic molecular processes can be visualized at the atomic and femtosecond scale (9). Under equilibrium conditions, the dynamics of molecular interactions are typically studied using calorimetry or surface plasmon resonance techniques and high-resolution fluorescence techniques (10). Importantly, many cellular processes take place under nonequilibrium conditions, (11) such as active intracellular transport processes and enforced binding and unbinding events. These processes can be quantified in vitro using nanoscopic techniques, such as optical tweezers (OT) (12, 13), magnetic tweezers (MT) (14, 15), biomembrane-force probe (BFP) (16, 17), or atomic force microscopy (AFM) (18, 19). These techniques typically resolve different force ranges: OT and MT at 0.1–100  pN, BFP at 1–100  pN, and AFM at 5  pN to 100 nN. In order to study the interaction properties of molecules, single-molecule force spectroscopy is commonly used. In these experiments, the interacting molecules typically are isolated from cells, purified, and bound to two opposing surfaces that can be brought into contact with each other. Unfortunately, some molecules cannot easily be extracted from cells and immobilized on surfaces. In particular, most membrane-anchored molecules are not soluble in water and hence are difficult to immobilize without destroying their functionality. Immobilizing such transmembrane proteins in artificial lipid membranes on a substrate provides

Measuring Cell Adhesion Forces: Theory and Principles

357

a possibility to access membrane protein unbinding events with force spectroscopy; however, this procedure is relatively complicated (20). An alternative approach to studying interactions of transmembrane proteins is to carry out the adhesion force experiments in  vivo using living cells. This method is often referred to as “single-cell force spectroscopy.” By binding a complete cell to an AFM cantilever and probing the cellular interactions with the extracellular space, transmembrane proteins can be studied in their native environment, and many of the complexities encountered with bilayer-based studies are bypassed. Cellular receptors are provided in their native conformation by the cell, and the applied force is coupled into the molecules by their “natural handles.” This approach opens a broad field to study cellular mechanisms and strategies that tune the adhesion strength of a cell on the molecular level (21–25). OT and MT cannot realize such cell-to-cell adhesion measurements, whereas BFP conceptually incorporates red blood cells into the force detection. Even though these techniques have better resolution in the low force regime, only AFM can study the interaction forces between arbitrary cells. AFM also has the advantage of resolving single molecular forces above 5 pN up to multimolecular interactions in the range of several nanoNewtons. In this chapter, we focus on single-cell force spectroscopy experiments with AFM, from single-molecule interactions to the interaction of intact cell layers. We first introduce the basic concepts of molecular adhesion, then describe experimental prerequisites and basic experimental methods, and finally present exemplary experiments, which demonstrate the power and limits of using this technique.

2. Molecular Concepts in Cell Adhesion

Cell adhesion is mediated by membrane-bound cell adhesion molecules (CAMs), for example integrins, cadherins, selectins, and proteins from the Ig-superfamily (26). These transmembrane proteins interact with their extracellular ligands through structural affinity according to the lock-and-key principle, where several weak bonds align to fit the lock and key together, to yield specificity in their interactions. Individual bonds formed during protein–protein interactions, i.e., Van der Waals interactions, hydrophobic interactions, hydrogen bonds, and electrostatic interactions, are several magnitudes weaker than covalent bonds and are continually competing with thermal energy. However, the relative weakness of specific bonds opens many possibilities for cells to easily manipulate the strength of their adhesion.

358

Benoit and Selhuber-Unkel

A parameter that describes the strength of a protein–protein interaction is affinity. Affinity is commonly characterized by the dissociation constant KD or the off-rate. In force experiments, binding affinity is also represented by the unbinding force and the molecular bond potential (17, 27). Furthermore, the following strategies can be employed by cells to modulate their adhesion strength: Avidity describes the number of binding competent molecules that are available in the membrane to be accessed by the binding partner. The cell can tune avidity by sterically hiding binding partners behind glycosylations or by modifying the expression level of an adhesion receptor. In some molecules (e.g., integrins), avidity can also be changed by switching the affinity state of the receptor. Anchoring refers to how the adhesion receptor is linked to the cell. Receptors that are freely diffusing in the membrane might reach the adhesion site faster than receptors that are spatially confined by a connection to the cytoskeleton. While a pure lipid anchor (e.g., a ceramide anchor) resists only approximately 20 pN, a transmembrane anchor (e.g., bitopic (a-helical) transmembrane proteins, polytopic a-helical proteins, or a transmembrane b barrel) is more resilient with up to approximately 100 pN resistance to failure (28). The strongest group of anchors connects the intracellular domain of transmembrane adhesion receptors to actin, tubulin, or other filaments of the cytoskeleton with binding forces up to the nanoNewton range. The anchorage of an adhesion molecule defines the mechanical micro-environment of the adhesion molecule in that it controls the lateral motility in the membrane and the loading rate of the applied force to the adhesion site (29). By changing the molecular anchoring, the cell has a wide variety of possibilities to modulate its adhesion state. For example, an adhesion molecule anchored in an actin-rich protrusion (microvillus) exposed on its tip has a higher probability to probe an object near the cell than the one situated within the retracted membrane regions. Also, the binding force of a nonanchored membrane protein is limited to the strength of the membrane anchor, even though the affinity of the binding site might be much stronger. When this type of bond is exposed to external forces, the membrane protein is either ripped out of the membrane or forms a membrane tether (see Fig. 1) depending on the membrane properties. A cell can tune the level of the tether force plateau by the lipid composition and thereby trigger adhesion and de-adhesion processes. Clustering is a strategic combination of avidity, affinity, and anchoring that increases the affinity of an adhesion spot with a strong anchor to the cytoskeleton by forming multimers of binding competent adhesion molecules. Integrin-mediated focal adhesions, the immunological synapse, desmosomes, gap ­junctions,

Measuring Cell Adhesion Forces: Theory and Principles

359

Tx membrane tension

tether force Fz

Ty

Fig. 1. A tether is a lipid membrane tube that forms due to applied force. The measured tether force consists of the membrane tension via the tube perimeter under static conditions, and pulling at constant velocity generates a viscous and frictional drag force due to rearrangements of membrane molecules at the foot of the tether.

and tight junctions are a few examples of these cell adhesive ­complexes. In general, adhesion clusters provide a more stable connection than randomly distributed bonds because clusters contain an internal “self-healing mechanism,” where it is possible to rebind broken bonds: bonds which dissociate are not pulled apart because the neighboring molecules in the cluster maintain a close proximity of the interacting surfaces and hence the binding partners (30, 31). With this mechanism, a cluster can withstand forces in the range of several nanoNewtons, i.e., a thousand times more than the strength of a single biomolecular bond.

3. Principles of AFM Experiments with Living Cells 3.1. Experimental Prerequisites of Single-Cell Force Spectroscopy

All of the cells use mechanisms to control adhesion strength mentioned in the previous section can be studied with single-cell force spectroscopy. Successful AFM experiments on living cells require control of a few parameters. An AFM that is employed for live cell experiments is ideally set up in conjunction with a light microscope so that the success of cell detachment can be verified optically and the viability of the studied cells can be monitored. In single-cell force spectroscopy experiments, a cell is bound to a cantilever and approached to a surface (see Fig. 3a). Hence, tipless cantilevers should be used in order to prevent harming the cell when it binds to the cantilever. The stiffness of the cantilevers should be chosen according to the studied processes. For example, the cantilevers should be as soft as possible when studying singlemolecule interactions on fragile cells. Furthermore, a highly important parameter for live cell experiments is the sample temperature. Most mammalian cells are optimized to metabolize at 37°C, and the viscosity of the cellular membrane depends strongly on temperature. Particularly, drastic

360

Benoit and Selhuber-Unkel

changes in the viscosity of the cell membrane take place at the transition temperature from liquid to solid phase of the particular lipid bilayer (32). So, it is essential to carry out the AFM experiments at constant temperature under physiological conditions. However, a disadvantage of carrying out an AFM experiment at a temperature other than room temperature is that drift effects are enhanced due to use of a temperature-sensitive force sensor. One solution to this conundrum is to install the whole microscope and AFM head into a heated box and start the experiments when the sample, the cantilever, and the sensor are thermally stabilized. Suitable functionalization of the cantilever for cell immobilization is also an essential prerequisite for all single-cell force spectroscopy experiments. As the composition of the outer cell membrane can differ greatly from one cell type to the other, coatings normally have to be adapted to each experiment. Several approaches to immobilize cells at the cantilever can be used. A very simple method is to coat the cantilever with lectins (e.g., concanavalin A, wheat germ agglutinin) that bind the glycocalyx of cells (18, 33, 34). A disadvantage of this method is that rupture events between the glycocalyx and the lectin may be mistaken for the actual receptor–ligand ruptures. To circumvent this uncertainty, it is useful to only consider the very last de-adhesion event as a valid measure of the bond of interest. Of course, this analysis requires that the cell is still connected to the cantilever after detachment from the surface. Another solution would be to covalently couple cells to the cantilever, e.g., using glutaraldehyde (35). In most experimental situations, the interaction between cells and artificial ligand-decorated substrates is studied. In order to avoid nonspecific interaction with the substrate, the surface should be passivated. Furthermore, the ligand should ideally be connected via a soft polymeric spacer so that it can freely rotate in space to explore its surroundings and prevent steric hindrance of binding. As the adhesion ligands on a substrate are exposed to forces, they should also be covalently coupled to this surface. One possible method to covalently immobilize proteins on a glass surface, which is the typical surface used in biological experiments, is to first add a layer of functional silanes (e.g., aminosilane) to the surface and then bind the proteins via bifunctional spacing molecules (e.g., glutaraldehyde or carboxy-PEG) (36–39). This technique also can be used with silicon or silicon nitride surfaces (cantilever tips are commonly made of these materials), and can be applied on gold surfaces by using thiol-functionalized molecules, e.g., alkanethiols, instead of silanes (40). In addition, spacer molecules are typically inserted into the system to avoid nonspecific interactions of the cell with the surface. Many cell types unspecifically bind to most surfaces. In order to avoid the nonspecific binding of the cells, it is important to passivate the substrate. Bovine serum ­albumin (BSA) is commonly used to inhibit cell and protein

Measuring Cell Adhesion Forces: Theory and Principles

361

adsorption to surfaces at short timescales. However, BSA often fails to block a surface (41). In such cases, more sophisticated passivation techniques are necessary. In the case of using carboxyPEG functionalization, one can inactivate the remaining ligand-free carboxyl-groups by ethanolamine leaving behind neutral polyethylene glycol (PEG) groups which prevent protein adsorption due to hydrogen bonding interactions with water. Along the same lines, covalent or electrostatic binding of PEG polymers to the surface can also be used (42, 43). A particularly elegant example is to use poly-l-lysine (PLL)-g-PEG for passivation. This polymer binds with its PLL backbone to negatively charged surfaces, thereby exposing PEG polymers to block nonspecific interactions with the surface. By changing the grafting ratio and length of the PLL and the PEG polymers, respectively, the conformation of the polymer on the surface is precisely controlled (43). 3.2. General Features of Cell Adhesion Force Data

Force traces in a cell adhesion experiment are typically highly complex (see Fig. 2). From the force traces, the following parameters can be extracted (34): The initial slope is the initial increase in adhesive force which is approximately linear. After subtracting the spring constant of the cantilever, this parameter basically represents the elastic elements of the cell (see Fig. 5, left arrowhead). The maximum adhesion force indicates the highest force reached in the force plot. This is a rough, first approximation of the adhesion strength. The slope prior to a de-adhesion event gives information on the mechanical properties of the cellular “spacer” (the mechanical

Fig. 2. This representative force–extension curve shows the detachment of a fibroblast (52) cell after 5 s of contact with fibronectin. The curve displays the following information: (1) the elastic response of the cell while pressing it to the surface, (2) single-molecule detachment events, (3) the last adhesion event following tether formation, and (4) maximum adhesion force. The gray area also represents the work of de-adhesion.

362

Benoit and Selhuber-Unkel

environment anchoring the bond(s) that open at the subsequent de-adhesion event). A slope close to zero indicates that a tether (viscous spacer mostly consisting of cell membrane as described below) has formed, whereas a steep slope results from a stiffer elastic spacer. The slope prior to the last de-adhesion event also defines the loading rate applied to the finally ruptured bond, r = dF/dt. The loading rate is a highly important parameter for adhesion force measurements because the strength of a biological bond increases logarithmically with increasing loading rate (44). In the force spectroscopy analysis, the bond rupture forces are plotted as a function of loading rate in order to assess the energy landscape of the adhesive interaction. The distance of a de-adhesion event from the original cell surface is a measure for the lifetime t of a bond, as t = distance/ pulling velocity. Determining the distribution of lifetimes provides a means to investigate the kinetics of biological bonds. The force step size of a de-adhesion event can be used as a lowend estimate of the actual unbinding force. While only the very last de-adhesion event is an exact measure of the unbinding force, the de-adhesion events occurring before the last event might be larger, but appear smaller due to force carrying connections still existing (see Fig.  3). These connections could be mediated by nonindependent cellular components bound between surface and cantilever. The area under the force trace has the dimension of energy. It reflects the work of de-adhesion, which is the energy dissipated by the separation of the cell from the surface rather than a summed “adhesion energy” contributed by individual molecular bonds. The adhesion probability cannot be determined by a single force measurement, but requires a set of at least 50 force curves in order to quantify the fraction of force curves with adhesion events. The bond formation probability is determined by the number of recognized adhesion events either per force curve or, in analogy to the adhesion probability, per all detectable single bonds of all force curves (including curves without adhesion) of the whole set of measurements. Tethers (lipid membrane tubes) are a feature often identified in force–extension curves. When a cell tries to adhere, for example a leukocyte in the blood stream, it initially utilizes membraneanchored weak adhesion molecules. These molecules would detach from the membrane if the adhesion force exceeded the anchorage force of these adhesion molecules. To avoid this phenomenon, cells employ tethers (45), which act similarly to gently releasing fishing line at constant force. In this way, both the force at the binding site and at the anchorage site in the cell membrane is held constant, since cells have a large reservoir of membrane that can be considered unlimited in this case.

Measuring Cell Adhesion Forces: Theory and Principles

363

Fig. 3. (a) This schematic presents the basic setup of the experiment and the homophilic interaction of two csA molecules. (b) The force–distance curve for the interaction of Dictyostelium with a Petri dish after a contact of 20 s shows strong interaction and large tethers with heights of about 100  pN. The zoomed inset shows intermediate de-adhesions that instantly are caught by the same tether. These de-adhesion events show the danger of taking de-adhesion events other than the last event into account, which for sure is not shortened by backup tethers. (c) Superposition of several de-adhesion force traces. They are sorted to show the range of no adhesion (lowest trace) to high adhesion forces. The upper trace in gray was collected from nonspecific adhesion to a Petri dish and displays a typical trace of a single membrane tether. The traces below were collected from csA binding measurements. The lowest traces show no tether formation because this process requires more force than one csA bond can hold (csA de-adhesion is the small shark fin pattern with 23 ± 7 pN). Higher traces form tethers with at least two bonds in parallel (50 ± 10 pN). The zoomed inset resolves such a twofold bond. Interestingly, the tethers formed on a Petri dish are about 100 pN and have a large variety in plateau force and slope.

364

Benoit and Selhuber-Unkel

A membrane tether (see Fig. 1) forms as a lipid bilayer tube, typically 10–200  nm in diameter, which must counterbalance membrane tension and membrane curvature energies (46–49). The diameter also depends on the composition of lipids and proteins in the membrane, the sample temperature (membrane stiffness), and the amount of molecules (e.g., actin filaments) that are pulled within the tube which also enlarge the tube diameter (50). Interestingly, neurons tend to pull extraordinary tethers up to millimeters in length before detaching (51). Typical forces needed to form a tether range from 10 to 100 pN. Tether failure takes place above 200 pN, depending on the maximum capable membrane tension and the membrane stiffness. In a force experiment where the cantilever is retracted with constant velocity, a membrane tether can be identified as a nearly constant force plateau in the force trace prior to a detachment event (see Fig. 2, 3, and 5). In this situation, the loading rate of the bond is close to zero and the tether acts as force clamp. When a tether is pulled at constant velocity, a steady flow of lipids into the growing tether is recruited from the cell membrane. This constant force is proportional to the pulling velocity and is determined by friction and viscosity at the origin of the tether on the cell membrane. This growth force exists in addition to the constant force generated by the membrane tension. The maximum membrane tension and bending rigidity of the tether largely depend on the lipid composition and temperature (32, 52, 53). Tethers pulled from cells can vary in diameter and visco-elastic behavior. If actin bundles or membrane proteins are pulled within the tether or if the membrane tension is low, the tether radius can increase to a few hundred nanometers.

4. Force Spectroscopy Experiments on Living Cells

Cells are the smallest units of life. As individuals (single-cell organisms such as amoebae or slime molds), they have adapted very well to the environment during evolution, and in multicellular organisms, they have to adequately react to several environmental changes in order to survive. Hence, a division of labor has been devised, where cells become specialists for certain tasks (e.g., immune cells, neurons, and endothelial cells) in order to provide a better means of adapting to environmental changes or even change the environment. Intercellular communication, differentiation, migration, and many other functions have to be maintained in such a multicellular organism. Therefore, some cellular reactions are universal while some reactions are present only in heart muscle cells, inner ear cells, red blood cells, and so on. Thus, there is no universal protocol for cell adhesion

Measuring Cell Adhesion Forces: Theory and Principles

365

e­ xperiments. In the worst case, a new protocol has to be ­established for each cell type. Nevertheless, a few basic principles and examples for cell adhesion measurements with AFM are described in the following sections. 4.1. Single Cell-to-Cell Measurements

The social amoeba Dictyostelium discoideum is a highly interesting organism biologically (54, 55). D. discoideum is fascinatingly diverse in appearance: if food runs short, the cell switches active genes in the nucleus and changes from a unicellular to a multicellular organism – a slug. During this process, Dictyostelium cells meet by a hot spot of a chemokine signal (cAMP) that is sent out from every cell undergoing the change. These switched cells then start to produce a Ca2+-independent lipid-anchored glycoprotein in the extracellular membrane called contact site A (csA). During the development of the multicellular Dictyostelium slug, csA plays an essential role as the homophilic binding between two csA molecules supports cell aggregation. In order to study this homophilic interaction between two individual csA molecules from different cells, cell–cell adhesion force measurements can be carried out. For studying such single-molecule interactions, the number of other interacting molecules should be as small as possible. One can reduce non-csA binding by removing Ca2+ from the environment for a few hours prior to the experiment because many adhesion proteins lose their binding function without divalent cations although csA remains active (56). After this treatment, the adhesion probability is typically below 3% (nonspecific interaction) for nonswitched wild-type cells after 0.1 s of cell–cell contact at 100  pN contact force. On the other hand, for csAexpressing Dictyostelium cells, the adhesion probability is typically 35% after 0.1 s cell–cell contact at 30 pN contact force. Typical force curves for the separation of interacting Dictyostelium cells show a variety of force signals that are received from the same molecular interaction (see Fig. 3). This variety could be due to the visco-elastic deformation and the plastic activity of the interacting cells causing a significant dissipation of energy. In order to receive the true information about csA–csA binding events, only de-adhesion events from the very last intercellular contact are taken into account in the analysis of adhesion forces (see Fig. 4b, arrowheads). For the interaction between two individual csA molecules, the adhesion force measurements shown here revealed that the most probable de-adhesion force was 23 pN. The forces varied between 19 pN at loading rates of 20 pN/s and 28 pN at 8 nN/s. After prolonged contacts of 1 and 2 s between two Dictyostelium cells, the force histograms showed pronounced force peaks at multiples of 23 pN. A critical issue when analyzing the measured forces is interpreting where the bond rupture occurs. The csA molecule is

366

Benoit and Selhuber-Unkel

b

a

csA expressed in the membrane

no csA 2sec

43 % 3 %

0.2sec force

0

100pN

100pN

1

3

1

2

3

0

1

2

3

4 [µm]

%

0

0

20 40 60 80 rupture force pN

100

force

0,2sec 100 pN

0

40 60 80 rupture force pN

100

66 %

1 0

20

3 %

0,2 sec 40 pN

2 1

100pN

%

100

1 sec 40 pN

2

96 % 2

40 60 80 rupture force pN

1

4 [µm] 0.2sec

3

20 40 %

0 0

2 sec 40 pN

2

0 0 1

2

3

4 [µm]

0

20

40 60 80 rupture force pN

100

Fig. 4. (a) Dictyostelium discoideum in the amoebous state do not express csA molecules, and the remaining intercellular interaction is effectively blocked by the added EDTA. (b) Starved Dictyostelium cells express csA, and the adhesion signal of individual interacting csA molecules is measured by the AFM. Arrows on the force traces indicate the last de-adhesion event counted in the histograms. The most probable de-adhesion force is extracted from the histograms (from 20 pN at short contacts to 23 pN for longer contacts). The percentage of force curves where no adhesion occurs is indicated by the truncated bar to the left side of the histograms. Contact time and contact forces are indicated above the histograms.

known to be only weakly anchored to the cell with a ceramide anchor in the external lipid layer of the membrane. In order to test whether the bond between two csA molecules breaks or if the anchor itself is extracted from the lipid bilayer in an AFM experiment, a genetically modified mutant of Dictyostelium expressing csA with transmembrane anchor was employed. In these experiments, no significant change in the adhesion pattern was observed. Therefore, the anchor is believed to be at least as strong as the molecular interaction, and thus csA is extracted from the membrane in less than 50% of the adhesion events (33). Tether formation is not often observed for csA-mediated Dictyostelium-to-Dictyostelium adhesion. However, when Dictyo­ stelium cells adhere to Petri dishes, tethers dominate the force curves. The pseudopodia-rich surface structure is thought to support tether formation (57). In this case, tethers can also reach lengths of several tens of micrometers and more (see Fig. 3b).

Measuring Cell Adhesion Forces: Theory and Principles

367

4.2. Spheres on Cells

Targeting therapeutic delivery to selected cells within an organism is a desired goal of pharmacology. In particular, designing vehicles loaded with a drug which targets only specific pathogenic cells is a strong aim in medical research. For the design of such vehicles, the molecular composition of the external surface of the vehicle should be specific for binding to a certain type of target cell. Medically relevant particles are often spherical and consist of polymers that are functionalized on the surface. In adhesion force experiments, a small vehicle can be mimicked by a sphere of 5 mm radius, for example, that is immobilized to an AFM cantilever. In order to quantify the interaction with different cell types and to find particle coatings specific for the binding to particular cell types, the initial binding force between the functionalized vehicle and a cell can be quantified. In the experiments described here, the adhesion of cells to two different types of surfaces, positively (NH2) and negatively (COOH) charged spheres, has been investigated. Two types of breast cancer cell lines were measured in this study, the noninvasive strain MCF-10A and the invasive strain MDA-MB-4355 (58). The analysis of the force traces with respect to adhesion probability and adhesion forces on the single molecular level (while maintained at 37°C in nutrient medium) showed a higher adhesion probability for the positively charged spheres. This result is likely due to the presentation of a negative net charge by these cells due to expression of carbohydrate groups in the glycocalyx, leading to stronger binding on positively charged spheres. The most probable “molecular” adhesion force is increased from 20 pN for negatively charged spheres to 25 pN for the positively charged spheres after contacts of 1 ms at 50 pN. Furthermore, an increase in adhesion probability from 20% for negatively charged spheres to almost 80% for positively charged spheres was observed for the MCF-10A cell line compared to an only moderate increase from 30% to almost 50% for the MDA-MB-4355 cell line. Hence, the MCF-10A cells appear much more negatively charged than the MDA-MB-4355 cells.

4.3. Cell Layer to Surface Measurements

The examples of cell adhesion force measurements presented so far sought to measure initial adhesion or fast molecular processes on the level of single molecules. These measurements are both important to our understanding of adhesion processes and they are feasible. In contrast, long-term adhesion processes are difficult to measure in a force experiment. For example, cells that initially seem to like a surface might decide to push it away after having explored it for an hour. But how will the adhesion forces involved in binding to artificial bones and implants develop? In this case, it is essential that long-term adhesion is stable and viable. With a bone cell layer on a cantilever, potential implant surfaces can be probed to find out the best surface for durable

368

Benoit and Selhuber-Unkel

force [nN]

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2

maximum force

20

10

tethering

56

57

58

59

60

61

62

63

distance [µm]

50

100

Fig. 5. Schematic series of a fibronectin-coated sphere microfabricated to the cantilever pulled back from a cell layer of epithelial cells (RL cells also described in Fig. 8) after a contact of 20 min (clock symbol ) at 5 nN. The corresponding force–distance trace shows a Hookean region (left arrow), the maximum adhesion force, and a descending shoulder consisting of many small force steps (zoomed inset shows the tether region) from individual molecular interactions.

acceptance of implants by the adhering cells (59). The cell-tosurface contacts can be prolonged to several minutes, maybe up to an hour, but then drift becomes a limiting factor and force spectroscopy is not applicable anymore. In force experiments with cell layers, the number of cells interacting with the surface is unknown and many adhesion molecules contribute to the force trace in parallel. A typical force graph of a fibronectin-coated sphere mounted to the cantilever after a contact of 20 min at 5 nN on a layer of confluent cells is shown in Fig. 5. After the contact, an almost Hookean stretching of the cell layer (left arrow) takes place until, by an increasing number of dissociating bonds and the progression of membrane and cytoskeleton disentanglement, the maximum force is reached. The measured maximum adhesion forces are up to three orders of magnitudes higher than in a single molecule experiment. The large maximum adhesion force of 20  nN is the sum of several hundred or thousands of single molecules, each contributing with its weak individual adhesion force. Some of these contributing molecules are still resolved as individual de-adhesion events when zooming into the tethering region in the force traces’ descending shoulder. An adhesive interaction length of several tens of micrometers and forces larger than 3  nN, as observed here, is not observed in single-molecule force measurements.

Measuring Cell Adhesion Forces: Theory and Principles

4.4. Cell Adhesion Clusters

369

The formation of molecular clusters is a ubiquitous mechanism that leads to strong mechanical interactions in cellular systems. Particularly relevant to the adhesion between cells and the extracellular matrix are focal adhesion clusters, often called “focal contacts.” Focal contacts are formed in cells subsequent to the activation and clustering of integrin transmembrane adhesion receptors (60). Integrin activation can be induced by binding to extracellular matrix proteins, such as fibronectin, and in response to integrin activation, further integrin molecules are recruited to the binding site and cluster. After this initial clustering process, a hierarchically organized plaque of intracellular proteins accumulates intracellularly at the integrin cluster. This plaque is also responsible for connecting the integrin cluster with the cellular cytoskeleton. Proteins in the plaque, such as talin, vinculin, and a-actinin, are very important to these structures because these proteins are supposed to be the first proteins that bind to the early integrin cluster (61, 62). Focal contacts have many biological functions. For example, they are responsible for a large number of signal transduction events, and they serve as cellular mechanosensors by “feeling” external forces and actively probing the mechanical properties of the cellular environment (3, 63). Due to the complex organization and function of focal contacts, it can be expected that their formation and function rely on a well-defined, hierarchical organization of the proteins in the cluster which is already critical during the initial binding and clustering processes. Hence, it is highly interesting to investigate the requirements necessary for the formation of the initial stages of focal contacts. Using nanolithographical techniques, it is possible to impose nanometer-sized binding sites for integrins on surfaces. A very elegant method is diblock-copolymer micelle nanolithography. With this method, nanometer-sized gold dots can be arranged in hexagonal patterns on standard glass coverslips, where the spacing between the individual gold dots can be defined between approximately 20 and 400 nm with nanometer precision (64–66). After deposition, the gold dots can be functionalized with thiolterminated adhesion ligands, e.g., RGD peptides so that they finally serve as nanometer-sized ligand patches for the binding of avb3 integrin adhesion receptors (see Fig. 6) (67). Using these nanostructured surfaces, it has been shown that the extent of integrin-mediated cell spreading and the formation of focal contacts are critically defined by the spacing between RGD ligand patches. In particular, above a ligand patch spacing of approx. 70  nm, cell spreading, proliferation, and focal contact formation are strongly inhibited (71). These results indicate that the molecular binding affinity between the avb3 integrin and the RGD is not sufficient for inducing focal contact formation, but their intermolecular spatial arrangement must also fulfill certain requirements.

370

Benoit and Selhuber-Unkel

Fig. 6. Left : Scanning electron microscope image shows the hexagonal pattern of gold nanodots on a glass surface generated using the block copolymer technique. Right : This schematic shows the most likely bio-functionalization pattern of the nanostructured surface used in studies of integrin-mediated adhesion. The polyethylene glycol (PEG) coating prevents nonspecific protein adsorption and cell adhesion to the glass surface, ensuring that the cellular integrins only bind to the gold dots.

In order to study the role of the intermolecular arrangement of integrins and the timescale of the onset of integrin cluster formation, we characterized cell detachment forces as a function of cell adhesion time using an AFM. In the experiments, a fibroblast cell (52) was coupled to a cantilever with concanavalin A and brought into contact with a nanostructured surface for a defined time span between a few seconds to several minutes. From the experiments, it is immediately clear that nanostructures providing integrin-binding site spacings larger than 60 nm lead to very different detachment forces compared to spacings smaller than 60  nm (see Fig.  7). For 35- and 55-nm integrinbinding site spacing, the detachment forces increase to more than 1 nN within 40 s. For larger spacings, detachment forces do not exceed 500  pN and stay almost constant with adhesion time. This result suggests that cell adhesion is reinforced for spacings smaller than 60 nm, whereas adhesion cannot normally develop for spacings larger than 60 nm (72). We believe that the observed reinforcement of adhesion results from a cooperative clustering of integrin receptors that can only take place for spacings smaller 60 nm. In a focal contact, the cooperative clustering of integrins is supported by secondary proteins that serve as cross-linkers between the individual integrin molecules. However, if the integrin molecules are located too far apart from each other, such an interconnection might fail to form. Prominent candidates for intracellularly cross-linking integrins are talin (length approx. 60  nm) and a-actinin (heterodimer length approx. 24 nm). However, biological proof of the involvement of these proteins in the reinforcement of adhesion on RGD nanostructures has not yet been demonstrated.

Measuring Cell Adhesion Forces: Theory and Principles

371

Fig. 7. These results of force versus inter-integrin binding site distance were taken from single-cell force microscopy measurements of fibroblasts on nanostructured surfaces. At 35 and 55 nm integrin binding site spacings, the cell detachment force increases to approximately 1 nN within 40 s of contact. Above 60 nm inter–integrin spacing, the detachment forces increase only slightly with time, and there is no significant difference observed between the forces at 70 and 103 nm. Error bars refer to the standard error of the mean.

4.5. Cell Layers on Cell Layers

Experiments between two cell layers present a highly natural cellular environment; however, this situation is extremely complex to analyze. In this layered configuration, cells can polarize, form intercellular connection centers, such as adhesion clusters, and carry out typical habits of epithelial cells. However, in these experiments, neither the surface geometry of the two layers is defined nor can parameters, such as the indentation force and the elasticity of the interacting cells, be calculated. When the two cell layers are brought into contact, they might start to communicate and also establish complex adhesion patterns since thousands of adhesion molecules are contributing to the measured de-adhesion forces (68). A biological situation where the adhesion between cell layers becomes relevant is the adhesion between trophoblast cells covering the few embryonic cells after fertilization and the uterine epithelial cell layer. In nature, these cells establish the homing of the embryo. For studying this interaction in AFM experiments, the JAR-cell line was used to form the spherical trophoblast cell layer. On a 60-mm sphere, the natural configuration of the trophoblast structure is resembled best in size, shape, and cellular arrangement of the apical region. Receptive uterine epithelial cells (RL95-2 cell line) or nonreceptive uterine epithelial cells (HEC1-A cell line) were cultured in Petri dishes and resemble either uterine epithelial layer. Both cell layers where held in contact with

372

Benoit and Selhuber-Unkel

a

b

JAR

force [nN] 15

HEC

10

40 min 20 min 10 min 1 min

5

15

RL

10

5

0

0

−5 distance

JAR

force [nN]

−5 20

40

µm60

distance 20

40

µm60

Fig. 8. (a) Representative force traces from nonreceptive HEC cells show increasing maximum adhesion forces with a contact time of 1–40 min after cell layer contact at 5 nN. The individual bonds open sequentially. (b) Representative force traces from receptive RL cells show a similar increase in the maximum adhesion force; however, the work of de-adhesion is drastically increased after 20 min of contact and is marked by strong de-adhesion events after mechanical cell stretching over tens of micrometers.

the trophoblast sphere for several minutes at 5  nN before the ­cantilever was retracted (see Fig. 8). An important question for fertilization and pregnancy is how long it would take to firmly arrest the trophoblast layer on either cell layer? When comparing the maximum adhesion forces for trophoblast spheres to HEC or RL cell layers at adhesion times between 1 and 10 min, a stronger adhesion to the HEC cells is observed. Furthermore within 10 min, the adhesion energy dissipation is enhanced for contacts between RL and JAR cell layers. After 20  min, the RL cells firmly connect their adhesion molecules into clusters so that they are strongly connected with the cytoskeleton of the JAR cell layer (69). Importantly, the maximum adhesion force was not able to report this phenomenon, indicating that it is indeed essential to compare different analysis methods in such experiments. Analyzing the force value of the dominating de-adhesion rupture event revealed a force of 15 nN, which suggests this rupture is due to the failure of molecular clusters. From the force measurements of single integrins (a4b1), a typical detachment force appears to be between 20 and 60 pN at loading rates of 10–100 pN/s. Hence, approximately 100–1,000 molecular bonds contribute in the measured clusters. All of these molecules must be connected to the cytoskeleton, or the cluster would separate from the cell and form a tether that breaks at a force of about 300 pN (41).

Measuring Cell Adhesion Forces: Theory and Principles

373

When checking the viability of the cells after such a strong de-adhesion event, at least one of the cell layers appeared to be severely damaged (cells of either layer were found loosened or even sticking to the opposite layer) so that the experiment cannot be carried out repeatedly as in single-molecule experiments. Although these experiments, therefore, require more preliminary setup time and equipment than single-molecule experiments, the technique provides useful information and quantification for cellular communication between cell layers that would otherwise not be obtained.

5. Perspectives In this chapter, we have described the experimental principles of single-cell force spectroscopy and how it can be applied for answering biophysical questions from the level of single molecules to molecular clusters and even multicellular interactions. Although the described experiments may look straightforward at first glance, it is important to remember that each cell might react differently due to the cell cycle, last feeding period, temperature changes during preparation, or the exerted force. Furthermore, a cell as a “spacer” for the molecular bond to be investigated contains several billions of molecules. This complex interplay among all of these factors renders the mechanical characteristics of the cell, since it is the linker between the adhesion molecules and the substrate or the force sensor. Hence, such cell experiments are always associated with experimental and statistical errors, and a large number of experiments must be carried out in order to receive appropriately representative information. As cells might change their adhesion when reacting to their environment, a cell adhesion experiment can even serve as a reporter for the function of pharmaceuticals (drugs, hormones, and chemokines) that trigger an intracellular reaction which has an impact on adhesion (21). Force spectroscopy can not only characterize antibodies with respect to their interaction force with their specific ligand, but might also identify diseases caused by malfunction of cellular adhesion and optimize related medication. AFM can also be applied for experiments in electrophysiology. Planar patch-clamp technology turned out to be not only a perfect platform for AFM measurements on non-adherent cells, but also an extension toward simultaneous electrophysiological measurements. This application is extremely attractive for pharmacological research. Here, for example, the mechanical signal from the cell can be correlated in time with activity of membrane pores (70).

374

Benoit and Selhuber-Unkel

In conclusion, AFM experiments on cells are becoming more and more interesting for medical scientists and pharmacists, and many new aspects of cell adhesion might be discovered during the next years.

Acknowledgments We acknowledge the people involved in the projects described in this chapter, namely, Hermann Gaub, Günther Gerisch, Joachim Spatz, Horst Kessler, Angelika Kardinal, Thomas Nicolaus, and Michael Thie. Furthermore, we thankfully dedicate this chapter to Kristin Michael. References 1. Bischofs, I. and Schwarz, U. S. (2003) Cell organization in soft media due to active mechanosensing. Proc. Natl. Acad. Sci. U. S. A. 100, 9274–9279. 2. Discher, D. E., Janmey, P. and Wang, Y.-l. (2005) Tissue Cells Feel and Respond to the Stiffness of Their Substrate. Science 310, 1139–1143. 3. Geiger, B. and Bershadsky, A. (2002) Exploring the Neighborhood: Adhesion-Coupled Cell Mechanosensors. Cell 110, 139–142. 4. Tan, J. L., Tien, J., Pirone, D. M., Gray, D. S., Bhadriraju, K. and Chen, C. S. (2003) From the Cover: Cells lying on a bed of microneedles: An approach to isolate mechanical force. Proc. Natl. Acad. Sci. U. S. A. 100, 1484–1489. 5. Burton, K. and Taylor, D. L. (1997) Traction forces of cytokinesis measured with optically modified elastic substrata. Nature 385, 450–454. 6. Dogterom, M., Kerssemakers, J. W., RometLemonne, G. and Janson, M. E. (2005) Force generation by dynamic microtubules. Curr. Opin. Cell Biol. 17, 67–74. 7. Paul, R., Heil, P., Spatz, J. P. and Schwarz, U. S. (2008) Propagation of Mechanical Stress through the Actin Cytoskeleton toward Focal Adhesions: Model and Experiment. Biophys. J. 94, 1470–1482. 8. Yamada, S., Wirtz, D. and Kuo, S. C. (2000) Mechanics of Living Cells Measured by Laser Tracking Microrheology. Biophys. J. 78, 1736–1747. 9. Puchner, E. M., Alexandrovich, A., Kho, A. L., Hensen, U., Schafer, L. V., Brandmeier, B., Grater, F., Grubmuller, H., Gaub, H. E. and Gautel, M. (2008) Mechanoenzymatics of

titin kinase. Proc. Natl. Acad. Sci. U. S. A. 105, 13385–13390. 10. Vogelsang, J., Cordes, T., Forthmann, C., Steinhauer, C. and Tinnefeld, P. (2009) Controlling the fluorescence of ordinary oxazine dyes for single-molecule switching and superresolution microscopy. Proc. Natl. Acad. Sci. U. S. A. 106, 8107–8112. 11. Mizuno, D., Tardin, C., Schmidt, C. F. and MacKintosh, F. C. (2007) Nonequilibrium Mechanics of Active Cytoskeletal Networks. Science 315, 370–373. 12. Grier, D. G. (2003) A revolution in optical manipulation. Nature 424, 810–816. 13. Mehta, A. D., Rief, M., Spudich, J. A., Smith, D. A. and Simmons, R. M. (1999) SingleMolecule Biomechanics with Optical Methods. Science 283, 1689–1695. 14. Smith, A.-S., Sengupta, K., Goennenwein, S., Seifert, U. and Sackmann, E. (2008) Forceinduced growth of adhesion domains is controlled by receptor mobility. Proc. Natl. Acad. Sci. U. S. A. 105, 6906–6911. 15. Walter, N., Selhuber, C., Blümmel, J., Kessler, H. and Spatz, J. P. (2006) Cellular Unbinding Forces of Initial Adhesion Processes on Nanopatterned Surfaces Probed with Magnetic Tweezers. Nano Lett. 6, 398–402. 16. Prechtel, K., Bausch, A. R., Marchi-Artzner, V., Kantlehner, M., Kessler, H. and Merkel, R. (2002) Dynamic force spectroscopy to probe adhesion strength of living cells. Phys. Rev. Lett. 89, 028101. 17. Merkel, R., Nassoy, P., Leung, A., Ritchie, K. and Evans, E. (1999) Energy landscapes of receptor-ligand bonds explored with dynamic force spectroscopy. Nature 397, 50–53.

Measuring Cell Adhesion Forces: Theory and Principles 18. Grandbois, M., Dettmann, W., Benoit, M. and Gaub, H. E. (2000) Affinity imaging of red blood cells using an atomic force microscope. J. Histochem. Cytochem. 48, 719–724. 19. Rief, M., Gautel, M., Oesterhelt, F., Fernandez, J. M. and Gaub, H. E. (1997) Reversible unfolding of individual titin immunoglobulin domains by AFM. Science 276, 1109–1112. 20. Dewa, T., Sugiura, R., Suemori, Y., Sugimoto, M., Takeuchi, T., Hiro, A., Iida, K., Gardiner, A. T., Cogdell, R. J. and Nango, M. (2006) Lateral organization of a membrane protein in a supported binary lipid domain: direct observation of the organization of bacterial light-harvesting complex 2 by total internal reflection fluorescence microscopy. Langmuir 22, 5412–5418. 21. Schmitz, J. and Gottschalk, K. E. (2008) Mechanical regulation of cell adhesion. Soft Matter, 1373–1387. 22. Parot, P., Dufrene, Y. F., Hinterdorfer, P., Le Grimellec, C., Navajas, D., Pellequer, J. L. and Scheuring, S. (2007) Past, present and future of atomic force microscopy in life sciences and medicine. J. Mol. Recognit. 20, 418–431. 23. Ludwig, T., Kirmse, R., Poole, K. and Schwarz, U. S. (2008) Probing cellular microenvironments and tissue remodeling by atomic force microscopy. Pflugers Arch. 456, 29–49. 24. Helenius, J., Heisenberg, C. P., Gaub, H. E. and Muller, D. J. (2008) Single-cell force spectroscopy. J. Cell. Sci. 121, 1785–1791. 25. Selhuber-Unkel, C., Lopez-Garcia, M., Kessler, H. and Spatz, J. P. (2008) Cooper­ ativity in adhesion cluster formation during initial cell adhesion. Biophys. J. 95, 5424–31. 26. Brackenbury, R. (1989) Cell adhesion molecules. Annu. Rep. Med. Chem. 25, 235–244. 27. Oberbarnscheidt, L., Janissen, R. and Oesterhelt, F. (2009) Direct and Model Free Calculation of Force-Dependent Dissociation Rates from Force Spectroscopic Data. Biophys. J. 97 L01–03. 28. Oesterhelt, F., Oesterhelt, D., Pfeiffer, M., Engel, A., Gaub, H. E. and Muller, D. J. (2000) Unfolding pathways of individual bacteriorhodopsins. Science 288, 143–146. 29. Schmitz, J., Benoit, M. and Gottschalk, K. E. (2008) The viscoelasticity of membrane tethers and its importance for cell adhesion. Biophys. J. 95, 1448–1459. 30. Erdmann, T. and Schwarz, U. S. (2006) Bistability of cell-matrix adhesions resulting from nonlinear receptor-ligand dynamics. Biophys. J. 91, L60–62. 31. Erdmann, T. and Schwarz, U. S. (2004) Stability of adhesion clusters under constant force. Phys. Rev. Lett. 92, 108102 (DOI: 10.1103/

375

PhysRevLett.92.108102) (http://prl.aps.org/ abstract/PRL/v92/i10/e108102). 32. Heimburg, T. 2007. Front Matter. In Thermal Biophysics of Membranes. Wiley-VCH, Berlin. I–XV. 33. Benoit, M., Gabriel, D., Gerisch, G. and Gaub, H. E. (2000) Discrete interactions in cell adhesion measured by single-molecule force spectroscopy. Nat. Cell Biol. 2, 313–317. 34. Franz, C. M., Taubenberger, A., Puech, P. H. and Muller, D. J. (2007) Studying integrinmediated cell adhesion at the single-molecule level using AFM force spectroscopy. Sci STKE 2007, pl5. 35. Gladilin, E., Micoulet, A., Hosseini, B., Rohr, K., Spatz, J. and Eils, R. (2007) 3D finite element analysis of uniaxial cell stretching: from image to insight. Phys.l Biol. 4, 104–113. 36. Dettmann, W., Grandbois, M., Andre, S., Benoit, M., Wehle, A. K., Kaltner, H., Gabius, H. J. and Gaub, H. E. (2000) Differences in zero-force and force-driven kinetics of ligand dissociation from beta-galactoside-specific proteins (plant and animal lectins, immunoglobulin G) monitored by plasmon resonance and dynamic single molecule force microscopy. Arch. Biochem. Biophys. 383, 157–170. 37. Micoulet, A., Spatz, J. P. and Ott, A. (2005) Mechanical Response Analysis and Power Generation by Single-Cell Stretching. ChemPhysChem 6, 663–670. 38. Stenger, D. A., Georger, J. H., Dulcey, C. S., Hickman, J. J., Rudolph, A. S., Nielsen, T. B., McCort, S. M. and Calvert, J. M. (1992) Coplanar molecular assemblies of amino- and perfluorinated alkylsilanes: characterization and geometric definition of mammalian cell adhesion and growth. J. Am. Chem. Soc. 114, 8435–8442. 39. Kufer, S. K., Puchner, E. M., Gumpp, H., Liedl, T. and Gaub, H. E. (2008) Singlemolecule cut-and-paste surface assembly. Science 319, 594–596. 40. Yousaf, M. N. and Mrksich, M. (1999) DielsAlder Reaction for the Selective Immobilization  of Protein to Electroactive Self-Assembled Monolayers. J. Am. Chem. Soc. 121, 4286–4287. 41. Thie, M., Rospel, R., Dettmann, W., Benoit, M., Ludwig, M., Gaub, H. E. and Denker, H. W. (1998) Interactions between trophoblast and uterine epithelium: monitoring of adhesive forces. Hum. Reprod. 13, 3211–3219. 42. Blümmel, J., Perschmann, N., Aydin, D., Drinjakovic, J., Surrey, T., Lopez-Garcia, M., Kessler, H. and Spatz, J. P. (2007) Protein repellent properties of covalently attached

376

Benoit and Selhuber-Unkel

PEG coatings on nanostructured SiO2-based interfaces. Biomaterials 28, 4739–4747. 43. Pasche, S., Textor, M., Meagher, L., Spencer, N. D. and Griesser, H. J. (2005) Relationship between Interfacial Forces Measured by Colloid-Probe Atomic Force Microscopy and Protein Resistance of Poly(ethylene glycol)Grafted Poly(L-lysine) Adlayers on Niobia Surfaces. Langmuir 21, 6508–6520. 44. Evans, E. and Ritchie, K. (1997) Dynamic strength of molecular adhesion bonds. Biophys. J. 72, 1541–1555. 45. Waugh, R. E. and Hochmuth, R. M. (1987) Mechanical equilibrium of thick, hollow, liquid membrane cylinders. Biophys. J. 52, 391–400. 46. Marcus, W. D. and Hochmuth, R. M. (2002) Experimental studies of membrane tethers formed from human neutrophils. Ann. Biomed. Eng. 30, 1273–1280. 47. Raucher, D. and Sheetz, M. P. (1999) Characteristics of a membrane reservoir buffering membrane tension. Biophys. J. 77, 1992–2002. 48. Harmandaris, V. A. and Deserno, M. (2006) A novel method for measuring the bending rigidity of model lipid membranes by simulating tethers. J. Chem. Phys. 125, 204905. 49. Sun, M., Graham, J. S., Hegedus, B., Marga, F., Zhang, Y., Forgacs, G. and Grandbois, M. (2005) Multiple membrane tethers probed by atomic force microscopy. Biophys. J. 89, 4320–9. 50. Hosu, B. G., Sun, M., Marga, F., Grandbois, M. and Forgacs, G. (2007) Eukaryotic membrane tethers revisited using magnetic tweezers. Phys. Biol. 4, 67–78. 51. Hochmuth, F. M., Shao, J. Y., Dai, J. and Sheetz, M. P. (1996) Deformation and flow of membrane into tethers extracted from neuronal growth cones. Biophys. J. 70, 358–69. 52. Sackmann, E. 1995. Physical basis of self-organization and function of membranes: physics of vesicles. In Structure and Dynamics of Membranes. R. Lipowsky and E. Sackmann, editors. Elsevier, Amsterdam. 213–298. 53. Seifert, U. and Lipowsky, R. 1995. The morphology of vesicles. In Structure and Dynamics of Membranes. R. Lipowsky and E. Sackmann, editors. Elsevier, Amsterdam. 403–463. 54. Bozzaro, S., Fisher, P. R., Loomis, W., Satir, P. and Segall, J. E. (2004) Guenther Gerisch and Dictyostelium, the microbial model for ameboid motility and multicellular morphogenesis. Trends Cell Biol. 14, 585–8. 55. Jin, T. and Hereld, D. (2006) Moving toward understanding eukaryotic chemotaxis. Eur. J. Cell Biol. 85, 905–913.

56. Beug, H., Katz, F. E. and Gerisch, G. (1973) Dynamics of antigenic membrane sites relating to cell aggregation in Dictyostelium discoideum. J. Cell Biol. 56, 647–688. 57. Gerisch, G. and Weber, I. (2007) Toward the structure of dynamic membrane-anchored actin networks: an approach using cryo-electron tomography. Cell Adh. Migr. 1, 145–148. 58. Munoz Javier, A., Kreft, O., Piera Alberola, A., Kirchner, C., Zebli, B., Susha, A. S., Horn, E., Kempter, S., Skirtach, A. G., Rogach, A. L., Radler, J., Sukhorukov, G. B., Benoit, M. and Parak, W. J. (2006) Combined atomic force microscopy and optical microscopy measurements as a method to investigate particle uptake by cells. Small 2, 394–400. 59. Benoit, M. and Gaub, H. E. (2002) Measuring cell adhesion forces with the atomic force microscope at the molecular level. Cells Tissues Organs 172, 174–189. 60. Cluzel, C., Saltel, F., Lussi, J., Paulhe, F., Imhof, B. A. and Wehrle-Haller, B. (2005) The mechanisms and dynamics of alphavbeta3 integrin clustering in living cells. J. Cell Biol. 171, 383–292. 61. Geiger, B., Bershadsky, A., Pankov, R. and Yamada, K. M. (2001) Transmembrane extracellular matrix-cytoskeleton crosstalk. Nat. Rev. Mol. Cell Biol. 2, 793–805. 62. Zaidel-Bar, R., Itzkovitz, S., Ma’ayan, A., Iyengar, R. and Geiger, B. (2007) Functional atlas of the integrin adhesome. Nat. Cell Biol. 9, 858–867. 63. Vogel, V. and Sheetz, M. (2006) Local force and geometry sensing regulate cell functions. Nat. Rev. Mol. Cell Biol. 7, 265–275. 64. Haupt, M., Miller, S., Ladenburger, A., Sauer, R., Thonke, K., Spatz, J. P., Riethmüller, S., Möller, M. and Banhart, F. (2002 ) Semiconductor nanostructures defined with self-organizing polymers. J Appl. Phys. 91, 6057–6059. 65. Glass, R., Möller, M. and Spatz, J. P. (2003) Block copolymer micelle nanolithography. Nanotechnology 14, 1153–1160. 66. Spatz, J. P., Mößmer, S., Hartmann, C., Möller, M., Herzog, T., Krieger, M., Boyen, H., Ziemann, P. and Kabius, B. (2000) Ordered Deposition of Inorganic Clusters from Micellar Block Copolymer Films. Langmuir 16, 407–415. 67. Wolfram, T., Belz, F., Schoen, T. and Spatz, J. P. (2007) Site-specific presentation of single recombinant proteins in defined nanoarrays. Biointerphases 2, 44–48. 68. Pierres, A., Prakasam, A., Touchard, D., Benoliel, A. M., Bongrand, P. and Leckband, D.

Measuring Cell Adhesion Forces: Theory and Principles (2007) Dissecting subsecond cadherin bound states reveals an efficient way for cells to achieve ultrafast probing of their environment. FEBS Lett. 581, 1841–1846. 69. Thie, M., Herter, P., Pommerenke, H., Dürr, F., Sieckmann, F., Nebe, B., Rychly, J. and Denker, H.-W. (1997) Adhesiveness of the free surface of a human endometrial monolayer as related to actin cytoskeleton. Mol. Hum. Reprod. 3, 275–283. 70. Pamir, E., George, M., Fertig, N. and Benoit, M. (2008) Planar patch-clamp force microscopy on living cells. Ultramicroscopy 108, 552–557.

377

71. Arnold, M., Cavalcanti-Adam, A., Glass, R., Blümmel, J., Eck, W., Kessler, H., and Spatz, J. P. (2004) Activation of integrin function by nanopatterned adhesive interfaces, Chem.Phys. Chem. 3, 383. (http://onlinelibrary.wiley. com/doi/10.1002/cphc.200301014/ abstract). 72. Selhuber-Unkel, C., López-García, M., Kessler, H., and Spatz, J. P. (2008) Cooperativity in adhesion cluster formation during initial cell adhesion. Biophysi. J. 95, 5424–5431. (http:// www.cell.com/biophysj/abstract/S0006-3495 (08)78965-3)].

wwwwwwwwwwwwww

Chapter 22 Nanoscale Investigation on E. coli Adhesion to Modified Silicone Surfaces Ting Cao, Haiying Tang, Xuemei Liang, Anfeng Wang, Gregory W. Auner, Steven O. Salley, and K.Y. Simon Ng Abstract Bacterial infection is a major challenge in biomaterials development. The adhesion of microorganisms to the material surface is the first step in infectious conditions and this quickly leads to the formation of biofilms on a material surface. A unique attribute of atomic force microscopy (AFM) is that it reveals not only the morphology of cells and the surface roughness of the substrate, but it can also quantify the adhesion force between bacteria and surfaces. We have shown that fluoroalkylsilane (FAS) and octadecyltrichlorosilane (OTS)-coated silicone samples exhibit greater potential for reducing E. coli JM 109 adhesion than heparin- and hyaluronan-modified samples. The force curves obtained from AFM can be used as a primary indicator in predicting bacterial adhesion. Key words: AFM, Bacterial adhesion, E. coli, SEM, Chemical vapor deposition

1. Introduction Bacterial infection is one of the major challenges in medical implants, such as heart valves, as it may result in severe complications (1, 2). The formation of a biofilm on a material surface is the primary factor in initiating infections, and it begins with the adhesion of microorganisms to the surface of the implant (3). Many approaches have been used to evaluate and quantify the extent of bacterial adhesion, such as bacterial counting, microscopy, and morphological observation (4–10). However, those techniques provide limited fundamental information toward the understanding of the mechanism of bacterial adhesion. A unique attribute of atomic force microscopy (AFM) is that it provides not only information on the morphology of cells (surface roughness),

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_22, © Springer Science+Business Media, LLC 2011

379

380

Cao et al.

but also quantifies the adhesion force between bacteria and ­modified surfaces (11–13). We have shown that FAS- and OTScoated silicone samples exhibit greater potential for reducing E. coli JM 109 adhesion than heparin- and hyaluronan-modified samples (14). The force curves obtained from AFM can be used as primary indicators in predicting bacterial adhesion. The adhesion force between E. coli and modified surfaces can be measured by an E. coli modified AFM tip. SEM allows the confirmation of successful modifications. These protocols can also be applied to other bacteria or cells with various surface compositions.

2. Materials 2.1. Bacteria Cultures

1. E. coli strain K12 JM 109. 2. Luria broth (LB) Miller. 3. Innova® 42 Incubator shaker.

2.2. Chemicals

1. Silastic silicone (Nonreinforced vulcanized gloss/gloss with 0.015 in. thickness). 2. Octadecyltrichlorosilane (OTS) 97.5%. 3. 1,3-Dimethylaminopropyl-3-ethylcarbodiimide ride 98% (water-soluble carbodiimide, WSC).

hydrochlo-

4. Heparin (sodium salt from porcine intestinal mucosa, ~170 USP units/mg). 5. Hyaluronic acid potassium salt (from human umbilical cord named Hyaluronan in this chapter). 6. 4-azidoanilinehydrochloride 97%. 7. Fluoroalkylsilane (FAS) 97%. 8. Ultrapure grade tris(hydroxymethyl)aminomethane (Tris). 9. Phosphate-buffered saline (PBS): 136 mM NaCl, 2.68 mM KCl, 10.1 mM Na2PO4, and 1.37 mM KH2PO4, pH = 7.5. 10. Glutaraldehyde. 11. Poly(ethyleneimine) (PEI, MW = 1,200).

3. Methods 3.1. Bacteria Culture 3.1.1. Bacterial Plating

1. Prepare the agar plate by melting Luria Broth Agar in boiled deionized water (DI water) following the ratios given in the descriptions accompanying the product.

Nanoscale Investigation on E. coli Adhesion to Modified Silicone Surfaces

381

2. Sterilize the agar solution with an autoclave at a temperature of 121°C for 60 min. Turn the steam control to slow exhaust until the pressure gauge reads zero and the temperature is below 80°C. 3. Take out the sterilized agar solution and place it in a laminar flow hood. Allow the agar solution to cool to about 45°C and then pour into a sterile plastic Petri dish. Rotate the dishes to evenly distribute the agar and wait until the agar solidifies. 3.1.2. Bacterial Colony

1. Hold the tube with E. coli strain in one hand and hold the needle or loop in the other hand. 2. Quickly pass the open tube mouth through the flame. Put a drop of strain on the edge of the solid agar plate surface. 3. Heat the needle in the flame until it becomes red. Allow it to cool for a couple of seconds before streaking the plate. 4. Streak the drop of the strain back and forth from edge to edge in parallel lines until halfway across the plate. Flame the needle, rotate the plate  90°, and streak the plate halfway across the plate again. 5. Repeat the above two procedures until the whole plate is streaked. Incubate the agar plate in an oven at 37°C until isolated individual colonies can be observed by the naked eye.

3.1.3. Bacterial Culture

1. Prepare the broth by melting Luria broth in boiled DI water using the ratios described in the directions accompanying the products. 2. Sterilize the broth according to the procedures described above. Cool the broth in a bio-hood. Flame the needle while holding the streak plate in the other hand. 3. Use a needle to select one obviously individual colony and transfer it into a tube with the sterile broth. 4. Incubate the tube at 37°C with a shaker speed of 150 rpm.

3.1.4. Bacterial Harvest

E. coli bacteria are harvested in the late exponential/early stationary phase.

3.2. Preparation of Silicone Surface for Modification

Put silicone samples in a 50-mL glass beaker. Clean samples with ethyl alcohol (100%) by ultrasonic cleaning for 5 min and remove from the beaker with tweezers. Then, blow-dry with compressed nitrogen.

3.2.1. Silicone Sample Surface Cleaning 3.2.2. Plasma Treatment on Silicone Surfaces

Place cleaned silicone samples in a Petri dish and treat with oxygen plasma for 5  min. The plasma-treated silicone samples are then placed into a sealed chamber. They are ready to be coated with OTS or FAS.

382

Cao et al.

Table 1 Characterization of surfaces via contact angle measurements with DI water and 1 mM Tris solution. (Reproduced from ref. 10 with permission from John Wiley & Sons, Inc.) Contact angle (°) ± Standard deviation Advancing (DI water)

Receding (DI water)

Equilibrium (DI water)

Equilibrium (1 mM Tris)

Silicone

119.1 ± 4.2

96.9 ± 5.1

112.6 ± 5

111.3 ± 1.6

OTS/silicone

120.7 ± 4.5

92.9 ± 2.8

110.3 ± 3.4

98.3 ± 1.0

FAS/silicone

119.8 ± 1.5

97.8 ± 3.4

116.2 ± 1.2

109.0 ± 1.7

Hep/OTS/silicone

95.9 ± 2.3

46.0 ± 1.0

  57.6 ± 2.1

51.3 ± 1.5

Hyaluronan/OTS/silicone

94.2 ± 2.4

50.6 ± 1.5

  59.1 ± 1.3

52.8 ± 1.6

N/A

0

Mica a

N/A

0

 Mica was cleaved just before use

a

3.2.3. Contact Angle Measurements

These procedures assume the use of a contact angle goniometer. Take contact angle measurements on three pieces of treated ­silicone samples and obtain advancing, receding, and equilibrium (under both DI water and 1 mM Tris solution) contact angles. Examples of contact angle values of plasma-treated silicone are listed in Table 1.

3.3. OTS and FAS Modification on Silicone Surfaces

1. Place plasma-treated silicone samples and a Teflon Petri dish into a desiccator. The chamber should be connected to a ­vacuum pump.

3.3.1. OTS Modification on Pretreated Silicone Surfaces

2. Fill the Petri dish with about 1 mL of OTS and seal the chamber (see Note 1). 3. Apply vacuum until the desiccator filled with OTS vapor is at 10−3  torr, then continue to evacuate for 4  h at room temperature. 4. After sealing the desiccator, turn off the vacuum pump and leave samples in a sealed dessicator for another 12 h with OTS (at about 10−2 torr).

3.3.2. FAS Modification on Pretreated Silicone Surfaces

1. Place the plasma-treated silicone samples into a desiccator and fill the Petri dish with about 0.1 mL of FAS. 2. Turn on the vacuum pump and keep the desiccator containing the FAS at 10−3 torr for 5 min at room temperature. 3. Keep the samples in the sealed desiccator for another 4 h with FAS at 0.3 torr.

Nanoscale Investigation on E. coli Adhesion to Modified Silicone Surfaces 3.3.3. Contact Angle Measurements

383

Take contact angle measurements on three pieces of OTS- or FAS-coated silicone samples and obtain advancing, receding, and equilibrium (under both DI water and 1 mM Tris solution) contact angles. Examples of contact angle values are listed in Table 1 (see Note 2).

3.4. Heparin- and Hyaluronan-Modified Silicone Surface

1. Prepare a 0.5% heparin solution by mixing heparin, WSC, and 4-azioanilinehydrochloride at a weight ratio of 2.35:1.29:1 in DI water.

3.4.1. Prepare Solution for Photo-immobilization

2. Adjust the pH of the solution to 4.70–4.75 by 2.3, and then 0.1 N NaOH solutions. Stir the solution for 24 h at 4°C. 3. Prepare a 0.2% aryl azido-modified hyaluronan solution by the same method but using modified amounts of hyaluronan, WSC, and 4-azioanilinehydrochloride at weight ratios of 2.16:1.64:1. 4. Keep the solution in the dark at 4°C before photo-­ immobilization (see Note 3).

3.4.2. Photoimmobilization of Heparin or Hyaluronan on OTS-Modified Silicone Surfaces

1. Set up the mercury vapor UV lamp (175 W, Regent Lighting, Burlington, NC) at a distance of 10  cm to a Teflon container. 2. Immerse OTS-coated silicone samples in a container filled with aryl azido-modified heparin or hyaluronan solution. Illuminate the samples with the UV lamp for 5 min (see Note 4). 3. Immerse heparin or hyaluronan immobilized silicon samples in DI water for 48 h. Refresh the water at least every 12 h and rinse the samples at the same time.

3.4.3. Contact Angle Measurements

Take contact angle measurements on three pieces of Heparin- or Hyaluronan-coated silicone samples and obtain the advancing, receding, and equilibrium (under both DI water and 1 mM Tris solution) contact angles. Examples of contact angle values are listed in Table 1.

3.5. AFM Tip Modification by E. coli

1. E. coli JM109 are harvested and washed in 10  mM PBS solution.

3.5.1. Pretreatment of E. coli

2. The bacteria are then treated with 2.5  vol% glutaraldehyde (adjusted to pH 7.5) for 2.5 h at 4°C with a final concentration of 0.6–0.8 DCW/mL. 3. The bacteria are copiously rinsed by PBS and resuspended in 1 mM Tris solution and pelletized by centrifuge at 4000 ´ g for 10 min.

3.5.2. Pretreatment of AFM Probes

4. Choose four silicon nitride integral tips (NP type) (Veeco Instrument, Santa Barbara, CA) which have a length of 115 mm and a width of 17 mm. Wash the tips with DI water and ethanol sequentially, and then illuminate them under UV light from

384

Cao et al.

a fiber optical illuminator (Dolan-Jenner Industries, Inc., Lawrence, MA) before use (see Note 5). 5. Attach the AFM probe on a flat Petri dish using double-sided adhesive tape that has been treated with oxygen plasma for 5 min. 6. Prepare 10  mL 1% PEI solution. Immerse the treated AFM probes in several drops of 1% PEI solution for 2 h (see Note 6). 7. After PEI treatment, carefully remove the AFM probe and rinse it with DI water and store at 4°C. 3.5.3. Immobilization of E. coli on AFM Probes

1. Place the AFM probes on a cleaned flat glass Petri dish and transfer a pellet of bacteria onto the tip of the PEI-coated probes using a pipette (see Note 7). 2. Next treat with a drop of 2.5% glutaraldehyde at 4°C to fix the bacteria onto the tip. 3. After about 1 h, rinse the probes with DI water to eliminate the weakly immobilized bacteria. Store the probes at 4°C.

3.5.4. SEM Images on AFM Tips Before and After E. coli Modification

1. This instruction assumes the use of a scanning electron microscope (S-2400, Hitach). To insure the successful immobilization of bacteria on top of the AFM tips, one extra AFM probe should be treated using the procedures described above. 2. Take SEM images of the tips with and without the bacteria immobilization by bonding the AFM probe to mounts using silver conductive paint (Hatfield, PA). 3. Sputter-coat both the silicon nitride and the ­bacteriaimmobilized tip with gold using an Ernest sputter coater (Latham, NY) at 50 mA for 25 s before taking the SEM images. Samples of SEM images are shown in Fig. 1 (see Note 8).

Fig. 1. SEM photograph of a standard AFM silicon nitride tip (a) before and (b) after being coated with E. coli JM 109 bacteria. (Reproduced from ref. 10 with permission from John Wiley & Sons, Inc.).

Nanoscale Investigation on E. coli Adhesion to Modified Silicone Surfaces

3.6. AFM Force Measurements 3.6.1. Load the Tip and Sample

385

1. These instructions assume the use of a Nanoscope IV-Dimension 3100 scanning probe microscope or Multimode (Veeco Instrument, Santa Barbara, CA). Remove modified silicone samples from the desiccators, and attach samples to a stainless steel sample disk using double-sided adhesive tape (see Note 9). 2. Carefully load the E. coli modified tip on the fluid tip holder. Install the tip holder and protective skirt following the instructions in the manual (see Note 10).

3.6.2. Calculation of the Spring Constant of the E. coli Modified Tip Probe

1. Use the technique described in the manual to align laser on the modified tip, adjust the photodetector, and focus on the sample surfaces (see Notes 11 and 12). 2. Manually tune the E. coli modified tip in air without applying external driving force and obtain the resonance curve of the same cantilever with bacteria. 3. The resonance frequencies of the cantilever before and after E. coli modification are used to calculate the spring constant of the cantilever by the Cleveland method (15) (see Note 13).

3.6.3. Obtain Force Curves in Contact Mode Under 1 mM Tris Solution

1. Carefully add a drop or more of 1 mN Tris solution onto the sample surface and lower the scanner head allowing the tip to enter the liquid. A fluid meniscus is formed between the fluid tip holder and sample surface (see Notes 14 and 15). 2. Readjust the laser alignment before taking any force measurements (see Note 16). 3. Engage the tip and switch to force calibration mode. Adjust the set point value, ramp size, and Z scan start to move the trace. Verify that the entire ramp shows on the graph (see Note 17).

3.6.4. Perform Three Independent Experiments and Collect the Force Curve Data

1. Move the sample stage to five to ten different locations to measure the interaction between bacteria and the modified silicone samples. 2. About 10 force measurements should be taken at each location. 3. Change another two modified AFM probes to collect the force curve on the same sample.

3.6.5. SEM Images on AFM Tips After Experiment

To ensure the presence of bacteria on the tips, use SEM to examine the tips after each experiment. A typical image is shown in Fig. 1.

3.7. AFM Data Analysis

1. Export the force curve data as ACSII to Excel. Both approach and retraction force measurements are recorded in one graph in terms of diode voltage (V) versus piezo position.

3.7.1. Determine Zero Separation Distance and Zero Force

2. To locate the position of zero force, identify the piezo position at a large separation where the tip deflection is constant. 3. Determine zero distance by identifying the region where the tip shows a linear deflection with the expansion of the piezo (see Note 18).

386

Cao et al.

Fig. 2. AFM retraction curves of E. coli JM 109-coated AFM tip on hydrophobic ­substrates. Measurements were performed in 1 mM Tris solution. FAS-coated silicone exhibited the smallest pull-off force and pull-off distance in hydrophobic surfaces. (Reproduced from ref. 10 with permission from John Wiley & Sons, Inc.).

4. Examples of “transformed” AFM approaching and retraction curves of E. coli JM109-coated AFM tip on hydrophobic ­substrates are shown in Fig. 2. 3.7.2. Calculate Pull-off Force

In retraction, the curves reflect the movement of the bacteriamodified tip pulling away from the substrates. This is the pull-off force.

3.7.3. Calculate Pull-off Distance

The point at which the pull-off occurs is the pull-off distance.

4. Notes 1. The formation of an OTS film is very sensitive to water. Even a trace amount of water can form multilayer or three-dimensional aggregates. Carefully transfer the OTS during this process. 2. The advancing and receding contact angle, often regarded as the most important surface characteristics, is influenced by the roughness, heterogeneity, and chemistry of surfaces. 3. The solution is strongly affected by light. It is easy to wrap the container with aluminum foil and keep it in ice water bath. All reactions should be executed in a dark room (16, 17). 4. The spring constant of a cantilever is important in force calculation. To determine the spring constant of cantilevers, the resonances curves of the cantilever used for attaching bacteria need to be recorded before immobilizing bacteria onto it. 5. The AFM probes become hydrophilic after plasma treatment. They can easily attach to and move with the solution. Adding an adequate amount of solution and keeping the AFM probe attached to the flat Petri dish help prevent probe damage.

Nanoscale Investigation on E. coli Adhesion to Modified Silicone Surfaces

387

6. The bacterial solution is viscous and difficult to transfer from centrifuge tubes because of the high bacterial concentrations. It is better to use a 100–1,000-mL pipette tip for transferring the bacteria than a small size pipette tip. Be careful not to touch the tip. The whole procedure needs to be done in a hood. 7. The bacteria are prone to solidify if exposed to ambient air. To keep it moist, cover the Petri dish with parafilm. 8. For the final 3 AFM tips, SEM images should be taken after force measurements. 9. The silicone is very soft. When mounting the silicone samples onto the sample disk, be careful to keep the sample surface flat and avoid small bumps. 10. When setting the tip holder on Dimension 3100, use the protective skirt to prevent the scanner from getting wet. The scanner head should not be immersed in more than 3 mm of fluid. 11. Usually the silicon nitride AFM probe has two tips on each side. Carefully check the laser when focusing on the tip that is used in the examination. 12. Distinguish the true signal from the signal reflected from fluid cell. 13. In this step, the resonance curves of the functionalized tip are recorded. The resonance frequencies obtained are used to calculate the spring constant of the cantilever. 14. Clean the fluid cell and protective skirt before adding fluid to avoid contamination. 15. The amount of liquid added to the surface depends on the ­sample hydrophobicity. Several drops of buffer are adequate for hydrophobic surfaces and more for hydrophilic surfaces. Slowly and carefully add liquid to avoid undesirable air bubbles. 16. The added fluid changes the laser reflection. 17. To achieve accurate force measurement, record the force curves until the curves presented on the graph become stable. Verify that the curves are in the whole graph. 18. Convert the tip deflection data to force using the model provided by Ducker et al., assuming that the cantilever is more complaint than the substrates (18).

Acknowledgments The author would like to thank Professors Guangzhao Mao and Gina Shreve, Chemical Engineering and Materials Science Department of Wayne State University, for their help in using the equipment. Financial support of the research by TACOM (contract no. DAAE07-03-C-L140) is gratefully acknowledged.

388

Cao et al.

References 1. Camesano T and Logan B. (1998) Influence of fluid velocity and cell concentration on the transport of motile and nonmotile bacteria in porous media, ENVIRONMENTAL SCIENCE & TECHNOLOGY;32(11):1699–1708. 2. Barnett B and Stephens D. (1997), Urinary Tract Infection: An Overview, AMERICAN JOURNAL OF THE MEDICAL SCIENCES;314(4):245–249. 3. Busscher HJ and Weerkamp AH. (1987), Specific and non-specific interactions in bacterial adhesion to solid substrata, FEMS Microbiology Letters;46(2):165–173. 4. An Y and Friedman R. (1997), Laboratory methods for studies of bacterial adhesion, JOURNAL OF MICROBIOLOGICAL METHODS;30(2):141–152. 5. Bowen W, Fenton A, Lovitt R, and Wright C. (2002), The measurement of Bacillus mycoides spore adhesion using atomic force microscopy, simple counting methods, and a spinning disk technique, BIOTECHNOLOGY AND BIOENGINEERING;79(2):170–179. 6. Kockro RA, Hampl JA, Jansen B, Peters G, Scheihing M, Giacomelli R, Kunze S, and Aschoff A. (2000), Use of scanning electron microscopy to investigate the prophylactic efficacy of rifampin-impregnated CSF shunt catheters, J Med Microbiol;49(5): 441–450. 7. Simhi E, van der Mei H, Ron E, Rosenberg E, and Busscher H. (2000), Effect of the adhesive antibiotic TA on adhesion and initial growth of E. coli on silicone rubber, FEMS MICROBIOLOGY LETTERS;192(1): 97–100. 8. Cao T, Wang A, Liang X, Tang H, Auner G, Salley S, and Ng K. (2008), Functionalization of AlN surface and effect of spacer density on Escherichia coli pili-antibody molecular recognition, COLLOIDS AND SURFACES B-BIOINTERFACES;63(2):176–182. 9. Cao T, Wang A, Liang X, Tang H, Auner G, Salley S, and Ng K. (2008), Patterned Immobilization of Antibodies in Mechanically Induced Cracks, JOURNAL OF PHYSICAL CHEMISTRY B;112(9):2727–2733.

10. Tang H, Cao T, Wang A, Liang X, Salley S, McAllister J, and Ng K. (2007), Effect of surface modification of silicone on Staphylococcus epidermidis adhesion and colonization, JOURNAL OF BIOMEDICAL MATERIALS RESEARCH PART A;80A(4):885–894. 11. Ong Y, Razatos A, Georgiou G, and Sharma M. (1999), Adhesion Forces between E. coli Bacteria and Biomaterial Surfaces, LANGMUIR;15(8):2719–2725. 12. Emerson R and Camesano T. (2004), Nanoscale Investigation of Pathogenic Microbial Adhesion to a Biomaterial, APPLIED AND ENVIRONMENTAL MICROBIOLOGY;70(10):6012–6022. 13. Cao T, Wang A, Liang X, Tang H, Auner G, Salley S, and Ng K. (2007), Investigation of spacer length effect on immobilized Escherichia coli pili-antibody molecular recognition by AFM, BIOTECHNOLOGY AND BIOEN­ GINEERING; 98(6):1109–1122. 14. Cao T, Tang H, Liang X, Wang A, Auner G, Salley S, and Ng K. (2006), Nanoscale ­investigation on adhesion of E. coli to surface modified silicone using atomic force ­microscopy, BIOTECHNOLOGY AND BIOENGINEERING;94(1):167–176. 15. Cleveland JP, Manne S, Bocek D, and Hansma PK. (1993), A nondestructive method for determining the spring constant of cantilevers for scanning force microscopy, Review of Scientific Instruments;64(2):403–405. 16. Wang A, Cao T, Tang H, Liang X, Salley S, and Ng K. (2005), In vitro haemocompatibility and stability of two types of heparin-immobilized silicon surfaces, COLLOIDS AND SURFACES B-BIOINTERFACES; 43(3-4): 245–255. 17. Wang A, Cao T, Tang H, Liang X, Black C, Salley S, McAllister J, Auner G, and Ng K. (2006), Immobilization of polysaccharides on a fluorinated silicon surface, COLLOIDS AND SURFACES B-BIOINTERFACES; 47(1):57–63. 18. Ducker WA, Senden TJ, and Pashley RM. (1992), Measurement of forces in liquids using a force microscope, Langmuir;8(7).

Part V Investigating Drug Action

wwwwwwwwwwwwww

Chapter 23 Imaging Bacterial Shape, Surface, and Appendages Before and After Treatment with Antibiotics Pier Carlo Braga and Davide Ricci Abstract Antibiotics are particular type of drugs that are able to interfere in different ways to the metabolic ­pathways of bacteria. This causes also morphostructural alterations of cell wall and surface appendages (flagella, fimbriae or pili, and filaments). Atomic force microscopy (AFM) is extremely useful for analyzing the three-dimensional structure of the surface of biological specimens, particularly bacteria. A step-by-step AFM methodology to be applied to different type of bacteria is reported and visual examples of the action of antibiotics are shown. Although scanning electron microscopy is still frequently used, the introduction of the AFM technique offers substantial benefits in real quantitative data acquisition in three dimensions, minimal sample preparation times, flexibility in ambient operating conditions (i.e., no vacuum is necessary), and effective three-dimensional magnification at submicron level. Key words: AFM, E. coli, H. pylori, B. cereus, S. pyogenes, Rokitamycin, Cefodizime, Daptomycin

1. Introduction Bacteria are typically smaller than eukaryotic cells. The average diameter of Staphylococcus aureus is 1 ± 0.5 mm while Escherichia coli is on average 0.5 × 25 mm. The bacterial cell is also characterized by the presence of a complex external rigid structure called cell wall, which protects the internal protoplast and gives also the cellular shape, that generally falls into one of the three basic morphologic categories: spherical (cocci), rod-shaped (bacilli), and spiral. Some bacteria show atypical bacterial shape. Bacteria are also able to extrude some material that collects outside the cell wall to form an additional surface layer. Many genera of bacteria possess also filamentous structures projecting through the cell wall to form the so-called surface appendages. Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_23, © Springer Science+Business Media, LLC 2011

391

392

Braga and Ricci

The most commonly observed bacterial appendages are flagella, fimbriae or pili, and filaments. Antibiotics are a particular type of drugs that are able to interfere in different ways to the metabolic pathways of bacteria. This also causes, directly or indirectly, changes in the structure of cell wall and consequent alterations in the shape of bacteria. The integrity of cell wall and bacterial shape are important to maintain the vitality and the virulence of bacteria. Morphostructural alterations not only cause bacteria to lose cytoplasm, but also to be more easily phagocytized and killed by human phagocytic cells. A  large amount of basic and clinical researches in microbiology, chemotherapy, infectivology, etc., have been performed in order to investigate the morphology and structure of bacteria. These studies have been previously done by means of optical microscopy and scanning electron microscopy (SEM). Optical and scanning (or transmission) electron microscopes are classified as “far-field microscopes” because the distance between the sample and the point at which the image is obtained is long in comparison with the wavelengths of the photons or electrons involved. In this case, the image is a diffraction pattern and its resolution is wavelength limited (1, 2): in optical microscopy, resolution is determined by the Nyquist relation to the wavelength of the light used (typically about 1 gm); in a general purpose SEM, it is limited by the properties of the electromagnetic lenses (typically about 50 Å) (3). In 1986, a completely new type of microscopy was proposed: without lenses, photons, or elections, it involves the mechanical scanning of samples (4) and opened up unexpected possibilities for the surface analysis of biological specimens. Initially called the scanning force microscope (SFM), it was a development of the previous scanning tunneling microscope (STM) (5) which ­provided information about atomic resolution of specimens that are electrically conducing. Because SFMs involve interactions between atomic forces (about 10−9 Newton), they are also and more frequently called atomic force microscopy (AFM) (3). These new types of scanning probe microscopes (SPMs) are based on the concept of “near field microscopy,” which overcomes the problem of the limited diffraction-related resolution inherent in conventional microscopes. Located in the immediate vicinity of the sample itself (usually within a few nanometers), the probe records the “intensity” and not the “interference signal,” and this greatly improves resolution (1). As shown in Fig. 1, AFM explores the surface of a sample not by means of a system of lenses that form an image using the diffraction patterns of rays of different wavelengths, but by means of a very small sharp-tipped probe located at the free end of a cantilever driven by the interatomic repulsive or attractive forces (van der Waals forces) between the molecules at the probe tip and

Imaging Bacterial Shape, Surface, and Appendages Before and After Treatment

393

Fig.  1. Comparative schematic view of the elements characterizing light microscopy (LM), scanning electron microscopy (SEM), and atomic force microscopy (AFM), together with their specific technical parameters.

those on the surface of the specimen. This can be done by scanning the sample laterally (xy) while a closed-loop control system keeps the tip in proximity to the surface by adjusting the z  position of the sample. In most AFMs, tip movements are ­monitored by reflecting a laser beam from the back of the cantilever on to a position-sensitive photo-diode (6). AFM is extremely useful for analyzing the three-dimensional structure of tire surface of biological specimens, particularly bacteria. Although SEM is still frequently used, the introduction of the AFM technique offers substantial benefits: real quantitative

394

Braga and Ricci

data acquisition in three dimensions, minimal sample preparation times, flexibility in ambient operating conditions (i.e., no vacuum is necessary), and effective three-dimensional magnifications at submicron level (7, 8). To investigate the shape and the surface of bacteria offers the possibility of investigating the efficacy and the mechanism of action of antibiotics that disrupt this structure as an epiphenomenon of internal biochemical action (9–13), and at the same time the possibility of investigating their lack of activity, as in the case of resistance.

2. Materials 1. Test organisms. Both Gram-positive and Gram-negative bacteria are suitable for AFM. 2. Triptic soy broth or other suitable medium. 3. Phosphate-buffered saline (PBS) (0.02  M phosphate and 0.15 M NaCl, pH 7.3). 4. Glutaraldehyde: 2.5%.in 0.1 M cacodylate buffer, pH 7. 5. Graded alcohols (60%, 70%, 80%, 90%, 100%). 6. Incubator. 7. Centrifuge. 8. Micropipette and sterilized disposables for culturing bacteria. 9. Round glass coverslides, diameter 6–7 mm (or mica). 10. AFM (including probe tips, software for processing signals and three-dimensional rendering, and a computer).

3. Methods 1. The cultures of chosen microorganism is prepared according to common standard procedures 2. Wash the test microorganism from the suspension in broth (i.e., 106 cells/ml) three times with PBS 3. Resuspend the final pellet in 1–2 ml of PBS 4. Collect 0.1 ml (or less) of this suspended bacteria wing a micropipette and place it on round glass coverslide (see Note 1) 5. Dry the coverslip in air 6. Fix with 2.5% glutaraldehyde in 0.1  M cacodylate buffer (pH = 7.1) 7. Dehydrate in graded alcohols

Imaging Bacterial Shape, Surface, and Appendages Before and After Treatment

395

8. Dry the coverslip in air (see Note 2) 9. Repeat steps 1–8 incubating bacteria with various supra-MICs or sub-MICs of antibiotic. 10. AFM observation (see Notes 3 and 4). A typical AFM imaging session begins by firmly fixing the sample cover slide to the microscope holder in order to avoid even the slightest movement (see Note 5), and then positioning it under the probe tip and locating the area of interest by moving the x–y table. 11. A good quality on-axis optical microscope is essential in order to be able to position the probe tip in the proximity of a bacterium to be imaged by AFM. As bacteria are about 1 mm in size, it is necessary to have appropriate lighting conditions to distinguish them from any debris on the slide surface. In the experiments described here, a reflection optical microscope equipped with long range objectives was used. Although the cantilever bearing the probe partially obstructs the optical view of the underlying bacteria, it does allow the probe to be positioned sufficiently accurately in the area of interest (see Notes 6 and 7). Commercial AFM instrumentation coupled to a transmitted light optical microscope offers a higher degree of precision in the first approach of the probe to the sample (see Note 8). 12. Once an area has been located after the tip-to-sample approach, a first large scan (i.e., 30 by 30 mm) using a high scan speed and small number of pixels per line can be made in order to assess its exact position within the scanner coordinate system, identify the nature of the bacteria, and select an interesting one. Further smaller scans may be necessary in order to position the bacterium exactly at the center of the scanning area. 13. Record high-resolution images (see Note 6) using appropriate instrument settings depending on the imaging mode selected (contact, intermittent contact, and noncontact) (see Note 7). In general, accurate feedback setting is necessary in order to obtain the maximum possible gain for the resolution of bacterial surface structures while avoiding oscillation when scanning along the cell sidewalls (see Note 2). 14. Acquire image (typically acquired at 512 × 512  pixels) and process by means of plane fitting, high-frequency filtering, and three-dimensional shaded rendering (Fig. 2). 15. Post-processing analysis and the spatial representation of AFMgenerated data is essential in order to extract all of the available information from the image dataset. As the recorded data is an intrinsically three-dimensional digital matrix (the height of the sample recorded at each x, y coordinate), the software makes it easy to obtain numerical data of cross-sections of interesting features expressed with subnanometer accuracy (see Note 8). The same software allows three-dimensional rendering of the

Fig.  2. Atomic force pictures of various bacteria. (a) Common morphology of E. coli without exposure to antibiotic. (b, c) Different alterations induced in E. coli by exposure to cefodizime. (d) Example of common morphology of S. pyogenes phenotype M without exposure to antibiotic. (e, f) After incubation with rokitamycin. (g) Example of untreated common rod-shaped morphology with flagella of B. cereus. (h, i) Incubation

Imaging Bacterial Shape, Surface, and Appendages Before and After Treatment

397

surface and rotation in space so that only one acquisition is needed to be able to observe the same object from many different points of view (see Notes 9–11).

4. Notes 1. It is better to use low concentrations of bacteria because they tend to concentrate in small areas during the air-drying phase, whereas a single bacterium provides a clearer image. Be sure to mark the location of your specimen on the upper surface of your round glass coverslide in order to avoid wasting time investigating the wrong side. 2. If the sample is kept dry, repeated sessions can generally be performed without any loss of resolution. 3. Bacterial sample preparation for AFM is very simple and rapid. There is no need for critical point drying, which also avoids shrinkage effects; there is no need for gold sputtering, a procedure that covers and smooths fine surface details. There is no need for vacuum conditions as with SEM. 4. A recent technical evolution has also opened up the possibility of using AFM on wet samples, i.e., living cells immersed in biological fluids in culture chambers (14, 15). 5. Care must be taken when choosing the adhesive used for ­fixing the glass slide to the sample holder. Avoid using thick double-sided adhesive tape, as this can expand for a long time after pressure and thus cause instability in the vertical position of the tip. The specially produced sticky tabs made by different manufacturers are fine. 6. To obtain the best results, it is necessary to be thoroughly familiar with the characteristics of different cantilevers and tips, how these can be used and how they suit to the different kinds of samples investigated. For high-resolution work, tip sharpness is essential: tip properties can vary significantly within the same batch of cantilevers. Fine tuning of the feedback loop and set point, together with the chosen scan speed, is critical for good surface tracking.

Fig.  2. (continued) of morphostructural alterations with daptomycin. (l) Common ­morphology of S. aureus without exposure to antibiotic. (m, n) Different alterations induced in S. aureus by exposure to moxifloxacin. (o) Common morphology of M.  catarrhalis without exposure to antibiotic. (p, q) Different alterations induced by exposure to moxifloxacin. (r) Common morphology of H. pylori without exposure to antibiotic. (s, t) Different alterations induced in H. pylori by exposure to rokitamycin. The bar at the bottom right corresponds to 500 nanometers.

398

Braga and Ricci

7. As mentioned above, AFM offers different imaging modes for investigating the sample. There is the “contact mode” in which the tip of the probe makes soft “physical contact” with the sample, which should be used with harder and stiffer materials than biological samples, as it can easily give rise to undesirable effects due to tip-to-sample interactions. Tip pressure can indent and deform the sample surface, and lateral forces can stretch the sample, drag away loosely bound fragments, or even detach the whole bacterium from the substrate (3). These drawbacks of the contact AFM mode are overcome by using the “intermittent–contact mode” also called “tapping mode.” In this case, the AFM feedback loop constantly dampens the high-frequency oscillations of the vibrating cantilever due to the tip coming into contact with the surface for a very short time (16). For this reason, indentation effects are less invasive, lateral forces are greatly reduced, and a high lateral resolution can be maintained. In the third “noncontact mode,” small amplitude and high-frequency oscillations induced on the cantilever allow the feedback control loop to maintain the tipto-sample distance within the range of attractive Van der Waals forces. Tip-to-sample interactions are greatly reduced at the expense of lateral resolution and the scanning speed (2). For biological specimens, the noncontact and intermittent–­contact are the most suitable, although the contact mode may be used for high-resolution work on very small areas. 8. In order to make accurate dimensional measurements, the calibration of the AFM’s piezoelectric scanner has to be periodically checked. The procedures are usually described in the instrument manual. Lateral dimension calibration is relatively straightforward, but special care must be taken when calibrating height. We used a VLSI standard calibration grid (NIST traceable) with a 100 nanometer nominal step height and an ­in-house developed statistical analysis procedure for calibration. 9. The images may sometimes be blurred as a result of poor washing procedures, an electrostatic charge on the specimen, improper feedback parameter settings, debris on the tip, or an eroded tip. 10. The images of spherical bacteria, such as S. aureus, suffer from little lateral resolution along the perimeter, due to the perpendicular direction of analysis. In general, the shape of the tip and its lateral walls limit the detection of steep elevated features (Fig. 3). 11. After image acquisition, the built-in software allows the rendering of the picture to be greatly improved by means of shadowing, rotation, different illumination, and different points of view.

Imaging Bacterial Shape, Surface, and Appendages Before and After Treatment

399

Fig. 3. Example of the AFM rendering of a spherical bacterium.

Acknowledgments We would like to thank M. Dal Sasso for preparing the bacterial samples. This study was partially supported by a grant from MIUR (60%). References 1. Heckl, W.M. (1995) Scanning the thread of life, in The human genome (Fisher, E.P., Klose, S.), R. Piper GmbH & Co. KG, Munchen, pp. 99–146. 2. Braga, P.C. and Ricci, D.(1998) Atomic Force Microscopy: Application to investigation of Escherichia coli morphology before and after exposure to cefodizime. Antimicrob. Agents Chemother. 42, 18–22. 3. Strausser, Y.E. and Heaton, M.G. (1994) Scanning probe microscopy technology and recent innovations. American Laboratory, May 1–7. 4. Binning, G., Quate, C.F. and Gerber, C. (1986) Atomic force microscope. Whys. Rev. Lett. 12, 930–933. 5. Binnig, G.and Rohrer, H. (1982) Scanning tunnelling microscopy. Helv. Phys. Acta. 55, 726–735. 6. Mc Donnel, L. and Phelan, M. (1998) The scanned cantilever AFM: a versatile tool for industrial application. Microscopy and Analysis (European ed.) 52, 25–27. 7. Ratneshwar, L. and Scott, A.J. (1994) Biological applications of atomic force microscopy. Am. J. Physiol. 266, C1–C21. 8. Campbell, P.A., Gordon, R. and Walmsley, D.G. (1998) Active surface modification by scanning tunnelling microscopy. Microscopy and Analysis (European ed.) 56, 25–27. 9. Lorian, V. (1986) Effect of low antibiotic concentrations on bacteria: effects on ultrastructure, their virulence and susceptibility to

10.

11.

12.

13.

14. 15.

16.

immunodefenses, in Antibiotics in Laboratory Medicine (Lorian, V.), The Williams & Wilkins Co., Baltimore, pp. 596668. Lorian, V., Atkinson, B., Walushacka, A. and Kim, Y. (1982) Ultrastructure, in vitro and in vivo, of staphylococci exposed to antibiotics. Curr. Microbiol., 7, 301–304. Braga, P.C. and Ricci, D. (2000) Atomic Force Microscopy: detection of rokitamycin induced morphological alterations in Helicobacter pylori. Chemotherapy, 46, 15–22. Braga, P.C. and Ricci, D. (2002) Differences in the susceptibility of Streptococcus pyogenes to rokitamycin and erythromycin revealed by morphostructural atomic force microscopy investigation. J. Antimicrob. Chemother. 50, 475–460. Braga, P.C., Ricci, D., Dal Sasso, M. and Thorne, G. (2002) Bacillus cereus morphostructural damage by daptomycin: atomic force microscopy investigation. J. Chemother. 14, 336–341. Nagao, E. and Dvorak, J.A. (1999) Developing the atomic force microscope for studies of living cells. Intern. Lab., January, 21–23. Schaus, S.S. and Henderson, E.R. (1997) Cell viability and probe-cell membrane interactions of XR1 glial cells imaged by atomic force microscopy. Biophys. J., 73, 1205–1214. Howland, R. and Benatar, L. (1997) A practical guide to scanning probe microscopy. Park Scientific Instrument Ed., pp. 1–73.

wwwwwwwwwwwwww

Chapter 24 Thymol-Induced Alterations in Candida albicans Imaged by Atomic Force Microscopy Pier Carlo Braga and Davide Ricci Abstract Thymol, a constituent of thyme essential oil that has been credited with interesting antimicrobial and antifungal effects, acts by interfering with the envelope of Candida albicans and this activity has been investigated by means of atomic force microscopy (AFM). Candida culture samples incubated with 1, 1/2, and 1/4 MIC of thymol or vehicle were taken at time 0 and after 1, 2, and 4 h, the envelopes of 100 cells in each of five randomly chosen fields were analysed by means of AFM. Our AFM findings show that thymol affects the envelope of C. albicans cells. The cells showed major morphostructural deformities with envelope damage becoming greater at increasing thymol concentrations and longer times of incubation, including the number of flattened cells with surface folds, cells with holes, and collapsed cells and ghosts. Thymol is an amphipathic monoterpene, which suggests that it affects cell membrane structure by generating asymmetries and membrane tensions. This is confirmed by the fact that terpenes alter cell permeability by entering between the fatty acyl chains making up the membrane lipid bilayers, disrupting lipid packing, and changing membrane fluidity. All of these phenomena lead to major surface alterations and deformities that also reduce the ability of fungi to adhere to mucosal cells, and decrease their virulence and infectiousness. Key words: Thymol, Candida albicans, AFM, Envelope damage

1. Introduction Candida albicans is a pleomorphic commensal or opportunistic pathogenic yeast that can cause a number of infections ranging from surface diseases of the skin and mucosae to deep tissue ­infections. It has both a budding and a filamentous life cycle: blastoconidia are unicellular forms of the fungus and may be seen as buddings, whereas the filamentous forms are hyphae and pseudohyphae (1, 2). It has been postulated that more than 90% of healthy ­individuals carry a single strain of Candida at different body sites for a long Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_24, © Springer Science+Business Media, LLC 2011

401

402

Braga and Ricci

time (3, 4). The human infections caused by C. albicans and several related species range from relatively trivial conditions, such as oral and genital thrush, to severe systemic infections (4). A number of virulence factors promote C. albicans colonisation or invasion of host tissues. The cell wall, which gives shape and form, is one of the most widely investigated because it is related to growth, protection against mechanical stress, and osmotic lysis, and provides passive protection against potentially harmful compounds. It is also where cell surface ligands bind the receptors that promote adhesion and the subsequent colonisation of host cells and tissues, whereas proteolytic enzymes are involved in tissue penetration (4, 5). The essential oils of many aromatic plants have antifungal effects (6, 7). Such oils are very complex mixtures of various components with different chemotypes (terpenes, aldehydes, alcohols, acyclic esters, etc.), and a number of studies have found that thyme oils are highly antimicrobial (8, 9) mainly because of their high phenol content (9, 10). The specific anti-Candida activity of essential oils is well established, as is that of thymol, carvacrol, and eugenol, the major phenolic components of thyme, oregano, and clove oils. Thymol has been credited with a series of pharmacological properties, including antimicrobial and antifungal effects (9–11). As it has been observed that one important factor underlying the virulence of fungi (and bacteria) is their ability to maintain the functional architecture of their envelopes, this study shows the interference induced by thymol on the envelope of C. albicans. Studies of this kind have so far been performed using optical and scanning electron microscopes, but a new family of useful instruments has now been introduced: atomic force microscopes (AFM). In this study, we used an AFM with a resolution of more than 0.1 nm. Rather than exploring the surface of a sample through a system of lenses that form an image by means of the diffraction patterns of rays of different wavelengths, AFMs have a very sharp-tipped probe located at the free end of a cantiliver driven by the interatomic repulsive or attractive (van der Walls) forces between the molecules at the tip and those on the surface of the specimen (12). AFMs are extremely useful for analysing the three-dimensional structure of the surface of biological specimens, such as bacteria and fungi, because of their real quantitative data acquisition in three dimensions, minimal sample preparation times, flexibility in ambient operating conditions (i.e. no vacuum is necessary), and effective three-dimensional magnification at submicron level. Investigating the shape and the surface of fungi (and bacteria) makes it possible to verify not only the efficacy and mechanisms of action of the antifungal drugs that disrupt it as an epiphenomenon of their surface or internal biochemical action, but also their lack of activity as in the case of resistance (12).

Thymol-Induced Alterations in Candida albicans Imaged by Atomic Force Microscopy

403

2. Materials 1. Test organisms C. albicans strains (or related species) from the ATCC or clinically isolated. 2. Sabouraud dextrose agar medium for isolating Candida. 3. Sabouraud broth or another suitable medium for Candida broth culture. 4. Phosphate buffered saline (PBS): 0.02  M phosphate and 0.15 M NaCl, pH 7.2. 5. Glutaraldehyde: 2.5% in 0.1 M cacodilate buffer, pH 7.0. 6. Incubator. 7. Centrifuge. 8. Optical microscope. 9. All sterilised disposables for culturing fungi (including sixwell plates). 10. Round glass cover slide, diameter 6 mm. 11. AFM (including probe-tips, software for signal processing and three-dimensional rendering and a computer). 12. Thymol.

3. Methods 1. A stock suspension of yeast cells is prepared by inoculating three or four colonies from agar plates in 6 ml of 2% glucose Sabouraud broth, and incubated at 32°C for 24 h. This procedure gives a suspension of budding blastoconidia. 2. The MIC of thymol for the Candida strain is determined using the broth macrodilution method. 3. The C. albicans culture is incubated in a six-well plate without or with 1× MIC or sub-MICs of thymol. One round glass cover slide is inserted in each well (see Note 1). 4. After 4 h of incubation, the medium is carefully withdrawn and each round cover slide was carefully removed from each well and gently washed (submerged) in PBS. 5. Each cover slide with attached Candida cells is fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, pH 7.0 at 4°C for at least 4 h. 6. After washing with distilled water, the cover slide is simply dried in air. None of the samples underwent critical point or gold sputtering procedures (see Note 2).

404

Braga and Ricci

7. A typical AFM imaging session begins by firmly fixing the cover slide to the microscope holder to avoid even the slightest movement, and positioning it under the probe tip, and locating the area of interest by moving the x–y table (see Note 3). 8. A good quality on-axis optical microscope is essential in order to be able to position the probe tip in the proximity of a fungal cell to be imaged by AFM. In the experiments described here, a reflection optical microscope equipped with long range objectives was used. Although the cantilever bearing the probe partially obstructs the optical view of the underlying fungi, it does allow the probe to be positioned sufficiently accurately in the area of interest (see Note 4). Commercial AFM instrumentation coupled to a transmitted light optical microscope offers a higher degree of precision in the first approach of the probe to the sample (see Note 5). 9. Once an area has been located after the tip-to-sample approach, a first large scan (i.e. 30 by 30 mm) using a high scan speed and small number of pixels per line can be made in order to assess its exact position within the scanner coordinate system, identify the nature of the fungal cell, and select one that is interesting. Further smaller scans may be necessary in order to position the cell exactly at the centre of the scanning area. 10. Record high-resolution images (see Note 6) using appropriate instrument settings depending on the imaging mode selected (contact, intermittent contact, and non-contact) (see Note 7). In general, an accurate feedback setting is necessary in order to obtain the maximum possible gain for the resolution of cell surface structures while avoiding oscillation when scanning along the side walls of the cell (see Note 8). 11. A PSIA XE-100 (PSIA Corp., Korea) advanced scanning probe microscopy system is used to record all of the AFM images. 12. The size and shapes of the C. albicans cells (which would normally have been too difficult to scan via conventional AFMs) are tracked by using the greater vertical range of the independent Z-piezo scanner of the PSIA XE-100 and its very fast time response (13). The AFM is mounted on a Halcyonics M1-plus active vibration isolation table (Halcyonics GmbH, Germany). The x, y, and z scanner calibration was checked and found to be accurate to within 1% of nominal values using a reference pattern from VLSI Standards (USA) that has a pitch of 3 mm and a step height of 18 nm. 13. The PSIA NCHR cantilevers for high-resolution non-contact mode were chosen because of their high aspect ratio and sharpness, which makes them especially suitable for imaging globular objects and deep pits. 14. We used the high-magnification (1,000×) on-axis reflection optical video microscope that is part of the PSIA XE-100 SPM

Thymol-Induced Alterations in Candida albicans Imaged by Atomic Force Microscopy

405

to locate and distinguish normal and deformed cells, and could therefore easily position the AFM tip on specific areas of interest. AFM imaging was performed using true non-contact mode under ambient conditions, thus allowing tip-to-sample interactions to be reduced to a minimum (14) (Figs.  1–4). Special care was taken to tune set point and feedback gain in conjunction with the scanning rate in order to allow the system to ­follow the relatively large cell structures whilst maintaining high resolution on local details. 15. We typically acquired 512 × 512 or 1,024 × 1,024 pixel images at a scanning rate of 0.5 Hz on areas ranging from 5 × 5 to 15 × 15 mm, using a high feedback gain and minimum interaction set point.

Fig. 1. AFM images of C. albicans cells. (a) After 4 h, the cells have smooth and round surfaces. Examples of various morphological alterations in C. albicans cells after incubation with thymol, (b–d) flattened and fold cells, (e–h) cells with holes, and (k) collapsed cells and ghosts (AFM, bar = 1 mm).

406

Braga and Ricci

Fig. 2. Overall view of randomly chosen fields of C. albicans under normal conditions (a), and after incubation with thymol 1 MIC for 4 h (b). The cells marked with an asterisk have normal envelopes while the others have envelopes with thymol-induced alterations (AFM, bar = 1 mm).

16. In order to obtain the best results in three-dimensional ­renderings, the raw AFM data is pretreated using the proprietary software (XEI by PSIA Corp. Korea) by means of x and y background global plane subtraction followed by a line-byline x-direction offset subtraction while excluding cells from computation by means of histogram selection. The same ­software was used to obtain cross sections along selected

Thymol-Induced Alterations in Candida albicans Imaged by Atomic Force Microscopy

407

Fig. 3. Different points of view obtained by rotating the sample after acquiring only one digital image (bar = 1 mm).

Fig.  4. Example of height (nm) and size measurement of a selected cross section of ­damaged Candida cells.

directions of cell morphology. The data were then rendered in three dimensions using WSxM freeware (Nanotec Electronica S.L., Spain) set for textured shaded solids with ambient and diffuse lighting and specular reflection effects. The software allowed snapshots of the acquired data to be obtained at any viewing angle and with any direction of illumination.

408

Braga and Ricci

4. Notes 1. It is better to use low concentrations of Candida (5 × 105–1 × 106 cells) because they tend to aggregate in small areas during the air-drying phase, whereas a single cell provides a clearer image. Be sure to mark the location of the specimen on the upper surface of the round glass cover slide in order to avoid wasting time investigating the wrong side. 2. Candida sample preparation for AFM is very simple and rapid. There is no need for critical point drying, which also avoids shrinkage effects; there is no need for gold sputtering, a procedure that covers and smooths fine surface details. There is also no need for vacuum conditions as with SEM. If the sample is kept dry, repeated sessions can generally be performed without any loss of resolution. 3. Care must be taken when choosing the adhesive used to fixing the glass slide to the sample holder. Avoid using thick double-sided adhesive tape, as this can expand for a long time after pressure and thus cause instability in the vertical position of the tip. The specially produced sticky tabs made by different manufacturers are fine. 4. To obtain the best results, it is necessary to be thoroughly familiar with the characteristics of the different cantilevers and tips, how these can be used, and how they suit different kinds of samples. For high-resolution work, tip sharpness is essential: tip properties can vary significantly within the same batch of cantilevers. Fine tuning of the feedback loop and set point, together with the chosen scan speed, is critical for good surface tracking. 5. The images may sometimes be blurred as a result of poor washing procedures, an electrostatic charge on the specimen, improper feedback parameter settings, debris on the tip, or an eroded tip. 6. As mentioned above, AFM offers different imaging modes for investigating the sample. There is the “contact mode” in which the tip of the probe makes soft “physical contact” with the sample, which should be used with harder and stiffer materials than biological samples as it can easily give rise to undesirable effects due to tip-to-sample interactions. Tip pressure can indent and deform the sample surface, and lateral forces can stretch the sample, drag away loosely bound fragments, or even detach the whole cell from the substrate. These drawbacks of the contact AFM mode are overcome by using the “intermittent–contact mode”, also called the “tapping mode”. In this case, the AFM feedback loop constantly

Thymol-Induced Alterations in Candida albicans Imaged by Atomic Force Microscopy

409

dampens the high frequency oscillations of the vibrating cantilever due to the tip coming into contact with the surface for a very short time (15). For this reason, indentation effects are less invasive, lateral forces are greatly reduced, and a high lateral resolution can be maintained. In the third “noncontact mode”, the small amplitude and high frequency oscillations induced on the cantilever allow the feedback control loop to maintain tip-to-sample distance within the range of attractive Van der Waals forces. Tip-to-sample interactions are greatly reduced at the expense of lateral resolution and the scanning speed. For biological specimens, the non-contact and intermittent–contact mode are the most suitable, although the contact mode may be used for high-resolution work on very small areas. 7. In order to make accurate dimensional measurements, the calibration of the AFM’s piezoelectric scanner has to be periodically checked. The procedures are usually described in the instrument manual. Lateral dimension calibration is relatively straightforward, but special care must be taken when calibrating height. We use a VLSI standard calibration grid (NIST traceable) with a 100-nm nominal step height and an inhouse statistical analysis procedure for calibration. 8. The images of spherical cells, such as C. albicans blastoconidia, suffer from little lateral resolution along the perimeter, because of the perpendicular direction of analysis. In general, the shape of the tip and its lateral walls limits the detection of steeply elevated features. 9. After image acquisition, the built-in software allows the rendering of the picture to be greatly improved by means of shadowing, rotation, different illumination, and different points of view. References 1. Odds, F.C. (1979) Morphogenesis in Candida with special reference to C. albicans. In “Candida and Candidosis”. Leicester University Press, Leicester; p. 31. 2. Sonnex, C. and Lefort, W. (1999) Microscopic features of vaginal candidiasis and their relation to symptomatology. Sex. Transm. Inf. 75, 417–419. 3. Odds, F.C. (1987) Candida infections: an overview. Crit. Rev. Microbiol. 15, 1–5. 4. McCullough, M.J., Ross, B.C. and Reade P.C. (1996) Candida albicans: a review of its taxonomy, epidemiology, virulence attributes, and methods of strain differentiation. Int. J. Oral. Maxillofac. Surg. 25, 136–144.

5. Calderone, R. and Brawn, P.C. (1991) Adherence and receptor relationships of Candida albicans. Microbiol. Rev. 55, 1–20. 6. Knoblock, K., Pauli, A., Iberl, N., Wqeigand, N. and Weiss, H.M. (1989). Antibacterial and antifungal properties of essential oil components. J. Essent. Oil Res. 1, 119–128. 7. Salgueiro, L.R., Cavaleiro, C., Pinto, E. et  al. (2003) Chemical composition and antifungal activity of the essential oil of Origarum virens on Candida species. Planta Medica 69, 871–874. 8. Rasooli, I. and Mirmostafa, S.A. (2002) Antibacterial properties of Thymus pubescens and Thymus serpyllum essential oils. Fitoterapia 73, 244–250.

410

Braga and Ricci

9. Cosentino, S., Tuberoso, C.I., Pisano, B., Satta, M., Moscia, V., Arredi, E,. and Palmas, F. (1999) In vitro antimicrobial activity and chemical composition of Sardinian Thymus essential oils. Lett. Appl. Microbiol. 29, 130–135. 10. Juliano, C., Mattana, A. and Usai, M.(2000) Composition and in vitro antimicrobial activity of essential oil Thymus herba-barona Loisel growing wild in Sardinia. J. Essential Oil Res. 12, 516–522. 11. Pauli, A. (2006) Anticandidal low molecular compounds from higher plants with special reference to compound from essential oils. Med. Res. Rev. 2, 223–268. 12. Braga, P.C. and Ricci, D. (2004) Imaging bacterial shape, surface, and appendages

before and after treatment with antibiotics. In Atomic Force Microscopy: Biomedical Methods and Applications (Braga, P.C. and Ricci, D. eds), Humana Press Inc, Totowa, NJ, pp. 179–188. 13. Kwon, J., Hong, J., Kim, Y.S., Lee, D.Y., Lee, K., Lee, S.M. and Park, S. (2003) Atomic force microscope with improved scan accuracy, scan speed, and optical vision. Rev. Sci. Instrum. 74, 4378–4383. 14. Martin, Y., Williams C.C. and Wickramasinghe, H.K. (1987) Atomic force microscope-force mapping and profiling on a sub 100 Å scale. J. Appl. Physics 61, 4723–4729. 15. Howland, R. and Benatar, L. (1997) A practical guide to scanning probe microscopy. Park Scientific Instrument, Sunnyvale, CA, pp. 1–73.

Chapter 25 Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix Interactions in Vascular Smooth Muscle and Endothelial Cells Zhe Sun and Gerald A. Meininger Abstract The interactions between cell surface integrins and extracellular matrix (ECM) play important roles in the function of vascular smooth muscle and endothelial cells. Atomic force microscopy (AFM) has emerged as a powerful tool to mechanically engage cell surface integrins through functionalized probes, and to apply mechanical forces directly to cells or to specific protein–protein receptor ligand interactions, such as integrin–ECM interactions. In the example of integrins, this approach allows more accurate evaluation of the regulation of integrin adhesive activities, and provides a unique approach to access and investigate integrin-mediated cellular mechanical responses. In addition, the AFM is also useful for the measurement of the cell topographic features and cell and cytoskeletal mechanical properties, such as stiffness/ elasticity. Key words: Atomic force microscopy, Vascular smooth muscle cells, Endothelial cells, Integrin, Extracellular matrix, Stiffness, Adhesion force

1. Introduction The regulation of integrin–extracellular matrix (ECM) interactions in vascular smooth muscle cells (VSMCs) and endothelial cells (ECs) underlies many important physiological and pathophysiological functions of the blood vessel wall. For example, modulation of integrin–ECM interactions in arterioles can alter vascular tone (1) and thereby influence regulation of blood pressure and blood flow. These alterations in vascular tone can be linked to integrin signaling events in both VSMCs and ECs (2, 3). As a further example, changes in integrin–ECM interaction in ECs have also been shown to significantly affect vascular permeability (4).

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_25, © Springer Science+Business Media, LLC 2011

411

412

Sun and Meininger

As alterations in blood flow, vascular resistance, and permeability are linked to many cardiovascular diseases, improved ­methods for assessing cellular and molecular responses to mechanical forces and for understanding processes governing cell adhesion are increasingly necessary. In this regard, atomic force microscopy (AFM) has emerged as a powerful tool for applying forces to cells, for investigating molecular interactions between integrin and ECM to assess regulation of adhesion (5, 6) and to evaluate cell topographical features and cell mechanical properties, such as stiffness, as indices of cytoskeletal properties (7–9). Here, we describe a detailed protocol using AFM to perform these types of studies on VSMCs and ECs.

2. Materials 2.1. Cell Culture

1. Dulbecco’s Modified Eagle’s Medium (DMEM/F-12) was supplemented with 10–20% fetal bovine serum (FBS), 2 mM l-glutamine, 1 mM sodium pyruvate, 100 unit/ml penicillin, 100  mg/ml streptomycin, 0.25  mg/ml amphotericin B, 10 mM HEPES, and 25 unit/ml Heparin for EC culture only were added into the medium. 2. 35 mm tissue culture dishes or 50 mm tissue culture dishes with a #1 glass cover slip bottom were used for culturing cells for AFM experiments. 3. 60  mm tissue culture dishes were used for maintaining cell cultures. 4. Cells were cultured in a humidified incubator (Heraeus Instruments, Inc., Newtown, CT) with 5% CO2 at 37°C.

2.2. AFM Probe Labeling

1. AFM cantilevers (model: MLCT, Veeco Metrology Inc., Santa Babra, CA), spring constant 0.01 N/m. 2. 0.5 mg/ml Human plasma fibronectin. 3. 0.5 mg/ml rat tail collagen. 4. 0.5 mg/ml human natural vitronectin. 5. 0.5 mg/ml mouse laminin. 6. 0.5 mg/ml bovine serum albumin. 7. Glass microsphere (5  mm diameter) (Structure Probe, Inc., West Chester, PA), Epoxy (Type: Low V). 8. Glass slides and coverslips. 9. Dulbecco’s phosphate buffered saline (DPBS). 10. 10 mg/ml polyethylene glycol (PEG).

Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix

2.3. Cell Treatment

413

1. Function-blocking antibodies against a5 integrin (HMa5-1; 16–100 mg/ml), b1 integrin (HMb1-1, 16–100 mg/ml), and b3 integrin (F11, 16–50 mg/ml) (all from BD Pharmingen, San Jose, CA). 2. For endothelial cell treatments, histamine (10 mM, Sigma).

2.4. Instrumentation

1. Bioscope AFM System (Model IIIa, Veeco) mounted on an Axiovert 100 TV inverted microscope (Carl Zeiss, Thornwood, NY). 2. Bioscope AFM System (Model IVa, Veeco) mounted on an IX-81 inverted microscope (Olympus, NY).

3. Methods Selection of the appropriate AFM probe characteristics and subsequent modification of the AFM probe tip to alter its biofunctional properties as well as the operation mode of the AFM is key to designing a successful AFM experiment. As outlined below, different AFM probe designs, modifications, and modes of operation are necessary to tailor the AFM for the measurement of different cell characteristics or to assess different cell functions. Topography:  To map the cell topography, a clean conical or pyramidal-tipped AFM cantilever probe is needed. Be aware that any contamination of the probe tip introduces error into the image results. The AFM was operated in contact scanning mode, such that the AFM tip was driven to raster scan across the cell surface to determine the cell’s height features (Fig. 1a–d) (10). An example of noise that can create errors in the image data caused by a contaminated AFM tip is shown in Fig. 1e. Adhesion force: To measure the integrin adhesion forces on cell surface, AFM probes need to be functionalized or coated with desired ECM proteins. When functionalized probes are brought into contact with cell surface, the ECM proteins on the probe can interact with cell surface integrins. The force required to rupture these bonds can be detected by using the AFM force mode (5, 6, 11). In this mode of operation, the AFM tip was driven to approach and retract from cell surface (0.5  Hz) in z-axis, and the cantilever deflection force in relation with the z-axis cantilever movement was measured (Fig. 2). Application of force to cells and measurement of cell generated forces: To measure the smooth muscle cell response to AFM pulling forces (12), a strong focal adhesion between the probe and cell surface needs to be formed before the AFM pulling force (800– 1,600 pN) can be applied. To achieve this, we modified the AFM

414

Sun and Meininger

Fig. 1. Illustration of measuring cell topography using AFM. (a) Schematic representation of an AFM probe scanning the surface of a cell in a raster-like fashion. (b) Height image of a vascular smooth muscle cell obtained using contact mode scanning. The scan area was 110 × 110  mm and 256 × 256  pixels. (c) Topographical image of the VSMC showing the height map of the cell in color scale. The height data was extracted from AFM height image and plotted in MATLAB. The high region of the cell was ~4 mm above the substrate. (d) Deflection image of the cell was collected simultaneously with the height image. Cell cortical features, such as cytoskeleton, were illustrated clearly. (e) Deflection image of a cell was collected using a contaminated AFM tip. White boxes illustrate the area of imaging errors caused by the contaminated tip. ((a) is reproduced from ref. 10 with permission from Taylor & Francis).

probes by gluing a glass microsphere onto the tip of the cantilever, and then coated the microsphere with ECM proteins. The microsphere (5 mm diameter) provided a much larger contact area with VSMC surface than regular cantilever tips (20–50 nm diameter), and enabled the formation of a strong focal adhesion at the microsphere–cell contact site. AFM was operated in contact mode with the scanning size set to zero, and the cell mechanical response can be recorded in the height images (Fig. 3). Measurement of cell mechanical properties: Both regular AFM probes (20–50 nm tip diameter) and AFM probes modified with microspheres (5 mm diameter) can be used to measure the cell elasticity (Young’s Modulus). In this type of measurement, the AFM was operated in Force mode, and the approach force curves were fitted with Hertz model to determine the cell elasticity (7, 8, 13) (Fig. 4). 3.1. Cell Preparation for AFM Measurement

1. Cells were cultured for 36–48 h before an AFM experiment. Immediately before an experiment, cells were washed twice with fresh medium and were incubated with either new

Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix

415

Fig. 2. Measurement of rupture forces of integrin–FN adhesions on VSMC cell surface. Typical nonadhesion (a) and adhesion (b) force curves generated using atomic force microscopy (AFM). A schematic of the AFM probe movement during the process of force curve generation is shown in the right. AFM FN-coated probes were controlled to repeatedly contact and retract from VSMC surface at a speed of 800 nm/s and 0.5 Hz frequency. During retraction, when a specific adhesion occurred, the rupture of this adhesion was detected as a sharp deflection shift in the retraction curve (b). (c and d): Force–adhesion events distributions and integrin–FN binding probabilities (black bar) of control cells (c) and cells blocked with a function-blocking antibody against b1 integrin (d). The force–adhesion events distributions were plotted in histograms and fitted with multiple Gaussian distributions (gray lines). Black lines represent the fitted model of distributions of adhesion events. (Figure reproduced from ref. 5 with permission from American Physiological Society).

Fig.  3. Measurement of VSMC response to pulling forces. (a) Displacement of FN-coated bead on VSMC surface in response to the step increases of pulling force applied by AFM. FN-coated bead fused to AFM cantilever was brought into contact with the VSMC surface to form molecular connections. Upper panel : step-increase of pulling force (Z-direction) applied onto the bead–cell connection site using AFM. Lower panel: the corresponding Z-direction movement of FN-coated bead, indicating the VSMC mechanical response to pulling forces. (b) Schematics of the VSMC response to a step increase of pulling force. (Figure reproduced from ref. 11 with permission from American Physiological Society).

416

Sun and Meininger

Fig. 4. Measurement of endothelial cell elasticity with AFM. (a) A typical extension force curve acquired on VSMC surface. The portion from point 2 to 3 was used to calculate indentation–force relationship and fitted with Hertz Model, with Ind = Zt − dt = 130 nm. Ct (between point 1 and point 2) is the distance an AFM tip was approached before touching the cell surface. The arrow depicts moving direction of AFM probe. (b) A schematic showing the indentation of cell surface by AFM probe in force mode. Point 2 is where the tip starts to touch cell surface, from point 2 to 3, the cantilever was driven down by Zt, the corresponding cell indentation is Ind, and the corresponding cantilever deflection is dt. (c) Elastic modulus of aortic endothelial cells. Histamine treatment doubled the elastic modulus of endothelial cells. ((c) is reproduced from ref. 9 with permission from Elsevier).

serum-containing medium or DMEM (without serum), depending on whether serum is required or not. Both were pre-warmed to room temperature. 2. For antibody blocking studies, integrin antibodies (HMa5-1, F11 or HMb1-1) were diluted in the experimental media and incubated with cells for 45 min. Appropriate incubation times need to be experimentally determined for antibodies used. 3. For evaluating drugs or chemical agents, appropriate concentrations are determined by concentration response studies and all agents are diluted in the experimental media. Incubation times vary with agent being evaluated. For example, to evaluate histamine effects on ECs, cells were incubated for 5 (adhesion force experiments) or 30 min (topography experiments).

Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix

3.2. Attaching Glass Bead onto AFM Cantilever

417

1. Glass slide should be thoroughly cleaned prior to the ­procedure for attaching glass bead to the AFM cantilever. The glass slides were washed with warm soap and water and rinsed thoroughly with water, followed by 1 rinse with 70% ethanol, 1 rinse with 95% ethanol, and 2 rinses with 100% ethanol and air-dried (see Note 1). 2. The AFM cantilever also needs to be cleaned before the procedure. AFM cantilevers were first rinsed in acetone, air-dried, and then rinsed in 100% ethanol and air-dried. 3. A circle (~10  mm diameter) was marked with a thin point marker on one side of the glass slide at the center of the slide. The slide was inverted and a very small amount of glass microspheres of the desired diameter were dispersed within the circled region. The marked circle provided a method for identifying the area of the slide that contained the microspheres. 4. A two part marine epoxy was thoroughly mixed and a small drop was placed on the slide ~10 mm from the circled region containing the glass microspheres. A clean coverslip was used to smear the epoxy so that it formed a very thin layer (a few micrometers thick) on the slide adjacent to the microspheres. 5. The slide was placed under the AFM with a cantilever mounted. To apply epoxy onto the cantilever, the tip of the selected cantilever was dipped into the edge of the epoxy layer and then retracted. The slide position was then adjusted to place a glass microsphere under the cantilever. The cantilever tip was then lowered onto a microsphere and following subsequent AFM retraction the glass microsphere adhered to the cantilever. 6. To cure the epoxy, the cantilever was set aside with the glass microsphere facing upward for 24 h.

3.3. Labeling of AFM Probe with ECM Proteins

1. Labeling of regular AFM probes (5, 11): The AFM probe was first rinsed with acetone and allowed to air-dry. The probe was then mounted on AFM cantilever holder and a drop of PEG (10 mg/ml, 10–20 ml) was added on the cantilever to immerse the probe for 5 min. The PEG solution was carefully removed and the probe was washed with purified water for five times in a similar manner. A drop of ECM protein (0.5 mg/ml, 10–20 ml) solution was then added to immerse the probes for 5  min. The protein solution was carefully removed and the probe washed with DPBS five times (see Notes 2 and 4). 2. To label AFM probes modified with glass microspheres, the AFM probe was mounted on cantilever holder, and a drop of ECM protein (0.5  mg/ml, 10–20  ml) solution was then

418

Sun and Meininger

added to immerse the probes for 5 min. The protein solution was then carefully removed and the probe washed with DPBS five times (see Note 4). 3.4. Topography Measurement of Smooth Muscle and Endothelial Cells

To record the cell topography, the AFM was operated in contact scanning mode (5, 9). In this mode of operation, the AFM probe is scanned across the cell surface while a constant compression force was applied to maintain contact between the AFM tip and cell surface. This allows the AFM to track cell surface features. An illustration of the scanning process is shown in Fig. 1a. The critical parameters for this operation includes the deflection set point, which determines the compression force between probe and the cell, the scanning velocity and the proportional and integral gains, which determine the cantilever adjustment at each point of contact. The topographic data includes a height map of the cell surface (Fig. 1b, c) and the deflection image created by display of deflection errors generated during height mapping to provide a contrast image that allows visualization of the cell cortical cytoskeletal fibers (Fig. 1d). Figure 1e shows a deflection image collected using a contaminated probe. Substantial noise can be observed here as illustrated by the white box (see Note 3). 1. A cell culture dish was placed on the AFM stage, a cell or a group of cells was selected by observation through the optical microscope, and cells were positioned under the AFM cantilever. 2. The softest cantilever was usually selected (typical spring constant: 0.1  N/m) and worked well for these measurements. The contact mode is selected. 3. Adjust the deflection set point so that the compression force between the probe and cell surface is 400–500 pN, adjust the initial scanning area to less than 10 mm, and the proportional gains to less than 0.5, also set the integral gain to less than 0.3 (values may vary among different systems). Start to engage. 4. After engaging the cell, gradually increase the scan size and decrease the scan rate until the scan range covers the whole area of interest while the scanning velocity is kept under 40 mm/s. Scanning in this mode with these parameters of the whole image usually took 15–30 min.

3.5. Adhesion Force Measurement on Smooth Muscle and Endothelial Cells

Force mode operation was used to determine the rupture force of specific integrin–FN interactions (5, 9). In this procedure, AFM probes were labeled with FN and allowed to repeatedly touch and retract from the surface of VSMC or endothelial cells. Two typical force curves are shown in Fig. 2a, b, and the changes of cantilever deflection are depicted in the middle column. As the probe approaches the cell (from point 1 to point 2), the cantilever is

Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix

419

straight, and the deflection force of cantilever remained at zero level. At the point that the probe contacted the cell surface (point 2), an increasing resistance force is observed, and the ­cantilever starts to bend or deflect. The resistance force gradually increases (point 2 to point 3) until the probe stops approaching and begins to retract. As probe retraction occurs (right pointing arrow-gray), the resistance force declines (point 3 to point 4). When there is no adhesion between FN and VSMC, the retraction curve resembles the approach curve (Fig. 2a). In contrast, if there were adhesions between the FN-coated AFM probe and the VSMC surface, pulling forces corresponding to the adhesion strength would reverse the deflection and bend the cantilever downward in the opposite direction. When adhesions rupture, quick deflection shifts (point 5) are observed in the retraction curve (Fig. 2b). When all the adhesions between the AFM probe and VSMC have ruptured, the cantilever returns to an unbent or straight position and the retraction curve again resembles the approach curve (see Note 5). 1. A cell culture dish is placed on the stage of the AFM, and a cell is selected by observation through the optical microscope and is positioned under the AFM cantilever. The softest cantilever is selected (typical spring constant: 0.1 N/m). Contact mode is selected in the Nanoscope software, and the scan size was set to 1 nm. 2. FN-coated probes can be applied to the surface of VSMC at randomly selected locations on the cell as desired or dictated by protocol. After engagement of the AFM probe, force mode is selected. Force curve collection rate was set at 0.5 Hz, and the z-axis ramp distance was set at 800 nm. 3. To determine the molecular specificity of the adhesions between FN and cell surface proteins, function-blocking antibodies were used to block the specific FN–integrin bindings. This same strategy is effective for other cell surface receptors. 4. For each experiment, the same AFM probe is used to obtain data from control and treated (antibody blocking) cells. The order of sampling is randomized. With each probe, 500–600 force curves are sampled from ten randomly selected cells (50–60 curves per cell) for each of the control and treated groups. 5. After collection of the force curves, the AFM cantilever should be applied to the dish bottom surface to measure the deflection sensitivity (s: units nm/v) and then lifted again to measure the cantilever spring constant (k: units N/m) using the Nanoscope software. The spring constant is measured using thermal noise method.

420

Sun and Meininger

6. All deflection shifts detected in retraction force curve are quantified to determine the adhesion force of integrin–FN bindings using the following equation:

f = d ·s ·k

(1)

where f is adhesion force, d is cantilever deflection, s is sensitivity, and k is cantilever spring constant. A proprietary program, NforceR (14), was used to automatically detect and calculate the adhesion forces. The observed adhesion forces and the corresponding number of events are plotted as histograms, and can be fitted with multiple Gaussian curves to analyze the distribution of adhesion events. (Fig. 2c, d) In addition to the adhesion force distribution, the binding probabilities can be determined as (number of force curves with adhesions)/ (number of total force curves sampled). As shown in Fig. 2c, d, the occurrence of FN–VSMC adhesion events and binding probability was significantly reduced by blocking of integrin b1 with specific function antibody (HMb1-1, 16 mg/ml). However, the distribution shown in the histogram of the adhesion forces was not significantly changed (5). 3.6. Elasticity Measurement on Smooth Muscle and Endothelial Cells

Measurement of the elasticity of VSMC or endothelial cells is also achieved by force mode operation (9). The collection of force curve is similar to that used in the adhesion force measurement. However, there are a few significant differences. The AFM probe does not require any biofunctional treatment, and the approach curve instead of retraction curve is used for elasticity calculation (7) (see Notes 5 and 6). 1. A cell culture dish is placed on the stage of the AFM, and a cell is selected and positioned under the AFM cantilever. The softest cantilever is selected (0.1  N/m). Contact mode is used in the Nanoscope software, and the scan size is set to 1 nm. 2. AFM probes are applied to the surface of the cell at randomly selected locations midway between the nucleus and cell margin or any position of choice. After engaging the AFM probe, the force mode is selected. To monitor changes in endothelial cell elasticity induced by histamine, force curves were continuously collected over 30 min to monitor the dynamic elasticity changes with or without histamine treatments. 3. After collection of the force curves, the AFM cantilever was applied on the dish bottom surface to measure the deflection sensitivity (s: units nm/v) and was lifted again to measure the cantilever spring constant (k: units N/m). 4. The relationship between deflection force and cell surface indentation was calculated for each force curve using

Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix

421

NforceR (14). The principles of these evaluations are described below and illustrated in Fig. 3a, b. The deflection force could be determined by using Eq.  1, and the probe indentation on cell surface could be determined by the following equation: Ind = Zt − dt



(2)

where Ind is indentation of cell surface, Zt is piezo travel distance from contact point (point 2, Fig.  3a) to any point beyond that (point 3, Fig.  3a, b), and dt is the amount of cantilever deflection from point 2 to 3 (Fig. 3a, b). The deflection force–indentation relationship was then fitted with Hertz Model using NforceR software to calculate the cell elasticity E: E · (Idt ) 2 f = · = k ·dt p 1 − u 2 tan (a ) 2



(

)

(3)

where f is force, E is Young’s Modulus, the cell elasticity, Idt is probe indentation, dt is cantilever deflection, k is the spring constant of the cantilever, a is the half opening angle of the tip, which is 35° for the probe we used, and n is the Poisson ratio of the cell, usually assumed as 0.5. As shown in Fig. 3c. Histamine treatment increased the elasticity of aortic endothelial cells from 8 ± 2 KPa for control cells to 20 ± 5 KPa (9). 3.7. Measuring Cellular Response to Pulling Forces on Smooth Muscle Cells

1. A glass-bottomed cell culture dish was placed on the stage of  the AFM, and a VSM cell was selected and positioned under the AFM cantilever. The AFM probe tip is modified with glass microspheres and coated with ECM proteins. The contact scanning mode is selected from the Nanoscope software menu and the scan size was set to 1 nm; this stops the probe movement in x–y axis and only records the probe movement in z-axis. Turn off the automatic plane fit and line fit in the software menu. This step is critical since the plane/line fit command could overwrite any height changes due to the cell response (see Note 3). 2. The glass microsphere at the tip of AFM cantilever can be engaged onto the cell surface in a region of choice on the cell and incubated in contact with the cell for 20 min to permit the cell to attach firmly to the microsphere. 3. Step pulling forces are then applied by adjusting the deflection set point to a smaller value. In our system, the sensitivity of the cantilever was usually ~160  nm/v, therefore, a step reduction of the set point by 0.5  n would correspond to a

422

Sun and Meininger

step increase of pulling force of ~800 pN. The set point was maintained for 5 min after each adjustment, which kept the pulling force constant for that period. The cell mechanical response would cause movement of the glass microsphere and the cantilever in the z-axis which was recorded in the height images. 4. After the final pulling force was applied, the probe was withdrawn from cell surface, and the vertical deflection voltage of the freed cantilever (VF) was recorded. The probe was then applied to the dish bottom surface to measure the deflection sensitivity (s: units nm/v) and lifted again to measure the cantilever spring constant (k: units N/m). 5. The AFM height data can be exported and extracted using Matlab or excel software. The pulling force can be determined by using Eq. 1, and the deflection of cantilever at each pull was calculated as:

di = Vi − V F

(4)

where di is the deflection voltage of pulling step i, Vi is the deflection set point of pulling step i, and VF is the vertical deflection voltage of the freed cantilever. If di < 0, AFM applied a pulling force on the cell; if di > 0, AFM applied a compression force on the cell, in which case, the height data recorded during step i cannot be used for analysis of the cell response to pulling forces. 6. An example of the VSMC cell response to pulling force is shown in Fig.  4a and a schematic of the cell response is shown in Fig. 4b. In response to the AFM applied pulling force, the bead was initially lifted for a short distance but remained connected with the cell, and then the VSM cell generated a counteracting force which pulled the bead downward (12).

4. Notes 1. To glue the microspheres onto the tip of AFM cantilevers, a  thoroughly cleaned glass slide surface is very important. Non-thorough cleaning promotes moisture condensation on the surface of the slide, and the surface tension formed by moisture around the microsphere makes it more difficult to attach the microsphere to the cantilever. Therefore, after cleaning, the slides should be stored in 100% ethanol until ready to use.

Atomic Force Microscope-Enabled Studies of Integrin–Extracellular Matrix

423

2. One problem of coating the AFM probe with ECM proteins is the formation of air bubbles on the cantilever. Trapped air bubbles make it difficult to align laser on the cantilever, and change the deflection property of the cantilever. If an air ­bubble is formed on a cantilever, the cantilever needs to be removed, dried, and recoated. Cleaning the probe with acetone is an effective way of preventing the air bubble formation. In our experience, after acetone cleaning, air bubbles rarely formed during probe coating. Careful attention needs to be exercised so that acetone does not touch the cantilever holder. This dissolves and damages the holder surface. 3. When AFM probe is initially immersed into the cell culture medium, thermal drifts of the vertical deflection are typically observed. Thermal drift causes bending of cantilever until the system reaches thermal equilibrium. The drift will usually stabilize after incubating the probe in cell culture media for approximately 30 min to an hour. It is best to allow thermal equilibrium to occur before proceeding to an actual experiment. 4. Attention needs to be paid to not touch the cantilever when removing liquid from the cantilever during the procedure of protein coating. We usually use a piece of folded lint free absorbent paper (Kimwipes) and use the edge that is free of fibers to absorb the drop of solution or buffer from the cantilever holder. The absorbent paper should be positioned 4–5 mm away from the cantilever to avoid the possibility of touching and damaging the cantilever. 5. Before engaging a cell with an AFM probe, the vertical deflection of cantilever is checked. It typically will fluctuate in a range of 0.02–0.05  n in our system. If a signal fluctuation higher than this level is observed, excessive noise will contaminate AFM measurements. Care and attention to reduce the system noise to normal level is important for all AFM studies. Noise can be introduced from a number of sources, such as a loose culture dish, a damaged or bad cantilever, loud environmental noises, air handling systems, and building vibration, as examples. 6. When calculating cell elasticity from force–indentation relationships, the indentation used to calculate the Young’s modulus should be limited to less than 150 nm. It has been shown that when indentation is greater than 150  nm, the Hertz Model is no longer suitable for calculating the Young’s modulus of elasticity for the cell (13).

424

Sun and Meininger

References 1. Martinez-Lemus, L. A., Crow, T., Davis, M. J., and Meininger, G. A. (2005) alphavbeta3and alpha5beta1-integrin blockade inhibits myogenic constriction of skeletal muscle resistance arterioles, Am J Physiol Heart Circ Physiol 289, H322–329. 2. Mogford, J. E., Davis, G. E., and Meininger, G. A. (1997) RGDN peptide interaction with endothelial alpha5beta1 integrin causes sustained endothelin-dependent vasoconstriction of rat skeletal muscle arterioles, J Clin Invest 100, 1647–1653. 3. Mogford, J. E., Davis, G. E., Platts, S. H., and Meininger, G. A. (1996) Vascular smooth muscle alpha v beta 3 integrin mediates arteriolar vasodilation in response to RGD peptides, Circ Res 79, 821–826. 4. Lum, H., and Malik, A. B. (1996) Mechanisms of increased endothelial permeability, Can J Physiol Pharmacol 74, 787–800. 5. Sun, Z., Martinez-Lemus, L. A., Trache, A., Trzeciakowski, J. P., Davis, G. E., Pohl, U., and Meininger, G. A. (2005) Mechanical properties of the interaction between fibronectin and alpha5beta1-integrin on vascular smooth muscle cells studied using atomic force microscopy, Am J Physiol Heart Circ Physiol 289, H2526–2535. 6. Li, F., Redick, S. D., Erickson, H. P., and Moy, V. T. (2003) Force Measurements of the {alpha}5{beta}1 Integrin-Fibronectin Inter­ action, Biophys. J. 84, 1252–1262. 7. Radmacher, M. (1997) Measuring the elastic properties of biological samples with the AFM

8.

9.

10.

11.

12.

13.

14.

[see comments], IEEE Eng Med Biol Mag 16, 47–57. Rotsch, C., and Radmacher, M. (2000) ­Drug-induced changes of cytoskeletal structure and mechanics in fibroblasts: an atomic force microscopy study, Biophys J 78, 520–535. Trache, A., Trzeciakowski, J. P., Gardiner, L., Sun, Z., Muthuchamy, M., Guo, M., Yuan, S. Y., and Meininger, G. A. (2005) Histamine effects on endothelial cell fibronectin interaction studied by atomic force microscopy, Biophys J 89, 2888–2898. Martinez-Lemus, L. A., Sun, Z., Trache, A., Trzciakowski, J. P., and Meininger, G. A. (2005) Integrins and regulation of the microcirculation: from arterioles to molecular studies using atomic force microscopy, Microcirculation 12, 99–112. Lehenkari, P. P., and Horton, M. A. (1999) Single integrin molecule adhesion forces in intact cells measured by atomic force microscopy, Biochem Biophys Res Commun 259, 645–650. Sun, Z., Martinez-Lemus, L. A., Hill, M. A., and Meininger, G. A. (2008) Extracellular matrixspecific focal adhesions in vascular smooth muscle produce mechanically active adhesion sites, Am J Physiol Cell Physiol 295, C268–278. Costa, K. D., and Yin, F. C. (1999) Analysis of indentation: implications for measuring mechanical properties with atomic force microscopy, J Biomech Eng 121, 462–471. Trzeciakowski, J. T., and Meininger, G. A. (2005) NforceR, Copyright Registration Number TXu1-328–659.

Chapter 26 Atomic Force Microscopy Studies on Circular DNA Structural Changes by Vincristine and Aspirin Zhongdang Xiao, Lili Cao, Dan Zhu, and Zuhong Lu Abstract In this chapter, we have presented materials and methods to study the interaction between DNA and small molecule drugs by AFM. The detailed AFM imaging of the circular DNA after incubation with ­various concentrations of vincristine and aspirin have been demonstrated. The immobilization of DNA fragments on mica surface as well as the force between tip and sample plays an important role for successful imaging of DNA–drug complexes. How to quantitatively describe the conformations and structures of circular DNA molecules and their changes is also introduced. Our work indicates that the AFM is a powerful tool in studying the interaction between DNA and small molecules. Key words: AFM, Circular DNA, DNA–drug complex, Vincristine, Aspirine

1. Introduction AFM has been widely used to examine the surface characteristics ranging from 100  mm down to the molecular scale. Compared with other conventional microscopic techniques, AFM has several unique advantages for studying biological samples. Firstly, AFM is not only able to get morphological information of biological samples at molecular resolution, but also can exert force at ­piconewton on their surfaces and make measurements simultaneously. Secondly, most native biological molecules and cells can be imaged directly by AFM without the sample prepreparation of staining, shadowing, or labeling. In addition, AFM can be operated in aqueous solutions with high resolution, which is very important to biological science because it is possible to obtain the native information for living things. In fact, AFM has become a powerful tool in the field of biomedicine (1, 2), and substantial achievements have been obtained in past years. Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_26, © Springer Science+Business Media, LLC 2011

425

426

Xiao et al.

The structure and conformation of DNA molecules are important to their functions because their change can cause the loss of DNA biological functions. Many factors can induce the change of structure and conformation of DNA. For example, double DNA chains are denatured at high temperature, and some ligands can bind to DNA molecules altering their structure and conformation (3–5). In order to assess the biological role of supercoiling, small circular DNA is frequently used as a model system for studying DNA–ligand complex conformation. Conventionally, many methods, such as X-ray diffraction patterns, NMR, and various spectroscopies, are usually employed to characterize the structure and conformation of DNA–ligand complex. Although conventional techniques have made much success, the detailed information on the structural alternation of large DNA fragments induced by ligands remains unclear. AFM can provide simple, straightforward, and high resolution results, especially to assess the effects of small molecules on DNA structures (6–9). Most importantly, AFM is unique in the possibility to study biochemical processes at a single molecule level under physiological conditions. Many small molecules, such as drugs, are able to recognize and bind to single- or double-stranded DNA with high affinity and selectivity, which could induce DNA structural alterations (10, 11). The main interaction between DNA and small molecules can be classified into electrostatic interaction, intercalation, or major and minor grooves binding. Vincristine (see Fig. 1a), a classical anticancer drug, can kill tumor cells during their replication phase (12). Aspirin (see Fig. 1b), as a cyclooxygenase inhibitor, can affect a number of physiological and biochemical regulation functions in humans (13). The two drugs are considered to bind DNA fragments and alter DNA conformation, which have been shown to occur through an intercalative action or groove binding using vibrational spectroscopic (14) and electrochemical techniques, respectively. But there is a lack of detailed information on the structural alternation of DNA fragments on

Fig. 1. The chemical structures of vincristine and aspirin.

Atomic Force Microscopy Studies on Circular DNA Structural Changes

427

single molecule level. AFM can give direct images of DNA–drug interactions demonstrating the binding states and processes of vincristine and aspirin on the DNA fragments (9). This chapter is devoted to a detailed report on studying the interaction between DNA and small molecule drugs with AFM.

2. Materials 2.1. Supports and Chemicals

1. Mica sheets with a thickness of about 0.3  mm and 13 mm × 18 mm size. 2. Aminopropyltriethoxysilane (APTES). 3. Ultrapure water of 18 MW. 4. Scotch tape.

2.2. DNA Solution and Drugs

1. Puc19 DNA (~0.5 mg/mL) solution. 2. Vincristine sulfate salt. 3. Acetylsalicylic acid.

2.3. AFM and Accessories

1. A commercial AFM equipped with a 8 mm × 8 mm scanning range and a 3-mm Z-range scanner (Agilent 5400 AFM, 5301 Stevens Creek Blvd, Santa Clara, CA 95051, United States). 2. Oxide-sharpened Si3N4 microcantilevers of 110 mm in length and a nominal force constant of k = 7.5 N/m (NSC 35/AIBS, MikroMasch).

3. Methods 3.1. Preparation of Mica Support for Sample Immobilization

1. Cleave mica with scotch tape. Make sure that the cleaved mica surface for sample preparation is completely smooth. 2. Freshly cleaved slabs of mica are exposed to an APTES atmosphere by suspending them in a glass close chamber which contained a small pool (20 ml) of APTES at 120°C for 2 h to functionalize the mica surface with amino groups (see Note 1). 3. Wash away the APTES that are not firmly attached to the mica by ultrapure water. Repeat the washing procedure at least three times. 4. The substrates are dried at 60°C in a glass vessel.

428

Xiao et al.

3.2. Preparation of Drug–DNA Complex

1. Dilute the DNA solution with ultrapure water. Please note that salt ion may affect the deformation of DNA molecules, so water was used as dilute solvent (see Note 2). 2. Make vincristine and aspirin solutions with concentration of 2 × 10−4, 2 × 10−5, 2 × 10−6 and 2 × 10−7  mol/L in ultrapure water, respectively. 3. Mix the above diluted DNA solution the above drugs solution of different concentrations by 1:1 (V/V), respectively. 4. Mix the above diluted DNA solution and equal volume of ultrapure water as control samples. 5. The drug–DNA complex and comparison are incubated for 30 min at 37°C.

3.3. Adsorption of DNA to Mica

1. The drug–DNA complex is rapidly spread on the center of a functionalized mica surface and incubated for 20 min at room temperature. 2. Rinse mica with ultrapure water several times to remove the complex that is not firmly attached to mica surface. 3. Dry under N2 gas and keep in an airtight container. 4. Transport the support onto the AFM sample holder and mount onto a microscope.

3.4. Operation of AFM

3.5. Quantitative Analysis of AFM Images

Samples are visualized by the Nanoscope AFM, operated in the tapping mode at room temperature under ambient conditions (see Note 3). Prior scanning the surface of substrate, the operating point of the instrument is set to forces below 1 nN. During scanning, the forces are kept as small as possible (<1 nN) and corrected manually to compensate for thermal shift. Two frames of 512 by 512 pixel are simultaneously recorded either showing topography or phase signal. The phase imaging often detects variations in composition, adhesion, friction, viscoelasticity, and perhaps other properties. This allows deformation of the sample in the fast scan direction to be detected and to be minimized by lowering the force applied to the stylus. Typically, the scan speed is set to 1–2 Hz (lines per second) and the tapping frequency of the cantilever was 200–400 kHz. 1. Quantitative analysis of AFM images can easily be carried out with the aid of computer programs or manual operation to design heights, lengths, and junction numbers of circular DNA fragments (see Note 4). 2. Make sure to collect enough AFM images because the results should be based on the statistic analysis of the population of different forms of DNA–drugs complex (see Notes 5–7).

Atomic Force Microscopy Studies on Circular DNA Structural Changes

429

3. When comparing results from different experiments, variability between the tips should be considered, especially to measure the width and length of DNA fragments.

4. Notes 1. Sample preparation is crucial to investigate the DNA–ligands interaction by AFM. One complicating factor is the necessity to immobilize molecules on a very flat substrate, to be able to recognize the molecules. Mica has been widely used as substrates for AFM imaging because the mica surface is atomically flat after freshly cleaved. Due to the negative charged surface in neutral solution, mica isn’t able to absorb DNA molecules that have the same negative surface charges. Therefore, mica surfaces have to be modified with positive charges for immobilizing DNA molecules or DNA–ligand complexes for AFM imaging. Bivalent cation ions, e.g., Ni2+ and Mg2+, are usually used as binding matter to immobilize DNA molecules by simply dipping into ion solution for several minutes. This method has made success in immobilizing DNA molecules for AFM study although such binding is not strong as well as salt aggregates sometimes are left on such surfaces (Fig.  2a). Some silanes, e.g., APTES, can form robust self-assembly monolayers through strong covalent

Fig. 2. The modified mica surfaces for DNA immobilization: (a) DNA fragments immobilized on Ni2+ surface, some salt particles could be often observed on the surface; (b) APTES-modified mica through vapor method showed smooth surfaces.

430

Xiao et al.

bonds while leave amino groups on the surfaces that can immobilize DNA molecules. APTES monolayer can be ­prepared by dipping freshly cleaved mica into APTES toluene and then washed with dry toluene. This approach often makes some large stains on surface because silane groups are liable to hydrolyze and aggregate together if water in toluene is not removed completely. A simple alternative method is to make mica exposing to an APTES atmosphere by suspending them in a close glass chamber that has a small pool (20 ml) of APTES inside and then to anneal at 120°C for 2 h. Relative smooth monolayer can be obtained by this method (Fig. 2b). Although there sometimes are some small aggregates on the surface by this method, these small aggregates do not influence the imaging of longer DNA chains. The roughness of modified surface can be easily measured in AFM images by software to illustrate the quality of monolayers. Contact angle measurement is another convenient way to characterize the APTES monolayers. 2. It is better to use low concentrations of mixture of DNA and drugs because they tend to concentrate in small areas during the air-drying phase, whereas single separate DNA–drug combination provides a clearer image. Be sure to position AFM tip at the location of the specimen on the mica surface to avoid wasting time. 3. In general, three modes of AFM can be used to image sample surfaces: contact mode, noncontact mode, and tapping mode. In contact mode, the overall force between the tip and the surface is repulsive during scanning, and thus may induce sample degradation effects. One should be careful to control the contact force when imaging the soft samples, such as DNA molecules, which is sometimes difficult for a new operator. In noncontact mode, an AFM is operated in attractive force as the tip of the cantilever does not contact the sample surface, so the liquid layer absorbed on DNA may influence the imaging of DNA molecules themselves when scanning in atmosphere. In tapping mode, the tip is oscillated close to the resonant frequency of the cantilever so that the tip makes contact with the sample only for a short duration in each cycle of oscillation. The tip is lowered until it approaches the surface, and due to the interaction with the surface the oscillation amplitude decreases. This reduction of the amplitude is kept constant in a feedback loop, and the feedback signal is used for imaging. This mode can minimize the loading forces on sample surface, particularly lateral forces. So tapping mode AFM tends to be more applicable and widely used to general imaging of soft samples, such as DNA–drugs complexes.

Atomic Force Microscopy Studies on Circular DNA Structural Changes

431

4. The resolution perpendicular to the surface mainly depends on the sensitivity of the cantilever and the piezo and is of the order 0.1 nm. The lateral resolution is more difficult to determine, and is mainly determined by the geometrical convolution of the tip and the sample. The apparent size of an object in an AFM image can be obtained as approximately rR by simple calculation, where r is the effective radius of the sample and R the effective radius of the tip. The size of microfabricated Si3N4 tips is not well defined and varies between batches. It usually has an end radius of approximately 20–50 nm. DNA molecules with a diameter of ~2 nm then have an apparent width of ~10 nm. The apparent height of flat DNA chain on mica surface was measured to be 0.4–0.5 nm in a tapping AFM image, less than the theoretical value ~2  nm. DNA molecules fatten when absorbed onto a mica surface, decreasing the height of DNA. Another reason is that AFM tip can indent into the DNA molecules after penetrating the liquid layer on the DNA surface. The measured height is close to the theoretical value when using noncontact mode because it images both the liquid and surface. 5. Some useful data can be obtained quantitatively from a ­successful image of DNA molecules. The contour length of every DNA molecule can be determined from their skeletons in AFM images. This may be done by either simple hand tracing or some special software. If DNA molecules were broken by UV or drugs, it is easy to be clearly distinguished from undamaged DNA fragments based on their contour lengths. Figure 3a shows many broken DNA pieces after incubating circular chains with drugs, and the resulting contour length distribution is usually used to describe the broken extent of DNA fragments (Fig. 3b), showing 300 nm length on average. If other molecules such as proteins and drugs bind onto DNA chains, it is easy to reveal the binding place by measuring the height of DNA in AFM image (Fig.  3c). Long or circular DNAs usually have irregular supercoiled conformations that can be clearly demonstrated in AFM images. To quantitatively describe the supercoiled conformation is a very difficult thing, but counting the nodes may be an easy way to go to this aim because the number of nodes reflects the characteristic of DNA molecules and their interaction with the surface. Figure 3d shows the supercoiled conformations of a plasmid DNA and their node number distributions. 6. Concentration is an important factor to study the influence of vincristine and aspirin on conformation of DNA fragments. When using low vincristine concentration, for example 10−7 M, circular DNA chains became thicker, invariably irregular and folded with no defined superhelix axis compared to the naked DNA (Fig. 4a vs. Fig. 3d). Increasing the

432

Xiao et al.

Fig. 3. The examples illustrating the data obtained from a DNA image: (a) and (b) show the contour length and its distribution, respectively. (c) The height of DNA fragment revealing the binding place of other molecules. (d) The supercoiled conformations of a plasmid DNA and their node number distributions.

vincristine concentration to 10−6 M or 10−5  M showed an obvious change in the circular conformation, since the molecules broke into small fragments although the length distribution of broken fragments was different between these two concentrations (Fig. 4b, c). The length and node distribution are obtained from many AFM images. An interesting phenomenon was observed when the vincristine concentration increased to 10−4  M. Almost no change on the conformation of circular DNA occurred under this vincristine concentration compared to naked DNA (Fig.  4d). This may be determined by the properties of vincristine.

Atomic Force Microscopy Studies on Circular DNA Structural Changes

433

Fig. 4. Conformations and structures of circular DNA molecules under various concentrations of vincristine and their quantitative analysis: (a) 10−7 M, (b) 10−6 M, (c) 10−5 M and (d) 10−4 M, respectively.

As is well known, vincristine is an anticancer drug by interaction with DNA chain. The present result implies that very high concentration of vincristine might have no pharmacological effect. Therefore, the concentration of drug must be firstly considered when studying the interaction between drug and DNA molecules. It is notable that the interaction of vincristine with DNA molecules may vary on patterns and concentrations between linear, circular, or different sizes of DNA (9). 7. DNA–aspirin complexes were imaged by AFM under ambient conditions. At low aspirin concentration of 10−7 M, most DNA chains can keep a defined superhelix axis, presenting no obvious change compared with the naked DNA. Some DNA

434

Xiao et al.

chains have the broader and higher regions that were considered to be aspirin-binding fractions (Fig.  5a), and the number of DNA molecules having binding regions increases with the aspirin concentrations (Fig. 5b, c) The DNA–aspirin binding regions were not homogenous on the whole DNA fragment, and they gradually extended along the binding sites with increasing drug concentrations. The average number of nodes also increases with increasing drug concentrations. At high concentration of 10−4 M, some DNA chains intertwined into sticks, showing more complex conformations (Fig.  5d). Keeping proper scanning parameters is very important for quantitatively determining the DNA–aspirin complexes. Moderate tapping amplitude and set point can get clear

Fig. 5. Conformations and structures of circular DNA molecules under various concentrations of aspirin and their quantitative analysis: (a) 10−7 M, (b) 10−6 M, (c) 10−5 M and (d) 10−4 M, respectively.

Atomic Force Microscopy Studies on Circular DNA Structural Changes

435

images as well as accepted indention of AFM tip. A large number of images should be collected for statistical analysis. It also is notable that the interaction of DNA–aspirin may vary on patterns and concentrations between linear, circular, or different sizes of DNA (9).

Acknowledgments This work was financially supported by the National Basic Research Program of China (973 Program: 2007CB936300), NSFC (No. 20875014 and 30870626) and 2008DFA51180. References 1. Osmulski, P. A., Gaczynska, M. (2008) AFM of biological complexes: What can we learn? Current Opin. Colloid Interface Sci. 13, 351–367. 2. Gu, J. H., Xiao, Z. D., Yam, C. M., Qin, G. T., Deluge, M., Boutet, S., Cai, C. Z. (2005) Attaching single biomolecules selectively to the apex of AFM tips for measuring specific interactions. Biophys. J. 89, L31–L33. 3. Krautbauer, R., Fischerlander, S., Allen, S., Gaub, H. E. (2002) Mechanical fingerprints of DNA drug complexes. Single Mol. 3, 97–103. 4. Viglasky, V., Valle, F., Adamcik, J., Joab, I., Podhradsky, D., Dietler, G. (2003) Anthracycline-dependent heat-induced transition from positive to negative supercoiled DNA. Electrophoresis 24, 1703–1711. 5. Banerjee, T., Mukhopadhyay, R. (2008) Structural effects of nogalamycin, an antibiotic antitumour agent, on DNA. Biochem. Biophys. Res. Commun. 374, 264–268. 6. Bussiek, M., Mucke, N., Langowski, J. (2003) Polylysine-coated mica can be used to observe systematic changes in the supercoiled DNA conformation by scanning force microscopy in solution. Nucleic Acids Res. 31, e137. 7. Borovok, N., Molotsky, T., Ghabboun, J., Cohen, H., Porath D., Kotlyar A. (2007) Poly(dG)–poly(dC) DNA appears shorter than poly(dA)–poly(dT) and possibly adopts an A-related conformation on a mica surface under ambient conditions. FEBS Lett 581, 5843–5846.

8. Tseng, Y. D., Ge, H. F., Wang, X. Z., Edwardson, J. M., Waring, M. J. Fitzgerald, W. J., Henderson, R. M. (2005) Atomic Force Microscopy Study of the Structural Effects Induced by Echinomycin Binding to DNA. J. Mol. Biol. 345, 745–758. 9. Zhu, Y., Zeng, H., Xie, J. M., Ba, L., Gao, X., Lu, Z. H. (2004) Atomic force microscopy studies on DNA structural changes induced by vincristine sulfate and aspirin. Microsc. Microanal. 10, 286–290. 10. Krasnoslobodtsev, A. V., Shlyakhtenko, L. S., Lyubchenko, Y. L. (2007) Probing interactions within the synaptic DNA-Sfil complex by AFM force spectroscopy. J. Mol. Biol. 365, 1407–1416. 11. Sorel, I., Pietrement, O., Hamon, L., Baconnais, S., Le Cam, E., Pastre, D. (2006) The EcoRIDNA complex as a model for investigating ­protein-DNA interactions by atomic force microscopy. Biochem. 45, 14675–14682. 12. Waterhouse, D.N., Dos Santos, N., Mayer, L.D. & Bally, M.B. (2001) Drug-drug interactions arising from the use of liposomal vincristine in combination with other anticancer drugs. Pharm Res 18, 1331–1335. 13. Frantz, B., O’Neill, E.A., Ghosh, S. & Kopp, E. (1995) The effect of sodium salicylate and aspirin on NF-kB. Science 270, 2017–2019. 14. Neault, J.F., Naoui, M., Manfait, M. & TajmirRiahi, H.A. (1996). Aspirin–DNA interaction studied by FTIR and laser Raman difference spectroscopy. FEBS Lett 382, 26–30.

wwwwwwwwwwwwww

Part VI Atomic Force Microscopy as a Nanotool

wwwwwwwwwwwwww

Chapter 27 Combined Atomic Force Microscopy and Fluorescence Microscopy Miklós S.Z. Kellermayer Abstract The atomic force microscope (AFM) is a high-resolution scanning-probe instrument which has become an important tool for cellular and molecular biophysics in recent years, but lacks the time resolution and functional specificities offered by fluorescence microscopic techniques. The advantages of both methods may be exploited by combining and synchronizing them. In this paper, the biological applications of AFM, fluorescence, and their combinations are briefly reviewed, and the assembly and utilization of a spatially and temporally synchronized AFM and total internal reflection fluorescence microscope are described. The application of the method is demonstrated on a fluorescently labeled cell culture. Key words: Fluorescence, Mechanics, Topography, Monolayer cells, Actin filaments, Evanescent field

1. Introduction The atomic force microscope (AFM) is a high-resolution ­scanning-probe device in which sample topography is explored by scanning a sharp tip at the end of a flexible cantilever across its surface (1). Ever since the introduction of AFM, there have been numerous attempts to combine it with other imaging modalities, including fluorescence microscopy (2–4). The motivation for combining AFM with fluorescence stems from the desire to exploit the complementary advantages of either technique. Whereas AFM provides high spatial resolution and the possibility of mechanical manipulation, fluorescence offers high temporal resolution, sensitivity to local physical chemistry, and the possibility of functional imaging via specific labeling procedures. In the present paper, the current state of combined AFM and fluorescence methodologies are briefly reviewed; then, the steps of assembling and running a temporally and spatially synchronized Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_27, © Springer Science+Business Media, LLC 2011

439

440

Kellermayer

AFM and total internal reflection fluorescence microscope (TIRFM) are described and an example of its use on a fluorescently labeled cell culture is presented. In AFM, the atoms at the cantilever tip surface and sample surface interact in an attractive or repulsive manner which deflects the cantilever. The miniscule motions of the cantilever are detected by directing a laser beam on the cantilever which is reflected and projected onto a position-sensing photodiode. By monitoring the position of the laser beam, sub-angstrom cantilever deflections can be detected, and a surface topographical image is reconstructed. Because of the high sensitivity by which cantilever deflections can be detected and the sample positioned, AFM allows the exploration of atomic and molecular spatial dimensions. In recent years, the AFM has been increasingly used in biological applications owing to improvements in sample preparation, development of different data acquisition modes (contact, noncontact, intermittent contact) and experimental geometries (e.g., force spectroscopy), and the ability of functionalizing the AFM tip that interacts with the sample. High-resolution structure of proteins and protein–DNA complexes can be obtained under aqueous buffer (5). The AFM offers special advantages in membrane biology, where otherwise difficult-to-study membrane proteins can be directly visualized and even manipulated (6–8). Tip functionalization provides a particularly wide array of opportunities for exploring interactions between molecules (9). By functionalizing the tip with carbon nanotubes, even greater spatial resolution and topographical access can be obtained (10–13). AFM can also be used to mechanically manipulate the sample surface or individual molecules in a method that has become known as dynamic force spectroscopy (14–20). In this measurement mode, the surface is not scanned; rather the cantilever is moved in perpendicular to it. Any molecule captured between the surface and the tip of a calibrated (21) cantilever becomes mechanically loaded, allowing the investigation of a range of phenomena (22), such as molecular elasticity (23), protein folding/ unfolding (24), conformational transitions (25), or the dynamic strength of biomolecular bonds (26). In many of the commercially available AFMs, the imaging and force-measuring modalities are conveniently coupled. In spite of the unparalleled advantages offered by AFM, it has its deficiencies. One of the main limitations is the low image-acquisition rate. Although there are attempts to devise high-speed AFMs by using custom-made, high-resonance-frequency scanners and cantilevers (27, 28), usual AFM image acquisition is on the minute time scale. The other drawback of AFM imaging is that the sample is accessed from the surface. Therefore, in dense and complex molecular environments, such as the living cell, effective resolution is very low and individual molecules cannot be easily identified. These limitations

Combined Atomic Force Microscopy and Fluorescence Microscopy

441

may be overcome by combining the AFM with light-microscopic imaging modalities, out of which fluorescence microscopy offers the most diverse functionality. Fluorescence is the radiation-accompanied relaxation of a suitable molecular system from an excited singlet electronic state to the ground state (29). Sample fluorescence may originate from intrinsic fluorophores (such as tryptophan), but most commonly external fluorophores attached via chemical or genetic means to specific sites are utilized (5). To avoid the uncertainty of nonspecific labeling, various site-specific conjugation strategies have evolved (30, 31). Fluorescent labeling of protein molecules can also be achieved by genetic conjugation to the green fluorescent protein (GFP) (32) or its spectral variants (33). The main advantage of this approach is that in vivo measurements can easily be carried out without the need for complicated delivery mechanisms of the protein into the cell (34). Quantum dots, which are semiconductor nanocrystals (35), represent a family of novel fluorescent labels with the advantages of having broad excitation spectra, tunable emission spectra, and a very high resistance to photobleaching, thereby gaining increasing utilization in in  vivo imaging (36–38). The excitation radiation is ­delivered to the sample in either a conventional wide-field epifluorescence, confocal, or total TIRF microscopic arrangement (5). Fluorescence emission is detected by either two-dimensional detectors (ever-improving wide-field CCD cameras) or pointsensors (photomultiplier tubes, avalanche photodiodes). Emitted photons may be segregated according to various criteria: count, wavelength, polarization, and arrival time. Special fluorescence modalities, such as single-pair fluorescence resonance energy transfer (spFRET) (39–41) or fluorescence lifetime imaging microscopy (FLIM) (42–45), open the way to explore molecular interactions, measuring nanoscale distances and their fluctuations, and the local chemical environment of the fluorophore. The main limitation of fluorescence microscopy is its low spatial resolution. Being a diffraction-limited technique, the spatial resolution is fundamentally limited by the Rayleigh criterion, dmin = (0.6l) / (n sina ), where dmin is the minimal resolved distance, l is the wavelength of the light used for imaging, and n sin a is the effective numerical aperture of the imaging system (46). Recently, there have been numerous attempts to tackle the limitations set by the Rayleigh criterion, leading to the development of sophisticated imaging technologies. By diluting the examined fluorophores either physically (fluorescence imaging with one nanometer accuracy, FIONA (47, 48)) or photochemically (Stimulated Emission Depletion, STED (49); PhotoActivation Localization Microscopy, PALM (50)), the resolution problem is converted into a position-sensing problem, and effective resolution­in the nanometer range may be obtained.

442

Kellermayer

There have been many early attempts to combine AFM and fluorescence imaging (2–4, 51–55). Due to the vast differences in acquisition rates and modes of imaging, the acquisition of AFM and fluorescence is most often sequential. Membranes, synthetic or cellular, are particularly well suited for investigation with combined AFM and fluorescence microscopies (56–60). These methodologies have been also applied to the imaging of nanoparticles (61, 62) and supramolecular assemblies, such as amyloid fibrils (63, 64), myosin thick filaments (65, 66), fibrin (23, 67), chromatin and chromosomes (68, 69), protein–DNA complexes (70), and the cytoskeletal system (71–74). In the case of cells, besides imaging (75–78), the application of mechanical forces have been employed in combined AFM and fluorescence systems (79–84). Temporally synchronized AFM and fluorescence measurements have typically been carried out in force spectroscopy or evanescent field mapping measurements (54, 85–87). Multiple modalities, such as AFM-TIRF-confocal imaging (88) or AFM-fluorescence-electrical potential measurements (90), undoubtedly yield unprecedented insights into the inner workings of the living cell. Total TIRF Microscopy (TIRFM) (89), due to the narrow excitation field and its ability to detect and follow individual fluorophores, may offer special advantages in AFM–fluorescence combinations. First, the cantilever tip, which may itself be fluorescent, is not excited because it stays above the evanescent field. Second, sensitive force measurements are not influenced by the excitation light impinging on the cantilever. In TIRFM, an evanescent field is formed in the lower index of refraction phase at a boundary between two phases of different refraction indices, provided that the light beam arrives at this boundary at an angle greater than the critical angle. The intensity of the evanescent field decays exponentially with the distance from the boundary according to I z = I 0 e− z / d , where I0 is the intensity at the boundary, Iz is the intensity measured at a distance z from the boundary, and d is decay length that depends on the relative refractive indices (n1, n2) and the angle of incidence (q), and scales with the −1/ 2 wavelength (l) of the light used as d = (l 4p )n12 sin 2 q − n 22  . Two main types of TIRFM geometries have been implemented depending on how the excitation beam is coupled into the microscope: the prism method and the objective-based method. Although the prism method has been used in AFM-fluorescence combination (67), the objective-based TIRFM lends itself more conveniently to adding an AFM to the fluorescence microscope (54, 87). In the following sections, the assembly and utilization of a spatially and temporally synchronized AFM and TIRFM are described, along with an application example of imaging cells in which the actin filaments are fluorescently labeled.

Combined Atomic Force Microscopy and Fluorescence Microscopy

443

2. Materials 2.1. Preparation of Cleaned Coverslips

Factory coverslips are usually covered with fluorescent contaminants. It is, therefore, important to make every effort to clean coverslips particularly when imaging single molecules, but it is also advised in the case of handling cell cultures. 1. Coverslips are placed vertically in a slide-staining container, in which all the rinsing steps can be carried out. For cell ­cultures, round coverslips with a diameter of 25 mm and #1 thickness are used. 2. Sonication in absolute ethanol for 20 min, followed by rinsing three times (3×) in MilliQ water (deionized water, ~20 MW conductivity). For sonication, the container is placed in a sonicating water bath. 3. Sonication in spectroscopic grade acetone for 20  min, ­followed by rinsing 3× with MilliQ water. 4. Incubation in concentrated nitric acid (HNO3) for 20 min, followed by rinsing 3× with MilliQ water. 5. Incubation in filtered 6N KOH solution for 20 min, followed by rinsing 3× with MilliQ water. Coverslips may be stored in MilliQ water at this stage. Prior to further use, coverslips are blown dry with high-purity N2 gas, and then illuminated with a mercury lamp (HBO50) for 30  min to photobleach any remaining background fluorescence. For cell culture, cleaned coverlips are dipped in absolute ethanol and sterilized by burning ethanol away with a gas flame.

2.2. Buffers and Solutions

1. Cell culture medium: Dulbecco’s Modified Eagle Medium (DMEM) containing 10% fetal bovine serum, 2% l-glutamine, 1% penicillin–streptomycin at 37°C in a humidified 5% CO2 atmosphere. 2. Phosphate-buffered saline (PBS): 10 mM K2HPO4/KH2PO4, pH 7.4, 140 mM NaCl. 3. Fixing solution: 2% paraformaldehyde in PBS, freshly prepared. 4. Blocking solution (PBS–BSA): 1% bovine serum albumin (BSA) in PBS. 5. Detergent solution: 0.1% Triton X-100 in PBS–BSA solution. 6. Washing solution: 10 mM ethanolamine in PBS. 7. Labeling solution: 200  nM tetramethylrhodamineisothiocyanate­ (TRITC)-phalloidin (Sigma Chemical Co, St. Louis, MO) in detergent solution. TRITC-phalloidin is stored as a 100  mM stock solution in ethanol at −20°C. Labeling solution is prepared fresh by adding 4 ml TRITCphalloidin stock solution to 2 ml detergent solution.

444

Kellermayer

2.3. Equipment 2.3.1. Microscopic Beads

Fluorescent microscopic beads may be used for aligning purposes. Fluorescent polystyrene beads (Polysciences Inc, Warrington, PA), 100 nm in diameter are used. Beads are washed with MilliQ water by centrifugation and resuspension three times. 5 ml of bead suspension is pipetted onto a clean coverslip and incubated for 10 min at room temperature. The coverslip is gently rinsed with MilliQ water and dried with a stream of clean N2 gas.

2.3.2. Cantilevers

For dry samples, Olympus AC160TS cantilevers (Si3N4, resonance frequency ~300  kHz) are used. For samples under liquid or for mechanical manipulation, usually Olympus BioLevers (lever B, Au-coated on both sides, resonance frequency ~35 kHz) are used.

2.3.3. Epifluorescence Microscope

In principle, any high-quality inverted epifluorescence microscope may work. Because of environmental considerations (see below), it is recommended to use a high-stability motorized microscope which can be remote controlled. If the control procedures can be programmed and interfaced with the AFM control, then a highly versatile system can be constructed. In the present work, we used an Olympus IX81 motorized inverted epifluorescence microscope with TIRF attachment coupled to two customattached lasers (488  nm, 15  mW, Novalux, Sunnyvale, CA and 532 nm YAG, 50 mW, JDS Uniphase) via an optical fiber (Point Source, UK). The microscope is ideally equipped with a grayscale or color CCD camera for viewing the sample and the cantilever position, and with a photon-counting detector (Avalanche photodiode, APD, EG&G Canada). For viewing fluorescent images, a sensitive EMCCD camera is attached (Andor, Ireland).

2.3.4. Atomic Force Microscope

While most brand-name AFMs can be combined with the fluorescence microscope one way or the other, for fully synchronized AFM and fluorescence data acquisition, a stage-scanning AFM is best suited because in this case the co-alignment of the optical axis, defined by the light microscope, and the mechanical axis, defined by the cantilever tip, can be maintained throughout the measurement (Fig. 1a). In the present work we used an Asylum Research MFP3D instrument (Asylum Research, Santa Barbara, CA). The AFM, together with the underlying epifluorescence microscope, is placed in a vibrationally isolated environment (see Note 1).

2.3.5. Accessory Electronics

The APD is connected to both a custom-built rate meter and a digital module (Asylum Research). The rate meter integrates photon count for user-adjustable time windows (20 ms–500 ms) and generates a voltage signal (0–10 V) proportional to fluorescence intensity. The digital module allows the acquisition of photon count directly by the AFM driver and software, enabling

y’

a

AFM head x’

Mechanical (z’) axis

AFM cantilever

y’’

Fluorescent sample stage x’’

Objective lens Optical (z) axis

Excitation beam

Emission path

IRfilter pinhole Point detector (avalanche photodiode)

b

532 nm YAG laser

Longpass dichroic

TIRF Objective lens (1.45 NA)

488 nm diode laser Lens

Lens

Coupling lens Moveable Field diaphragm fiber-coupling assembly Fiberoptic cable

Long-pass dichroic filter Emission filter Focusing lens IR filter

PC

AFM controller

Digital module Rate meter

40 µm pinhole Avalanche photodiode

Oscilloscope

Fig. 1. Schematics of the combined, synchronized AFM and fluorescence microscope system. (a) Concept of instrumental arrangement. The optical (z ) axis is defined by the optics of the microscope, whereas the mechanical (z ¢) axis by the cantilever tip. The cantilever, together with the AFM head, can be positioned in the horizontal plane, thereby enabling the superposition of the z and z ¢ axes. The sample can be positioned independent of either the optical (z ) or mechanical (z ¢) axes. Excitation light arrives via the objective lens. If the incident angle of the excitation beam exceeds the critical angle, then total internal reflection (TIR) occurs, and an evanescent field builds up in the aqueous compartment. Emission is collected via the objective lens and projected onto a photon-counting module (avalanche photodiode). (b) Arrangement of the instrumental components of the TIR fluorescence microscope (TIRFM). In this objective-based TIRFM two lasers are coupled, via an optical fiber, into a relay lens system and then into the microscope. The emitted photons are detected by an avalanche photodiode (APD), the signal of which is either directly coupled, via a digital module, into the AFM controller, or entered into a rate meter that converts photon counts to intensity.

446

Kellermayer

pixel-level synchrony of parallel data acquisition. It is advised that the signals are displayed on an oscilloscope for an easy real-time visualization and inspection.

3. Methods 3.1. Fluorescent Labeling of Monolayer Cells

1. Cultured HeLa cells are rinsed with warm (37°C) PBS twice. 2. Incubate in detergent solution for 30 min at room temperature. Wash with PBS twice. 3. Incubate in fixing solution for 15 min at room temperature. Wash with PBS twice. 4. Rinse with washing solution for 5 min at room temperature. Wash with PBS twice. 5. Stain with labeling solution in the dark for 20 min at room temperature. Wash with PBS three times, by rinsing for 5 min each time. Add 10  mM ß-mercaptoethanol for reducing photobleaching. 6. Optionally, dry the sample with a stream of clean N2 gas.

3.2. Combining AFM with Fluorescence Imaging

The schematics of the combined AFM and TIRFM are shown in Fig.  1b. The objective-based TIRF microscope is built on an inverted motorized epifluorescence microscope with a TIRF module and a 60×, 1.45 NA oil-immersion objective lens. Light from a green laser (YAG 532  nm, 50  mW) and a blue laser (488 nm, 15 mW) is joined with a long-pass dichroic mirror and coupled to the TIRF module via an optical fiber. The angle of incidence is adjusted by positioning the free fiber end in the plane conjugate to the back focal plane of the objective lens. The size of the illuminated field is adjusted with a field diaphragm conjugate with the sample plane. Emitted fluorescence is collected via a long-pass dichroic, emission filter, collector lens, IR filter, and a confocal aperture (40 mm), and detected with an APD. The signal is processed by either counting the photons with a digital access module or by calculating intensity with a custom-built rate meter. The sample is positioned on a special, double XY microscope stage (Fig. 1a). The lower, mechanical stage permits the movement of the AFM head (together with the upper, sample stage) relative to the optical axis. The upper, sample stage is a closedloop XY-piezo stage. The rate-meter signal, proportional to fluorescence intensity, is entered into the AFM controller via its BNC panel. Photon count signal is entered via the digital port of the controller.

Combined Atomic Force Microscopy and Fluorescence Microscopy

3.3. Imaging with Combined AFM and TIRFM

447

1. Clean the TIRF objective with spectroscopic grade acetone. Attach one to two drops of low-fluorescence immersion oil. 2. Mount sample on stage. Round coverslips containing the fluorescently labeled cell culture are mounted in a fluid cell module. Alternatively, the coverslip may be attached, using silicon grease, onto the bottom of a flat plexiglas ring. 3. Position the AFM head on the stage. Lower the cantilever under feedback contol. It is best to lower the cantilever onto a region devoid of cells, so that the sample itself is not damaged. 4. Under visual control and brightfield illumination, using the CCD camera image displayed on a TV screen, position the cantilever tip to the center of the field of view. The cantilever tip should be gently placed on the glass surface. By focusing on the cantilever tip, the coverslip surface is brought into focus as well. If the cantilever is far out of the field of view, then it helps to switch to a lower magnification (e.g., 4–10×) objective lens. Once the cantilever tip is centered, disengage it from the sample surface (see Note 2). 5. Illuminate the sample with the exciting laser beam and adjust for total internal reflection. Open the shutter of the TIRF attachment module and, under visual control, adjust the incident angle with the micrometer. While increasing the angle of incidence of the exciting laser beam, fluorescence emission suddenly increases as the critical angle is reached. Upon increasing the angle of incidence further, total internal reflection is obtained, in the case of which a high-contrast, lowbackground fluorescent image is observed. 6. Adjust the sample position. While the cantilever tip remains disengaged, bring the desired part of the sample into the field of view under visual control. Only the sample stage should be moved, and the AFM head stage and thereby the cantilever tip location must remain stationary. Close the field of view as much as possible in order to avoid photobleaching of large areas of the sample during scanning. The field diaphragm should be centered and co-aligned with the cantilever tip. 7. Record a fast TIRFM image by scanning rapidly (up to 4 Hz) about a large field (up to 90 × 90  mm2) (see Note 3). It is advised that the sample focus is adjusted during this process. 8. Engage the tip and the sample. Adjust the scanning frequency (0.5–1 Hz) and start scanning. Photon counts and user input signal (rate-meter voltage proportional to fluorescence

448

Kellermayer

Fig. 2. Representative image acquired with synchronized AFM/TIRFM. (a) Height contrast data corresponding to the topography of the sample. The sample is a monolayer HeLa cell culture labeled with TRITC-phalloidin, fixed and dried. (b) Fluorescence emission intensity information collected in synchrony with the topography data. (c) Three-dimensionalrendered surface topography colored with the fluroescence data.

i­ntensity) are collected in parallel with the AFM imaging modalities (height, amplitude, phase, and deflection) (see Figs. 2 and 3 and Notes 4 and 5). 3.4. Manipulation with Combined AFM and TIRFM

1. Attach fluorescently labeled molecules to the cantilever tip. We use TRITC-labeled titin (87) to coat the cantilever by pipetting an aliquot (5 ml) of sample (0.05 mg/ml) onto the

Combined Atomic Force Microscopy and Fluorescence Microscopy

449

Fig. 3. Problem of focusing in AFM/TIRFM. (a) Amplitude contrast AFM (left ) and fluorescence image (right ) of a monolayer HeLa cell. The AFM image is properly collected in spite of the slight out-of-focus fluorescence image. (b) Scanning TIRFM image demonstating the effect of focusing during data acquisition. (c) Height contrast AFM (left ) and fluorescence image (right ) of the periferal ctoplasm of a Triton X-100 treated HeLa cell. Because of the proper focus and low drift, fine detail can be discerned in the images.

cantilever tip and incubating at room temperature for 10 min. Remove unbound molecules by successive washes with buffer (PBS). 2. Calibrate cantilever. Cantilever stiffness is calibrated by the thermal method (21). 3. Align cantilever tip to the center of the field of view. 4. Press the cantilever to a clean coverslip surface, and simultaneously record the fluorescence and force data. By plotting the force and fluorescence data as a function of cantilever displacement, the evanescent field is mapped along the optical axis (see Fig. 4 and Note 6).

450

Kellermayer

Intensity (AU)

3.0

2.0

1.0

coverslip evanescent field 15

Force (nN)

AFM cantilever 10

Fluorescent molecule

Movement direction

5

Microscopecoverslip coverslip

0

0.0

0.5

1.0

1.5

2.0

Distance (µm)

Fig. 4. Mapping the evanescent field with AFM. By moving the cantilever carrying a fluorescently labeled molecule along the vertical axis, the axial distribution of the excitation intensity can be mapped. The upper curve shows the fluorescence emission intensity and the lower the force (proportional to cantilever deflection) as a function of cantilever displacement. Inset, arrangement of the cantilever moving relative to the microscope coverslip surface.

4. Notes 1. Environmental aspects. For high-quality images and force spectra, the AFM workstation is positioned in a vibrationally and acoustically isolated environment. In addition, because sensitive photonic detectors are used, the instrument is enclosed in a light-sealed cabinet. A custom-built cabinet lined with dark acoustic foam may be assembled for this purpose. Bright LEDs on instruments near the microscope are covered with tape. 2. Co-aligning the optical and mechanical axes. For proper spatial synchrony between AFM and fluorescence images, the optical and mechanical axes need to be co-aligned. Alignment is carried out by positioning the AFM cantilever, hence the entire AFM head, relative to the optical axis defined by the microscope objective lens. The co-alignment is best done by ­positioning a fluorescently labeled cantilever tip with the aid of the APD signal. Because the APD senses only signals arriving from the exact center of the field, the cantilever is first roughly ­centered under video control. If the cantilever is far out of the field of view of the TIRF objective lens, then a

Combined Atomic Force Microscopy and Fluorescence Microscopy

451

lower-­magnification objective is first used. Notably, in the combined Olympus IX81 TIRF microscope and Asylum Research MFP3D system used here, the objective lens must be pulled all the way down (away from the stage) to avoid jamming into the microscope stage. The fluorescently labeled cantilever tip is engaged with the surface of a clean coverslip and brought into the center of the field of view. Then the tip position is finely co-aligned with the optical axis by following the fluorescence intensity measured by the APD and displayed on the oscilloscope. The AFM stage micrometers are adjusted so that the fluorescence intensity is maximized along both the X and Y axes. It is recommended to mark the position of the cantilever tip on the video screen for easy subsequent ­co-alignment. The success of co-alignment can be monitored by model samples, such as fluorescently labeled microscopic beads mounted on a coverslip surface. It is best to use beads which are smaller than the diffraction limit. Alignment of the fluorescent and topography data can be done a posteriori as well, by correcting for the Dx and Dy values measured between the datasets. However, because of temporally synchronized data acquisition, spatial misalignments amount to temporal differences as well. 3. Imaging with scanning TIRFM. The instrument employed here (Fig. 1) is essentially a stage-scanning, TIRF-illuminated confocal microscope. Although it utilizes the AFM controller hardware and software, the instrument can nevertheless be used for fluorescence image acquisition independent of AFM imaging. By disengaging the AFM cantilever tip, TIRF images can be acquired faster than AFM images by rapily scanning the stage. Scanning frequency is limited by the speed of closedloop piezo stage movement. In order to avoid photobleaching of the sample over a wide area, the field diaphragm is closed. At full closure, approximately a 5-mm wide cirular area of the sample­is illuminated. Note that under these settings, much of the excitation intensity is discarded. Excitation intensity can be further reduced, if necessary, by placing neutral density filters in the excitation beam path. The depth of the excitation field (evanescent field) is controlled by tilting the angle of incident excitation laser beam. Tilting the incident beam is accomplished by positionig the tip of the fiberoptic cable in a plane that is conjugate to the back focal plane of the TIRF objective lens. If the fiberoptic cable is moved away from the optical axis, then the angle of incidence is increased. If it is aligned with the optical axis, then a wide-field epifluorescence excitation is obtained. Therefore, it is possible to record wide-field fluorescence microscopic images, too, using a sensitive EMCCD camera­ attached to the microscope. Wide-field TIRF images may also be recorded in this setting. However, due to hard-toavoid contamintants along the optical path, interference­

452

Kellermayer

­ atterns usually arise, causing uneven illumination of the ­sample p across the field of view. By using stage-scanning, the problems arising from uneven illumination are avoided. 4. Combining AFM and fluorescence data. Figure  2 shows ­typical AFM and flurorescence images obtained with the combined AFM and TIRFM instrument. The HeLa cell culture seen in the figure was labeled with TRITC-phalloidin to visualize the actin-filament system. Figure 2a shows the height contrast image, whereas (Fig.  2b) shows the fluorescence image generated by recording fluorescence emission intensities. Data acquisition was fully synchronized by collecting the different image modalities at the same time. To simultaneously display topography and fluorescence, the three-­ dimensional-rendered AFM height image was colored with the fluorescence intensity information (Fig. 2c). The differences between the spatial resolution of AFM and fluorescence can be discerned from the image: the fluorescence data is blurred across the sharp detail of the AFM topography. 5. Focusing. Because of the low image-acquisition rate and the narrow excitation field, focusing is often difficult. Refocusing during scanning requires precautions because movement of the oil-immersion objective, which is in viscous coupling with the sample, causes noise in simultaneous AFM scanning. Because AFM images are obtained in feedback, proper images can be recorded even if the TIRF images are out of focus (Fig. 3a). Focusing in TIRFM may be obtained through the scan of an entire field of view (Fig. 3b). The upper part of Fig. 3b is out of focus, and lower part is in focus. Scanning was from the upper left corner to the lower right corner of the image. Once the often cumbersome focusing process is properly carried out, high-resolution images can be obtained in which individual fluorescently labeled stress fibers can be visualized (Fig. 3c). 6. Mapping the evanescent field with AFM. Figure 4 shows the fluorescence and force data as a function of cantilever displacement obtained in an experiment, in which a fluorescently labeled cantilever was moved down and up within the evanaescent field. The cantilever was labeled with TRITC-labeled titin. For excitation, 532 nm was used. Fluorescence emission was recorded by the rate meter as a voltage signal proportional to intensity. Fluorescence and force data were recorded simultaneously. When the cantilever tip is positioned on the glass surface, maximum fluorescence intensity is observed. Fluorescence intensity was slightly reduced with increasing compression forces (up to 15 nN) in a reversible manner, suggesting that force-driven conformational transitions in the fluorescently labeled titin may lead to fluorescence quenching. As the tip is pulled away from the surface, cantilever bending and hence force is reduced to zero. In this displacement

Combined Atomic Force Microscopy and Fluorescence Microscopy

453

regime, the cantilever tip moves vertically within the ­evanescent field, and is expected to sense an excitation field intensity which becomes exponentially reduced with the distance from the coverslip surface. Fluorescence intensity indeed decays as a function of cantilever displacement (Fig. 4 upper panel), mapping out the evanescent field along the optical axis.

Acknowledgments This work was supported by grants from the Hungarian Science Foundation (OTKA K73256), and the Hungarian National Office of Research and Technology (NANOAMI KFKT-1-2006-0021, OMFB-380/2006). I am indebted to László Grama, Attila Nagy, Árpád Karsai, Tamás Huber, Zsolt Mártonfalvi, Balázs Kiss, Ünige Murvai, Csaba Niedetzky and Roger Proksch for assisting at various stages of the instrumental development and its validation with a range of biomolecular and cellular systems. References 1. Binnig G, Quate CF, Gerber C. (1986) Atomic force microscope. Physical Review Letters 56:930–3. 2. Gorelik J, Shevchuk A, Ramalho M, et  al. (2002) Scanning surface confocal microscopy for simultaneous topographical and fluorescence imaging: application to single virus-like particle entry into a cell. Proc Natl Acad Sci USA 99:16018–23. 3. Mathur AB, Truskey GA, Reichert WM. (2000) Atomic force and total internal reflection fluorescence microscopy for the study of force transmission in endothelial cells. Biophys J 78:1725–35. 4. Nishida S, Funabashi Y, Ikai A. (2002) Combination of AFM with an objective-type total internal reflection fluorescence microscope (TIRFM) for nanomanipulation of single cells. Ultramicroscopy 91:269–74. 5. Kellermayer MSZ. (2005) Visualizing and manipulating individual protein molecules. Physiol Measurement 26:R119–R53. 6. Muller DJ. (2008) AFM: a nanotool in membrane biology. Biochemistry 47:7986–98. 7. Muller DJ, Janovjak H, Lehto T, Kuerschner L, Anderson K. (2002) Observing structure, function and assembly of single proteins by AFM. Prog Biophys Mol Biol 79:1–43. 8. Oesterhelt F, Oesterhelt D, Pfeiffer M, Engel A, Gaub HE, Muller DJ. (2000) Unfolding pathways of individual bacteriorhodopsins. Science 288:143–6.

9. Stroh C, Wang H, Bash R, et  al. (2004) Single-molecule recognition imaging ­microscopy. Proc Natl Acad Sci USA 101:12403–507. 10. Cheung CL, Hafner JH, Lieber CM. (2000) Carbon nanotube atomic force microscopy tips: direct growth by chemical vapor deposition and application to high-resolution imaging. Proc Natl Acad Sci USA 97: 3809–13. 11. Hafner JH, Cheung CL, Woolley AT, Lieber CM. (2001) Structural and functional imaging with carbon nanotube AFM probes. Prog Biophys Mol Biol 77:73–110. 12. Katz E, Willner I. (2004) Biomoleculefunctionalized carbon nanotubes: applications in nanobioelectronics. Chemphyschem 5:1084–104. 13. Woolley AT, Cheung CL, Hafner JH, Lieber CM. (2000) Structural biology with carbon nanotube AFM probes. Chem Biol 7: R193–204. 14. Evans E, Ritchie K. (1997) Dynamic strength of molecular adhesion bonds. Biophys J 72:1541–55. 15. Fernandez JM, Li H. (2004) Force-clamp spectroscopy monitors the folding trajectory of a single protein. Science 303:1674-8. 16. Kellermayer MS, Bustamante C, Granzier HL. (2003) Mechanics and structure of titin oligomers explored with atomic force microscopy. Biochim Biophys Acta 1604:105–14.

454

Kellermayer

17. Rief M, Gautel M, Oesterhelt F, Fernandez JM, Gaub HE. (1997) Reversible unfolding of individual titin immunoglobulin domains by AFM. Science 276:1109–12. 18. Rief M, Grubmuller H. (2002) Force spectroscopy of single biomolecules. Chemphyschem 3:255–61. 19. Rief M, Oesterhelt F, Gaub HE. (1997) Single molecule force spectroscopy on polysaccharides by AFM. Science. 20. Rief M, Pascual J, Saraste M, Gaub HE. (1999) Single molecule force spectroscopy of spectrin repeats: low unfolding forces in helix bundles. J Mol Biol 286:553–61. 21. Hutter JL, Bechhoefer J. (1993) Calibration of atomic-force microscope tips. Rev Sci Instrum 64:1868–73. 22. Santos NC, Castanho MA. (2004) An overview of the biophysical applications of atomic force microscopy. Biophys Chem 107:133–49. 23. Carlisle CR, Coulais C, Namboothiry M, Carroll DL, Hantgan RR, Guthold M. (2009) The mechanical properties of individual, ­electrospun fibrinogen fibers. Biomaterials 30:1205–13. 24. Borgia A, Williams PM, Clarke J. (2008) Single-molecule studies of protein folding. Annu Rev Biochem 77:101–25. 25. Marszalek PE, Li H, Oberhauser AF, Fernandez JM. (2002) Chair-boat transitions in single polysaccharide molecules observed with force-ramp AFM. Proc Natl Acad Sci USA 99:4278–83. 26. Evans E. (2001) Probing the relation between force--lifetime--and chemistry in single molecular bonds. Annu Rev Biophys Biomol Struct 30:105–28. 27. Ando T, Kodera N, Takai E, Maruyama D, Saito K, Toda A. (2001) A high-speed atomic force microscope for studying biological macromolecules. Proc Natl Acad Sci USA 98:12468–72. 28. Ando T, Uchihashi T, Kodera N, et al. (2008) High-speed AFM and nano-visualization of biomolecular processes. Pflugers Arch 456:211–25. 29. Lakowicz JR. (1983) Principles of fluorescence spectroscopy: Plenum Publishing Corporation. 30. Kapanidis AN, Ebright YW, Ebright RH. (2001) Site-specific incorporation of fluorescent probes into protein: hexahistidine-­­ tag-mediated fluorescent labeling with (Ni(2+):nitrilotriacetic Acid (n)-fluorochrome conjugates. J Am Chem Soc 123:12123–5. 31. Kapanidis AN, Weiss S. (2002) Fluorescent probes and bioconjugation chemistries for

single-molecule fluorescence analysis of biomolecules. J Chem Phys 117:10953–64. 32. Tsien RY. (1998) The green fluorescent protein. Annu Rev Biochem 67:509–44. 33. Miyawaki A, Sawano A, Kogure T. (2003) Lighting up cells: labelling proteins with fluorophores. Nat Cell Biol Suppl:S1–7. 34. Iino R, Koyama I, Kusumi A. (2001) Single molecule imaging of green fluorescent proteins in living cells: E-cadherin forms oligomers on the free cell surface. Biophys J 80:2667–77. 35. Bruchez M, Jr., Moronne M, Gin P, Weiss S, Alivisatos AP. (1998) Semiconductor nanocrystals as fluorescent biological labels. Science 281:2013–6. 36. Alivisatos AP, Gu W, Larabell C. (2005) Quantum dots as cellular probes. Annu Rev Biomed Eng 7:55–76. 37. Hotz CZ. (2005) Applications of quantum dots in biology: an overview. Methods Mol Biol 303:1–17. 38. Walling MA, Novak JA, Shepard JR. (2009) Quantum dots for live cell and in vivo imaging. Int J Mol Sci 10:441–91. 39. Deniz AA, Dahan M, Grunwell JR, et  al. (1999) Single-pair fluorescence resonance energy transfer on freely diffusing molecules: observation of Forster distance dependence and subpopulations. Proc Natl Acad Sci USA 96:3670–5. 40. Ha T, Enderle T, Ogletree DF, Chemla DS, Selvin PR, Weiss S. (1996) Probing the interaction between two single molecules: fluorescence resonance energy transfer between a single donor and a single acceptor. Proc Natl Acad Sci USA 93:6264–8. 41. Ha T, Ting AY, Liang J, et al. (1999) Singlemolecule fluorescence spectroscopy of enzyme conformational dynamics and cleavage mechanism. Proc Natl Acad Sci USA 96:893–8. 42. Dong CY, French T, So PT, Buehler C, Berland KM, Gratton E. (2003) Fluorescencelifetime imaging techniques for microscopy. Methods Cell Biol 72:431–64. 43. Esposito A, Wouters FS. (2004) Fluorescence lifetime imaging microscopy. Curr Protoc Cell Biol Chapter 4:Unit 4 14. 44. Chang CW, Sud D, Mycek MA. (2007) Fluorescence lifetime imaging microscopy. Methods Cell Biol 81:495–524. 45. Festy F, Ameer-Beg SM, Ng T, Suhling K. (2007) Imaging proteins in  vivo using fluorescence lifetime microscopy. Mol Biosyst 3:381–91. 46. Slayter EM, Slayter HS. (1992) Light and electron microscopy. Cambridge: Cambridge University Press.

Combined Atomic Force Microscopy and Fluorescence Microscopy 47. Snyder GE, Sakamoto T, Hammer JAI, Sellers JR, Selvin PR. (2004) Nanometer localization of single green fluorescent proteins: evidence that myosin V walks hand-over-hand via telemark configuration. Biophys J 87:1776–83. 48. Yildiz A, Forkey JN, McKinney SA, Ha T, Goldman YE, Selvin PR. (2003) Myosin V walks hand-over-hand: single fluorophore imaging with 1.5-nm localization. Science 300:2061–5. Epub 03 Jun 05. 49. Hein B, Willig KI, Hell SW. (2008) Stimulated emission depletion (STED) nanoscopy of a fluorescent protein-labeled organelle inside a living cell. Proc Natl Acad Sci USA 105:14271–6. 50. Gould TJ, Verkhusha VV, Hess ST. (2009) Imaging biological structures with fluorescence photoactivation localization microscopy. Nat Protoc 4:291–308. 51. Horton H, Lehenkari P, Charras G. (2003) Commbined atomic force and confocal microscopy for biological processes. Microscopy and Analysis (Euro) 86:17–9. 52. Mathur AB, Truskey GA, Reichert WM. (2000) Atomic force and total internal reflection fluorescence microscopy for the study of force transmission in endothelial cells. Biophys J 78:1725–35. 53. Micic M, Hu D, Suh YD, Newton G, Romine M, Lu HP. (2004) Correlated atomic force microscopy and fluorescence lifetime imaging of live bacterial cells. Colloids Surf B Biointerfaces 34:205–12. 54. Sarkar A, Robertson RB, Fernandez JM. (2004) Simultaneous atomic force microscope and fluorescence measurements of protein unfolding using a calibrated evanescent wave. Proc Natl Acad Sci USA 101:12882–6. Epub 2004 Aug 23. 55. Yamada T, Afrin R, Arakawa H, Ikai A. (2004) High sensitivity detection of protein molecules picked up on a probe of atomic force microscope based on the fluorescence detection by a total internal reflection fluorescence microscope. FEBS Lett 569:59–64. 56. Bennun SV, Faller R, Longo ML. (2008) Drying and rehydration of DLPC/DSPC ­symmetric and asymmetric supported lipid bilayers: a combined AFM and fluorescence microscopy study. Langmuir 24:10371–81. 57. Chiantia S, Ries J, Chwastek G, et al. (2008) Role of ceramide in membrane protein organization investigated by combined AFM and FCS. Biochim Biophys Acta 1778:1356–64. 58. Ira, Zou S, Ramirez DM, et  al. (2009) Enzymatic generation of ceramide induces membrane restructuring: Correlated AFM and fluorescence imaging of supported bilayers. J Struct Biol 168:78–89.

455

59. Oreopoulos J, Yip CM. (2009) Combinatorial microscopy for the study of protein-membrane interactions in supported lipid bilayers: Order parameter measurements by combined polarized TIRFM/AFM. J Struct Biol 168:21–36. 60. Thormann E, Simonsen AC, Nielsen LK, Mouritsen OG. (2007) Ligand-receptor interactions and membrane structure investigated by AFM and time-resolved fluorescence microscopy. J Mol Recognit 20:554–60. 61. Lee SY, Nakaya K, Hayashi T, Hara M. (2009) Quantitative study of the gold-enhanced fluorescence of CdSe/ZnS nanocrystals as a function of distance using an AFM probe. Phys Chem Chem Phys 11:4403–9. 62. Touryan LA, Baneyx G, Vogel V. (2009) Exploiting fluorescence resonance energy transfer to probe structural changes in a macromolecule during adsorption and incorporation into a growing biomineral crystal. Colloids Surf B Biointerfaces. 63. Choucair A, Chakrapani M, Chakravarthy B, Katsaras J, Johnston LJ. (2007) Preferential accumulation of Abeta(1-42) on gel phase domains of lipid bilayers: an AFM and fluorescence study. Biochim Biophys Acta 1768:146–54. 64. Vestergaard M, Hamada T, Saito M, et  al. (2008) Detection of Alzheimer’s amyloid beta aggregation by capturing molecular trails of individual assemblies. Biochem Biophys Res Commun 377:725–8. 65. Brockerhoff SE, Davis TN. (1992) Calmodulin concentrates at regions of cell growth in Saccharomyces cerevisiae. J Cell Biol 118:619–29. 66. Brown AE, Hategan A, Safer D, Goldman YE, Discher DE. (2009) Cross-correlated TIRF/ AFM reveals asymmetric distribution of forcegenerating heads along self-assembled, ­“synthetic” myosin filaments. Biophys J 96:1952–60. 67. Peng L, Stephens BJ, Bonin K, Cubicciotti R, Guthold M. (2007) A combined atomic force/ fluorescence microscopy technique to select aptamers in a single cycle from a small pool of random oligonucleotides. Microsc Res Tech 70:372–81. 68. Di Bucchianico S, Venora G, Lucretti S, Limongi T, Palladino L, Poma A. (2008) Saponaria officinalis karyology and karyotype by means of image analyzer and atomic force microscopy. Microsc Res Tech 71:730–6. 69. Huang J, Ma L, Sundararajan S, Fei SZ, Li L. (2009) Visualization by atomic force microscopy and FISH of the 45S rDNA gaps in mitotic chromosomes of Lolium perenne. Protoplasma 236:59–65.

456

Kellermayer

70. Ebenstein Y, Gassman N, Kim S, Weiss S. (2009) Combining atomic force and fluorescence microscopy for analysis of quantum-dot labeled protein-DNA complexes. J Mol Recognit 22:397–402. 71. Barfoot RJ, Sheikh KH, Johnson BR, et  al. (2008) Minimal F-actin cytoskeletal system for planar supported phospholipid bilayers. Langmuir 24:6827–36. 72. Jung SH, Park JY, Yoo JO, Shin I, Kim YM, Ha KS. (2009) Identification and ultrastructural imaging of photodynamic therapyinduced microfilaments by atomic force microscopy. Ultramicroscopy. 73. Santacroce M, Orsini F, Perego C, et al. (2006) Atomic force microscopy imaging of actin cortical cytoskeleton of Xenopus laevis oocyte. J Microsc 223:57-65. 74. Zhou D, Jiang X, Xu R, et al. (2008) Assessing the cytoskeletal system and its elements in C6 glioma cells and astrocytes by atomic force microscopy. Cell Mol Neurobiol 28: 895–905. 75. Grzywa EL, Lee AC, Lee GU, Suter DM. (2006) High-resolution analysis of neuronal growth cone morphology by comparative atomic force and optical microscopy. J Neurobiol 66:1529–43. 76. Mangold S, Harneit K, Rohwerder T, Claus G, Sand W. (2008) Novel combination of atomic force microscopy and epifluorescence microscopy for visualization of leaching bacteria on pyrite. Appl Environ Microbiol 74:410–5. 77. Murakoshi M, Iida K, Kumano S, Wada H. (2009) Immune atomic force microscopy of prestin-transfected CHO cells using quantum dots. Pflugers Arch 457:885–98. 78. Oh YJ, Jo W, Lim J, Park S, Kim YS, Kim Y. (2008) Micropatterning of bacteria on twodimensional lattice protein surface observed by atomic force microscopy. Ultramicroscopy 108:1124–7. 79. Birukova AA, Arce FT, Moldobaeva N, et al. (2009) Endothelial permeability is controlled by spatially defined cytoskeletal mechanics: atomic force microscopy force mapping of pulmonary endothelial monolayer. Nanomedicine 5:30–41.

80. Gunning AP, Chambers S, Pin C, Man AL, Morris VJ, Nicoletti C. (2008) Mapping specific adhesive interactions on living human intestinal epithelial cells with atomic force microscopy. Faseb J 22:2331–9. 81. Kidoaki S, Matsuda T. (2007) Shapeengineered fibroblasts: cell elasticity and actin cytoskeletal features characterized by fluorescence and atomic force microscopy. J Biomed Mater Res A 81:803–10. 82. Li QS, Lee GY, Ong CN, Lim CT. (2008) AFM indentation study of breast cancer cells. Biochem Biophys Res Commun 374:609–13. 83. Riethmuller C, Schaffer TE, Kienberger F, Stracke W, Oberleithner H. (2007) Vacuolar structures can be identified by AFM elasticity mapping. Ultramicroscopy 107:895–901. 84. Silberberg YR, Pelling AE, Yakubov GE, Crum WR, Hawkes DJ, Horton MA. (2008) Mitochondrial displacements in response to nanomechanical forces. J Mol Recognit 21:30–6. 85. Eckel R, Walhorn V, Pelargus C, et al. (2007) Fluorescence-emission control of single CdSe nanocrystals using gold-modified AFM tips. Small 3:44–9. 86. Gumpp H, Stahl SW, Strackharn M, Puchner EM, Gaub HE. (2009) Ultrastable combined atomic force and total internal fluorescence microscope. Rev Sci Instrum 80:063704. 87. Kellermayer MS, Karsai A, Kengyel A, et  al. (2006) Spatially and temporally synchronized atomic force and total internal reflection fluorescence microscopy for imaging and manipulating cells and biomolecules. Biophys J 91:2665–77. 88. Trache A, Lim SM. (2009) Integrated microscopy for real-time imaging of mechanotransduction studies in live cells. J Biomed Opt 14:034024. 89. Callies C, Schon P, Liashkovich I, et al. (2009) Simultaneous mechanical stiffness and ­electrical potential measurements of living vascular endothelial cells using combined atomic force  and epifluorescence microscopy. Nanotechnology 20:175104. 90. Axelrod D. (2003) Total internal reflection fluorescence microscopy in cell biology. Methods Enzymol 361:1–33.

Chapter 28 Chemical Modifications of Atomic Force Microscopy Tips Régis Barattin and Normand Voyer Abstract Atomic force microscopy (AFM) works by scanning a very tiny tip over a surface with great precision. The microscope tips can be chemically functionalized to improve the images obtained. Well-defined chemical functionalization of AFM tips is especially important for experiments, such as chemical force microscopy and single molecule recognition force microscopy, to examine specific interactions at the single molecular level. In this chapter, we present an overview of chemical modifications of tips that have been reported to date with regards to the proper fixation of probe molecules, focusing particularly on chemical procedures developed to anchor biological molecules on AFM tips. Key words: Atomic force microscopy, Tip functionalization, Silicon and gold-coated tips, Silane, Alkanethiols, Alkenes

1. Introduction Since its invention (1), the atomic force microscope has been increasingly used to obtain images of surfaces of interest (2), and to investigate mechanical forces and molecular interactions. In atomic force microscopy (AFM), the microscope tips are chemically modified to make them sensitive to a specific molecular interaction. It is thus possible to investigate chemically sensitive imaging using chemical force microscopy (CFM; ref. 3), where AFM tip surfaces are modified with specific chemical functional groups that interact with complementary functionalities on the surface. In single molecular recognition force microscopy (SMRFM, Refs. 4–6), tips are functionalized with one or more probe molecules that can recognize a specific type of target molecule supported on the surface, enabling the detection of specific interactions at the single molecule level. This way, AFM serves to Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_28, © Springer Science+Business Media, LLC 2011

457

458

Barattin and Voyer

investigate a wide variety of single receptor-ligand binding ­phenomena. Biological ligands are thus anchored on the AFM tip and their corresponding receptors are supported on the surface, or vice versa. By measuring force-distance cycles between these functionalized tips and the surface, it is possible to estimate specific biological recognition forces. For both of these AFM applications, surfaces of AFM tips have to be chemically modified. To obtain significant and reliable results, it is mandatory to ensure a stable, strong and well-defined chemical modification of AFM tips (7, 8). In this chapter, we present chemical modifications of tips reported to date with regards to the proper fixation of probe molecules. The first part focuses on general chemical procedures allowing the functionalization of bare tips. The second part describes particular chemical procedures developed to anchor biological molecules on AFM tips.

2. General Chemical Procedures to Modify AFM Tip Surfaces

This part of the chapter reviews the main general chemical strategies to functionalize AFM tips. Particularly, we focus on the chemical modifications of bare or gold-coated silicon tips, as they are the most widely used tips for conventional AFM experiments. Commercially available AFM tips consist of microfabricated pyramids of silicon (Si) or silicon nitride (Si3N4) that are naturally oxidized and thus covered with a silicon oxide layer bearing silanol groups (Si–OH). In this case, chemical functionalization of tips is based on the reactivity of these surface silanol groups. The silicon oxide layer can also be removed by a treatment with hydrofluoric acid, that generates a hydrogenated silicon surface (Si–H) on which other kinds of functionalization can be done using the reactivity of the Si–H group.

2.1.1. Silanization

One of the main chemical modifications of silicon tips is silanization. It is based on the reaction of alkylchlorosilanes (or alkylalkoxysilanes) with surface silanol groups of silicon tips, to form organosilane monolayers (Fig. 1).

SiO2

Si / Si3N4

2.1. Chemical Functionalization of Silicon Tips

OH OH OH

R

Si

X

RR R = Cl, OCH3, OCH2CH3

O Si O O Si O O Si

X X X

Fig. 1. General scheme for the silanization process.

459

Chemical Modifications of Atomic Force Microscopy Tips H

Si / Si3N4

OH H2SO4 / H2O2 OH

OH OH

Cl

Si

Cl Cl Step 1

X

OH Cl

Si

OH Cl Cl OH

H H H

- HCl Step 2

O

H

O Si

O

X

O O

X H

Step 3

O Si

O H

∆

X

O

O Si O O Si O O Si

X X X

H

Fig. 2. Mechanism of the silanization reaction (10).

The mechanism of this grafting method using alkylchlorosilanes involves four steps (Fig. 2; Refs. 9, 10): 1. The oxide layer is activated to generate more silanol groups. 2. The silane is physisorbed onto the hydroxylated silica surface. 3. The trichloro groups are hydrolyzed so the molecule forms hydrogen bonds with the surface silanol groups and with close neighbor molecules. 4. Water is eliminated to create polysiloxane monolayers. This last step is also called reticulation. Prior to the reaction, silicon tip surfaces have to be activated to generate more surface hydroxyl groups and to remove organic contaminants such as hydrocarbons. This activation is generally done by washing the tip with Piranha solution (H2SO4/30% H2O2: 7/3) or by treatment with ozone followed by washing with alkaline and acid solutions. After activation, silicon tips are immersed in a silane solution for a few hours to allow chemisorption of silane compounds (Fig. 2, Steps 1 and 2). Tips are then heated to complete the formation of organosilane monolayers on the surface (Fig. 2, Step 3). For examples of functionalization of silicon and silicon nitride tips with organosilanes, see refs. 11, 12. 2.1.1.1. Materials and Method

1. Pretreatment of Si/Si3N4 tips with ozone. (a) Wash the tip with ethanol and water; dry it in an oven for 10 min at 120°C. (b) Treat the tip with ozone (100  mg/h; supplied by an ozone generator fed with oxygen) for 30 min. (c) Immerse the tip successively in 0.5 M NaOH for 20 min, 0.1 M HCl for 20 min and 0.5 M NaOH for 10 min. (d) Wash the tip with 0.1 M HCl and water; dry it in an oven for 10 min at 120°C. 2. Pretreatment of Si/Si3N4 tips with Piranha solution (alternative to ozone treatment). (a) Immerse the tip in Piranha solution (7:3 concentrated H2SO4: 30% H2O2) for a few minutes (see Note 1). (b) Wash the tip several times with distilled and deionized water; dry it in an oven for 10 min at 120°C.

460

Barattin and Voyer

3. Silanization of Si/Si3N4 tips. (a) After preteatment, immerse the tip in a 1–5 mM silane ­solution in toluene for several hours (1–24 h depending on the concentration of the silane used) at room temperature (see Note 2). (b) Wash the tip with toluene to remove unreacted silane materials. (c) Heat the tip in an oven for 30 min at 120°C to complete the monolayer formation. (d) Wash the tip with ethanol and water; dry it in an oven for 10 min at 120°C. 2.1.1.2. Notes

1. Treatment with Piranha solution: Caution! Piranha solution may be prepared by adding the peroxide to the acid. Mixing the solution is highly exothermic. Piranha solution reacts ­violently with organic compounds and should be handled with great care, and using appropriate personal protective measures. Waste should not be stored in closed containers and should be disposed of by approved procedures. 2. Silanization: Alkylchlorosilanes and alkylalkoxysilanes are very sensitive to moisture and may be hydrolyzed before they are physisorbed on the silicon surface. Therefore, a solution of the corresponding silane has to be prepared in dehydrated solvent (toluene in most cases). An alternative chemical modification of silicon tips is the functionalization with alcohol (usually ethanolamine), which reacts with silanol groups to form an amino-terminated monolayer (Fig. 3). This grafting method has the advantage to be performed at very low concentrations, allowing the surface density of ligands on the tip to be adjusted to a sufficiently dilute value. For example, by treating a silicon tip with a solution of ethanolamine at 0.55  g/mL, the surface density of primary amine on the tip is determined to be around 700 groups/mm2. For examples of functionalization of AFM tips using etherification with ethanolamine, see refs. 13–15.

SiO2

Si / Si3N4

2.1.2. Etherification

OH HO OH

NH2 DMSO, ∆

OH

O NH2 O O

NH2 NH2

Fig. 3. Silicon surface etherification with ethanolamine.

Chemical Modifications of Atomic Force Microscopy Tips 2.1.2.1. Materials and Method

461

1. Prepare a solution of ethanolamine hydrochloride in dry DMSO by heating at 100°C. 2. Add molecular sieve beads (pore size 0.3 nm) to the solution (see Note 1). 3. Wash the silicon tip with chloroform; dry it under a nitrogen flow (see Note 2). 4. Incubate the tip overnight in the ethanolamine-DMSO solution over molecular sieve beads. 5. Rinse the tip four times in DMSO (75°C) and twice in ethanol (room temperature). 6. Dry the tip under a nitrogen flow.

2.1.2.2. Notes

1. The competing reaction to the etherification between silanol groups and ethanolamine is the hydrolysis of the silanol groups. For this reason, molecular sieve beads are used to extract all the water from the ethanolamine-DMSO solution. 2. To obtain a high density surface of primary amines, the tip can be treated with ozone or with Piranha solution in order to generate more silanol groups at the surface (see Subheading 2.1.1).

SiO2

HF 1% Step 1

Si / Si3N4

Hydrosilylation is another approach to modifying silicon tips that is based on the reaction of 1-alkenes (and 1-alkynes) with ­hydrogencovered silicon surfaces to form strong Si–C bonds (Fig. 4, Refs. 16–18). The silicon surface must first be etched by cleaning with Piranha solution to remove organic contaminants. Then the oxide layer is removed with hydrofluoric acid (HF) to generate a hydrogen silicon surface (Fig. 4, Step 1). Reaction of the unsaturated bond with the resulting silicon-hydride group can then be achieved by a thermally-induced hydrosilylation performed by immersion of the etched silicon surface into refluxing solution of the appropriate 1-alkene (Fig. 4, Step 2). This grafting method allows the formation of dense and well-ordered alkane monolayers via very stable covalent Si–C bonds. Applying this strategy, Cai et  al. modified silicon AFM tips with an oligo ethylene glycol (OEG) derivative to obtain an

Si / Si3N4

2.1.3. Hydrosilylation

H

R

H H

∆

R

H

Step 2

R R R

Fig. 4. General scheme for hydrosilylation process of silicon tips.

Barattin and Voyer

Si / Si3N4

462

H H H H

8

O

O

8

3

∆

8

8

O

O 3

O 3

O

O O

3

Fig. 5. Formation of an OEG-terminated monolayer on a silicon tip using hydrosilylation and Si–C bond formation (19).

OEG-terminated monolayer attached to the tip surface, in order to decrease nonspecific interactions of the tip surface with proteins (Fig. 5, ref. 19). 2.1.3.1. Materials and Method

1. Immerse the tip in Piranha solution (7:3 concentrated H2SO4: 30% H2O2) for a few minutes (see Subheading 2.1.1, Note 1). 2. Wash the tip several times with distilled and deionized water. 3. Immerse the tip in ~2% HF solution for 1 min (see Notes 1 and 2). 4. Rapidly rinse the tip with distilled and deionized water; dry it immediately with a stream of dry nitrogen (see Note 2). 5. Right after the etching process, immerse the tip in a degassed solution of the appropriate 1-alkene in freshly distilled mesitylene under nitrogen. Let the solution gently reflux at 180°C for 2 h (see Notes 3 and 4). 6. Allow the solution to cool to room temperature and wash the tip successively with petroleum ether, ethanol and dichloromethane. Dry the tip under a nitrogen flow.

2.1.3.2. Notes

1. HF is exceedingly dangerous even at very low concentrations and should be handled with great care in a well-ventilated hood with appropriate personal protective measures. 2. HF etching: Hydrogen-terminated silicon surfaces, which are obtained after the HF etching process, are very sensitive to oxygen and oxidize to form an inert silicon oxide layer. Precautions must be taken to limit oxygen traces. Thus, it is highly advisable to perform the HF etching in a nitrogen blow-dried glove-box. As well, the distilled and deionized water that is used for washing the tip after HF treatment must be degassed just before use. 3. Grafting solution: For hydrosilylation, it is necessary to use an apolar hydrophobic solvent to improve the reaction. The use of dry freshly distilled mesitylene is advised to solubilize the alkene. Depending on the alkene solubility, dry nitrobenzene could also be used.

463

Chemical Modifications of Atomic Force Microscopy Tips

4. Hydrosilylation of 1-alkene on the Si–H surface: To avoid traces of oxygen during the grafting process, the solvent has to be freshly distilled and the solution of 1-alkene must be thoroughly degassed and maintained under a nitrogen atmosphere during the whole process. 2.2. Chemical Functionalization on Gold-Coated Silicon Tips

The most popular approach to modifying AFM tips is through the immobilization of thiol-based monolayers on gold-coated tips. This strategy requires a preliminary coat of gold on the silicon AFM tips.

2.2.1. Gold Coating

Coating silicon tips with gold is a two step process (Fig. 6): a thin chromium adhesive layer is first deposited (<5 nm), followed by a vapor deposition of a thicker gold layer (40–100 nm, or less). The initial chromium layer is essential to ensure proper adhesion of the gold surface to the silicon or silicon nitride surface. Metal coatings may be performed using several vapor deposition processes (chemical vapor deposition (CVD), physical vapor deposition (PVD), including e-beam evaporation, sputtering, etc.) that is not detailed in this chapter. For more information on coating procedures, see ref. 20. Gold-coated silicon AFM tips are also commercially available.

2.2.2. Formation of Self-Assembled Monolayers and Mixed SAMs with Thiols

Alkane thiols can be used for the functionalization of gold-coated AFM tips (21, 22). Gold has a strong specific interaction with sulfur that allows the formation of stable and densely-packed self-­assembled monolayers (SAMs) (Fig. 7). Thiol-based monolayers are ­prepared by immersing cleaned gold-coated tips into a thiol solution for several hours. For examples of functionalization of gold-coated AFM tips with alkanethiol monolayers, see refs. 23–25. Au

Si / Si3N4

1) Cr deposition 2) Au deposition

Si / Si3N4

Cr

Fig. 6. General scheme for gold coating.

b

R

S

R

S

Au

Si / Si3N4

a

HS

R

S S S S S S

R R R R R R

R

S

R

S

R

R S

S

S S

Fig. 7. Formation of self-assembled monolayers of alkanethiols on gold using (a) thiols or (b) disulfide compounds that show a similar reactivity toward gold surfaces (26).

Barattin and Voyer

Si / Si3N 4 Au

464

HS

99% alkanethiol 1%

HS ligand of interest

S S S S S S

Fig. 8. Schematic representation of mixed SAMs preparation.

Examples are also found in the literature where biological molecules of interest are modified to incorporate thiols in order to obtain functionalized monolayers on the tip with biomolecules. The preparation of SAMs of neoglycoconjugates (27) or of peptides (28) on gold-coated tips has been achieved this way. By using two different alkanethiols, mixed SAMs can be prepared on gold-coated tips (Fig. 8). This strategy can be very useful to decrease the density of a specific ligand at the surface of the tip and, consequently, would facilitate the detection of single interactions and prevent unwanted steric hindrance in the recognition process. In this case, a low ratio of ligands vs. alkanethiols has to be used (generally 1:99) to sufficiently dilute the ligand within the matrix formed by the alkanethiols. Applying this method, AFM tips have been functionalized with mixed SAMs of DNA-thiols (29), of proteins (30) to study interactions with cells, or of antibodies (31) for studying complementary antigen interactions. 2.2.2.1. Materials and Method

1. Clean the gold-coated tip in a Piranha etching solution (7:3 concentrated H2SO4:30% H2O2) for a few minutes (see Subheading 2.1.1, Note 1). 2. Wash the tip several times with distilled and deionized water and with ethanol; dry it under a nitrogen flow. 3. Immerse the tip in a 1 mM thiol solution in ethanol for several hours (12–24  h depending on the thiol used) at room temperature (see Notes 1 and 2). 4. Rinse the tip with ethanol; dry it under a nitrogen flow.

2.2.2.2. Notes

1. Depending on the solubility of the thiol used, a mix of ethanol and chloroform may also be used. 2. For the formation of a mixed SAM, prepare a 1 mM solution in ethanol with the desired ratio of thiols (usually from 1:99 to 1:20 depending on the wanted density of ligand).

Chemical Modifications of Atomic Force Microscopy Tips

3. Particular Chemical Procedures to Anchor Biological Probes on AFM Tips

465

In this section, we describe multi-step coupling procedures that are used to anchor biological probes on AFM tips. Two approaches are exploited. The modification can be done either directly on the silicon surface, or after a prior functionalization. In the latter approach, the tip is first functionalized following one of the general chemical procedures described above to give a reactive surface. In a second step, this reactive functionality of the surface is used to couple the ligand of interest. Bovine serum albumin (BSA) has been shown to irreversibly bind to silicon and silicon nitride surfaces (32). Using this property, a three step procedure has been developed to modify AFM tips surfaces with more complex biological architectures (Fig. 9):

3.1. BSA–Biotin Coupling Method

1. The first step involves adsorbing biotin-labeled BSA on the microscope tip, thus creating a biotin-terminated surface. 2. In the second step, the biotinylated surface allows the fixation of avidin on the tip as avidin (or streptavidin) binds strongly to that ligand. During this step, it is likely that the tetrameric avidin molecule is attached to the tip via two biotin binding sites. Consequently, by incubating the biotin-terminated tip in an avidin solution, an avidin surface with free biotin binding sites is obtained (33).

BSA

Biot-BSA

Si / Si3N4

Si / Si3N4

3. In the third step, the free biotin binding sites are then used to anchor biotin-labeled biological molecules of interest, for example biotin-labeled oligonucleotides (34), antibodies (35), or proteins (e.g. Concanavalin A, which is used to attach cells together) (36).

Avidin

Biotin-labelled ligand

Biotin-labelled antibody

Biotin Avidin

=

Biotin-labelled oligonucleotide

BSA Biotin-labelled Concanavalin A

Fig. 9. Modification of AFM tips by a three step coupling procedure using biotin-avidin binding properties.

466

Barattin and Voyer

3.1.1. Materials and Method

1. Clean the tip by immersion in acetone for 5 min and irradiate the tip with ultraviolet light for 30 min. 2. Prepare a 1 mg/mL biotinylated bovin serum albumin (BSA) solution in phosphate buffer saline (PBS; 20  mM PO43−, 150 mM NaCl, pH 7.2). 3. Incubate the tip in 50 mL of the biotin-BSA solution for 12 h at 37°C. 4. Rinse the tip six times with phosphate buffer saline (PBS; 20 mM PO43−, 150 mM NaCl, pH 7.2) (see Note 1). 5. Prepare a 1  mg/mL avidin (or streptavidin) solution in ­phosphate buffer saline (PBS; 20 mM PO43−, 150 mM NaCl, pH 7.2). 6. Incubate the biotinylated tip in 50 mL of the avidin solution for 5 min at room temperature. 7. Rinse the tip with PBS to remove unbound avidin molecules (see Note 2). 8. Incubate the avidin-labeled tip for a few minutes in a 1 mg/mL solution of the biotin-labeled biological molecule of interest.

3.1.2. Notes

1. After incubation of the tip in the biotin-BSA solution, the functionalized intermediate tip may be stored at 4°C until needed. 2. After incubation with avidin (or streptavidin), functionalized tips have to be immediately used for the next step.

3.2. Coupling Method on AzideFunctionalized Tips

A new method of functionalization of AFM tips has recently been described (25). It is a simple two step procedure based on the formation of an azide-terminated surface on the tip (Fig.  10). This resultant azide group is then used to anchor a biological molecule of interest previously modified with an alkyne group, using a click reaction between the azide and alkyne moieties. This cycloaddition reaction developed by Sharpless (37) provides stereospecifically stable triazoles in high yields and is compatible with a variety of solvents and functional groups.

3.2.1. Formation of the Azide Surface

As described by Chen et al. (25), an azide-modified surface may be obtained by using an azide-teminated thiol on a gold-coated tip (Fig. 10, Step 1 – way A) according to the formation of SAMs on gold tips as described in Subheading  2.2.2. In this specific case, a polyethylene glycol (PEG) derivative modified with a thiol at one end and an azide group at the other was used. Alternatively, commercial silanes bearing azido groups are available and can be used to functionalize silicon and silicon nitride AFM tips (Fig. 10, Step 1, way B), according to the procedure described in Subheading  2.1.1, in order to form azideterminated surfaces on the tip.

467

Chemical Modifications of Atomic Force Microscopy Tips

The azide-terminated surface is then employed in a Cu(I)-catalyzed 1,3-dipolar cyclization with an alkyne moiety of an appropriate biomolecule (Fig. 10, Step 2). For this coupling method, the biological molecule of interest has to be previously modified to insert the alkyne functional group essential to the click reaction (38). Chen et al. used this method to modify AFM tip surfaces with anti-ricin antibodies to detect trace amounts of ricin (Fig.  11, ref. 25) (a poison).

3.2.2. Click Chemistry Coupling Reaction

1. Prepare SAMs using appropriately modified OEG bearing a thiol and an azido group according to procedure described in Subheading 2.2.2.

n Way A

N3

S

n

S

Au

Si / Si3N4

N3

HS

Au

Si / Si3N4

3.2.2.1. Materials and Method

n

S

n

S

N3 N3 N3

n

SiO2

Si / Si3N4

= OH

MeO MeO

OH

Si

N3

O Si O O Si

n

OMe Way B

O O Si

OH

Step 1: Formation of the azide surface

n

N3

n

N3

n

N3

N3

N3

N3

N

N3

N N3

N

Step 2: Click chemistry coupling reaction

Fig. 10. General scheme of click chemistry coupling on azide-terminated tip surfaces.

Fig.  11. Schematic representation of AFM tip modification using the two step click chemistry procedure. [Reprinted with the permission of the authors from ref. 25. Copyright 2005 American Chemical Society].

468

Barattin and Voyer

2. Prepare a 0.1  mg/mL solution of the alkyne-modified biomolecule in water. 3. Place the tip in 200 mL of the alkyne-modified biomolecule solution and add sodium ascorbate (3 mL; 1 M) and copper(II) sulfate (1 mL; 0.3 M) (see Note 1). 4. Let the click reaction take place for 10–12 h in the dark at 0°C (see Note 2). 5. Wash the tip with buffer and store it in buffer at 4°C until needed. 1. The click chemistry reaction is catalyzed by copper(I). Sodium ascorbate is essential to reduce copper(II) sulfate to copper(I). Instead of these reactants, copper(I) iodide may also be used alone.

3.2.2.2. Notes

2. Copper(I) is highly light sensitive and must to be protected from light to avoid degradation. 3.3. Coupling Methods Using the Amino Functionality

In this section, coupling methods using amino functionalities are described. They usually exploit the reactivity of an intermediate amine-terminated tip surface.

3.3.1. Formation of an Amino Surface

By using the general chemical procedures described in Subheading 2, several possibilities are offered to produce amineterminated surfaces (Fig. 12; Refs. 39, 40). For example, exploiting the reactivity of silanol groups on untreated silicon tips covered with a layer of silicon oxide, ­silanization with 3-aminopropyltriethoxysilane APTES (Fig. 12a; ref. 41) or etherification with ethanolamine (Fig.  12b; see Subheading  2.1.2) can be used efficiently to obtain amino-­ terminated monolayers. With hydrogen-terminated silicon tips obtained after HF treatment, hydrosilylation with an amino-alkene may also be

b O Si O O Si

NH2 APTES

O O Si

OH

NH2

H N

H H H H

H N

CH3

n

O

NH2

H O CH3 N

NH2 n

O

HO

O

NH2

NH2 O O

NH2

d n

n

OH

NH2

CH3

n

OH

OH

NH2 NH2

c Si / Si3N4

SiO2

Si

EtO OEt

Au

OH

Si / Si3N4

SiO2

Si / Si3N4

OH

EtO

Si / Si3N4

a

HS

NH2 n

S

NH2 n

S

n

S

n

S

n

NH2 NH2 NH2

Fig. 12. Formation of amine-terminated surfaces: (a) by silanization with 3-aminopropyltriethoxysilane (APTES); (b) by etherification with ethanolamine; (c) by hydrosilylation of protected-amino-alkene; (d) by chemisorption of aminoalkanethiols on gold-coated surfaces.

469

Chemical Modifications of Atomic Force Microscopy Tips

applied (Fig.  12c). In this case, the amino group has to be ­previously protected with an acetate functional group (or other protecting group) to avoid the reaction of the amine with the hydrogen-terminated silicon surface. After grafting, the acetate is deprotected to generate the reactive amino surface (42). Commercial amino-functionalized thiols (such as 11-aminoundecanethiol) are available to perform chemisorption on gold surfaces to yield amino-terminated self-assembled monolayers (Fig. 12d). 3.3.1.1. Materials and Method

1. Silanization with APTES (see Subheading 2.1.1). (a) Immerse the tip in a 20 mM freshly distilled APTES solution in toluene for 12 h at room temperature. (b) Wash the tip with toluene to remove unreacted silane materials. (c) Heat the tip in an oven for 10 min at 100°C to complete the monolayer formation. (d) Wash the tip in ethanol and water; dry it in an oven for 10 min at 120°C. 2. Etherification with ethanolamine (see Subheading 2.1.2). 3. Hydrosilylation with amino-alkenes (see Subheading 2.1.3). (a) Immediately after the HF etching process, immerse the tip in a degassed solution of N-acetylamino-alk-1-ene in freshly distilled mesitylene under nitrogen. Let the solution gently reflux (180°C) for 2 h. (b) Allow the solution to cool to room temperature and wash the tip successively with petroleum ether, ethanol and dichloromethane. (c) To deprotect the acetate functional group and generate amino surfaces, immerse the tip in a 1 N HCl solution at 50°C for 16 h. (d) Wash the tip successively with water and dichloromethane; dry the tip under a nitrogen flow. 4. Chemisorption of Subheading 2.2).

3.3.2. Amine–Succinimide Coupling Method

thiols

on

gold-coated

tips

(see

Biologically relevant probes can be anchored on amino-­terminated tips using the coupling reaction between amine and N-hydroxysuccinimide (NHS) ester functional groups, which forms stable amide bonds (Fig. 13). This strategy implies that the ligand of interest has to be previously modified to introduce the NHS function. If the probe contains an acidic functional group, introduction of the NHS function can be easily done following a classical esterification reaction between the carboxylic acid and NHS.

Barattin and Voyer

This procedure has been recently reported by Riener et al. to couple a biotinylated PEG derivative to an amino-modified tip via its NHS ester end in order to detect biotin-avidin interactions (Fig. 14, ref. 40). NHS

O N

NH2 NH2 NH2 NH2

O

O

AFM tip

O

AFM tip

470

NH2 O NH NH2

Fig. 13. General scheme for the amine–succinimide coupling reaction.

Fig. 14. Schematic representation of the coupling of a biotinylated polyethylene glycol derivative to an amino surface via an NHS group. [Reprinted with the permission of the authors from ref. 40. Copyright 2003 Elsevier].

471

Chemical Modifications of Atomic Force Microscopy Tips

H

H

NH2

NH2

Glutaraldehyde

NH2

O H

N NH2

H2N

NH2

AFM tip

AFM tip

NH2

O

AFM tip

O

NH2

N N NH2

Fig. 15. Coupling of amino-containing biomolecule onto an amine-terminated tip surface via glutaraldehyde. 3.3.2.1. Materials and Method

1. Prepare a 1 mg/mL solution of the NHS-ligand in chloroform containing 0.5% (v/v) of triethylamine. 2. Incubate the amino-modified tip in the NHS-ligand solution for 2 h at room temperature. 3. Rinse the tip several times in chloroform and dry it under a nitrogen flow.

3.3.3. Amine–Aldehyde Coupling Method

In this coupling strategy developed by Weetall et  al. (43), the amino-terminated tip surface is activated with glutaraldehyde to allow the formation of a highly reactive aldehyde surface by the formation of an imine bond (Fig. 15). This aldehyde functionality is then exploited to anchor proteins on the surface through the formation of a second imine bond between an amino group at the surface of the protein and the terminal aldehyde group of the silicon-modified tip. Using this methodology, AFM tips have been functionalized with antibodies (44), ferritin (45) or enzymes (46).

3.3.3.1. Materials and Method

1. Prepare a 10% (v/v) glutaraldehyde solution in 100  mM phosphate buffer (pH 7). 2. Immerse the amino-terminated tip in the glutaraldehyde solution for 45 min. 3. Rinse the tip thoroughly with distilled water. 4. Incubate the activated tip in the appropriate solution of protein to complete the fixation of the biological probe of interest (see Note 1).

3.3.3.2. Notes

1. After incubation with the biological probe, the unreacted glutaraldehyde can be blocked via immersion of the tip in a 100  mM ethanolamine solution (pH 8) for 1  h at room temperature. The amino/aldehyde coupling method has been recently used in a different multi-step strategy to immobilize proteins (Fig. 16). Hahn et al. reported the functionalization of a goldcoated tip with a disulfide compound bearing an aldehyde functional group, allowing the formation of an aldehyde-terminated surface. Then the aldehyde’s reactivity is exploited to couple proteins via their amino functionalities (26).

Barattin and Voyer

H N

S S

O

H O

H2N

Gold-coated AFM tip

Gold-coated AFM tip

472

H N

S O

S

N

Fig. 16. Formation of an aldehyde-terminated tip surface to bind proteins via their surface amino groups (26).

"NHS" NH2

O

1

O O N O

HN

O

PEG

NH2

HN

O

O

H

2

goat IgG biotin avidin

mica

Fig. 17. Schematic representation of the linking of an antibody via a lysine-aldehyde imine coupling using a heterobifunctional PEG derivative. [Reprinted with the permission of the authors from ref. 47. Copyright 2007 American Chemical Society].

Also, Ebner et al. reported a two step procedure to link antibodies to an AFM tip (Fig. 17, ref. 47). The cross-linker used is a heterobifunctional PEG derivative bearing an aldehyde moiety at one end and an NHS-activated ester at the other. In this strategy, the amino-terminated tip surface reacts with the NHS ester to allow the binding of the PEG derivative leaving the aldehyde end free (the amine-NHS ester reaction is faster than the amine–aldehyde reaction). An antibody can then be stably bound to the aldehyde end of the PEG derivative through the formation of an imine link, using one of its lysine residues.

Chemical Modifications of Atomic Force Microscopy Tips

473

3.4. Coupling Methods Using the Thiol Functionality

Along the same lines, thiol reactivity can be exploited to couple biological ligands onto AFM tips. Indeed, the thiol group has been reacted with maleimide, disulfide or vinylsulfone groups to create stable covalent links.

3.4.1. Formation of Thiol-Modified Surfaces

A prerequisite in this strategy is to prepare a thiol-modified tip surface. Different approaches have been used for that purpose. Kienberger et al. described an interesting method to obtain such a surface on a silicon tip (Fig. 18a; ref. 48). The three step procedure begins with the prior functionalization of the tip surface with ethanolamine (according to procedure described in Subheading  2.1.2), allowing the formation of an amino-end ­surface. In the second step, the resulting amines react with the NHS ester of N-succinimidyl-3-(S-acetylthio)proprionate SATP (according to the procedure described in Subheading  3.3.2). Subsequent deprotection of the acetyl groups with hydroxylamine (NH2OH) yields the thiol-terminated tip surface. In a simpler approach using gold-coated tips, chemisorption of alkane-dithiol, such as 1,8-octanedithiol, allows the formation of a thiol-end monolayer (Fig. 18b; ref. 30). 1. Coupling reaction between amine-end surface N-succinimidyl-3-(S-acetylthio)proprionate (SATP).

3.4.1.1. Materials and Method

and

(a) Prepare a 1  mg/mL solution of N-succinimidyl-3(S-acetylthio)proprionate in chloroform containing 0.5% (v/v) of triethylamine. (b) Incubate the amino-modified tip in the SATP solution for 2 h at room temperature. (c) Rinse the tip several times in chloroform.

O

AFM tip

NH2 NH2 NH2

O

S

AFM tip

N

NH2

O

O SATP

O

S

NH

O

O

S

NH

O

O

NH2OH

AFM tip

O

a

SH

NH O

SH

NH

HS

6

SH

Gold-coated AFM tip

Gold-coated AFM tip

b S

6

SH

S

6

SH

S

6

SH

S

6

SH

Fig. 18. Formation of thiol-terminated surfaces: (a) by reaction of SATP on amino-end surface (48); (b) by chemisorption of alkane-dithiols on gold-coated surfaces (30).

474

Barattin and Voyer

2. Deprotection of S-acetyl-terminated tips with hydroxylamine. (a) Prepare a solution of hydroxylamine NH2OH (500 mM NH2OH·HCl/25  mM EDTA pH 7.5) by dissolving 1.74 g hydroxylamine hydrochloride, 0.355 g anhydrous Na2HPO4 and 0.466  g EDTANa2 · 2H2O in 40  mL of water, adjusting the pH to 7.5 with NaOH. (b) Incubate the S-acetyl-terminated tip in the hydroxylamine­ solution for 1 h. (c) Wash the tip with PBS buffer (150  mM NaCl/5  mM Na2HPO4, pH 7.5). 3. Chemisorption of Subheading 2.2).

dithiols

on

gold-coated

tips

(see

3.4.2. Thiol–Maleimide Coupling Method

With a thiol-end tip surface, binding of biological molecules of interest may be achieved using the coupling reaction between the thiol and maleimide functional groups (Fig. 19). The nucleophilic thiol group reacts with the electrophilic maleimide function according to a nucleophilic conjugated addition reaction to form a stable covalent link. This strategy has been used by Wieland et al. to bind proteins to gold-coated tips, following the procedure illustrated in Fig. 20 ref. (30). The gold-coated tip is first functionalized with mixed SAMs of alcohol-end and thiol-end alkanethiols as described above. The resulting thiol surface groups react with the maleimide function of the heterobifunctional PEG spacer, whereas the free NHS activated esters of the spacer are exploited to bind proteins through their free amino groups at their surface.

3.4.2.1. Materials and Method

1. Coupling reaction between thiol-end surface and maleimide. (a) Prepare a 1  mg/mL solution of the maleimide-tagged molecule in PBS buffer (100  mM NaCl/50  mM Na2HPO4/1  mM EDTA, pH 7.4 adjusted by adding 1 M NaOH). (b) Incubate the tip in the maleimide solution for 20 min.

Maleimide

SH

O

AFM tip

AFM tip

SH

O SH

N

SH SH

O

N S SH

Fig. 19. General scheme for the thiol–maleimide coupling reaction.

O

Chemical Modifications of Atomic Force Microscopy Tips

475

Fig. 20. Reaction scheme for the covalent attachment of proteins on gold-coated surfaces via thiol-maleimide coupling. [Reprinted with the permission of the authors from ref. 30].

(c) Rinse the tip with PBS buffer (100 mM NaCl/50 mM Na2HPO4/1  mM EDTA, pH 7.4 adjusted by adding 1 M NaOH). 3.4.3. Thiol–Disulfide Coupling Method

Ligands of interest may also be linked to a thiol-end surface using the reactivity of the 3-(2-pyridyl)-dithiopropionyl (PDP) group. Indeed, the thiol group reacts with a PDP group to form a stable disulfide bond (Fig. 21). With this strategy, it is necessary to use the ligands of interest that have been previously derivatized with a PDP group. This step is generally performed by coupling N-succinimidyl 3-(2-pyridyl)-dithiopropionate (SPDP) with a primary amine from the ligand (49). Using this strategy, Riener et al. linked a PDP-modified PEG derivative on a thiol-terminated tip surface to anchor a protein (Fig. 22a; ref. 50). Likewise, in a variant of this methodology, a PDP-terminated surface can be prepared on the tip. Indeed, Kamruzzahan et al. used a heterobifunctional PEG derivative bearing an NHS ester group at one end that reacts with an amino-terminated tip surface. Then, the 3-(2-pyridyl)-dithiopropionyl (PDP) group at the other end is further used to anchor an SATP-labeled antibody (Fig. 22b; ref. 49).

476

Barattin and Voyer PDP

SH SH SH

SH N

S

S

AFM tip

AFM tip

SH

S S SH

Fig. 21. General scheme for the thiol-disulfide coupling reaction using PDP derivatives.

Fig. 22. Schematic representation of different functionalizations of tips using thiol-PDP coupling reactions. [Reprinted with the permission of the authors from refs. 49 and 50. Copyright 2006 American Chemical Society and Copyright 2003 Elsevier].

3.4.3.1. Materials and Method

1. Coupling reaction between a thiol-end surface and 3-(2-pyridyl)dithiopropionyl (PDP) labeled molecules. (a) Prepare a solution of a PDP-labeled molecule in PBS buffer (100 mM NaCl/50 mM Na2HPO4/1 mM EDTA, pH 7.5 adjusted by adding 1 M NaOH). (b) Immerse the thiol-terminated tip in the PDP-molecule conjugate solution for 1 h. (c) Rinse the tip several times with PBS buffer (100  mM NaCl/50 mM Na2HPO4/1 mM EDTA, pH 7.5 adjusted by adding 1 M NaOH).

Chemical Modifications of Atomic Force Microscopy Tips

477

2. Coupling reaction between a PDP-terminated surface and SATP-labeled molecules. (a) Prepare a solution containing the 50 mL of SATP-labeled molecule, 25  mL of hydroxylamine NH2OH reagent (500  mM NH2OH·HCl/25  mM EDTA pH 7.5) and 50  mL of PBS buffer (100  mM NaCl/50  mM Na2HPO4/1  mM EDTA, pH 7.5 adjusted by adding 1 M NaOH) (see Note 1). (b) Immerse the PDP-terminated tip in the SATP solution for 1 h. (c) Wash the tip with PBS buffer (150  mM NaCl/5  mM Na2HPO4, pH 7.5). 1. The hydroxylamine reagent NH2OH is prepared by dissolving 1.74 g hydroxylamine hydrochloride, 0.355 g anhydrous Na2HPO4 and 0.466 g EDTANa2·2H2O in 40 mL of water, then adjusting the pH to 7.5 with 1 M NaOH.

3.4.3.2. Notes

2. Deprotection of these S-acetyl compounds using hydroxylamine (see Subheading 3.4.1) and the coupling reaction can be done in one pot. In this way, tips are incubated in a solution containing both SATP-labeled ligand and hydroxylamine. Another strategy is to prepare a vinylsulfone-end tip surface that can react with thiol groups of biological molecules of interest. Like maleimide, vinylsulfone is an electrophilic group with which thiol reacts rapidly by a nucleophilic conjugated addition reaction to form stable covalent links (Fig. 23). Puntheeranurak et al. used this strategy to anchor glucose on a microscope tip to study sugar transport by Na+/glucose cotransporter SGLT1 (Fig.  24; ref. 51). A heterobifunctional PEG derivative bearing an NHS activated ester at one end and a vinylsulfone function at the other was used to functionalize an aminoterminated tip. A spacer was linked to the amine surface via an ester linkage providing a vinylsulfone termination on the tip. A thiol derivative of glucose was then bound by reaction with the sulfone group.

Vinylsulfon-end O

HS

S O

Fig. 23. General scheme for the thiol-vinylsulfone coupling reaction.

AFM tip

AFM tip

3.4.4. Thiol–Sulfone Coupling Method

O S O

S

478

Barattin and Voyer cantilever NH2 NH2

NH2 NH2 NH2

1

O

O O N

PEG HN

NHS-end

O O Vinylsulfon-end

O S O 2 SH

Fig. 24. Schematic representation of the functionalization of tips using the thiol-­vinylsulfone coupling reaction. [Reprinted with the permission of the authors from ref. 51. Copyright 2007 American Chemical Society].

A similar procedure has been followed by Li et al. to anchor antibodies to tips to investigate the molecular interactions of angiotensin II type 1 receptors (52). 3.4.4.1. Materials and Method

1. Coupling reaction between a vinylsulfone-terminated tip and SATP-labeled molecules. (a) Prepare a solution containing the 50 mL of SATP-labeled molecule, 25  mL of hydroxylamine NH2OH reagent (500  mM NH2OH · HCl/25  mM EDTA pH 7.5) and 50  mL of PBS buffer (100  mM NaCl/50  mM Na2HPO4/1  mM EDTA, pH 7.5 adjusted by adding 1 M NaOH). (b) Immerse the vinylsulfone-terminated tip in the SATP solution for 1 h. (c) Wash the tip with PBS buffer (150  mM NaCl/5  mM Na2HPO4, pH 7.5).

3.5. N-Nitrilotriacetic Acid–Histidine Coupling Method

This coupling procedure is based on the NTA–His tag ­complex properties. The tridentate N-nitrilotriacetic acid (NTA) forms a hexagonal complex with divalent metal ions, such as Ni2+, occupying four of the six coordination sites. The remaining two coordination/binding sites are accessible to electron donor groups such as the imidazole of histidine. This way, an NTA-terminated tip surface can be exploited to bind a histidine-tagged ligand of interest. This approach implies that the tip must be derivatized first with an NTA-terminated monolayer so this functionality can

479

Chemical Modifications of Atomic Force Microscopy Tips

efficiently bind the histidine-tagged molecule of interest. Two examples of this strategy are presented below. Riener et al. developed a heterobifunctional PEG derivative bearing a 3-(2-pyridyl)-dithiopropionyl (PDP) group at one end and an NTA functional group at the other (Fig.  22a, ref. 50). The spacer is linked to a thiol-end tip surface via the PDP function providing an NTA termination on the tip. A histidine-tagged ligand MBP-VLDLR1-3 is then tightly bound to the NTA-Ni2+ function to study human rhinovirus serotype 2 (HRV2) recognition phenomena. Au tip

S

O O

O

O O

O

O O

OH

O

O

HN

Ni

O

O

O

O O O O

O

O O

O

O

O

O

O

O

O

O

O

O

O

O

O

O

HO

HO

O O

Ni

N

N

N

H

N

N

N

O

OH

OO O O

O

O

NTA alkanethiols

OH

O

HN

O

OH OH

O

EG alkanethiols

S S S S S S S SS

O

3.5.1. Example on a Thiol-Modified Si Tip

N N

N

N

His-tagged Fv Lyso HN

O

HN

O

COOH alkanethiols + NHS/EDC H O H O HO H O HO H O

H O HO H O

HO HO HO

OH alkanethiols

Au substrate

S

S

S

S

S

S

S

S

S

S

S

S

S

S

Fig. 25. Scheme of the strategy using NTA/Histidine couple to functionalize gold-coated AFM tips and studying recognition phenomena. [Reprinted with the permission of the authors from ref. 31. Copyright 2005 American Chemical Society].

480

Barattin and Voyer

3.5.2. Example on a Gold Tip

3.5.2.1. Materials and Method

Berquand et al. recently used the NTA/Ni2+/His tag system to anchor an antibody (antilysozyme antibody Fv) to explore the binding force of this antibody with its complementary antigen (lysozyme), as illustrated in Fig. 25 (31). The gold-coated tip was functionalized with mixed SAMs from ethylene glycolalkane thiols and NTA-alkane thiols in a 95:5 ratio. (For the synthesis of NTA-alkane thiols, see refs. 53, 54). Fv antibodies were anchored to the NTA functionality via the formation of the NTA–His complex. 1. Formation of the Ni–NTA complex. (a) Incubate the NTA-functionalized tip for 5 min in a 1 mM NaOH solution and for 60  min in a 40  mM aqueous solution of NiSO4 (pH 7.2). (b) Rinse the tip with phosphate buffer saline (PBS; pH 7.0; 150 mM NaCl, 3 mM KCl, 4.3 mM Na2HPO4 · 7H2O, 1.4 mM KH2PO4). 2. Coupling with Histidine-tagged molecules. (a) Incubate the Ni-NTA-tip in a 25 mg/mL histidine-tagged biomolecule solution in PBS for 2 h. (b) Rinse the tip several times with PBS.

4. Conclusion In this chapter, the main chemical strategies to functionalize AFM tips reported to date have been described. For bare silicon tips, silanization and hydrosilylation reactions are the most widely used, while for gold-coated tips chemisorption of alkanethiols is the most popular approach. A large variety of coupling methods (including amine and thiol functionalities, and biotin/avidin or NTA/histidine couples) have been used to anchor a biological probe on a tip. With all procedures presented, a stable and strong chemical functionalization is obtained. However, for most of these strategies, precise control over the density of the target ­molecules on the tip remains challenging. Only a few procedures, such as the use of mixed SAMs or the modification of silicon tip with ethanolamine at very low concentrations, facilitate a lower density of molecules on the tip. This actual limitation could be an important drawback to investigating biological recognition phenomena at the single molecule level. Therefore, finding synthetic procedures that provide easily and reproducibly modified tips of predetermined spatial distribution is a promising and challenging research area.

Chemical Modifications of Atomic Force Microscopy Tips

References 1. Binnig, G., Quate, C. F. and Gerber, C. (1986) Atomic force microscope. Phys. Rev. Lett. 56, 930–933. 2. Hansma, H. G. and Hoh, J. H. (1994) Biomolecular imaging with the atomic force microscope. Annu. Rev. Biophys. Biomol. Struct. 23, 115–139. 3. Noy, A., Vezenov, D. V. and Lieber, C. M. (1997) Chemical force microscopy. Annu. Rev. Mater. Sci. 27, 381–421. 4. Ludwig, M., Rief, M., Schmidt, L., Li, H., Oesterhelt, F., Gautel, M. and Gaub, H. E. (1999) AFM, a tool for single-molecule experiment. Appl. Phys. A 68, 173–176. 5. Zlatanova, J., Lindsay, S. M. and Leuba, S. H. (2000) Single molecule force spectroscopy in biology using the atomic force microscope. Prog. Biophys. Mol. Biol. 74, 37–61. 6. Hugel, T. and Seitz, M. (2001) The study of molecular interactions by AFM force spectroscopy. Macromol. Rapid Commun. 22, 989–1016. 7. Barattin, R. and Voyer, N. (2008) Chemical modifications of AFM tips for the study of molecular recognition events. Chem. Commun. 1513–1532. 8. Ebner, A., Wildling, L., Zhu, R., Rankl, C., Haselgrübler, T., Hinterdorfer, P. and Gruber, H. J. (2008) Functionalization of probe tips and supports for single-molecule recognition force microscopy. Top. Curr. Chem. 285, 29–76. 9. Wasserman, S. R., Tao, Y-T. and Whitesides, G. M. (1989) Structure and reactivity of alkylsiloxane monolayers formed by reaction of alkyltrichlorosilanes on silicon substrates. Langmuir 5, 1074–1087. 10. Silberzan, P., Léger, L., Ausserré, D. and Benattar, J. J. (1991) Silanation of silica surface. A new method of constructing pure or mixed monolayers. Langmuir 7, 1647–1651. 11. Ito, T., Namba, M., Bühlmann, P. and Umezawa, Y. (1997) Modification of silicon nitride tips with trichlorosilane self-assembled monolayers (SAMs) for chemical force microscopy. Langmuir 13, 4323–4332. 12. Wenzler, L. A., Moyes, G. L., Olson, L. G., Harris, J. M. and Beebe Jr, T. P. (1991) Singlemolecule bond-rupture force analysis of interactions between AFM tips and substrates modified with organosilanes. Anal. Chem. 69, 2855–2861. 13. Hinterdorfer, P., Baumgarten, W., Gruber, H. J., Scilcher, K. and Schindler, H. (1996)

481

Detection and localization of individual antibody-antigen recognition events by atomic force microscopy. Proc. Natl. Acad. Sci. 93, 3477–3481. 14. Klein. D. C. G., Stroh, C. M., Jensenius, H., van Es, M., Karuzzahan, A. S. M., Stamouli, A., Gruber, H. J., Oosterkamp, T. H. and Hinterdorfer, P. (2003) Covalent immobilization of single proteins on mica for molecular recognition force microscopy. ChemPhysChem 4, 1367–1371. 15. Avci, R., Schweitzer, M., Boyd, R. D., Wittmeyer, J., Steele, A., Toporski, J., Beech, I., Arce, F. T., Spangler, B., Cole, K. M. and McKay, D.S. (2004) Comparison of antibodyantigen interactions on collagen measured by conventional immunological techniques and atomic force microscopy. Langmuir 20, 11053–11063. 16. Linford, M. R., Fenter, P., Eisenberger, P. M. and Chidsey, C. E. D. (1995) Alkyl monolayers on silicon prepared from 1-alkenes and hydrogen-terminated silicon. J. Am. Chem. Soc. 117, 3145–3155. 17. Sieval, A. B., Vleeming, V., Zuilhof, H. and Sudhölter, E. J. R. (1999) An improved method for the preparation of organic monolayers of 1-alkenes on hydrogen-terminated silicon surfaces. Langmuir 15, 8288–8291. 18. Buriak, J. M. (2002) Organometallic chemistry on silicon and germanium surfaces. Chem Rev. 102, 1271–1308. 19. Yam, C-M., Xiao, Z., Gu, J., Boutet, S. and Cai, C. (2003) Modification of silicon AFM cantilever tips with an oligo(ethylene glycol) derivative for resisting proteins and maintaining a small tip size for high-resolution imaging. J. Am. Chem. Soc. 125, 7498–7499. 20. Shon-Roy, L., Wiesnoski, A. and Zorich, R. (1998) Deposited films and planaration processes, in Advanced semiconductor fabrication handbook (Phillips, W., ed.) Integrated Circuit Engineering Corporation, Arizona, USA. 21. Ulman, A. (1996) Formation and structure of self-assembled monolayers. Chem. Rev. 96, 1533–1554. 22. Love, J. C., Estroff, L. A., Kriebel, J. K., Nuzzo, R. G. and Whitesides, G. M. (2005) Self-assembled monolayers of thiolates on metals as a form of nanotechnology. Chem. Rev. 105, 1103–1169. 23. Noy, A., Frisbie, C. D., Rozsnyai, L. F., Wrighton, M. S. and Lieber, C. M. (1995) Chemical force microscopy: Exploiting chemically-modified tips to quantify adhesion, friction, and functional group distributions in

482

Barattin and Voyer

molecular assemblies. J. Am. Chem. Soc. 117, 7943–7951. 24. Wang, M. S., Palmer, L. B., Schwartz, J. D. and Razatos, A. (2004) Evaluating protein attraction and adhesion to biomaterials with the atomic force microscopy. Langmuir 20, 7753–7759. 25. Chen, G., Ning, X., Park, B., Boons, G-J. and Xu, B. (2009) Simple, clickable protocole for atomic force microscopy tip modification and its application for trace ricin detection by recognition imaging. Langmuir 25, 2860–2864. 26. Hahn, C.D., Leitner, C., Weinbrenner, T., Schlapak, R., Tinazli, A., Tampé, R., Lackner, B., Steindl, C., Hinterdorfer, P., Gruber, H. J. and Hölzl, M. (2007) Self-assembled monolayers with latents aldehydes for protein immobilization. Bioconjugate Chem. 18, 247–253. 27. Tromas, C., Rojo, J., de la Fuente, J. M., Barrientos, A. G., Garcia, R. and Penadés, S. (2001) Adhesion forces between Lewis determinant antigens as measured by atomic force microscopy. Angew. Chem. Int. Ed. 40, 3052–3055. 28. Nakamura, C., Takeda, S., Kageshima, M., Ito, M., Sugimoto, N., Sekizawa, K. and Miyake, J. (2004) Mechanical force analysis of peptide interactions using atomic force microscopy. Biopolymers 76, 48–54. 29. Boozer, C., Chen, S. and Jiang, S. (2006) Controlling orientation on mixed ssDNA/ OEG SAMs. Langmuir 22, 4694–4698. 30. Wieland, J. A., Gewirth, A. A. and Leckband, D. E. (2005) Single molecule adhesion measurements reveal two homophilic neural cell adhesion molecule bonds with mechanically distinct properties. J. Biol. Chem. 280, 41037–41046. 31. Berquand, A., Xia, N., Castner, D. G., Clare, B. H., Abbott, N. L., Dupres, V., Adriaensen, Y. and Dufrêne, Y. F. (2005) Antigen binding forces of single antilysosyme Fv fragments explored by atomic force microscopy. Langmuir 21, 5517–5523. 32. Florin, E-L., Moy, V. T. and Gaub, H. E. (1994) Adhesion forces between individual ligand-receptor pairs. Science 264, 415–417. 33. Yuan, C., Chen, A., Kolb, P. and Moy, V. T. (2000) Energy landscape of streptavidin-biotin complexes measured by atomic force microscopy. Biochemistry 39, 10219–10223. 34. Florin, E-L., Lehmann, H., Ludwig, M., Dornmair, C., Moy, V. T. and Gaub, H. E. (1995) Sensing specific molecular interactions with the atomic force microscopy. Biosens. Bioelectron. 10, 895–901.

35. Moy, V. T., Florin, E-L. and Gaub, H. E. (1994) Adhesive forces between ligand and receptor measured by AFM. Colloids Surfaces A: Physicochem. Eng. Aspects 93, 343–348. 36. Li, F., Redick, S.D., Erickson, H. P. and Moy, V. T. (2003) Force measurements of the a5b1 integrin-fibronectin interaction. Biophys. J. 84, 1252–1262. 37. Kolb, H. C., Finn, M. G. and Sharpless, K. B. (2001) Click chemistry: Diverse chemical function from a few good reactions. Angew. Chem. Int. Ed. 40, 2004–2021. 38. Moses, J. E. and Moorhouse, A. D. (2007) The growing applications of click chemistry. Chem. Soc. Rev. 36, 1249–1262. 39. Ebner, A., Hinterdorfer, P. and Gruber, H. J. (2007) Comparison of different aminofunctionalization strategies for attachment of single antibodies to AFM cantilevers. Ultramicrosopy 107, 922–927. 40. Riener, C. K., Stroh, C. M., Ebner, A., Klampfl, C., Gall, A. A., Romanin, C., Lyubchenko, Y. L., Hinterdorfer, P. and Gruber, H. J. (2003) Single test system for single molecule recognition force microscopy. Anal. Chim. Acta 479, 59–75. 41. Howarter, J. A. and Youngblood, J. P. (2006) Optimization of silica silanization by 3aminopropyltriethoxysilane. Langmuir 22, 11142–11147. 42. Sieval, A. B., Linke, R., Heij, G., Meijer, G., Zuilhof, H. and Sudhölter, E. J. R. (2001) Amino-terminated organic monolayers on hydrogen-terminated silicon surfaces. Langmuir 17, 7554–7559. 43. Weetall, H. H. (1976) Covalent coupling methods for inorganic support materials. Methods Enzymol. 44, 134–148. 44. Brogan, K. L., Ho Shin, J. and Schoenfisch, M. H. (2004) Influence of surfactants and antibody immobilization strategy on reducing nonspecific protein interactions for molecular recognition force microscopy. Langmuir 20, 9729–9735. 45. Allen, S., Chen, X., Davies, J., Davies, M. C., Dawkes, A. C., Edwards, J. C., Roberts, C. J., Sefton, J., Tendler, S. J. B. and Williams, P. M. (1997) Detection of antigen-antibody binding events with the atomic force microscope. Biochemistry 36, 7457–7463. 46. Vinckier, A., Gervasoni, P., Zaugg, F., Ziegler, U., Lindner, P., Groscurth, P., Plückthun, A. and Semenza, G. (1998) Atomic force microscopy detects changes in the interaction forces between GroEL and substrate proteins. Biophys. J. 74, 3256–3263.

Chemical Modifications of Atomic Force Microscopy Tips 47. Ebner, A., Wilding, L., Kamruzzahan, A. S. M., Rankl, C., Wruss, J., Hahn, C. D., Hölzl, M., Zhu, R., Kienberger, F., Blaas, D., Hinterdorfer, P. and Gruber, H. J. (2007) A new, simple method for linking of antibodies to atomic force microscopy tips. Bioconjugate Chem. 18, 1176–1184. 48. Kienberger, F., Ebner, A., Gruber, H. J. and Hinterdorfer, P. (2006) Molecular recognition imaging and force spectroscopy of single biomolecules. Acc. Chem. Res. 39, 29–36. 49. Kamruzzahan, A. S. M., Ebner, A., Wilding, L., Kienberger, F., Riener, C. K., Hahn, C. D., Pollheimer, P. D., Winklehner, P., Hölzl, M., Lackner, B., Schörkl, D. M., Hinterdorfer, P. and Gruber, H. J. (2006) Antibody linking to atomic force microscope tips via disulfide bond formation. Bioconjugate Chem. 17, 1473–1481. 50. Riener, C.K., Kienberger, F., Hahn, C. D., Buchinger, G. M., Egwin, I. O. C., Haselgrübler, T., Ebner, A., Romanin, C., Klampfl, C., Lackner, B., Prinz, H., Blaas, D., Hinterdorfer, P. and Gruber, H. J. (2003) Heterobifunctional crosslinkers for tethering

483

single ligand molecules to scanning probes. Anal. Chim. Acta. 497, 101–114. 51. Puntheeranurak, T., Wimmer, B., Castaneda, F., Gruber, H.J., Hinterdorfer, P. and Kinne, K. H. (2007) Substrate specificity of sugar transport by rabbit SGLT1: Single-molecule atomic force microscopy versus transport studies. Biochemistry 46, 2797–2804. 52. Li, G., Xi, N. and Wang, D. H. (2005) Investigation of angiotensin II type 1 receptor by atomic force microscopy with functionalized tip. Nanomed: Nanotech, Biol & Med. 1, 306–312. 53. Luk, Y-Y., Tingey, M. L., Hall, D. J., Israel, B. A., Murphy, C.J., Bertics, P. J. and Abbott, N. L. (2003) Using liquid crystals to amplify protein-receptor interactions: Design of surfaces with nanometer-scale topography that present histidine-tagged protein receptors. Langmuir 19, 1671–1680. 54. Schmitt, L., Ludwig, M., Gaub, H. E. and Tampé, R. (2000) A metal-chelating microscopy tip as a new toolbox for single-molecule experiments by atomic force microscopy. Biophys. J. 78, 3275–3285.

wwwwwwwwwwwwww

Chapter 29 Atomic Force Microscopy as Nanorobot Ning Xi, Carmen Kar Man Fung, Ruiguo Yang, King Wai Chiu Lai, Donna H. Wang, Kristina Seiffert-Sinha, Animesh A. Sinha, Guangyong Li, and Lianqing Liu Abstract Atomic force microscopy (AFM) is a powerful and widely used imaging technique that can visualize single molecules under physiological condition at the nanometer scale. In this chapter, an AFM-based nanorobot for biological studies is introduced. Using the AFM tip as an end effector, the AFM can be modified into a nanorobot that can manipulate biological objects at the single-molecule level. By functionalizing the AFM tip with specific antibodies, the nanorobot is able to identify specific types of receptors on the cell membrane. It is similar to the fluorescent optical microscopy but with higher resolution. By locally updating the AFM image based on interaction force information and objects’ model during nanomanipulation, real-time visual feedback is obtained through the augmented reality interface. The development of the AFM-based nanorobotic system enables us to conduct in situ imaging, sensing, and manipulation simultaneously at the nanometer scale (e.g., protein and DNA levels). The AFM-based nanorobotic system offers several advantages and capabilities for studying structure–function relationships of biological specimens. As a result, many biomedical applications can be achieved by the AFMbased nanorobotic system. Key words: AFM, Augmented reality, Nanomaniplation, Nanorobot, Single-molecule recognition

1. Introduction The investigation of cellular processes with current methods is a significant challenge. New technologies are needed for investigation of biomolecule interactions. Nanomanipulation tools would be one of such novel approaches for the task and provide a novel platform for understanding the location, structure, and molecular dynamics of these molecules at the single-molecule level. This process plays an important role in intracytoplasmic sperm injection, pronuclei DNA injection, gene therapy, and other biomedical areas. Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5_29, © Springer Science+Business Media, LLC 2011

485

486

Xi et al.

1.1. Application of AFM to Biomedical Research

The development of atomic force microscopy (AFM) (1) offers new ways to investigate the dynamic changes in cellular structures, and study the functions of the single molecules of genes and proteins at single-molecule level. Although AFM was initially developed to study surface topography at the nanometer scale, it is now being increasingly used in the biological sciences (2–4). AFM offers several advantages: native biomolecules can be directly studied; it provides three-dimensional images of surface topography and quantitative measures of biological properties; and it can be performed on living tissues. In addition to the capability of AFM to characterize surfaces on a nanometer scale, it has been recently demonstrated that AFM can be employed as a nanorobot to modify surfaces and manipulate objects in nanosize by using the AFM tip as an end effector (5, 6). This nanorobot can also be used in biomedical studies by combining imaging and manipulation, allowing precise and controlled modifications and examination of biological ­systems at an unparalled level of detail. The first demonstration of AFM-based biomanipulation was performed on genetic material (7). Subsequently, several groups have demonstrated the similar application on genetic manipulation by AFM. For example, chromosomes have been dissected by controlling the applied force (8) and a combined technique using contact and noncontact mode of operations (9). More recently, AFM-based nanomanipulation has been demonstrated to extract mRNAs from living cells (10). In addition, the AFM-based nanorobot allows the isolation and manipulation of other biological objects, including single cell (11, 12) and protein molecules (13–15). All these examples illustrate that it is possible to perform a variety of molecular manipulation and investigation with AFM-based nanorobot.

1.2. Current Challenges and Difficulties in Using AFM as a Nanorobot in Biomedical Research

The technical difficulties in this strategy are extremely tough and not obvious (see Note 2). For example, the precise manipulation requires an accurate control of the end effector and the bio­ molecules. In recent years, various kinds of AFM-based nanomanipulation schemes have been developed (16, 17). The main problem with these manipulation schemes is the lack of real-time visual feedback. Each operation has to be verified by another new image scan before the next operation. Obviously, this scandesign-manipulation-scan cycle is time consuming and less accurate, which makes monitoring the dynamic interactions and biological events difficult. Recently, some researchers have been trying to combine the AFM with both haptic technique and virtual reality interface to facilitate nanomanipulation (18, 19). Although ­virtual reality that displays a static virtual environment has been ­constructed, it does not display any real-time changes in the environment during manipulation. Therefore, the operator is still blind because he/she cannot see the environmental changes

Atomic Force Microscopy as Nanorobot

487

in real time. Thus, any methods that can update the AFM image in  real time during manipulation help the operator to perform operations efficiently. Furthermore, most studies on nanomanipulation using AFM were carried out in ambient conditions and few of them are performed in liquid which is usually the environment for biological specimens. Studies of living cells in liquid using high-resolution AFM are hampered by cell deformation and tip contamination (20), which easily damage and destroy the sample during manipulation. Different approaches have been used to obtain high-­resolution images of soft biological materials. At low temperature, cells stiffen and high-resolution imaging becomes feasible (21). Cells also become stiff after chemical fixation (22). These circumstances, however, can hardly be called physiological. Another solution is to use the tapping mode AFM (TMAFM) in liquid, which gives a substantial improvement in imaging quality and stability over standard ­contact mode AFM (23). Because of the viscoelastic properties of the plasma membrane, the cell may behave like a “hard” material when responding to externally applied high frequency vibration, and it is effectively less susceptible to deformation (20). Recent progress in the spatial resolution of AFM technology has rendered topographical imaging of single protein a routine procedure (24, 25). However, it is still impossible to recognize specific proteins like receptors only from the topographical information. Because the interaction between ligands and receptors is highly specific and possesses a high degree of spatial and orientational specificity, the technique to functionalize an AFM tip with certain molecules has opened a promising way to recognize single specific molecules. It has been proven that single receptors can be recognized by an AFM tip functionalized with antibodies through a force mapping technique (26–28). However, all these results are obtained by imaging well-prepared samples on substrate surfaces rather than under physiological condition. In practice, it is still very hard to obtain a clear image at a molecular level when imaging living cells by AFM. The objective of our work differs from previous reports in that we aim to develop an AFM-based nanorobotic system with the capabilities of videolized AFM imaging and real-time haptic displaying for chemical and biological sensing, and manipulation in a nanobiological environment, and to use such robotic tool for the investigation of dynamic and complex interactions between membrane proteins at the single-molecule level. The AFM-based nanorobot comprises of a controllable nanorobotic arm, functionalizable and adaptable nano end effector, videolized real-time display system, haptic feedback device, and human command generator. Ultimately, the nanorobotic tool can augment a nanoenvironment such that the molecular level activities inside a biological system can be directly visualized, sensed, and manipulated by humans.

488

Xi et al.

2. Development of AFM-Based Nanorobots

We have developed a prototype of an AFM-based nanorobotic system. The AFM-based nanorobotic system includes two subsystems: the AFM system and the augmented reality environment (29–38). The AFM system (left side, Fig. 1) is the main part of the nanorobotic system and is designed for imaging and manipulation. The augmented reality environment (right side, Fig.  1) provides the operator a real-time interactive environment to view the real-time AFM image, directly control the tip motion, and receive the force feedback of the tip and sample interaction through a haptic joystick. The real-time visual display is a dynamic AFM image of the operating environment, which is locally updated, based on the real-time force and the local scanning information as well as the object’s behavior models. The AFM system called Bioscope system (Veeco Metrology Inc., Santa Barbara, CA) has a specially designed scanning head that can be used to image biological samples in liquid. The closedloop scanning head (Veeco Metrology Inc.) with a maximum XY scan range of 90 mm × 90 mm and a Z range of 5 mm is connected to an AFM controller and the main computer. The AFM ­controller is responsible for running the main control program and providing­ an interface for users to change control parameters and view

Fig. 1. Configuration of an AFM-based nanorobotic system.

Atomic Force Microscopy as Nanorobot

489

r­ eal-time data. A signal access module (SAM) provides an interface for external devices to access the AFM controller signal. Peripheral devices, including an optical microscope and a charge-coupled device (CCD) camera, are connected to the system. The inverted optical microscope and the CCD camera help the operator to locate the tip and adjust the AFM laser gun, and also search for interesting areas on the biological sample. To perform nanomanipulation, an actively controlled AFM cantilever (active probe) is employed as the adaptable end effector of the robotic manipulator. The active probe is not only used to sense and image the sample, but also can be used as a tool to perform manipulation (39). The augmented reality interface is implemented in a ­computer equipped with a haptic device (PhantomTM from Sensable Co.). Through the SAM, the cantilever deflection signal can be directly sent to the A/D acquisition card inside the computer. The augmented reality environment provides enhanced media for the operator to view the real-time videolized AFM image and feel the force feedback during nanomanipulation. The videolized ­real-time visual display is a continually dynamic AFM image of the operating environment. With the advantages of real-time visual and force feedback, the nanorobot has broad applications in biomedical areas, such as manipulating DNA and protein molecules, in living cells (40–45). Another broad application of the nanorobot is nanoassembly and nano­ fabrication. The AFM-based nanorobotic system has brought nano­ manipulation into real applications for nanoimprinting (29–38), fabricating, and assembling nanodevices, such as single carbon nanotube (CNT)-based nanosensors (46–50). Nanoimprints can be inscribed on a soft surface using the AFM-based nanorobotic system. Figure 2 shows the nanoimprinting using the AFM-based nanorobotic system, which reveals that the final result (Fig. 2b) matches the display in the augmented reality environment (Fig. 2a). Hence, the

Fig. 2. Nanoimprinting on a polycarbonate surface. The scanning range is 10 mm. (a) Real-time display in the augmented reality environment during nanoimprinting. (b) The new scanned AFM image after nanoimprinting.

490

Xi et al.

Fig.  3. Assembling single carbon nanotube-based nanoelectronic devices. The AFM images (a) before and (b) after nanomanipulation.

nanorobotic system is able to provide the operator an accurate visual feedback. In addition, an efficient nanomanufacturing process was developed for building a single CNT-based sensor by depositing CNTs on the substrate surface and aligning them to bridge the electrode gap using the AFM-based nanorobot, as shown in Fig. 3. 2.1. Hardware Control: Adaptable End Effector with Haptic Force Feedback

In the AFM-based robotic system, the AFM cantilever tip functions as the end effector of the robotic manipulator. The property of the AFM cantilever is important for performing imaging, sensing, and manipulation. A softer and flexible AFM cantilever is desired to achieve a sensitive imaging and sensing in pushing nano-objects. However, the tip is very easy to slip over the nano-objects due to flexibility of the cantilever and makes nanomanipulation inefficient. On the other hand, a relatively rigid cantilever may overcome the tip slipping, but it is insensitive to manipulation force. Therefore, the rigid cantilever is not suitable for generating a haptic feedback. To overcome this hurdle, an actively controlled AFM ­cantilever (active probe) is developed (39). The rigidity of the cantilever can be changed via a piezo-actuating layer and a ­controller for different operations. In the imaging mode, the ­cantilever can become very soft, and the sensing speed and sensitivity of the cantilever can be increased. In the manipulation mode, the cantilever can become more rigid such that the positioning accuracy during nanomanipulation can be improved. The active probe can also be used to improve the force sensitivity of the haptic feedback and the accuracy of nanomanipulation. Besides, the AFM tip can also be chemically or biologically functionalized for biomedical investigations. There are two methods to functionalize the AFM-based nanorobot end effector with functional agents: directly coating the agents on a silanized tip or tethering the agents on a tip using a linker. The direct coating

Atomic Force Microscopy as Nanorobot

491

Fig. 4. The process of tip functionalization with antibody via a linker.

method is simple and results in high lateral resolution. The tethering method needs a much more complicate protocol, but it results in better binding between the tip and the functional agents. This also provides a high degree of spatial and orientation specificity. The process and design to functionalize the AFM tips are illustrated in Fig. 4 (39). 2.2. Software Development: Augmented Reality Enhanced Interface

In current nanomanipulation systems, each operation is designed off-line based on the static AFM image, and then downloaded to the AFM system to implement the operation in open loop. The result of each operation is verified by a new image scan. Obviously, this scan-design-manipulation-scan cycle is very time consuming. Combining the AFM with virtual reality interface and haptic devices provides a solution to this problem. However, the operator is still blind during the manipulation because the real-time AFM image is not available during the manipulation. Therefore, a new image scan is still necessary after each operation. The augmented reality enhanced interface aims to provide the operator a real-time visual display and force feedback during the nanomanipulation. The augmented reality interface enables the operator to view the real-time AFM image and feel the force feedback during nanomanipulation. The real-time visual display is a dynamic AFM image of the operating environment which is locally updated based on the environment models, real-time force information, and the local scanning information (34). We have developed

492

Xi et al.

a videolized AFM-based nanorobotic system assisted by the ­augmented reality (29–32). By enhancing the augmented reality with a videolized AFM system and a haptic feedback system, the operator can feel the real-time three-dimensional (3D) interaction forces through the haptic system and observe the real-time changes of the nanoenvironment through the videolized AFM image simultaneously.

3. Applications of AFM-Based Nanorobot in Biomedical Research

3.1. Imaging and Cutting Living Neural Cells’ Brunches

The AFM-based nanorobot offers various important biomedical applications. The most recent and significant demonstration in biomedical studies is outlined, which includes (1) imaging and cutting living neural cells’ brunches, (2) imaging and manipulation of DNA molecules, and (3) detailed characterization and visualization of the desmosomal junctions of human epithelial cells. Imaging and manipulation of living neuron cells under liquid were performed using the AFM-based nanorobot. Living cell images were obtained in their physiologic environments using TMAFM. Under the assistance of the augmented reality system, manipulations of living neuron cells were performed at the nanoscale level (42–44). The living cell samples were prepared and grown on glass ­coverslips. The cells originated in the dorsal root ganglia (DRG) tissue of male Wistar rats (body weight, 125–200  g). The DRG tissues from the cervical, thoracic, lumbar, and sacral levels were removed aseptically and collected in F12 medium (Gibco/BRL, Gaithersburg, MD). The trimmed DRG tissues were digested in 0.25% collagenase (Boerhinger Mannheim, Indianapolis, IN) in F12 medium at 37°C for 90 min. After a 15-min incubation in PBS containing 0.25% trypsin (Gibco/BRL), the tissues were ­triturated with a pipette in F12 medium containing DNAse (Sigma-Aldrich, 80 mg/mL), trypsin inhibitor (Sigma-Aldrich, 100 mg/mL), and 10% heat-inactivated horse serum (Hyclone, Logan, UT). The cells were then seeded in a 12-well culture plate with polyornithinecoated glass coverslides inside. The cells were cultured in a humid incubator at 37°C with 5% carbon dioxide and 9% air. The cells were ready for AFM scanning after 7–10 days of culture. The glass coverslip with a monolayer of DRG cells grown on the surface was put into a Petri dish containing F12 medium. A single cell was located using the optical microscope and then moved underneath the cantilever tip by adjusting the AFM stage. The image of the living cells was obtained using the TMAFM

Atomic Force Microscopy as Nanorobot

493

Fig. 5. The AFM image of living neuron cells with scanning range of 90 mm. (a) Topographic image and (b) phase image. (c) Real-time image displayed in the augmented reality interface.

Fig. 6. Final result of the cutting operation obtained from AFM image with a scanning range of 90 mm. (a) Height image. (b) Phase image.

(Fig. 5a, b). After the AFM image of the living cells was obtained, manipulation was performed under the assistance of the ­augmented reality system. The tip can be injected into the cell or used to cut the cell membrane at certain locations. The real-time image ­displayed in augmented reality environments during manipulation is shown in Fig. 5c. The big circle in Fig. 6 is the first try at cutting a large neuron cell axon. The small circle in Fig. 6 is the second attempt to cut a small neuron cell branch. It can be seen that the

494

Xi et al.

final results of the cutting operation obtained from the AFM image are consistent with those displayed in the augmented reality environment. 3.2. Imaging and Manipulation of DNA Molecules

In addition, using the AFM-based nanorobotic system, kinks and deformation of DNA molecules can be created artificially by controlling the pushing force between the tip and the sample surface (40–44). The DNA molecules or DNA bundles can be either broken or deformed (Fig. 7). A large pushing force in the normal direction usually breaks the DNA molecule, and a small pushing force may only deform the DNA molecule without breaking it. In Fig.  7b, the big scratches on the surface indicate large pushing force applied on the AFM tip, and small scratches imply small pushing force used. The DNA bundle was broken when a big pushing force was applied, but only deformed when a small pushing force was applied. Manipulation of DNA molecules can be displayed in real time in the augmented reality environment. An example of DNA manipulation is shown in Fig. 8, in which Fig. 8a shows the DNA molecules in their original shapes, Fig. 8b shows the manipulation of DNA molecules displayed in the augmented reality ­environment, and Fig. 8c shows an AFM image after manipulation. It can be seen that several kinks have been created by slightly pushing the DNA molecules or bundles, and the kinks created in the augmented reality environment are relatively identical to the real results.

Fig. 7. (a) AFM image of DNA ropes in its original shape. (b) DNA ropes are cut by the AFM tip. The pushing force can be controlled to cut the DNA rope or only deform the DNA rope. The big scratches on the surface indicate strong pushing force applied, and small scratches imply a small pushing force. Arrows indicate the pushing directions.

Atomic Force Microscopy as Nanorobot

495

Fig. 8. Pushing DNA on a polycarbonate surface (scanning range of 3 mm). (a) Image of DNA before pushing. (b) Real-time display on the augmented reality during pushing. (c) A new scanning image after several pushing operations.

3.3. Investigation of Desmosomal Junctions of Human Epithelial Cells

Desmosomal junctions are specialized structures important for cell–cell adhesion of epithelial tissues. The development of autoantibodies to specific desmosomal proteins in certain disease states, such as Pemphigus vulgaris (PV), is associated with acantholysis or the loss of cell–cell adhesion. Desmosomal junctions are not easily accessible to existing techniques for analyses and thus, their activity in healthy and disease conditions is not well understood. We employed AFM to image the detailed 3D structure of the cell junction at high magnification (45). Additionally, we show that antibodies directed against desmoglein 3 (a major component of the desmosomal structural unit) are associated with changes at the cell surface of the human keratinocytes, supporting the hypothesis that cell structures and junctions are modified by antibody binding. This work indicates that the structure of gap junctions can be studied by AFM to help illuminate a more detailed understanding of disease mechanisms.

3.3.1. Imaging of HaCaT Cells

We used AFM to visualize desmosomes in the immortalized human keratinocyte cell line HaCaT in situ. HaCaT cells were used and prepared for AFM imaging (see Note 1). First, a low magnification AFM image was obtained (Fig. 9a) to estimate the positions of the junctions between two cells. Next, high magnification images with a scan size down to 4 mm2 (Fig. 9b) and 1 mm2 (Fig. 9c) were acquired after zooming into the areas of adjacent cells. High-resolution AFM imaging reveals intercellular structures in parallel organization bridging adjacent keratinocytes (Fig.  9d). From the high-resolution AFM image, the height of the cell junction is measured to be approximately 55 nm. Our AFM images correspond well to alternative techniques that have been used to visualize the cell–cell junctions. A low

496

Xi et al.

Fig. 9. Visualization of HaCaT cells by AFM. (a) Low magnification height imaging with scan size of 30 × 30 mm2 reveals the location of intercellular structures (boxed regions). High magnification images with a scan size of (b) 4 × 4 mm2 and (c) 1.3 × 1.3 mm2 were obtained by zooming into the highlighted regions within the imaged cells. (d) The three-­dimensional view of the 1.3 × 1.3  mm2 high magnification image. In high magnification (b, c, and d), intercellular fibers bridging ­adjacent cells with a size down to submicron are observed between the cells.

magnification­ scanning electron microscopy (SEM) image (Fig. 10a) displays the “cobblestone pattern” typical for ­keratinocyte cultures that are connected to each other. High magnification SEM (Fig.  10b) reveals further details of the intercellular junctional structures. SEM highlights the presence of intercellular fibers bridging adjacent cells as we observed by AFM. We have also performed immunofluorescence (IF) staining for cytokeratin 18 in HaCaT cells. A lacelike network of filaments is seen throughout the cells and extending to intracellular structures that bridge adjacent cells, displaying an architectural ­structure similar to that revealed in the SEM and AFM images (Fig. 10c, 40× magnification; Fig. 10d, digital zoom of the 40× magnification).

Atomic Force Microscopy as Nanorobot

497

Fig. 10. Visualization of HaCaT cells by SEM and immunofluorescence microscopy. (a) Low magnification SEM imaging (scale bar: 20 mm) reveals the typical “cobblestone” architecture of cultured keratinocytes. (b) High magnification SEM imaging (scale bar: 1  mm) reveals intercellular structures in parallel organization bridging adjacent keratinocytes. (c) Immunofluorescence microscopy shows stained keratin filaments (cytokeratin 18) throughout the cells (×40 magnification). (d) A digitally zoomed image from the ×40 immunofluorescence image reveals intercellular structures that bridge adjacent keratinocytes.

In contrast to SEM and IF, which must be performed on fixed cells, AFM allows us to study the behavior of such cellular structures. It provides a much higher spatial resolution than other methods, and imaging does not depend on the use of antibodybound staining reagents or membrane permeability for fluorescent dyes. Furthermore, AFM can be used to measure nanomechanical properties of desmosomal structures quantitatively, for example, elasticity, to provide a more comprehensive examination of biological behavior. 3.3.2. Effect of Anti-Dsg3 Antibody Binding to Living HaCaT Cells

We next investigated the effect of specific antibody binding to a critical structural component of the keratinocyte cell junction, Dsg3. When living HaCaT cells were incubated without antibody, structures likely representing intercellular adhesion structures were clearly visualized between the cell membranes (Fig.  11a). When a second batch of living cells was incubated with anti-Dsg3 antibody for 24  h, the structure of the cell junctions changed

498

Xi et al.

Fig. 11. Effect of anti-Dsg3 antibody treatment on intercellular structures. (a) HaCaT cells were visualized at time 0 h, or (b) incubated with anti-Dsg3 antibody for 24 h before visualization by AFM. Structures likely representing intercellular adhesion structures are clearly visible between keratinocytes (a). These intercellular structures partly disappear after 24-h antibody treatment (b). This observation was confirmed in three independent experiments.

s­ ignificantly (Fig.  11b). We demonstrated that the ­intercellular structures linking two adjacent cells are no longer clearly ­detectable, indicating that the molecular structure between the cells was demonstrated altered by anti-Dsg3 antibody binding. In summary, the development of an AFM-based nanorobot opens many exciting possibilities in biomedical research. While the ­single molecules can be identified in their physiologic condition, manipulation of single molecules on a living cell membrane ­surface also becomes possible. By combining high-resolution imaging and the controlled manipulation of single biomolecules, AFM nanorobot is a powerful tool to study the molecular dynamics and interactions, thereby revealing the functionality of an ­individual biomolecule. We anticipate that the research that seeks to understand and exploit the interaction forces between nanoprobing mechanisms will provide a leap forward for biomedical research, whose progress is limited by the cumbersome and static multistep methods currently available. The future applications of AFM nanorobot in biomedical research are expanding and will have a profound impact on ­disease prevention, diagnosis, and treatment. Using AFM nanorobot for biosensing and actuation is one of the exciting developments. The AFM manipulator can be modified and used to image, probe, and identify particular biological and chemical entities of a biological sample simultaneously. In the future, they can be used to simulate various physiological conditions and look at the interaction between biomolecules in a wide range of ­circumstances. This opens the door to more effective uses of AFM nanorobot in drug

Atomic Force Microscopy as Nanorobot

499

discovery, disease detection, and cell and gene therapy. Moreover, researchers are exploring the use of AFM as a nanorobot for nanosurgery and controlled drug delivery. Nanosurgery techniques may potentially be developed using AFM-based manipulation system with a nanoneedle and femtosecond laser surgery. The AFM nanorobot can also be modified and integrated into a fluidic delivery system such that the selected molecules can be dispensed into individual cells for drug delivery and testing. There are many other emerging applications of AFM-based nanorobot, and researchers can now build the path to new discoveries of cellular structure and function.

4. Notes 1. Cell plating for AFM imaging and manipulation: the sample preparation procedures used in different studies vary widely. For the experiments presented here, the key steps are described as follows: (a) A glass coverslip is coated with poly-l-ornithine (SigmaAldrich, St. Louis, MO) to enhance primary cell adhesion to the glass coverslip. (b) The cells reach confluency after 7–10 days of cell plating and are ready for AFM imaging. (c) For experiments with fixed cells, the glass coverslip carrying the cells is washed with PBS and then fixed by adding 3.75% formaldehyde (Sigma-Aldrich) for at least 15 min. (d) For experiments with living cells, the glass coverslip is transferred to the AFM instrument directly after washing with PBS. (e) The glass coverslip is then glued to a magnetic stainless disc for experiments. (f) To avoid drying and dying of the cells during the AFM scanning, a thin layer of DMEM (Gibco-Invitrogen, Carlsbad, CA) is added to the sample immediately before visualization. 2. Potential problems during AFM imaging and manipulation, and solutions: (a) Before starting the scanning, a calibration of the cantilever is necessary to avoid the variation of the data obtained from the AFM image. The variation in the repulsive force with sample indentation is dependent upon the material properties of the tip and the sample, as well as the shape of the tip. A feature for determining the spring constant

500

Xi et al.

is provided by the AFM software. The calibration can also be performed by obtaining the reference force curve for a relatively stiff material (e.g., glass substrate) with known mechanical properties. (b) Another important step to successful imaging of living cells in fluid is the selection of an appropriate probe. This choice is largely dependent on the sample characteristics (hardness, roughness, etc.). Based on our work, a silicon nitride cantilever with a spring constant of 0.38 N/m is used. (c) The time duration for the study must be optimized to avoid living cells suffering from nonphysiological conditions during the analysis. For the experiment presented here, the study is set and completed within 2 h. (d) Because changes in temperature and environmental condition during the long imaging intervals will affect ­viability and metabolism of cells, the resulting images may be affected. In this case, a commercially available fluid cell module is used to maintain the consistent amount of DMEM supplied to the sample during AFM scanning. For some AFM systems, such as Bioscope (Veeco Metrology Inc., Santa Barbara, CA), a specifically designed fluid cell module is needed. Moreover, the AFM is placed in an acoustic and vibration isolation chamber. (e) The positioning problem is one of the most critical issues in AFM-based nanomanipulation. The positioning errors are mainly caused by the thermal drift. The position error due to the thermal drift can be reduced using the fast local scans before and after each operation. Before starting the manipulation operation, the actual position of the manipulated nano-object can be identified by a quick local scan. Nanomanipulation is then performed immediately after the local scan. From the AFM image, the local position of a nano-object is obtained. A local scanning pattern is generated for the nano-object. The scanning pattern is fed to the imaging interface to scan the surface. If the nanoobject is not found, a new scanning pattern will be generated. The process continues until the nano-object is discovered. The actual position of the nano-object can then be computed. The manipulation path is then adjusted based on the actual position of the nano-object. (f) Biological samples can be imaged with AFM in either of two imaging modes: contact mode or tapping mode. In contact mode, the tip is brought in contact with the ­surface and the cantilever deflection is kept constant during the scanning by the feedback loop. However, when using contact mode to study the biological sample, such as ­living cells or biomolecules, the sample is easily damaged­

Atomic Force Microscopy as Nanorobot

501

due to large shear force applied by the tip. On the other hand, in tapping mode, the cantilever is driven to oscillate up and down at its resonance frequency by a small piezoelectric element mounted in the AFM tip holder. Tapping mode, therefore, is more suitable for studying living cells in liquid.

Acknowledgments This research work is partially supported by NSF Grants IIS-0713346 and DMI-0500372, and ONR Grants N00014-04-1-0799 and N00014-07-1-0935. References 1. Binnig, G., Quate, C. F., and Gerber, C. (1986) Atomic force microscope. Phys Rev Lett. 56, 930–933. 2. Engel, A., and Müller, D.J. (2000) Observing single biomolecules at work with the atomic force microscope. Nature Structural Biology. 7(9), 715–718. 3. Hörber, J.K.H., and Miles, M.J. (2003) Scanning probe evolution in biology. Science 302, 1002-1005. 4. Dufrêne, Y.F. (2004) Using nanotechniques to explore microbial surfaces. Nature Reviews Microbiology 2, 451–460. 5. Schaefer, D.M., Reifenberger, R., Patil, A., and Andres, R.P. (1995) Fabrication of twodimensional arrays of nanometer-size clusters with the atomic force microscope. Appl. Phys. Lett. 66, 1012–1014. 6. Junno, T., Deppert, K., Montelius, L., and Samuelson, L. (1995) Controlled manipulation of nanoparticles with an atomic force microscope. Appl. Phys. Lett., 66(26), 3627–3629. 7. Hansma, H.G. (1992) Reproducible imaging and dissection of plasmid DNA under liquid with the atomic force microscope. Science 256, 1180–1184. 8. Vesenka, J., Guthold, M., Tang, C.L., Keller, D., Delaine, E., and Bustamante, C. (1992) Substrate preparation for reliable imaging of DNA molecules with the scanning force microscope. Ultramicroscopy 42–44, 1243–1249. 9. Thalhammer, S., Stark, R.W., Müller, S., Wienberg, J., and Heckl, W.M. (1997) The atomic force microscope as a new microdissecting­ tool for the generation of genetic probes. J Struct Biol 119, 232–237.

10. Toshiya, O., Hironori, U., Hyonchol, K., and Atsushi, I. (2003) mRNA analysis of single living cells. J Nanobiotech. 1, 2. 11. Sagvolden, G., Giaever, I., Pettersen, E.O., and Feder. (1999) J Cell adhesion force microscopy. Proc Natl Acad Sci USA 96, 471–476. 12. Charras, G., and Horton, M. (2002) Single cell mechanotransduction and its modulation analyzed by atomic force microscope indentation. Biophys J. 82, 2970–2981. 13. Oesterhelt, F. (2000) Unfolding pathways of individual bacteriorhodopsins. Science 288, 143–146. 14. Viani, M.B., Pietrasanta, L.I., Thompson, J.B., Chand, A., Gebeshuber, I.C., Kindt, J.H., Richter, M., Hansma, H.G., and Hansma, P.K. (2000) Probing protein–­protein interactions in real time. Nat Struct Biol. 7, 644–647. 15. Bustamante, C., Macosko, J., and Wuite, G. (2000) Grabbing the cat by the tail: manipulating molecules one by one. Nat Rev Mol Cell Biol. 1, 130–136. 16. Hansen, L.T., Kuhle, A., Sorensen, A.H., Bohr, J., and Lindelof, P.E. (1998) A technique for positioning nanoparticles using an atomic force microscope. Nanotechnology 9,337–342. 17. Resch, R., Baur, C., Bugacov, A., Koel, B.E., Madhukar, A., Requicha, A.A.G., and Will, P. (1998) Building and manipulating three- dimensional and linked two-­ dimensional structures of nanoparticles using scanning force microscopy. Langmuir 14(23), 6613–6616.

502

Xi et al.

18. Sitti, M., and Hashimoto, H. (1998) ­Tele-nanorobotics using atomic force microscope. Proc. IEEE Int. Conf. Intelligent Robots and Systems 1739–1746. 19. Guthold, M., Falvo, M.R., Matthews, W.G., Washburn, S.S., Paulson, and Erie, D.A. (2000) Controlled manipulation of molecular samples with the nanomanipulator. IEEE/ ASME Trans. on Mechatronics 5(2), 189–198. 20. Putman, C.A.J., van der Werf, K.O., de Grooth, B.G., van Hulst, N.F., and Greve, J. (1994) Viscoelasticity of living cells allows high resolution imaging by tapping mode atomic force microscopy. Biophys. J. 67, 1749–1753. 21. Prater, C.B., Wilson, M.R., Garnaes, J., Massie, J., Eling, V.B., and Hasma, P.K. (1991) Atomic force microscopy of biological samples at low temperature. J. Vac Sci. Tech. B. 9, 989–991. 22. Butt, H.J., Wolff, E.K., Gould, S.A.C., Northern, B.D., Peterson, C.M., and Hansma, P.K. (1990) Atomic force microscopy of biological samples at low temperature. J. Struct. Biol. 105, 54–61. 23. Hansma, P.K., Cleveland, J.P., Radmacher, M., Walters, D.A., Hillner, P.E., Bezanilla, M., Fritz, M., Vie, D., and Hansma, H.G. (1994) Tapping mode atomic force microscopy in ­liquids. Appl. Phys. Lett. 64, 1738–1740. 24. Baker, A.A., Helbert, W., Sugiyama, J., and Miles, M.J. (2000) New insight into cellulose structure by atomic force microscopy shows the ia crystal phase at near-atomic resolution. Biophys. J. 79, 1139–1145. 25. Fotiadis, D., Scheuring, S., Müller, S.A., Engel, A., and Müller, D.J. (2002) Imaging and manipulation of biological structures with the AFM. Micron 33, 385–397. 26. Willemsen, O.H., Snel, M.M.E., van der Werf, K.O., de Grooth, B.G., Greve, J., Hinterdorfer, P., Gruber, H.J., Schindler, H., van Kooyk, Y., and Figdor, C.G. (1998) Simultaneous height and adhesion imaging of antibody-antigen interactions by atomic force microscopy. Biophys. J. 75, 2220–2228. 27. Ludwig, M., Dettmann, W., and Gaub, H.E. (1997) Afm imaging contrast based on molecular recognition. Biophys. J. 72, 445–448. 28. Raab, A., Han, W., Badt, D., Smith-Gill, S.J., Lindsay, S.M., Schindler, H., and Hinterdorfer, P. (1999) Antibody recognition imaging by force microscopy. Nat. Biotechnol. 17, 902–905. 29. Chen, H., Xi, N., and Li, G. (2006) CADguided automated nanoassembly using atomic force microscopy-based nanorobotics. IEEE Transactions on Automation Science and Engineering 3(3), 208–217. 30. Li, G., Xi, N., Yu, M., and Fung, W. K. (2003) Augmented reality system for real-time

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

­ anomanipulations. Proc. IEEE Int. Conf. n Nanotechnology 2, 64–67. Li, G., Xi, N., and Yu, M. (2004) Development of augmented reality system for afm based nanomanipulations. IEEE/ASME Trans. on Mechatronics 9, 358–365. Zhang, J., Xi, N., Liu, Li., Chen, H., Lai, K. W. C., and Li, G. (2008) Atomic force yields a master nanomanipulator. IEEE Nano­ technology Magazine 2(2), 13–17. Li, G., Xi, N., Chen, H., Pomeroy, C., and Prokos, M. (2005) “Videolized” atomic force microscopy for interactive nanomanipulation and Nanoassembly. IEEE Transactions on Nanotechnology 4, 605–615. Liu, L., Luo, Y., Xi, N., Wang, Y., Zhang, J., and Li, G. (2008) Sensor referenced real-time videolization of atomic force microscopy for nanomanipulations. IEEE/ASME Transactions on Mechatronics 13(1), 76–85. Li, G., Xi, N., Wang, Y.C., Yu, M., and Fung, W.K. (2004) Planning and control of 3-D nano-Manipulation. Acta Mechanica Sinica. 20(2), 117–124. Li, G., Xi, N., and Yu, M. (2004) Augmented Reality Enhanced “Top-Down” nanomanufacturing. Proc. of IEEE Conference on Nanotechnology 352–354. Li, G., Xi, N., and Yu, M. (2004) Calibration of AFM based nanomanipulation system. Proc. of IEEE International Conference on Robotics and Automation 422–427. Li, G., Xi, N., Chen, H., Saeed, A., and Yu, M. (2004) Assembly of nanostructure using AFM based nanomanipulation system. Proc. of IEEE International Conference on Robotics and Automation 428–433. Zhang, J., Xi, N., Li, G., Chan, H., and Wejinya, U.C. (2006) Adaptable end effector for atomic force microscopy based nanomanipulation. IEEE Transactions on Nano­ technology 5(6), 628–642. Li, G., Xi, N., Chen, H., Saeed, A., Li, W.J., Fung, C.K.M., Chan, R.H.M., Zhang, M., and Tarn, T.J. (2004) Nano-assembly of DNA based electronic devices using atomic force microscopy. Proc. of IEEE/RSJ International Conference on Intelligent Robotics and Systems 583–588. Li, G., Xi, N., Chen, H.P., Saeed, A., Zhang, J., Li, W.J., Fung, C.K.M., Chan, R.H.M., Zhang, M., and Tarn, T.J. (2004) Experimental studies of DNA electrical properties using AFM based nano-manipulator. Proc. of IEEE Conference on Nanotechnology 308–310. Li, G., Xi, N., and Wang, D.H. (2005) Investigation of angiotensin II type 1 receptor by atomic force microscopy with functionalized probe. Nanomedicine: Nanotechnology, Biology, and Medicine 1(4), 302–312.

Atomic Force Microscopy as Nanorobot 43. Li, G., Xi, N., and Wang, D.H. (2006) Probing membrane proteins using atomic force microscopy. J. Cellular Biochem. 97, 1191–1197. 44. Li, G., Xi, N., and Wang, D.H. (2005) In situ sensing and manipulation of molecules in biological samples using a nanorobotic system. Nanomedicine: Nanotechnology, Biology, and Medicine 1(1), 31–40. 45. Fung, C.K.M., Kristina, S., Lai, K.W.C., Yang, R., Panyard, D., Zhang, J., Xi, N., and Sinha, A. A. (2010) Investigation of human keratinocyte cell adhesion using atomic force microscopy (AFM). Nanomedicine: Nanotechnology, Biology, and Medicine, 6(1), 191–200. 46. Zhang, J., Xi, N., and Lai, K. W. C. (2007) Fabrication and testing of a nanoinfrared detector using a single carbon nanotube

47.

48.

49.

50.

503

(CNT). SPIE Newsroom. (online at: http:// spie.org/x8489.xml). Zhang, J., Xi, N., Chen, H. Z., Lai, K. W. C., and Li, G. (2008) Photovoltaic effect in single carbon nanotube based Schottky diodes. International Journal of Nanoparticles 1(2), 108–118. Zhang, J., Xi, N., Chen, H. Z., Lai, K. W. C., and Li, G. (2009) Design, manufacturing and testing of single carbon nanotube based infrared sensors. IEEE Transactions on Nanotechnology 8(2), 245–251. Chen, H. Z., Xi, N., Lai, K. W. C., and Zhang, J. (2008) Infrared detection using carbon nanotube field effect transistor. Proc. of IEEE Conference on Nanotechnology 88–91. Fung, C. K. M., Xi, N., Shanker, B., and Lai, K. W. C. (2009) Nanoresonant signal booster for carbon nanotube based infrared detectors. Nanotechnology 20, 185201.

wwwwwwwwwwwwww

Index

A Aminofunctionalization.................. 225, 227–229, 235, 238 Amplitude and phase distance (APD) curves............. 75–78 Amyloid fibrils apolipoprotein A-I................................................ 82, 83 b2-microglobulin.......................................82–88, 91–92 Antibiotics sub-MICs................................................................. 395 supra-MICs.............................................................. 395 Apolipoprotein A-I, purification................................ 87, 88 Artifacts creep..................................................................... 37–38 double image...................................................... 34, 129 hysteresis,.............................................................. 37, 38 image processing,.......................................32, 39, 41–42 linearization,......................................................... 36–37 sample tilting,....................................................... 39–40 scanner............................................................ 32, 34–41 thermal drift......................................................... 38, 40 testing......................................................................... 42 tip�������������������������������������������� 32–34, 42, 115, 129, 350 vibration..................................................................... 40 Aspirin.................................................................... 425–435 Atomic force microscopy (AFM) biomanipulation........................................................ 486 biomedical research......................68, 486–487, 492–499 calibration....................... 21, 27, 35–37, 90, 93, 200–201, 246, 247, 312, 313, 319, 320, 398, 404, 409, 499, 500 cell mechanics....................................305–307, 311–315 constant temperature................................................ 360 contact mode...................16, 21–24, 41, 58, 61, 66, 104, 105, 121, 134, 139, 142, 147–149, 155–157, 161, 163–165, 172, 173, 216–218, 255, 262, 310, 342, 398, 408, 409, 414, 419, 420, 430, 487, 500 data analysis.......................................315–318, 385–386 deflection mode.......................................................... 22 deflection sensor................................................... 11–12 dynamic force spectroscopy................ 27, 202, 204, 209, 332, 350, 440 electrophysiology...................................................... 373 fluorescence microscopy.............................254, 439–453 force-clamp................ 331–336, 338–344, 347, 348, 350

force curves................................... 26–28, 202, 380, 385, 414, 415, 418–420 force measurements.............. 20, 26, 311, 335–339, 385, 387, 418–420, 442 force modulation................................................. 25, 304 high-speed imaging.......................................... 285–299 image formation................................................... 12–13 imaging.................... 4, 19, 31, 47, 61, 87, 109, 119, 133, 153, 171, 212, 224, 243, 263, 285, 310, 335, 392, 404, 414, 428, 439, 457, 486 indentation................... 27, 245–247, 249–253, 255, 256, 310–318, 320–322, 398, 409, 416, 420, 421, 423, 499 intermittent contact mode....................24–25, 112, 212, 217–219, 229, 398, 404, 408, 440 measurement......... 4, 19, 36, 57, 89, 156, 174, 200, 224, 255, 260, 306, 331, 365, 384, 398, 413, 425, 440 mechanical testing.....................................308–311, 313 noncontact mode............................ 19, 23–24, 263, 398, 409, 430, 431, 440, 486 parachuting........................................287, 289, 290, 296 performance range.................................................... 5–6 phase imaging mode....................................25, 212, 310 relative humidity................................................... 69, 70 rendering.............................41, 186, 248, 394, 399, 403, 406, 409 resolution...........................4, 22, 32, 56, 61, 81, 97, 110, 123, 134, 153, 174, 211, 226, 243, 259, 285, 309, 338, 357, 392, 402, 425, 439, 487 sample��������������������������4, 19, 31, 48, 61, 81, 98, 110, 120, 134, 154, 172, 199, 211, 227, 243, 259, 287, 308, 333, 359, 385, 392, 402, 419, 425, 439, 487 scanner................. 6–9, 12–15, 17, 26, 27, 32, 34–41, 50, 65, 66, 99, 121, 155, 203, 216, 246, 255, 287–290, 296, 298, 310, 312, 321, 398, 404, 409, 427, 440 setting.................... 23, 40, 41, 66–67, 91, 104, 105, 140, 156, 236, 246–247, 404, 408, 451 tapping mode...................48, 53, 83–86, 89, 91, 99, 104, 105, 118, 121, 122, 139, 148, 149, 172, 173, 175, 176, 192, 224, 227, 230, 232, 275, 287, 296, 310, 398, 408, 428, 430, 431, 487, 492, 500, 501 time-lapse........................................................... 97–106 vibrations.....................15, 40, 66, 68, 99, 128, 162, 163, 296, 312, 342, 404, 423, 444, 450, 487, 500

Pier Carlo Braga and Davide Ricci (eds.), Atomic Force Microscopy in Biomedical Research: Methods and Protocols, Methods in Molecular Biology, vol. 736, DOI 10.1007/978-1-61779-105-5, © Springer Science+Business Media, LLC 2011

505

Atomic Force Microscopy in Biomedical Research: Methods and Protocols 506  Index

  

B Bacteria adhesion forces......................................................... 380 B. cereus..................................................................... 396 culture........................................................349, 380–381 Gram negative.......................................................... 394 Gram positive........................................................... 394 shape................................................................. 391–399 surface........................................................391–399, 402 Bell model.............................................................. 332, 346 b2-microglobulin aggregation................................................83–88, 91–92 fibrillization.................................................... 84–86, 91 preparation............................................................ 87, 91 purification................................................................. 91 Biomechanics, subcellular................................. 314–315 Biomolecular processes, imaging.............285–291, 294–297 Biotin������������������������������ 292, 298, 332, 348, 349, 465, 480

C Candida albicans measurements................................................... 407, 409 shapes....................................................................... 404 size��������������������������������������������������������������������404, 407 Cantilever composition.............................................................. 336 dimension.................................. 174, 288, 296, 309, 336 noise limit......................................................... 338–339 reflective coating....................................................... 336 resonance frequency........................ 11, 27, 89, 104, 112, 163, 440, 444, 501 rigidity...................................................................... 490 slope deflection......................................................... 342 spring constant............................. 11, 21, 50, 65, 69, 87, 111, 112, 155, 173, 202–204, 216, 217, 226, 232, 244, 246, 247, 255, 288, 296, 309, 310, 315, 334, 336, 338, 342, 343, 350, 361, 385–387, 412, 418–422, 500 thermal drift................................................40, 336, 423 Capsids............................ 172, 177–181, 183–188, 190–192 Cefodizime..................................................................... 396 Cell adhesion................98, 319, 331, 355–374, 412, 495, 499 elastic properties............................................... 318, 356 endothelial.................................................364, 411–423 epithelial.............254, 307, 318, 368, 371, 492, 495–499 layers..........................................357, 367–368, 371–373 mechanical properties....................... 243, 303, 306, 313, 315, 318, 361, 369, 412, 414 surface........................176, 182, 192, 212, 230, 250, 251, 255, 362, 402, 404, 413, 415, 416, 418–422, 495 yeast.......................................................................... 403 Chromatin................109, 264, 265, 267, 268, 275, 282, 442

Chromosome metaphase......................................................... 109–115 native................................................................ 110–112 observation in liquid......................................... 111, 112 Clusters molecular.......................................................... 369, 373 integrin............................................................. 369, 370 Collagen assembly.............................................................. 98–100 high-resolution................................................... 97–106 imaging..................................................................... 104 matrices...............................................97–101, 103–105 reshaping.......................................................... 101, 105 self-assembly....................................................... 97–106 Coupling method amine/aldehyde................................................. 471–472 amine/succimide............................................... 469–471 amino surface.................................................... 468–469 NTA/histidine.................................................. 478–480 thiol surfaces..................................................... 473–474 thiol/disulfide................................................... 475–477 thiol/maleimide................................................ 474–475 thiol/sulfone..................................................... 477–478

D Daptomycin.................................................................... 397 Dictyostelium discoideum.......................................... 365, 366 DNA double-stranded................. 63, 68–70, 72, 187–190, 426 drug interactions....................................................... 427 gyrase.......................................................................... 68 quantitative analysis...................................431, 433, 434 sample preparation.....................................425, 427, 429

E Erythrocytes............................................223–239, 306, 307 Escherichia coli.......................... 156–160, 379–387, 391, 396 Essential oils................................................................... 402

F Fibrillogenesis.................................................................. 81 Fibroblast....................................................................... 199 Flagellum................................................................ 278–281 Fluorescence beads................................................................. 444, 451 imaging......................................................133, 141, 446 labeling............................................................. 320, 441 Force spectroscopy experimental prerequisites................................ 359–361 living cells......................................................... 364–373 single cell................................... 357, 359–361, 371, 373 single molecule................................. 154, 163, 197, 332, 336–338, 348, 356 Fungi�������������������������������������������������������������� 305, 402–404



Atomic Force Microscopy in Biomedical Research: Methods and Protocols 507 Index     

G Glutaraldehyde..................99, 105, 172, 175, 213–215, 226, 229, 237, 244, 245, 360, 380, 383, 384, 394, 403, 471

H Halobacterium halobium................................................... 157 Heparin...........................................................380, 383, 412 Hertzian theory.............................................................. 315 Histamine................................................413, 416, 420, 421 H. pylori.......................................................................... 397 Hyaluronan......................................................380, 382, 383 Hysteresis....................................................37, 38, 319, 335

I Immunoglobulin G antibodies physisorption.................................................. 65–66, 73 preparation............................................................ 64, 68 setting parameters................................................. 66–67 Integrin, adhesion forces..................................413, 415, 420 Interactions receptor-ligand....................................48, 198, 331–350 streptavidin-biotin............................................ 332, 347

J Junctions desmosomal...............................................492, 495–499 proteins..................................................................... 495

K Keratinocyte........................................................... 495–498

L Lectins . .......................................................................... 360 Lymphocytes...................................................111, 114, 182

M M. catarrhalis.................................................................. 397 Membrane receptors analysis......................................................48, 49, 54–56 antibody-receptor......................................54, 55, 58, 59 channels...................................................................... 48 imaging................................................................. 47–59 isolation.......................................................... 48, 50–52 proteins.......................................... 47, 48, 50–54, 56–58 Mica support������������������������ 52, 65, 98, 103, 104, 121, 123, 129, 154, 155, 292–293, 427 Microscopes diffraction limit..........................................3, 4, 441, 451 electron-based............................................................... 4 far field..................................................................... 392

radiation-based............................................................. 4 scanning electron.................5, 6, 11, 110, 214–216, 220, 293, 370, 380, 384, 385, 387, 392, 393, 397, 402, 408, 496, 497 scanning probe.........................4, 17, 31, 41, 64, 65, 260, 385, 392, 404 Microscopy fluorescence...............................................254, 439–453 lateral force........................................................... 22–23 Microtubules...........................216, 218, 253, 278, 286, 295, 299, 305, 308 Mitochondria clusters.......................................................139–142, 148 images................ 133–136, 139–142, 144–149, 161–162 isolation.....................................................133, 137–138 membranes.........133–135, 144, 148, 149, 154, 161–162 treatment.......................................................... 138–139 Moxifloxacin................................................................... 397

N Nanolithography............................................................. 369 Nanomanipulation...........................485–487, 489–491, 500 Nano-Newton...........................................63, 335, 357–359 Nanorobot DNA molecules.........................................492, 494–495 living neuron..................................................... 492–494 Nanoscribe.............................................................. 101, 105 Neurons chick embryo.................................................... 244, 245 culture....................................................................... 245 growth cone...................................................... 243–256 imaging...................................... 245–250, 252, 254, 255 mechanical properties........................243, 244, 250–253 topography........................................................ 250–253

P Pico-Newton............................................................ 63, 425 Piezoelectric element.......................8, 39, 40, 246, 333–335, 338, 501 feedback control....................................................... 335 Poly-lysine slides.................................................... 139, 248 Porin . ......................................................135, 154, 156–161 Proteasoma analysis..............................................118, 123–128, 130 crystal structure.......................... 118, 119, 123, 125, 130 imaging............................................................. 120–128 preparation.................................................121–122, 129 Proteins affinity....................................... 228, 291, 292, 357, 358 immobilizing............................. 121, 341, 349, 360, 471 membrane........................... 47–49, 53, 56–58, 133–135, 144, 148, 149, 153–165, 224, 225, 229, 230, 233, 357, 358, 364, 440, 487

Atomic Force Microscopy in Biomedical Research: Methods and Protocols 508  Index

  

R Red blood cell (RBC) attachment.................................................225, 227–229 imaging...................................... 224–227, 229, 232, 235 membrane.......................... 224–227, 230–234, 237, 238 preparation................................. 224–226, 229, 230, 232 Rokitamycin........................................................... 396, 397

S Sabouraud broth............................................................. 403 Saccharomyces cerevisiae............................................ 128, 130 Sample adhesion.......................................................93, 192, 255 holder............... 6, 8, 16, 88, 99, 162, 246, 397, 408, 428 immobilization..........................................154, 155, 427 preparation....................... 4, 6, 53, 64, 68, 71, 83, 87, 88, 91, 134, 147, 162–164, 181–185, 245, 259, 394, 397, 402, 408, 425, 427, 429, 440, 499 Sensors capacitive....................................................... 9, 335 Silanization...................................... 458–460, 468, 469, 480 Silicone surface....................................................... 379–387 Single molecule force-spectroscopy........ 154, 163, 197, 332, 336–338, 348, 356 Smooth muscle culture............................................................... 412, 414 elasticity measurements.................................... 420–421 pulling forces.................................................... 421–422 topography........................................................ 414, 418 vascular............................................................. 411–423 Specimen, biological.........171, 392, 393, 398, 402, 409, 487 Spectroscopy, dynamic................................................ 27–28 Sperm imaging......................................................263–264, 275 measurement..............259, 260, 262–263, 265, 267, 282 preparation.................................................259, 261–262 temperature............................... 260–262, 264, 265, 267, 271–275, 277, 280–282 Spermatozoa flagellum........................................................... 278–281 head.................................................................. 264–275 neck........................................... 264, 265, 270, 275–278 S. pyogenes....................................................................... 396 Staphylococcus aureus.........................................391, 397, 398 Stiffness.........................25–27, 62, 246, 247, 252, 253, 263, 304, 305, 307, 314–316, 319, 336, 349, 359, 364, 412, 449 Strain gauges.............................................................. 9, 335 Streptavidin.............................286, 291, 292, 297, 332, 341, 344, 465, 466

T Tethers.....................157, 206, 336, 349, 358, 359, 361–364, 366, 368, 372, 490, 491

Thymol................................................................... 401–409 Tip artifact.......................................................... 32–34, 115 azide-functionalized......................................... 466–468 etherification......................................460, 461, 468, 469 functionalization............... 333, 336, 339–341, 458–464, 466, 471, 473, 476, 478, 491 geometrical shape................................................. 32, 93 gold-coated.......................................333, 339–341, 458, 463, 464, 466, 469, 471, 473, 474, 479, 480 hydrosilylation...................................461–463, 468, 480 modification...............................383–384, 413, 457–480 sample interaction................................. 4, 20, 21, 24, 27, 28, 59, 63, 67, 68, 164, 253, 286, 289, 290, 295, 398, 405, 408, 409, 488 silicon.................................. 10, 458–464, 468, 473, 480 surfaces............................ 199, 336, 457–465, 467, 468, 471–475, 477–479 Trypanosoma cruzi cultivated medium.................................................... 214 imaging............................................................. 216–219 preparation................................................................ 216

V Vincristine.............................................................. 425–435 Virus architecture........................................176, 177, 180, 183 brome mosaic.............................................175, 179, 188 cauliflower mosaic..................................................... 176 classification...................................................... 176, 188 crystalline form................................................. 175, 176 herpes simplex...........................................176, 178, 179 human rhinovirus (HRV2), ...............198, 200, 207, 479 imaging............................................................. 175–176 internal structure imaging................................. 176, 186 measures................................................................... 176 mimivirus.......................................... 175, 176, 179, 180, 183–186, 193 Moloney Mouse Leukemic...............177, 178, 181–183, 186, 187, 191 mutant.......................................................181–185, 191 nucleic acids............................... 172, 179, 187–191, 193 PBCV–1 .......................................... 175, 177, 179, 180, 183, 184, 189 population analysis................................................... 185 preparation........................................................ 181–185 receptor interaction................................................... 205 receptor kinetic................................................. 205, 208 Tipula iridescent....................................................... 176 tomato bushy stunt................................................... 175 turnip yellow mosaic.................. 175, 179, 186, 188, 189 vaccinia.............................. 175–177, 183, 186–190, 193 Voltage-dependent anion channel (VDAC) imaging....................................................... 161–162