Biosensors and Biodetection: Methods and Protocols Volume 1: Optical-Based Detectors (Methods in Molecular Biology Vol 503)

Biosensors and Biodetection Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hatfield,...

Author: Avraham Rasooly | Keith E. Herold

290 downloads 970 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Biosensors and Biodetection

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

For other titles published in this series, go to www.springer.com/series/7651

METHODS

IN

MOLECULAR BIOLOGY™

Biosensors and Biodetection Methods and Protocols Volume 503: Optical-Based Detectors

Edited by

Avraham Rasooly* and Keith E. Herold† *FDA Center for Devices and Radiological Health, Silver Spring, MD, USA and National Cancer Institute, Bethesda, MD, USA † Fischell Department of Bioengineering, University of Maryland, College Park, MD, USA

Editors Avraham Rasooly FDA Center for Devices and Radiological Health Silver Spring, MD USA and National Cancer Institute Bethesda, MD USA [email protected]

Keith E. Herold Fischell Department of Bioengineering University of Maryland College Park, MD USA [email protected]

ISBN: 978-1-60327-566-8 e-ISBN: 978-1-60327-567-5 ISSN: 1064-3745 e-ISSN: 1940-6029 DOI: 10.1007/978-1-60327-567-5 Library of Congress Control Number: 2008941063 © Humana Press, a part of Springer Science+Business Media, LLC 2009 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 springer.com

Preface 1. Biosensor Technologies In recent years, many types of biosensors have been developed and used in a wide variety of analytical settings, including biomedical, environmental, research, and others. A biosensor is defined by the International Union of Pure and Applied Chemistry (IUPAC) as a “device that uses specific biochemical reactions mediated by isolated enzymes, immunosystems, tissues, organelles, or whole cells to detect chemical compounds usually by electrical, thermal, or optical signals” (1). Thus, almost all biosensors are based on a two-component system: a biological recognition element (ligand) that facilitates specific binding to or biochemical reaction with a target, and a signal conversion unit (transducer). Although it is impossible to fully cover the fast-moving field of biosensing in one publication, this publication presents some of the many types of biosensors to give the reader a sense of the enormous potential for these devices. An early reference to the concept of a biosensor is from Dr. Leland C. Clark, who worked on biosensors in the early 1960s (2) developing an “enzyme electrode” for glucose concentration measurement with the enzyme glucose oxidase, a measurement that is important in the diagnosis and treatment of disorders of carbohydrate metabolism in diabetes patients. Still today, the most common biosensors used are for glucose analysis. A large number of basic biosensors, all combining a biological recognition element and a transducer, were subsequently developed. Currently, the trend is toward more complex integrated multianalyte sensors capable of more comprehensive analyses. Advances in electronics and microelectrical and mechanical systems (MEMS) have enabled the miniaturization of many biosensors and the newest generation biosensors include miniaturized multianalyte devices with high-throughput capabilities and more than 1,000 individually addressable sensor spots per square centimeter. A useful categorization of biosensors is to divide them into two groups: direct recognition sensors, in which the biological interaction is directly measured, and indirect detection sensors, which rely on secondary elements for detection. Figure 1 shows a schematic of the two groups of biosensors. In each group, there are several types of transducers including optical, electrochemical, and mechanical. For all of these technologies, the recognition ligand plays a major role. Although the most commonly used ligands are antibodies, other ligands are being developed including aptamers (protein-binding nucleic acids) and peptides. In the literature and in practice, there are numerous types of biosensors, and the choice of a suitable system for a particular application is complex, based on many factors such as the nature of the application, the label molecule (if used), the sensitivity required, the number of channels (or area), cost, technical expertise, and the speed of detection needed. A primary purpose of this book is to provide more access to the

v

vi

Preface

Fig. 1. General schematic of biosensors: (a) direct detection biosensors where the recognition element is label free; (b) indirect detection biosensors using a “sandwich” assay where the analyte is detected by labeled molecule. Direct detection biosensors are simpler and faster but typically yield a higher limit of detection compared with indirect detection systems

technical methods involved in using a variety of biosensors to facilitate such decision making. Direct detection biosensors utilize direct measurement of the biological interaction. Such detectors typically measure physical changes (e.g., changes in optical, mechanical, or electrical properties) induced by the biological interaction, and they do not require labeling (i.e., label free) for detection. Direct biosensors can also be used in an indirect mode, typically to increase their sensitivity. Direct detection systems include optical-based systems (most common being surface plasmon resonance) and mechanical systems such as quartz crystal resonators. Indirect detection sensors rely on secondary elements (labels) for detection. Examples of such secondary elements are enzymes (e.g., alkaline phosphatase or glucose oxidase) and fluorescently tagged antibodies that enhance detection of a sandwich complex. Unlike direct detectors, which directly measure changes induced by biological interactions and are “label free,” indirect detectors require a labeled molecule to bind to the target. Most indirect sensors based on optical detection are designed to measure fluorescence. The detection system can be based on a charge coupled device (CCD), photomultiplier tube (PMT), photodiode, or spectrometer. Electrochemical transducers, which measure the oxidation or reduction of an electroactive compound on the secondary ligand, are another common type of indirect detection sensor. Several types of electrochemical biosensors are in use including amperometric devices, which measure the electric current as a function of time while the electrode potential is held constant. Ligands are recognition molecules that bind specifically with the target molecule to be detected. The most important characteristics for ligands are affinity and specificity. Various types of ligands are used in biosensors. Biosensors that use antibodies as recognition elements (immunosensors) are common because antibodies are highly specific, versatile, and bind strongly and stably to the antigen. Several limitations of antibodies are long-term stability, and manufacturing costs, especially for multitarget biosensor applications where many ligands are needed.

Preface

vii

Other types of ligands that show promise for high-throughput screening and chemical synthesis are aptamers and peptides. Aptamers are protein-binding nucleic acids (DNA or RNA molecules) selected from random pools on the basis of their ability to bind other molecules with high affinity. Peptides can be selected for affinity to a target molecule by display methods (phage display and yeast display). However, in general, the binding affinity of peptides is lower than the affinity of antibodies or aptamers.

2. Biosensor Applications Biosensors have several potential advantages over other methods of biodetection, especially increased assay speed and flexibility. Rapid, essentially real-time analysis can provide immediate interactive information to users. This speed of detection is an advantage in essentially all applications. Applications of biosensors include medical, environmental, public security, and food safety areas. Medical applications include clinical, pharmaceutical and device manufacturing, and research. Biosensor-based diagnostics might facilitate disease screening and improve the rates of earlier detection and attendant improved prognosis. Such technology may be extremely useful for enhancing health care delivery in the community setting and to underserved populations. Environmental applications include spill clean-up, monitoring, and regulatory instances. Public safety applications include civil and military first responders as well as unattended monitoring. Food safety applications include monitoring of food production, regulatory monitoring, and diagnosis of food poisoning. Biosensors allow multitarget analyses, automation, and reduced costs of testing. The key strengths of biosensors are the following: • Fast or real-time analysis: Fast or real-time detection provides almost immediate interactive information about the sample tested, enabling users to take corrective measures before infection or contamination can spread. • Point-of-care detection: Biosensors can be used for point-of-care or on-site testing where state-of-the-art molecular analysis is carried out without requiring a stateof-the-art laboratory. • Continuous flow analysis: Many biosensor technologies can be configured to allow continuous flow analysis. This is beneficial in food production, air quality, and water supply monitoring. • Miniaturization: Biosensors can be miniaturized so that they can be integrated into powerful lab-on-a-chip tools that are very capable while minimizing cost of use. • Control and automation: Biosensors can be integrated with on-line process monitoring schemes to provide real-time information about multiple parameters at each production step or at multiple time points during a process, enabling better control and automation of many industrial and critical monitoring facilities.

viii

Preface

3. Aims and Approach The primary aim of this book is to describe the basic types and the basic elements of biosensors from methods point of view. We tried to include manuscripts that represent the major technologies in the field and to include enough technical detail so that the informed reader can both understand the technology and also be able to build similar devices. The target audience for this book includes engineering, chemical, and physical science researchers, who are developing biosensing technologies. Other target groups are biologists and clinicians, who are the users and developers of applications for the technologies. In addition to supporting the research community, the book may also be useful as a teaching tool for bioengineering, biomedical engineering, and biology faculty and students. To better represent the field, most topics are covered by more than one chapter. The purpose of this “redundancy” is to try to include several alternative approaches for the topics, so as to help the reader choose an appropriate design.

4. Chapter Organization This publication is divided into two volumes: Vol. 503 is focused on Optical-Based Detectors and Vol. 504 is focused on Electrochemical and Mechanical Detectors, Lateral Flow, and Ligands for Biosensors. 4.1. Volume 503: Optical-Based Detectors

Optical detection is used in a broad array of biosensor technologies, including both direct and indirect style sensors. Volume 503 is organized in two parts. Part I focuses on direct optical detectors, while Part II concentrates on indirect optical detection. Probably, the most common approach for direct optical detection is based on evanescent wave physics, where the interaction between the evanescent wave and the bound target generates a detection signal. The most common technology in this group is surface plasmon resonance (SPR) and several chapters (see Chaps. 1–5) describe biosensors based on SPR. Other important optical direct detection methods including resonant mirror (see Chap. 6), optical ring resonator (see Chap. 7), interferometric sensors (see Chaps. 8 and 9) and grating coupler (see Chap. 10) are all included in Part I. The second part of Vol. 503 describes various indirect optical detectors. As discussed earlier, indirect detectors require a labeled molecule to bind to the target generating a signal. For optical sensors, the label molecule emits or modifies light. Most indirect optical detectors are designed to measure fluorescence. However, optical detectors can also measure optical density (densitometry), changes in color (colorimetric), and chemoluminesence, depending on the type of label used. Optical signals can be measured in various ways (described in Part II) including various CCD-based detectors, which are very versatile, inexpensive, and relatively simple to construct and use (see Chaps. 11–16 and 25). Other optical detectors discussed in Part II are photodiodes (see Chaps. 17–20), photomultipliers (see Chaps. 21–23),

Preface

ix

and spectrometers (see Chaps. 24 and 25). Photomultipliers may offer higher sensitivity, smaller footprint (the size of photodiode can be few millimeters). Spectrometers offer better interrogation of changes in light wavelengths. 4.2. Volume 504: Electrochemical and Mechanical Detectors, Lateral Flow, and Ligands

Volume 504 describes various electrochemical and mechanical detectors, lateral flow devices, and ligands for biosensors. As in Vol. 503, we describe several direct measurement sensors (in Part I), indirect methods (Parts II–III). Ligands are described in Part IV and two related technologies are described in Part V. In Part I, we describe several mechanical detectors that modify their mechanical properties as a result of biological interactions. Such mechanical direct biosensors typically sense resonance of the mechanical element, which changes when the target molecule binds to the surface. Piezoelectric biosensors (see Chaps. 1–3) employ a technology that is widely used in a variety of applications (e.g., vapor deposition of metals) and is thus readily available and relatively inexpensive. Cantilever-based systems (see Chaps. 4 and 5) can be miniaturized to micrometer dimensions with attendant benefits for system and sample size. In Part II, we describe several electrochemical detectors (see Chaps. 6–11). Electrochemical biosensors were the first biosensors developed and are the most commonly used biosensors today (e.g., glucose monitoring). Part III covers lateral flow technologies (see Chaps. 12–15). Although lateral flow devices are not “classical” biosensors, with ligands and transducers, they are included in this book because of their importance for biosensing. Lateral flow assays are simple immunodetection (or DNA hybridization) devices, which utilize competitive or sandwich assays. They are used mainly for medical diagnostics, including laboratory, home and point-of-care detection. A common format is a “dipstick” in which the test sample diffuses through a porous matrix via capillary action followed by detection by a colorimetric reagent bound to a secondary antibody. The primary antibody is bound to the matrix in a line, and the assay result is a color change at a particular location on the matrix. Lateral flow assays can be dependable and inexpensive. Part IV focuses on recognition ligands, which are key elements in any biosensor (see Chaps. 16–22). The recognition ligands bind specifically with the target molecule to be detected. Various ligands described in Part IV include antibodies, aptamers, and peptides. Antibodies are the most commonly used ligands but advances in selection methods for aptamers (SELEX) and peptides (phage and yeast display) are currently providing alternatives. Part V includes two papers on protein (see Chap. 23) and DNA preparation (see Chap. 24). These papers are relevant to the subject of biosensor technologies but did not fit elsewhere into the book organization outline. References 1. IUPAC Compendium of Chemical Terminology 2nd Edition (1997). (1992), International Union of Pure and Applied Chemistry: Research Triangle Park, NC. 2. Clark LC Jr., Lyons C (1962) Electrode systems for continuous monitoring in cardiovascular surgery. Ann N Y Acad Sci 102:29–45.

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contents of Volume 504. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

v xiii xvii

PART I: OPTICAL-BASED DETECTORS 1.

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced Fluorescence Spectroscopy for Sensitive Detection of Tumor Markers . . . . . . . . Yusuke Arima, Yuji Teramura, Hiromi Takiguchi, Keiko Kawano, Hidetoshi Kotera, and Hiroo Iwata 2. Surface Plasmon Resonance Biosensor for Biomolecular Interaction Analysis Based on Spatial Modulation Phase Detection . . . . . . . . . . . . . . . . . . . Xiang Ding, Fangfang Liu, and Xinglong Yu 3. Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection . . . . . . Jong Seol Yuk and Kwon-Soo Ha 4. Optical Biosensors Based on Photonic Crystal Surface Waves . . . . . . . . . . . . . . . Valery N. Konopsky and Elena V. Alieva 5. Surface Plasmon Resonance Biosensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Marek Piliarik, Hana Vaisocherová, and Ji í Homola 6. Label-Free Detection with the Resonant Mirror Biosensor. . . . . . . . . . . . . . . . . Mohammed Zourob, Souna Elwary, Xudong Fan, Stephan Mohr, and Nicholas J. Goddard 7. Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ian M. White, Hongying Zhu, Jonathan D. Suter, Xudong Fan, and Mohammed Zourob 8. Reflectometric Interference Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Guenther Proll, Goran Markovic, Lutz Steinle, and Guenter Gauglitz 9. Phase Sensitive Interferometry for Biosensing Applications . . . . . . . . . . . . . . . . Digant P. Davé 10. Label-Free Serodiagnosis on a Grating Coupler . . . . . . . . . . . . . . . . . . . . . . . . . Thomas Nagel, Eva Ehrentreich-Förster, and Frank F. Bier

3

21 37 49 65 89

139

167 179 189

PART II: INDIRECT DETECTORS 11. CCD Camera Detection of HIV Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . John R. Day 12. Simple Luminescence Detector for Capillary Electrophoresis . . . . . . . . . . . . . . . Antonio Segura-Carretero, Jorge F. Fernández-Sánchez, and Alberto Fernández-Gutiérrez 13. Optical System Design for Biosensors Based on CCD Detection . . . . . . . . . . . . Douglas A. Christensen and James N. Herron

xi

203 221

239

xii

Contents

14. A Simple Portable Electroluminescence Illumination-Based CCD Detector . . . . Yordan Kostov, Nikolay Sergeev, Sean Wilson, Keith E. Herold, and Avraham Rasooly 15. Fluoroimmunoassays Using the NRL Array Biosensor . . . . . . . . . . . . . . . . . . . . Joel P. Golden and Kim E. Sapsford 16. Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral and Polarization Imagers for Fluorescence and Spectroscopic Imaging . . . . . . . . Neelam Gupta 17. Photodiode-Based Detection System for Biosensors. . . . . . . . . . . . . . . . . . . . . . Yordan Kostov 18. Photodiode Array On-chip Biosensor for the Detection of E. coli O157:H7 Pathogenic Bacteria. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joon Myong Song and Ho Taik Kwon 19. DNA Analysis with a Photo-Diode Array Sensor . . . . . . . . . . . . . . . . . . . . . . . . Hideki Kambara and Guohua Zhou 20. Miniaturized and Integrated Fluorescence Detectors for Microfluidic Capillary Electrophoresis Devices . . . . . . . . . . . . . . . . . . . . . . . Toshihiro Kamei 21. Photomultiplier Tubes in Biosensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yafeng Guan 22. Integrating Waveguide Biosensor. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Shuhong Li, Platte Amstutz III, Cha-Mei Tang, Jun Hang, Peixuan Zhu, Yunqi Zhang, Daniel R. Shelton, and Jeffrey S. Karns 23. Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves and In-Fiber Microchannel Sensors . . . . . . . . Rudi Irawan and Swee Chuan Tjin 24. Multiplex Integrating Waveguide Sensor: Signalyte™-II. . . . . . . . . . . . . . . . . . . Shuhong Li, Yunqi Zhang, Platte Amstutz III, and Cha-Mei Tang 25. CCD Based Fiber-Optic Spectrometer Detection. . . . . . . . . . . . . . . . . . . . . . . . Rakesh Kapoor Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

259

273

293 307

325 337

361 375 389

403 423 435 447

Contributors ELENA V. ALIEVA • Institute of Spectroscopy, Russian Academy of Sciences, Troitsk, Moscow Region, Russia PLATTE AMSTUTZ • Creatv MicroTech, Inc., Potomac, MD, USA YUSUKE ARIMA • Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan FRANK F. BIER • Department of Molecular Bioanalytics & Bioelectronics, Fraunhofer Institute for Biomedical Engineering, Branch Potsdam-Golm, Potsdam, Germany Institute of Biochemistry and Biology, University of Potsdam, Potsdam, Germany DOUGLAS A. CHRISTENSEN • Department of Bioengineering and Department of Electrical & Computer Engineering, University of Utah, Salt Lake City, UT, USA DIGANT P. DAVÉ • University of Texas at Arlington, Arlington, TX, USA JOHN R. DAY • Gen-Probe Incorporated, San Diego, CA, USA XIANG DING • Department of Precision Instruments and Mechanics, Tsinghua University, Beijing, China EVA EHRENTREICH-FÖRSTER • Department of Molecular Bioanalytics & Bioelectronics, Fraunhofer Institute for Biomedical Engineering, Branch Potsdam-Golm, Potsdam, Germany SOUNA ELWARY • Biosensors Division, Biophage Pharma, Montreal, QC, Canada XUDONG FAN • Department of Biological Engineering, University of Missouri-Columbia, Columbia, MO, USA ALBERTO FERNÁNDEZ-GUTIÉRREZ • Department of Analytical Chemistry, Faculty of Sciences, University of Granada, Granada, Spain JORGE F. FERNÁNDEZ-SÁNCHEZ • Department of Analytical Chemistry, Faculty of Sciences, University of Granada, Granada, Spain GUENTER GAUGLITZ • Institute of Physical and Theoretical Chemistry, University of Tuebingen, Tuebingen, Germany NICHOLAS J. GODDARD • School of Chemical Engineering and Analytical Science (CEAS), The University of Manchester, Manchester, UK JOEL P. GOLDEN • Center for Bio/Molecular Science & Engineering, US Naval Research Laboratory, Washington, DC, USA YAFENG GUAN • Department of Instrumentation & Analytical Chemistry, Dalian Institute of Chemical Physics, Dalian, China NEELAM GUPTA • Army Research Laboratory, Adelphi, MD, USA KWON-SOO HA • Department of Molecular and Cellular Biochemistry and Nanobio Sensor Research Center, Kangwon National University College of Medicine, Chuncheon, Kangwon-do, Korea JUN HANG • Creatv MicroTech, Inc., Potomac, MD, USA

xiii

xiv

Contributors

KEITH E. HEROLD • Fischell Department of Bioengineering, University of Maryland, College Park, MD, USA JAMES N. HERRON • Fischell Department of Bioengineering and Department of Electrical & Computer Engineering, University of Utah, Salt Lake City, UT, USA JIrˇÍ HOMOLA • Institute of Photonics and Electronics, Academy of Sciences of the Czech Republic, Prague, Czech Republic RUDI IRAWAN • BioMedical Engineering Research Centre, Singapore-University of Washington Alliance, Nanyang Technological University, Singapore Department of Physics, University of Lampung, Bandar Lampung, Indonesia HIROO IWATA • Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan HIDEKI KAMBARA • Central Research Laboratory, Hitachi Ltd., Tokyo, Japan TOSHIHIRO KAMEI • National Institute of Advanced Industrial Science and Technology (AIST), Ibaraki, Japan RAKESH KAPOOR • Department of Physics, University of Alabama at Birmingham, Birmingham, AL, USA JEFFREY S. KARNS • Environmental Microbial Safety Laboratory, U.S. Department of Agriculture-Agricultural Research Service, Beltsville, MD, USA KEIKO KAWANO • Advanced Software Technology and Mechatronics Research Institute of Kyoto, Kyoto, Japan VALERY N. KONOPSKY • Institute of Spectroscopy, Russian Academy of Sciences, Troitsk, Moscow Region, Russia YORDAN KOSTOV • Center for Advanced Sensor Technology, University of Maryland Baltimore County (UMBC), Baltimore, MD, USA HIDETOSHI KOTERA • Department of Microengineering, Graduate School of Engineering, Kyoto University, Kyoto, Japan HO TAIK KWON • Celltek Co., Ltd., Ansan-si, South Korea SHUHONG LI • Creatv MicroTech, Inc., Potomac, MD, USA FANGFANG LIU • Department of Precision Instruments and Mechanics, Tsinghua University, Beijing, China GORAN MARKOVIc • Institute of Physical and Theoretical Chemistry, University of Tuebingen, Tuebingen, Germany STEPHAN MOHR • School of Chemical Engineering and Analytical Science (CEAS), The University of Manchester, Manchester, UK THOMAS NAGEL • Department of Molecular Bioanalytics & Bioelectronics, Fraunhofer Institute for Biomedical Engineering, Branch Potsdam-Golm, Potsdam, Germany Institute of Biochemistry and Biology, University of Potsdam, Potsdam, Germany MAREK PILIARIK • Institute of Photonics and Electronics, Academy of Sciences of the Czech Republic, Prague, Czech Republic GUENTHER PROLL • Institute of Physical and Theoretical Chemistry, University of Tuebingen, Tuebingen, Germany AVRAHAM RASOOLY • FDA Center for Devices and Radiological Health, Silver Spring, MD, USA, National Cancer Institute, Bethesda, MD, USA

Contributors

xv

KIM E. SAPSFORD • Center for Bio/Molecular Science & Engineering, US Naval Research Laboratory, Washington, DC, USA ANTONIO SEGURA-CARRETERO • Department of Analytical Chemistry, Faculty of Sciences, University of Granada, Granada, Spain NIKOLAY SERGEEV • FDA Center for Devices and Radiological Health, Silver Spring, MD, USA DANIEL R. SHELTON • Environmental Microbial Safety Laboratory, U.S. Department of Agriculture-Agricultural Research Service, Beltsville, MD, USA JOON MYONG SONG • Research Institute of Pharmaceutical Sciences and College of Pharmacy, Seoul National University, Seoul, South Korea LUTZ STEINLE • Institute of Physical and Theoretical Chemistry, University of Tuebingen, Tuebingen, Germany JONATHAN D. SUTER • Biological Engineering Department, University of Missouri-Columbia, Columbia, MO, USA HIROMI TAKIGUCHI • Advanced Software Technology and Mechatronics Research Institute of Kyoto, Kyoto, Japan CHA-MEI TANG • Creatv MicroTech, Inc., Potomac, MD, USA YUJI TERAMURA • Department of Polymer Chemistry, Graduate School of Engineering, Kyoto University, Kyoto, Japan SWEE CHUAN TJIN • Photonics Research Centre, School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore HANA VAISOCHEROVÁ • Institute of Photonics and Electronics, Academy of Sciences of the Czech Republic, Prague, Czech Republic IAN M. WHITE • Biological Engineering Department, University of Missouri-Columbia, Columbia, MO, USA SEAN WILSON • University of Maryland Baltimore County (UMBC), Baltimore, MD, USA Xinglong Yu • Department of Precision Instruments and Mechanics, Tsinghua University, Beijing, China JONG SEOL YUK • Department of Molecular and Cellular Biochemistry and Nanobio Sensor Research Center, Kangwon National University College of Medicine, Chuncheon, Kangwon-do, Korea YUNQI ZHANG • Creatv MicroTech, Inc., Potomac, MD, USA GUOHUA ZHOU • Central Research Laboratory, Hitachi Ltd., Tokyo, Japan HONGYING ZHU • Biological Engineering Department, University of Missouri-Columbia, Columbia, MO, USA PEIXUAN ZHU • Creatv MicroTech, Inc., Potomac, MD, USA MOHAMMED ZOUROB • Biosensors Division, Biophage Pharma, Montreal, QC, Canada

Contents of Volume 504 Preface Contributors Contents of Volume 503

PART I: MECHANICAL DETECTORS 1. 2. 3.

4.

5.

A Set of Piezoelectric Biosensors Using Cholinesterases Carsten Teller, Jan Halámek, Alexander Makower, and Frieder W. Scheller Piezoelectric Biosensors for Aptamer–Protein Interaction Sara Tombelli, Alessandra Bini, Maria Minunni, and Marco Mascini Piezoelectric Quartz Crystal Resonators Applied for Immunosensing and Affinity Interaction Studies Petr Skládal Biosensors Based on Cantilevers Mar Álvarez, Laura G. Carrascosa, Kiril Zinoviev, Jose A. Plaza, and Laura M. Lechuga Piezoelectric-Excited Millimeter-Sized Cantilever Biosensors Raj Mutharasan

PART II: ELECTROCHEMICAL DETECTORS 6.

Preparation of Screen-Printed Electrochemical Immunosensors for Estradiol, and Their Application in Biological Fluids Roy M. Pemberton and John P. Hart 7. Electrochemical DNA Biosensors: Protocols for Intercalator-Based Detection of Hybridization in Solution and at the Surface Kagan Kerman, Mun’delanji Vestergaard, and Eiichi Tamiya 8. Electrochemical Biosensor Technology: Application to Pesticide Detection Ilaria Palchetti, Serena Laschi, and Marco Mascini 9. Electrochemical Detection of DNA Hybridization Using Micro and Nanoparticles María Teresa Castañeda, Salvador Alegret, and Arben Merkoçi 10. Electrochemical Immunosensing Using Micro and Nanoparticles Alfredo de la Escosura-Muñiz, Adriano Ambrosi, Salvador Alegret, and Arben Merkoçi 11. Methods for the Preparation of Electrochemical Composite Biosensors Based on Gold Nanoparticles A. González-Cortés, P. Yáñez-Sedeño, and J.M. Pingarrón

PART III: LATERAL FLOW 12. Immunochromatographic Lateral Flow Strip Tests Gaiping Zhang, Junqing Guo, and Xuannian Wang

xvii

xviii

Contents of Volume 504

13. Liposome-Enhanced Lateral-Flow Assays for the Sandwich-Hybridization Detection of RNA Katie A. Edwards and Antje J. Baeumner 14. Rapid Prototyping of Lateral Flow Assays Alexander Volkov, Michael Mauk, Paul Corstjens, and R. Sam Niedbala 15. Lateral Flow Colloidal Gold-Based Immunoassay for Pesticide Shuo Wang, Can Zhang, and Yan Zhang

PART IV: LIGANDS 16. Synthesis of a Virus Electrode for Measurement of Prostate Specific Membrane Antigen Juan E. Diaz, Li-Mei C. Yang, Jorge A. Lamboy, Reginald M. Penner, and Gregory A. Weiss 17. In Vivo Bacteriophage Display for the Discovery of Novel Peptide-Based Tumor-Targeting Agents Jessica R. Newton and Susan L. Deutscher 18. Biopanning of Phage Displayed Peptide Libraries for the Isolation of Cell-Specific Ligands Michael J. McGuire, Shunzi Li, and Kathlynn C. Brown 19. Biosensor Detection Systems: Engineering Stable, High-Affinity Bioreceptors by Yeast Surface Display Sarah A. Richman, David M. Kranz, and Jennifer D. Stone 20. Antibody Affinity Optimization Using Yeast Cell Surface Display Robert W. Siegel 21. Using RNA Aptamers and the Proximity Ligation Assay for the Detection of Cell Surface Antigens Supriya S. Pai and Andrew D. Ellington 22. In Vitro Selection of Protein-Binding DNA Aptamers as Ligands for Biosensing Applications Naveen K. Navani, Wing Ki Mok, and Yingfu Li

PART V: PROTEIN AND DNA PREPARATION 23. Immobilization of Biomolecules onto Silica and Silica-Based Surfaces for Use in Planar Array Biosensors Lisa C. Shriver-Lake, Paul T. Charles, and Chris R. Taitt 24. Rapid DNA Amplification Using a Battery-Powered Thin-Film Resistive Thermocycler Keith E. Herold, Nikolay Sergeev, Andriy Matviyenko, and Avraham Rasooly Index

Chapter 1 Surface Plasmon Resonance and Surface Plasmon Field-Enhanced Fluorescence Spectroscopy for Sensitive Detection of Tumor Markers Yusuke Arima, Yuji Teramura, Hiromi Takiguchi, Keiko Kawano, Hidetoshi Kotera, and Hiroo Iwata Summary Surface plasmon resonance (SPR), which provides real-time, in situ analysis of dynamic surface events, is a valuable tool for studying interactions between biomolecules. In the clinical diagnosis of tumor markers in human blood, SPR is applied to detect the formation of a sandwich-type immune complex composed of a primary antibody immobilized on a sensor surface, the tumor marker, and a secondary antibody. However, the SPR signal is quite low due to the minute amounts (ng–pg/mL) of most tumor markers in blood. We have shown that the SPR signal can be amplified by applying an antibody against the secondary antibody or streptavidin-conjugated nanobeads that specifically accumulate on the secondary antibody. Another method employed for highly sensitive detection is the surface plasmon field-enhanced fluorescence spectroscopy-based immunoassay, which utilizes the enhanced electric field intensity at a metal/water interface to excite a fluorophore. Fluorescence intensity attributed to binding of a fluorophore-labeled secondary antibody is increased due to the enhanced field in the SPR condition and can be monitored in real time. Key words: Surface plasmon resonance, Immunosensing, Tumor marker, Signal amplification, Polyclonal antibody, Surface plasmon field-enhanced fluorescence spectroscopy.

1. Introduction Surface plasmon resonance (SPR)-based sensing has been used to study interactions between biomolecules (1). The SPR method is a sensitive technique to detect changes in the local refractive index near the surface of a thin metal (typically gold) film (2).

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_1

3

4

Arima et al.

Figure 1 shows an SPR apparatus of Kretchmann configuration (a) and its schematic representation (b). A beam of p-polarized light is used to illuminate the back side of a gold thin film on glass, and the front side of the film faces a solution of interest. When the incident angle exceeds the critical angle, total internal reflection occurs. An evanescent wave is generated on the surface facing the solution, which has a lower refractive index than glass. At a specific incident angle, the evanescent wave of the incoming light is able to couple with the free oscillating electrons (plasmons) in the metal film, and the surface plasmon is resonantly excited. This excitation causes energy from the incident light to be lost to the metal film, resulting in a reduction in the intensity of reflected light (Fig. 2a). The resonance angle is a function of the refractive index at the interface of the metal film and solution. Thus, a shift in resonance angle reflects events at the interface, such as protein adsorption on the surface and antigen–antibody interactions. SPR offers rapid, label-free, and real-time monitoring of binding events between biomolecules.

(a)

10 cm

(b) Photodiode

Lens Photodiode

Biaxial rotation stage Prism Flow cell Protein solution

Lens Iris Beam splitter Glass plate

Iris He-Ne laser (632.8 nm)

Gran-Thomson prism

Fig. 1. a A surface plasmon resonance (SPR) apparatus and b its schematic representation.

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

5

Recently, SPR-based immunoassays for the clinical detection of biomarkers in human blood have been investigated (3, 4). The SPR-based immunoassay detects the specific interaction of a biomarker with an antibody immobilized on the SPR sensor (Fig. 2b). Primary antibodies against a particular biomarker are immobilized onto a gold surface modified with a self-assembled monolayer (SAM) of alkanethiols. A blood sample is brought into contact with the sensor, and the biomarker in blood specifically binds to the antibodies immobilized on the sensor surface. If the concentration of the biomarker in blood is high and its molecular weight is large, the SPR resonance angle shift can be detected easily without further modification. However, the concentrations of most tumor markers are in the range ng/mL–pg/mL (Table 1). Therefore, amplification of SPR signal intensity is needed to detect most tumor markers in clinical samples. Several methods for amplifying SPR signal intensity are shown schematically in Fig. 3. In the method shown in Fig. 3a, the sensor surface is incubated with a solution containing

(a)

(b) Secondary antibody

Reflectance

Primary antibody Tumor marker Self-assembled monolayer

Gold thin film q

Incident light

Reflected light

Incident angle q

Fig. 2. a Relationship between incident angle q and intensity of reflected light before (solid line) and after (dashed line) protein adsorption. For real-time monitoring, the intensity of reflected light is monitored at a fixed angle throughout the measurement (arrow). b Schematic representation of SPR-based sandwich-type immunoassay.

Table 1 Concentrations of some tumor markers Molecular weight

Cut-off level (ng/mL)

Cut-off level (pmol/L)

Tumor marker

Name

PSA

Prostate-specific antigen

34,000

4

118

AFP

α-Fetoprotein

70,590

10

142

CEA

Carcinoembryonic antigen

180,000

5

28

6

Arima et al.

(a)

(b) Antibody against secondary antibody Secondary antibody Secondary antibody Tumor marker

Tumor marker

Primary antibody

(c)

Primary antibody

(d) Biotin-labeled anti-streptavidin antibody Streptavidin-conjugated magnetic beads (50 nm)

Detector for fluorescence Fluorophore-conjugated secondary antibody

Biotin-labeled secondary antibody

Tumor marker Primary antibody

Fig. 3. Signal amplification methods for detection of a minute amount of tumor marker. a Binding of secondary antibodies to tumor marker captured by immobilized primary antibodies. b Binding of polyclonal antibodies to secondary antibody. c Accumulation of streptavidin-conjugated nanobeads and biotin-labeled antistreptavidin antibodies. d Detection of fluorophore-conjugated secondary antibodies by surface-plasmon field-enhanced fluorescence spectroscopy.

secondary antibodies, which bind to the tumor marker previously captured by the primary antibody on the SPR sensor surface. As shown in Fig. 3b, further amplification can be achieved by applying polyclonal antibodies against the secondary antibody (5). Because the immobilization of nanobeads causes a large change in the refractive index at the metal/solution interface, nanobeads are expected to be useful for inducing a substantial shift in the SPR resonance angle (Fig. 3c). In this method, biotin-labeled secondary antibodies are bound to tumor markers, which are trapped by primary antibodies immobilized on a sensor surface. Streptavidin-conjugated nanobeads (50 nm in diameter) and biotin-labeled anti-streptavidin antibodies are alternately layered on the surface via the specific biotin–streptavidin interaction (6). All three of the previously described amplification methods utilize an SPR angle shift caused by changes in the local refractive index near the sensor surface. A fourth interesting method to amplify signal intensity utilizes the strong increase in surface light intensity at the metal/water interface around the SPR resonance angle, i.e., surface plasmon field-enhanced fluorescence spectroscopy (SPFS) (7). In this method, SPFS is employed to detect the

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

7

fluorescence due to binding of a fluorophore-labeled secondary antibody (Fig. 3d). Here, we introduce the methods of SPR-based and SPFSbased immunoassays for the detection of tumor markers.

2. Materials 2.1. Optical System for Surface Plasmon Resonance

1. Hemicylindrical prism (custom-made, diameter: 25 mm, width: 10 mm; Sigma Koki Co., Ltd, Tokyo, Japan). 2. Index-matching fluid (n = 1.515; Cargille Laboratories, Ceder Grove, NJ). 3. Silicone rubber tubes (inner diameter: 0.5 mm, outer diameter: 1 mm) (Fig. 4). 4. Silicone rubber heater regulated by a digital controller (E5EK; Omron Corp., Kyoto, Japan). 5. Peristaltic pump (MP-3 N; Tokyo Rikakikai Co., Ltd, Tokyo, Japan). 6. He–Ne laser beam (l = 633 nm, 05-LHP-151; Melles Griot, Carlsbad, CA). 7. Beam splitter (NPCH-10-6328; Sigma Koki). 8. Gran-Thomson prism (GTPC-08-20AN; Sigma Koki). 9. Photodiode detector (S3590-01; Hamamatsu Photonics K.K., Shizuoka, Japan). 10. Lens (SLB-30-300PM, f = 300 mm; Sigma Koki). 11. Iris (IH-30; Sigma Koki). 12. Optical rail (OBA-500SH; Sigma Koki). 13. Optical table (MB-PH; Sigma Koki).

PVC plate

Silicone rubber Glass plate

12 mm

0.5 mm Hemicylindrical prism

Fig. 4. Assembly of a flow cell.

2 mm

8

Arima et al.

14. Lens (SLB-30-60PM, f = 60 mm; Sigma Koki). 15. Biaxial rotation stage (SGSP-120YAW-W; Sigma Koki), which is operated through an intelligent driver (CSG-522R; Sigma Koki) by homemade software. 2.2. Optical System for Surface Plasmon Field-Enhanced Fluorescence Spectroscopy

The basal components of an SPFS apparatus are the same as for an SPR apparatus (Fig. 1b), except for the addition of a CCD camera on the sensor surface: 1. Laser diode (LD, VHK laser diode module l = 635 nm, 0.95 mW; Coherent, Santa Clara, CA). 2. Lens (SLB-20-25P, f = 25 mm; Sigma Koki, Tokyo, Japan) for collimation of the laser light. 3. Polarizing filter (TS0851-G; Sugitoh, Tokyo, Japan). 4. Neutral density filter (AND-20C-10, T = 10%; Sigma Koki). 5. Guide rail (OBS-200G; Sigma Koki) with appropriate holders. 6. Rotational stage (MM-40θ; Chuo Seiki, Tokyo, Japan). 7. Lens (SLB-20-25P; Sigma Koki, Japan). 8. Photodiode detector (S2281-01; Hamamatsu Photonics, Hamamatsu, Japan). 9. Two-axis stage controller (QT-CM2; Chuo Seiki). 10. Objective lens (SLWD Plan20×; Nikon, Tokyo, Japan). 11. Interference filter (l. = 670 nm, transmittance 75%, full-width half length max (FWHM) = 7 nm; Optical Coatings Japan, Tokyo, Japan). 12. Extension barrel (TS0155-H; Sugitoh, Tokyo, Japan). 13. CCD camera with a charge multiplier (MC681-SPD; Texas Instruments, Dallas, TX) (see Note 1). 14. Image capture board (MT-PCI2; Micro-Technica, Tokyo, Japan). 15. Homemade intensity scanning software (see Note 2). 16. A glass plate (S-LAL10, refractive index: 1.720; Sigma Koki) with a thin gold film (49 nm in thickness). 17. Triangular prism (25 × 25 × 25 mm, S-LAL10; Sigma Koki). 18. Index-matching fluid (n = 1.72, Cargille Laboratories).

2.3. Working Solutions

1. Ethanol is deoxygenated with bubbling nitrogen gas for 20–30 min before use. 2. (1-Mercaptoundecanoic-11-yl)tri(ethylene glycol) (TEG: HS–(CH2)11–(OCH2CH2)3–OH) and (1-mercaptoundecanoic11-yl)hexa(ethylene glycol)carboxylic acid (HEG:

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

9

HS–(CH 2 ) 11 –(OCH 2 CH 2 ) 5 –OCH 2 CH 2 OCH 2 COOH) (SensoPath Technologies, Inc., Bozeman, MT) are dissolved in deoxygenated ethanol solution at 0.9 and 0.1 mM, respectively (see Note 3). 3. α-Fetoprotein (AFP; Morinaga Institute of Biological Science, Inc., Yokohama, Japan) is dissolved in phosphate buffer at 10 µg/mL and stored in single-use aliquots at −80°C. 4. Whole blood is collected by drawing venous blood from healthy donors into Venoject® II blood collection tubes containing EDTA-2 Na (TERUMO Co., Tokyo, Japan). To separate plasma, the tubes are centrifuged at 3,000 × g at 4°C for 30 min. After centrifugation, supernatant is collected and stored at −80°C until use. 5. Working solutions of AFP at prescribed concentrations are prepared by dilution in human plasma (see Note 4). 6. A tablet of PBS-Tween® (10 mM phosphate buffer, 140 mM NaCl, 3 mM KCl, 0.05% Tween® 20; Calbiochem, Inc., Darmstadt, Germany) is dissolved in 1 L pure water (one tablet per 1 L). 7. Phosphate buffer (pH 6.6) is prepared by dissolving 33 mM Na2HPO4 and 33 mM KH2PO4 in pure water at a volume ratio of 2:1. This solution is degassed by a water aspirator and used for sample preparation. 8. Anti-AFP antibody (mouse monoclonal, Clone number: ME-101, affinity = 6 × 109; Abcam Ltd, Cambridge, UK) is dissolved in phosphate buffer at 3.2 mg/mL and stored in single-use aliquots at −80°C. A primary antibody solution is prepared by diluting the solution to 10 µg/mL in phosphate buffer (see Note 5). 9. Lyophilized powder of anti-AFP antibody with various salts (rabbit polyclonal; Monosan Ltd, Uden, Netherland) is dissolved in pure water at 1.0 mg/mL and stored at 4°C. A secondary antibody solution is prepared by diluting the solution to 10 µg/mL in phosphate buffer (see Note 5). 10. Anti-rabbit IgG antibody (goat polyclonal, 1.5 mg/mL; Zymed Laboratories, Inc., South San Francisco, CA) is dissolved at 10 µg/mL in phosphate buffer just before use (see Note 4). 11. A blocking solution is prepared by dissolving bovine serum albumin (BSA fraction V; Sigma-Aldrich, Inc., St. Louis, MO) at 10 mg/mL in phosphate buffer (see Note 6).

10

Arima et al.

3. Methods 3.1. Surface Plasmon Resonance-Based Immunoassay 3.1.1. Setup of SPR Apparatus

Various SPR instruments are commercially available from companies such as Biacore AB and Moritex Corp. The optical construction of an SPR instrument is simple, as shown schematically in Fig. 1b. We assembled an SPR instrument from optical parts as described (8). The glass plate is coupled to a hemicylindrical prism (custom-made, diameter: 25 mm, width: 10 mm; Sigma Koki Co., Ltd, Tokyo, Japan) with an index-matching fluid (n = 1.515; Cargille Laboratories, Ceder Grove, NJ), and the SPR flow cell is set on the glass plate. The flow cell is assembled with silicone rubber and a PVC plate and is connected to silicone tubes (inner diameter: 0.5 mm, outer diameter: 1 mm) (Fig. 4). The temperature of the flow cell is kept constant with a silicone rubber heater regulated by a digital controller (E5EK; Omron Corp., Kyoto, Japan). A peristaltic pump (MP-3 N; Tokyo Rikakikai Co., Ltd, Tokyo, Japan) delivers the liquid sample to the flow cell at the rate of 4 mL/min. A He–Ne laser beam (l = 633 nm, 05-LHP-151; Melles Griot, Carlsbad, CA) is separated into two by a beam splitter (NPCH10-6328; Sigma Koki). One beam is linearly p-polarized using a Gran-Thomson prism (GTPC-08-20AN; Sigma Koki), and the other is guided to a photodiode detector (S3590-01; Hamamatsu Photonics K.K., Shizuoka, Japan) to monitor the fluctuation of incident light intensity. The polarized light is focused by a lens (SLB-30-300PM, f = 300 mm; Sigma Koki) onto the backside of a gold thin film evaporated on a glass plate. The He–Ne laser, iris (IH-30; Sigma Koki), Gran-Thomson prism, beam splitter, and lens are placed on an optical rail (OBA-500SH; Sigma Koki) and fixed on an optical table by optical bases (MB-PH; Sigma Koki). The reflected light passes through a lens (SLB-30-60PM, f = 60 mm; Sigma Koki), and its intensity is measured by a photodiode detector. Reflectance is calculated from the intensities of the incident and reflected light (voltage), which are converted from the current of each photodiode. The sample stage and the detector for reflected light are rotated at intervals of 0.03° using a biaxial rotation stage (SGSP-120YAW-W; Sigma Koki), which is operated through an intelligent driver (CSG-522R; Sigma Koki) by homemade software, to obtain the relationship between incident angle and reflectance (Fig. 2a). To precisely determine a resonance angle, a quadratic function was fitted to an incident angle–reflectance profile ranging from 0.4° lower to 0.1° higher than an apparent minimum in reflectance, and the minimum of the quadratic function was regarded as the resonance angle. An SPR control and analysis program was developed on an integrated development environment using Object Pascal language (Boland Delphi 4 Pro. Jpn. ed.; Inprise Corp., Tokyo, Japan).

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

11

For real-time monitoring of binding processes, the change in reflected light intensity is monitored at a fixed angle (in our system, 0.5° lower than the resonance angle) during a measurement (Fig. 2a, arrows). Finally, the change in reflected light intensity is converted to SPR angle shift by homemade software. The amount of adsorbed protein is determined by the SPR angle shift using the following relationship: The amount of adsorbed protein (ng/cm2) = 500 × increase in resonance angle (degree), where the refractive index and density of protein are assumed to be 1.45 and 1.0 g/cm3, respectively. 3.1.2. Preparation of SPR Sensor

1. Glass plates for SPR (material: BK7, refractive index: 1.515, 25 mm × 25 mm × 1 mm) are purchased from Arteglass Associates (Kyoto, Japan). 2. Piranha solution, a 7:3 mixture of concentrated sulfuric acid and 30% hydrogen peroxide solution, is prepared (see Note 7). 3. BK7 glass plates are immersed in piranha solution for 5 min, rinsed twice with deionized water, rinsed with 2-propanol, and stored in 2-propanol until use. 4. Glass plates are dried with a stream of nitrogen gas. 5. Glass plates are placed in a thermal evaporation apparatus (V-KS200; Osaka Vacuum, Osaka, Japan) (see Note 8). 6. A gold wire (purity: 99.99%, f = 0.5 mm) and a chromium piece (purity: 99.99%) are placed in separate tungsten baskets. 7. The pressure in the evaporation chamber is decreased to less than 3 × 10−4 Pa. 8. A chromium layer of 1-nm thickness is deposited on the glass plate at 0.02 nm/s (see Note 9). 9. A gold layer is deposited on the glass plate at 0.05 nm/s for 4 nm, 0.3 nm/s for 38 nm, and 0.05 nm/s for 7 nm (total thickness: 49 nm) (see Note 10). 10. The gold-coated plates are immersed immediately in the TEG/ HEG mixture (see Subheading 1.2.3) for approximately 24 h at room temperature to form a mixed SAM (see Note 11). 11. The glass plates modified with TEG/HEG-mixed SAM are washed thoroughly with pure water and 2-propanol before use (see Note 12).

3.1.3. Determination of AFP Concentration in Human Plasma

1. A TEG/HEG-mixed SAM-coated glass plate is dried with a N2 gas stream and placed on a hemicylindrical prism with index-matching fluid (n = 1.515; Cargille Laboratories, Cedar Grove, NJ). 2. A flow cell chamber composed of a glass plate with prism, a 0.5-mm-thick silicone rubber spacer, and an upper

12

Arima et al.

plate is assembled and is placed in the SPR instrument (see Note 13). 3. The flow cell chamber and solutions are kept at 30°C (see Note 14). 4. For stabilization of baseline in the SPR instrument, degassed phosphate buffer solution is circulated by a peristaltic pump at 4.0 mL/min for 30 min before the start of measurements. 5. A mixture of 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC; Dojindo Laboratories, Kumamoto, Japan) and N-hydroxylsuccinimide (NHS; Nacalai Tesque, Kyoto, Japan) (stored as a mixed powder) is dissolved in degassed phosphate buffer at 0.1 and 0.05 M, respectively, just before use. 6. The mixture of 0.1 M EDC and 0.05 M NHS in phosphate buffer is flowed for 15 min to activate the COOH groups of the TEG/HEG-mixed SAM, and then the primary antibody solution (10 µg/mL) is immediately flowed for 25 min to achieve covalent immobilization. 7. A BSA blocking solution is flowed for 15 min to block nonspecific adsorption and to deactivate unreacted NHS ester groups on the surface (see Note 15). 8. Human plasma containing AFP (50–500 ng/mL) or a clinical sample is flowed into the SPF flow cell for 30 min. 9. PBS-Tween® solution is flowed for 15 min after flowing of plasma to remove nonspecifically adsorbed proteins (see Note 15). 10. A secondary antibody solution (10 µg/mL, polyclonal) is applied for 30 min, followed by a solution of anti-rabbit IgG antibody (10 µg/mL, polyclonal) for 30 min for the SPR signal enhancement (see Note 16). 11. As control experiments, the same procedures (steps 8–10) are performed in the absence of AFP in human plasma. 12. In this system, sample and buffer solutions (3 mL each) are circulated through the flow cell at 4.0 mL/min by a peristaltic pump. Between the different solution injections, the flow cell is washed with phosphate buffer for at least 5 min, except for between the two injections in step 6 (see Notes 17 and 18). 3.1.4. SPR Sensorgram

The SPR immunoassay is based on the formation of a sandwichtype immune complex with two kinds of antibody (Fig. 1.2b), which is also the basis of the widely used enzyme-linked immunosorbent assay (ELISA). Primary antibody is immobilized onto a SAM surface bearing carboxylic groups via covalent amide bonding by a NHS/EDC coupling method. After blocking treatment

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

13

with BSA or inert materials to prevent nonspecific adsorption of serum proteins, AFP in human plasma and secondary antibody (rabbit IgG) are applied sequentially. The SPR signal shift can be further enhanced by binding of anti-rabbit IgG antibody (polyclonal) (Fig. 1.5; Subheading 1.3.1.3). The concentration of a tumor marker in blood can be determined from the calibration curve, which is obtained from resonance angle shifts for solutions with various concentrations of tumor marker. Amplification of the SPR signal to detect low concentrations of AFP (25 pg/mL–1 ng/mL) can be accomplished using streptavidin-conjugated nanobeads and biotin-labeled antistreptavidin antibodies, according to the method reported for detection of brain natriuretic peptide (BNP) (6). Namely, after flowing of biotin-labeled secondary antibody, streptavidinnanobeads and biotin-labeled antistreptavidin antibody are flowed sequentially, followed by the flowing of 0.05% Tween 20 in PBS solution for the removal of nonspecific adsorbed proteins and beads. These procedures can be repeated to further amplify the SPR signal (see Note 19).

(a)

(b)

EDC / Primary NHS antibody

AFP in plasma

Secondary Tween20 antibody

700

3500

2500 2000

[AFP] = 500 ng / mL Secondary antibody

1500

1440

1000

1430 1420 1410

500

1400

[AFP] = 50 ng / mL

1390

0

1380 153

0

163

173

183

20 40 60 80 100 120 140 160 180

Time (min)

Angle shift value (mDA)

Angle shift value (mDA)

[AFP] = 500 ng / mL

BSA

3000

600 500 400 300

[AFP] = 50 ng / mL

200

Control (no AFP)

100 0

0

5

10

15

20

25

30

35

Time (min)

Fig. 5. SPR-based immunoassay for AFP in human plasma. a SPR profiles of sequential reactions during AFP detection in human plasma containing 50 or 500 ng/mL AFP (5). After immobilization of primary antibody by EDC/NHS and blocking treatment with BSA, plasma containing AFP was perfused. The sensor surface was washed with PBS-Tween®. Then, a solution of secondary antibody was flowed. An SPR signal shift (155 mDA) was clearly observed when the solution of secondary antibody was applied. The inset shows an SPR signal shift for the plasma containing 50 ng/mL AFP. b Amplification of SPR signal using goat polyclonal antibody against the secondary antibody (rabbit IgG). An SPR sensor surface was exposed to plasma supplemented with AFP ([AFP] = 50 or 500 ng/mL) or without AFP addition. The sensor surface was washed with PBS-Tween®, and then a solution of secondary antibody was flowed. The time course of the SPR profile during application of the solution of anti-rabbit IgG antibody was recorded.

14

Arima et al.

3.2. Surface Plasmon Field-Enhanced Fluorescence Spectroscopy-Based Immunoassay

The basal components of an SPFS apparatus are the same as for an SPR apparatus (Fig. 1.1b), except for the addition of a CCD camera on the sensor surface. When the incident angle approaches the SPR angle, the surface electric field intensity at the water/ metal interface strongly increases, as depicted by the dashed line in Fig. 1.6a (7). Peak intensities relative to the incoming intensity can reach an enhancement factor of ~16 for gold and incident light l = 635 nm. This enhanced field can be utilized to excite a fluorophore to detect fluorophore-labeled secondary antibodies in the SPFS-based immunoassay (Fig. 1.6b) (9). The intensity of fluorescence by SPFS is increased relative to that of the conventional total internal reflection due to the enhanced field in the SPR condition.

3.2.1. Setup of SPFS Apparatus

Figure 1.7a is a photo of our SPFS apparatus. The optical construction of an SPFS instrument is simple as shown schematically in Fig. 1.7b. A laser diode (LD, VHK laser diode module l. = 635 nm, 0.95 mW; Coherent, Santa Clara, CA), a lens (SLB-20-25P, f = 25 mm; Sigma Koki, Tokyo, Japan) for collimation of the laser light, a polarizing filter (TS0851-G; Sugitoh, Tokyo, Japan), and a neutral density filter (AND-20C-10, T = 10%; Sigma Koki) are placed on the same guide rail (OBS200G; Sigma Koki) with appropriate holders. The guide rail with these parts is placed on an arm of a rotational stage (MM-40θ; Chuo Seiki, Tokyo, Japan). Optical parts for detection of the reflected light, such as a lens (SLB-20-25P; Sigma Koki, Japan) and a photodiode detector (S2281-01; Hamamatsu Photonics, Hamamatsu, Japan), are placed on a guide rail and attached to an arm of another rotational stage. These rotational stages are operated by a two-axis stage controller (QT-CM2; Chuo Seiki). Reflectance is determined from the intensities of incident and

(a)

(b) 1

0.6 10 0.4 5

0.2 Incident angle, q

0

Field enhancement

Reflectance

15

Fluorophore-conjugated secondary antibody

Primary antibody

Tumor marker

Self-assembled monolayer

θ

0.8

0

Detector for fluorescence

20

Incident light

Gold thin film Reflected light

Fig. 6. Principle of surface plasmon field-enhanced fluorescence spectroscopy (SPFS) and its application to immunoassays. a Reflectance (solid line) and electric field enhancement (dashed line) relative to the incoming intensity as a function of incident angle q at an Au/water interface. b Schematic representation of SPFS-based sandwich-type immunoassay.

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

15

(a)

5 cm

(b)

SPFS PC

Stage controller

CCD camera Interference filter (670 nm) Objective (x20)

Glass plate (S-LAL10, Au:Cr = 49:1 nm)

Flow cell Prism (S-LAL10, n = 1.72)

θ

Polarizing filter LD (635 nm, 0.95 mW)

Photodiode

Lens

SPR Rotational stage

Fig. 7. a A surface plasmon field-enhanced fluorescence spectroscopy apparatus and b its schematic representation.

reflected light (voltage), which are converted from currents of the photodiode detector by the same algorithm used for the SPR measurement apparatus. Fluorescent light from fluorophores on a sensor surface is collected by an objective lens (SLWD Plan20×; Nikon, Tokyo, Japan), passed through an interference filter (l = 670 nm, transmittance 75%, FWHM = 7 nm; Optical Coatings Japan, Tokyo, Japan) fixed in a extension barrel (TS0155-H; Sugitoh, Tokyo, Japan), and then guided to a CCD camera with a charge multiplier (MC681-SPD; Texas Instruments, Dallas, TX) (see Note 1). A fluorescence image is acquired by an image capture board (MT-PCI2; Micro-Technica, Tokyo, Japan) and analyzed by homemade intensity scanning software (see Note 2). The light intensity in a fixed area (100 × 100 pixels) at the center of the laser light is monitored for a certain integration time, and the average intensity is recorded.

16

Arima et al.

The glass plate (S-LAL10, refractive index: 1.720; Sigma Koki) with a thin gold film (49 nm in thickness) is coupled to a triangular prism (25 × 25 × 25 mm, S-LAL10; Sigma Koki) with an index-matching fluid (n = 1.72, Cargille Laboratories). The flow cell illustrated in Fig. 1.4 is assembled on the glass plate. Instead of a hemicylindrical prism and a PVC plate, a triangular prism and a transparent PMMA plate are employed (see Note 20). 3.2.2. Preparation of Working Solutions

1. Phosphate-buffered saline (PBS) consists of 10 mM phosphate buffer with 140 mM NaCl and 3 mM KCl in pure water (pH 7.4). 2. Anti-AFP antibody used as a secondary antibody (1D5, mouse monoclonal; Japan Clinical Laboratories, Inc., Kyoto, Japan) is dissolved at 2.5 mg/mL in PBS and stored at 4°C. 3. Anti-AFP antibody used as a secondary antibody (6D2, mouse monoclonal; Japan Clinical Laboratories, Inc.) is dissolved at 2.5 mg/mL in PBS and stored at 4°C. 4. The secondary antibody is conjugated with Alexa Fluor 647 dye according to the standard protocol of a labeling kit from Molecular Probes (Eugene, OR), and the concentration of conjugate is <1 mg/mL in PBS. 5. Other working solutions are prepared by the methods mentioned in Subheading 1.2.3.

3.2.3. Preparation of SPFS Sensor

1. Glass plates (S-LAL10, refractive index: 1.720, 25 mm × 25 mm × 1 mm) are purchased from Sigma Koki (Tokyo, Japan). 2. S-LAL10 glass plates are cleaned with oxygen plasma using a plasma reactor (PA300AT; O-kuma Engineering, Fukuoka, Japan) under 5 Pa for 1 min (see Note 21). 3. Glass plates with a gold thin layer are prepared, a TEG/HEGmixed SAM is formed on the gold surface, and primary antibodies are immobilized on the mixed SAM by the methods described in Subheading 1.3.1.3, steps 1–7. 4. Immobilization of primary antibody and blocking with BSA are monitored in the SPR mode.

3.2.4. Determination of AFP Concentration in Human Plasma

1. The incident angle is adjusted to the SPR condition by monitoring the intensity of reflected light (see Notes 22 and 23). 2. For one-step measurement of the AFP concentration in plasma, 20 µL of the 1 µg/mL Alexa Fluor 647-labeled secondary antibody solution is added to 4,980 µL of human plasma containing AFP and left for 20 min at room temperature to form an AFPsecondary antibody complex (see Note 24). 3. After incubation, human plasma is introduced into the flow cell at a rate of 1.3 mL/min for 5 min.

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced AFP and secondary antibody mixture

17

PBS-Tween®

250000

Intensity (a.u.)

200000 150000 100000 [AFP] = 10 ng / mL 50000 [AFP] = 1 ng / mL Control (no AFP)

0 0

2

4

6

8

10

Time (min)

Fig. 8. SPFS-based immunoassay for AFP (0, 1, or 10 ng/mL). Primary antibodies were immobilized on a sensor surface as shown in Fig. 5a. After incubation of AFP and the secondary antibody for 20 min, the mixture was infused into the SPFS flow cell. The sensor surface was washed with PBS-Tween®.

4. The surface is rinsed with PBS for 5 min to remove plasma from the flow cell (see Note 25). 5. Fluorescence intensity is monitored during the infusions of plasma and PBS (Fig. 1.8). 3.2.5. SPFS Sensorgram

The SPFS immunoassay is based on the formation of a sandwichtype immune complex with two different antibodies (Fig. 1.7b). Primary antibody is immobilized on the sensor surface by an NHS/EDC coupling method, and the surface is treated with a BSA solution. Immobilization of the primary antibody and blocking with BSA are monitored by SPR as described above. A mixture of the AFP-containing plasma and the fluorophore-labeled secondary antibody is introduced into the flow cell. The fluorescent intensity excited by the enhanced electric field is increased upon introduction of the plasma (Fig. 1.8). After washing for 10 min with PBS, the specific complex bound to the surface remains. The AFP concentration in plasma can be determined from the fluorescence intensity. The SPFS-based immunoassay is a sensitive, straightforward method for detecting tumor markers at low concentrations.

4. Notes 1. A photomultiplier (PMT) or CCD camera is available for the SPFS measurement.

18

Arima et al.

2. Although the intensity scanning software, which is used to determine the average pixel intensity of a fixed area in the fluorescence image, was programmed in our laboratory, the software can be substituted by other software. 3. Carboxylic groups of HEG are accessible without steric hindrance by TEG chains on the surface because the chain length of HEG is longer than that of TEG. 4. Numerous competitive reagents are available from other commercial sources. 5. We have used two kinds of antibodies that are recommended for ELISA. Numerous competitive reagents are available from other commercial sources. 6. Numerous competitive blocking reagents are available from other commercial sources. 7. Piranha solution reacts violently with many organic materials and should be handled with extreme care. 8. Electron beam evaporation may be employed in place of thermal evaporation. 9. A chromium underlayer is needed to improve attachment of the gold layer to the glass plate. A titanium underlayer may also be employed. 10. A 49-nm-thick gold layer is optimal for this wavelength of incident light (633.5 nm). Optimal thickness varies with the wavelength of incident light. 11. Although a SAM of alkanethiol on gold is formed within a few minutes, we usually choose to leave gold substrates in a solution of alkanethiol for approximately 24 h. 12. Unless otherwise stated, all solutions should be prepared in water that has a resistance higher than 18.2 MΩ and a total organic content of less than five parts per billion. This standard is referred to as “water” in the text. 13. SAM-coated glass plates should be used within a few days. Otherwise, the glass plate is stored in a vacuum desiccator. 14. A temperature change in the solution or flow cell also drifts the resonance angle. A temperature increase of 1°C causes a decrease in resonance angle of 0.015°. To improve the signal/ noise ratio, temperatures should be kept constant. 15. The apparent increase in SPR signal during the flowing of BSA and plasma solution is due to an increased refractive index caused by the bulk effect of high concentrations of components. The increase in baseline signal after flowing plasma and washing with PBS-Tween® is caused not only by the binding of AFP to primary antibody but also by nonspecific adsorption of plasma components on the surface.

Surface Plasmon Resonance and Surface Plasmon Field-Enhanced

19

16. Polyclonal antibody binds polyvalently to AFP on the surface, which leads to a large enhancement of SPR signal. 17. When the solution is exchanged, the introduction of air into the circulation tube should be avoided. Otherwise, the baseline is shifted substantially. 18. Activated COOH groups are hydrolyzed and inactivated in the presence of water. Therefore, the sensor surface is washed with phosphate buffer for only 30 s after flowing EDC/NHS. 19. Although the SPR signal can be amplified by repeating this procedure, the background noise level is also increased. 20. When the collimated light is used to illuminates the back side of a gold thin film on a glass plate, a triangular prism is used to adjust the incident angle same over the illuminated area. 21. Because S-LAL10 cannot resist strong acid, the plates must be cleaned by plasma or UV-ozone treatment. 22. Although there is a slight difference between the angle at the maximum enhancement of the electric field and the SPR angle, the angle of the incident light is set at the SPR condition. It is important to precisely set the incident angle at the SPR angle for the SPFS measurement because the intensity of fluorescence is highly sensitive to the electric field intensity. 23. Although the electric field enhancement by the surface plasmons is strong, the excess intensity of the incident light causes photobleaching of the fluorescent dye. For SPFS measurements in this system, the incident light is attenuated to 0.1 mW (~7.85 mW/cm2) with a neutral density filter (10% transmittance; Sigma Koki). 24. It is essential that an excess amount of the secondary antibody for AFP is present in the mixture. 25. As shown in Fig. 1.8, the mixture of AFP and Alexa Fluorlabeled secondary antibody is flowed for 5 min and adsorbed to primary antibody on the surface. After the washing step, the remaining fluorescence reflects the concentration of AFP, and quantitative analysis is possible.

References 1. Homola, J. (2003) Present and future of surface plasmon resonance biosensors. Anal. Bioanal. Chem. 377, 528–539 2. Raether, H. (1988) Surface Plasmons on Smooth and Rough Surfaces and on Gratings. Springer, Berlin 3. Besselink, G. A. J., Kooyman, R. P. H., van Os, P. J. H. J., Engbers, G. H. M., and Schas-

foorta, R. B. M. (2004) Signal amplification on planar and gel-type sensor surfaces in surface plasmon resonance-based detection of prostate-specific antigen. Anal. Biochem. 333, 165–173 4. Chou, S. F., Hsu, W. L., Hwang, J. M., and Chen, C. Y. (2004) Development of an immunosensor for human ferritin, a nonspecific

20

Arima et al.

tumor marker, based on surface plasmon resonance. Biosens. Bioelectron. 19, 999–1005 5. Teramura, Y. and Iwata, H. (2007) Label-free immunosensing for α-fetoprotein in human plasma using surface plasmon resonance. Anal. Biochem. 365, 201–207 6. Teramura, Y., Arima, Y., and Iwata, H. (2006) Surface plasmon resonance-based highly sensitive immunosensing for brain natriuretic peptide using nanobeads for signal amplification. Anal. Biochem. 357, 208–215 7. Liebermann, T. and Knoll, W. (2000) Surface-plasmon field-enhanced fluores-

cence spectroscopy. Colloid Surf. A 171, 115–130 8. Hirata, I., Morimoto, Y., Murakami, Y., Iwata, H., Kitano, E., Kitamura, H., and Ikada, Y. (2000) Study of complement activation on well-defined surfaces using surface plasmon resonance. Colloid Surf. B 18, 285–292 9. Yu, F., Persson, B., Löfås, S., and Knoll, W. (2004) Surface plasmon fluorescence immunoassay of free prostate-specific antigen in human plasma at the femtomolar level. Anal. Chem. 76, 6765–6770

Chapter 2 Surface Plasmon Resonance Biosensor for Biomolecular Interaction Analysis Based on Spatial Modulation Phase Detection Xiang Ding, Fangfang Liu, and Xinglong Yu Summary Surface plasmon resonance (SPR) biosensor is a powerful tool for biomolecular interaction analysis in proteomics research and drug discovery. But when it is used to analyze small molecules, the sensitivity still needs enhancement. Phase detection is a potential solution, for phase changes more abruptly than intensity when SPR is excited. An SPR system is developed based on spatial modulation phase detection (SMPD). In this system, collimated monochromatic light is used to excite SPR, and the phase of the reflected light is spatially modulated to generate an interference pattern. By processing the interference pattern by certain algorithms, instantaneous phase distribution of the whole sensing area can be obtained. Continuously detecting the phase change, the whole process of biomolecular interaction can be recorded in the form of phase change, based on which we can make kinetic analysis and get kinetic parameters. Antibody array chip is tested by this system. Experimental results indicate that this technique is capable of array detection and is more sensitive than intensity detection. Key words: Surface plasmon resonance (SPR), Spatial modulation, Phase detection, Biomolecular interaction, Antibody array.

1. Introduction Present commercialized surface plasmon resonance (SPR) biosensors are mostly based on intensity detection, which takes an advantage of simple system configuration. But due to drifts of the light source, the photoreceiver, and the amplification circuit, the detection precision is degraded, commonly no better than 10−6 refractive index unit (RIU).

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_2

21

22

Ding, Liu, and Yu

In proteomics research and drug discovery, the voice for techniques of small molecule detection is getting louder and louder. Small molecules are usually defined as those whose molecular weights are below several hundred Daltons. When interacting with other molecules, they generate very limited refractive index change, which is difficult to be detected. So it is necessary to enhance the sensitivity of SPR biosensing techniques. There are two solutions: one is to modify the analyzed molecules, for example, nanoparticle modification; the other is to excavate the potential of sensors. As we know, modification to molecules may lead to a lot of problems: poor repeatability, time wasting, influence on molecular activity, etc. An approach to higher sensitivity without modification to molecules is preferable. When SPR is excited, both the reflected light’s intensity and phase vary rapidly, but the phase changes much more abruptly than the intensity. It means that detecting phase of the reflected light instead of its intensity is a simple way to enhance the sensitivity of SPR biosensor. Theoretical analysis indicates that the sensitivity of phase detection is higher than that of intensity detection by 1–2 magnitudes (1). There are many phase detection techniques, including ellipsometry, interferometry, and heterodyne, which have the simplest system configuration and the highest sensitivity among them. Our group developed a SPR biosensing system and used heterodyne method to detect the interaction between ricin and its antibody (2). However, it is difficult to realize array detection by heterodyne. Interferometry can detect the phase distribution over the whole sensor area, and is appropriate for array detection. We performed array detection based on spatial modulation phase detection (SMPD), and obtained phase distribution of the antibody-array chip and phase–time curves of all sensing units. Experimental results indicate that this technique has a high sensitivity in array detection. SPR technique can be used to analyze the interactions of a broad variety of biomolecules. Among all biomolecular interaction model systems, antigen–antibody interaction model system is one of the most common and simple one. So we choose rabbit IgG as an antigen and goat anti-rabbit-IgG IgG (as an antibody), a typical model system which can reflect common characteristics of antigen– antibody interaction, to illustrate the SPR biosensing technique.

2. Materials 2.1. Reagents and Solutions

1. Piranha solution: 7:3 (volume ratio) mixture of concentrated sulfuric acid and 30% hydrogen peroxide. 2. MUA solution: 11-mercapto-undecanoic acid (MUA, Sigma) dissolved in ethanol at 1 mM and stored at room temperature.

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

23

3. Activation solution: 1:1 (volume ratio) mixture of N-hydroxy succinimide (NHS, 0.1 M, Sigma) solution and fresh 1-ethyl-3[3-dimethylaminopropyl]carbodiimide (EDC, 0.4 M, Sigma) solution (see Note 1). 4. Phosphate buffered saline (PBS) solution: 8 mM Na2HPO4, 1.5 mM KH2PO4, 27 mM KCl, and 1.37 M NaCl in highpurity H2O, pH 7.4. 5. Regeneration solution: 0.2 M glycine, pH adjusted to 2.0 with hydrochloric acid. 6. Rabbit IgG solution (Biodee, CN): dissolved in PBS at 1 mg/mL. 7. Goat anti-rabbit-IgG IgG solution (Biodee, CN): dissolved in PBS at 1 mg/mL. 8. Bovine serum albumin (BSA, Amresco) solution: dissolved in PBS at 10 mg/mL. 9. Diiodomethane (CH2I2, Sigma): used as the refractive index oil, stored in the dark. 2.2. Sensing Chip

1. The sensing chip is made of ZF5 glass, the same material as the triangle prism in the Kretschmann configuration. 2. The chip has a size of 12 mm × 12 mm × 1.2 mm, with both surfaces polished (see Note 2). 3. A chromium (Cr) layer with a thickness of 2 nm is evaporated onto one surface of the chip, at a speed of 0.1 nm/s, in 5 × 10−3 Pa vacuum (see Note 3). 4. A 40-nm gold (Au) layer is evaporated onto the chip surface with predeposition of Cr, at a speed of 0.1 nm/s, in 5 × 10−3 Pa vacuum (see Note 4). 5. Chips are stored in airtight container, protected from dirt and moisture.

2.3. Preparation of Antibody Array Chip

1. Before use, the gold surface of the chip should be cleaned thoroughly. First, the chip is dipped in Piranha solution for 5 min, rinsed by pure water, and blown dry. Then it is cleaned ultrasonically in ethanol for 5 min. 2. Immerse the chip in MUA solution over 24 h to form a carboxyl terminated self-assembled monolayer on the gold surface. Then rinse the chip by pure water and blow it dry. 3. Carboxyl groups on the gold surface are activated by immersing the chip in the activation solution for 30 min before they can immobilize antibody molecules. Rinse the chip by pure water and blow it dry. 4. Rabbit IgG solution (1 mg/mL) is manually spotted on the activated carboxyl terminated surface by a pipette. A 2 × 2

24

Ding, Liu, and Yu

array of rabbit IgG spots is fabricated, with spot diameters of about 1 mm. The antibody array chip is incubated at 35°C for 30 min and rinsed by PBS solution thoroughly. 5. The chip is immersed in BSA solution (10 mg/mL) at 35°C for 30 min to block residual activated carboxyl groups, and rinsed by PBS thoroughly. 6. The antibody array chip is glued to the flow cell. 7. Fill the flow cell with PBS solution to keep the chip surface from becoming dry. 2.4. Instrumental Configuration

The configuration of the SPR system is shown in Fig. 2.1, which can be divided into three parts: the incident arm, the SPR sensing configuration, and the reflective arm. The incident arm consists of the following (numbers correspond to those in Fig. 2.1): 1. A He–Ne laser (200-mm length and 50-mm diameter, purchased from Qufu Normal University, CN), which generates frequency-stabilized laser with a beam diameter of about 2 mm, at wavelength of 632.8 nm. 2. A pinhole of 10-µm diameter at the common focus of the lenses in the beam expander, made of aluminum. 3. A beam expander that contains a microscope objective with the focus length of 4.65 mm (Daheng, Inc., CN, GCO-2105) and a lens with the focus length of 30 mm (Daheng, Inc., CN, GCL-0102). 4. A 1/2 wave plate with the diameter of 25.4 mm (Daheng, Inc., CN, GCL-0606). 5. A rectangular diaphragm that has a size of 12 mm × 12 mm.

Fig. 1. Configuration of SPR system based on SMPD.

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

25

The SPR sensing configuration consists of the following: 6. A Kretschmann prism with an equiangular triangle section and three 20 mm × 12 mm side faces, made of ZF5 glass, whose refractive index is 1.740. 7. A sensing chip (details in Subheading 2.2.2). 8. A flow cell (details in Subheading 2.3.3). The reflective arm consists of the following: 9. A Wollaston prism that has a size of 12 mm × 12 mm × 12 mm and a splitting angle of 0.3° (Qufu Normal University, CN, LSP-03). 10. A polarizing prism that has a size of 10 mm × 10 mm × 27 mm (Qufu Normal University, CN, LGP-01). 11. An imaging lens with focal length of 200 mm and diameter of 38.1 mm (Daheng, Inc., CN, GCL-0101). 12. A 1392 × 1040 CCD camera with 4.65 µm × 4.65 µm pixels, which is connected to the computer through a PCI interface (Roper Scientific, Inc., US, PHOTOMETRICS CoolSNAP cf). 13. A computer.

3. Methods 3.1. SPR Biosensing Based on Phase Detection 3.1.1. Principle of SMPD

When SPR occurs, phase of p-polarized light varies abruptly against the refractive index near the gold surface (shown in Fig. 2.2), while phase of s-polarized light remains approximately constant. Hence, it provides an approach to measuring the phase change of p-polarized light by using s-polarized light as the reference. Phase difference between the two light components can not be measured directly. But after passing through a polarizer, their phase difference is connected with the light intensity, defined as: I = I1 + I 2 cos(j), where I is the intensity of interference light, I1 and I2 are determined by the intensities of the p-polarized light and the s-polarized light, and j is their phase difference. Since I1 and I2 are unknown, j cannot be obtained by just detecting I. In SPR system based on SMPD, the Kretschmann configuration (3) is employed. The Kretschmann configuration is the simplest method to excite SPR and has been widely used in commercialized and laboratory SPR systems, shown in Fig. 2.3. In SMPD, the core optical element is a Wollaston prism, splitting the p-polarized component and s-polarized component of the reflected light by a small angle. After passing through a polarizer,

26

Ding, Liu, and Yu

Fig. 2. Phase of p-polarized light against refractive index on SPR sensor surface at different gold film thicknesses. The light wavelength is 632.8 nm. The prism’s dielectric constant is 3.08 and the gold’s dielectric constant is −10.92 + 1.49i.

Fig. 3. The Kretschmann configuration and interference image in SPR system based on SMPD. The reflected light passes through a Wollaston prism and a polarizing prism and forms interference stripes on screen.

the two light components interfere with each other and form interference stripes on screen (see Fig. 2.3), whose intensity is determined by: I (x ,y ) = I 1 + I 2 cos[j (x ,y ) + fx ], where f is spatial frequency of the stripes (see Note 5), x and y are rectangular coordinates on the screen. Collecting the interference images by a CCD camera and processing them with algorithms, phase distribution can be obtained. Several algorithms can be used to calculate the phase difference, such as sinusoidal fitting, correlation, and Fourier transform process (FTP). FTP

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

27

gives spatially resolved phase distribution and is appropriate to SPR array detection (4). 3.1.2. SMPD Applied to SPR Biosensing

Biomolecular interaction is the biomolecules’ adsorption or desorption to the chip surface, which leads to the change in surface mass density on the chip surface. In fact, surface mass density determines the refractive index. In SPR biosensing, phase of the reflected light is extremely sensitive to the refractive index on the chip surface, which implies that it is sensitive to the biomolecular interaction taking place on the sensor surface. Because of the spatial resolved ability of SMPD in detecting phase distribution, different biomolecules can be analyzed on one chip at the same time. Figure 2.4 shows how SMPD is applied to SPR array detection of biomolecular interaction. Several different biomolecules are immobilized on a sensing chip. In SMPD, biomolecules’ adsorption or desorption to the chip surface leads to change in refractive index and thereby displacement of the interference

Fig. 4. The principle of SPR biosensing based on SMPD: (a) Four different samples are spotted on a chip. (b) Biomolecular interaction leads to displacement of interference stripes. (c) Change in phase distribution is calculated by algorithms, giving direct positive and negative results. (d) Phase–time curve of a sample is obtained, which records the whole process of the sample interacting with the analyte.

28

Ding, Liu, and Yu

stripes. Change in phase distribution caused by biomolecular interaction can be calculated by algorithms. It gives direct results whether there is something binding to or dissociating from the chip surface. Continuously detecting the phase change, the whole process of biomolecular interaction can be recorded in the form of phase change, based on which we can make kinetic analysis and get kinetic parameters. 3.1.3. Optics

Functions of the three parts are as follows: the incident arm generates expanded and collimated monochromatic light beam, the SPR sensing configuration excites SPR and provides a place for biomolecular interaction, and the reflective arm introduces spatial interference and collects image data. In the incident arm, polarized light from the He–Ne laser 1 is expanded and collimated by the beam expander 3, and filtered by the pinhole 2. The intensity ratio of p-polarized and s-polarized components of the light is adjusted by rotating the ½ wave plate 4. The rectangular diaphragm 5 makes the beam’s cross section become a rectangular shape. The light beam is projected into the Kretschmann prism to excite SPR, illuminates the sensing chip 7, and is reflected at the glass–gold interface. To eliminate the air gap between the Kretschmann prism 6 and the sensing chip 7, a kind of matching oil (CH2I2, diiodomethane) having the same refractive index as the prism 6 is filled in the interface. The reflected light contains information of biomolecular interaction. In the reflective arm, the Wollaston prism 9 and the polarizing prism 10 cooperate to generate interference between p-polarized light and s-polarized light. The interference pattern is imaged onto the CCD camera 12 by the imaging lens 1. The images are collected and sent to the computer 13 for processing and analysis. When SPR is excited, biomolecular interaction, such as antigen– antibody interaction on the sensing chip 7, leads to great change in the phase of the reflected light. Thus, the SPR interference pattern, which is determined by phase distribution of the reflected light, changes correspondingly to the biomolecular interaction. By collecting the interference images by the CCD camera 13 and processing the data of gray values in the computer 14, we can get quantitative and time-resolved information of biomolecular interaction.

3.2. Mounting and Adjusting the Optic System

Figure 2.5 shows the photo of the optical system. 1. Scribe the base and draw lines parallel to the optic axis, one in the incident arm and the other in the reflective arm. 2. Mount the laser 1 to the base. Ensure that the light beam is parallel to the optical axis (see Note 6). 3. Mount the pinhole 2 and the expander 3. Adjust the distance between the pinhole 2 and the microscope objective in the

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

29

Fig. 5. Photos of the optical system: (a) The incident arm: The dashed line is part of the optical axis. Five white triangles indicate the adjusting screws 1–5 from left to right, respectively. The pinhole within the beam expander and the rectangular diaphragm is adhered to ½ the wave plate so that they are not marked in the picture. (b) The reflective arm: The dashed line is part of the optical axis. The white triangle indicates the adjusting screw 6.

expander 3 by the adjusting screw 1, until the light beam from the expander 3 has a maximum light intensity and a uniform intensity distribution. 4. Adjust the orientation and the position of the lens in the expander 3 by the adjusting screws 2 and 3 and make the light beam a collimated beam and parallel to the optical axis. 5. Mount the ½ wave plate 4 and two reflectors (the orientation of reflector 1 is fixed and that of the reflector 2 can be adjusted). 6. Mount the prism holder to the base. 7. Mount two reflectors (one reflector’s orientation is fixed and that of the other’s can be adjusted) in the reflective arm, and rotate the reflector 4 to make the light beam parallel to the optical axis. 8. Mount the Wollaston prism 9, the polarizing prism 10, the imaging lens 11, and the CCD camera 12 to the base according

30

Ding, Liu, and Yu

to the screw holes in the base. Connect the CCD camera to the computer with cables. 3.3. The Fluidic System

1. The fluidic system consists of an injecting system, a four-way valve, a flow cell, and a tubing system, as shown in Fig. 2.6 (see Note 7). 2. The injecting system involves a homemade injecting device and four syringes (two 10-mL syringes and two 2.5-mL syringes, only three of them are used in the experiments). 3. The four-way valve is made of polymethyl methacrylate (PMMA) with three inlets and an outlet. 4. The structure of the flow cell in shown in Fig. 2.7. The flow cell is made up of two parts: a flat plate and a plate with flow channels processed in it. The two parts of the flow cell are made of PMMA and adhered together by CHCl3 (trichloromethane). The flow cell has a volume of 30 µL.

Fig. 6. A photo of the fluidic system.

Fig. 7. The flow cell. The whole flow cell consists of three parts: the chip, the upper part of the flow cell (a flat PMMA plate), and the lower part of the flow cell (a PMMA plate with channels processed in it).

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

31

5. The chip is adhered to the flow cell by 502glue, with the gold surface toward the inside of the cell. 6. The tubing system is made of polytetrafluoroethylene (PTFE) tube with the inner diameter of 0.7 mm. 3.4. Experimental Procedures

1. Before experiments, fill the syringes with PBS solution, goat anti-rabbit-IgG IgG solution (1 mg/mL), and regeneration solution, respectively. 2. Clean the glass surface of the sensing chip with ethanol. Add about 2-µL diiodomethane onto the glass surface of the chip and mount the flow cell to the Kretschmann configuration (see Note 8). 3. Rotate the reflector 2 by the adjusting screw 4 in front of the prism to adjust the incident angle to excite SPR and observe the intensity of the reflected light by CCD. The resonant angle is slightly lower than 60°, corresponding to the minimum intensity of the reflected light. 4. Adjust the prism height by the adjusting screw 5 in the prism holder and rotate the reflector 4 by the adjusting screw 6 corresponding to the adjustment to the incident angle. 5. Rotate the polarizers and adjust the exposure time of the CCD to get interference image with good contrast. 6. Start to collect interference images at a speed of 0.5 frame/s. 7. When starting to collect images, inject PBS solution into the flow cell continuously, lasting 80 s at a flow rate of 200 µL/min. 8. Inject goat anti-rabbit-IgG IgG solution slowly to form antigen–antibody complex, lasting 300 s at a flow rate of 200 µL/min. 9. Inject regeneration solution to dissociate the antigen–antibody complex, lasting 100 s at a flow rate of 200 µL/min. 10. Inject PBS solution for 120 s at a flow rate of 200 µL/min. 11. Stop collecting images and save data in computer.

3.5. Signal Processing and Result Analysis

1. Use FTP algorithm to calculate the phase distribution of the reflected light (see Note 9). 2. Subtract the phase distribution at t = 1 s from the phase distribution at t = 380 s to get the phase change caused by goat anti-rabbit-IgG IgG binding to rabbit IgG, shown in Fig. 2.8 (see Note 10). In Fig. 2.8, four white regions can be distinguished obviously, representing four rabbit IgG spots. The color bar has the unit of degree. Higher phase change value indicates that there are more molecules binding to the chip surface. Maximum phase change is higher than 12°,

32

Ding, Liu, and Yu

while background (BSA, dark region) displays fairly low noise level, about 2–3°. 3. Calculate the phase–time curves of rabbit IgG spots and background (see Note 11). Figure 2.8 shows the phase distribution after rabbit IgG binds to goat anti-rabbit-IgG IgG, with four sample areas S1, S2, S3, and S4 (rabbit IgG) and two reference areas R1 and R2 (BSA) (see Note 12). 4. Phase–time curves of sample area A and background area B are shown in Fig. 2.9a. Background signal is generated by bulk refractive change, nonspecific binding, and highfrequency noise. Directly subtracting the background signal from the sample signal, phase change generated by bulk refractive change and nonspecific binding is completely removed from the signal of sample area, and high-frequency noise can also be eliminated to a great extent. The phase resolution of the system is higher than 0.2°, equivalent to 3 × 10−5 RIU in refractive index (see Note 13). Figure 2.9b shows the processed signal

Fig. 8. Phase distribution after rabbit IgG binds to goat anti-rabbit-IgG IgG, with four sample areas S1, S2, S3, and S4 (rabbit IgG) and two reference areas R1 and R2 (BSA).

Fig. 9. Phase curves of the sample area S2 (rabbit IgG) and the reference area R1 (BSA): (a) Original phase curves. (b) Phase curve after background is subtracted.

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

33

Fig. 10. Curves of sample signals minus background signals.

by subtracting the background signal from the sample signal. From 80 to 380 s, antigen–antibody binding process is revealed clearly by the ascending phase curve, and the phase signal rapidly drops to 0° when regeneration solution is injected. 5. Figure 2.9 shows the phase–time curves of four sample area minus two reference areas, respectively. In Fig. 2.10, x-axis is the time (unit: second) and the y-axis is the phase change (unit: degree). Curves of these areas vary almost consistently (see Note 14). 6. As for kinetics analysis, goat anti-rabbit-IgG IgG at a series of concentrations is tested to repeat the above experiment cycle and get a group of phase–time curves. By fitting these curves with proper exponential function, association rate constant, dissociation rate constant, and affinity constant can be obtained. 7. In practical application, a number of different biomolecules can be analyzed at the same time by spotting them onto one chip (see Note 15).

4. Notes 1. EDC powder is stored at −20°C and protected from moisture. Make sure that only fresh solution is used because EDC is moisture sensitive and quickly hydrolyzed in water. 2. Herein surface refers in particular to the surface of the chip with size of 12 mm × 12 mm, and the same below.

34

Ding, Liu, and Yu

3. Predepositing a Cr layer makes the gold layer adhere to the chip surface firmly. 4. The thickness of the gold layer is chosen to be 40 nm in order to get a high sensitivity and a large dynamic range (2). 5. The spatial frequency f is the reciprocal of the period of the interference stripes d, f = 1/d. The period d can be theoretically calculated by equation d = a /l, in which a is the splitting angle of the Wollaston prism and l is the light wavelength. In this system, a is 0.3° and l is 632.8 nm. So the calculated period of the interference pattern is 120.9 µm, equivalent to the width of 26.0 pixels in a line on the CCD camera. The experimental period value d can also be obtained by fitting the interference pattern in horizontal direction with ideal sinusoidal function, and the fitted period value is about 24.9 pixels, which is consistent with the theoretically calculated value. The tiny difference between the theoretically calculated value and experimentally obtained value may originate from the manufacturing error in the Wollaston prism’s splitting angle. 6. Herein parallel to the optical axis means parallel to the base and simultaneously parallel to the scribed line on the base. All parts and components are mounted to the base through screws holes in the base. 7. In order to get a better repeatability of the flow rate and automatically handle the fluid injection, commercialized automatic injecting system can be substituted. Commercialized valves and tubes can also be found. 8. Cleanness of the glass surface is very important since any impurity even moisture remaining between the prism and the chip could intervene in the optical path and influence the accuracy of the results. It is also necessary to eliminate the air bubble between the prism and the chip. 9. The spatial frequency of the interference stripes is obtained by fitting the gray values with standard sinusoidal curve, which is consistent with the theoretically calculated spatial frequency. Parameters in the algorithm such as frequency shift and cut frequency are determined according to the fitted spatial frequency value. 10. The phase distribution remains invariable from t = 1 to 80 s. Antigen–antibody complex begins to form at t = 80 s and ends at t = 380 s. Only phase change tells the biomolecular binding or dissociation. Absolute phase value makes no sense since it is determined by the light beam’s inherent phase distribution and the optical system. 11. The phase–time curve is calculated by averaging the phase values in an area of 100 × 100 pixels in the sample area or background area. In practice, the size of the averaged area is

Surface Plasmon Resonance Biosensor for Biomolecular Interaction

35

determined by the spot size when fabricating the array chip. The larger the area used to calculate the average value, the higher precision will be obtained. 12. It is better to choose a background area that is close to the sample area in respect that it gives better noise reduction performance when subtracting the background signal from the sample signal. 13. The phase resolution of the system is determined by NaCl solution experiments in which NaCl solution of different concentrations is tested. The lowest concentration of the NaCl solution that can be detected is 0.02%, which generates 0.2° phase change. In form of refractive index, the detection limit is as low as 3 × 10−5 RIU. As in protein–protein interaction analysis, the system can detect 3 ng/cm2 of absorbed protein on the chip surface. 14. Curve S1-R1 shows a higher noise level comparative to the other curves, which is generated by optical noise and can be avoided by system improvement. 15. In SMPD, several periods of interference stripes are needed to calculate the phase change. Thereby, there is a limit to the number of biomolecules spotted onto one chip, typically not more than 20.

Acknowledgments This research is supported by the National Natural Science Foundation of China (NSFC, Grant No. 30727001) and the Specialized Research Fund for the Doctoral Program of Higher Education (SRFDP, Grant No. 20050003056).

References 1. Yu X., Wang D., and Yan Z. (2003) Simulation and analysis of surface plasmon resonance biosensor based on phase detection. Sensors and Actuators B 91, 285–290 2. Yu X., Zhao L., Jiang H., Wang H., Yin C., and Zhu S. (2001) Immunosensor based on optical heterodyne phase detection. Sensors and Actuators B 76, 199–202

3. Kretschmann E., and Raether H. (1968) Radiative decay of non-radiative surface plasmons excited by light. Zeitschrift fur Naturforschung 23A 2135–2136 4. Liu J., Ding X., Yu X., and Wang D. (2005) Data analysis of surface plasmon resonance biosensor based on phase detection. Sensors and Actuators B 108, 778–783

Chapter 3 Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection Jong Seol Yuk and Kwon-Soo Ha Summary Spectral SPR biosensor is a useful system for a rapid analysis of protein arrays, as the biosensor with a fiber optic spectrometer can be easily aligned with the reflected light from protein arrays. The spectral SPR biosensor was constructed by Kretschmann geometry, based on the wavelength interrogation with various modules such as protein arrays, optical unit, programs for data acquisition and processing, and a motorized x–y stage for scanning. Protein arrays consist of glass/gold film/linker layer/protein/air. The surface of gold arrays was modified with poly(diallyldimethylammonium chloride) and 11-mercaptoundecanoic acid to immobilize mumps virus. The virus arrays were used to analyze anti-mumps virus in a buffer or human serum by the line-scanning mode of the spectral SPR biosensor. The array-based spectral SPR biosensor has a strong potential for a high-throughput serodiagnosis of infectious diseases. Key words: Spectral SPR biosensor, Line scanning mode, Protein array, Mumps virus, Serodiagnosis, Resonance wavelength, PDDA.

1. Introduction Surface plasmon resonance (SPR)-based biosensors have been widely used for biomolecular interactions, as the optical method allows real-time measurement of biomolecular interactions without labeling and the device is simple (1–3). The SPR method is an optical technique that uses the evanescent wave to measure changes in a refractive index on the sensor surface (4, 5). Biomolecular interactions have been intensively investigated by three types of SPR-based biosensors such as angular (angular interrogation-based) SPR biosensors, spectral (wavelength

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_3

37

38

Yuk and Ha

interrogation-based) SPR biosensors and surface plasmon microscopy biosensors (2, 6, 7). Spectral SPR biosensors are based on wavelength interrogation and scan wavelengths at a constant incidence angle to analyze biomolecular interactions (8). Recently, we have demonstrated the possible use of the spectral SPR biosensor in a high-throughput analysis of protein interactions and serodiagnosis in an array format (9, 10). In this chapter, system integration of the spectral SPR biosensor and analysis of protein arrays are introduced. Protein arrays were prepared by immobilizing mumps virus onto the modified surface of gold arrays and analyzed by the line-scanning mode of the spectral SPR biosensor.

2. Materials 2.1. Optical Components

1. 10–100W Series Q quartz tungsten halogen (QTH) Source (#66892) (ORIEL, Stratford, CT) (see Note 1). 2. Achromatic lenses (#42540, #42570, #42650) (ORIEL, Stratford, CT). 3. Maximum reflection flat mirrors (#44152) (ORIEL, Stratford, CT). 4. Iris diaphragms (#62030, #71400) (ORIEL, Stratford, CT) ). 5. Glan-Taylor polarizing prism (#03 PTA 401) (Melles Griot, Carlsbad, CA). 6. Precision beamsteering assembly and StableRod™ horizontal mounting platform with fine height adjustment (#07 BSF 503, #07 BSB 003, #07 DSS 511) (Melles Griot, Carlsbad, CA). 7. Rotation stage (#07 TRT 507) (Melles Griot, Carlsbad, CA). 8. Precision flexure mirror mount (#07 MFM 513) (Melles Griot, Carlsbad, CA). 9. Right angle UV fused silica prism (#46161) (ORIEL, Stratford, CT). 10. Optical table (#10770) (ORIEL, Stratford, CT).

2.2. Fiber Optic Spectrometer

1. Fiber optic spectrometer (Ocean Optics, Dunedin, FL) (see Note 2). 2. Fiber optic cable with SMA termination for the fiber optic spectrometer (Ocean Optics, Dunedin, FL). 3. ADC2000-PCI+ A/D Converter for PCI-bus (Ocean Optics, Dunedin, FL).

Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection

2.3. Motorized Translational Stage

39

1. Motorized translation stages (40 mm traveling range with a resolution of 20 µm) (#MTS-40) (Cheung Won Mechatronics, Korea) (see Note 3). 2. PCI-GPIB, NI-488.2 for Windows Me/9x (#778032-01) (National Instruments, Austin, TX).

2.4 .Gold Arrays

1. Two types (metallic and Cr) of masks for preparing gold arrays with 1 or 1.5 mm spots (e.g. 5 × 10, 7 × 16 and etc). Metallic (stainless still) and Cr masks are prepared by a drill machine and the traditional semiconductor technology, respectively. 2. Gold arrays are prepared by depositing Ti/Au (5/45 nm) onto pyrex glasses with the masks by an RF-magnetron sputtering apparatus at a vacuum of 3 × 10−6 Torr (see Note 4). 3. Cleaning solution of gold arrays: H2O2/NH4OH/ milliQ water (1:1:5, v/v).

2.5. Antigen Arrays

1. Mumps virus (strain 98-10) (the Kangwon National University College of Medicine, Korea): was prepared in VeroE6 cells (American Type Culture Collection, CRL 1585). 2. Poly(diallyldimethylammonium chloride) (PDDA) (Sigma, St. Louis, MO), Mw ~100,000: 1 mg/mL solution in 0.5 M NaCl. 3. poly L-lysine: 1 mg/mL solution in milliQ water. 4. 11-mercaptoundecanoic acid (MUA) (Sigma, St. Louis, MO): 1 mM solution in ethanol. 5. Phosphate-buffered saline (PBS): 8.1 mM Na2HPO4, 1.2 mM KH2PO4, 2.7 mM KCl and 138 mM NaCl, pH 7.4. 6. 0.1% Tween 20 in PBS. 7. 3.7% formaldehyde in PBS. 8. 0.2% Triton X-100 in PBS.

2.6. Antibodies

1. The hybridoma cell line producing a monoclonal antibody against mumps virus was established by fusion of Sp2/0-Ag14 mouse myeloma cells (ATCC, CRL-1581) with spleen cells of Balb/c mice immunized with mumps virus strain 98–40 (11) (see Note 5). 2. Monoclonal anti-GST (Boditech, Korea).

3. Methods The spectral SPR biosensor is very useful for a high-throughput analysis of biomolecular interactions and serodiagnosis in

40

Yuk and Ha

an array format, as the method is simple and label-free. The array-based spectral SPR biosensor consists of various modules such as protein arrays, signal acquisition, motion control, and data processing. The schematic diagram and optical components of the SPR system are shown in Figs. 1 and 2, respectively. Gold arrays are modified with octadecyltrichlorosilane to prevent reaction solutions from being mixed between spots during incubation, and protein arrays are prepared by immobilizing mumps virus onto the modified surface of gold arrays (10). To obtain reproducible results, it is essential to perform beam alignment to obtain a parallel and p-polarized light, to determine an incidence angle above the critical angle for the total internal reflection, considering the sensitivity of the system and the dynamic ranges of a spectrometer, and to calculate the resonance wavelength with a polynomial curve fitting technique (5, 8). Protein arrays are analyzed in the line-scanning mode of the spectral SPR biosensor and the obtained results are displayed by a color spectra (10).

Fig. 1. Schematic diagram of array-based spectral SPR biosensor. Protein arrays consist of various proteins on the gold surface. Collimated white light enters the polarizer and passes through the right angle prism contacted with a thin Au metal (45 nm thick) of protein arrays at the fixed angle of qsp. The reflected light is collected into a fiber optic spectrometer. Scanning, data acquisition and data processing are performed by self-developed LabVIEW™ software. The arraybased spectral SPR biosensor can be used for various biological applications such as profiling of protein expression, diagnosis of diseases and proteome researches.

Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection

41

Fig. 2. Optical components of the spectral SPR biosensor. Array-based spectral SPR biosensor is configured by the Kretschmann geometry of the attenuated total reflection method. A polarizer is positioned at the input light path to obtain transverse magnetic polarized light. Reflected light from the protein arrays is collected into an optical fiber and analyzed by the fiber optic spectrometer. (a) A schematic diagram. (b) A photograph.

3.1. Integration of Spectral SPR Bio sensor; Optical Components, a Fiber Optic Spectrometer and Motorized Translation Stages

1. These instructions assume the use of a white light lamp, a fiber optic spectrometer and motorized translation stages. A 20 W quartz halogen lamp is installed in a lamp housing (see Note 6). Figure 2 shows the components of the spectral SPR biosensor. 2. Iris diaphragms are installed in front of the light source to obtain parallel light.

42

Yuk and Ha

3. Achromatic convex lenses are located in front of iris diaphragms by monitoring the status of collimated light on the screen (see Note 7). 4. To obtain p-polarized light, a Glan-Taylor polarizing prism is positioned in front of a beam steering device. The beam steering device consists of a horizontal mounting platform and a rotation stage combined with a precision flexure mirror mount. The rotation stage is used to change an incidence angle. 5. Incidence angle is set at 48° using the beam steering device. 6. A PCI-GPIB card is installed in the PC slot to communicate with a motion controller. 7. Motorized translation stages are installed on the optical table and connected with the GPIB cable located between the PCIGPIB card and the motion controller. The motorized translation stage is used to move the sample holder. 8. A gold array is put onto the translation stage (see Note 8). 9. A fiber optic spectrometer is installed to collect the reflected light from the gold arrays with the same incidence angle (see Note 9). 3.2. Calculation of a Resonance Wavelength; Polynomial Curve Fitting

1. A wavelength range of the experimental SPR spectrum is selected to calculate a resonance wavelength. Generally, the wavelength range below 30% of maximal intensity is selected. 4

2. The fourth polynomial curve fit technique such as y = å c j x j , j =0

which is a matrix type of Ac = y, is selected. x is a wavelength and y is the intensity of experimental SPR spectrum. 3. The polynomial fit coefficient (c) is calculated with the linear algebra c = A−1·y. 4. The fourth polynomial curve fit is evaluated with the calculated coefficients and the index of a minimal value is calculated with a polynomial curve fitted SPR spectrum. 5. The index of a minimal value of the polynomial curve fitting data is converted into a resonance wavelength. 6. The best polynomial fitting data is compared with an experimental SPR spectrum. 3.3. Interface and Data Analysis

1. When the hardware is installed into the computer to interface between the devices and the software, it is essential to prevent electrostatic discharge from damaging the device and the components by using a grounding strap or by holding a conduction material such as the computer chassis. 2. GPIB driver software for LabVIEW™ is installed in the computer. 3. Power is turned off and the computer is unplugged.

Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection

43

4. Any static electricity, that might be on one’s clothes or body, is discharged by touching the metal part of the power supply case inside the computer. 5. The GPIB card-edge connector is lined up with the expansion slot receptacle. 6. The computer is plugged in and the installation is verified. 7. The Pprocedure (1–6) is repeated until an SPR spectrum is acquired with the fiber optic spectrometer with ADC2000PCI+ A/D Converter for PCI-bus (see Note 9). 8. Data acquisition is performed by the fiber optic spectrometer when reflected light is collimated into the optical fiber entrance. 9. Resonance wavelength is calculated by the fourth-order polynomial curve fitting method (see Subheading 3.2). 10. Statistical analysis of protein array is performed by the line scanning mode (see Subheading 3.7). Each protein array is scanned every 100 µm along the central line by moving the x–y translation stage, which is systemically controlled by the motion controller. Average resonance wavelengths and standard deviations of each spot are calculated by the collected SPR data. 3.4. Calibration of SPR Spectrum

1. The SPR spectrum is very sensitive to the changes of the refractive index on the metal surface and so the spectrum is easily affected by various experimental environments. It is important to find an exact resonance wavelength from the experimental SPR spectrum to monitor biomolecular interactions. To obtain an exact resonance wavelength, the experimental SPR spectrum is calibrated with a theoretical SPR spectrum when the QTH lamp is replaced. 2. A resonance wavelength of 577 nm is calculated with 3-phase Fresnel’s equations (fused silica prim/Au (45 nm) film/air) at the incidence angle of 48° (12). 3. An exact incidence angle is adjusted to find the resonance wavelength of 577 nm for the 45 nm-thick Au coated chip with the beam steering device at the incidence of angle 48° (see Note 10).

3.5. Surface Modification of Gold Arrays

1. The Au arrays are washed with the cleaning solution at 70°C for 10 min (see Note 11). 2. The arrays are incubated with a mixture of cyclohexane/ carbon tetrachloride/octadecyltrichlorosilane (40:10:0.08, v/v) to generate a hydrophobic glass surface (hydrophobic wells) at 45°C for 30 min. 3. The arrays are then washed with a mixture of cyclohexane and carbon tetrachloride (4:1, v/v), carbon tetrachloride and ethanol in order.

44

Yuk and Ha

4. The arrays are again cleaned with the cleaning solution at 70°C for 10 min. 5. Then, the Au arrays are incubated with 1 mM MUA in ethanol for 16 h and washed with ethanol to remove the excess MUA. 6. The gold arrays are incubated with 50 mM NaOH in milliQ water for 10 min to make the MUA surface negative. 7. The modified arrays are incubated with 1 mg/mL PDDA in 0.5 M NaCl or 1 mg/mL poly L-lysine in milliQ water for 30 min and then washed with milliQ water. 3.6. Preparation of Protein Arrays

1. Mumps virus is immobilized on the PDDA surface of gold arrays at 37°C for 2 h. 2. The virus arrays are incubated with 3.7% formaldehyde and 0.2% Triton X-100 each for 30 min (see Note 12). 3. The virus arrays are washed with PBS and incubated with 10 mg/mL of bovine serum albumin in PBS for 30 min to reduce nonspecific interactions. 4. The virus arrays are washed with 0.1% Tween 20 in PBS. 5. Then, the virus arrays are incubated with various concentrations of anti-mump virus or anti-GST antibody in PBS or human serum for 2 h at 37°C. For analysis of mumps virus infection, human sera are diluted ten times with PBS and applied to the virus arrays for 2 h at 37°C. 6. The virus arrays are washed with 0.1% Tween 20 in PBS and milliQ water, and dried under N2 gas. 7. The arrays are analyzed by the line-scanning mode of the spectral SPR biosensor to determine the amount of antimumps virus in the samples that has interacted with the virus arrays (see Note 13).

3.7. Analysis of Protein Arrays by the Line Scanning Mode of the Spectral SPR Biosensor; Display by Color Spectra

1. Protein arrays are analyzed by the line-scanning mode of the spectral SPR biosensor and the resulting resonance wavelengths are displayed by the color spectra. Protein arrays are horizontally divided into a plurality of pixels. The pixels are determined by the resolution of the stage and the diameter of the optical fiber. 2. Protein arrays are sequentially scanned by moving an x–y motorized translation stage. 3. Reflected spectra are sequentially acquired and SPR spectra are selected by considering the spectral characteristics of dip shape. 4. Resonance wavelengths are calculated by the fourth polynomial curve fitting method. 5. The net shift of each resonance wavelength is sequentially encoded into a one-dimensional array pixel, which is assigned

Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection

45

Fig. 3. High-throughput analysis of anti-mumps virus antibody in human serum on the viral arrays by the spectral SPR biosensor. Mumps virus arrays were prepared by immobilizing mumps virus on the PDDA surface of arrays. The virus arrays were incubated with the indicated concentrations of monoclonal antibodies against mumps virus and glutathione in human serum for 2 h at 37°C. Following washing, the arrays were analyzed by the line-scanning mode of the spectral SPR biosensor and the shift of resonance wavelengths was were expressed by color spectra (a). The results are expressed as means ± S.D. from three independent experiments (b).

to an appropriate color from the GRB color tool bar in proportion to the wavelength. 6. This procedure is repeated until the end of a plurality of pixels (7) (see Note 14). Figure 3 shows analysis of anti-mumps virus in human serum by the line-scanning mode of the spectral SPR biosensor.

4. Notes 1. A 20 W QTH lamp was used since as it provides a stable spectral irradiance curve. 2. The spectrometer used was an S2000 series with a spectral range of 500–700 nm, 25 µm slit and 1,200 lines/mm.

46

Yuk and Ha

3. A two-axis controller of the stage is connected with a GPIB card and controlled by IEEE-488.2 command. 4. The surface of the pyrex glass is required to cleaned before depositing Ti/Au films. 5. The specificity of the monoclonal antibody is established by immunofluorescence staining and immunoblotting. 6. When the lamp is installed, it is required to handled the lamp with poly-vinyl gloves (powder-free). 7. Generally, several achromatic convex lenses are required to obtain collimated parallel light. If the filament shadow of the quartz tungsten halogen lamp appears during beam collimation, beam focusing is inappropriate for obtaining collimated light. This process is called “alignment of the optics”. 8. A sample holder is designed to install protein arrays contacted with a prism. 9. To obtain an SPR spectrum with LabVIEW™ software, the OOILVD LabVIEW driver (32-bit) (Ocean Optics, Dunedin, FL) is used. 10. If the shape of the SPR spectrum is not similar to that of inverse Gaussian distribution, it is necessary to align the optical lens system again. If the intensity of the SPR spectrum is over 60%, it is necessary to check the thickness of gold arrays. 11. The quality of gold arrays is controlled by checking the gold surface after cleaning the arrays. 12. To determine the titer of mumps virus in solution, virusimmobilized arrays are analyzed without incubating with 3.7% formaldehyde or 0.2% Triton X-100. 13. The protein arrays can be recycled several times after cleaning with the cleaning solution. 14. The scanning speed of the spectral SPR biosensor is mainly dependent on the performance of the hardware motorized stages.

Acknowledgments This work was supported in part by a grant of the Korea Health 21 R&D Project, Ministry of Health, Welfare & Family Affairs, Republic of Korea (A030003).

Array-Based Spectral SPR Biosensor: Analysis of Mumps Virus Infection

47

References 1. Liedberg, B., Nylander, C., and Lundström, I. (1983) Surface plasmon resonance for gas detection and biosensing. Sens. Actuators B 4, 299–304 2. Yuk, J. S. and Ha, K. -S. (2005) Proteomic applications of SPR biosensors: analysis of protein arrays. Exp. Mol. Med. 37, 1–10 3. Ha, K. -S. and Yuk, J. S. (2006) Wavelength interrogation-based surface plasmon resonance biosensors, in Encyclopedia of Sensors (Grimes, C. A., Dickey, E. C., and Pishko, M. V., eds), American Scientific, Stevenson Ranch, CA, pp. 433–442 4. Homola, J., Yee, S. S., and Gauglitz, G. (1999) Surface plasmon resonance sensors: review. Sens. Acuators B 54, 3–15 5. Yuk, J. S., Jung, S. -H., Jung, J. -W., Hong, D. -G., Han, J. -A., Kim, Y. -M., and Ha, K. -S. (2004) Analysis of protein interactions by a wavelength interrogation-based surface plasmon resonance biosensor. Proteomics 4, 3468–3476 6. Röthenhausler, B., and Knoll, W. (1998) Surfac-plasmon microscopy. Nature 332, 615–617 7. Yuk, J. S., Kim, H. -S., Jung, J. -W., Jung, S. -H., Han, J. A., Kim, Y. -M., and Ha, K. -S. (2006) Analysis of protein interactions on protein arrays by a novel spectral surface plasmon resonance imaging. Biosens. Bioelectron. 21, 1521–1528

8. Yuk, J. S., Yi, S. -J., Lee, H. J., Lee, H. G., Kim, Y. -M., and Ha, K. -S. (2003) Characterization of surface plasmon resonance wavelength by changes of protein concentration on protein chips. Sens. Actuators B 94, 161–164 9. Jung, J. -W., Jung, S. -H., Kim, H. -S., Yuk, J. S., Park, J. -B., Kim, Y. -M., Han, J. -A., Kim, P. -H., and Ha, K. -S. (2006) Highthroughput analysis of GST-fusion protein expression and activity-dependent protein interactions on GST-fusion protein arrays by using a spectral surface plasmon resonance biosensor. Proteomics 6, 1110–1120 10. Kim, H. -S., Jung, S. -H., Kim, S. -H., Suh, I. B., Kim, W. J., Jung, J. -W., Yuk, J. S., Kim, Y. -M., and Ha, K. -S. (2006) High-throughput analysis of mumps virus and the virus specific monoclonal antibody on the arrays of a cationic polyelectrolyte with a spectral SPR biosensor. Proteomics 6, 6426–6432 11. Kim, S. H., Kee, S. H., Ahn, J. -E., Song, J. -W., Song, K. -J. (2003) Production and characterization of monoclonal antibodies to mumps virus isolated in Korea. J. Bacteriol. Virol. 33, 203–208 12. Yuk, J. S. and Ha, K. -S. (2004) Analysis of immunoreactions on protein arrays by using wavelength interrogation-based surface plasmon resonance sensors. J. Korean Phys. Soc. 45, 1104–1108

Chapter 4 Optical Biosensors Based on Photonic Crystal Surface Waves Valery N. Konopsky and Elena V. Alieva Summary Optical biosensors have played a key role in the selective recognition of target biomolecules and in biomolecular interaction analysis, providing kinetic data about biological binding events in real time without labeling. The advantages of the label-free concept are the elimination of detrimental effects from labels that may interfere with fundamental interaction and the absence of a time-consuming pretreatment. The disadvantages of all label-free techniques – including the most mature one, surface plasmon resonance (SPR) technique, are a deficient sensitivity to a specific signal and undesirable susceptibilities to non-specific signals, e.g., to the volume effect of refraction index variations. These variations arise from temperature fluctuations and drifts and they are the limiting factor for many state-of-the-art optical biosensors. Here we describe a new optical biosensor technique based on the registration of dual optical s-polarized waves on a photonic crystal surface. The simultaneous registration of two different optical modes from the same surface spot permits the segregation of the volume and the surface signals, while the absence of metal damping permits an increase in the propagation length of the optical surface waves and the sensitivity of the biosensor. The technique was tested with the binding of biotin molecules to a streptavidin monolayer that has been detected with a signal/noise ratio of about 15 at 1 s signal accumulation time. The detection limit is about 20 fg of the analyte on the probed spot of the surface. Key words: Label-free optical biosensors, Photonic crystal surface waves, Biotin–streptavidin binding, Streptavidin postbinding conformational change.

1. Introduction Registration of optical waves propagating along the surface under investigation is the most used method in label-free optical biosensors (1, 2). In the SPR technique (3), these waves are surface plasmonpolaritons (4) propagating along a gold or silver surface, while in the resonant mirror technique (5) the waves are waveguide modes Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, Totowa, LLC 2009 DOI: 10.1007/978-1-60327-567-5_4

49

50

Konopsky and Alieva

excited in a high refractive index dielectric layer via the frustrated total internal reflection (TIR) from a low refractive index spacer. In both cases, an evanescent field of the optical wave (with penetration depth in water ~100 nm) is sensitive not only to biomolecular interactions at the surface but also to changes in the volume refraction index (RI) of the liquid due to variations of the liquid temperature, composition and so on. For example, a water temperature change of 0.01°C gives a water RI change of about 10−6. Therefore, a need exists for a biosensor technique that would be able to segregate the volume and the surface contributions from an analyte in detected signals. To obtain these two parameters, one needs to detect at least two optical waves with different characteristics (e.g., with different penetration depths) simultaneously. In our work (6), we exploited a bulk optical wave propagating above the sensing surface as a reference of the volume RI variations. Our goal was to overcome the variations caused by temperature fluctuations and drifts that are a problem for many state-of-the-art optical biosensors. The weakness of this method is the need for a complicated flow cell design at small flow cell height (because of using the bulk optical wave). Slavík et al. (7) tried to use the excitation of long-range and short-range plasmons at the same surface spot with polychromatic light at a fixed incident angle (so-called wavelength interrogation – the spectrum of the reflected light is examined) to separate bulk from surface effects. The authors claimed the noise-limited resolution of their method is 11 times worse than the ordinary SPR method with wavelength interrogation (which itself is less sensitive than an angular interrogation method). The reason is the very small propagation length of the short-range plasmons. Moreover, the excitation of both modes at the same incident angle means that these modes differ little in their penetration depths, because the penetration depth difference here originates from a wavelength difference of these modes only (see Eq. 1 below). In the dual-waveguide interferometric technique (8) the measurement of propagation constants of two modes with s- and p-polarizations is used to seek an adsorption layer thickness and its RI. It should be noted that the exploitation of the modes with the orthogonal polarizations is a some weakness of the method. This is because of an implicit assumption in this method that the adlayer is an isotropic substance, while the adlayer is almost always anisotropic (and birefringent to some extent) as a result of the adlayer binding to the surface. Here we present a technique based on the simultaneous registration of two s-polarized optical surface waves on a one-dimensional photonic crystal surface. Photonic crystals (PCs) are materials that possess a periodic modulation of their refraction index on the scale of the wavelength of light (9). Such materials can exhibit photonic band gaps that are very much like the electronic band gaps for electron waves travelling in the periodic potential of the crystal.

Optical Biosensors Based on Photonic Crystal Surface Waves

51

In both cases, frequency intervals exist where the wave propagation is forbidden. This analogy may be extended (10) to include surface levels, which can exist in band gaps of electronic crystals. In PCs, they correspond to optical surface waves with dispersion curves located inside the photonic band gap. The one-dimensional photonic crystal (1D PC) is a simple periodic multilayer stack. Optical surface modes in 1D PCs were studied in the 1970s, both theoretically (11) and experimentally (12). Twenty years later, the excitation of optical surface waves in a Kretschmann-like configuration was demonstrated (13). Despite several theoretical proposals (13, 14) that suggested that the photonic crystal surface waves (PC SWs) have the potential to be superior alternatives in sensor applications to surface plasmons (due to low damping of PC SWs), there are no experimental demonstrations of such applications to date. In our opinion, the reason is that the limiting factor for the SPR technique is not the instrumental sensitivity but the temperature fluctuations and drifts as mentioned earlier. From this point of view, the increase of a propagation length of surface waves itself is ineffective without a concurrent compensation of the fluctuations of the liquid. We show that in addition to the low loss propagation (which is common for other all-dielectric biosensors), the technique presented here, based on dual optical surface waves in 1D PCs, has some additional advantages over all the earlier-mentioned biosensor techniques. Unique tunable properties of 1D PCs permit the design of a 1D PC structure that can support two longrange surface modes at the same wavelength (this is impossible in the SPR technique), with one mode exited very close to the angle of TIR from the water (this is unfeasible in any other waveguide technique). The mode, in which the exited angle is infinitesimally close to the angle of TIR from the external medium, has a very large penetration depth in this medium (e.g., water) and may be used as a reference of bulk RI fluctuations. Indeed, the weak localization of this mode reduces its sensitivity to overlayers and increases its sensitivity to changes in the bulk RI. Simultaneous detection of two modes, with one of them being more sensitive to changes of the RI of the liquid than the other, permits us to derive both the RI of the liquid, ne = ne(r1, r2), and the adlayer thickness, da = da(r1, r2), as functions of the detected angular parameters r1 and r2 of two PC SWs.

2. Materials All biochemicals (except streptavidin) were purchased from Sigma-Aldrich (Germany) and were used immediately after preparation.

52

Konopsky and Alieva

1. A dialkoxy aminosilane 3-(2-Aminoethylamino) propyldimethoxymethylsilane [molecular weight – 206.36] was used to convert OH-terminated SiO2 surface to NH2-terminated one (15). 2. Biotin-XX, SSE [Sulfosuccinimidyl Ester sodium salt or SulfoNHS-LC-LC-Biotin; molecular weight – 669.74; mass added to target – 452.6] was used for the biotinylation of the aminoterminated surface. 3. The streptavidin from Amersham (UK) [molecular weight Mstr~60 000] was deposited on the biotinylated surface. 4. The free biotin [vitamin H; molecular weight Mb = 244.31] was used as a test for the detection of small molecules binding with a streptavidin monolayer. All experiments were done in the PBS [phosphate-buffered saline; pH = 7.2] except absolute angle measurements of PC SWs excitation, which were done in pure water. 2.1. Photonic Crystal Structure

The following 1D PC structure was used: substrate/(LH)3L´/water, where L is a SiO2 layer with thickness d1 = 154.0 nm, H is a Ta2O5 layer with d2 = 89.4 nm and L´ is a SiO2 layer with d3 = 638.5 nm. The SiO2/Ta2O5 7-layers structure (started and finished by SiO2 layers) was deposited by ion sputtering. The prism and substrate were made from BK-7 glass. The RIs of the substrate, SiO2, Ta2O5 and water at l = 532 nm, were n0 = 1.52, n1 = n3 = 1.49, n2 = 2.12 and ne = 1.335, correspondingly. The RIs at other wavelengths were derived using dispersion data presented by Palik (16).

2.2. Flow Cell

The flow cell was made from a glass slide with two holes in which two glass tubes were fitted, serving as inlet and outlet, respectively. The inlet tube was connected to a small tank with the solution under study. Due to the force of gravity the solution flowed to the outlet. The flow velocity was controlled by an elevation difference of the inlet tank level and the outlet end. The height of the cell was determined by the thickness of the Teflon film, which served as a sealing gasket and as a spacer between the sample and the glass slide. We used Teflon films of 35 or 100 µm thickness. The flow cell volume was 3.5 µL or 10 µL, respectively. The dead volume of the flow cell system was approximately 25 µL. Gravity flows of streptavidin or biotin solutions and pure PBS buffer were used with a volumetric flow rate of up to 1 mL/min.

3. Methods 3.1. Absolute Angle Measurements

The excitation angles of the optical surface waves, indicated as white pentagrams in Fig. 1, were experimentally measured with

Optical Biosensors Based on Photonic Crystal Surface Waves

53

log10( Ie / I0) 0.4 2 0.45

442 nm 1

λ, µm

0.5

0.55

0

532 nm

−1 0.6 −2 0.65 ρTIR = ne → 0.7 1.25

1.3

ρ2

ρ1 1.35 ρ = no⋅sin(θo)

1.4

−3

1.45

Fig. 1. The calculated dispersion of the 7-layers PC structure in water and measured experimental points (white pentagrams) at l = 532 nm and l = 442 nm laser wavelengths. The two optical surface modes are clearly seen as dark curves (with an enhancement about 1,000) inside the band gap (light areas with an enhancement much less than 1).

an angular accuracy of ±1´ by parallel laser beam at two wavelengths: l = 532 nm (second harmonic of Nd-YAG laser) and l = 442 nm (He–Cd laser). A calculated dispersion of our 1D PC structure in water is presented in Fig. 1 as the logarithm of optical field enhancement (i.e., as lg[(EeEe*)/(E0E0*)]) in the external medium near the structure. Good correspondence is seen between experimental points (white pentagrams) and the calculated dispersion curves of the surface modes. The dispersion is presented in coordinate l(r), where l is an optical wavelength and r is a numerical aperture r = n0sin(q0). The numerical aperture r may be used as an angle variable instead of angles qj in different layers. This is a unified angle variable for all layers since, according to Snell’s law, r = n0·sin(q0) = nj·sin(qj), for any layer j. The angular parameter r, at which the excitation of a surface mode occurs, is equal to an effective RI of the mode. Therefore, the two dark curves in Fig. 1 present the dispersion of the two optical surface modes (i.e., the dependence of theirs effective RI from the wavelength). From Fig. 1, one can see that it is possible to excite one of the PC SWs in close proximity to the TIR angle from the water by appropriately choosing the laser wavelength and/or by appropriately choosing the PC structure. The penetration length of the evanescent wave intensity (i.e., EeEe*) in the external medium, which is

54

Konopsky and Alieva

le =

l 2 4p r12 - rTIR

,

(1)

may be very large for this mode if the difference (r1−rTIR) = (r1−ne) is small. This is a unique property of PC SWs, because in any standard waveguide techniques (5, 8), the numeric aperture or (in other words) the effective RI of the waveguide mode rmode is always more than the RI of the low refractive index spacer nspacer. Therefore, the difference (rmode−rTIR) ≥ (nspacer−ne) cannot be made small in the standard waveguides, taking into account the RI of the water (ne ≅ 1.33) and the RI of the spacer (usually made from SiO2, nspacer ≅ 1.49). Another unique property of PC SWs is the possibility to excite them in the structure, where the final dielectric layer (the silicon oxide layer in our case) may have a low RI, while the standard waveguide has a high RI layer on a low RI spacer. This simplifies the procedures of biochemical modification of the external surface, which is now the standard SiO2 surface. 3.2. Sample Preparation

Samples (i.e., their top silicon oxide layers with thickness 638.5 nm) were cleaned as follows: first, they were sonicated in ethanol and acetone for 5 min each and then immersed in a piranha solution (H2SO4[95%]:H2O2[30%] = 3:1) for 15 min (caution: piranha solution reacts violently with organic solvents). The glass slides were then exposed to UV-ozone (185 nm and 254 nm) for 45 min and finally thoroughly rinsed with DI water. The precleaned glass slides (with expected OH bonds on the SiO2 surface) were immersed in 1% aminosilane solution in 95% acetone/water for 5 min. The slides were then rinsed with acetone and baked for 30 min at 120°C. Then the sample was mounted in the flow cell, and further sample treatment was made in situ. For the biotinylation of the NH2-terminated surface of the slides, Sulfo-NHS-LC-LC-Biotin (2 mg/mL in PBS) was flowed over the flow cell for several minutes, then the fluid flow was stopped for several hours or even overnight, and the biotinylation of the surface was monitored in real time. Then the flow cell system was thoroughly rinsed with PBS.

3.3. Angular Resonance Curves Measurements

In Fig. 2 the biosensor setup scheme and a typical raw experimental signal from the setup are shown. A laser beam from second harmonic of Nd-YAG laser (LCM-T-111, Laser-export (17)) was sent to the sensor surface through a polarization maintaining fiber cable (FCPP-532-1-FC/APC2, OFR (18)) to improve the quality of a beam profile. The angular resonance curves in Fig. 2b were measured by focusing both parts of the split laser beam (with diameter D ≅ 3 mm) in the same spot on the structure surface with the objective of a focal length of f = 60 mm, and detecting the intensity distribution of reflected light with a 512-pixel

Optical Biosensors Based on Photonic Crystal Surface Waves

55

fiber

mm

laser,532nm

16

n θ01 θ02

n1 n2

n0

ρ1(θ01)

1D photonic crystal

ρ2(θ02)

ne photodiode array

ρ1

1.4

↓

1.2 1

↓

0.6

←−

←−

IR, a.u.

ρ2 0.8

∆θ =0.038

0.4 0.2

P1

0

↓ 0

50

100

150

↓ 200

250

300

P2

350

400

450

P, pixels Fig. 2. The biosensor scheme (a) and a typical raw experimental signal from the photodiode array (b).

photodiode array placed 385 mm| 442 mm (r1| r2) apart from the structure as shown in Fig. 2a. The Hamamatsu (19) photodiode array (S3904-512Q) was used to record both the experimental signals – angles of two PC SWs (r1 and r2) – simultaneously. The dynamic range of the angular measurements is ±D/(2f) = ± 0.025 rad that corresponds to the external media RI change ∆n ≅ ± 0.035 or to the adlayer thickness deposition ∆d ≅ ± 120 nm.

56

Konopsky and Alieva

The interference near resonance curves is the distinguishing feature of long-range PC SWs propagation. We observed similar interference in our work (20), dealing with long-range surface plasmon-polaritons propagation. The appearance of such interference means that the propagation distance of the PC SW becomes much more than the waist of an incident Gaussian beam at the surface (this also may be easily seen on the sample surface by the naked eye). In ref. 20, the origin of this interference is described in more detail. Note that the resonance peaks in Fig. 2b are very sharp (due to the long-range PC SWs propagation), and this allows measurement of the resonance peaks position change with high precision. 3.4. Data Handling

Data acquisition from the photodiode array, data processing and presentation were done with homemade software we wrote on a personal computer running under Windows. The RS232 computer interface was used to connect the personal computer and the photodiode array driver circuit (CDP Corp. (21)). The changes of the resonance peak positions P1 and P2 on the photodiode array (see Fig. 2b) may be converted to changes of the resonance angles ∆r1 and ∆r2. To derive the changes of the RI of the liquid and the adlayer thickness from the changes of the resonance angles of two PC SWs, we used two independent methods, which give similar results if the changes are small. The first method is based on an analysis of influence of the adlayer deposition and the bulk RI changes on the dispersion relation of the two optical surface modes. This method needs absolute values of the resonance angular parameters r1 and r2, and generates the absolute values of the RI of the external medium, ne = ne(r1, r2), and the adlayer thickness, da = da(r1, r2). The equations derived in the first method are cumbersome and not presented here. The second method is a pure linear method, based on the Taylor expansion. It needs relative changes ∆r1, ∆r2 and generates the relative changes of the RI of the liquid, ∆n = ∆n(∆r1, ∆r2), and relative changes of the adlayer thickness, ∆d = ∆d(∆r1, ∆r2). In the second method, we take Taylor expansion of both resonance angle changes in terms of ∆n and ∆d: Dr1 = Dn r1 + D d r1 =

∂ r1 ∂r Dn + 1 Dd , ∂n ∂d

(2a)

Dr2 = Dn r2 + D d r2 =

∂ r2 ∂r Dn + 2 Dd. ∂n ∂d

(2b)

From these equations we obtain the desired values as functions of measured ∆r1 and ∆r2:

Optical Biosensors Based on Photonic Crystal Surface Waves

Dr - Dr2 / K d r , Dn r1 = Dn ∂ 1 = 1 ∂n 1 - Kn / K d

D d r2 = Dd

∂ r2 Dr2 - K n Dr1 , = 1 - Kn / K d ∂d

57

(3a)

(3b)

where Kd and Kn are the ratio of the corresponding partial derivatives: Kd =

∂ r2 ∂d

∂ r1 , ∂d

(4a)

Kn =

∂ r2 ∂n

∂ r1 . ∂n

(4b)

To obtain a dimensionless value proportional to ∆n in one channel, and a dimensionless value proportional to ∆d in another channel, we need only Kd and Kn coefficients (see the right-hand side of Eqs. 3a, b). If we want to have ∆n in RI units and ∆d in length units, we also need ∂r1/∂n and ∂r2/∂d respectively. All these coefficients may be obtained, for example, from a theoretical simulation of the real 1D PC structure. For the presented structure, these coefficients are: Kd = 0.415; Kn = 0.1; ∂r1/∂n = 0.5 and ∂r2/∂d = 0.06 [1/µm] (assuming that the adlayer RI is na = 1.43). All ∆n and ∆d data presented below are derived by using the second (linear) method, with the given coefficients. 3.5. Streptavidin Monolayer Deposition

To verify the sensitivity of the biosensor and to compare it with existing label-free methods, we present the unsmoothed experimental data of free biotin binding on the streptavidin monolayer. Initially (Fig. 3), we present the build-up of the streptavidin monolayer on the biotinylated surface. Streptavidin (diluted in PBS to a concentration of cstr = 16 µg/mL) was run through the flow cell with volumetric flow rate vstr = 0.4 mL/min. Then the flow cell was rinsed with PBS. In Fig. 3, one can see that the adlayer thickness increases to 6.2 nm during streptavidin binding to the biotinylated surface.

3.6. Biotin Binding to the Streptavidin Monolayer

Figure 4c presents the ∆d (adlayer thickness changes) during free biotin binding to the streptavidin monolayer, while Fig. 4d shows ∆n (RI changes of the analyte) during these biotin solution injections. Biotin (in a concentration of cb = 3 µg/mL) was injected into the PBS running through the flow cell with volumetric flow rate vb = 0.6 mL/min. Figure 4c shows that the

Konopsky and Alieva 7 Streptavidin injection

↓

6

Streptavidin

∆d, nm

5 4

Biotinylated PC surface

3

SiO2

2 1 0 0

200

400

600

800

1000

1200

t, sec Fig. 3. Immobilization of streptavidin on a biotinylated surface. The measurement time is 1 s per point (no posterior data averaging and smoothing). In the grayscale inset the corresponding process is illustrated.

a

4.06

363.1

−→

(t=1500 sec)

363.08

4.05 363.04

∆ ρ2 ×104

4.055

363.06 P2, pixels

58

4.045

363.02 363

↑ Biotin injection 1

362.98

1350

1400

1450

1500

4.04

↑ Biotin injection 2

1550

1600

1650

1700

Fig. 4. Initial data from the biosensor and theirs treatment by the linear method during biotin binding to the streptavidin monolayer. (a) Movement of the second resonance mode in terms of pixels (left axis P2) and in terms of angle changes (right axis ∆r2). (b) Movement of the first resonance mode (which is more sensitive to the bulk RI) in terms of pixels (left axis P1) and in terms of angle changes (right axis ∆r1). (c) Calculated changes of the adlayer thickness in terms of ∆dr2 (right axis) and in nanometers (left axis). In the grayscale inset the corresponding process is illustrated. (d) Calculated changes of the bulk RI in terms of ∆nr1 (right axis) and in RI units (left axis).

Optical Biosensors Based on Photonic Crystal Surface Waves

b

59

159.05 159 158.95

−→

158.9

1.015

∆ ρ1 × 103

P1, pixels

(t=1500 sec)

158.85 158.8 158.75 1.01 158.7 158.65

↑

↑

Biotin injection 1

Biotin injection 2

158.6 158.55 1350

1400

1450

1500

1550

1600

1650

1.005 1700

t

c

free biotin

SiO2

Fig. 4. (continued)

streptavidin monolayer at first increases its thickness, but then contracts to a value slightly less than the initial one. At the same time, Fig. 4d shows that the external medium RI is not changed until the second biotin injection (from 1,501s until 1,600s ne ≅ const). So, in Fig. 4c, in this time period, we observe the act of streptavidin conformation during (or after) biotin molecule penetration into binding pockets of streptavidin molecules.

60

Konopsky and Alieva

d 1.15

x 10−4

x 10−5

1.1

5.5

1

5.0

→

∆n ρ 1

∆n

1.05

(t=1500 sec)

0.95

0.9

4.5

↑

0.85 1350

↑

Biotin injection 1

1400

1450

1500

Biotin injection 2

1550

1600

1650

1700

t

Fig. 4. (continued)

The second biotin injection did not result in the same streptavidin conformation, because most streptavidin subunits are already occupied by biotin molecules. 3.6.1. Mass-Transport Kinetics

Before proceeding further, it is worth estimating the characteristic time of the processes under study. The upper limit of a masstransport kinetics (by a combination of convection and diffusion) in our flow cell may be estimated by the so-called Smoluchowski– Levich approximation (22, 23) æDö j 0 = ckm = c ç ÷ èhø

2/3

1/3

æ v ö çè ÷ wx ø

,

(5)

where j0 [molecules/(s cm2)] is the deposition rate of molecules (with concentration c [molecules/cm3] and with a diffusion coefficient D [cm2/s]) on the sensing surface, which is considered an ideal collector. Parameters of our flow cell are as follows: h = 0.01 cm – height, w = 1 cm – width, x = 0.4 cm – distance from flow chamber entrance and v is the volumetric flow rate (cm3/s). The diffusion coefficient of a streptavidin molecule may be estimated by the Stokes–Einstein equation: Dstr =

kT 6ph Rstr

7 ´ 10 -7 cm 2 s,

(6)

where we take the streptavidin molecule diameter 2Rstr = 6.2 nm (see above), water viscosity h ≅ 10−2 g/(cm s) and Boltzmann constant k = 1.38 × 10−16 g cm2/(s2 grad). The diffusion coefficient

Optical Biosensors Based on Photonic Crystal Surface Waves

61

of a biotin molecule may be estimated by assuming that its effective radius is Rb = Rstr(Mb/Mstr)1/3. In this case Db = 4.3 × 10−6 cm2/s. Now we can estimate the time it takes the combination of convection and diffusion to supply a required number of molecules N for the monolayer cover of a probed spot on the surface. It is t=

N , j 0S

(7)

where S is the area of a probed spot on the surface. The probed spot on the surface is determined by the size of the laser beam focus ω0 ≅ 50 µm and by the propagation length of the optical surface waves L ≅ 2 mm. Therefore, an area of the probed spot is S = w0 L ≅ 0.1 mm2. Assuming that one streptavidin molecule occupies a square of 10 × 10 nm2 (i.e., N/S ≅ 1010 molecules/mm2), we deduce that Nstr ≅ 109 molecules were detected on our probed spot during the deposition presented in Fig. 3. Using Eqs. 5–7 we obtain the characteristic time of the mass transport of 109 streptavidin molecules to the probed spot of the surface: tstr ≅ 14 s, that corresponds well to the time of the linear increase in Fig. 3. From the same equations, we receive the next characteristic time of the mass transport of Nb = 2Nstr ≅ 2 × 109 biotin molecules to the probed spot of the surface: tb ≅ 0.16 s, i.e., less than the time of the single measurement (1 s). 3.6.2. Biotin–Streptavidin Binding Kinetics

The biotin–streptavidin couple has extremely high binding affinity KA = kon/koff ~ 1013 [1/M] and, therefore, the characteristic time of biotin–streptavidin binding estimated through its association constant kon ~ 7.5 × 107 [1/(M s)] (24) is in the millisecond range (at the concentration cb we used). However, the biotin– streptavidin association involved several transient intermediate steps, and the simple framework based on the single association constant appears insufficient for a detailed description of this binding. The transient intermediate steps include desolvation of five bound water molecules in each biotin binding site. Then a flexible loop in streptavidin becomes immobilized after biotin binding in a biotin binding pocket and closes the biotin binding pocket (and, hence, shields a biotin molecule from competition with the solvent) (25). The decrease of the streptavidin thickness in Fig. 4c at 1,501s to 1,600s may be this streptavidin postbinding conformational change needed for a stable interaction.

3.7. Conclusion

We have employed the two different optical modes on the photonic crystal surface for the optical sensing of biomolecular interactions. Unique properties of photonic crystals were used for the excitation of optical waves along the photonic crystal surface so that the evanescent field of one wave penetrates much deeper into the liquid volume. This wave is used as a reference for the RI of the liquid.

62

Konopsky and Alieva

The simultaneous registration of the two modes makes it possible to derive both the RI of the liquid and the adlayer thickness. This permitted us to segregate the volume and the surface signals from the analyte, increase the sensitivity of biomolecule detection and record the act of streptavidin conformation during the binding of biotin molecules. Independent registration of the adlayer thickness and the temperature-dependent RI of the liquid may be potentially useful for designing a temperature-controlled flow cell, which may be considered a biochemical reactor for evaluation of temperature dependency of reactions at the surface.

4. Notes 1. Measurements of adlayer thickness changes against a background of bulk RI changes. It may appear that the sharp thickness increase in Fig. 4c (at 1,499 s) is a data evaluation artefact due to a sharp change in the bulk RI, because the kinetics of the increase looks very similar to that of the decrease of the refractive index in Fig. 4d. To be sure that it is not the case and that the adlayer thickness really increases after biotin injection, we present the initial data of resonance peak changes: Fig. 4a, b, while Fig. 4c, d are the treatment of these data by our second method (i.e., by Eq. 3a, b). From Fig. 4a, one can see that the second resonance angle r2 increases its value (at 1,499s to 1,500s) after biotin solution injection, reaches its maximum at 1,500s and only then decreases, while the first resonance angle r1 (which is more sensitive to the bulk RI) is sharply decreasing at 1,499s to 1,500s (Fig. 4b). Considering Eq. 2b and keeping in mind that all partial derivatives in this equation are positive (∂r2/∂d, ∂r2/∂n > 0) one can deduce that the increase of r2 at 1,499s to 1,500s may be the result of either the adlayer thickness increase or a bulk RI increase. But to exclude the latter interpretation we have decreased the RI of the biotin solution in respect to RI of the PBS by adding 20 µL of the pure water in 1 mL of PBS with biotin (before the biotin solution injection). Since nPBS − nH2O ≅ 0.0012, the change of the RI of the biotin solution (due to water injection) in respect of the RI of PBS is ∆n ≅ −2 × 10−5, while the biotin itself in such a small concentration practically does not change the RI of the PBS (∆nbiotin < 10−7). This procedure was needed for us to be sure that the injection of the biotin solution will cause only the bulk RI to decrease. Therefore, the increase of r2 at 1,499s to 1,500s may be the result of only the adlayer thickness increase. The exact value of the corresponding calculated adlayer thickness increase ∆d (but not its sign) in Fig. 4c depends on the exactness of the numerical values of the coefficients pointed

Optical Biosensors Based on Photonic Crystal Surface Waves

63

after Eq. 4. But, inasmuch as the expected bulk RI change ∆n ≅ −2 × 10−5 corresponds well to the results from Fig. 4d (calculated with the same coefficients), we expect that the calculated value of the adlayer thickness increase ∆d is also reasonably accurate. It may be noted here that both changes of buffer RI in Fig. 4d are in good agreement with the calculated values. During biotin injection1, 1 mL of biotin solution in PBS with 20 µL of pure water was added into the flow cell system (0.2 mL of pure PBS was at this time in the system) – the expected decrease of RI is equal to 2 × 10−5. After biotin injection 2, the same amount (1 mL of biotin solution in PBS with 20 µL of pure water) was added to 0.2 mL of the solution RI of which already was decreased. The calculated decrease of RI in this case equal to 3 × 10−6 is in good agreement with the data in Fig. 4d. 2. Measurement noises and mass detection limits. In compliance with (26), we suppose that the process of the biotin–streptavidin binding is a good candidate for comparison of the signal/ noise ratio of different label-free techniques. We believe that for comparison of the signal/noise ratio it is also very important always to point out the time of the measurement and the fact of posterior data averaging and/or smoothing (which increase the effective measurement time). In other words, the noise should be reduced to 1/√Hz value. In our experiments, the signal accumulation time was 1 second per point and no posterior data averaging or smoothing was done. The noise (i.e., standard deviation – std) of the thickness measurement was equal δd = std(da) ≅ 1.3 pm/√Hz. The noise of the measurement of the external medium RI was δn = std(ne) ≅ 5 × 10−7/√Hz. In Fig. 4c, one can see that we detected the streptavidin conformation process during free biotin binding with a (signal/noise)b ratio of about 15. The deposition of the streptavidin monolayer was detected with a (signal/noise)str ratio of about 5,000. Taking into account the (signal/noise) ratio, we obtain a minimal quantity of streptavidin str molecules that may be detected at the probed spot of our str N min = Nstr/(signal/noise)str ⯝ 200 000 streptavidin setup: molecules. This corresponds to a mass detection limit str str mmin = N min M str / N A @ 2 ´ 10 -14 g = 20 fg of the analyte on the probed spot of the surface (NA ≅ 6 × 1023 is Avogadro’s number). For biotin molecules we have a minimal detectable b str 8 quantity equal to N min = 2N / (signal/noise)b @ 1.3 ´ 10 b b biotin molecules or mmin = N min M b / N A @ 50 fg of the analyte on our probed spot. We believe that the noise of this technique could be further decreased by improving the quality of the dielectric multilayer coating and by decreasing the laser noise.

64

Konopsky and Alieva

Acknowledgement The authors thank S. Grachev for the kind donation of some biochemicals and for helpful advice about surface preparation. This work was partly supported by the European Network of Excellence, NMP3-CT- 2005-515703-2. References 1. Robinson, G. (1995) The commercial development of planar optical biosensors. Sens. Actuators B 29, 31–36 2. Cooper, M. A. (2003) Label-free screening of bio-molecular interactions. Anal. Bioanal. Chem. 377, 834–842 3. Homola, J., Yee, S. S. and Gauglitz, G. (1999) Surface plasmon resonance sensors: review. Sens. Actuators B 54, 3–15 4. Raether, H. (1988) Surface Plasmons. Springer, Berlin 5. Cush, R., Cronin, J., Stewart, W., Maule, C., Molloy, J. and Goddard, N. (1993) The resonant mirror – a novel optical biosensor for direct sensing of biomolecular interactions. I. Principle of operation and associated instrumentation. Biosens. Bioelectron. 8, 347–353 6. Alieva, E. V. and Konopsky, V. N. (2004) Biosensor based on surface plasmon interferometry independent on variations of liquid’s refraction index. Sens. Actuators B 99, 90–97 7. Slavík, R., Homola, J. and Vaisocherová, H. (2006) Advanced biosensing using simultaneous excitation of short and long range surface plasmons. Meas. Sci. Technol. 17, 932–938 8. Cross, G., Reeves, A., Brand, S., Swann, M., Peel, L., Freeman, N. and Lu, J. (2004) The metrics of surface adsorbed small molecules on the Young’s fringe dual-slab waveguide interferometer. J. Phys. D Appl. Phys. 37, 74–80 9. Yablonovitch, E. (1993) Photonic band-gap structures. J. Opt. Soc. Am. B 10, 283–295 10. Kossel, D. (1966) Analogies between thinfilm optics and electron band theory of solids. J. Opt. Soc. Am. 56, 1434–1434 11. Yeh, P., Yariv, A. and Hong, C.-S. (1977) Electromagnetic propagation in periodic stratified media. I. General theory. J. Opt. Soc. Am. 67, 423–438 12. Yeh, P., Yariv, A. and Cho, A. Y. (1978) Optical surface waves in periodic layered media. Appl. Phys. Lett. 32, 104–105 13. Robertson, W. M. and May, M. S. (1999) Surface electromagnetic waves on one-dimensional photonic band gap arrays. Appl. Phys. Lett. 74, 1800–1802

14. Villa, F., Regalado, L., Ramos-Mendieta, F., Gaspar-Armenta, J. and Lopez-Rios, T. (2002) Photonic crystal sensor based on surface waves for thin-film characterization. Opt. Lett. 27, 646–648 15. Li, J., Wang, H., Zhao, Y., Cheng, L., He, N. and Lu, Z. (2001) Assembly method fabricating linkers for covalently bonding DNA on glass surface. Sensors 1, 53–59 16. Palik, E. D. (1985) Handbook of Optical Constants of Solids. Academic, London 17. http://www.laser-export.com 18. http://www.ofr.com 19. http://www.hamamatsu.com 20. Konopsky, V. N. and Alieva, E. V. (2006) Long-range propagation of plasmon polaritons in a thin metal film on a one-dimensional photonic crystal surface. Phys. Rev. Lett. 97, 253904 21. http://www.cdpsystems.com 22. Elimelech, M. (1994) Particle deposition on ideal collectors from dilute flowing suspensions: Mathematical formulation, numerical solution, and simulations. Sep. Technol. 4, 186–212 23. Myszka, D. G., He, X., Dembo, M., Morton, T. A. and Goldstein, B. (1998) Extending the range of rate constants available from BIACORE: Interpreting mass transport-influenced binding data. Biophys. J. 75, 583–594 24. Hyre, D. E., Trong, I. L., Merritt, E. A., Eccleston, J. F., Green, N. M., Stenkamp, R. E. and Stayton, P. S. (2006) Cooperative hydrogen bond interactions in the streptavidin– biotin system. Protein Sci. 15, 459–467 25. Freitag, S., Trong, I. L., Klumb, L., Stayton, P. S. and Stenkamp, R. E. (1997) Structural studies of the streptavidin binding loop. Protein Sci. 6, 1157–1166 26. Zybin, A., Grunwald, C., Mirsky, V. M., Kuhlmann, J., Wolfbeis, O. S. and Niemax, K. (2005) Double-wavelength technique for surface plasmon resonance measurements: Basic concept and applications for single sensors and two-dimensional sensor arrays. Anal. Chem. 77, 2393–2399

Chapter 5 Surface Plasmon Resonance Biosensing Marek Piliarik, Hana Vaisocherová, and Jirˇí Homola Summary Surface plasmon resonance (SPR) biosensors belong to label-free optical biosensing technologies. The SPR method is based on optical measurement of refractive index changes associated with the binding of analyte molecules in a sample to biorecognize molecules immobilized on the SPR sensor. Since late 1990’s, SPR biosensors have become the main tool for the study of biomolecular interactions both in life science and pharmaceutical research. In addition, they have been increasingly applied in the detection of chemical and biological substances in important areas such as medical diagnostics, environmental monitoring, food safety and security. This chapter reviews the main principles of SPR biosensor technology and discusses applications of this technology for rapid, sensitive and specific detection of chemical and biological analytes. Key words: Optical biosensors, Affinity biosensing, Biorecognition elements, Detection of chemical and biological species, Bioassays.

1. Introduction Optical affinity biosensors based on surface plasmon resonance (SPR) present one of the most advanced label-free optical sensing technologies (1–3). Their ability to monitor the interaction between a molecule immobilized on the surface of the sensor and the interacting molecular partner in a solution have made SPR sensors a very powerful tool for biomolecular interaction analysis and biomolecular research in general (4). In recent years, SPR biosensors have been increasingly used also for the detection of chemical and biological substances related to medical diagnostics, environmental monitoring, food safety and security (5).

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_5

65

66

Piliarik, Vaisocherová, and Homola

This chapter deals with SPR biosensor technology and its application in the quantification of chemical and biological species. 1.1. Principles of SPR Biosensing

SPR affinity biosensors are devices that consist of three main subsystems: sensor hardware (optical reader), biorecognition element, and sample preparation and delivery system (Fig. 1). In the optical reader of an SPR sensor, a light wave excites a special mode of electromagnetic field – surface plasmon (SP). Surface plasmon propagates along a thin metal film and its field probes the medium adjacent to the metal surface. Any change in the refractive index in the proximity of the metal surface results in a change in the velocity of the surface plasmon. This change in the propagation can be determined from the characteristics of the light wave coupled to the surface plasmon. Biorecognition elements specific to analyte molecules are immobilized on the surface of the metal. If a liquid sample is brought in contact with the sensor surface, molecules of analyte are captured by biorecognition molecules (Fig. 2). The binding gives rise to a refractive index change close to the sensor surface, which can be measured by the optical reader. The liquid sample is brought to the sensor surface using a sample preparation and a delivery system.

Optical reader change of refractive index Biorecognition element Interaction with analyte molecules Sample delivery system Fig. 1. Principal components of SPR affinity biosensor.

Surface plasmon

Sample with analyte

Biorecognition elements, n

Biorecognition elements, n+δn

Metal film

Metal film

Optical reader

ve

a tw

gh

Li

Fig. 2. Principle of operation of SPR affinity biosensors.

SP

Li

gh

tw

av

e

Surface Plasmon Resonance Biosensing

67

The following sections present the underlying principles and characteristics of the main parts of an SPR affinity biosensor. 1.2. Optical Reader

Nowadays, numerous SPR readers are available as commercial systems or research prototypes. Although their designs may differ considerably, the underlying physical principles of the SPR method remain the same. Therefore, the first part of this section is dedicated to fundamentals of surface plasmons and SPR sensing while the following part illuminates possible differences among the SPR sensing platforms and gives examples of the most frequent SPR sensing platforms.

1.2.1. Surface Plasmons and Their Optical Interrogation

A surface plasmon is an electromagnetic wave, which propagates along an interface between a dielectric and a metal and is characterized by the propagation constant and electromagnetic field distribution (6). The field of a surface plasmon is transversemagnetic (TM) polarized (the vector of magnetic intensity is perpendicular to the plane of incidence) and evanescently decays into both the media while the major part of the field is located in the dielectric. The propagation constant of SP bSP is determined by the optical constants of the surrounding media (permittivity eM of the metal and the refractive index of the dielectric nD) as follows: b SP =

e MnD2 w w nef = , c c e M + nD2

(1)

where w is the angular frequency, c is the speed of light in vacuum and nef denotes the effective refractive index of the surface plasmon. Although several metals can support surface plasmons at optical frequencies, all the main existing SPR (bio)sensors use gold due to its excellent chemical stability. The electro-magnetic field of the SP is strongly localized at the metal surface and its penetration depth into the dielectric Lpd (see Note 1) is typically 150–400 nm depending on the operating wavelength (Fig. 3). A change in the refractive index of the dielectric within a distance h from the metal surface produces a change in the effective refractive index of the surface plasmon nef. The magnitude of the change depends on the thickness of the layer h within which the refractive index change occurs, operating wavelength, and a refractive index distribution. If the thickness of the layer is much higher than that of the penetration depth Lpd of the SP field, the change in the effective refractive index of the surface plasmon can be calculated as Dnef =

nef3 (DnD )h > > L pd. nD3

(2)

Piliarik, Vaisocherová, and Homola 1.0

Magnetic field intensity [arb. u.]

68

Metal

Dielectric

Wavelength

0.8

630 nm 760 nm 850 nm

0.6

0.4 1/e

0.2

0.0 0.0

0.2 0.4 0.6 0.8 Perpendicular coordinate [µm]

1.0

Fig. 3. Distribution of electromagnetic field of a SP at the gold–water interface for three different wavelengths.

The change in the effective refractive index of a surface plasmon due to a refractive index change within a thin layer h <
2h nef3 (DnD )h < < L pd. L pd nD3

(3)

If the refractive index change is produced by the binding of specific molecules at the sensor surface, the change of refractive index ∆nb b occurring within a layer of thickness h can be expressed as æ dn ö æ dn ö DG Dnb = ç ÷ Dc b = ç ÷ , è dc ø vol è dc ø vol h

(4)

where (dn/dc)vol is the refractive index increment, ∆cb is the wt/vol concentration of bound molecules within the sensitive layer with the thickness h and ∆Γ is the corresponding surface concentration (mass per surface area). The linear relation between the refractive index change and the surface concentration of bound molecular mass in Eq. 4 is often referred to as de Feijter formula (7). The refractive index increment (dn/dc)vol is a well-characterized property for most of the biochemical species and ranges typically from 0.1 to 0.3 cm3g−1 (8). Proteins and nucleic acids typically exhibit (dn/dc)vol = 0.18 cm3g−1 with variability less than 8%. As follows from Eqs. 2–4, a change in the effective index of the surface plasmon due to the capture of analyte can be expressed as

Surface Plasmon Resonance Biosensing

Dnef = K DG,

69

(5)

where K is a constant. The optical reader of the SPR sensor measures changes in a characteristic of a plasmon-coupled light wave resulting from changes in the effective refractive index of a surface plasmon. Several coupling mechanism are used to couple a light wave to a surface plasmon. The most common mechanisms include the attenuated total reflection in prism couplers and diffraction on diffraction grating couplers (6). Most SPR sensors are based on the Kretschmann configuration of the attenuated total reflection method. In this geometry, a light wave passes through a high refractive index prism and is totally reflected at the prism base covered with a thin gold film (Fig. 4). Light evanescently tunnels through the thin metal film and can excite an SP at the outer boundary of the gold if the incident light wave and SP are closely phase-matched. The phasematching condition can be written as n p sin q = nef ,

(6)

where np is the refractive index of the coupling prism and q is the angle of incidence on the metal film (in the prism). The coupling of incident light wave to an SP is accompanied by a transfer of energy and results in a drop of the intensity of the reflected light wave. As the coupling occurs only within a narrow range of angles of incidence (or wavelengths), the excitation of SP produces a narrow dip in the angular (or wavelength) spectrum of the reflected light. Based on these characteristics, the reflected light wave is measured in the SPR sensor hardware.

Light wave Coupling prism

Metal film Sample

SP

Fig. 4. Excitation of surface plasmons in the Kretschmann geometry of the attenuated total reflection method.

70

Piliarik, Vaisocherová, and Homola

Polychromatic light source

Array detector Monochromatic light source

Monochromatic light source

Spectrometer

Photodetector

Convergent light beam

n

n+∆n

∆λ Wavelength

c)

n

n+∆n

∆θ Angle of incidence

Relative intensity

b)

Relative intensity

Relative intensity

a)

n

∆R n+∆n Wavelength or angle

Fig. 5. SPR sensors based on modulation of (a) wavelength, (b) angle of incidence, and (c) light intensity.

SPR sensors are classified as: (a) SPR sensors with wavelength modulation (angle of incidence is fixed and the coupling wavelength serves as a sensor output), (b) SPR sensors with angular modulation (coupling wavelength is fixed and the coupling angle of incidence serves as a sensor output), (c) SPR sensors with intensity modulation (both the angle of incidence and the wavelength of incident light are fixed at nearly resonant values and the light intensity serves as a sensor output) (Fig. 5). A change in the refractive index at the metal surface results in a change in the effective refractive index of an SP and a change in the coupling condition (6). Consequently, the change in the refractive index produces a change in the position of the SPR dip in the spectrum of reflected light (SPR sensors with angular or wavelength modulation) or a change in the intensity of reflected light (SPR sensors with intensity modulation) (Fig. 5). The relation between the refractive index change and the sensor output depends on the spatial distribution of the refractive index change and design of the SPR reader (coupling method, modulation method, and operating wavelength). Table 1 shows sensor responses to changes in the bulk refractive index of the sample and changes in the refractive index at the surface of the metal. Changes in the bulk refractive index are given in refractive index units (RIU) and changes in the surface refractive index are expressed in an equivalent amount of captured protein. It can be concluded from Table 1 that the change of the sensor response induced by the adsorption of 10 pg × mm−2 of a protein is comparable to the change induced by the change of the bulk refractive index above the sensor surface of 10−5 RIU. This relation depends on the resonant wavelength and the method of

Surface Plasmon Resonance Biosensing

71

Table 1 Comparison of responses of prism-based SPR sensors with different modulation approaches

Bulk change 10−5 RIU −2

Surface change 10 pg × mm

Wavelength modulation (nm)

Angular modulation (deg)

Intensity modulation (%)

0.055

1 × 10−3

0.12

0.063

1.15 × 10

−3

0.14

(dn/dc)vol = 0.18 cm3g−1 operating wavelength – 750 nm

coupling and the presented comparison applies to the operating wavelength of 750 nm. High-performance SPR sensors used for the most demanding applications are able to resolve changes in the surface concentration as small as 0.1 pg mm−2. This suggests (Table 1) that the optical system of the sensor has to resolve changes of resonant wavelength smaller than 0.5 pm or the resonant angle of incidence smaller than 10−5 degree. 1.2.2. Implementations of SPR Readers

Since the first commercial SPR sensors appeared on the market in 1990, numerous SPR platforms have been introduced by several companies. The most widely used SPR sensors are sensors developed by Biacore (since 2006 Biacore has been a part of GE Healthcare) which offer a wide range of SPR sensors for biomolecular interaction research. Their product portfolio includes, for instance, high-resolution angular-modulation systems with four sensing channels and automated flow delivery, and sample handling, such as Biacore 3000 and Biacore T100. Biacore Flexchip, IBIS–iSPR (Ibis, The Netherlands) and Proteomic Processor (Lumera Corporation, USA) represent array SPR sensors based on intensity modulation (SPR imaging) with several hundreds of sensing channels. Autolab ESPRIT (Eco Chemie B.V., Netherlands) system is also based on angular modulation and fast scanning and sample delivery is realized by direct autosampling in two measuring cuvettes. The Multiskop system (Optrel GbR, Germany) can be used either in the angular scanning mode for accurate measurements or with a quadrant-cell photodiode for real-time kinetic measurements. Examples of research prototypes include β-SPR (Sensia, Spain) which is an angular modulation based system, with two independent sensing channels and Plasmon IV (Institute of Photonics and Electronics, Czech Republic), a four-channel SPR sensor with wavelength modulation.

72

Piliarik, Vaisocherová, and Homola

An SPR reader comprises four key subsystems (Fig. 6) – light source module, sensor head with coupling optics, detector module, and data processing unit. The choice of a light source depends on the modulation method used and operating range required. In SPR sensors with angular and intensity modulations, monochromatic light or quasi monochromatic light is used and therefore laser diodes or narrowband light emitting diodes are the light sources of choice. In SPR sensors with wavelength modulation, polychromatic light sources such as halogen lamps are most commonly employed. The light source needs to be stable in terms of emitted intensity and spectral distribution for optimum performance of the SPR sensor, (see Note 2). The sensor head performs multiple functions and therefore comprises several subsystems. Illuminating optics deliver light beams of appropriate spatial distribution into the SPR coupling element (parallel beam in intensity and wavelength modulation based SPR sensors and wedge-shaped beam in angular modulation based sensors). The coupling optics typically consist of a coupling prism and an SPR sensor chip which is optically matched to the prism (see Note 3). As SP is a TM electromagnetic wave, polarization of light coupling to the SP is controlled with polarizing optics which is, typically, a linear polarizer. Light reflected from the SPR chip passes through the coupler and is collected at the detector module. Signals from individual channels are separated either by optical means in the detector module or by software means in the data processing unit (see Note 4). As the temperature fluctuations can induce changes in the refractive index of samples and thus interfere with SPR biosensing,

Optical reader

Light source

Input light

Coupling optics

Data processing unit Data acquisition and digitization

Light wave with encoded SPR signal

Detector module

modulation of SPR Biorecognition element Interaction with target molecules Sample handling system

Fig. 6. Scheme of an SPR biosensor and main subsystems of optical SPR reader.

Surface Plasmon Resonance Biosensing

73

precise control of temperature at the sensor surface is required for reliable results. High-performance SPR sensors are therefore equipped with temperature stabilization of the coupling optics and the analyzed sample. The detector module measures the intensity of light transmitted through the sensor head or its spectral or angular distribution. In the intensity modulation-based sensors, array detectors such as 2D CCD cameras are frequently used to allow simultaneous measurements in multiple areas of the sensing surface (see Note 5). In angular modulation-based sensors, the angular spectrum is usually spread across an array detector and different sensing channels (if more than one channel is present) are measured with one or more detector lines. SPR sensors with wavelength modulation are usually equipped with a separate spectrometer for each sensing channel. For more details, see Fig. 5. In the process of detection of light, the detector module produces a set of intensity values (the number of intensities for an individual channel depends on the type of detector). The data corresponding to individual sensing channels are processed separately. In SPR sensors based on angular or wavelength modulation, spectra for individual sensing channels are extracted and an appropriate algorithm locates the position of the SPR dip (9). In SPR sensors based on intensity modulation, the sensor output is usually defined as an averaged intensity over the area of the detector corresponding to a sensing channel. The sensor output can be calibrated to a change in the refractive index, or surface concentration of bound molecules. 1.3. Biorecognition Element

In SPR affinity biosensors, the biomolecular recognition element is immobilized on the solid surface of the SPR sensor reader and captures analyte molecules. The choice of biorecognition elements (BRE) and immobilization method presents an important step in the development of an SPR biosensor and has a direct impact on the key performance characteristics of the developed SPR biosensor system such as specificity and limit of detection (10). Numerous molecules can serve as biomolecular recognition elements. The key factors for their selection include high affinity to the target analyte, stability of biological activity, specificity, and availability of functional groups for their direct immobilization. Antibodies remain by far the most frequently used biorecognition elements (11). However, development of high-quality antibodies is still a rather expensive and laborious process. Other molecules that have been commonly employed in SPR sensors include antibody fragments, antigens, engineered affibodies, peptides, oligonucleotides and molecular imprinted polymers (12, 13). Surface chemistries used for immobilization of biorecognition elements on the SPR reader are required to provide a high density of biorecognition elements on the surface while maintaining their

74

Piliarik, Vaisocherová, and Homola

biological activity. Given the complexity and variability of biorecognition elements, there is no universal immobilization method. The choice of immobilization chemistry is therefore made based on the properties of a specific biorecognition element (5). Nonspecific binding of non-target molecules in complex samples such as bodily fluids, food samples, etc. presents another challenge for surface chemistry of SPR biosensors. Since the beginning of SPR biosensors research in early 1990’s various approaches to the immobilization of biomolecules on the SPR sensor surface have been developed. These include procedures based on physical adsorption via hydrophobic and electrostatic interactions (14), strong covalent binding (15), or attachment of tagged proteins by a site-specific non-covalent interaction between the tag and a capture molecule. The attachment to the surface can be mediated by a molecular recognition event such as biotin–avidin interaction (16) or DNA hybridization (17). The covalent attachment requires the presence of reactive groups on the support (usually electrophilic groups such as aldehydes or succinimidyl esters) able to react with nucleophilic groups (amino, thiol, hydroxyl) on the ligand molecules (15, 18). One of the most remarkable techniques in surface chemistry is the spontaneous self-organization of n-alkylthiols or disulfides on gold surfaces (Fig. 7). Self-assembled monolayers (SAMs) have been employed in many immobilization methods for spatially controlled attachment of biomolecular recognition elements (19). Threedimensional matrices that contain binding sites in a structured 3D environment (e.g. hydrogels such as carboxymethyl dextran) present an interesting alternative for the immobilization of biomolecular recognition elements (15). This approach offers a high number of binding sites for covalent attachment of biomolecular recognition elements and relatively non-fouling background. However, the binding of analyte molecules to the immobilized elements may be slowed down due to the diffusion of the analyte through the hydrogel matrix (20–22).

Gold Layer Substrate Fig. 7. Scheme of n-dodecanethiolate self-assembled on gold substrate. The assembly is held together by bonds between the sulfur head groups and the gold surface and van der Waals and hydrophobic interactions between neighboring hydrocarbon chains.

Surface Plasmon Resonance Biosensing

1.4. Sample Delivery System

75

A liquid sample is brought to the sensor surface using a sample preparation and delivery system. In principle, the sample is either injected into a cuvette that is interfaced with a sensing surface and the interaction takes place in a fixed volume under no-flow conditions or the sample is continuously flowed through a special chamber (a flow cell unit) attached to the sensor surface (23). In cuvette-based SPR sensors, the analyte contained in a fixed volume of liquid sample (cuvette volume is typically 10–200 µL) interacts with biorecognition elements on the sensing surface. This interaction may be affected by mass transport of the analyte from the sample volume to the sensing surface; therefore SPR sensors employing cuvettes usually allow stirring of the sample in the cuvette during the measurement. In flow cell-based SPR sensors, the measured sample flows through a flow cell (volume of a flow cell chamber is 10 nL–10 µL) at flow rates ranging from 1–100 µL/min (see Note 6). In the flow cell, the analyte molecules diffuse to the sensor surface and interact with biorecognition elements on the sensing surface. In order to provide a constant flow of sample of a desired flow rate, flow cells are connected with peristaltic or syringe pumps. Advanced sample delivery systems offer complex and automated sample management (24).

2. Materials 2.1. Immobilization of Biorecognition Elements

1. The buffer solutions used for immobilization of biorecognition elements include: PB (10 mM phosphate buffer, pH 7.0 at 20°C), PBS (10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, pH 7.4 at 20°C), PBM (10 mM phosphate buffer, 15 mM MgCl2, pH 7.2 at 20°C); PBNa (10 mM phosphate buffer, 0.75 M NaCl, 2.7 mM KCl, pH 8.0 at 20°C), PBST (10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, 0.05% Tween-20, pH 7.4), Tris–HCl (10 mM Tris, 50 mM NaCl, pH 8.0 at 20°C), and SA (1–10 mM sodium acetate, pH 5.0 at 20°C). Phosphate buffers can be prepared using standard procedure from stock monobasic and dibasic sodium and potassium phosphate solutions with a ratio corresponding to a desired pH. When appropriate, pH of the buffer may be adjusted with HCl or NaOH. 2. Alkylthiols (AT), i.e. C11-chained tetra(ethylene glycol)- terminated alkanethiol (AT-EG4), C11-chained tetra(ethylene glycol) -COOH- terminated alkanethiol (AT-EG4–COOH), and C16chained biotin-terminated (AT-BAT), and C11-chained NH2terminated thiols (AT–NH2) can be obtained from Prochimia, Poland (www.prochimia.com). 11-Amino-1-undecanethiol, hydrochloride (AT-NH2) can be obtained from Dojindo Laboratories, Japan (www.dojindo.com).

76

Piliarik, Vaisocherová, and Homola

3. Reagents for activation of carboxylic groups N-hydroxysuccinimide (NHS, catalog number: 56480), N-Ethyl-N′-(3dimethylaminopropyl) carbodiimide hydrochloride (EDC, catalog number: E7750), N,N,N′,N′-Tetramethyl-O-(N-succinimidyl) uronium tetrafluoroborate (TSTU, catalog number: 85972), and 4-(dimethylamino)-pyridin (DMAP) are available from SigmaAldrich, USA (www.sigma.com). Trifluoroacetic acid, ammonium hydroxide solution, acetic acid and triethylmaine (all of these reagents should be reagent grade or better) are available, for instance, from Sigma-Aldrich, USA (www.sigma.com). 4. Standard solvents such as absolute ethanol, anhydrous dioxane, and anhydrous dimethylformamide (DMF) can be obtained from Sigma-Aldrich, USA. 5. Tween 20, streptavidin, caseine, ethanolamine, and bovine serum albumin can be obtained from Sigma-Aldrich, USA.

3. Methods 3.1. Immobilization of Biorecognition Elements

Immobilization of biorecognition elements on the gold surface of the SPR reader can be performed either in the SPR system using the microfluidic channels or outside the SPR instrument on the whole surface of the sensor chip. In the first approach, the whole functionalization process or its final steps involving delivery of biorecognition elements to a selected sensing channel are performed with the sensor microfluidics. This approach makes it possible to observe the SPR sensor response to the immobilization of biorecognition elements and determine the surface concentration of the immobilized molecules. Alternatively, a spatiallyresolved functionalization of the SPR chip by biorecognition elements can be achieved by techniques such as microspotting, in which small droplets of a solution containing biorecognition elements are delivered on the reactive sensor surface. This approach is particularly useful for producing SPR arrays with a large number of sensing channels which cannot be individually addressed by microfluidic devices. In this chapter, we present protocols for four selected immobilization methods for in situ attachment of oligonucleotides and proteins on the gold surface of an SPR chip.

3.1.1. Immobilization of Antibodies (Proteins) via Covalent Attachment to COOH/OEG Self-Assembled Monolayer (SAM) (15, 25, 26)

Suitable for immobilization of proteins and other molecules containing primary amines. Advantages: Stability of immobilized biorecognition element layer under a wide range of environmental conditions (pH, temperature) and good resistance to non-specific binding of molecules from complex media such as serum, urine, and food matrix.

Surface Plasmon Resonance Biosensing

77

Disadvantages: Immobilization level of biorecognition elements depends on accessibility of primary amines for covalent binding to N-hydroxysuccinimidyl esters created on SAM, not-oriented immobilization. 1. Prepare the 1 mM stock solutions of AT-EG4–COOH and AT-EG4 thiols in absolute ethanol. Seal the tubes well with parafilm. Store them at a temperature of −20°C before use (see Note 7). 2. Rinse the gold sensor surface with water and absolute ethanol. Dry with nitrogen stream. Clean the gold sensor surface in UV ozone cleaner for 20 min, wash it with water (see Note 8) and dry with nitrogen stream (see Note 9). 3. Immerse the clean sensor chip in solution mixed of AT-EG4– COOH and AT-EG4 thiols mixed in the ratio of 3:7 at a room temperature with a total thiol concentration of 200 µM. Use PP tube or glass Petri dish. Add trifluoroacetic acid to the final concentration of 2%. Seal the tube or dish or with parafilm. After 10 min of incubation at 40°C, store the chip in this solution in a dark place at a room temperature overnight. Immersed chip can be stored for up to 2 weeks. 4. Rinse the chip with ethanolic solution of ammonium hydroxide (10% v/v), ethanol, water and ethanol and dry with nitrogen stream (see Note 10). 5. Immerse the sensor chip in a solution of TSTU in argondeaerated DMF (2 mg/mL), add DMAP (0.02 M) and shake it for 1 h under inert gas. Alternatively, expose the chip to freshly prepared water solution of NHS (0.05 M) and EDC (0.2 M) for 10 min. Alternatively, dissolve NHS in anhydrous dioxane (2.02 mg/mL) and EDC in water (100 mg/mL), mix them in the ratio of 49:1 and incubate the sensor chip in this solution for 1 h at room temperature. Storage in inert gas is recommended. 6. Rinse the sensor chip with ethanol and water, dry it with nitrogen and mount it into the SPR instrument immediately. 7. Flow 10 mM sodium acetate buffer, pH 5.0 (SA) at 25°C at higher flow rate (e.g. 50 µL/min, depending on the sensor fluidics, Note 11) through the sensor flow cell for a few minutes. 8. Inject antibody or protein in SA buffer at a concentration of 5–50 µg/mL and let it flow through the flow cell for 15 min at 30 µL/min. (Optimal flow rate can differ depending on the design of the used flow cell; optimal concentration can differ depending on immobilized ligand reactivity with surface-bound N-hydroxysuccinimidyl esters.) Replace the solution with SA buffer. For proteins sensitive to low pH, use PB buffer (10 mM phosphate, pH 7.0) instead of SA.

78

Piliarik, Vaisocherová, and Homola

9. Inject 1 M ethanolamine, pH 8.0 at 20°C or Tris buffer for 10 min to deactivate the residual N-hydroxysuccinimidyl esters. Flow 10 mM phosphate buffer, 0.75 mM NaCl, pH 8.5 (PBNa) along the functionalized surface for 5 min to remove all non-covalently bound biorecognition elements. 10. Switch to appropriate detection/assay buffer such as 10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, pH 7.4 (PBS). Inject suitable blocking agent (e.g. bovine serum albumin, caseine, dry milk) in PBS at a concentration of 0.2– 10 mg/mL to block the sensor surface. The concentration of the blocking agent dependas on the complexity of the analyte solution; in general, the more complex the solution, the higher should be the concentration of the blocking agent. 3.1.2. Immobilization of Biotinylated Ligands to Streptavidin Coupled via Covalent Attachment to COOH/OEG Self-Assembled Monolayer (26, 27)

Suitable for immobilization of biotin-tagged ligands (proteins, antibodies, oligonucleotides) Advantages: Site-specific immobilization, high level of immobilized biotinylated biorecognition elements, stability of immobilized biorecognition element layer under a wide range of environmental conditions (pH, temperature), and good resistance to non-specific binding of molecules from complex media such as serum, urine, and food matrix. Disadvantages: Biotinylation of biorecognition elements required before immobilization. 1. Prepare 1 mM stock solutions of AT-EG4-COOH and AT-EG4 thiols in absolute ethanol. Seal the tubes well with parafilm. Store them at a temperature of −20°C before use (see Note 7). 2. Rinse the gold sensor surface with water and absolute ethanol. Dry it with nitrogen stream. Clean the gold sensor surface in UV ozone cleaner for 20 min, wash it with water (see Note 8) and dry it with nitrogen stream (see Note 9). 3. Immerse the clean sensor chip into solution of AT-EG4COOH and AT-EG4 thiols mixed in the ratio of 3:7 at room temperature with a total thiol concentration of 200 µM. Use PP tube or glass Petri dish. Add trifluoroacetic acid to the final concentration of 2%. Seal the tube or dish with parafilm. After 10 min of incubation at 40°C, store the chip in this solution in a dark place at room temperature overnight. The immersed chip can be stored for up to 2 weeks. 4. Rinse the chip with ethanolic solution of ammonium hydroxide (10% v/v), ethanol, water and ethanol and dry it with a stream of nitrogen (see Note 10). 5. Immerse the sensor chip in a solution of TSTU in argondeaerated DMF (2 mg/mL), add DMAP (0.02 M) and shake it for 1 h under inert gas. Alternatively, expose the chip to

Surface Plasmon Resonance Biosensing

79

freshly prepared water solution of NHS (0.05 M) and EDC (0.2 M) for 10 min. Alternatively, dissolve NHS in anhydrous dioxane (2.02 mg/mL) and EDC in water (100 mg/mL), mix them in the ratio of 49:1 and incubate the sensor chip in this solution for 1 h at room temperature. Storage in inert gas is recommended. 6. Rinse the sensor chip with ethanol and water, dry it with nitrogen and mount it in the SPR instrument immediately. 7. Flow 10 mM sodium acetate buffer, pH 5.0 (SA) at 25°C at a higher flow rate (e.g. 50 µL/min, depending on the sensor fluidics, Note 11) through the sensor flow cell for a few minutes. 8. Inject streptavidin in SA buffer at concentration of 25 µg/ml and let it flow through the flow-cell for 15 min at 30 µl/min. (Optimal flow rate can differ depending on design of the used flow cell.) Replace the solution with SA buffer. 9. Inject 1 M ethanolamine, pH 8.0 at 20°C or Tris buffer for 10 min to deactivate the residual N-hydroxysuccinimidyl esters. Flow 10 mM phosphate buffer, 0.75 mM NaCl, pH 8.5 (PBNa) along the streptavidin-coated surface for 5 min to remove all non-covalently bound streptavidin. Flush the sensor surface with SA; at this point the streptavidin immobilization level can be determined. 10. Switch to 10 mM phosphate buffer, 15 mM MgCl2 buffer, pH 7.2 (PBM) (oligonucleotide immobilization) or to 10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, pH 7.4 (PBS) (protein immobilization). Inject solution of biotinylated oligonucleotide probes with hexa(ethyleneglycol) spacer at concentration of 50 nM in PBM or biotinylated antibodies at concentration of 10 µg/mL in PBS until a plateau is reached (typically 10 min at a flow rate of 30 µL/ min). Replace the solution with PBM buffer or PBS buffer, respectively. When a stable level is reached, the immobilization of biorecognition element on the sensor surface is completed and the chip is ready for use. A typical sensorgram illustrating this protocol is shown in Fig. 8 which shows sensor response to the immobilization of biotinylated oligonucleotides. 3.1.3. Immobilization of Biotinylated Ligands via BAT/OEG Method (16, 17, 28)

Suitable for immobilization of biotin-tagged ligands (proteins, antibodies, oligonucleotides). Advantages: Site-specific immobilization, high level of immobilized biotinylated biorecognition elements, stability of immobilized biorecognition element layer under a wide range of environmental conditions (pH, temperature), good resistance to non-specific binding of molecules from complex media such as serum, urine, and food matrix.

2.8

SA 2.3

PBM

1.8 1.4

Streptavidin surface coverage

Biotin-23mer ON probe 50 nM

PBNa

PBM 0.9 0.5

SA Streptavidin 50 µg/ml

SA

ON probe surface coverage

Piliarik, Vaisocherová, and Homola

Sensor response [mRIU]

80

0.0 15

30

45 TIme [min]

60

75

Fig. 8. A typical sensorgram corresponding to immobilization of biotinylated oligonucleotides on the SPR chip via COOH/OEG method. The resulting amounts of bound streptavidin and oligonucleotide (ON) are indicated. Arrows indicate injections of solutions.

Disadvantages: Biotinylation of biorecognition elements required before immobilization. 1. Prepare the 5 mM stock solution of four(ethylene glycol) thiol (AT-EG4) and 1 mM stock solution of biotin-terminated four(ethylene glycol) thiol (AT-EG4-BAT) in absolute ethanol. Seal the tubes well with parafilm. Store them at a temperature of −20°C before use (see Note 7). 2. Rinse the gold sensor surface with absolute ethanol. Dry it with nitrogen stream. Clean the gold sensor surface in UV ozone cleaner for 20 min, wash it with water and dry it with nitrogen stream (see Note 9). 3. Immerse the clean sensor chip in solution mixed of AT-EG4BAT and AT-EG4 thiols mixed in the ratio of 1:9 at a room temperature with a total thiol concentration of 100 µM. The stock AT solution is typically used (see Subheading 2.1, item 2). In a glass Petri dish, combine absolute ethanol (14,860 µl), AT-EG4-BAT (50 µl, 1 mM) and AT-EG4 (90 µl, 5 mM). Use PP tube or glass Petri dish. Immerse chip immediately in this thiol solution. Seal the tube/dish well with parafilm. Store overnight or for up to 2 weeks at room temperature. 4. Rinse the chip with absolute ethanol, blow it dry with a stream of nitrogen and mount it in the SPR instrument immediately. 5. Inject streptavidin in SA buffer at a concentration of 5 µg/mL and let it flow through the flow cell for 15 min at 30 µL/min. (Optimal flow rate can differ depending on the design of the used flow cell.) Replace the solution with SA buffer. 6. Switch to 10 mM phosphate buffer, 15 mM MgCl2 buffer, pH 7.2 (PBM) (oligonucleotide immobilization) or to 10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, pH 7.4 (PBS)

Surface Plasmon Resonance Biosensing

81

(protein immobilization). Inject solution of biotinylated oligonucleotide probes with hexa(ethyleneglycol) spacer at concentration of 50 nM in PBM or biotinylated antibodies at concentration of 5 µg/mL in PBS until a plateau is reached (typically 10 min at a flow rate of 30 µL/min). Replace the solution with PBM buffer or PBS buffer, respectively. When a stable level is reached, the immobilization of the biorecognition element on the sensor surface is completed and the chip is ready for use. 3.1.4. Immobilization of Proteins via Physical Adsorption to Positively Charged Surface of SAM (25)

Suitable for immobilization of a wide range of biorecognition elements (proteins, antibodies, oligonucleotides), particularly suitable for biomolecules containing negatively charged groups/epitopes. Advantages: Simple procedure, very high level of immobilized biorecognition elements. Disadvantages: Low resistance to non-specific binding from complex media, surface blocking is required. 1. Prepare 1 mM stock solution of AT-NH2 thiols in absolute ethanol. Seal the tubes well with parafilm. Store it at a temperature of −20°C before use. 2. Rinse the gold sensor surface with water and absolute ethanol. Dry it with nitrogen stream. Clean the gold sensor surface in UV ozone cleaner for 20 min, wash it with water (see Note 8) and dry with nitrogen stream (see Note 9). 3. Immerse the clean sensor chip into a solution of AT-NH2 with a total thiol concentration of 200 µM and at room temperature. Use PP tube or glass Petri dish. Add triethylamine to the final concentration of 3%. Seal the tube or dish with parafilm. Store the chip in this solution in a dark place at room temperature overnight. The immersed chip can be stored for up to 2 weeks. 4. Rinse the chip with ethanolic solution of acetic acid (10% v/v), ethanol, water and ethanol and dry it with nitrogen stream (see Note 10). Mount it in the SPR instrument immediately. 5. Flow 10 mM phosphate buffer, pH 7.0 (PB) at 25°C at higher flow rate (e.g. 50 µL/min, depending on the sensor fluidics, Note 11) through the sensor flow cell for a few minutes. 6. Inject antibody or protein in PB buffer at a concentration of 2–20 µg/mL and let it flow through the flow cell for 15 min at 50 µL/min. (Optimal flow rate can differ depending on design of the used flow cell.) Replace the solution with PB buffer. 7. Flow 10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, 0.05% Tween, pH 7.4 (PBST) along the functionalized surface for 15 min to remove all weakly bound biorecognition elements. Flush the sensor surface with PB; at this point the biorecognition element immobilization level can be determined.

82

Piliarik, Vaisocherová, and Homola

8. Switch to 10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, pH 7.4 (PBS). Inject suitable blocking agent (e.g. bovine serum albumin, caseine, dry milk) in PBS at a concentration of 0.2–10 mg/mL to block the sensor surface. The used concentration of blocking agent is dependent on the complexity of analyte solution; in general, the higher complexity of analyte solution, the higher concentration and the more careful selection of blocking agent are necessary. In typical biosensing experiments, one of the sensing channels is functionalized with a biorecognition element against the specific target while the other channel (reference) is coated with molecules which are structurally similar to the used biorecognition element but are expected to have no interaction with the sample. Response of the reference channel is then subtracted from the sensing channel to reduce the effect of interferences. 3.2. Detection of Chemical and Biological Analytes 3.2.1. Direct Binding Assay

Direct detection is usually preferred in applications when binding of the analyte in the concentration range of interest provides a sufficient sensor response. Direct detection is typically well suited for medium and large analytes (>5,000 Da) for which detection limits in the ng/mL or sub-ng/mL range can be achieved (5, 29, 30). In the direct detection format analyte molecules in a sample bind to the biorecognition elements immobilized on the sensor surface (e.g. antibody, Fig. 9), and the sensor response is proportional to the concentration of the captured analyte acquired. In the direct detection format, the flow cell of the SPR sensor is initially loaded with buffer (both the sensing and reference channels) and when the baseline is established, a sample is injected and incubated with the sensor surface for a certain period of time (see Note 12). Subsequently, the sample is replaced with the buffer. Typical responses of the SPR sensor are shown in Fig. 10. The response of the sensor to the analyte present in the sample can be quantified by the following metrics: ● Slope of the sensor response immediately after the injection; the slope is determined using data from a fixed time interval (see Note 13)

Analyte

Sensor surface coated with antibodies

Fig. 9. Detection formats used in SPR biosensors: direct detection.

Surface Plasmon Resonance Biosensing 15 Slope of the res ponse

200nM

12 Sensor response [RU]

83

9

Equilibrium sensor response Buffer e

of ns o e op esp l S er th

6 hCG

20nM

3

0 0

(A)

5 (B)

10

15

20

Time [min]

Fig. 10. Typical sensorgram showing response of the sensor to direct detection of two concentrations of hCG hormone. The slopes of sensor response after the injection are plotted. The initial slope of the sensor response (time A), typical example of a fixed-time measurement (time B) and the equilibrium sensor responses are denoted. Arrows indicate injections of solutions.

●

Equilibrium sensor response, i.e. sensor response when the association and dissociation processes are in equilibrium and sensor response no longer increases with time

Fixed-time response, i.e. response of the sensor observed in a fixed time after the sample was injected (see Note 13) If the appropriate calibration curve is available (Fig. 11), the sensor response (the slope, the equilibrium response or the fixedtime response, respectively) can be translated to the concentration of analyte. The ability of the sensor to detect the analyte is characterized by the limit of detection (LOD). LOD is defined as a concentration of the analyte that produces a sensor response equal to three standard deviations of the sensor response for blank sample (31); the LOD for this model oligonucleotide system was approximately 100 pM (Fig. 11).

●

3.2.2. Sandwich Assay

The lowest detection limits of the direct SPR sensors can be improved by using a sandwich assay. A sandwich assay is generally used to improve the limit of detection beyond that available in the direct detection format and improve specificity of the detection. This approach has also been demonstrated to be suitable for detection in complex matrices (29). In the sandwich assay detection format, the primary biorecognition element immobilized on the sensor surface captures the analyte from the sample

Piliarik, Vaisocherová, and Homola

Slope of the response [min−1]

84

10−2

10−3

10−4

Limit of detection

0.1

0.2

0.4 0.6 0.8 1

2

4

6 8 10

20

40

Concentration [nM]

Fig. 11. Determination of concentration of analyte from the calibration curve and the limit of detection. The calibration curve is quantified by the slope of the sensor response.

Analyte

Secondary antibody

Sensor surface coated with antibodies

Fig. 12. Detection formats used in SPR affinity biosensors: sandwich assay.

and the subsequently-injected secondary biorecognition element binds to the previously captured analyte (Fig. 12). 3.2.3. Binding Inhibition Assay

Competitive assays are commonly used for detections of low concentrations of small analytes, whose binding to surface-immobilized recognition molecules does not produce a measurable sensor response. The binding inhibition assay has been successfully used for small chemicals such as pesticides and toxins with detection limits below 0.1 ng/mL (32, 33). In the binding inhibition assay (Fig. 13) analyte molecules or their derivatives are immobilized on the sensor surface. A known quantity of biorecognition elements is incubated with the sample and allowed to bind to the analyte in a sample. The mixture is then flowed over the sensor surface and the unreacted biorecognition elements are captured at the sensor surface. As the concentration of the analyte increases, the concentration of free epitopes on the biorecognition elements decreases and therefore the sensor response is inversely proportional to the concentration of the analyte.

Surface Plasmon Resonance Biosensing

Sample + antibody mixture

85

Sensor surface coated with target analyte

Fig. 13. Detection formats used in SPR affinity biosensors: binding inhibition assay.

4. Notes 1. Penetration depth is defined as the distance from the interface, where the amplitude of the field decreases by a factor of 1/e ≈ 0.37. 2. In SPR systems, the stability of the wavelength spectrum significantly influences the performance of the sensor. Therefore, special attention needs to be paid to the light source condition: ●

●

The spectrum is influenced by the changes of temperature. Therefore, the light source module requires typically 20–60 min to stabilize or warm up before the experiment. The spectrum can change or become unstable when the lifetime of the light source is exceeded.

3. The coupling prism and the sensor chip are typically made of material with identical or very close refractive indices (e.g. the standard optical glass NBK7). The chip must be optically matched to the coupling prism using a layer of refractive index matching material (e.g. refractive index matching liquid or polymer). 4. Signals can be directed into different detectors such as spectrometers, or line CCD arrays, or different portions of a single 2D array detector. This is often achieved using standard imaging optics, or an array of lenses aligned with the array of sensing channels. 5. The approach based on the intensity measurement is usually referred to as SPR imaging. A 2D array of sensing channels is illuminated with monochromatic light under the angle of incidence close to the resonance and the reflected light is projected on a CCD camera. The spatial distribution of light intensity can be mapped on the spatial distribution of sensing channels. 6. The choice of a flow rate depends mostly on the geometry of the fluidics. Faster flow rates can slightly improve the mass transport when very fast binding needs to be monitored;

86

Piliarik, Vaisocherová, and Homola

slower flow rates may reduce consumption of sample when the binding is too slow even for a relatively high concentration of analyte. 7. Instead of four (ethylene glycol), general oligo(ethylene glycol) thiol may be used. However, shorter OEG groups are likely to lead to coatings with a lower resistance to nonspecific binding. 8. Unless stated otherwise, all solutions should be prepared in water that has a resistivity of 18.2 MΩ cm and total organic content of less than five parts per billion. 9. Alternatively, different cleaning procedures can be used (such as plasma cleaner, piranha solution, etc.) unless the sensor chip is stored since production in clean and inert atmosphere. To clean the gold sensor surface in Piranha solution, use 3:1 mixture of sulfuric acid and 30% hydrogen peroxide in a glass or teflon vial. The solution must be mixed before cleaning or directly applied to the gold surface, applying the sulfuric acid first, followed by the peroxide. Piranha solutions are extremely energetic and may result in explosion or skin burns if not handled with extreme caution. 10. During the manipulations with the coated gold sensor chip, it is necessary to avoid touching the sensor area with the tweezers. It could destroy the molecular coatings or contaminate the sensor surface. 11. Air bubbles are strongly undesirable in all running buffers. Buffer solutions should be degassed before preparation of all solutions and injection to the sensor. In order to eliminate the risk of air bubbles, certain SPR sensor systems incorporate inline degassers or bubble traps. 12. The time of incubation depends on details of the specific application and factors such as affinity between the biorecognition element and analyte, concentration of analyte, desired limit of detection and operating range of the sensor, etc. 13. The time of detection should be adjusted based on the required detection limits and stability of the sensor output. While, in general, longer detection times yield better limits of detection, at long detection times the limits of detection may be limited by the stability of sensor output. References 1. Homola, J. (2003) Present and future of surface plasmon resonance biosensors. Analytical and Bioanalytical Chemistry 377, 528–539 2. Homola, J., Vaisocherová, H., Dostálek, J., and Piliarik, M. (2005) Multi-analyte surface

plasmon resonance biosensing. Methods 37, 26–36 3. Boozer, C., Kim, G., Cong, S.X., Guan, H.W., and Londergan, T. (2006) Looking towards label-free biomolecular interaction analysis in a high-throughput format:

Surface Plasmon Resonance Biosensing

4.

5. 6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

a review of new surface plasmon resonance technologies. Current Opinion in Biotechnology 17, 400–405 Rich, R.L., and Myszka, D.G. (2006) Survey of the year 2005 commercial optical biosensor literature. Journal of Molecular Recognition 19, 478–534 Homola, J. (2006) Surface plasmon resonance based sensors. Springer-Verlag Raether, H. (1988) Surface-plasmons on smooth and rough surfaces and on gratings. Springer Tracts in Modern Physics 111, 1–133 de Feijter, J.A., Benjamins, J., and Veer, F.A. (1978) Ellipsometry as a tool to study the adsorption of synthetic and biopolymers at the air– water interface. Biopolymers 17, 1759–1772 Tumolo, T., Angnes, L., and Baptista, M.S. (2004) Determination of the refractive index increment (dn/dc) of molecule and macromolecule solutions by surface plasmon resonance. Analytical Biochemistry 333, 273–279 Nenninger, G.G., Piliarik, M., and Homola, J. (2002) Data analysis for optical sensors based on spectroscopy of surface plasmons. Measurement Science & Technology 13, 2038–2046 Piliarik, M., Vaisocherová, H., and Homola, J. (2007) Towards parallelized surface plasmon resonance sensor platform for sensitive detection of oligonucleotides. Sensorors and Actuators B Chem 121, 187–193 Tomizaki, K.Y., Usui, K., and Mihara, H. (2005) Protein-detecting microarrays: current accomplishments and requirements. Chembiochem 6, 782–799 Angenendt, P. (2005) Progress in protein and antibody microarray technology. Drug Discovery Today 10, 503–511 Elia, G., Silacci, M., Scheurer, S., Scheuermann, J., and Neri, D. (2002) Affinitycapture reagents for protein arrays. Trends Biotechnology 20, S19–S22 Koubová, V., Brynda, E., Karasová, L., Škvor, J., Homola, J., Dostálek, J., Tobiška, P., and Rošický, J. (2001) Detection of foodborne pathogens using surface plasmon resonance biosensors. Sensorors and Actuators B Chem 74, 100–105 Lofas, S., Johnsson, B., Edstrom, A., Hansson, A., Lindquist, G., Hillgren, R.M.M., and Stigh, L. (1995) Methods for site controlled coupling to carboxymethyldextran surfaces in surface-plasmon resonance sensors. Biosensors & Bioelectronics 10, 813–822 Busse, S., Scheumann, V., Menges, B., and Mittler, S. (2002) Sensitivity studies for specific

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

87

binding reactions using the biotin/streptavidin system by evanescent optical methods. Biosensors & Bioelectronics 17, 704–710 Ladd, J., Boozer, C., Yu, Q., Chen, S., Homola, J., and Jiang, S. (2004) DNAdirected protein immobilization on mixed self-assembled monolayers via a streptavidin bridge. Langmuir 20, 8090–8095 Oshannessy, D.J., Brighamburke, M., and Peck, K. (1992) Immobilization chemistries suitable for use in the biacore surface-plasmon resonance detector. Analytical Biochemistry 205, 132–136 Knoll, W., Liley, M., Piscevic, D., Spinke, J., and Tarlov, M.J. (1997) Supramolecular architectures for the functionalization of solid surfaces. Advances in Biophysics 34, 231–251 Myszka, D.G., He, X., Dembo, M., Morton, T.A., and Goldstein, B. (1998) Extending the range of rate constants available from BIACORE: interpreting mass transportinfluenced binding data. Biophysical Journal 75, 583–594 Sikavitsas, V., Nitsche, J.M., and Mountziaris, T.J. (2002) Transport and kinetic processes underlying biomolecular interactions in the BIACORE optical biosensor. Biotechnology Progress 18, 885–897 Witz, J. (1999) Kinetic analysis of analyte binding by optical biosensors: hydrodynamic penetration of the analyte flow into the polymer matrix reduces the influence of mass transport. Analytical Biochemistry 270, 201–206 Ward, L.D., and Winzor, D.J. (2000) Relative merits of optical biosensors based on flow cell and cuvette designs. Analytical Biochemistry 285, 179–193 Sjölander, S., and Urbanitzky, C. (1991) Integrated fluid handling system for biomolecular interaction analysis. Analytical Chemistry 63, 2338–2345 Wang, H., Chen, S., Li, L., and Jiang, S. (2005) Improved method for the preparation of carboxylic acid and amine terminated self-assembled monolayers of alkanethiolates. Langmuir 21, 2633–2636 Lahiri, J., Isaacs, L., Tien, J., and Whitesides, G.M. (1999) A strategy for the generation of surfaces presenting ligands for studies of binding based on an active ester as a common reactive intermediate: a surface plasmon resonance study. Analytical Chemistry 71, 777–790 Vaisocherová, H., Zítová, A., Lachmanová, M., Štepánek, J., Kralíková, S., Liboška, R.,

88

Piliarik, Vaisocherová, and Homola

Rejman, D., Rosenberg, I., and Homola, J. (2006) Investigating oligonucleotide hybridization at subnanomolar level by surface plasmon resonance biosensor method. Biopolymers 82, 394–398 28. Pinkel, D., Segraves, R., Sudar, D., Clark, S., Poole, I., Kowbel, D., Collins, C., Kuo, W.L., Chen, C., Zhai, Y., Dairkee, S.H., Ljung, B.M., Gray, J.W., and Albertson, D.G. (1998) High resolution analysis of DNA copy number variation using comparative genomic hybridization to microarrays. Nature Genetics 20, 207–211 29. Homola, J., Dostálek, J., Chen, S.F., Rasooly, A., Jiang, S.Y., and Yee, S.S. (2002) Spectral surface plasmon resonance biosensor for detection of staphylococcal enterotoxin B in milk. International Journal of Food Microbiology 75, 61–69 30. Oh, B.K., Kim, Y.K., Lee, W., Bae, Y.M., Lee, W.H., and Choi, J.W. (2003) Immunosensor

for detection of Legionella pneumophila using surface plasmon resonance. Biosensors & Bioelectronics 18, 605–611 31. Thomsen, V., Schatzlein, D., and Mercuro, D. (2003) Limits of detection in spectroscopy. Spectroscopy 18, 112–114 32. Gobi, K.V., Tanaka, H., Shoyama, Y., and Miura, N. (2004) Continuous flow immunosensor for highly selective and real-time detection of sub-ppb levels of 2-hydroxybiphenyl by using surface plasmon resonance imaging. Biosensors & Bioelectronics 20, 350–357 33. Dostálek, J., and Homola, J. (2008) Surface plasmon resonance sensor based on an array of diffraction gratings for highly-parallelized observation of biomolecular interactions, Sensors and Actuators B: Chemical 129, 303–310

Chapter 6 Label-Free Detection with the Resonant Mirror Biosensor Mohammed Zourob, Souna Elwary, Xudong Fan, Stephan Mohr, and Nicholas J. Goddard Summary The resonant mirror (RM) biosensor is a leaky waveguide-based instrument that uses the evanescent field to probe changes in the refractive index at the sensing surface.The RM can therefore be used to monitor in real-time and label-free the interaction between an analyte in solution and its biospecific partner immobilized on the waveguide surface.The RM has been used in studying the interaction of a variety of moieties including proteins, carbohydrates, cells, nucleic acids and receptors, leading to applications in areas such as clinical diagnostics, homeland security, and pharmaceutical and biomolecular interactions. This chapter will review the principle of this biosensor, and the recent advances in instrumentation, different immobilization chemistries, and kinetic studies, as well as some applications. Key words: Resonant mirror biosensor, Leaky waveguide biosensor, Label-free. Biomolecules immobilization, Refractive index, Immunosensors, Cell detection, Nucleic acid hybridization, Kinetics, Biomedical application, Industrial applications.

1. Introduction Optical evanescent biosensors greatly facilitate the study of the interaction between a range of interactants. The resonant mirror was developed in 1993 (1, 2) and demonstrated in a number of applications, including concentration determination (2), molecular recognition (3, 4), kinetics association and dissociation (5, 6), epitope mapping for different analytes including proteins (7–9), nucleic acids (10–12), carbohydrates (13–16), cells (17–19), and receptors (20, 21), in addition to industrial applications (22–25). In principle, the user links one of the interactants to the sensor

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_6

89

90

Zourob et al.

surface, adds the other reactants and follows the binding response. The optical signal is generated by the change in the refractive index as the binding species displace the water molecules from the interrogated surface. It will provide information about the strength and rate of binary complex formation and dissociation. This chapter will cover some background about the resonant mirror, recent developments, different protocols used for immobilizations, and some applications and kinetics measurements. 1.1. The Resonant Mirror Instrument

The resonant mirror (RM) was first commercialized by Affinity Sensors (IAsys, Cambridge, UK), then Thermo Electron acquired the technology, becoming Thermo Labsystems, and now it has been transferred to NeoSensors (Durham, UK). In the RM sensor, FTIR is used to couple the light in and out of a highindex waveguiding layer. The RM is effectively a prism coupler where the air gap has been replaced by a low-index dielectric layer. Figure 1 shows the RM device integrated with a microfluidic and stirrer. The RM chip structure, consists of a high-index substrate prism (n = 1.72), a thin low-index spacer (about 550 nm of silica, n = 1.45) and a very thin monomode waveguiding layer (about 80 nm of Si3N4, n = 2.0) (1, 2). The high-index resonant layer acts as both a waveguide and the sensing layer. The light incident above the critical angle on the substrate–spacer interface is coupled with the waveguiding layer via the evanescent field in the spacer, when the propagation constants in the substrate prism and the waveguide match. For monochromatic light (angular scan

1.1.1. The RM Structure

Stirrer Aspirator tube

Outlet

Inlet

High index resonant layer

Low index coupling layer

Polarizer

light source

Polarizer

Detector

Q Prism block

Fig. 1. Schematic showing the structure of the resonant mirror integrated in a microfluidic and stirrer.

Label-Free Detection with the Resonant Mirror Biosensor

91

using constant wavelength), this occurs over a very narrow range of incidence angles, typically spanning considerably less than 1°. The light decoupled out via the coupling layer emerges to strike the detector. The resonance angle is very sensitive to the change in the refractive index within the evanescent field at the surface and will not detect changes in the bulk outside the evanescent field. Thus by immobilizing recognition receptors (ligand), it is possible to measure only those molecules (ligates) that bind to or dissociate from the ligand. To resolve the resonance angle, the optical components are arranged so that a 90° rotation of polarization occurs for the light to travel along the waveguide. Another polarizer is placed in front of the detector to enable only the light that has traveled along the waveguide to reach the detector. A sharp peak of intensity is seen at the discrete angle of interest and this is extremely sensitive to the surface binding/dissociation. In the IAsys the laser light is scanned over an angle of 14°, the middle 10° of which is used for detection purposes. The other 4° is used to ensure the laser is moving at a constant speed over the measurement region. An encoder grating located on the end of the laser arm enables the position of the arm to be known at any time. The time at which the light falls on the detector is monitored and converted into a resonance angle, the resulting sensorgram displaying the resonance angle vs. time in the IAsys software (25). The resonant mirror can be operated using two instrumental configurations, the angular scan at constant wavelength (Fig. 2a) and wavelength scan at constant angle (Fig. 2b) (26). It was shown that the two configurations of the RM (the angular scan at constant wavelength and wavelength scan at constant angle) have equal sensitivity in biosensing (26). The choice of the operating mode is governed more by the availability of suitable light sources and detectors than by any fundamental difference between the two modes of operation. Goddard et al. demonstrated a grating modified resonant mirror for chemical/biosensing application (27). This was achieved by fabricating a grating (closely-spaced strips) parallel to the direction of the incoming light, which allows miniaturization of the RM optical setup and eliminates the need for polarized light, as on resonance there is a π phase difference between the light reflected from the stripes and the area between the stripes (27). However, the IAsys instrument commercially available uses the prism coupling and angle scanning mode (monochromatic light). The resonant mirror modes are dispersive as well as leaky, causing the coupling angle for a particular mode to shift as the illumination wavelength changes. This is a particular problem when using coherent illumination light sources e.g., laser diodes, as these are prone to mode hopping as the diode temperature changes. This leads to sudden changes in the wavelength of the

92

Zourob et al.

(a)

H

I

F

J

E

K

G

D C B

A

60mm

100mm

100mm

100mm

(b) H

I

J

E C

L G

B

A

M K

Fig. 2. Schematic diagram of the resonant mirror configurations for (a) angular scan (b) wavelength scan. A light source (LED or tungsten lamp); B collimating lens; C aperture stop; D filter; E input polarizer; F plano-convex cylindrical lens; G equilateral SF 10 prism; H RM sensor chip; I quarter wavelength plate; J output polarizer; K detector; L grating; and M collimating lens.

laser output, causing a similar abrupt change in the measured resonance angle. The internally-referenced resonant mirror has been developed to overcome problems in changing illumination wavelength (28, 29). In this approach, an additional buried RM waveguide layer was incorporated into the sensor structure. The buried waveguiding layer provides a second resonance that has the same sensitivity to wavelength changes as the conventional surface RM resonance, but is considerably less sensitive to surface refractive index changes (28, 29). 1.1.2. The Sample Cuvette

The IAsys resonant mirror chip is mounted onto a cuvette structure such that the sensing surface is exposed to the sample contained within the cuvette lumen (Fig. 1). The sample volume within the cuvette chamber for single channel is approximately 200 µL and 80 µL for the dual channel. The cuvette body is coated with a biocompatible coating to prevent the non-specific binding. A vibro-stirring mechanism and a vacuum aspiration system are

Label-Free Detection with the Resonant Mirror Biosensor

93

inserted into the cuvette after placement in the instrument. The vibro-stirrer is used to ensure efficient mass transport of materials to the sensor surface and rapid equilibration of the system. As the stirrer vibrates, liquid is forced towards the surface of the sample chamber and back up through the centre of the stirrer cone. The stirrer frequency is fixed at 126 Hz but it can be adjusted via the instrument operating software. Reagents and samples can be added either by a pipette or by a syringe using the auto-sampler. The samples can also be removed either by a pipette or by the vacuum-driven aspiration systemto waste bottles. As the refractive index measurements are temperature sensitive, the cuvette is clamped tightly against peltier-driven thermal pads when inserted in the instrument. Good thermal conductivity is ensured between the peltier pads and the sample by the configuration of the cuvette body. Thermal equilibration time is typically 5–10 min for a temperature change of + or −10 °C. The rapid temperature equilibration allows the real-time study of the effect of temperature in binding studies. The use of real-time reference signal subtraction in the RM experiments allows the elimination of baseline shifts due to operating temperature changes. Furthermore, temperature changes can be easily incorporated into IAsys auto + scripts. The use of the unique real-time decision making function (ASK) allows the IAsys software to wait for complete baseline re-equilibration to be established before continuing an experiment. 1.1.3. Data Acquisition

IAsys machines have two data acquisition software, FASTplot and GraFit, which enable rapid and comprehensive analysis of data. It has a number of templates that can be used for kinetic and affinity analysis of data obtained from the IAsys. The supplied templates enable users to go from binding curves to values for affinity and kinetic constants easily. In-built flexibility enables users to modify existing templates, for example, constraining a fit to a value of kdiss, to creating their own templates. The templates make use of hard coded equations, which are also supplied as separate items, to simplify the fitting. Using ‘hard-coding’ removes the requirement on users to supply initial estimates to the non-linear fitting engine. FASTplot enables the relevant data from one or more experiments to be accumulated. Coupled with the latest versions of the operating software, automatic region selection can be obtained from the Event log. Concentrations entered during the experiment are carried through, speeding up the processing. Linear correction to baselines can be applied to the data during the accumulation, and subtraction from control experiments can also be performed in FASTplot, prior to the data analysis. Once the data have been accumulated, the data to be fitted can be rapidly exported directly into a ready-made fitting template in GraFit, and results generated without further

94

Zourob et al.

user input. GraFit also enables users to generate their own templates and analysis methods. GraFit (Erithacus software) is a fullyfunctioning, spreadsheet-based graphing and data analysis software. 1.1.4. Sensitivity

IAsys instruments measure the mass bound at the sensor surface; therefore the sensitivity can be considered in terms of the lowest amount of mass detectable per unit area. From the calibration factors (200 arcsec = 1 ng/mm2 for matrix surfaces and 600 arcsec = 1 ng/mm2for planar surfaces) and the signal to noise, it is possible to calculate the sensitivity to mass of the IAsys range of instruments. As mentioned above, the minimum sensitivity of the system can be considered in terms of the minimum surface concentration that can be resolved. This will depend upon the smallest response that can be detected over and above that of the baseline. The latest range of IAsys instruments can be divided into two types, standard and enhanced. One source of noise is caused by disturbances to the laser light incident on the resonant mirror structure. The enhanced range of instruments have modifications that fully enclose the input light, shielding it from external influences such as convection currents, thereby reducing the background noise. From the calibration factors and the arc second response corresponding to two times the standard deviation of the baseline noise, we can calculate the minimum detectable surface concentration of a protein. With the matrix cuvettes the detection limit of the standard instrument is 0.4 pg/mm2 and that of the enhanced instrument is 0.1 pg/mm2. With the planar surfaces the minimum detection level is 0.14 pg/mm2 for the standard instrument and 0.04 pg/mm2 for the enhanced instrument. It is however important to remember that the matrix surface is threedimensional, and so in absolute terms will have a higher binding capacity per unit area of surface. As the detection method used in IAsys is mass based, the larger the molecule, the larger the response. This means that in order to directly detect small molecules (100s of Daltons molecular weight range), a larger number of binding sites and higher solution concentrations are required. Competition studies can be performed in which a small molecule is allowed to compete against a conjugate of the small molecule coupled to a large molecule such as a protein. The binding of the protein conjugate is then detected, giving an enhanced response. It is possible to ‘magnify’ a response by building up a complex on the sensor surface, increasing the apparent bound mass and thereby increasing the sensitivity of the measurement. This is possible by using a sandwich molecule, for instance, an antibody to a second epitope on the target molecule. The signal can be further enhanced if the second antibody is labeled with colloidal gold. The potential gain in sensitivity, in this case approximately 1,000-fold,

Label-Free Detection with the Resonant Mirror Biosensor

95

in using colloidal gold-labeled secondary antibodies is shown in Subheading 3.17. 1.1.5. Resonant Mirror Surfaces

To ensure the specificity of the measurement, a recognition species (ligand) is immobilized at the sensor surface. To facilitate the immobilization, the IAsys RM chips have been derivatized with different matrices, e.g., 3D-carboxy methyl dextran hydrogel layer, and planar surfaces derivatized with amino, carboxylate, biotin, or hydrophobic groups. To ensure efficient mass transport of materials to the sensor surface and rapid equilibration of the system, a vibrating stirrer assembly is inserted into the IAsys cuvette after placement in the instrument.

2. Materials 2.1. Components for RM Setup

1. 670 nm laser diodes or high-intensity red light emitting diodes. 2. 20 W tungsten–halogen lamp. 3. Plano-convex with 25 mm diameter, EFL: 60 mm (45-509, Edmund Optics, York, UK). 4. 350 µm pinholes (53 mm outer diameter and 1.2 mm minimum aperture (32-618, Edmund Optics, York, UK). 5. Input-polarizer with 25 mm diameter (47-216, Edmund Optics, York, UK). 6. Plano-convex cylindrical lens with 25 mm diameter, EFL 100 mm (46-018, Edmund Optics, York, UK). 7. SF10 prism (n = 1.72, Comar instruments, Cambridge, UK). 8. Index-matching fluid (Cargille Labs Refractive Index Liquid Series M RI 1.730). 9. Output polarizer with 25 mm diameter (47-216). 10. Convex lens with 25 mm diameter, EFL: 60 mm (45-509, Edmund Optics, York, UK). 11. Grating (a 1,200 lines/mm replicated Ruled grating (364323) supplied by Ealing Electro-Optics (Watford, Hertfordshire, UK).

2.2. Materials for the Preparation of Dextran Coated Surfaces and Immobilization of Amine-Containing Ligand to CMD

1. Dextran MW 45,000 (Sigma)(see Note 1). 2. 0.1 M and 2 M sodium hydroxide. 3. 1 M bromoacetic acid. 4. 500 mL phosphate buffered saline pH 7.4, 0.05% Tween 20 (PBST).

96

Zourob et al.

5. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 6. NHS stock (200 mg in 15 mL high-purity water, (see Note 2). 7. 5 mL 10 mM sodium acetate, pH 5.0. 8. 200 µL 0.05 mg/mL IgG in PBS. Storage buffers must be free of reducing agents, Tris and azide. 9. Blocking agent: 5 mL 1 M ethanolamine, pH 8.5; or 1 mM Tris–HCl, pH 8.0. 10. 25 mL 10 mM HCl. 2.3. Materials for Immobilization of Thiol-Containing Ligand to CMD

1. CMD chip (Neosensors, Durham, UK, CUV991100). 2. 500 mL 0.01 M phosphate buffered saline pH 7.4, 0.05% Tween 20 (PBST) (Sigma, P3563). 3. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 4. NHS stock (200 mg in 15 mL high-purity water), (see Note 2). 5. 1 M ethylenediamine, pH 8.5 (Sigma, E1521). 6. 1 mg/mL Sulfo-SMCC in PBS (Pierce, 22322). 7. 1 mg/mL Cys-GM-CSF, thiol containing peptides in PBS.

2.4. Materials for Polymerizing the Glutaraldehyde and Immobilization on Amino-Silane Functionalized Surface

1. Amino-silane (AS) chip (Neosensors, Durham, UK, CUV991400). 2. 5 mL of 5% (v/v) glutaraldehyde (Sigma, G5882). 3. 500 µL 0.1 M NaOH. 4. 500 µL 0.1 M HCl. 5. 500 mL phosphate buffered saline pH 7.7 (PBS). 6. 200 µL 1 mg per/mL Rabbit anti-mouse Fc (RAMFc) (ICN Flow). 7. 25 mL 1 M formic acid. 8. 500 µL 1 mg/mL BSA in PB.

2.5. Materials for Immobilization of Amine-Containing Ligand to Carboxylate Surface

1. Carboxylate functionalized chip (Neosensors, Durham, UK, CUV991300). 2. 500 mL of phosphate buffered saline (PBS) pH 7.4,without Tween. 3. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 4. NHS stock (200 mg in 15 mL high-purity water), (see Note 2). 5. 5 mL 10 mM sodium acetate, pH 5.5. 6. 200 µL of 1 mg/mL Rabbit anti-mouse Fc (RAMFc). Storage buffers must be free of reducing agents, Tris and azide. 7. 200 µL of 2 mg/mL β-casein (Sigma, C6905). 8. 25 mL of 1 M formic acid.

Label-Free Detection with the Resonant Mirror Biosensor

2.6. Materials for Immobilization of Thiol-Containing Ligand to Carboxylate Surface

97

1. Carboxylate functionalized chip (Neosensors, Durham, UK, CUV991300). 2. 500 mL phosphate buffered saline pH 7.4, (PBS). 3. 10 mL phosphate buffered saline pH 7.4, 0.05% Tween 20 (PBST). 4. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 5. NHS stock (200 mg in 15 mL high-purity water), (see Note 2). 6. 5 mL of 1 M ethylenediamine, pH 8.5 (Sigma, E1521). 7. 5 mL of 1 M Tris pH 8.5. 8. 1 mg/mL Sulfo-SMCC in PBS (Pierce, 22322). 9. 1 mg/mL Cys-GM-CSF, thiol-containing peptides in PBS.

2.7. Materials for Immobilization of Avidin and Biotinylated Ligand Capture

1. Biotin functionalized chip (Neosensors, Durham, UK, CUV991200). 2. 500 mL phosphate buffered saline (PBS) pH 7.4, 0.05% Tween 20 (PBST). 3. 200 µL of 2 mg/mL avidin (Pierce, 21121) dissolved in highpurity water. 4. 200 µL of 0.2 mg/mL biotinylated protein G (Pierce, 2988) dissolved in high-purity water. 5. 200 µL of 0.1 mg/mL IgG dissolved in PBST, pH 7.4.

2.8. Materials for Immobilization of Thiol Containing Ligands to Avidin

1. Biotin functionalized chip (Neosensors, Durham, UK, CUV991200). 2. 500 mL of 20 mM HEPES-KOH, 138 mM NaCl, 2.7 mM KCl, 2 mM EDTA, pH 7.4 plus 0.05% Tween-20. (HSTE). 3. 100 µL of 5 mg/mL avidin (Pierce, 21121) dissolved in highpurity water. Streptavidin and NeutrAvidin from Pierce also can be used instead of avidin. 4. 500 µL of 50 mM N-succinimidyl 3-(2-pyridyldithio) propionate (SPDP) (Sigma, P-3415 or Pierce, 21857) dissolved in DMSO. 5. 5 mL of 10 mM sodium acetate pH 5.5. 6. 25 mL of 100 mM HCl. 7. Reduced and thoroughly desalted IgG at 0.1 mg/mL was used for this example. Storage buffer must be free of reducing agents, Tris and azide. 8. 2 mg/mL IgG dissolved in HBSTE. 9. D-salt Dextran plastic desalting column (Pierce, 43230).

2.9. Materials for Immobilization of Amine Containing Ligands to Avidin

1. Biotin functionalized chip (Neosensors, Durham, UK, CUV991200). 2. 500 mL of 20 mM HEPES-KOH, 138 mM NaCl, 2.7 mM KCl, 2 mM EDTA, pH 7.4 plus 0.05% Tween-20. (HBSTE).

98

Zourob et al.

3. 100 µL of 5 mg/mL avidin (Pierce, 21121) dissolved in high-purity water. Streptavidin and NeutrAvidin can be used instead of avidin. 4. 250 µL of 50 mM N-sulfosuccinimidyl 6-(3′-(2-pyridyldithio)propionamido) hexanoate (Sulfo-LC-SPDP) (Pierce, 21650) dissolved in water. 5. 5 mL of 10 mM sodium acetate pH 5.0. 6. 25 mL of 100 mM HCl. 7. 0.5 mL of 5 mM D-biotin dissolved in HBSTE. 8. 5 mM dithiothreitol (DTT) in water (for amine coupling option). 9. 1 M ethanolamine, pH 8.5. 10. 0.2 mg/mL IgG was used as example. Storage buffer must be free of reducing agents, Tris, and azide. 2.10. Materials for Formation of a Lipid Monolayer on Hydrophobic Surface

1. Hydrophobic chip ( (Neosensors, Durham, UK, CUV991500). 2. 500 mL of phosphate buffered saline (PBS) containing 0.025% (w/v) sodium azide and 1 mM EDTA (PBS/AE) pH 7.4. (Sigma, P-3813). 3. 25 mL of 0.1 M HCl. 4. 25 mL of 10 mM NaOH. 5. 50 mL of 2-propanol HPLC grade (Aldrich, 27049-0). 6. 200 µL of 20 mg/mL Dioleoylphosphatidylcholine (DOPC) (Sigma, P-7212) in chloroform containing 0.1% (w/w) butylated hydroxytoluene, store at −18 °C or below. 7. 200 µL of 20 mg/mL Bovine serum albumin (BSA) (Sigma, A-7888) in PBS/AE, store at −18 °C.

2.11. Materials for Immobilizing Phospholipids on Non-Derivatized Surface

1. Non-derivatized CUV991600).

chip

(Neosensors,

Durham,

UK,

2.12. Materials for Capture of His-Tagged Proteins

1. Carboxylate functionalized chip (Neosensors, Durham, UK, CUV991300).

2. 500 mL of phosphate buffered saline (PBS) pH 7.4. 3. 50 mL of PBS + 1.25% OG (PBS/OG). 4. 5 mL of 2 mg/mL PLs in PBS/OG. In this example a 50:50% mix of dimyristoyl-phosphatidylethanolamine (DMPE) and dioleoyl-phosphatidylglycerol (DOPC).

2. 500 mL of 10 mM HEPES, 0.15 M NaCl, 50 µM EDTA, 0.005% Tween 20, pH 7.4, (HBSTE). 3. 30 mL of 50 mM HEPES buffer pH 8.5 (HEPES). 4. 5 mL of 5 m 00 µM NiCl2 in HBSTE.

Label-Free Detection with the Resonant Mirror Biosensor

99

5. 30 mL of 1 M ethanolamine pH 8.5. 6. 30 mL of 50 mM phosphate buffer, 10% glycerol, 30 mM imidazole, 300 mM NaCl, pH 6.8, (PGI wash buffer). 7. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 8. NHS stock (200 mg in 15 mL high-purity water), (see Note 2). 9. 10 mg/mL NTA (Qiagen, 34491) in 380 µL HEPES buffer. 10. His-tagged protein. 2.13. Materials for Capture of FLAG-Tagged Proteins

1. Carboxylate functionalized chip (Neosensors, Durham, UK, CUV991300). 2. 30 mL of 10 mM acetate buffer pH 4.5. 3. 25 mL of 10 mM HCl (regenerant). 4. Anti-FLAG M2® monoclonal antibody (Sigma, F3165). 5. An irrelevant monoclonal antibody as a control. 6. FLAG peptide (Sigma, F3290). 7. FLAG-BAP control protein (Sigma, P7457). 8. 500 mL of 50 mM phospate buffer saline, 0.05% Tween-20, buffer pH 7.4 (PBS/T). 9. 30 mL of 1 M ethanolamine pH 8.5. 10. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 11. NHS stock (200 mg in 15 mL high-purity water), (see Note 2).

2.14. Materials for Preparation of Colloidal Gold-Protein Complexes

1. 100 mL of 25 mM potassium carbonate. 2. 10 mL of 10% (w/v) BSA adjusted to pH 9 with NaOH, filtered. 3. Protein to complex in salt free solution. 4. 30 nm colloidal gold sol (Biocell, Cardiff, UK). 5. Rotary mixer and microcentrifuge.

2.15. Materials for Determination the Concentration of Theophylline in Buffer and Serum

1. CMD chip (Neosensors, Durham, UK, CUV991100). 2. 500 mL 10 mM phosphate buffered saline (PBS), 2.7 mM KCl, 137 mM NaCl, pH 7.4, 0.05% Tween. PBS/T). 3. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 4. NHS stock (200 mg in 15 mL high-purity water), (see Note 2). 5. 5 mL 10 mM sodium acetate, pH 4.5. 6. 10 µg/mL Rabbit anti-mouse Fc (RAMFc) were supplied by FAST. 7. Thyophyline-8-butyric acid lactam (Calbiochem) (theophyllineprotein-conjugate) were supplied by FAST.

100

Zourob et al.

8. 100 µg/mL MAb was from Biogenesis (Bournemouth, UK). 9. Prepare solutions of theophylline (Aldrich, Dorset, UK) in the range 0–40 µg/mL in PBS/T. 10. Mix theophylline solutions with equal volumes of theophyllineprotein-conjugate solution (final concentration 200 µg/mL). 11. Prepare solutions containing only conjugate. 12. Prepare solutions of theophylline (Aldrich, Dorset, UK) in the range 0–40 µg/mL in serum and dilute it in 1:10 in PBS/T prior mixing with theophylline-protein-conjugate. 2.16. Materials for Protein–Carbohydrate Recognition

1. CMD chip (Neosensors, Durham, UK, CUV991100). 2. 500 mL 10 mM phosphate buffered saline (PBS), 138 mM NaCl, 2.7 mM KCl, pH 7.4, 0.05% Tween. (PBS/T). 3. 2 M NaCl, 10 mM Na2HPO4, pH 7.2. (used as generation buffer between analysis of different samples of the same ligate). 4. 2 M guanidine HCl (Post regeneration when different ligate type was used). 5. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 6. NHS stock (200 mg in 15 mL high-purity water), (see Note 2). 7. N-hydroxysuccinimide amino caproate (LC) biotin (Pierce). 8. Human bFGF was purified as described (30). Different concentrations were prepared (20–330nM). 9. Heparin from porcine intestinal mucosa (Sigma). 10. Streptavidin (Sigma). 11. bFGF peptides (−55 to −43, 52–61, 117–126, and 127–140) were synthesized and purified as described in (31). Different concentrations were prepared (2–93 µM). 12. Amino groups of heparin were biotinylated as described (31).

2.17. Materials for Nucleic Acid Hybridization

1. CMD chip (Neosensors, Durham, UK, CUV991100) 2. 200 µg/mL streptavidin (Sigma) in 10 mM acetate buffer pH 5.0. 3. 20 mM HCl. Biotinylated oligonucleotides were prepared using biotin phosphoramidite using Applied Biosystem Synthesiser resulting in the biotinylation at the 5′ end. 4. Oligo-1. A biotinylated 19mer oligonucleotide with 53% GC content.Tm = 58 °C (1). 5. Oligo-2. A 40mer oligonucleotide with 50% GC content, containing at its 3′ end a 19 base sequence complementary to oligo-1. 6. Oligo-3. A 50-mer oligonucleotide with no complementary to oligo-1. Approximately 50% GC content.

Label-Free Detection with the Resonant Mirror Biosensor

101

7. Oligo-4. A biotinylated 19-mer oligonucleotide with no complementary to oligo-2. 53% GC content. 8. 500 mL of 10 mM phosphate buffered saline (PBS), 2.7 mM KCl, 137 mM NaCl, pH 7.4, 0.05% Tween, PBS/T). 2.18. Materials for Fermentation Monitoring

1. CMD chip (Neosensors, Durham, UK, CUV991100). 2. 500 mL of 10 mM phosphate buffered saline (PBS), 0.5 M NaCl, 2.7 mM KCl, pH 7.4, 0.05% Tween. (PBS/T). 3. 10 mM acetate buffer pH 5. Acetate buffer was made by titrating sodium acetate (AnalaR grade) with 2 M acetic acid. 4. Hen leg lysozyme (Sigma, L-7001). 5. D1.3 Fv produced by batch fermentation of E. coli following a protocol described in (32).

2.19. Materials for Bacterial Cells Detection

1. AS chip (Neosensors, Durham, UK, CUV991400). 2. 500 mL 10 mM phosphate buffered saline (PBS), 137 mM NaCl, 2.7 mM KCl, pH 7.4. 3. Staphylococcus aureus strain (Cowan-1) and (Wood 46) (Sigma). 4 Human immunoglobulin IgG (ICN). 5. Bovine serum albumin (BSA) (Sigma). 6. 25% Glutaraldehyde in water (Sigma). 7. Colloidal gold sol (30 nm diameter) (Biocell). 8. Prepare colloidal gold human IgG conjugate as described in Subheading 6.3.9 (33). Dilute the colloidal gold-IgG particles to a constant O.D.530 in PBS buffer. 9. Haemocytometer for determining cell concentrations in stock suspension.

2.20. Materials for Receptor–Cell Interactions

1. AS chip (Neosensors, Durham, UK, CUV991400). 2. Bis[sulfosuccinimidyl] suberate (BS3) (Pierce, 21580). 3. 1 mg/mL BS3 in PBS (PBS/BS3). 4. 500 µL 0.1 M HCl. 5. 500 mL 10 mM sodium phosphate buffered saline, pH 7.4, 134 mM NaCl, 3 mM KCl, (PBS). 6. 500 mL 10 mM sodium phosphate buffered saline, pH 7.4, 134 mM NaCl, 3 mM KCl, (PBS/T), 0.05% Tween 20. Tween 20 (Pierce and Wariner, Surfact-Amps® 20, 28320). 7. 100 µg/mL BSA in PBS (Sigma, A-6003). 8. Blocking solution 2 mg/mL BSA in PBS. 9. 130 µg/mL PR3B10 antibody (34) in PBS. 10. Cells were grown in DMEM medium with 10% v/v fetal calf serum and re-suspended in PBS/BSA. Murine L cells,

102

Zourob et al.

co-transfected with a full length CEA cosmid clone (35), and pSVneo, selected for neomycin resistance, were used as a source of CEA bearing cells. 11. Untransformed cells were used as control. 2.21. Materials for Determination of the Kinetic Constants

1. CMD chip (Neosensors, Durham, UK, CUV991100). 2. 500 mL 10 mM phosphate buffered saline (PBS), 137 mM NaCl, 2.7 mM KCl, pH 7.4, 0.05% (v/v) Tween 20. (PBS/T). 3. Staphylococcus aureus protein A (Sigma, P 3838). 4. Prepare different concentrations (2–60 nM) of human immunoglobulin IgG (ICN) in PBS/T. 5. Bovine serum albumin (BSA) (Sigma). 6. 10 mM acetate buffer pH 4.5 was prepared by titrating sodium acetate (AnalaR grade) with 2 M acetic acid. 7. EDC stock (1.15 g in 15 mL high-purity water), (see Note 2). 8. NHS stock (200 mg in 15 mL high-purity water), (see Note 2).

3. Methods 3.1. RM Sensors Chip Fabrication

The resonant mirror sensor chips were fabricated on 1 mm thick Schott SF10 glass substrates, optically polished on both sides (Gooch and Housego Ltd, Illminster, UK). The substrates were cleaned successively in Decon-90 solution and acetone, then dried at 80 °C for 30 min. Deposition of silica spacer layers and titania–hafnia waveguide layers was performed using e-beam assisted vacuum deposition. The sensor substrates were cut into rectangles 10 × 12.5 mm by scribing and fracturing after deposition of the spacer and waveguide layers.

3.2. RM Setup Instrumentation

Figure 2a, b show a schematic of the optical arrangement used for the angular and wavelength scan instrumental configurations. The same optical arrangement used has been detailed elsewhere in more detail (1, 2, 19–26). For the angular scan instruments, a monochromatic light source, such as 670 nm laser diodes or high-intensity red light emitting diodes with an interference filter (see Note 3), was used and a low power (20 W) tungsten–halogen lamp with suitable filters for the wavelength scan (see Note 4). On the input side, a light source (LED or halogen–tungsten lamb), with some collimating lens (plano-convex with 25 mm diameter, EFL: 60 mm) to provide a ‘clean’ beam of light and 350 µm pinholes (53 mm outer diameter and 1.2 mm minimum aperture) was used. This was passed through an optional filter (see Note 5),

Label-Free Detection with the Resonant Mirror Biosensor

103

then an input-polarizer with 25 mm diameter, oriented to provide equal TE and TM intensities. In the angular scan mode, the input beam is then passed through a plano-convex cylindrical lens with 25 mm diameter, EFL 100 mm to provide a converging wedge beam, whereas in the wavelength scan mode, the collimated beam is unmodified. The incoming beam then passes into a high-index (SF10) prism (n = 1.72, Comar instruments, Cambridge, UK) and into the sensor substrate via an index-matching fluid (Cargille Labs Refractive Index Liquid Series RI = 1.730). On the output side, the beam is passed through a quarter-wavelength plate, which acts to remove any additional ‘background’ phase shift between TE and TM modes. The beam then passes through the output polarizer with 25 mm diameter, which is oriented at about 90° to the input polarizer, to remove all but the light, which has undergone resonance. For the angular scan instrument, this then results in a substantially parallel beam of light covering a narrow range of angles, which can be focused on to a detector using a convex lens (25 mm diameter, EFL:60 mm). As the surface conditions change, the output angle changes and hence the position of the peak on the detector changes. The wavelength scan instrument is more complex. The output beam is still well collimated, but only contains a narrow range of wavelengths, corresponding to those wavelengths that undergo resonance at the fixed input angle. The output beam is allowed to fall on to a grating (a 1,200 lines/mm replicated ruled grating (Ealing Electro-Optics (Watford, Hertfordshire, UK), and the resulting diffracted beam is then focused on to a detector. In both instances, as the surface refractive index changes, the position of the peak on the detector changes. Two kinds of charged coupled device (CCD) detectors were employed. For the angular scan, a linear CCD detector was used to monitor the device output, consisting of a 5,000-pixel CCD (Sony ILX506A, pixel pitch 7 µm). The CCD output was digitized to a 12-bit resolution and collected via the parallel port of a host computer with software written in-house using C++. For the wavelength scan, a 2-D CCD detector with associated image processing functions was used, based on a Sony ICX039AL CCIR standard monochrome camera chip, resolution 752 horizontal pixels by 582 lines. The output from the CCD sensor was digitized to 8-bit resolution using a flash analogue-to-digital converter at 14.1875 MHz, and passed to a hard-wired signal averager, which allowed up to 256 successive fields to be averaged together to improve the signal-to-noise ratio of the input. The pixel data were stored in 1 MB of fast (35ns) static random access memory (RAM), which could also be accessed by the programmable image processing system. This was based on an Inmos 20 MHz T805 transputer, with 4 MB of program RAM and a high-speed serial link to a host computer. The transputer-based

104

Zourob et al.

image processing system eliminated the need for large amounts of data to be transferred to the PC for processing. 3.3. Carboxy Methyl Dextran

This is a three dimensional matrix of 200 nm thick immobilized carboxymethylated dextran on the sensor surface. The dextran is a polymer of glucose residues with a low degree of branching, low polydispersity and average weight of approximately MW = 500,000. The dextran is carboxylated to give approximately one carboxyl group (pKa = 3.5) per two glucose residues. Hence, the Carboxy Methyl Dextran (CMD) surface is negatively charged in typical buffers used that have a pH > 5. If the pH of the buffer is less than the isoelectric point (PI) of a protein, the protein will exhibit a net positive charge so it will promote its electrostatic pre-concentration into the dextran prior to covalent immobilization. The CMD enables simple coupling of amino-containing molecules to the carboxylate group of the matrix through succinimidyl ester chemistry resulting in the formation of a peptide bond between the amino-containing molecules and the matrix (Fig. 3a). The advantage of the CMD coated surface is that the 3-D nature of the hydrophilic matrix increases the ligand immobilization loading of the system compared to a planar layer. In addition, it keeps immobilized ligands away from direct contact with the surface, resulting in greater retention of the biological activity of biomolecules, reducing protein denaturation and non-specific adsorption. The disadvantage of this surface is that the orientation of immobilized recognition molecules will not be uniform due to the random nature of the immobilization chemistry. The second disadvantage is that this matrix can only be used for small analytes as the large analytes such as cells may be size-excluded. Thirdly, the molecule’s binding will not be sensed equally in the matrix due to the exponential nature of the evanescent field. In addition, from the kinetics point of view, molecules bound to sites farthest from the surface may sterically hinder the binding of other molecules to sites within the matrix (36, 37).

3.3.1. Procedure for Dextran Coating

1. Silanized chips are immersed overnight in 23% (w/w) dextran solution in 0.1 M NaOH (see Note 1). 2. The following day the chip is washed extensively with water to remove excess dextran. 3. The chip is treated with 1 M bromoacetic acid in 2 M NaOH for 6 h. 4. Rinse the chip extensively with water and dry; then it is ready to use.

3.3.2. Procedure for the Immobilization of AmineContaining Ligand to CMD

This protocol describes the immobilization of the protein on a dextran coated cuvette in the IAsys instrument. 1. Insert a CMD cuvette into the instrument and wash with PBST for 10 min for equilibration.

OH

(a) NHS

N

O

O NH2

+

EDC

CH3

Cl

N C N

O

+

O

OH O

+

NH

O O

OH OH

O

N

CH3

O

O

OH

OH OH

OH

NH

O

O

O

OH

O

OH OH

OH

OH

OH

CMD coated sensor surface O

(b)

O

N

S

N

O

O

Amino-modified CM-Dextran O

NH2

O

HN

HO

NH

O

O

O OH

HN

HO

OH

O

O

OH

O OH

NH

HO

OH

OH

O

Sulpho-SMCC

O OH

NH

O

O H

OH

SH2

CMD coated sensor surface

(c) (1)

H

Basic pH

OH

O

O

H

O

O

O

H

Glutaraldehyde

polymerized glutaraldehyde

(2) H

OH

O

O

O

H

(c)

O

H

+

H

O HN

H H

O

O

O

H NH2 NH2

NH2

NH2

H

NH2 NH

O

O

O

H

H NH2

NH2

H NH2

NH2 NH

NH2

NH2

Fig. 3. (a) Reaction scheme for carboxymethyl dextran with EDC/NHS for the immobilization of amine containing ligands. (b) Reaction scheme for Sulpho-SMCC modified immobilization of thiol containing ligand. (c) Reaction scheme (A) polymerization of glutaraldehyde and (B) the cross-linking of a ligand to an amino modified sensor surface.

106

Zourob et al. OH

(d) NHS

N

O

O

+ EDC

Cl

N C N

CH3

NH

+ O

HO OH

CH3

O N O

OH

O

O OH

OH

(e)

O

O

OH

S

O

N

O

O

HN

OH

OH

O

NH2

OH O

O

O

N

NH O

OH OH

O

O

NH

O

NH2

OH

O

HN

NH O

OH O

OH

Sulpho-SMCC

O

NH O

OH O

OH

SH2

Amino-modified carboxyle groups

(f)

Biotinylated ligand

Streptavidin

Biotinylated surface

Fig. 3. (continued) (d) Reaction scheme for carboxylate surface with EDC/NHS for the immobilization of amine containing ligands. (e) Reaction scheme for sulpho-SMCC mediated immobilization of thiol containing ligands. (f) Schematic diagram of capturing biotinylated ligand onto biotin functionalized surface via streptavidin.

Label-Free Detection with the Resonant Mirror Biosensor

(g)

107

O N O O N

OH

O

O

N S

S

O

N S S

+

S S

O

NH2 NH2 NH2

NH2

NH2 HN

NH2

O

NH2

HS

NH2 HN

NH2

NH2

Fig. 3. (continued) (g) Schematic reaction diagram of surface amines with SPDP, and subsequent reaction with free, accessible thiol from protein. (h) Formation of a lipid monolayer from solution onto hydrophobic sensor surface.

2. Start data acquisition and collect baseline data for 3 min. 3. Introduce a mixture of 200 µL of EDC and 200 µL of NHS for 7 min (see Notes 2 and 6). 4. Wash with PBST and collect baseline data for 3 min (see Note 7). 5. Change the PBST to 10 mM acetate buffer pH 5.5 to get a baseline (see Note 8). 6. Introduce IgG in acetate buffer to the cuvette for 5 min to begin protein electrostatic uptake prior covalent binding. 7. Wash with PBST and collect baseline data for 3 min. 8. Block the unreacted NHS-ester with 1 M ethanolamine pH 8.5 for 3 min (see Note 9). 9. Wash with PBST to get the baseline. 10. Wash the cuvette with 10 mM HCl for 3 min to remove the non-covalently bound ligand to the matrix. 11. Wash with PBST and collect baseline data for 3 min. 12. Calculate the amount of immobilized protein by subtracting the baseline level at step 4 from that at step 11.

108

Zourob et al.

3.3.3. Procedure for the Immobilization of Thiol-Containing Ligand to CMD

Protocol

This protocol describes a method for immobilization of a thiol containing ligand to CMD coated sensor surface as shown in Fig. 3b. This approach will include the modification of the CMD with a primary amine, followed by reaction with an amine–thiol reactive heterobifunctional cross-linking agent Sulfosuccinimide 4-(N-maleimidomethyl)-cyclohexane-1-carboxylate (Sulfo-SMCC). Covalent immobilization is achieved by the direct addition of thiol-bearing ligands to the modified sensor surface. Thiol groups offer a good approach to achieving orientation of a ligand, especially with cysteine incorporation into one terminal of a peptide. 1. Insert CMD cuvette into the instrument and wash with 200 µL of PBST and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Incubate the surface with 50 µL of a mixture of 1:1 (v/v) EDC/ NHS for 10 min (see Notes 2 and 6). 4. Wash with 200 µL of PBST and collect baseline data for 3 min. 5. Incubate the surface again twice with 50 µL of a mixture of 1:1 (v/v) EDC/ NHS for 7 min. 6. Wash with 200 µL of PBST and collect baseline data for 3 min. 7. Incubate the surface with 50 µL of 1 M ethylene diamine pH 8.5 for 10 min to modify the carboxymethyl groups with amines. 8. Wash with 200 µL of PBST and collect baseline data for 2 min. 9. Activate the amine group with Sulfo-SMCC by washing and incubating the surface with 200 µL of 1 mg/mL SulfoSMCC for 10 min. 10. Wash with 200 µL of PBST and collect baseline data for 2 min. 11. Incubate the surface with 50 µL of Cys-GM-CSF to covalent binding for 8 min. 12. Wash with 200 µL of PBST and collect baseline data for 2 min. 13. Calculate the amount of immobilized protein by subtracting the baseline level after step 10 from that after step 12.

3.4. Aminosilane Functionalized Surface

The biosensor surface is coated with an aminosilane (AS) derivative, resulting in the introduction of amino groups to the surface via an alkyl chain linker. The amino derivatized surfaces are particularly useful for immobilization of ligands with PI < 4, where methodologies based on electrostatic concentration into a CMD matrix are not effective. The advantage of this chemistry is that

Label-Free Detection with the Resonant Mirror Biosensor

109

it allows large ligates such as membrane fragments, liposomes, viruses and cells which are unable to penetrate the 3D-hydrogel matrix to enter the most sensitive area of the evanescent field. Highly charged molecules can be used with this surface, which might be electrostatically repelled from the CMD matrix. This surface can be activated with cross-linkers such as glutaraldehyde or bis(sulfosuccinimidyl)suberate. The disadvantage of this method is that there is very little electrostatic concentration of ligand to the surface unlike the charged CMD matrix. This requires higher concentrations of ligand to give satisfactory surface coverage. Typically, concentrations in the range of 0.1–1 mg/mL are required. This protocol will describe immobilization of ligands via their amino groups to the surface amino-functionalized groups using polymerized glutaraldehyde (Fig. 3c). 3.4.1. Procedure for Polymerizing the Glutaraldehyde

1. Polymerized glutaraldehyde (PG) is prepared by adding 500 µL of 0.1 M NaOH to 5 mL of 5% (v/v) glutaraldehyde as shown in Fig. 3c(A). 2. Leave for 30 min to polymerize before neutralizing with 500 µL 0.1 M HCl. Presence of PG can be determined from a wavelength scan between 200 and 350 nm. An absorption peak is seen around 234 nm, which represents the polymer, whereas a peak at 280 nm is the monomer. The resulting polymerized glutaraldehyde solution must be stored at or below −18 °C.

3.4.2. Immobilization Procedure

1. Insert new AS cuvette into the instrument and wash with PBS for 10 min for equilibration (see Note 10). 2. Start data acquisition and collect 3 min of baseline data. 3. Introduce PG to the cuvette for 30 min. 4. Wash with PBS to get stable baseline. 5. Introduce the ligand solution 100 µg/mL and leave it for about 30 min (see Note 11). 6. Wash the cuvette with 1 M formic acid for 2 min to remove the non-covalently bound ligand to the surface. 7. Wash with PBS for about 5 min to get the baseline. 8. Wash the cuvette with 1 mg/mL BSA for 5 min to block unreacted PG (see Note 12). 9. Wash with PBS to get the baseline. 10. Calculate the amount of immobilized protein by subtracting the baseline level after step 4 from that after step 9.

3.5. Carboxylate Surface

The planar carboxylate surface allows the user to analyze biomolecular interactions in the absence of the 3-D CMD hydrogel layer. This surface is useful for large molecules (> 1,000 KDa),

110

Zourob et al.

particles or cells binding to the recognition molecules closer to the surface, where the evanescent field is more intense. Amino and thiol containing ligands can be immobilized on the carboxylic acid functionalized surface. 3.5.1. Immobilization of Amine-Containing Ligand to Carboxylate Surface

Protocol

The biosensor surface contains carboxylic acid functional groups that can react with the primary amine of the ligand via amide bond using EDC/NHS chemistry (Fig. 3d) as described in Subheading 3.13.1. 1. Insert carboxylate cuvette into the instrument and wash with PBS for 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Introduce a mixture of 200 µL of EDC and 200 µL of NHS for 7 min (see Notes 2 and 6). 4. Wash with PBST to get stable baseline. 5. Change the PBST to 10 mM acetate buffer pH 5.5 to get stable baseline (see Note 8). 6. Introduce RAMFc in acetate buffer to the cuvette for 10 min to begin electrostatic uptake of protein prior covalent coupling. 7. Wash with PBS for 3 min to get stable baseline (see Note 11). 8. Block un-reacted NHS-ester and uncovered surface with 2 mg/mL β-casein for 3 min. 9. Wash with PBS and collect 3 min of baseline data (see Note 12). 10. Wash residual non-covalently bound ligand with 1 M formic acid for 2 min. 11. Wash with PBS for 3 min to get the baseline. 12. Calculate the amount of immobilized protein by subtracting the baseline level after step 4 from that after step 11.

3.5.2. Immobilization of Thiol-Containing Ligand to Carboxylate Surface

Protocol

This protocol describes a method for immobilization of a thiol containing ligand to carboxylate functionalized sensor surface as shown in Fig. 3e. This approach includes modification of the carboxylate modified surface with a primary amine, followed by a reaction with an amine–thiol reactive heterbifunctional cross-linking agent Sulfosuccinimide 4-(N-maleimidomethyl)cyclohexane-1-carboxylate (Sulfo-SMCC). Covalent immobilization is achieved by the direct addition of thiol-bearing ligands to the modified sensor surface. Thiol groups offer a good approach to achieving orientation of a ligand, especially with cysteine incorporation into one terminal of a peptide. 1. Insert new carboxylate cuvette into the instrument and wash with 200 µL of PBS and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data.

Label-Free Detection with the Resonant Mirror Biosensor

111

3. Wash and incubate the surface with 50 µL of a mixture of 1:1 (v/v) EDC/ NHS for 10 min (see Notes 2 and 6). 4. Wash with 200 µL of PBS and collect baseline data for 3 min. 5. Wash twice again and incubate the surface with 50 µL of a mixture of 1:1 (v/v) EDC/ NHS for 7 min. 6. Wash with 200 µL of PBST and collect baseline data for 3 min. 7. The activated carboxyl group is then converted to amines by washing and the surface is incubated with 50 µL of 1 M ethylene diamine pH 8.5 for 10 min. 8. Wash with 200 µL of PBS and collect baseline data for 5 min. 9. Block any remaining activated carboxyl groups by washing twice and incubating the cuvette with 200 µL of 1 M Tris pH 8.5 for 5 min. 10. Wash with 200 µL of PBS and collect baseline data for 5 min. 11. Wash extensively with 200 µL of PBST and collect baseline data for 5 min. 12. Activate the amine group with Sulfo-SMCC by washing and incubating the surface with 200 µL of 1 mg/mL SulfoSMCC for 10 min. 13. Wash with 200 µL of PBST and collect baseline data for 2 min. 14. Incubate the surface with 50 µL of Cys-GM-CSF to covalent binding for 10 min. 15. Wash with 200 µL of PBST and collect baseline data for 2 min. 16. Calculate the amount of immobilized protein by subtracting the baseline level after step 13 from that after step 15. 3.6. Biotin Surface

3.6.1. Immobilization Procedure of Streptavidin

This method is widely used for immobilization of oligonucleotides, DNA, antibodies, peptides and other proteins. Usually, in this process biotinylated biomolecules are attached to the biotin functionalized sensor surface by depositing avidin or streptavidin or NeutrAvidin before introducing the biotinylated biomolecules (Fig. 3f). The advantage of this method is the convenient and rapid immobilization of biotinylated molecules. Since the interaction between the avidin and biotin is non-covalent, the surface can be regenerated by completely removing the avidin-captured molecules. Two examples will be described here, of immobilizing biotinylated protein ligands (Subheading 3.6.2), and biotinylated DNA (Subheading 3.15) onto a biotin-functionalized surface using streptavidin. 1. Insert new biotin cuvette into the instrument and wash with PBST for 10 min for equilibration (see Note 13).

112

Zourob et al.

2. Start data acquisition and collect baseline data for 3 min. 3. Introduce avidin in PBST and allow the binding to occur for 10 min (see Note 14). 4. Wash with PBST to get a baseline (see Note 15). 5. Calculate the amount of immobilized avidin by subtracting the baseline level after step 2 from that after step 4. 3.6.2. Immobilization Procedure for Biotinylated Ligand

The following steps begin after the streptavidin coating and washing steps as described in Subheading 3.6.1. 6. Introduce biotinylated protein G in PBST and allow the binding to occur for about 5 min. 7. Wash with PBST for 3 min to establish a baseline. 8. Introduce IgG in PBST and allow binding to occur for about 5 min. 9. Wash with PBST for 5 min to get a baseline. 10. Calculate the amount of immobilized protein G by subtracting the baseline level after step 5 from that after step 7.

3.6.3. Immobilization of Thiol Containing Ligands to Avidin

Protein Preparation

This protocol describes a method for the immobilization of proteins through available surface thiol groups using an amine-to-thiol heterobifunctional linker to the amine groups in an immobilized avidin protein layer (Fig. 3g). The remaining carboxylates on the avidin layer then serves to electrostatically attract a protein ligand. Covalent coupling through its free thiol groups then follows. The use of this protein layer in the immobilization of biomolecules allows ligands of interest to be coupled through a variety of crosslinking chemistries. The advantage of using this protein layer is that it is suitable for the immobilization of biomolecules of all sizes and reduces the non-specific binding. 1. Dissolve 500 µL of 0.5 mg/mL IgG in HBSTE. 2. Reduce the IgG protein to have free thiol by addition of DTT to a final concentration of 10 mM, and room temperature incubation for 30 min. 3. Desalt the product using a D-salt™ column (Pierce, 43230) using HBSTE as the pre-equilibrium and running buffer.

Desalting Protein

1. Equilibrate the column with 200 mL of HBSTE at room temperature, then add 500 µL of IgG mixture to the column head. 2. After loading, fill the funnel with HBSTE. Collect a number of fractions of drops (260 µL) each; the protein will be present in fraction numbers 6–9. 3. Wash the columns with 150 mL of HBSTE to remove lowmolecular weight compounds prior to re-use.

Label-Free Detection with the Resonant Mirror Biosensor Immobilization Procedure

113

1. nsert new biotin cuvette into the instrument and wash with 200 µL of HBSTE and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Introduce 40 µL of avidin in HBSTE in the cuvette and allow binding to occur for about 3 min. 4. Wash with 200 µL of HBSTE and allow excess avidin to dissociate for about 3 min. 5. Wash extensively with 200 µL of 100 mM HCl for 2 min followed by washing with 150 µL of HBSTE for 3 min. 6. Add 60 µL of SPDP in HBSTE in the cuvette and allow the reaction to continue for 10 min. 7. Wash with 200 µL of HBSTE and collect baseline data for 3 min. 8. Wash with 200 µL of 10 mM acetate buffer pH 5.5, and establish baseline data for about 3 min. 9. Introduce 20 µL of protein containing free thiol (reduced IgG in this example) and incubate it for 15 min or desired level. 10. Wash with 200 µL of HBSE and collect baseline data for 3 min. 11. Wash with 200 µL of 100 mM HCl and incubate for 2 min to remove excess ligand. 12. Wash with 200 µL of HBSTE and collect baseline data for 3 min. 13. Add 600 ng/µL of a specific antibody to detect the level of binding to check if the immobilized protein is still active towards its cognate binding partner. 14. Then wash the free antibody from the cuvette and allow it to dissociate for 10 min. 15. Treat the surface with 100 mM HCl for 2 min to remove the bound antibody. 16. Wash and refill the cuvette with HBSTE to reveal a new baseline which matches that at step 14.

3.6.4. Immobilization of Amine Containing Ligands to Avidin

This protocol describes a method for the immobilization of proteins through available surface amino groups using an amine-to-thiol heterobifunctional linker to the amine groups in an immobilized avidin protein layer. The remaining carboxyls on the avidin layer then serve to electrostatically attract a protein ligand. Depending on the mode of activation, covalent coupling may then proceed through thiol groups as described in Subheading 3.6.3, or through the amine as described in this protocol. As previously mentioned, the use of a protein layer in the immobilization of biomolecules allows ligands of interest of all sizes to be coupled through a variety of cross-linking chemistries and reduces the non-specific binding.

114 Protocol

Zourob et al.

1. Insert new biotin cuvette into the instrument and wash with 200 µL of HBSTE and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Introduce 40 µL of avidin in HBSTE in the cuvette and allow binding to occur for about 3 min. 4. Wash with 200 µL of HBSTE and incubate to allow excess avidin to dissociate for about 3 min. Wash again with 200 µL of HBSTE to establish a baseline. 5. Block biotin binding sites by addition of 20 µL of 5 mM biotin for 1 min. 6. Wash extensively with 200 µL of 100 mM HCl for 2 min followed by washing with 150 µL of HBSTE for 3 min and incubate for 3 min to allow excess avidin to dissociate. Wash again with 200 µL of HBSTE to establish a baseline. 7. Add 60 µL of SPDP in HBSTE in the cuvette and allow the reaction to continue for 10 min. 8. Wash with 200 µL of HBSTE and collect baseline data for 3 min. 9. Wash with 200 µL of 5 mM DTT, and allow thiol reduction to proceed for 7 min. 10. Wash with 200 µL of HBSTE and collect baseline data for 3 min. 11. Repeat the addition of 200 µL of SPDP and allow the activation to continue for 10 min. 12. Wash with 200 µL of HBSTE and collect baseline data for 3 min. This baseline will be used as starting place for the calculation of net immobilized protein. 13. Wash with 200 µL of 10 mM acetate buffer pH 5.5, and establish baseline data for about 3 min. 14. Introduce 20 µL of protein (1 µg IgG as example for electrostatic uptake), and allow reaction to occur for about 15 min or desired level. 15. Wash with 200 µL of HBSE and collect baseline data for 3 min. 16. Wash with 200 µL of 100 mM HCl and incubate for 2 min to remove excess ligand. (Optional step as might be the protein will not withstand this treatment). 17. Wash with 200 µL of HBSTE and collect baseline data for 3 min. 18. Wash with 200 µL of 1 M ethanolamine, pH 8.5 for 3 min to block residual NHS esters. 19. Wash with 200 µL of HBSTE and establish baseline data for about 3 min.

Label-Free Detection with the Resonant Mirror Biosensor

115

20. Add 100 ng/µL of a specific antibody to detect the level of binding to check if the immobilized protein still active towards its cognate binding partner. 21. Then wash the free antibody from the cuvette and allow to dissociate for 5 min. 22. Treat the surface with 100 mM HCl for 2 min to remove the bound antibody. 23. Wash and refill the cuvette with HBSTE to reveal a new baseline. 3.7. Immobilization of Phospholipids on Hydrophobic Surface

Protocol

The RM hydrophobic cuvette provides a substrate onto which hydrophobic molecules can be easily deposited. Lipids can be simply deposited from a solution avoiding the preparation of liposomes (Fig. 3h). This surface is used for the immobilization of lipids from solutions, obviating the requirement for liposomes (38). The surface is simply exposed to the required mixture of lipids in the solvent, the solvent is diluted out by the addition of buffer, and the lipids are forced to assemble on the surface. Before starting this protocol, it is essential to clean the IAsys instrument with ultra pure water, and aqueous solution containing 1 mM NaOH, 1% (w/w) SDS, 1% Tween 20, followed by extensive washing with water. To obtain reproducible results it is essential that the lipids used are not oxidized. It is recommended that stock lipid solutions contain butylated hydroxytoluene 0.1% (w/v) and are stored at less than −60 °C for no longer than 2 weeks. 1. Insert a new hydrophobic cuvette into the instrument and wash extensively with 250 µL of 2-propoanol. 2. Start data acquisition and collect 3 min of baseline data. 3. Wash extensively with 200 µL of PBS/AE and collect data for about 10 min (see Note 16). 4. Wash with 300 µL of 2-propanol and collect data for about 3 min. 5. Add 30 µL of DOPC and leave for 2 min (see Note 17). 6. Wash rapidly with 320 µL of PBS/AE and collect data for about 6 min (see Note 18). 7. Wash with 250 µL of 0.1 M HCl for 5 min. 8. Wash with 350 µL of PBS/AE and collect data for about 5 min. 9. Wash with 250 µL of NaOH and leave for 1 min. 10. Wash with 370 µL of PBS/AE and collect data for about 5 min. Measure the amount of lipid on the surface (see Note 19). 11. Add BSA solution to a final concentration of 1.5 mg/mL.

116

Zourob et al.

12. Wash with 350 µL with PBS/AE and leave for 3 min. 13. Steps 10–11 test for coverage of the sensor surface with lipid. In the presence of a complete lipid monolayer no BSA binding should be apparent after washing with PBS/AE. It is recommended that if saturated lipids (e.g. DPPC and DSPC) are used a rinse (wash three times with 200 µL) with 5% sucrose (w/w) in water is performed before washing with PBS/AE. Saturated lipid micelles will stick to the surface preventing monolayer formation. The increased density of the sucrose enables the micelles to float to the surface and be washed away. 3.8. Immobilization of Lipids on NonDerivatized Surface

Protocol

The non-derivatized cuvette offers an alternative surface to the hydrophobic cuvette for the immobilization of phospholipids (PLS). Single or mixed lipids can be immobilized from the solution in the detergent n-octyl b-D-glucopyranoside (OG). The lipid-coated surface can then be used for analysis of protein/lipid interactions, or of lipid-associated proteins with other soluble components. 1. Insert a new non-derivatized cuvette into the instrument and wash with 150 µL of PBS and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Wash with 180 µL of PBS/OG, and collect data for about 3 min. 4. Sonicate the PLs for 10 s in a sonicating cleaning bath and introduce 100 µL to the instrument of the PLS solution till the response is about at a plateau. 5. Add 30 µL of PBS without removing the PLS solution. Leave until the binding response has approached a plateau. 6. Wash three times with 50 µL of PBS. 7. The amount of lipid immobilized can be determined by subtracting the initial baseline level in PBS from the final level after stage 6 (see Note 20).

3.9. Capture of His-Tagged Proteins

Protocol

Poly-histidine is a commonly used tag on recombinant proteins. This protocol describes a method for the chelation of His-tagged proteins on a nickel charged surface onto a carboxymethyl dextran coated surface. The immobilization process includes coupling of nitrotriacetic acid (NTA) to the dextran using EDC/NHS and charging the NTA with nickel ions. This complex will then be used to capture His-tagged protein from solutions. 1. Insert new carboxylate cuvette into the instrument and wash with 200 µL of HEPES and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data.

Label-Free Detection with the Resonant Mirror Biosensor

117

3. Wash and incubate the surface with 180 µL of a mixture of 1:1 (v/v) EDC/NHS for 15 min. 4. Wash with 200 µL of HEPES and collect baseline data for 3 min (see Note 2). 5. Incubate 50 µL of NTA solution for 8 min (see Note 21). 6. Wash with 200 µL of HEPES and collect baseline data for 3 min (see Note 22). 7. Incubate the cuvette with 1 M ethanolamine pH 8.5 for 8 min to block unreacted NHS-esters (see Notes 9 and 23). 8. Wash with 200 µL of HEPES and collect baseline data for 5 min. 9. Wash three times and incubate the surface with 200 µL of 500 µM of NiCl2 in HBSTE for 6 min. 10. Wash with 200 µL of HEPES and collect baseline data for 5 min. Optional

11. Wash three times with 200 µL of PGI wash buffer and leave for 5 min to remove loosely associated metal ions from the residual carboxyl’s on the CMD. 12. Wash with 200 µL of HEPES and collect baseline data for 3 min. 13. Spike in the His-tagged sample and allow binding 14. Wash with 200 µL of HEPES and collect baseline data for 3 min. If high concentration of His-tagged material has been used, it is possible that a very distinct dissociation phase will follow. This can take several minutes to settle down. It is possible to speed up the process by washing loosely associated material of the surface using PFI wash buffer (optional). 15. Wash and incubate the surface with 200 µL of PGI for 2 min to clear the his-tag from the surface (see Note 24). 16. Wash with 200 µL of HEPES and collect baseline data for 3 min.

3.10. Capture of FLAGTagged Proteins

Protocol

FLAG is a commonly used tag on recombinant proteins. The tag consists of eight amino acids in the sequence Asp-Tyr-Lys-Asp-AspAsp-Asp-Lys. This protocol describes a method for immobilization of anti-FLAG antibody to CMD for the capture of FLAG-tagged proteins. Captured recombinants can be easily eluted with competing FLAG peptides at nM concentrations (43). 1. Insert new carboxylate cuvette into the instrument and wash with 200 µL of PBS/T and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Wash and incubate the surface with 180 µL of a mixture of 1:1 (v/v) EDC/ NHS for 10 min.

118

Zourob et al.

4. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 5. Replace PBS/T with acetate buffer pH 4.5 and leave to equilibrate for 4 min. 6. Introduce Anti-FLAG antibody between 10 and 20 µg/ mL and leave for 5 min. An irrelevant antibody can be used immobilized at the same level as a control. 7. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 8. Incubate the cuvette with 1 M ethanolamine pH 8.5 for 4 min to block unreacted NHS-esters. 9. Wash with 200 µL of PBS/T and collect baseline data for 5 min. 10. Wash extensively with 200 µL of 50 mM HCl and leave to equilibrate for 5 min to remove loosely bonded antibodies from the CMD matrix. 11. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 12. Repeat steps 9 and 10 twice 13. Introduce 740 nM FLAG-BAP control peptide and incubate it for 5 min. 14. Wash with 200 µL of PBS/T and collect baseline data for 5 min. 15. Wash with 200 µL of 50 mM HCl and leave for 2 min to regenerate the surface. 16. Wash with 200 µL of PBS/T and collect baseline data for 5 min. 17. Repeat steps 13–16 to assess reproducibility of activity (see Note 25). 3.11. Sensitivity Enhancement Using Colloidal Gold Complexes

Protein-colloidal conjugate cause a higher refractive index and thickness change per binding event than native protein or other biomolecule. As a result, the colloidal gold nano-particles can be used to enhance the sensitivity of the assays. This protocol describes a general method for the preparation of colloidal gold-protein complexes and its use to improve the detection limit of an assay. Gold complexes can be prepared using the De Mey method. (33). The bioactivity of the colloidal gold complexes can be checked by immunoblotting with the appropriate antibody (Biocell gold conjugates technical information and guidelines for use in electron microscopy, Light Microscopy and Immunoblotting.). The average number of protein molecules per colloidal gold particle can be determined from radioactive studies (40).

Label-Free Detection with the Resonant Mirror Biosensor

119

Colloidal gold is susceptible to aggregation; therefore all glassware and plasticware should be thoroughly cleaned and if possible, silliconized. Protocol

1. Adjust the pH of the colloidal gold to 0.5 pH units above the PI of the protein with potassium carbonate. 2. Determine the number of gold particles by measuring the absorbance at λmax or 520 nm (A520 = 1 is 1.7 × 1011 particles/mL for 30 nm particles). 3. Add protein stepwise to colloidal gold to give a final molar ratio of 100:1 or 10:1, depending on how many protein molecules are required by gold particles. Typical number of particles will be 1 × 1012 particles/mL final concentration. 4. Add BSA to the solution to final concentration between 0.1 and 0.5% of BSA (w/v) to block any remaining exposed sites on the gold. 5. Leave on a rotary mixer for 15 min. 6. Spin the mixture in a microcentrifuge for 20 min. 7. Carefully remove the clear supernatant leaving loose red sediment. 8. Resuspend the gold complex in fresh BSA solution to remove the more free protein. 9. Repeat steps 6–8. A good final volume is 1/10th of the starting volume. 10. It is recommended at this stage to retest the concentration of the colloid as described in step 2. 11. Store the product in an air-tight siliconized glass vial at 4 °C until it used for assay enhancement (see Subheading 3.17).

3.12 Measuring Affinity and Rate Constants

Kinetic information indicates how fast the interactants come together, and how fast the resulting complex breaks down. Kinetic information is given by the rate constants of the forward and reverse reaction as shown in Eq. 1. The interaction of the ligate (L) with the immobilized ligand (G) on the sensor surface can be represented as: K on G + L ¾¾¾ ® GL.

G+L

kass kdiss

GL.

The rate of complex (GL) formation is = Kass (A)(B) The rate of the reverse reaction (dissociation) = Kdiss (AB) KD =

| G || L | kdiss 1 = = , GL kass KA

120

Zourob et al.

where KD and KA are the affinity constants (dissociation equilibrium constant in units (M) and association equilibrium constant in units (M−1) respectively). The parameters Kass (M−1 S−1) and Kdiss (S−1) refer to the association and dissociation rate constants respectively. By measuring (G), (L), and (GL) at equilibrium, the affinity can be determined. The biosensor gives rise to a signal that directly correlates to the amount of complex (GL) which is being formed. The dissociation equilibrium constant KD refers to the concentration at which 50% of the binding sites are occupied. KD and KA are opposite each other. If KD is low, the affinity of the interaction is high and in case the KA is high, the affinity is high. Upon exposure of the ligate to the ligand, initially the forward reaction predominates and GL is formed. With time, as the concentration of GL increases, the reverse reaction becomes significant. Eventually, at equilibrium, the time taken to reach this depends on the kinetics of the interaction and the concentration of the reactants, and the rate of the forward and reverse reaction, will be the same. GL continues to be formed from G and L, and GL continues to dissociate, and concentrations of GL, G and L remain constant. R eq =

Rmax [L] . K D + [L]

The classical route to determine equilibrium constants, is measuring the amount of complex GL at equilibrium from a fixed value of ligand. Rmax is the response when all possible ligand sites are occupied. At each concentration of ligate, a value of the response at equilibrium (Req) is obtained, equivalent to the amount of complex GL formed at equilibrium. At high concentrations of ligate L compared with the ligand G, the assumption that the concentration of L in solution is constant can be made. This is termed pseudo-first order condition. At high concentrations, Req is the same as Rmax. By plotting Req against the free ligate (L) concentration, the concentration required to saturate 50% of the available ligand sites can be determined; this value is the concentration which give rise to the response of Rmax/2. Rate of complex (GL) formation: d[GL] = kass [G][L] - kdiss [GL]. dt The concentration of free ligand sites decreases with ligate binding, and thus the concentration of free ligate sites at time t, is equal to the maximum number of sites (G)0 minus the sites bound (GL)t or ( (G)0−(GL)t); d[GL] = kass [G]0 - [GL]t [L] - kdiss [GL]t . dt

Label-Free Detection with the Resonant Mirror Biosensor

121

As complex (GL) formation results in a change in response (Rt), which is directly proportional to (GL), Rmax is equivalent to (L)0 and the above equation can be rewritten as: dRt = kass (Rmax - Rt )[L ] - kdiss Rt . dt Multiplying out the brackets and rearranging gives: dRt = kass (Rmax [L] - Rt (K ass [L] + K diss ). dt Originally this led to biosensor data being analyzed by linearization of the data for each ligate concentration (L) by plotting dR/dt against Rt. This should give a straight line of slope kass [L] + kdiss. Plotting slopes against concentrations of L will then give a straight line of slope equal to the association rate constant (Kass), and the intercept is equal to the dissociation rate constant (Kdiss). However, a problem with linearization is that the errors are compounded. By integration the equation can be transformed into a form suitable for application of non-linear regression:

(

)

Rt = Req 1 - e - (kass [L]kdiss )t . Using the above equation, binding data can be fitted at different concentrations to determine Kon (where Kon = Kass [L] + Kdiss. A plot of Kon against ligate concentration (L), will give a straight line with a slope of Kass and an intercept Kdiss. The determination of Kdiss from the intercept of Kon vs. (L) often has a high extrapolation error. A more robust approach is to initiate dissociation by removing the ligate with a buffer wash. Under these conditions (i.e. L = 0) dissociation of the ligate from the immobilized ligand is described by Eq. 3: Here the concentration of L is taken to be zero and the first order decay in response is given by: dR = - K diss Rt . dt On integration it gives: Rt = R0 e - (kdiss t ,

(3)

where R0 is the response at the initiation of dissociation. This simple equation is not applicable to all dissociation profiles due to incomplete dissociation which has be attributed to rebinding of the dissociated ligate. 3.13. Determination of the Concentration of Theophylline in Buffer and Serum

Theophylline has a molecular weight of 180 Da. The response of the RM for the direct detection of any analyte less than 5,000 Da will be too small to be measured reliably. Alternatively, indirect assay format must be employed. The assay is based on competition

122

Zourob et al.

between theophylline and a protein-theophylline conjugate for binding to a theophylline monoclonal antibody (Mab) captured on the sensor surface by an immobilized specific rabbit-mouse antibody (RAM-Fc) through the Fc region. The theophylline conjugate used in this assay is IgG (150,000 Da); it contains approximately 20 theophylline residues per protein molecule. 3.13.1 Assay Procedure in PBS/T

1. Insert new CMD cuvette into the instrument and wash with 200 µL of PBS/T and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Wash and incubate the surface with 180 µL of a mixture of 1:1 (v/v) EDC/ NHS for 10 min. 4. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 5. Introduce 10 µg/mL of RAM Fc for immobilization on CMD using the EDC/NHS chemistry. 6. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 7. Incubate the cuvette with 1 M ethanolamine pH 8.5 for 4 min to block unreacted NHS-esters. 8. Wash with 200 µL of PBS/T and collect baseline data for 5 min. 9. Introduce 100 µg/mL of MAb for 10 min. 10. Wash with PBS/T and collect baseline data for 5 min. 11. Incubate a solution containing one of the theophylline/conjugate mixtures into the cuvette for 10 min. 12. Wash with PBS/T and collect baseline data for 5 min. 13. Regenerate the surface to remove the anti-theophylline and theophylline-protein-conjugate with 50 mM HCl for 3 min, followed by re-equilibration into PBS/T buffer. 14. Steps 5–13 were then repeated for further theophylline/ conjugate mixtures. 15. Construct the calibration curve.

3.13.2 Assay Procedure in Serum

Repeat the same experimental procedure as described for PBS/T. With the exception of theophylline, solutions were prepared in serum and were diluted 1:10 in PBS/T prior to mixing with the theophylline-protein-conjugate. In addition, a control for nonspecific binding was performed with a digoxin-protein-conjugate (see Note 26).

3.14 Protein– Carbohydrate Recognition

The importance of protein–carbohydrate recognition in modulating key cellular activities, such as receptor binding, adhesion, and mitogenesis, has become increasingly recognized in recent

Label-Free Detection with the Resonant Mirror Biosensor

123

years. Basic fibroblast growth factor (bFGF) is the archetype of the FGF family of growth factors and is involved in the regulation of growth and morphogenetic events as diverse as the induction of mesoderm and limb buds to the growth of skeleton and the mammary gland (41). To elicit their cellular effects, bFGF, like the other FGFs, must interact with a dual receptor system consisting of the receptors tyrosine kinases and heparan sulfate proteoglycans; bFGF binds to the heparin sulfate chains of the proteogylcans (42). Heparan sulphate is a sulphated glycosaminoglycan present on the surface of most mammalian cells. Heparin, a specialized form of heparan sulphate produced by mast cell, contains binding sequences for most heparan sulphate binding proteins, including bFGF (41). A number of peptides derived from the protein sequence of bFGF have been identified that are implicated in heparin binding (41). The RM used to study the interaction of heparin to bFGF was capable of quantifying the very low association rates and affinities of bFGF-derived peptides for heparin. Novel insights on structure and function were therefore obtained. A full report of this work can be found in (43). Protocol

1. Insert new CMD cuvette into the instrument and wash with 200 µL of PBS/T and allow 10 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Deposit streptavidin at a concentration of 200 µg/mL in 10 mM acetate buffer pH 5.0 for 15 min as described in Subheading 3.6.1. 4. Wash with 200 µL of 20 mM HCl for 3 min to remove noncovalently attached streptavidin. 5. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 6. Introduce 100 µL of 1 mg/mL of biotinylated heparin and incubate for 30 min. 7. Regenerate the surface with Na2HPO4 buffer for 5 min. 8. Wash with PBS/T and collect baseline data for 5 min. 9. For association 100 µL of minimum five different concentrations per ligate (bFGF and peptides) in the PBS/T were analyzed for 5–10 min on at least two different surfaces CMD surfaces and one AS surface. 10. For dissociation run 200 µL of PBS/T for at least 5 min. 11. Regenerate the surface with Na2HPO4 buffer for 3 min, followed by re-equilibration in PBS/T. 12. Apply bFGF and bFGF-derived peptides as control analysis onto streptavidin surfaces in the absence of captured heparin (should not show significant interaction).

124

Zourob et al.

13. Calculate the kinetic association and dissociation constants (kass and kdiss) and the dissociation equilibrium constant (KD) as described in Subheading 3.12. Using a series of profiles for bFGF and bFF-peptide binding to heparin to generate the kinetics and equilibrium constants shown in Table 1 (see Note 27). 3.15 Nucleic Acid Hybridization

This protocol describes the use of RM for the rapid and unlabeled single stranded DNA–DNA hybridization using a biotinylated oligonucleotide, attached to CMD coated sensor surface via streptavidin (44). This approach is a convenient, generic approach and biotinylated oligonucleotides can be easily synthesized. The binding of biotin to streptavidin is virtually irreversible, requiring extreme conditions to separate the two molecules (45, 46). 1. Insert new CMD cuvette into the instrument and wash with 200 µL of PBS/T and allow 10 min for equilibration.

Protocol

2. Start data acquisition and collect 3 min of baseline data. 3. Streptavidin was immobilized at a concentration of 200 µg/ mL streptavidin in 10 mM acetate buffer pH 5.0 for 15 min as described in Subheading 3.6.1. 4. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 5. Wash with 200 µL of 20 mM HCl for 3 min to remove noncovalently attached molecules. 6. Wash with 200 µL of PBS/T and collect baseline data for 3 min.

Table 1 kinetic and equilibrium constants for the interaction of bFGF with heparin Ligate Constant (units) a

a

BFGF (±SE)

BFGF (127–140) (±SE)

KD (M)

−9

84 (±55) × 10

30 (±4) × 10−6

KD (M)b

74 (±20) × 10−9

35 (±7) × 10−6

Kass (M−1S−1)c

9.3 (±2.5) × 10−4

4 (±0.16) × 10−2

Kdiss (S−1)d

6.8 (±0.4) × 10−3

1.4 (±0.38) × 10−2

KD calculated from five concentrations of ligate in three experiments (two carried out on CM-Dextran and one on AS) by binding curve analysis. KD was calculated from KD = (Kdiss/Kass) c Mean of three determinations (±SE) d Mean of nine determinations (±SE). (Reproduced from ref. 48 with permission from NeoSensors Ltd.)

Label-Free Detection with the Resonant Mirror Biosensor

125

7. Introduce 2 mg/mL Oilgo-1 in PBS/T and incubated for 15 min. 8. Monitor the binding response after washing with 200 µL of PBS/T and collect baseline data for 3 min. 9. Wash the surface with hybridization buffer (pH 8, 5× saline, sodium citrate, 5× Denhardts solution), followed by 5 mM sodium phosphate, 0.1% v/v Tween 20. 10. Introduce 4 µg/mL complementary oligonucleotides (Oilgo2) in hybridization buffer and incubate for 15 min. 11. Wash the surface with hybridization buffer and monitor the shift in resonance position (Fig. 4) (see Note 28). 12. Run control experiments on new streptavidin derivatized surfaces using (see Note 28): (a) Oligo-3 (5 µg/mL, 303 nM) with immobilized Oligo-1 at the sensor surface (Fig. 4). (b) Oligo-4 (2 µg/mL) with immobilized Oligo-2 (4 µg/ mL, 303 nM) at the sensor surface. Introduce Oligo-2 (4 µg/mL, 303 nM) in the absence of an immobilized specific oligonucelotide. 3.16. Fermentation Monitoring

The fermentation of micro-organisms and mammalian cells for the production of recombinant proteins is an area of increasing industrial interest. It could benefit greatly from the rapid assay to quantify the bioproducts of interest during the fermentation process. Here is one example reported for the use of the RM for

Fig. 4. Specific hybridization of Oligo-2 to captured complementary biotinylated Oligo-1 and the absence of hybridization of non-complementary Oligo-3 with Oligo-1. Arrows indicate: (a) addition of hybridization buffer, (b) addition of oligonucleotides. (Reproduced from ref. 46 with permission from NeoSensors Ltd.).

126

Zourob et al.

monitoring the recombinant antibody fragment, D1.3 Fv (specific for hen egg lysozyme, HEL) (47) produced by periplasmic secretion in E. coli fermentation broths (48). Protocol

1. Insert new CMD cuvette into the instrument and wash with 200 µL of PBS/T and allow 10 min for equilibration. 2. Start data acquisition and collect baseline data for 3 min. 3. Immobilize HEL (30 µg/mL) in acetate buffer pH 5 on CMD matrix for 5 min. 4. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 5. Incubate the surface with 50 µL of 1 M ethylene diamine pH 8.5 for 10 min to modify the carboxymethyl groups with amines. 6. Wash with 200 µL of PBST and collect baseline data for 3 min. 7. Construct calibration curve using (negative) E. coli fermentation broth (one does not express D1.3 Fv that mimicked the process broth in terms of operation and final protein concentration (49) and by spiking different concentration of purified D1.3 into this broth. 8. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 9. Incubate the sample from the fermenter at regular intervals for 5 min after removing the particulates using microcentrifuge 11,600 × g for 5 min. 10. Wash with 200 µL of PBS/T and collect baseline data for 3 min. 11. Regenerate the surface from the immobilized ligand with 100 mM HCl for 4 min. 12. Data can be analyzed using two methods either measuring the absolute change in response after incubating the sample with the ligand for 5 min or measuring the initial binding rate (using linear regression using FASTfit™ software) (50).

3.17. Bacterial Cell Detection

Protocol

The RM has been used to distinguish between bacterial strains on the basis of their surface proteins. Staphylococcus aureus (Cowan-1) cells, which express protein A on their surface, could be detected by binding to Human IgG (51, 52). Staphylococcus aureus strain (Wood-46) cells do not interact significantly with Human IgG as they do not express protein A on their surface (53) (see Note 29). 1. Insert AS cuvette into the instrument and wash with 200 µL of PBS and allow 10 min for equilibration.

Label-Free Detection with the Resonant Mirror Biosensor

127

2. Start data acquisition and collect 3 min of baseline data. 3. Activate the surface with 5% (v/v) glutaraldehyde in water for 30 min. 4. Wash with 200 µL of PBS and collect baseline data for 3 min. 5. Incubate the surface with 100 µg/mL IgG in PBS for 8 min. 6. Wash with 200 µL of PBS and collect baseline data for 3 min. 7. Block the surface with 3 mg/mL BSA for 5 min. 8. Wash with 200 µL of PBS and collect baseline data for 3 min. 9. Introduce S. aureus (Wood-46) and incubate for 20 min. 10. Wash with 200 µL of PBS and collect baseline data for 3 min (see Fig. 5a for the results). 11. Repeat the same procedure from steps 1 to 10 for Wood-46 strain (see Fig. 5a for the results). 12. Repeat the same procedure from steps 1 to 10 for S. aureus (Cowan-1) on BSA immobilized surface as control surface (see Fig. 5a for the results).

Fig. 5. (a) Calibration curve showing the response against cell concentration. The control data for bacterial cell strain Wood-46 and for the interaction between the Cowan-1 and immobilized BSA are also shown. (Reproduced from ref. 19 with permission from Analytical Chemistry). (b) Comparison of the direct binding assay with the colloidal gold sandwich assay. (Reproduced from ref. 19 with permission from Analytical Chemistry).

128

Zourob et al.

Fig. 5. (continued)

13. For sandwich assay using gold complex: repeat the same procedure as in direct assay from steps 1 to 10 followed by the addition of colloidal gold human IgG and incubate for 10 min. 14. Wash with 200 µL of PBS and collect baseline data for 3 min (see Fig. 5b for the results). 3.18. Receptor–Cell Interactions

Protocol

Carcinoembryonic antigen (CEA) (54) has been identified as immunocytochemical on tumor cell membrane from a variety of tissues, making it one of the most useful human tumor markers (55). Cell lines have been developed expressing CEA to study the etiology of carcinomas. In the present protocol the RM is used for the real-time detection of the binding of L cells bearing a cellexpressed CEA antigen, to anti-CEA antibody immobilized on AS sensor surface (see Note 29). 1. Insert new AS cuvette into the instrument and wash with PBS for 5 min for equilibration. 2. Start data acquisition and collect 3 min of baseline data. 3. Introduce PBS/BS3 to the cuvette for 10 min 4. Wash with PBS to get stable baseline. 5. Introduce the antibody solution 130 µg/mL and leave it for about 10 min (see Note 8). 6. Block remaining activated groups with 2 mg/mL BSA in PBS for 10 min.

Label-Free Detection with the Resonant Mirror Biosensor

129

7. Wash the cuvette with 20 mM HCl for 2 min 8. Wash with PBS/BSA for about 5 min to get the baseline. 9. Introduce cell solution. 10. Wash with PBS/BSA and collect 5 min of baseline data. 11. Regenerate the surface with 0 mM HCl for 2 min, PBS/T for 2 min, and finally for another 2 min with 20 mM HCl (see Fig. 6) (see Note 30). 12. Repeat steps 2–11 with non-expressing cells (Reference cells) (see Fig. 6). 3.19. Determination of the Kinetic Constants

Protocol

The use of RM for measuring equilibrium binding constants is attractive due to the label free detection unlike commonly used techniques such as ELISA, RIA, hapten inhibition, and fluorescence methods (56). This protocol describes the use of RM to determine the rate and affinity constants of the binding interaction between immobilized Staphylococcus aureus protein A and different concentration of human IgG in solution (57, 58). 1. Insert CMD cuvette into the instrument and wash with 200 µL of PBS and allow 5 min for equilibration.

Fig. 6. Binding of L cells to an AS surface derivatized with an anti-CEA antibody. (Reproduced from ref. 56 with permission from NeoSensors Ltd.).

130

Zourob et al.

2. Start data acquisition and collect 3 min of baseline data. 3. Introduce a mixture of 200 µL of EDC and 200 µL of NHS for 7 min. 4. Wash with PBST and establish a stable baseline. 5. Change the PBST to 10 mM acetate buffer pH 5.5 and collect 3 min baseline data. 6. Introduce 200 µL of 12.5 µg/mL Staphylococcus aureus protein in acetate buffer to the cuvette for 15 min. 7. Block the unreacted NHS-ester with 1 M ethanolamine pH 8.5 for 3 min. 8. Wash with PBST to get the baseline. 9. Wash the cuvette with 10 mM HCl for 3 min to remove the non-covalently bound ligand to the matrix. 10. Wash with PBST for 5 min to get the baseline and to determine the amount of protein A. 11. Incubate the first concentration of HIgG in PBS/T for 30 min. 12. Wash with PBS/T to get the baseline. 13. Regenerate the surface with 10 mM HCl for 3 min. 14. Re-establish baseline with PBS/T 15. Repeat steps 11–14 with different concentrations of HIgG (Fig. 7a) (see Note 31). 16. Plot dR/dt vs. R if the maximum binding response is known and the association between the two interactants is carried out under pseudo first order conditions for which the overall rate is: dR = kass (Rmax - R)[C] - kdiss R, dt dR = kass (Rmax [C] - R(K ass [C] + K diss ). dt However, if the maximum response is not known the ka can be obtained by plotting the slope, ks, of the plot dR/dt vs. R (Fig. 7b) for all concentrations of HIgG. The slope of this new plot is the ka and the intercept, in theory, is the dissociation rate constant (kd) (Fig. 7c). Ks = kaC + kd

.

Ka was calculated for HIgG and protein A 1.13 × 105M−1s−1 and the intercept value, kd, was 1.83 × 10−4s−1. Another way to determine the dissociate rate constant, which is independent of concentration, is to replace the sample HIgG with buffer and follow the dissociation phase. The dissociation

Label-Free Detection with the Resonant Mirror Biosensor

131

Fig. 7. (a) The biding curve of HIgG to immobilized SpA. (Reproduced from ref. 59 with permission from NeoSensors Ltd.). (b) Plots of dR/dt vs R for all HIgG concentrations. (Reproduced from ref. 59 with permission from NeoSensors Ltd.). (c) Slope Ks of dR/dt vs. R plotted against HIgG concentration. Determination of ka for SpA:HIgG. (Reproduced from ref. 52 with permission from NeoSensors Ltd.). (d) Plot of ln(R0/Rn) vs. tn − t0 to determine the dissociation rate constant for protein A:IgG. (Reproduced from ref. 59 with permission from NeoSensors Ltd.).

132

Zourob et al.

Fig. 7. (continued)

phase should follow the following equation, when rebinding is negligible due to complete removal of the dissociating analyte: dR = - K diss R. dt Integrating the above equation gives: dR = kd (tn - t0 ), dt

Label-Free Detection with the Resonant Mirror Biosensor

133

where R0 is the response at t0, and Rn is the response at tn. By constructing a plot of ln(R0/Rn) vs. tn − t0, kd is determined from the slope of the plot (Fig. 7d). The kd value was found to be 1.09 × 10−4. The overall affinity constant (kd/ka) of the HIgG and protein A was therefore calculated to be 1.62 × 10−9 M.

4. Notes 1. Silanization can be carried out by exposing the cleaned RM chip to 3-mainopropyletriethoxysilane vapour at 145 °C and about 3 mm Hg for 1 h (2). 2. NHS coupling kit is commercially available from Neosensors (Durham, UK, CUV999900) if you do not want to prepare it. EDC/NHS aliquots should be stored at or below −18 °C. 3. The use of coherent illumination proved troublesome, as dust particles within the instruments caused light scattering, producing a characteristic speckle pattern on the detector. 4. The finite bandwidth of these sources has the effect of broadening the apparent width of the resonance peaks in the angular scan mode, while the finite size of the sources reduces the degree of collimation, which also broadens the peak in the wavelength scan mode. 5. For the angular scan, an interference filter with 670 ± 10 nm wavelength, 25 mm in diameter (Comar instruments, Cambridge, UK), and in case of wavelength scan a band-pass interference filters of wavelengths 512, 531, 550, 570, 596, 605, 632, 645 and 670 nm (10 nm FWHM bandwidth, Ealing Electro-optics, Watford, UK) was used to provide light of different wavelengths. 6. Fresh dissolved EDC should be used. In case degraded EDC has been used the cuvette can be treated with 1 M Tris, 3 M NaCl, pH 8.0 to release any accumulated materials. 7. In case of highly charged ligand or ligates, it is desirable to reduce the charges on the matrix by activating the matrix with EDC and NHS, followed directly by blocking with ethanolamine prior to carrying out the protocol as written. 8. The pH of the acetate buffer should be below the PI of the protein ligand, but above a value of 3.5. 9. 1 M Tris–HCl, pH 8.0 may also be used to block the unreacted NHS ester.

134

Zourob et al.

10. Avoid exposing the AS surface to buffers containing detergent prior to immobilization as it may reduce or variable the levels of ligand immobilization. 11. High concentration of ligands are needed as there is very little electrostatic concentration of ligand to the As surface unlike CMD surface. 12. Smaller and less charged proteins such as β-casein or cytochrome c can be used. Be sure that the sensor surface is completely coated as failure to coat the surface with protein results in an increase in the non-specific binding when ligates are added. 13. It is advisable after the PBST wash to further treat the immobilized avidin layer with the same regenerants that may be used in subsequent ligate binding experiments. 14. Streptavidin and NeutrAvidin from Pierce can be used instead of avidin. 15. Biotin cuvette can be reused several times after removing the avidin layer by treating the surface for 2 min with saturated KOH outside the instrument. It is essential to wash the surface thoroughly and rapidly to remove all traces of KOH prior to reuse. 16. Prevent de-wetting of the sensor surface by extensive washing. 17. To obtain reproducible results it is essential that the lipids used are not oxidized. It is recommended that the stock lipid solution contain 0.1% (w/w) butylated hydroxytoluene and are stored at less than −60 °C for no longer than 2 weeks. 18. It is recommended in case of using saturated lipids to wash the surface with 50% sucrose (w/w) in water before washing back into the PBS/AE. The sucrose solution will float and washed away the formed sticky micelles instead of the formation of a monolayer. 19. The resonance scan should be a symmetrical peak, indicating uniformity immobilization throughout the sensed area. If the peak is not symmetrical, stages 6–9 can be repeated. 20. The PLs can be stripped off from the surface by incubating the cuvette with 2% OG in high quality water for 3 min followed by washing with PBS. 21. Incubation time can be extended up to 15 min to ensure that a reasonable quantity of NTA is immobilized. 22. Due to the expensive nature of the TNA, it is recommended to recover NTA solution as it can be re-used.

Label-Free Detection with the Resonant Mirror Biosensor

135

23. It is recommended to extend the blocking time due to the extension of the activation time. 24. Alternatively, the sample can be removed by washing with 10 mM HEPES, 0.15 M NaCl, 0.35 M EDTA, 0.005% Tween 20, pH 8.3. 25. This procedure can be used at any time during the experiment to test the activity of the surface. The protocol was also found to work with FLAG-Ubiquitin (Sigma, U5382). 26. No significant non-specific binding of a digoxin-protein conjugate was observed. 27. The dissociation equilibrium constant for the interaction of peptide bFGF (127–140) with heparin is a 1,000-fold higher than that observed for native bFGF, its association rate constant and dissociation equilibrium constants for the peptide were readily determined by the RM. The results showed how RM can be used to analyze protein and a very low affinity peptide–carbohydrate interactions and provides evidence that a part of the recognition site of bFGF for heparin lies within the amino-acid sequence of 127–140 of bFGF (43). 28. The results from Fig. 3 show the specific hybridization of complementary oligonucleotides (Oligo-2 to immobilized Oligo-1) under average stringency conditions, while the non-complementary DNA (Oligo-3) did not give any significant response. 29. The RM is unsuitable for cell detection due to the short penetration of the evanescent field at the sensor surface, which places the majority of the cells outside the field (59). However, the sensitivity of the RM technique can be increased 1,000-fold using a human immunoglobulin G (HIG)-colloidal gold complex (19). 30. Keep cells in ice between the bindings. 31. All experimental data shown were generated on a single cuvette.

Acknowledgements The authors thank W. Jones and S. Mian from Neosensors for their help and providing the material for this chapter. The views expressed here are those of the authors and do not represent those of Biophage Pharma Inc.

136

Zourob et al.

References 1. Cush, R., Cronin, J.M., Stewart, W.J., Maule, C.H., Molloy, J., and Goddard, N.J. (1993) The resonant mirror: a novel optical sensor for direct sensing of biomolecular interactions, Part I: Principle of operation and associated instrumentation. Biosens. Bioelectron. 8, 347–353 2. Buckle, P.E., Davies, R.J., Kinning, T., Yeung, D., Edwards, P.R., Pollard-Knight, D., and Lowe, C.R. (1993) The resonant mirror: a novel optical sensor for direct sensing of biomolecular interactions, Part II: Applications. Biosens. Bioelectron. 8, 355–363 3. Charalambous, B.M., Evans, J., Feavers, I.M., and Maiden, M.C. (1999) Comparative analysis of two meningococcal immunotyping monoclonal antibodies by resonant mirror biosensor and antibody gene sequencing. Clin. Diagn. Lab. Immunol. 6, 838–843 4. Kobayashi, K., Matsuura, E., Liu, Q., Furukawa, J., Kaihara, K., Inagaki, J., Atsumi, T., Sakairi, N., Yasuda, T., Voelker, D.R., and Koike, T., (2001) A specific ligand for β2-glycoprotein I mediates autoantibodydependent uptake of oxidized low density lipoprotein by macrophages. J. lipid Res. 42, 697–709 5. Zimmer, S., Stocker, A., Sarbolouki, M.N., Spycher, S.E., Sassoon, J., and Azzi, A. (2000) A novel human tocopherol-associated protein. Cloning, in vitro expression, and characterization. J. Biol. Chem. 275, 25672– 25680 6. Dmitriev, D.A., Massino, Y.S., and Segal, O.L. (2003) Kinetic analysis of interactions between bisepecific monoclonal antibodies and immobilized antigens using a resonant mirror biosensor. J. Immun. Meth. 280, 183–202 7. Rhodes, D.C. (2000) Binding of TammHorsfall protein to complement 1q measured by ELISA and resonant mirror biosensor techniques under various ionic-strength conditions. Immunol. Cell Biol. 78, 474–482 8. Nakra, P., Manivel, V., Vishwakarma, R.A., and Rao, K.V. (2000) Cell responses to a peptide epitope. X. epitope selection in a primary response is thermodynamically regulated. J. Immnol. 164, 5615–5625 9. Salek-Ardakani, S., Arrand, J.R., Shaw, D., and Mackett, M. (2000) Heparin and heparan sulfate bind interleukin-10 and modulate its activity. Blood 96, 1879–1888 10. Torigoe, H., Hari, Y., Sekiguchi, M., Obika, S., and Imanishi, T. (2001) 2′-O,4′-CMethylene bridged nucleic acid modification

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

promotes pyrimidine motif triplex DNA formation at physiological pH. Thermodynamic and kinetic studies. J. Biol. Chem. 276, 2354–2360 De Zutter, J., Forget, A., Logan, K., and Knight, K. (2001) Phe217 regulates the transfer of allosteric information across the subunit interface of the RecA protein filament. Structure 9, 47–56 Sassoon, J., Lilie, H., Baumann, U., and Kohli, J. (2001) Biochemical characterization of the structure-specific DNA-binding protein cmb1 from Schizosaccharomyces pombe. J. Mol. Biol. 309, 1101–1115 Despeyroux, D., Walker, N., Pearce, M., Fisher, M., Mcdonnell, M., Bailey, S.C., Griffiths, G.D., and Watts, P. (2000) Characterization of heterogeneity by electrospray mass spectrometry, capillary electrophoresis, and resonant mirror. Anal. Biochem. 279, 23–36 Hirmo, S., Artursson, E., Puu, G., Wadstorm, T., and Nilsson, B. (1998) Characterization of Helicobacter Pylori interactions with sialglycoconjugates using a resonant mirror biosensor. Anal. Biochem. 257, 63–66 Hirmo, S., Artursson, E., Puu, G., Wadstorm, T., and Nilsson, B. (1999) Helicobacter Pylori interactions with human gastric mucin studied with a resonant mirror biosensor. J. Microbiol. Meth. 37, 177–182 Sigurdsson, H.H., Loftsson, T., and Lehr, C.-M. (2006) Assessment of mucoadhesion by a resonant mirror biosensor. Int. J. Pharm. 325, 75–81 Woods, C.M., Neslund, G., Kornburst, E., and Flaim, S.F. (2000) Perflubron attenuates neutrophil adhesion to activated endothelial cells in vitro. Am. J. Physiol. Lung Cell Mol. Physiol 278, 1008–1017 Sheng, N., Fairbanks, M.B., Heinrikson, R.L., Canziani, G., Chaiken, I.M., Mosser, D.M., Zhang, H., and Colman, R.W. (2000) Cleaved high molecular weight kininogen binds directly to the integrin CD11b/CD18 (Mac-1) and blocks adhesion to fibrinogen and ICAM-1. Blood 95, 3788–3795 Watts, H.J., Lowe, C.R., and Pollard-Knight, D.V. (1994) Optical biosensor for monitoring microbial cells. Anal. Chem. 66, 2465– 2470 Nunomura, W., Takakuwa, Y., Parra, M., Conboy, C., and Mohandas, N. (2000) Regulation of protein 4.1R, p55, and glycophorin C ternary complex in human erythrocyte membrane. Biol. Chem. 275, 24540–24546

Label-Free Detection with the Resonant Mirror Biosensor 21. Nunomura, W., Takakuwa, Y., Parra, M., Conboy, C., and Mohandas, N. (2000) Ca2+dependent and Ca2+-independent Calmodulin Binding Sites in Erythrocyte Protein 4.1. Implications for regulation of protein 4.1 interactions with transmembrane proteins. J. Biol. Chem. 275, 6360–6367 22. Gill, A., Bracewell, D.G., Maule, C.H., Lowe, C.A., Hoare, M. (1998) Bioprocess. monitoring: An optical biosensor for rapid bioproduct analysis. J. Biotechnol. 65, 69–80 23. Rasooly, L., and Rasooly, A. (1999) Real time biosensor analysis of Staphylococcal enterotoxin A in food. Int. J. Food Microbiol. 49, 119–127 24. Skaldal, P. (1999) effect of methanol on the interaction of monoclonal antibody with free and immobilized atrazine studied using the resonant mirror-based biosensor. Biosens. Bioelectron. 14, 257–263 25. IAsys Technical note 3.1 BI (1994) The resonant mirror 26. Goddard, N.J., Pollard-Knight, D., and Maule, C.H. (1994) Real-time biomolecular interaction analysis using the resonant mirror sensors. Analyst, 119, 583–588 27. Goddard, N.J., Singh, K., Holmes, R.J., and Bastani, B. (1998) Resonant grating sensors using frustrated total-internal reflection. Sens. Actuators B 51, 131–136 28. Hulme, J., Malins, C., Singh, K., Fielden, P.R., and Goddard, N.J. (2002) Internally references resonant mirror for chemical and biochemical sensing. Analyst 127, 1233– 1236 29. Goddard, N.J., Singh, K., Hulme, J.P., Malins, C., and Holmes, R.J. (2002) Internally-referenced resonant mirror devices for dispersion compensation in chemical sensing and biosensing applications. Sens. Actuators A 100, 1–9 30. Ke, Y., Wilkinson, M.C., Fernig, D.G., Smith, J.A., Rudland, P.S., and Barraclough, R. (1992) A rapid procedure for production of human basic fibrobalst growth factor in E. coli. Biochim. Biophys. Acta 1131, 307–310 31. Kinsella, L., Chen, H.L., Smith, J.A., Rudland, P.S., and Fernig, D.G. (1998) Interactions of putative heparin-binding domains of basic fibrobalst growth factor and its receptor, FGFR-1, with heparin. Glycoconjugate J. 15, 419–422 32. Skerra, A., Pfitzinger, I., and Plückthun, A. (1991) The functional expression of antibody Fv fragemnts in Escherichia coli: Improved vectors and a generally applicable purification technique. Bio/Technology 9, 273–278

137

33. De Mey, J. (1986) The preparation and use of gold probes. Immunocytochemistry- Modern Methods and Applications, 2nd edition, Polak, J. M., Van Noorden, S., Wright and Sons, Bristol, UK 34. Richman, P.I., and Bodmer, W.F. (1987) Monoclonal antibodies to human colorectal epithelium: markers for differentiation and tumour characterization. Int J Cancer 39, 317–328 35. Willcocks, T.C., and Craig, I.W. (1990) Characterization of the genomic organization of human carcinoembryonic antigen (CEA): comparison with other family members and sequence analysis of 5′ controlling region. Genomics 8, 492–500 36. Edwards, P.R., Gill, A., Pollard-Knight, D., Hoare, M., Puckle, P.E., Lowe, C.R., and Leatherbarrow, R.J. (1995) Kinetics of protein–protein interactions at the surface of an optical biosensor. Anal. Biochem. 231, 210–217 37. Schuck, P. (1996) Kinetics of ligand binding to receptor immobilized in a polymer matrix, as detected with an evanescent wave biosensor. I. A computer simulation of the influence of mass transport. Biophys. J. 70, 1230–1249 38. Athanassopulou, N.R.J., Davies, P.R., Edwards, D.Yeung., Maule, C.H. (1998) Cholera toxin and GM1: a model membrane study with IAsys. Biochem. Soc. Trans. 27, 340–343 39. Brizard, B.L., Chubet, R.G., and Vizard, D.L. (1994) Immunoaffinity purification of FLAG epitope-tagged bacterial alkaline phosphatase using a novel monoclonal antibody and petide elution. BioTechniques 16, 730–735 40. IAsys Application note 1.3 (1994) Sensitivity enhancement using colloidal gold complexes 41. Fernig, D.G., and Gallagher, J.T. (1994) Fibroblast growth factors: an information network controlling tissue growth, morphogenesis and repair. Prog. Growth Factor Res. 5, 353–377 42. Turnbull, J.E., Fernig, D.G., Key, Y., Wilkinson, M.C., and Gallagher, J.T. (1992) Identification of the basic fibroblast growth factor binding sequence in fibroblast heparan sulphate. J. Biol Chem. 267, 10337–10341 43. IAsys application note 3.6 (1994) Molecular recognition, protein–carbohydrate interaction 44. Watts, H., Yeung, D., and Parkes, H. (1995) Real-time detection and quantification of DNA hybridization by an optical biosensor. Anal. Chem. 67, 4283–4289

138

Zourob et al.

45. Chaiet, I., and Wolf, F.J. (1964) The properties of streptavidin, a biotin-binding protein produced by Streptomycetes. Arch. Biochem. Biophys. 106, 1–5 46. IAsys application note 3.3 (1994) Molecular recognition Nucleic acid hybridization. 47. Ward, E.S., Gussow, D., Griffiths, A.D., Jones, P.T., Winter, G. (1989) Binding cativities of a reportoier of single immunoglobulin variable domains secreted from Escherichia coli. Nature 341, 544–546 48. IAsys application note 2.2 (1994), Kinetic analysis recombinant antibody fragment (D1.3 Fv) binding to immobilized hen egg lysozyme 49. Berry, M.J., Wattam, T.A., Willets, J., Lindner, N., de Graaf, T., Hunt, T., Gani, M., Davis, P.J., and Porter, P. (1994) Assay and purification of Fv fragments in fermenter cultures: design and evaluation of generic binding reagents. J. Immunol. Meth. 167, 173–182 50. Gill, A., Leatherbarrow, R.J., Hoare, M., Pollard-Knight, D.V., Lowe, P.A., Fortune, D.H. (1996) Analysis of kinetic data of antibody-antigen interaction from an optical biosensor by exponential curve fitting. J. Biotchnol. 18, 117–127 51. Grov, A., Myklestad, B., and Oeding, P. (1964) Immunochemical studies on antigen preparations from Staphaylococcus aureus.1. Isolation and characterization of antigen A. Acta Pathol. Microbiol. Scand 61, 588–596 52. Oeding, P., Grov, A., Myklestad, B. (1965) Antigenic properties of Staphylococci. Acta Pathol. Microbiol. Scand. 182, 183–190

53. Haukenes, G. (1974) Cellular antigen of Staphylococci. Acta Pathol. Microbiol. Scand. 236, 15–21 54. Gold, P., and Freedman, S.O. (1965) Demonstration of tumor-specific antigens in human colonic carcinomata by immunological tolerance and absorption techniques. J. Exp. Med. 121, 439–462 55. Goldenberg, D.M., Melville, M., and Carter, A.C. (1981) Carcinoembryonic antigen; its role as a marker in the management of cancer. A national Institute of Health Consensus Development Conference. Ann. Intern. Med. 94, 407–409 56. IAsys application note 5.2. (1994), Receptor–cell interactions, binding of L cells bearing the CEA antigen to an immobilized anti-CEA antibody 57. Absolom, D.R. and Van Oss, C.J. (1986) The nature of the antigen–antibody bound and the factors affecting its association and dissociation. CRC Crit. Rev. Immunol. 6, 1–46 58. Langone, J.J. (1982) Protein A of Staphylococcus aureus and related immunoglobulin receptors produced by Streptococci and Pneumococci. Adv. Immunol. 32, 157–252 59. IAsys application note 2.1 (1994), Kinetic analysis Protein A and Human IgG interaction 60. Zourob, M., Mohr, S., Treves-Brown, B.J., Fielden, P.R., McDonnell, M.B., and Nicholas, J.G. (2005) An integrated metal clad leaky waveguide sensor for detection of bacteria. Anal. Chem. 77, 232–242

Chapter 7 Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform Ian M. White, Hongying Zhu, Jonathan D. Suter, Xudong Fan, and Mohammed Zourob Summary Optical label-free detection prevents the cost and complexity of fluorescence and radio labeling while providing accurate quantitative and kinetic results. We have developed a new optical label-free sensor called the liquid core optical ring resonator (LCORR). The LCORR integrates optical ring resonator sensors into the microfluidic delivery system by using glass capillaries with a thin wall. The LCORR is capable of performing refractive index detection on liquid samples, as well as bio/chemical analyte detection down to detection limits on the scale of pg/mm2 on a sensing surface. Key words: Optical ring resonator, LCORR, Whispering gallery modes, Refractive index detection, Protease detection, DNA sequence detection.

1. Introduction The liquid core optical ring resonator (LCORR) sensing platform (1–8) integrates micro-capillary fluidics with label-free optical ring resonator sensing technology. Optical ring resonators have been studied for a decade for sensing applications (9–24). However, due to the use of the capillary as the ring resonator, the LCORR inherently integrates the sensor head with the sample fluidics, which can increase optical performance while simplifying the system design. The LCORR sensing platform is illustrated in Fig. 1. The ring resonator is formed in the circular cross section of the

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_7

139

140

White et al.

Fig. 1. LCORR sensing platform. (a) The ring resonator is defined in the cross-section of the LCORR capillary, and is excited by evanescent coupling from a fiber taper or waveguide; (b) An evanescent field at the inner surface interacts with the sample; (c) Detection of analytes causes a spectral shift in the resonant wavelength; the spectral position of the resonant wavelength over time forms the sensing signal.

capillary. The capillary wall acts as a waveguide; light travels repeatedly in a circle around the circumference of the capillary as the ring guides the light through total internal reflection. Light is evanescently coupled into the ring resonator using a fiber taper or a pedestal waveguide. Light that forms an integer number of wavelengths around the circumference of the ring is resonant; the resonating modes are called whispering gallery modes (WGMs). As shown in Fig. 1b, the WGM has an evanescent field that extends beyond the inner wall of the LCORR capillary, where it interacts with the sample as it moves through the inside of the capillary. Analytes are detected by immobilizing biorecognition molecules (e.g. antibodies) that capture the analytes at the inner LCORR surface, where they interact with the evanescent field of the WGM. The presence of the analytes in the optical field changes the effective refractive index experienced by the WGM. Thus the effective optical path length around the ring changes, which causes the resonating wavelength to shift spectrally. This change in the WGM spectral location over a period of time as the sample passes through the LCORR capillary is the sensor signal, as illustrated in Fig. 1c. In this chapter, the method for creating LCORRs is described, the preparation of the LCORR sensor setup is presented, and the steps for performing sensing measurements are given in detail. Protocols for some sensing applications of the LCORR are

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

141

presented. These include a bulk refractive index measurement, an assay for detecting protein adsorption and protease activity, and an assay for detecting the presence of a specific DNA sequence in a sample.

2. Materials 2.1. Components and Equipment for Producing LCORRs and Fiber Tapers

1. CO2 laser. #F48-2(S)W, 25W (Synrad, Inc., Mukilteo, WA). 2. Glass tubes. Various tubes have been used as the LCORR preform, including aluminosilicate tubes (#A120-85-10, Sutter Instrument, Novato, CA), silica tubes, (#Q120-90-10, Sutter Instrument), and silica capillary tubing (#TSP530660, Polymicro Technologies, Phoenix, AZ). 3. Fiber optic cable. Single mode fiber, #SMF28 (Corning, Inc., Corning, NY). 4. Motion controller. 4-axis motion controller, #NI PCI-7390 (National Instrument, Austin, TX). 5. Mechanical slide type 1. Belt-drive slide #ZF1 (Techno, Inc., New Hyde Park, NY). 6. Mechanical slide type 2. Sherline linear slide (no part #) (Sherline, Inc., Vista, CA). 7. Data acquisition card. PCI-based 37-pin NI-DAQ, #NI PCI-6221 (National Instrument, Austin, TX). 8. UV-curable glue. Norland Optical Adhesive #8101 (Norland, Cranbury, NJ). 9. Plain microscope slides, Fisherbrand #12-550A (Fisher Scientific, Pittsburgh, PA). 10. Gas torch. The Little Torch (Smith Equipment, Watertown, SD). 11. Fiber optic clamp. adjustable force magnetic clamp, #T711250, (Thorlabs, Newton, NJ).

2.2. Components and Equipment for the Experimental Setup

1. Pumps. Various pumps have been used for moving the samples through the LCORRs, including a syringe pump (#55-1140, Harvard Apparatus, Holliston, MA) and a peristaltic pump (Masterflex #7562-10, Cole-Parmer, Vernon Hills, IL). 2. Tunable laser. Various tunable lasers have been used in the experimental setup, including a butterfly-packaged 1,550 nm distributed feedback (DFB) laser (JDSU #CQF935, JDS Uniphase Corp., Milpitas, CA), a 785 nm DFB laser (#DL100, Toptica Photonics), and a 980 nm external cavity laser (Velocity #6309, New Focus, San Jose, CA).

142

White et al.

3. Photodetector. Large area IR photoreceiver, #2033 (New Focus). 4. Data acquisition card. PCI-based 37-pin NI-DAQ, #NI PCI6221 (National Instrument, Austin, TX). 5. Thermo-electric cooler (TEC). #DT6-6-01 (Marlow Industries, Dallas, TX). 6. TEC controller. #LDT5910-B (ILX Lightwave, Bozemen, MT). 7. Thermister. #YSI 44036RC (YSI Temperature, Dayton, OH). 8. Tubing. Tygon microbore tubing, 0.01″ inner diameter, 0.03´ outer diameter, #EW-06418-01 (Cole-Parmer). 2.3. Buffers and Reagents

1. 18 MW water (referred to as water throughout). Produced by the EASYpure UV, #D7401 (Barnstead/Thermolyne Corp., Dubuque, IA). 2. Ethanol. Absolute, 200 proof, #E7023 (Sigma-Aldrich, St. Louis, MO). 3. Methanol. Absolute, acetone-free, #M1775 (Sigma-Aldrich). 4. Phosphate buffered saline (PBS) tablets. #P-4417 (SigmaAldrich). 5. Hydrochloric acid (HCl). Certified A. C. S. plus, #A144-212 (Fisher Scientific). 6. Hydrofluoric acid (HF). 48% A. C. S. reagent, #244279 (Sigma-Aldrich). 7. 3× SSC buffer. Prepare 0.45 M sodium chloride (NaCl, #S271, Fisher Scientific) and 0.045 M sodium citrate (#S1804, SigmaAldrich) in water.

2.4. Silanes, Cross-Linkers, and Biomolecules

1. 3-aminopropyl-trimethoxylilane (3-APS). 97%, #281778 (SigmaAldrich), store at 4 °C. 2. Dimethyl adipimidate dihydrochloride (DMA). 99%, #285625 (Sigma-Aldrich), store at 4 °C. 3. 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide HCl (EDC): #22980 (Pierce Biotechnology, Rockford, IL), store frozen. 4. N-Hydroxysulfosuccinimide (sulfo-NHS): #24310 (Pierce Biotechnology), store at 4 °C. 5. Bovine serum albumin (BSA). Minimum 96%, #A2153 (Sigma-Aldrich). 6. Oligonucleotide probes. Purchased through Sigma-Genosys (Sigma-Aldrich) with the custom-designed sequence: 5′-CCAACCAGAGAACCGCAGTCACAAT; the 5′ end is aminated, and has a 6-Carbon spacer. 7. DNA samples. All DNA samples are custom designed SigmaGenosys oligonucleotides of length 25mer, 50mer, and 100mer (Sigma-Aldrich).

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

143

3. Methods This chapter describes in detail the process for first setting up an LCORR measurement and then the process for performing a sample analysis with the LCORR. At this stage of the development of the LCORR for use in bio/chemical analysis, it is necessary to assemble many of the components in the lab, and to produce LCORRs on site. Figure 2 presents a flow chart of all the steps necessary to conduct a single sample analysis, from producing the LCORR to analyzing the data. This chapter further explains how to create the experimental setup, as well as how to assemble on-site LCORR and fiber taper production systems. 3.1. Producing LCORR Capillaries

The primary component of the LCORR sensing system is a glass capillary that acts simultaneously as a microfluidic channel for sample delivery and as a ring resonator for sample detection. This

Fig. 2. The process for conducting a bio/chemical molecule detection experiment using the LCORR sensing platform.

144

White et al.

capillary is unlike the typical capillaries utilized for liquid or gas sample movement because the wall of the LCORR must be very thin. In fact, for sensing purposes, the LCORR capillary wall must be less than 5 µm thick. As capillaries with this dimension are not available today, it is required to produce LCORR capillaries in the lab before conducting any detection assays. To produce a capillary with a thin wall, we stretch a commercially available glass capillary to thin its dimensions (see Note 1). This is analogous to the process of drawing fiber optic cable from a preform. The preform glass capillary is heated to the softening point while one end is pulled. As wall thickness is critical, attention must be given to the temperature and pulling speed, as these parameters dictate the change in the aspect ratio (diameter to wall thickness) during pulling. Higher speed and lower temperature will combat the effects of surface tension, which is pulling the softened glass radially inward and decreasing the aspect ratio. Preservation of nearly 100% of the aspect ratio during capillary drawing has been exhibited with our LCORR drawing technique, and has been demonstrated in similar work (25). The configuration utilized for drawing LCORR capillaries and a photo of the setup are shown in Fig. 3. The entire apparatus is contained within a clear acrylic enclosure to reduce air currents, which can cause fluctuations in the temperature of

Fig. 3. (a) Diagram of LCORR drawing setup (reproduced from ref. 8 with permission from SPIE). (b) Photo of a constructed setup to draw LCORRs.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

145

the heating zone. CO2 lasers are used as the heat source to soften the glass capillary in the heating zone (see Note 2). Two CO2 lasers are used on opposite sides to provide more evenly distributed heating (see Note 3). The power of the lasers, which is computer-controlled using a data acquisition card, is dependent upon the type of glass used (see Notes 4 and 5). While the lasers soften the glass tube in one spot, the two stages holding each end of the glass tube are moved to stretch the glass. The tube can be taped onto each stage, although any temporary clamp will suffice. The stages are mechanical slides with stepper motors for movement. The movement of the stages is controlled with a PCIbased motion controller. One stage is pulled quickly away from the heating zone while the other is slowly pushed toward the heating zone, keeping constant the mass of glass in the heating zone. The ratio of pulling speed to feed-in speed controls the diameter of the pulled capillary (see Note 6). In this prototype implementation, a computer code controlling the stage movements and laser power are developed using LabView software (National Instruments). User inputs to the program include the desired laser power, pulling speed, and feed-in speed. The program operates through the DAQ card and the PCIbased motion controller to generate the appropriate voltage-based outputs to control the laser power and the pull/feed stages. The steps to be performed for pulling an LCORR capillary are listed below: 1. The preform glass capillary is bridged between the two stages, using clamps or tape to hold the glass in place. 2. The lasers are turned on to the power level used for softening the glass (see Note 7). 3. About 5 s are allowed to pass before activating the stages so that the glass has time to heat and soften. 4. The pulling and feed-in stages are moved at a constant speed until they reach the pre-set point, corresponding to the desired length. 5. The LCORR capillary is removed from the clamps by handling it with the remaining ends of the preform. With typical pulling conditions, the capillary is relatively robust, and can be handled easily. 6. The LCORR capillary can be cut from the final glass piece, although in some prototyping applications, it is desirable to keep at least one end of the remaining preform for handling purposes. The resonant region of the LCORR capillary extends from the thinned portion in the final heating zone to the portion of the capillary, where the equilibrium in the diameter is reached. Figure 4 illustrates the resonant region of a drawn LCORR capillary, and shows a photo of an LCORR drawn from a glass tube preform.

146

White et al.

3.2. Pulling Tapers

For prototyping and pre-clinical applications of the LCORR sensing system, it is sometimes practical to utilize a tapered fiber optic cable to excite WGMs in an LCORR (see Note 8). The operating principle of the fiber taper is that a fiber optic cable that is thinned to a few micrometers in diameter will have an evanescent field outside the cable. This evanescent field is capable of coupling light into the LCORR. In general, fiber tapers must be produced in the lab in which they will be used because of their fragility. Thus, taper production is presented here as a necessary component in the LCORR assay development. An illustration of the setup for pulling fiber tapers and a photo of the setup are presented in Fig. 5. The taper is produced by stretching a fiber optic cable under heat. Typical single mode fibers are utilized as the tapering fiber. The heat source is a clean flame provided by gas torch (see Note 9) (26). The fiber optic cable is stretched slowly from both sides of the heat zone while the heat zone is scanned back and forth by about 1 cm

Fig. 4. LCORR capillary after pulling from a preform.

Fig. 5. (a) Setup for pulling fiber tapers. (b) Photo of a constructed setup to pull tapers.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

147

in order to provide a region of constant diameter. Just as with the LCORR pulling apparatus, the stages are mechanical slides controlled by a computer via a code developed in LabView and a PCI-based motion controller. The LabView program is very similar to the one used for pulling LCORR capillaries. It may be desirable to pass laser light through the fiber and to measure the loss during taper pulling to monitor the health of the taper while it is produced. The steps to be performed for pulling fiber taper are listed below: 1. Prepare a length of fiber cable that is long enough such that the leads on each side of the taper can be connected to the experimental setup (see Subheading 3.3). One meter is generally sufficient. 2. In the center of the fiber cable, strip away the polymer jacket along a 4 cm region and clean away dust and polymer remains. This will be the tapering region. 3. Using fiber optic clamps mounted on the stages, mount the fiber into the setup. 4. Position the stage holding the torch such that the tube is centered with respect to the tapering region. 5. Turn on the gas for the torch and light the flame. Position the torch so that the tapering region is being heated to the desired temperature. 6. Allow enough time before pulling for the oven to heat to the silica’s softening temperature. This may take 5 s. 7. When the glass is softened, the stages pull in opposite directions at a speed of around 0.007 cm/s. Meanwhile, the torch is scanned back and forth at a speed of 0.1 cm/s, with a travel distance of 5 mm. 8. When the desired taper diameter is reached, the stages immediately stop pulling and the torch is turned off (see Notes 10 and 11). 9. Mount the taper on a rigid holder such that the taper is exposed and not in contact with anything, while being anchored on both sides for support (see Note 12). Figure 6 shows a drawing and a photo of an anchored taper. 3.3. Creating the Experimental Setup

The purpose of the experimental setup is to pass the sample through the LCORR while monitoring the sample’s effect on the spectral position of the WGMs. The experimental setup is given in Fig. 7. The LCORR is connected to tubing so that the sample can be passed through, using a pump. The taper is connected at one end to a tunable laser (see Note 13) and at the other to a photodetector. The tunable laser scans across a spectral range

148

White et al.

Fig. 6. (a) Sketch and (b) photo of a fiber taper anchored onto a portable mount.

Fig. 7. LCORR sensing platform experimental setup (reproduced from ref. 7 with permission from SPIE).

wide enough to detect a WGM and track its movement. Typically, 10–15 GHz (i.e., about 100 picometers (pm) for a center wavelength of 1,550 nm) of tuning range is sufficient for this. When the scanning laser passes through a resonant wavelength, destructive interference occurs on the fiber taper at the coupling point with the LCORR, resulting in an observable decrease in the output power. Thus, scanning the laser across one WGM will

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

149

produce a measured waveform like the one shown in Fig. 7 (see Note 14). Tunable laser control and data collection are performed by a data acquisition (DAQ) card, under the direction of the code written in LabView. The DAQ voltage output scans the output wavelength of the laser while the input samples the voltage at the output of the photodetector. After each scan, the sample values are stored in respective files on the computer. Also, the LabView program displays the measured voltage in real time. Following the experiment, a simple program written in Matlab (Mathworks) is used to scan each file for the spectral location of the voltage minimum (indicative of the WGM), and then a data set is created to represent the WGM spectral position over time. This data set is the sensorgram. An exemple of a sensorgram is shown in Fig. 1c. As with many label-free optical sensors, the signal from the LCORR sensor is temperature dependent (4). Therefore, a temperature control system is required in the experimental setup to suppress temperature fluctuations in the LCORR sensing region that would be translated to noise in the sensing signal. Our design is based on a thermo-electric cooler (TEC) and TEC controller. The setup is illustrated in Fig. 8.

Fig. 8. (a) Temperature control setup for the LCORR sensor. (b) Photo of a taper and LCORR mounted on the temperature control setup. A white line is drawn along the taper so that it can be visualized.

150

White et al.

The steps to be performed for creating the experimental setup are listed below. 1. Create the temperature control setup. Place the TEC heat sink and TEC on the bench top (or on a platform elevated above the bench top). It is recommended to use a thermally conductive adhesive between materials in this setup. Place a machined (as shown in Fig. 8) conductive block on top of the TEC. This is the TEC controlled mount. 2. Place a thermister onto (or inside a machined hole of) the TEC controlled mount. Mounting the thermister closer to the LCORR sensing region results in a better temperature control performance. 3. Connect the leads of the TEC and the thermister to the TEC controller. 4. Next, prepare the LCORR. Attach tubing to each side of the LCORR while anchoring the LCORR onto the mount. The tubing diameter should be just large enough to fit around the diameter of the LCORR and should be compatible with either a peristaltic pump or a syringe pump. The mount must leave the sensing region of the LCORR accessible while providing support (see Notes 15 and 16). A drawing and a photo are shown in Fig. 9. 5. Mount the LCORR on a 3-dimensional optical stage, such that the sensing region is accessible to a fiber taper. 6. Mount the fiber taper on a 3-deminsional stage such that the tapered region can be brought into contact with the LCORR. 7. Connect one end of the taper to the tunable laser output and the other to the photodetector input (splicing or free-space coupling can be used). 8. Connect the DAQ voltage output to the laser wavelength control (see Note 17); connect the DAQ voltage input to the photodetector output. 9. Begin scanning the laser and monitoring the photodetector signal. 10. Bring the LCORR in contact with the tapered region of the fiber while monitoring the WGMs (see Note 18). Use a microscope camera to assist in the positioning process. The tapered region of the fiber should contact the LCORR resonant region. The two objects are at approximately 90° with respect to each other, as shown in Figs. 7 and 8. Attempt to excite the LCORR resonance with different positions along the taper to optimize the coupling strength. Generally, the best coupling will occur at the thinnest point of the taper.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

151

Fig. 9. (a) Sketch and (b) photo of an LCORR with attached tubing, mounted on a portable mount.

11. When the preferred locations of the taper and LCORR are identified, lower the LCORR to be in contact with the LCORR heat sink as shown in Fig. 8. Then bring the taper into contact with the LCORR. 12. Connect one of the tubes that are connected to the LCORR to the pump and use the pump to move the sample through the LCORR. For liquid samples, a peristaltic pump or a syringe pump can be used for slow, constant flows (see Notes 19 and 20). 3.4. Preparing and Characterizing the LCORR for Sensing

After pulling the LCORR capillary, the wall thickness may not be as thin as desired for sensing purposes. Additionally, the sensitivity must be characterized before use. Sub-micron differences in wall thickness result in significant differences in the sensitivity of the LCORR, and thus it must be well-characterized. Therefore, once an LCORR is pulled, it is placed in the experimental setup for sensitivity optimization and characterization. To set the sensitivity of the LCORR, hydrofluoric acid (HF) is used to slowly etch away the inner capillary surface. The amount of glass to be removed depends on the preform and the capillary

152

White et al.

pulling process. For example, if the preform has an outer diameter of 1 mm and an inner diameter of 0.9 mm, and the LCORR’s outer diameter is 100 µm, then the thinnest that the wall can be is 10 µm. To achieve a good sensitivity, around 7 µm may need to be removed with HF. This is done by passing diluted concentrations of HF through the LCORR while intermittently characterizing the sensitivity. The sensitivity is characterized using solutions of known refractive indices, as the sensing mechanism of the LCORR is based on refractive index detection. We use solutions of ethanol in water (1), as the difference in refractive index of the varying concentrations is known (27). Determining the WGM spectral shift for a particular change in refractive index of the sample leads to the refractive index sensitivity of the LCORR sensor. The steps for optimizing and characterizing the sensitivity of the LCORR are listed here: 1. Prepare solutions of HF in water (see Note 21). The desired strength of HF depends on the type of glass used for the LCORR. Silica etches slowly, so stronger HF must be used. Aluminosilicate or other glasses etch much faster than silica. Typically, 5–10% HF is used for silica, while 1–2% may be used with aluminosilicate. 2. Prepare solutions of ethanol. The desired concentrations depend on the sensitivity goal for the LCORR. For example, if 10 nm/RIU (refractive index units) is desired, and considering that a 10 pm WGM spectral shift is practical to detect, the test samples should be prepared in increments of approximately 1 mRIU. This is equivalent to increments of about 2% (v/v) ethanol in water. 3. With the LCORR in contact with the taper, begin flowing HF through the LCORR (see Note 22). Monitor the WGM spectrum (as described in Subheading 3.3) while the LCORR is etching (see Note 23). When the LCORR becomes sensitive, the spectral position will drift to lower wavelengths. This is because the wall thickness is slowly decreasing, causing more light to interact with the sample, and thus changing the effective optical diameter of the LCORR. The speed of the WGM shift is an indicator of the wall thickness (see Note 24). 4. Once the spectral movement of the WGMs is easily visible, the sensitivity of the LCORR should be characterized. 5. Stop the etching by passing water through the LCORR. The WGM spectral movement should stop almost instantly after the water begins passing through the LCORR. 6. Once the WGM spectral position has stabilized, the refractive index sensitivity test can be conducted. While monitoring the WGM spectral position, change the sample solution in the LCORR from water to the lowest concentration of ethanol.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

153

Record the WGM spectral shift (an exemplary WGM spectral shift is shown in Fig. 10). When the refractive index inside the LCORR increases, the WGM spectral position should move to a longer wavelength (red shift). Allow the mode to stabilize, and then change the solution in the LCORR back to water, while recording the WGM spectral shift. 7. Repeat this a few times. Record the WGM spectral shift in picometers (pm). If no spectral shift is visible, return to the HF etching step. If the WGM spectral shift is apparent at all, proceed with the characterization process. 8. Repeat step 6 (water–ethanol solution–water) with the other concentrations of ethanol while recording the WGM spectral shift that occurs for each solution change. 9. Determine the average WGM shift for each of the ethanol solutions. 10. Plot the spectral shift vs. change in the refractive index and perform a linear fit of the data. The slope of this fit is the refractive index sensitivity (see Note 25). An exemplary refractive index sensitivity plot is presented in Fig. 11 for an LCORR with a sensitivity of 315 pm/RIU. 11. If the refractive index sensitivity is sufficient, then the optimization and characterization process is complete. If not, then return to the HF etching step. If the sensitivity is close, then it may be prudent to reduce the concentration of the HF. 12. Clean the inside surface of the LCORR thoroughly with any desired glass treatment, such as 1:1 methanol:HCl. At least

Fig. 10. The observed WGM shift in response to a 10% ethanol solution replacing water inside an LCORR with a sensitivity of 315 pm/RIU.

154

White et al.

Fig. 11. The measured refractive index sensitivity for an exemplary LCORR. Data is recorded for ethanol solutions of 10%, 20%, 30%, 40%, and 50%. The values are the result of an average across at least six observed WGM shifts. The slope of the linear fit is the refractive index sensitivity.

2 h of treatment is recommended. Caution should be exercised with the reactivity of any of the glass treatments with tubing used in the experimental setup. 3.5. Bulk Refractive Index Detection

The characterization procedure outlined in Subheading 3.4 illustrates that the LCORR sensing platform can be used to identify the refractive index of sample liquids. Once the sensitivity has been characterized, liquids of unknown refractive index can be passed through the LCORR while the WGM spectral position is monitored. This can be used, for example, to identify if a sample has a small amount of contaminants. The following procedure is used to measure the bulk refractive index. 1. Prepare the experimental setup described in Subheading 3.3. 2. Prepare a baseline solution. Ideally, this solution will be relatively close in refractive index to the sample. For example, if looking for the amount of water contamination in ethanol, then pure ethanol should be used as the baseline solution. 3. Use the pump to drive the baseline solution through the LCORR. It is recommended to move the liquid slowly, e.g., 10 µL/min. 4. Monitor the WGM spectral position. 5. Quickly switch to the sample, while trying to minimize any air gaps between the baseline liquid and the sample (see Note 26).

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

155

6. Record the WGM spectral shift that occurs as the sample replaces the baseline liquid in the LCORR. 7. Switch back to the baseline liquid, and repeat the measurement several times. 8. Compute the average value of the WGM spectral shift. 9. The refractive index difference between the baseline and the sample is easily computed by dividing the measured WGM spectral shift by the LCORR sensitivity measured in the characterization procedure. 3.6. Detection of Protein Binding and Proteolytic Activity

The LCORR can utilize its refractive index sensing capabilities to detect biomolecule analytes that are captured at the inner surface of the capillary. Typically, biomolecules have refractive indices around 1.45–1.55, while buffers are typically close to the range of 1.3–1.35. Thus, when biomolecules bind to the surface, the local refractive index in the region of the WGM evanescent field increases. This RI increase is reflected in the sensor signal by a red shift of the WGM spectral position. Figure 12 illustrates an exemplary sensor signal reflecting the binding of BSA molecules at the inner surface of the capillary. Because the total WGM spectral shift is proportional to the number of molecules that bind to the surface, the concentration of analytes in the sample can be determined when using the LCORR as a sensor. Through the same refractive index sensing mechanism that causes a red-shift for binding analytes, a blue shift in the WGM spectral position occurs when analytes are removed from the capillary surface. Thus, the LCORR can also be used to detect

Fig. 12. Sensorgram showing the binding of BSA molecules at the inner surface of the LCORR capillary. The BSA concentration is 1 mg/mL.

156

White et al.

proteolytic activity. Figure 13 shows the sensor signal when trypsin is introduced into the same LCORR following the binding of BSA (shown in Fig. 12). Trypsin is known to cleave BSA molecules at a number of residues, which will cause a significant amount of the BSA mass to be removed from the LCORR. As an example of protein detection, the steps for detecting BSA molecules in a sample are listed below: 1. Prepare a solution of 10% ethanol in 90% water (v/v). 5 mL of solution is sufficient. 2. Prepare the surface functionalization silane. Use 1% (v/v) of 3-aminopropyltrimethoxysilane (3-APS) in the 10/90 ethanol/water. It is recommended to prepare about 2 mL of 3-APS solution. 3. Prepare the phosphate buffered silane (PBS). Dissolve a PBS tablet in water to prepare 0.01 M PBS as indicated in the manufacturer instructions. 4. Prepare the BSA solution. Dissolve the desired concentration of BSA in PBS. The sample volume required is very small. However, for practical purposes, it may be desirable to preparerepare at least 100 µL of BSA solution. If very low concentrations are used, a higher volume, such as 1 mL, should be prepared. A cross-linker will be added to this solution, but it should not be done until just before the sample is used. Otherwise, the protein molecules may be cross-linked together before the experiment begins. 5. Prepare the LCORR for surface functionalization. Begin by passing 1:1 HCl:methanol through the LCORR for at least

Fig. 13. Sensorgram showing the proteolytic activity of trypsin acting to remove BSA bound at the inner surface of the LCORR capillary. The trypsin concentration is 10 µg/mL.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

157

30 min. Then rinse with methanol for at least 10 min. For all solutions throughout this measurement, the flow velocity should be kept below 10 µL/min. 6. Begin passing the 10/90 ethanol/water solution through the LCORR. Allow this solution to pass through the LCORR for approximately 15 min to establish a starting baseline for the WGM spectral position. Begin recording the WGM spectral position. 7. Quickly switch to 3-APS solution, minimizing the flow gap (without changing the pump speed) between the buffer solution and the 3-APS solution. The 3-APS should be passed through the LCORR for at least 15 min. The WGM spectral position should shift to a higher wavelength during the deposition, but should reach steady state after around 15 min (see Note 27). 8. Switch back to the 10/90 ethanol/water solution in the LCORR. This will cause some 3-APS to be released from the surface, resulting in a WGM blue shift. Once the blue shift is complete (10–15 min at most), allow another 10 min to establish the new signal baseline. The deposition of 3-APS should result in a red shift from the initial baseline WGM spectral position to the current position. 9. Begin passing PBS buffer through the LCORR and establish a new baseline position. 10. During this WGM stabilization, complete the preparation of the BSA solution. Add 1-Ethyl-3-[3-dimethylaminopropyl] carbodiimide hydrochloride (EDC) and N-hydroxysulfosuccinimide (sulfo-NHS) to the BSA/PBS solution to a concentration of 1 mg/mL each. EDC with sulfo-NHS is used to convert carboxyl groups to amine-reactive Sulfo-NHS esters (see Note 28). This solution should be prepared immediately before use. 11. Run the BSA solution through the LCORR for 30–45 min. A red shift with similar features to the one shown in Fig. 12 should occur. If the concentration of the BSA is high (greater than 0.1 mg/mL), then the total red shift may occur within as little as 5–10 min. 12. Rinse the LCORR with BSA. Again, the rinse should cause a blue-shift as some BSA will be removed from the surface of the LCORR. Once the blue shift is complete (10–15 min at most), allow another 10 min to establish the final signal baseline. The binding of BSA produces the red shift from the WGM spectral position after step 9 to the current WGM spectral position. 13. To determine the concentration of protein molecules in an unknown sample, the net WGM spectral shift should be

158

White et al.

compared to a calibration curve that plots the net WGM shift vs. protein concentration. The calibration curve is prepared using the same steps as above, but with samples that contain a known concentration of protein molecules. Naturally, the LCORR used to test the sample should have the same sensitivity as the LCORR used to produce the calibration curve. 14. Once the concentration in the sample is known, the reaction kinetics between the molecules in the sample and the capture molecules on the surface can be analyzed. The WGM spectral shift is related to the protein concentration in the sample by (14): dl WGM =

dlmax [protein] , K d + [protein]

where Kd is the dissociation constant and δlmax is the maximum WGM spectral shift in the calibration curve. Using this expression, the Kd can be determined for any protein sample. 15. If proteolytic activity is to be measured, the enzyme of interest should be dissolved in PBS to the desired concentration and passed through the LCORR. A blue shift in the WGM spectral position with features similar to the sensorgram shown in Fig. 13 should occur. 3.7. Detection of Specific DNA Sequences

One important application of the label-free biomolecule detection capability of the LCORR sensing platform is the identification of a specific DNA sequence in a sample. Similar to DNA microarray technology, an oligonucleotide probe that is designed to be the complement of the single-stranded target is immobilized on the sensor surface. If the target sequence exists in the sample, it will bind with an immobilized oligo probe at the LCORR inner surface. As shown in Subheading 3.6, the LCORR quantitatively detects biomolecules that bind to the surface. Surface chemistry for successful immobilization of the oligo probes is critical. First, the surface of the LCORR and the 5′ end of the oligo probe are functionalized with amine groups. Then, dimethyl adipimidate (DMA) is used to crosslink the oligo probes to the amino-functionalized LCORR surface. This surface chemistry is illustrated in Fig. 14. Upon immobilizing the probes, the LCORR is prepared to detect the presence of the complementary probe sequence in the sample. Figure 15 presents an exemplary sensorgram, which shows the WGM spectral shift during the functionalization, oligo probe immobilization, and sample analysis processes. The steps for conducting the DNA sequence detection measurement are outlined below: 1. Prepare a solution of 10% ethanol in 90% water (v/v). 5 mL of solution is sufficient.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

159

Fig. 14. Surface chemistry for immobilizing oligo probes to the LCORR surface.

Fig. 15. Exemplary sensorgram for 3-APS functionalization, 25 base-pair oligo probe immobilization, and sample detection (25 base pair DNA sequence). In this case, all steps, including the 3-APS deposition, were performed in 3× SSC buffer. Therefore, there are no WGM spectral shifts due to buffer changes.

2. Prepare the surface functionalization silane. Use 1% (v/v) of 3-aminopropyltrimethoxysilane (3-APS) in the 10/90 ethanol/water. It is recommended to prepare about 2 mL of 3-APS solution. 3. Prepare 3× SSC buffer. 3× SSC is 0.45 M NaCl and 0.045 M sodium citrate. 10 mL of buffer is sufficient. 4. Prepare oligonucleotide probe solution. Dissolve the probe pellet into 3× SSC to produce a concentration of 10 µM. Prior to opening the tube of pelleted DNA, it may be advantageous to centrifuge it for 10–15 min to ensure that there is no aerosol material lost. Sonicate the solution to ensure that the oligos are dissolved. DMA will be added to this solution

160

White et al.

immediately before probe immobilization onto the LCORR. It should not be done in advance to avoid DMA hydrolysis and cross-linking of the oligo probes. 5. Prepare the DNA sample in 3× SSC (see Note 29). In many cases, the DNA sample will be the output of RT-PCR targeted at the sequence of interest. 6. Prepare the LCORR for surface functionalization and probe immobilization. Begin by passing 1:1 HCl:methanol through the LCORR for at least 30 min. Then rinse with methanol for at least 10 min. For all solutions, the flow velocity should be kept below 10 µL/min. 7. Begin passing 10/90 ethanol/water through the LCORR. If interested in using the LCORR sensing functionality to monitor the surface functionalization and probe immobilization processes, begin recording the WGM spectral position at this point. Let the solution run for at least 15 min while recording the spectral position in order to establish an initial baseline WGM position. 8. Quickly switch to 3-APS solution, minimizing the flow gap (without changing the pump speed) between the buffer solution and the 3-APS solution (see Note 26). The 3-APS should be passed through the LCORR for at least 15 min. The WGM spectral position should shift to a higher wavelength during the deposition, but should have reached steady state after around 15 min. 9. Switch back to the 10/90 ethanol/water solution in the LCORR. This will cause some 3-APS to be released from the surface, resulting in a WGM blue shift. Once the blue shift is complete (10–15 min at most), allow another 10 min to establish the new signal baseline. 10. Begin passing 3× SSC buffer through the LCORR and allow a new baseline position to be established. 11. During this WGM stabilization, complete the preparation of the oligo probe solution. Add DMA to the oligo solution to a concentration of 5 mg/mL. This solution should be prepared immediately (within 10 min) before use. 12. Run the DMA/oligo probe solution through the LCORR for 30–45 min. Because the DMA loses its reactivity after approximately one hour, there is no benefit in running the solution any longer. The WGM spectral position should undergo a red-shift as the oligo probes are immobilized onto the surface of the LCORR. 13. Rinse the LCORR with SSC. Again, the rinse should cause a blue-shift as some oligo will be removed from the surface of the LCORR. Once the blue shift is complete (10–15 min at most), allow another 10 min to establish the new signal baseline.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

161

14. The LCORR is now prepared for the sample. Pass the sample through the LCORR while continuing to record the WGM spectral position. Allow the sample to pass through until the WGM has stabilized following the red shift. This may be as long as 1 h. Recycling the sample solution through the LCORR may enable more opportunities for the sample to bind at the LCORR surface. 15. Run a final SSC rinse through the LCORR. Once again, a blue shift will result as some target DNA is removed from the surface. After the blue shift is complete (10–15 min at most), allow another 10 min to establish the final WGM spectral position. 16. To determine the concentration of DNA molecules with the sequence of interest in an unknown sample, the net WGM spectral shift should be compared to a calibration curve that plots the net WGM shift vs. DNA concentration. The calibration curve is prepared using the same steps as above, but with samples that contain a known concentration of DNA molecules matching the sequence of interest. Naturally, the LCORR used to test the sample should have the same sensitivity as the LCORR used to produce the calibration curve. 17. Once the concentration in the sample is known, the reaction kinetics between the DNA molecules in the sample and the capture probe on the surface can be analyzed. The WGM spectral shift is related to the target DNA concentration in the sample by (14): dl WGM =

dlmax [DNA] , K d + [DNA]

where Kd is the dissociation constant and δlmax is the maximum WGM spectral shift in the calibration curve. Using this expression, the Kd can be determined for any sample.

4. Notes 1. Two different options have been used for the LCORR capillary preform: glass tubes with an outer diameter of 1.2 mm and an inner diameter of 0.9 mm, and silica capillaries with an outer diameter of 617 µm and an inner diameter of 535 µm. 2. A wire filament or heated ceramic oven can also be used as long as the temperature inside can be raised to the softening point of the glass without damaging the heating element.

162

White et al.

3. The lasers are slightly angled with respect to the perpendicular axis of the LCORR so that the beam of one laser is not hitting the opposite laser. 4. For quartz tubes with an outer diameter of 1.2 mm and an inner diameter of 0.9 mm (Suter Instruments), approximately 5 W from each laser has been used. For aluminosilicate tubes of the same size, approximately 2.5 W from each laser has been used. 5. It may be advantageous to have a feedback control design for the laser power, in which a laser micrometer monitors in real time the size and shape of the LCORR and invokes dynamic power adjustments of the lasers through computer control. 6. A typical pull speed is just over 1 cm/s while feeding at a speed of about 0.02 cm/s; this results in a diameter of approximately 115 µm from a glass tube with in initial diameter of 1.2 mm. 7. Extreme caution should be used with CO2 lasers. At these power levels, they can easily damage lab materials and are harmful to users. Moreover, the light is completely invisible. 8. Ultimately, in a clinical version of the LCORR, the capillaries would be mounted in a package on top of an array of pedestal waveguides, which would be used to excite WGMs (2–3). 9. A few options exist for pulling tapers (26, 28–30). 10. The desired diameter of the taper may depend on the wavelength of light utilized. Lower wavelengths will have a shorter evanescent field, and thus tapers for lower wavelength sources must be pulled thinner for sufficient evanescent exposure. 11. For SMF28, each stage pulls approximately 1.5 cm to produce a taper of approximately 3 µm. The difference in pulling length for a 1,500 nm source and a 600 nm source may be only about 1 mm. 12. A U-shaped holder made from glass microscope slides is assembled and UV-curable glue (Norland) is used to anchor the taper onto the arms of the mount. 13. Several options are available for tunable lasers. One common approach is to use distributed feedback (DFB) lasers, such as the Toptica LD100 or the JDSU CQF935. In this case, changing the laser gain current, tunes the output wavelength. Also, external cavity lasers, such as the New Focus Velocity 6300 series, can be used. For these lasers, applying a variable voltage signal onto a piezo-controlled grating or prism will tune the output wavelength. 14. Scanning the laser across 100 pm may result in the appearance of many WGMs.

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

163

15. UV-curable glue is used as the sealant. It is important to consider the sealant reactivity or solubility with any solutions that will pass through the LCORR. 16. The sealant tends to be pulled into the tubing by capillary action. It must be ensured that the sealant is not pulled far enough inside to block the sample’s path through the LCORR. 17. Two methods are commonly used for tuning lasers. In the case of distributed feedback (DFB) lasers, typically the gain current is modulated, which results in small wavelength shifts. In external cavity lasers, a piezo control for the external grating may be available for scanning the wavelength. 18. We use the DAQ and the LabView program to view the photodetector signal (and thus the WGMs) in real time. However, if this option is not available, a properly triggered oscilloscope will also work. 19. Some peristaltic pumps have been found to cause pressure variations inside the LCORR, which can add noise to the sensing signal (the WGM is extremely sensitive to the diameter of the LCORR, which can be altered slightly under varying pressure). 20. It is recommended to pull the sample instead of pushing it through the LCORR to reduce the possibility of breaking any of the seals along the fluidic path. 21. HF is a dangerous acid. Extreme caution should be exercised when using HF. Do not use glassware for containing the HF. 22. The LCORR etches more quickly if the HF solution is constantly moved through the LCORR, as opposed to sitting stagnant inside the LCORR. 23. A WGM spectral shift may occur very soon after the HF begins passing through the LCORR. However, this may be due to the temperature change caused by the reaction of the HF with the glass. 24. The spectral width of the WGM is expected to increase during the etching process. As the evanescent field in the liquid core increases, more light absorption occurs, which decreases the Q-factor and thus increases the spectral width of the mode. Furthermore, surface roughness of the LCORR due to etching can also reduce the Q-factor, especially as the evanescent field at the surface increases due to thinner walls. 25. First, convert the percentage of ethanol into a mole fraction of ethanol in water. Then use the expression from Ghoreyshi et al. (27) to find the difference in refractive index between water and the ethanol solution: 0.179258X - 0.380008X2 + 0.351867X3 – 0.124503X4

164

White et al.

where X is the mole fraction. This expression is validated for data taken at 632.8 nm at 25 °C. 26. Keeping the pump speed constant throughout the measurement is highly recommended, as pressure changes in the LCORR will cause shifts in the WGM spectral position. 27. Instead of pumping it through the LCORR, it can be made to sit still in the LCORR, allowing the 3-APS to diffuse to the surface. Changing the pump speed (or stopping the pump) corrupts the sensing signal, so this should only be done if the sensing signal is not desired for this step. 28. In the case of BSA deposition on an aminated surface, crosslinkers are not necessary, as the BSA will adsorb to the surface. Cross-linkers here are used for illustrative purposes, as in many typical measurements, they may be required to effectively bind the biomolecule to the surface. 29. If it is necessary to use a different buffer for the sample, then a baseline WGM spectral position must be established when switching buffers between the probe immobilization and sample deposition. This is similar to when the baseline is established for SSC buffer after the 3-APS deposition and 10/90 ethanol/water rinse.

References 1. White, I. M., Oveys, H., and Fan, X. (2006) Liquid core optical ring resonator sensors, Opt. Lett. 31, 1319–1321 2. White, I. M., Oveys, H., Fan, X., Smith, T. L., and Zhang, J. (2006) Integrated multiplexed biosensors based on liquid core optical ring resonators and antiresonant reflecting optical waveguides, Appl. Phys. Lett. 89, 191106191101–191106-1-3 3. White, I. M., Suter, J. D., Oveys, H., and Fan, X. (2007) Universal coupling between metal-clad waveguides and optical ring resonators, Opt. Express. 15, 646–651 4. Suter, J. D., White, I. M., Zhu, H., and Fan, X. (2007) Thermal characterization of liquid core optical ring resonator sensors, Appl. Opt. 46, 389–396 5. Zhu, H., White, I. M., Suter, J. D., Zourob, M., and Fan, X. (2007) An integrated refractive index optical ring resonator detector for capillary electrophoresis, Anal. Chem. 79, 930–937 6. Fan, X., White, I. M., Zhu, H., Suter, J. D., and Oveys, H. (2007) Overview of novel integrated optical ring resonator bio/chemical sensors, Proc. SPIE 6452, 64520M

7. White, I. M., Zhu, H., Suter, J. D., Oveys, H., and Fan, X. (2006) Liquid core optical ring resonator label-free biosensor array for lab-on-a-chip development, Proc. SPIE 6380, 63800F 8. White, I. M., Oveys, H., Fan, X., Smith, T. L., and Zhang, J. (2007) Demonstration of a liquid core optical ring resonator sensor coupled with an ARROW waveguide array, Proc. SPIE 6475, 647505 9. Laine, J. -P., Tapalian, H. C., Little, B. E., and Haus, H. A. (2001) Acceleration sensor based on high-Q optical microsphere resonator and pedestal antiresonant reflecting waveguide coupler, Sens. Actuators A 93, 1–7 10. Arnold, S., Khoshsima, M., Teraoka, I., Holler, S., and Vollmer, F. (2003) Shift of whispering-gallery modes in microspheres by protein adsorption, Opt. Lett. 28, 272–274 11. Vollmer, F., Arnold, S., Braun, D., Teraoka, I., and Libchaber, A. (2003) Multiplexed DNA quantification by spectroscopic shift of two microsphere cavities, Biophys. J. 85, 1974–1979 12. Hanumegowda, N. M., White, I. M., Oveys, H., and Fan, X. (2005) Label-free protease

Label-Free Detection with the Liquid Core Optical Ring Resonator Sensing Platform

13.

14.

15. 16.

17.

18.

19.

20.

21.

22.

sensors based on optical microsphere resonators, Sensor Lett. 3, 315–319 Hanumegowda, N. M., White, I. M., and Fan, X. (2006) Aqueous mercuric ion detection with microsphere optical ring resonator sensors, Sens. Actuators B 120, 207–212 Zhu, H., Suter, J. D., White, I. M., and Fan, X. (2006) Aptamer based microsphere biosensor for thrombin detection, Sensors 6, 785–795 Boyd, R. W. and Heebner, J. E. (2001) Sensitive disk resonator photonic biosensor, Appl. Opt. 40, 5742–5747 Blair, S. and Chen, Y. (2001) Resonantenhanced evanescent-wave fluorescence biosensing with cylindrical optical cavities, Appl. Opt. 40, 570–582 Armani, A. M. and Vahala, K. J. (2006) Heavy water detection using ultra-high-Q microcavities, Opt. Lett. 31, 1896–1898 Chao, C. -Y. and Guo, L. J. (2003) Biochemical sensors based on polymer microrings with sharp asymmetrical resonance, Appl. Phys. Lett. 83, 1527–1529 Krioukov, E., Greve, J., and Otto, C. (2003) Performance of integrated optical microcavities for refractive index and fluorescence sensing, Sens. Actuators B 90, 58–67 Ksendzov, A., Homer, M. L., and Manfreda, A. M. (2004) Integrated optics ring-resonator chemical sensor with polymer transduction layer, Electron. Lett. 40, 63–65 Ksendzov, A., and Lin, Y. (2005) Integrated optics ring-resonator sensors for protein detection, Opt. Lett. 30, 3344–3346 Chao, C. -Y., Fung, W., and Guo, L. J. (2006) Polymer microring resonators for biochemical sensing applications, IEEE J. Sel. Top. Quant. 12, 134–142

165

23. Yalcin, A., Popat, K. C., Aldridge, O. C., Desai, T. A., Hryniewicz, J., Chbouki, N., Little, B. E., King, O., Van, V., Chu, S., Gill, D., Anthes-Washburn, M., Unlu, M. S., and Goldberg, B. B. (2006) Optical sensing of biomolecules using microring resonators, IEEE J. Sel. Top. Quant. 12, 148–155 24. Yang, J. and Guo, L. J. (2006) Optical sensors based on active microcavities, IEEE J. Sel. Top. Quant. 12, 143–147 25. Fitt, A. D., Furusawa, K., Monro, T. M., and Please, C. P. (2001) Modeling the fabrication of hollow fibers: capillary drawing, J. Lightwave Technol. 19, 1924–1931 26. Sumetsky, M., Dulashko, Y., and Hale, A. (2004) Fabrication and study of bent and coiled free silica nanowires: self-coupling microloop optical interferometer, Opt. Express 12, 3521–3531 27. Ghoreyshi, A. A., Farhadpour, F. A., Soltanieh, M., and Bansal, A. (2003) Transport of small polar molecules across nonporous polymeric membranes, J. Membr. Sci. 211, 193–214 28. Laine, J. -P., Little, B. E., and Haus, H. A. (1999) Etch-eroded fiber coupler for whispering-gallery-mode excitation in high-Q silica microspheres, IEEE Photonic Tech. L. 11, 1429–1431 29. Knight, J. C., Cheung, G., Jacques, F., and Birks, T. A. (1997) Phased-matched excitation of whispering-gallery-mode resonances by a fiber taper, Opt. Lett. 22, 1129–1131 30. Tong, L., Gattass, R. R., Ashcom, J. B., He, S., Lou, J., Shen, M., Maxwell, I., and Mazur, E. (2003) Subwavelength-diameter silica wires for low-loss optical wave guiding, Nature 426, 816–819

Chapter 8 Reflectometric Interference Spectroscopy Guenther Proll, Goran Markovic, Lutz Steinle, and Guenter Gauglitz Summary Reflectometry is classified in comparison to the commercialized refractometric surface plasmon resonance (SPR). The advantages of direct optical detection depend on a sophisticated surface chemistry resulting in negligible nonspecific binding and high loading with recognition sites at the biopolymer sensitive layer of the transducer. Elaborate details on instrumental realization and surface chemistry are discussed for optimum application of reflectometric interference spectroscopy (RIfS). A standard protocol for a binding inhibition assay is given. It overcomes principal problems of any direct optical detection technique. Key words: Label-free optical biosensor, Reflectometric interference spectroscopy (RIfS).

1. Introduction The methods available for direct monitoring of biomolecular interaction can be divided into methods measuring changes in the refractive index of the interaction layer, and methods measuring changes in reflectometry at the layer (1). Regarding the first method, BiaCore (2) has opened the market by introducing surface plasmon resonance (SPR) (3) as a very promising tool in biomolecular interaction analysis (BIA). In contrast to refractometry, reflectometry concentrates on the measurement of changes in physical thickness. Reflectometry has been introduced many decades ago as ellipsometry using polarized light. Interference at the interfaces of the layer causes a change in the relative amount of amplitude of the two polarized radiation beams and in phase. This interferometric method has been applied to a very simple analytical method, called reflectometric interference

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_8

167

168

Proll et al.

spectroscopy (RIfS) (Figs. 1 and 2) to be used as an example of a very robust, simple optical detection principle in chemical and biochemical sensing (4). This label-free optical detection method for surface interactions is based on white light interference at transparent thin layers. At each interface of thin layers of different materials with negligible absorption, radiation is partially reflected and transmitted. If the optical path length through these layers is less than the coherence wavelength, the different partial beams interfere, and form an interference pattern depending on the wavelength, the optical thickness, which is given by the product of the refractive index of the layer and its physical thickness, the incident angle, and the refractive index of the surrounding medium (5, 6). In case of perpendicular incidence, a nonabsorbing layer, and low reflectances, the reflec-tance R is given by:

Fig. 1. Scheme of the RIfS detection principle. The left part shows the superimposition of the reflected light beams and the change in optical thickness during a binding event on the sensor surface. The right part shows the corresponding change of the characteristic interference spectrum and the resulting binding curve.

Halogen lamp Photo diode array

PC

Transducer with Interference Layer

Fig. 2. RIfS setup.

Reflectometric Interference Spectroscopy

R = R1 + R2 + 2 R1R2 cos(4p nd / l ),

169

(1)

where R1 and R2 denote the Fresnel reflectance at the two interfaces, d is the physical thickness of the film, n its refractive index, and l the wavelength of incident light. A typical interference pattern showing the modulation of reflectance with cos(1/l) is given in Fig. 1. The optical thickness nd can be determined from the position of an extremum with a given order value m by nd =

ml . 2

(2).

RIfS uses the change in the optical properties in or at the top layer of a given layer system as detection principle (Figs. 1 and 2). The binding of an analyte molecule or particle to the sensor surface causes a shift of the interference pattern in the wavelength domain. To evaluate the binding signal, the locus of an extremum is tracked over time; thus, the change of the interference spectrum results in a time-resolved binding curve representing the binding of the analyte molecule to the sensor surface. A major advantage of RIfS is its resistance to changes in temperature (7). Refractometric methods such as SPR and ellipsometry, on the other hand, are very sensitive to temperature variations due to the high impact of temperature on the refractive index. Thus, temperature changes during a measurement cause negative effects with these methods, and quick changes of temp-erature between measurements are technically challenging. Since the refractive index n is dependent on the density given from the Clausius–Mossotti equation, and the density is dependent on the thermal expansion, the refractive index decreases with increasing temperature. Because of the thermal expansion of the biopolymer layer, the physical thickness d increases with increasing temperature. These two contrary temperature-dependent effects result in a rather low influence of temperature on the optical thickness – the product of the refractive index and the physical thickness.

2. Materials Common chemicals of analytical grade are purchased from Sigma or Merck. Milli-Q water is deionized water with a conductivity of 18.2 MΩ cm−1. 2.1.Transducer

The RIfS standard transducer consists of a glass substrate (D 263 glass, Schott AG, Germany) (~1-mm thick) coated with a 10-nm

170

Proll et al.

layer of a material with high refractive index (usually Ta2O5 or Nb2O5), and a top layer of SiO2 (330 nm). As a reference use a transducer without SiO2 layer. 2.1.1. Glass Substrates

1. BK 7 glass, n = 1.51, Schott, Mainz. 2. WG 345 glass, n = 1.699, Schott, Mainz. 3. Interference transducer: D 263, 10-nm Ta2O5, 500-nm SiO2, Schott, Mainz. 4. Goethe glass: multilayer system: 1-m D 23, 45-nm Ta2O5, 20-nm SiO2, Schott, Mainz (cut in small squares (10 mm × 10 mm) ).

2.1.2. Parallel Setup

1. White light source 100 W/12 V, Osram, Munich. 2. Lenses, mirror, and positioning optics, Spindler & Hoyer, Göttingen. 3. Polymer light guides (PMAA, n = 1.490, coupling element 1 × 2 (50:50), 1 mm, fiber diameter with SMA 905 fiber connectors, Microparts, Dortmund. 4. Optical 4 × 1 multiplexer DiCon VX 500-C, laser components, Olching. 5. MMS diode row spectrometer with Liliput-PC, ZEISS, Jena. 6. Nineteen-inches industrial standard housing by RS Components, Walldorf-Mörfelden. 7. 486 PC (Windows 95).

2.1.3. Single Setup

1. Modified simultaneous spectrometer SPECKOL 1100, Zeiss, Jena. 2. 486 PC (Windows 95).

2.1.4. Liquid Handling

1. Ten-position valve VICI, Valco Europa, Schenkon, Switzerland. 2. HPCL three-way valve VICI, Valco Europa, Schenkon, Switzerland. 3. Six-position valve, Bischoff, Leonberg. 4. Peristaltic pump Reglo-Digital MS2/8–160 ISM 832, Ismatec, Wertheim. 5. Peristaltic pump MS Fixo, Ismatec, Wertheim. 6. High-grade steel capillaries, screws, and fittings from Rheodyne, USA.

2.1.5. Software

1. MeasureCR for capturing spectras and controlling the systems. 2. IFZCR for evaluating the interferograms. 3. MS-Excel 7.0 and Microcal Origin 5.0 for further processing of measured data.

Reflectometric Interference Spectroscopy

2.2. Surface Chemistry

171

1. GOPTS (3-glycidyloxypropyltrimethoxysilane) purum: Toxic. Store at room temperature; keep under argon, sensitive to humidity (Fluka). 2. AMD (Aminodextran): MW 100 kD, level of amination 50%. Store at 4 °C under dry conditions (Innovent e.V. Technologieentwicklung, Jena, Germany). 3. Dicarboxypoly- and diaminopoly(ethylene glycol) (PEG): MW 2,000 Da. Store at −20 °C under dry conditions (Rapp Polymere, Tuebingen, Germany). 4. NHS (N-hydroxysuccinimide) purum: Store under room temperature (Fluka). 5. DIC (N,N´-diisopropylcarbodiimide) purum: Toxic. Store at room temperature, keep under argon, sensitive to humidity (Fluka). 6. EMCS (6-maleimidohexanoic acid N-succinimidyl ester) purum: Store at −20 °C under dry conditions (Sigma). 7. TBTU (2-(1H-benzotriazol-1-yl)-1,1,3,3-tetramethyluronium tetrafluoroborate). 8. HOBT (1-hydroxybenzotriazole) hydrate purum: Store at −20 °C at dry conditions (Fluka). 9. DIPEA (N,N-diisopropylethylamine): Highly inflammable. Store at room temperature (Sigma).

3. Methods 3.1. Setup for RIfS

In the standard laboratory setup for RIfS (see Note 1), the white light is guided via an optical fiber (1-mm PMMA fiber) to the back of the transducer mounted in a microfluidic flow cell, which in turn is attached to a liquid handling system. The reflected light is gathered in the same waveguide. A schematic of sample handling with flow injection analysis (FIA) for RIfS is shown in Fig. 3. The fiber optics used is bifurcated (50:50 ratio), with one tail leading to the light source and the other to the UV/Vis spectrometer. Possible light sources are halogen lamps (e.g., 10-W halogen lamp with fiber in-coupling optics consisting of front surface spherical mirror, collimating lens, and an infrared absorption filter) or LEDs. For the spectral detection of the reflected light, diode array spectrometers are used normally. A gap (approximately 100 µm) between transducer chip and the fiber output is filled with glycerol (80%) for refractive index matching. Samples are handled by a flow system (e.g., two peristaltic pumps, injectload valve, and six-way valve). Moreover, this flow system can be equipped with an autosampler. The various samples are injected

172

Proll et al.

Carrier Flow

D

P1 V1

Waste

P2

Waste

Loading: Injection:

V2 Sample Loop

Flow Direction while loading: while injection:

Fig. 3 Schematic of the FIA system for RIfS.

by the autosampler into a sample loop. From there, the sample is driven in continuous flow, passing the prepared transducer. The raw spectra are corrected for the dark current of the spectrometer by subtraction (if necessary), and normalized to the reflectance spectrum from a glass chip without the SiO2- and bio-layer. The position of an interference extremum at approximately 550 nm is determined by a parabolic fit to one half-wave of the interference spectrum. Optical thickness is calculated according to Eq. 2 (see Note 2). The costs of the setup can be reduced by sequentially illuminating the transducer with different suitable wavelengths coming from LEDs or a white light source with appropriate filters, detecting the intensity without spectral resolution using a photodiode, and reconstructing part of the interference pattern by fitting a parabola through the interpolation points. This detection principle can be used to realize a parallel screening system, which allows optical online detection of specific biomolecular interaction in 96- or 384-well microplate formats. Therefore, the whole area of the plate bottom consisting of a RIfS transducer is illuminated by a halogen light source combined with a filter wheel, which allows the subsequent passage of monochromatic light of seven different wavelengths. This is not only to reduce the costs, but is necessary because a CCD camera is used as a detector, which is able to detect the intensity of the whole area of interest in a single shot. 3.1.1. Silanization

1. Transducer chips are cleaned with freshly prepared Piranha solution (mixture of 30% hydrogen peroxide and concentrated sulfuric acid at a ratio of 2:3; caution: hot and aggressive!) for 30 min in an ultrasonic bath to clean the chip and to generate silanol groups on the transducer surface.

Reflectometric Interference Spectroscopy

173

2. After rinsing with double-distilled water and drying in a nitrogen stream, the surface is immediately activated by GOPTS at a surface concentration of approximately 10 µL cm−2 for 1 h in dryness (see Notes 3) (by assembling two slides face-to-face placed in a tray with ground joint) followed by cleaning with dry acetone and drying in a nitrogen stream (see Notes 4 and 5). 3.1.2. Biopolymer

When investigating interactions of biomolecules on surfaces, nonspecific binding to the sensor must be minimized. Therefore, the glass slide is coated with a shielding layer, which prevents nonspecific binding and additionally provides a large number of functional groups for the immobilization of the binding partner. The standard shielding chemicals used are dextrans, which form a 3D hydrogel loaded with a large number of binding sites, and polyethylene glycols (PEG) forming a two-dimensional brushlike monolayer with less binding sites but with a more defined surface. Amino and carboxy groups are normally functional groups for the immobilization process since the well-established peptide chemistry methods can be applied. Other shielding chemicals can be used depending on the application or the surface chemistry needed. 1. Aminodextran: The coupling of aminodextran as an aqueous solution (1:3) (approximately 15 µL cm−2) to the silanized surface is carried out by incubating over night (minimum 16 h) in a water-saturated atmosphere (see Note 6). After thoroughly rinsing with double-distilled water and drying, the prepared chips are stable for several months. 2. Polyethylenglycol: The coupling of diamino- or dicarboxypoly(ethylene glycol) (PEG) as a 1 mM solution in dichloromethane (DCM) (approximately 10 µL cm−2) to the silanized surface is achieved by incubating overnight at 70 °C, followed by thorough rinsing with Milli-Q water and drying in a nitrogen stream. 3. Change of functional groups: For some applications (e.g., immobilization of DNA oligomers with an aminolinker) the follow-up functionalization via peptide chemistry works better with diaminopoly(ethylene glycol) treated with 5 M glutaric anhydride (GA) in N,N-dimethylformamide (DMF) (10 µL cm−2) for 6 h to generate carboxylic groups on the surface. The same protocol can be used to modify AMD surfaces with carboxylic groups.

3.1.3. Immobilization of Amino-Terminated Ligands

1. The covalent coupling of amino-terminated molecules/ligands to biopolymers with carboxylic groups is done by standard peptide chemistry via activated esters (see Note 7). Therefore, the previously modified transducers are activated with a solution of NHS and DIC (1 M NHS, 1.2 M DIC in DMF,

174

Proll et al.

10 µL cm−2) for 2–4 h as sandwich pairs in a DMF saturated atmosphere (see Note 5). The transducers are rinsed first with DMF and then with dry Aceton, and dried in a nitrogen stream (see Notes 4 and 5). 2. To achieve a high density of covalently bound molecules, it is necessary to use an excess of reagents. Therefore, the ligand is applied as a DMF solution (10 µL cm−2) (if your ligand is insoluble in DMF, use another aprotic solvent of appropriate polarity) at a concentration of 1–3 µM to the activated surface for 4 h as sandwich pairs in a DMF saturated atmosphere (see Note 6). Clean the transducers by rinsing them with DMF and Milli-Q water and drying them in a nitrogen stream (see Note 4). 3.1.4. Immobilization of Carboxy-Terminated Ligands

1. For activation of the carboxy-terminated molecules, the ligand is dissolved in DMF at a concentration of 1–3 µM together with 1.5 M DIC. After quick mixing, this solution is immediately applied to a transducer modified with a biopolymer providing amino groups at a concentration of 10 µL cm−2. Place a second transducer on top to form a sandwich pair (see Note 6). 2. The reaction is finished after 4 h in a DMF-saturated atmosphere (see Note 6). Clean the transducers by rinsing them with DMF and Milli-Q water and drying them in a nitrogen stream (see Note 4).

3.1.5. Immobilization of Thiol-Terminated Ligands

1. Dissolve 1-mg EMCS per 10-µL DMF, and apply this solution to a transducer modified with a biopolymer offering amino groups at a surface concentration of 10 µL cm−2 for 6–12 h in a DMF-saturated atmosphere (see Note 6). After rinsing with DMF (see Note 4), dissolve the ligand at a concentration of 1–3 µM in DMF, and apply this solution to the transducers at a surface concentration of 10 µL cm−2 for 6–12 h as sandwich pairs in a DMF saturated atmosphere (see Note 6). 2. Clean the transducers by rinsing them with DMF and Milli-Q water and drying them in a nitrogen stream (see Note 4).

3.1.6. Immobilization of Biotin

1. Biotin is immobilized by TBTU activation: D-biotin (1 mg, 4 mmol), TBTU (1.4 mg, 4.4 mmol), and DIPEA (4 mL, 23.3 mmol) are mixed with DMF (50 mL) until the active ester is formed and the reaction mixture appears homogeneous. This solution is dripped on to a transducer at a surface concentration of 10 µL cm−2 pretreated with a biopolymer providing amino groups. Two of such transducers are put together to form a sandwich. 2. After a reaction time of 4 h in a saturated DMF atmosphere (see Note 6), the sandwich is separated and both transducers are rinsed with water and dried in a nitrogen stream.

Reflectometric Interference Spectroscopy

175

3.2. Measurement Protocol for Standard Binding Inhibition Assay for the Determination of the Affinity of an Antigen– Antibody Interaction

Binding inhibition assay to determine the affinity constant of antigen–antibody interaction. The transducer is modified according to protocols given before either with a carboxy or an amino-terminated antigen. Then, fixed amounts of antibody are preincubated with fixed volumes of differently concentrated antigens (see Note 8) to reach binding equilibrium before injection of the mixture with the FIA system (see Notes 9–12) on the transducer. Antigens with affinity to the antibody in the preincubated solution inhibit the binding of the antibody to the surface-immobilized antigen. In the case of diffusion (mass transport) limited binding, the diffusion rate obeys first Fick’s law. This results in a constant binding rate of the antibody to the surface and a linear binding curve. The slope of the binding curve is determined by the concentration of free antibody binding sites in solution. The slope of the binding curves decreases with increasing concentration of antigen in solution and reaches zero for high antigen concentrations. Ovalbumin (final concentration of 200 µg/mL) should be added to all antibody solutions to avoid loss of antibody by nonspecific binding in the fluidic system. The model function describing the concentration of antibodies with free binding sites is fitted to the titration curve using a Marquart–Levenberg nonlinear least-square algorithm (software ORIGIN from Microcal, Northampton/USA).

3.3. Data Evaluation

1. Determination of the concentration of active antibody Measuring the concentration of active antibody in a sample working at mass-transport limited conditions is essential. Accordingly, the rate of diffusion to the surface is much slower than the binding reaction to the surface. The reaction kinetics can therefore be neglected. Mass-transport limited conditions can be achieved by a high binding capacity of the surface and a low antibody concentration in solution. Under mass-transport limited conditions according to first Fick’s law, the resulting binding curve is linear, and its slope is proportional to the concentration of functional antibody. If all samples contain the same amount of protein, determined by UV spectroscopy or Bradford assay, it is possible to calculate the active antibody concentration in a.u. by the different slopes of the samples, where the sample with the highest slope is assigned the a.u. 1 per µg used protein. 2. Determination of affinity and kinetic constants To determine the affinity and the kinetics of an antibody (see Note 13), mass transport to the surface must be much faster than the rate of binding to the surface. If this is the case, the mass transport can be neglected. This can be achieved by a high concentration of bulk antibody and a low surface coverage of hapten derivative. The time dependence of the surface coverage can then be described by a pseudo first-order binding reaction as follows:

176

Proll et al.

dΓ (t ) = kac Ab (Γ max - Γ (t )) - kd Γ (t ), dt

(3)

with cAb: antibody concentration, G(t): surface coverage at time t, Gmax: maximal surface coverage, ka: association rate constant, and kd: dissociation rate constant. Solving this differential equation, the surface coverage that is equal to the resulting signal curve is given by Γ (t ) = Γ Eq (1 - e - kobst ),

(4)

with Γ Eq = Γ max

K aff c Ab , 1 + K aff c Ab

(5)

the equilibrium coverage and kobs = kac Ab + kd ,

(6)

the observed rate constant. The affinity constant Kaff is given by the fraction of the association rate constant, ka, and the dissociation rate constant, kd. To obtain accurate values for GEq and kobs, it is necessary that only the kinetics-controlled part of the signal curve is approximated, and that the surface and bulk are nearly in equilibrium at the end of the measurement. Now it is possible to determine ka as the slope of the straight line of the plot with kobs as ordinate and active antibody concentration as abscissa. In order to determine Kaff, Eq. 5 can be rewritten as 1 1 1 1 = + × . Γ Eq Γ max K aff ×Γ max c Ab

(7)

With this formula, it is possible to obtain the value for Gmax with 1 1 plot. Inserting this value in Eq. 3, a Γ Eq c Ab nonlinear least square fit results in a value for Kaff. This method

a linear fit of the

æΓ ö gives nearly the same values as a Scatchard plot ç Eq Γ Eq ÷ . è c Ab ø

4. Notes 1. Classical RIfS setup (Figs. 1 and 2) 2. Very thick biolayers (more than approximately 100 nm) lead to measurements out of the linear correlation between

Reflectometric Interference Spectroscopy

177

change of optical thickness and shift of the interference spectrum. 3. Correct side of the RIfS transducers: Mark the uncoated side with a diamond pen to avoid mistakes during the surface chemistry steps. The coated side of the transducers appears colored because of the light interference. 4. In the case a transducer shows a gray shadow after rinsing and drying repeat this cleaning procedure. If this does not work, start the surface chemistry from the beginning (otherwise the modification will not be homogeneous and will show high nonspecific binding). 5. All surface chemistry protocols that produce highly reactive groups are sensitive to humidity (deactivation). This can be avoided by immediately applying the next step and working under dry ambient conditions. 6. Use a tray with ground joint. Place the transducer as sandwich pairs (by assembling two slides face-to-face) in the tray on a solid support and add the same solvent that is used in the reaction for a solvent-saturated atmosphere. 7. In the case of limited ligands it is possible to use commercially available piezo-based microdosing devices for printing a ligand solution in Milli-Q water. 8. Usually phosphate buffered saline (PBS, 150-mmol NaCl and 10-mmol dipotassium hydrogen phosphate in Milli-Q water at pH 7.4) is used for antigen–antibody interaction analysis. In general, all buffers can be used for RIfS measurements. Only restriction: do not use buffers with pH > 8.5 because of destruction of the surface modification. 9. Testing of the modified transducer for nonspecific binding: use ovalbumine (OVA) or bovine serum albumin (BSA) in excess (e.g., 1 mg mL−1) directly before the measurement. 10. Air disturbs the measurements: Use degassed buffers and avoid negative pressure (sucking) through the flow cell. In addition, it is helpful to keep the buffer under a slight positive Argon pressure. 11. If possible use the same buffers during a complete measurement to avoid artifacts because of changes in the refractive index. 12. After the interaction process, the transducer surface can be regenerated. To remove antibodies from their antigens, a solution of 0.5% SDS (sodium dodecyl sulphate) at pH 1.9 is applied via the FIA system. In the case of hybridization experiments, a solution of either 0.25% SDS at pH 2.5 or a solution containing 6 M guanidinium hydrochloride and 6 M urea at pH 2 can be used.

178

Proll et al.

13. RIfS is not restricted for biosensing; this technique can be used also for chemical sensing. Instead of biopolymers one can modify the transducer with, e.g., nanoporous polymer layers, rubber-like polymers, or molecular imprinted polymers (MIPs). The determination of kinetics is also possible. References 1. Gauglitz, G. (2005) Direct optical sensors: principles and selected applications. Anal Bioanal Chem. 381(1), 141–155 2. www.biacore.com 3. Homola, J., Yee, S.S., and Gauglitz, G. (1999) Surface plasmon resonance sensors: review. Sensor Actuator. B54, 3–15 4. Schmitt, H.M., Brecht, A., Piehler, J., and Gauglitz, G. (1997) An integrated system for optical biomolecular interaction analysis. Biosens Bioelectron. 12, 219–233

5. Brecht, A., Gauglitz, G., and Nahm, W. (1992) Interferometric measurements used in chemical and biochemical sensors. Analusis. 20, 135–140 6. Brecht, A., Gauglitz, G., Kraus, G., and Nahm, W. (1993) Chemical and biochemical sensors based on interferometry at thin layers. Sensor Actuator. 11B, 21–27 7. Proell, F., Moehrle, B., Kumpf, M., and Gauglitz, G. (2005) Label-free characterisation of oligonucleotide hybridisation using reflectometric interference spectroscopy. Anal Bioanal Chem. 382(2), 1889–1894

Chapter 9 Phase Sensitive Interferometry for Biosensing Applications Digant P. Davé Summary A simple yet highly sensitive implementation of an interferometric technique for a label-free molecular biosensing application is described. The intereferometric detection method is based on the phasesensitive detection of spectral interference fringes. The change in optical path length due to binding of biomolecules on functionalized optically clear substrates can be quantified by detecting the change in the phase of the spectral fringes. The common path interferometeric design permits measurement of sub- monolayer binding of biomolecules to the sensor surfaces. Key words: Interferometry, Biosensor, Phase-sensitive.

1. Introduction Optical techniques used for biosensing can be put in two distinct categories, namely labeled and label-free techniques. Labeling of biomolecules of interest using fluorescent tags is widely used in biosensing. Given the simplicity of detection and high sensitivity of labeled techniques, they are the preferred method of biosensing in a wide variety of biomolecular recognition applications. Despite the success of fluorescent tags for biosensing their limitations are well recognized. Ideally one would prefer to recognize the biomolecules of interest without labeling them since it is possible that the label may interfere with the activity of the biomolecule and its interaction with the recognition molecule. Moreover, robust chemistry needs to be developed to attach fluorescent tags to each biomolecule of interest. In the absence of a label, intrinsic properties of the biomolecule need to be exploited for detection. Optical properties that can Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_9

179

180

Davé

be potentially exploited for biosensing include, autofluorescnce, Raman scattering, refractive index, optical path length (product of refractive index and geometric size) and polarization. Autofluorescence and Raman scattering are typically weak signals for sensitive biosensing applications. Though Raman signal can be amplified using surface enhanced Raman spectroscopy (SERS), it requires nanostructured surfaces. A change in refractive index and optical path length (OPL) can be accurately measured with techniques such as surface plasmon resonance (SPR) and interferometry. SPR is now a well established method in biosensing, particularly to monitor and characterize the kinetics of biomolecular binding. The introduction of SPR instrumentation by Biacore enabled widespread use of SPR in biochemistry research and application. Using interferometry, a change in refractive index or OPL can be measured with a sensitivity that translates into biomolecule detection sensitivity that is comparable or even better than any other widely used biosensing technique. Despite the potential, the use of interferometry for biosensing has been limited until now. Sensitivity required for biomolecular sensing requires implementation of phase sensitive interferometry. In recent years a number of phase sensitive interferometeric techniques have been reported which are well suited for biosensing (1–19). 1.1. Phase Sensitive Interferometry

By measuring the phase on an interference signal, sub-wavelength changes in OPL can be measured. Although phase information is readily available in any interferometric setup, environmental noise corrupts the phase information, rendering it useless. For robust phase measurement, interferometer design should enable cancellation of common mode noise. A common path interferometric implementation thus enables common mode noise cancellation. In this chapter a simple spectral interferometric technique to measure OPL changes in the picometer range and its implementation for biosensing are described. In principle, the described technique is similar to reflectrometric interferometric spectroscopy implemented for biosensing. A significant advantage of the technique described is that it does not require the use of specially fabricated surfaces, and in fact can be easily interfaced with commonly used optically clear substrates. The technique is based on phase-sensitive implementation of low coherence interferometery in the spectral domain (19–26).

1.2. Spectral Domain Phase Sensitive Interferometry

Consider an optical setup in which light from a partially coherent single light source after traveling through two optical paths (reference and sample) is recombined and the spectrum is measured. The recorded spectrum is not only the superposition of the spectrum of light from the sample and the reference path but has an additional term that is present due to the

Phase Sensitive Interferometry for Biosensing Applications

181

non-zero cross-correlation of reference and sample spectrums. This non-zero cross-correlation term modulates the spectrum, giving rise to spectral interference. The modulation frequency is proportional to the optical path-length difference between reference and sample paths. Recorded spectral interference, due to mixing of light from the reference path and sample path can be written as S (k) = S ref (k) + S samp (k) + 2 S ref (k)S samp (k) ×m (k) ×cos(2p k Dp + j 0 ), (1) where k is the wave vector, m(k) is the spectral coherence function, and ∆p is the optical path length difference between the reference path and sample path. Sref(k) and Ssamp(k) are the spectrum of light from the reference and sample path, respectively. ∆p is a product of the geometrical path length difference (∆z) and refractive index of the medium separating reference and sample path which can be calculated by Fourier transformation of Eq. 1. For biosensing applications it is necessary to measure the change in optical path length with sub-wavelength resolution as the biomolecule of interest binds to the recognition molecule attached to a substrate. The change in the optical thickness of the biolayer is typically less than 10 nm depending on the size of the biomolecule. Sub-wavelength changes in ∆p can me measured by detecting the change in phase of the frequency component of interest of the modulation term in Eq. 1 as follows æ Im(F (S (k)) ö 2p n Dz j |z =n Dz = tan -1 ç = lo è Re(F (S (k)) ÷ø 1.3. Implementation of SD-PSI for Biosensing

(2)

Fiber-based implementation of SD-PSI is shown in Fig. 1. SDPSI system can be constructed with commercially available components and minimal effort in optical alignment.

Fig. 1. Setup of a fiber-based spectral domain phase sensitive interferometer for biosensing applications.

182

Davé

2. Materials The part numbers listed below were used to build and test a prototype SD-PSI system operating at 830 nm for biosensing applications. Similar SD-PSI can be built at other operating wavelengths also. All the components can be purchased off-the-shelf from various manufactures. Given below is the list of components needed to set up a SD-PSI system for biosensing. 1. Fiber coupled SLD (SuperlumDiodes Inc. – SLD-37-MP will require a controller and a diode mount). 2. Single mode 2 × 2 fiber coupler (Thorlabs). 3. Fiber Isolator (OFR-IO-F-830). 4. Spectrometer (see Note 1). 5. Fiber Collimator (OFR-PAF-X-15-830). 6. Focusing Lens (10× microscope objective). 7. Cage System (Thorlabs). ●

Cage Rods (4) (SR6)

●

Cage Plate (2) (CP02)

●

Kinematic Mirror Mount (2) (KC1)

3. Methods 3.1. Interferometer Setup

Light from the broadband light source is coupled into a 2 × 2 (50:50 split ratio) single mode fiber (SMF) splitter. Any broadband light source can be used as long as the spectrum of the light source is stable and sufficient light can be coupled into a single mode fiber to maintain a signal-to-noise ratio, necessary to detect a desired phase change in the spectral interference fringes. The bandwidth of the light source affects the smallest thickness of the functionalized transparent substrate that can be used and also the precision with which the thickness of the biolayer can be measured. Fiber coupled superlumincsent diodes (SLD) are an excellent choice as broadband light sources for SD-PSI (see Note 2). These compact sources have a stable spectral and power output. Given the potential of damage to the SLD due to backreflection from the optical setup, a fiber isolator should be inserted between the SLD source and the input port of the 2 × 2 SMF coupler. In the common path configuration only one port of the coupler is used. The port that is not used should be angle polished or angle cleaved to avoid any back reflection from the fiber–air interface into the coupler. As an alternative the fiber can be dipped in 60–80% glyc-

Phase Sensitive Interferometry for Biosensing Applications

Functionalized glass slide

183

Microscope objective Kinematic mirror mount

Coupler output fiber

Inline collimator

FC/APC connector Fig. 2. Photograph of the SD-PSI sample path configuration showing various components integrated together using a cage system.

erol solution to avoid back reflection. This back reflection will reduce the signal-to-noise ratio of the interferometric signal and hence the sensitivity of measured phase (see Note 3). The other port of the fiber coupler should be terminated with an FC/APC connector. Light from the fiber port is first collimated and then focused onto the sample. A simple and robust setup for the sample path optics would be to mount all the opto-mechanical components on a cage system (Fig. 2). Optical components are mounted on cage plates and kinematic mirror mounts, which are in turn mounted on cage rods which ensure co-linearity of the optical train. Fiber connector with a snap-on collimator should be mounted on a kinematic mirror mount to enable a fine adjustment of collimated beam tip and tilt orientation with respect to the focusing lens. Choice of the focusing lens is not critical if it is assumed that the biomolecule binding per unit area remains constant. Under this assumption, phase measuring sensitivity remains constant, as the phase measured is the ratio of the area (footprint) of all attached biomolecules to the area of focused light on the biolayer being interrogated. A suitable sample geometry for biosensing is any transparent substrate (glass or plastic) of suitable thickness functionalized with chemistry for recognition of biomolecules of interest. The thickness of the transparent substrate that can be used is only limited by the resolution of the spectrometer used in the setup. The spectral modulation frequency is proportional

184

Davé

to the thickness of the substrate and that of the functionalized biolayer. In the sample optical configuration shown in Fig. 3, light reflects back onto the interferometer from the air–substrate (interface 1), substrate–functionalized biolayer (interface 2), functionalized biolayer–target biomolecule (interface 3) and target biomolecule–buffer (interface 4). Spectral modulation of various frequencies will occur due to interference between light reflection from various interfaces (1 and 2, 2 and 3, 3 and 4, 1 and 3, 1 and 4, 3 and 4). The modulation depth of the spectral interference signal at various frequencies is a function of the ratio of the reflecti-vities of the two interfaces involved. Inference will also occur due to multiple reflection of light from the various interfaces but the modulation depth is far smaller than the primary interference signal and the modulation frequency will be a multiple of the primary modulation frequency. Each interference signal occurs at a fixed spatial frequency band that is

Fig. 3. Depiction of change in spectral inference signal with the composite change in optical thickness of functional biosensing substrate. Example of bovine serum albumin binding to recognition molecule (EDC) is shown in the graph. EDC is a spacer that binds to functionalized –COOH glass coverslip.

Phase Sensitive Interferometry for Biosensing Applications

185

proportional to the optical thickness of the interfaces involved and can be measured by Fourier transformation of the recorded spectral interference. In the case of functionalized substrate for biosensing (Fig. 3), recording of kinetics and ultimately the detection of the target biomolecule concentration can be achieved by measuring the optical thickness changes that occur due to the formation of layer 4 with the binding of target biomolecule to the recognition layer 3. The spectral interference signal of interest is the interference signal that occurs between interface 1 and 4 or 2 and 4 or 3 and 4. Note that layer 3 is not a true monolayer unless all the recognition sites are filled with the target biomolecules. In principle, even if one biomolecule attaches to the recognition site a change in the phase of the interference signal will occur. Practically, the limit of the number of biomolecules detected will depend on the SNR and also on the spectral characteristics of the sources (wavelength, spectral stability and bandwidth). Phase noise of the reported SD-PSI systems is in the tens of picometer range, although the theoretical values calculated put the lower limit in the sub picometer range. Additional sources of noise will include mechanical vibration of the substrate resulting in probe beam scanning over the surface roughness, statistical binding and bonding variation of the target biomolecule and recognition molecule. Sensitivity with which biomolecules can be detected using SD-PSI can be written as Sensitivity (gm / mm 2 ) »

j N wm ´ , j T Am

(3)

where jT and jN are the total phase change due to a complete monolayer of biomolecule being interrogated and phase noise standard deviation of the SD-PSI system. wm is the weight of the biomolecule and Am is the footprint of the biomolecule. The sensitivity in Eq. 3 does not take into consideration the dynamics of target biomolecule diffusion and binding with recognition molecule, which can be affected by the size of the sensor area and methods to promote binding like agitation and circulation of analyte. Consider the example of IgG binding to a functionalized layer of anti-IgG. For a SD-PSI system with jN equivalent to 1 mrad (45 pm at 800 nm), approximating the IgG molecule to a sphere of 5 nm radius, a bound monolayer of IgG will give total phase change of jT (∆p = 7 nm at 800 nm, refractive index of 1.4), the sensitivity of detection is about 20 pg/mm2. Stepwise detection of biomolecule sensing on funtionalized optically clear substrate with a layer of recognition molecules using SD-PSI is diagrammatically shown in Fig. 4.

186

Davé

Fig. 4. Interference signal formed due to interference of partially reflected light from various interfaces. Binding of the target molecule to the receptor molecule will change the phase of the interference signal.

4. Notes 1. A scanning monochromator should not be used in place of an array based spectrometer as large spectral phase noise can be introduced due to mechanical scanning of the grating. 2. If the SD-PSI setup is used only to quantify the total concentration of biomolecules bound to the functionalized sensor surface, care must be taken to thoroughly clean the sensor chip after hybridization so that none of the cleaning solution or unbound target biomolecule residue remains on the chip. A reference surface with only the functionalized layer on the sensor chip is necessary to quantify the thickness of the bound target biomolecules. 3. Phase artifacts can be introduced in the measured binding kinetics sensogram due to the presence of gas bubbles at the sensor–liquid interface, changes in temperature of liquid in contact with the sensor surface, and mechanical movement of sensor chip. For binding kinetics measurements, a flow cell arrangement similar to the one used in SPR setups with degasser and proper temperature control of the liquids will minimize phase artifacts. References 1. G. H. Cross, A. A. Reeves, S. Brand, J. F. Popplewell, L. L. Peel, M. J. Swann, N. J. Freeman, “A new quantitative optical biosensor for protein characterization,” Biosensors and Bioelectronics 19, 383 (2003) 2. V. S. -Y. Lin, K. Motesharei, K. -P. S. Dancil, M. J. Sailor, M. R. Ghadiri, “A porous siliconbased optical interferometric biosensor,” Science 278, 840 (1997)

3. D. J. Bornhop, J. C. Latham, A. Kussrow, D. A. Markov, R. D. Jones, H. S. Sørensen, “Molecular interactions studied backscattering interferometry,” Science 317, 1732 (2007) 4. L. Peng, M. M. Varma, W. Cho, F. E. Regnier, D. D. Nolte, “Adaptive interferometry of protein on a BioCD,” Applied Optics. 46, 5384 (2007)

Phase Sensitive Interferometry for Biosensing Applications 5. M. M. Varma, H. D. Inerowicz, F. E. Regnier, D. D. Nolte, “High-speed label-free detection by spinning-disk micro-interferometry,” Biosensors and Bioelectronics 19, 1371–1376 (2004) 6. M. M. Varma, D. D. Nolte, H. D. Inerowicz, F. E. Regnier, “Spinning-disk self-referencing interferometry of antigen–antibody recognition,” Optics Letters 29, 950–952 (2004) 7. K. Haupt, A. -S. Belmont, S. Jaeger, D. Knopp, R. Niessner, G. Gauglitz, “Molecularly imprinted polymer films for reflectometric interference spectroscopic sensors,” Biosensors and Bioelectronics 22(12), 3267–3272 (2007) 8. K. AddedKroger, J. Bauer, B. Fleckenstein, J. Rademann, G. Jung, G. Gauglitz, “Epitopemapping of transglutaminase with parallel label-free optical detection,” Biosensors and Bioelectronics 17(11–12), 937–944 (2002) 9. A. Brecht, G. Gauglitz, G. Kraus, G. Lang, J. Piehler, J. Seemann, “Application of reflectometric interference spectroscopy to chemical and biochemical sensing,” In Sensor 95, 355–360 (1995) 10. K. Schmitt, B. Schirmer, A. Brandenburg, “Development of a highly sensitive interferometric biosensor,” Proceedings of SPIE 5461, 22 (2004) 11. O. Birkert, G. Gauglitz, “Development of an assay for label-free high-throughput screening of thrombin inhibitors by use of reflectometric interference spectroscopy,” Analytical Bioanalytical Chemistry 372, 141 (2002) 12. J. Hast, H. Heikkinen, L. Krehut, R. Myllyla, “Direct optical Biosensor based on optical feedback interferometry,” IEEE, 177 (2005) 13. W. B. Nowall, N. Dontha, W. G. Kuhr, “Electron transfer kinetics at a biotin/avidin patterned glassy carbon electrode,” Biosensors and Bioelectronics 13, 1237 (1998) 14. C. J. Easley, L. A. Legendre, M. G. Roper, T. A. Wavering, J. P. Ferrance and J. P. Landers, “Extrinsic fabry-perot interferometry for noncontact temperature control of nanolitervolume enzymatic reactions in glass microchips,” Analytical Chemistry 77, 1038 (2005) 15. B. H. Schneider, J. G. Edwards, N. F. Hartman, “Hartman interferometer: versatile integrated optic sensor for label-free, real-time quantification of nucleic acids, proteins, and pathogens,” Clinical Chemistry 43, 1757 (1997)

187

16. K. Schmitt, B. Schirmer, C. Hoffmann, A. Brandenburg, P. Meyrueis, “Interferometric biosensor based on planar optical waveguide sensor chips for label-free detection of surface bound bioreactions,” Biosensors and Bioelectronics 22, 2591 (2007) 17. N. Kinrot, M. Nathan, “Investigation of a periodically segmented waveguide Fabry– Pérot interferometer for use as a chemical/ biosensor,” Journal of Lightwave Technology 24, 2139 (2006) 18. D. A. Markov, K. Swinney, D. J. Bornhop, “Label-free molecular interaction determinations with nanoscale interferometry,” J. Am. Chem. Soc. 126, 16659 (2004) 19. J. Lu, C. M. Strohsahl, B. L. Miller, L. J. Rothberg, “Reflective interferometric detection of label-free oligonucleotides,” Analytical Chemistry 76, 4416 (2004) 20. R. Leitgeb, C. K. Hitzenberger, A. F. Fercher, “Performance of fourier domain vs. time domain optical coherence tomography,” Optics Express 11, 889 (2003) 21. J. F. de Boer, B. Cense, B. H. Park, M. C. Pierce, G. J. Tearney, B. Bouma, “Improved signal-to-noise ratio in spectral-domain compared with time-domain optical coherence tomography,” Optics Letters 28, 2067 (2003) 22. M. Choma, M. Sarunic, C. Yang, J. A. Izatt, “Sensitivity advantage of swept source and Fourier domain optical coherence tomography,” Optics Express 11, 2183 (2003) 23. M. A. Choma, A. K. Ellerbee, C. Yang, T. L. Creazzo, J. A. Izatt, “Spectral-domain phase microscopy,” Optics Letters 30, 1162 (2005) 24. C. Joo, T. Akkin, B. Cense, B. H. Park, J. F. de Boer, “Spectral-domain optical coherence phase microscopy for quantitative phase-contrast imaging,” Optics Letters 30, 2131 (2005) 25. N. Nassif, B. Cense, B. H. Park, M. Pierce, S. Yun, B. Bouma, G. Tearney, T. Chen, J. F. de Boer, “In vivo high-resolution video-rate spectral-domain optical coherence tomography of the human retina and optic nerve,” Optics Express 12, 367 (2004) 26. B. H. Park, M. C. Pierce, B. Cense, S.-H. Yun, M. Mujat, G. Tearney, B. Bouma, J. F. de Boer, “Real-time fiber-based multi-functional spectral-domain optical coherence tomography at 1.3 µm,” Optics Express 13, 3931 (2005)

Chapter 10 Label-Free Serodiagnosis on a Grating Coupler Thomas Nagel, Eva Ehrentreich-Förster, and Frank F. Bier Summary The unique feature of the label-free measurement techniques for screening specific binding molecules against a certain ligand is that knowledge about the analyte is not required. Due to the direct monitoring of the binding event, no further detection step, e.g., by a fluorescently labeled antibody, is necessary. This technique enables not only the analysis of binding properties, but also applications in serodiagnosis and in primary screening in drug discovery. Especially when complex biological solutions such as blood serum are used as sample fluids, the minimization of unspecific attachment is the crucial point of the assay. In this chapter, the basic handling of the grating coupler as an example of a label-free transducer is described, together with a simple protocol to minimize unspecific attachment when measuring undiluted blood serum. Key words: Label-free detection, Grating coupler, Blood serum, Serodiagnosis, Passivation of glass surfaces, Protein coupling.

1. Introduction During the last two decades, a variety of transducers in the field of optical label-free detection methods have been developed: surface plasmon resonance, resonant mirror, interferometric sensors and reflectometric interference spectroscopy (1), as also the grating coupler, another effective label-free transducer that is based on surface bound refractive index changes (2). All these methods allow the observation of binding events to the sensor surface in real-time. Due to the evanescent field of the guided wave in the grating coupler only refractive index changes at the near vicinity of the surface are detected. Such changes occur when proteins or other molecules bind to the surface. Interactions Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_10

189

190

Nagel, Ehrentreich-Förster, and Bier

between different types of biomolecules such as proteins, DNA and in some cases even small molecules (down to some hundred Daltons) are detectable. The avoidance of modifying the molecules of interest, e.g., with a fluorescence label, enables the work with biomolecules in their native structure. Furthermore, the real-time observation of the binding event allows the determination of association and dissociation of rate constants and consequently, more information can be obtained than from an end-point assay. Apart from these classical applications, efforts are made to use the label-free approach in the field of serodiagnosis and as a screening tool in drug discovery. For these applications, end-point determinations are sufficient in the majority of cases, which have a specific binder to the immobilized ligand, to identify these samples in a direct way. The crucial points for such measurements are the throughput of samples and the feasibility to use complex solutions like blood serum as sample fluids. The development of label-free devices that allow a throughput of thousands of samples within a day are far advanced (3, 4) and the use of complex sample solutions requires a coupling chemistry that minimizes the unspecific attachment to the surface (5, 6). To determine picomolar or even femtomolar concentrations of the analyte in blood as sample fluid, an indirect competitive assay can be performed (7). In this chapter, we describe a method for the passivation of glass surfaces for measurements with complex solutions, using a grating coupler as the transducer. In addition to these, the chemistry described here can also be used with similar systems based on sensor chips with a glass surface, such as interferometric sensors or reflectometric interference spectroscopy. 1.1. Sensor Principle

Three different grating coupler setups have been investigated: the input (8, 9), the output (10–12) and the reflection grating coupler (13). The advantage of the grating coupler in the reflection mode as compared to the other setups is that moving parts are not necessary. The optical arrangement is shown in Figs. 1 and 2. A convergent HeNe laser beam (λ = 633 nm) is irradiated onto the grating and the reflected light intensity is detected by a charge-coupled device (CCD) line sensor. Under the specific coupling angle α0, a part of the light is coupled into the waveguide and the electromagnetic field of the light wave reaches into the medium which has a lower refracting index (Fig. 3). This results in an exponentially decaying field, the evanescent wave. The position of the reduced intensity in the reflected light is observed and evaluated (Fig. 4). Each binding effect at the sensor surface results in a change of the effective refractive index Neff, leading to a shift of the coupling angle. This relation is given by the coupling condition: N eff = n0 sin a 0 +

kl . L

(1)

Label-Free Serodiagnosis on a Grating Coupler

191

Fig. 1. Optical arrangement for the grating coupler in the reflection mode. The distance between the grating coupler chip and the CCD sensor is about 100 cm. The wedge plate in front of the grating coupler chip and the positioning system is not pictured.

Fig. 2. A photograph of the optical arrangement of the grating coupler in the reflection mode. In the front is the flow cell which presses the grating coupler chip against the wedge plate. The cylinder lens, the neutral density filter and the laser are arranged in the background. In the right upper corner is the mirror which deflects the light to the here not shown CCD sensor.

where n0 denotes the refractive index in air, k the diffraction order, l the wavelength and L the grating period. The limit of detection of the grating coupler system is the minimum resolution of refractive index changes being in the range of 3 × 10−6. This corresponds to an approximate mass coverage of 10 pg/mm2 (14). Exemplary, biomolecules with a molecular weight of 150 kDa, like immunoglobulin G, can be detected down to a concentration

192

Nagel, Ehrentreich-Förster, and Bier

Fig. 3. Schematic illustration of the grating coupler chip with the in-coupled light under coupling conditions

Fig. 4. Intensity distribution recorded by the CCD sensor under coupling conditions (by courtesy of Jörg Henkel, FhG-IBMT). The shift of the left slope of the intensity minimum is determined.

of 0.3 nM (15). Since the grating coupler is sensitive only to changes of the refractive index very close to the surface, the use of complex sample solutions like blood serum affects the noise of the signal only in a minimal way. It is therefore possible to detect antibody concentrations in the low nM-range. The limiting factor is the discrimination between a specific binding event and unspecific attachment of sample compounds both happening at the surface. Consequently, the minimization of unspecific

Label-Free Serodiagnosis on a Grating Coupler

193

attachment is the crucial point for the detection of low analyte concentrations. 1.2. Surface Functionalization and Coupling Strategies

Coupling of biomolecules to surfaces can be performed with a considerable number of different strategies. The simplest approach is the physical adsorption of the ligand to the surface by electrostatic bonds, van der Waals interactions, hydrogen bonds, and/or hydrophobic interactions. However, this construction is thus stable only to a limited extent against changes in the ionic strength and the pH-value. Stable binding is obtained by using covalent coupling protocols which generate a linkage between reactive groups on the surface and the biomolecule (see Subheading 3.3 and 3.4). Reactive groups on the surface of the Ta2O5 waveguide can be generated for example by functionalization with an aminosilane layer in order to expose amino groups on the surface (Subheading 3.2.). Another method presents the affine coupling via the biotin-streptavidin system. Biotinlabeled biomolecules react under the formation of a very stable non-covalent interaction with streptavidin, avidin or neutravidin. The coupling procedure for different biomolecules always has to be optimized to establish a functional surface with the intact ligand in an epitope presenting orientation. A good collection of feasible coupling reactions and protocols can be found in (16–18). Two simple but effective covalent coupling procedures with a good surface passivation for measurements with complex solutions are described here in detail. More complex strategies for surface passivation have been described and evaluated in (5).

1.2.1. PDITC Coupling

1,4-phenylene diisothiocyanate (PDITC), a homobifunctional crosslinker, can be used to couple a biomolecule covalently to the sensor surface in a two-step process (18). In the first step, one of the two isothiocyanate groups of the crosslinker reacts with an amine group at the sensor surface under the formation of a thiourea linkage. The para-position of the isothiocyanate groups prevents the reaction with another amine group on the surface. In the second step, the free isothiocyanate group reacts with amine groups of the protein. This coupling chemistry results in a very dense surface coverage with biomolecules and, hence, minimizes unspecific binding to the surface. Consequently, the PDITC coupling procedure is ideal for measurements on complex solutions like blood serum (6). A typical sensorgram of a measurement on undiluted clinical samples (blood serum) for the direct detection of a tuberculosis specific antibody is shown in Fig. 5. In this case the label-free measurement with the grating coupler shows a similar sensitivity and specificity like an established ELISA test (6).

1.2.2. EDC/Sulfo-NHS Coupling

The covalent linkage between a carboxylate and an amine can be achieved with the reagents N-Ethyl-N´-(3-dimethylaminopropyl)

194

Nagel, Ehrentreich-Förster, and Bier 800 B 700 B

∆Neff (*10−6)

600 500 400 R 300 200 R 100 1 0

2

B

0 2000

4000

6000

B 8000

time [s]

Fig. 5. Real-time measurement of clinical blood serum samples with the grating coupler. (1) tuberculosis negative blood serum; (2) tuberculosis positive blood serum; (R) regeneration (50 mM HCl); (B) PBS buffer.

carbodiimide hydrochloride (EDC) and N-hydroxysulfo succinimide (sulfo-NHS). EDC reacts with a carboxylate group to form an active ester group (O-acylisourea), but this intermediate is very unstable in aqueous solution. Sulfo-NHS increases the stability of the intermediate and rapidly reacts with amines forming a stable amid linkage (16). Beside the coupling of the carboxyl groups to the amines on the surface, self-polymerization occurs if amine groups are also present in the biomolecule. This self-polymerization yields in a dense surface coverage of the biomolecule.

2. Materials 2.1. Sensor Chip Modification 2.1.1. Cleaning and Activation of the Sensor Chips

1. Distilled H2O. 2. Ethanol 99%, denatured. 3. Piranha cleaning solution: hydrogen peroxide (30%) and sulfuric acid (95–97%) in a mixing ratio of 1:3. Always add the peroxide to the acid. An exothermic reaction occurs and the solution heats up above 100 °C. Mixing piranha solution with organic compounds may cause an explosion. Handle with care. Wear safety glasses and nitrile gloves and always work inside a fumehood. 4. Sodium hydroxide (5 M). Handle with care. Wear safety glasses and nitrile gloves and always work inside a fumehood.

Label-Free Serodiagnosis on a Grating Coupler

195

2.1.2. Silanization of the Surface

1. Aminopropyltriethoxysilane (APTES) solution: 10% (v/v) APTES in distilled H2O, adjust the pH with hydrochloric acid precisely to 3.45.

2.1.3. PDITC Coupling

1. Dimethylformamid, dry (DMF). 2. Phenylene diisothiocyanate, purum (PDITC): 300 mM in DMF. Handle with care. Wear safety glasses and nitrile gloves and always work inside a fumehood. 3. Phosphate-buffered saline (PBS): 137 mM sodium chloride, 10 mM di-sodium hydrogen phosphate, 2.7 mM potassium chloride, 2 mM potassium di-hydrogen phosphate, pH 7.4. 4. TRIS-HCL buffer: 1 M tris(hydroxymethyl)aminomethane, pH 7.4.

2.1.4. EDC/Sulfo-NHS Coupling

1. N-Ethyl-N´-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC). Protect from moisture. Store desiccated at −20 °C. 2. N-hydroxysulfosuccinimide (sulfo-NHS). 3. MES buffer: 50 mM 2-(N-morpholino)ethanesulfonic acid, pH 6. 4. Distilled H2O with 0.1% (v/v) Tween 20.

2.2. Measurement

1. Phosphate-buffered saline (PBS): 137 mM sodium chloride, 10 mM di-sodium hydrogen phosphate, 2.7 mM potassium chloride, 2 mM potassium di-hydrogen phosphate, pH 7.4. 2. Regeneration solution: 50 mM hydrochloric acid. 3. Human serum samples.

2.3. Instrumentation 2.3.1. Grating Coupler

The optical arrangement of the grating coupler is shown in Fig. 1 and Fig. 2. 1. Laser: 10 mW HeNe laser, λ = 633 nm, # 1137P (Uniphase, Manteca, CA, USA). 2. Laser power supply: #1202–2 (Uniphase, Manteca, CA, USA). 3. Neutral density filter: T = 10%, #063462 (Linos, Göttingen, Germany). 4. Cylinder lense: f = 40 mm, #063422 (Linos, Göttingen, Germany). 5. Chip holder: a. Wedge plate: #334482 (Linos, Göttingen, Germany). b. Positioning system consists of a X-Y stage, a Z-axis stage and a rotary stage (Linos, Göttingen, Germany). c. Grating coupler chips: Grating period = 0.75 µm, Ta2O5 waveguide, #ASI 3200 (Artificial Sensing Instruments, Zürich, Switzerland).

196

Nagel, Ehrentreich-Förster, and Bier

6. Mirror: 1/10 wave accuracy, Ø = 31,5 mm, #340008400 (Linos, Göttingen, Germany). 7. CCD line sensor: 1728 Pixels, #TH 7803ACC (Thomson, Saint-Egreve, France). 2.3.2. Flow System

1. Flow cell: PVC, Volume ∼ 9 µL (Fraunhofer IBMT, Potsdam, Germany). 2. Tubing: Teflon tube, Øi = 0.5 mm (ERC, Riemerling, Germany). 3. Peristaltic pump: Perimax 12 (Spetec, Erding, Germany).

2.3.3. Data Evaluation

1. Standard PC with a CCD sensor specific converter card (Thomson, Saint-Egreve, France). 2. The Linux based evaluation software (Fraunhofer IBMT, Potsdam, Germany) determined the shift of the minimum by the evaluation of the left slope of the measured signal. Fig. 4 shows a typical signal under coupling conditions. After the calculation of the inflection point of the slope, the minimum position of the inflection is determined and the shift of this position is recorded (5 Hz).

3. Methods 3.1. Cleaning and Activation of the Sensor Chips

All steps should be performed under gentle agitation. 1. Clean the chips with distilled H2O – ethanol – distilled H2O two times (each step 2 min). 2. Incubate in piranha cleaning solution for 5 min (see Note 1). 3. Wash intensively with distilled H2O. 4. Dip in 5 M sodium hydroxide for maximal 30 s (see Note 2). 5. Wash intensively with distilled H2O, avoid air contact and go on to the next step (see Note 3).

3.2. Silanization of the Surface

1. Incubate the chips in APTES-solution for 2 h at 80 °C (see Note 4). 2. Wash intensively with distilled H2O. 3. Dry at 110 °C for 1 h. 4. Store them dry and under vacuum.

3.3. PDITC Coupling

1. Dry the chips at 110 °C for 10 min. 2. Prepare a glass petri dish with a filter paper at the bottom. 3. Rinse the cooled chips with DMF and dry them in a nitrogen stream.

Label-Free Serodiagnosis on a Grating Coupler

197

4. Put the chips in a petri dish and pipette 150 µL PDITC solution onto the grating. Replace the air in the petri dish with a nitrogen stream and seal it with Parafilm. Incubation time: 2 h (see Note 5). 5. Wash the chips with DMF and dry them in a nitrogen stream. 6. Pipette the protein solution (in PBS) at a concentration in the range of 1 mg/mL onto the PDITC activated area (see Note 6). 7. Incubate overnight in a humidity chamber (sealed petri dish with 1 mL PBS buffer). 8. Wash with distilled H2O. 9. Deactivate the remaining isothiocyanate groups with TRISHCL buffer buffer for 2 h. 10. Wash with distilled H2O. 3.4. EDC/Sulfo-NHS Coupling

1. Mix the protein solution with MES buffer to a final concentration in the range of 0.1–1 mg/mL (see Note 6). 2. Add EDC (0.2 M) and sulfo-NHS (0.05 M) to the protein solution. 3. Pipette this mixture onto the sensor chip. 4. Incubate it for 3 h in an humidity chamber (sealed petri dish with 1 mL MES buffer). 5. Wash with distilled H2O containing 0.1% Tween 20. 6. Wash with distilled H2O.

3.5. Measurement

Avoid temperature variations and air bubbles in the flow system. Use always well degassed running buffer. The sample flow should be from the bottom to the top of the flow cell to ensure a complete filling with the sample (see Notes 7 and 8).

3.5.1. Chip Mounting and Basic Settings

1. Position the chip on the wedge plate (with a little droplet of refractive index oil between glass substrate and wedge plate) and fix it with the flow cell (see Fig. 2). 2. Fill the flow system with PBS buffer. 3. Irradiate the laser light on the grating and adjust the angle of the incident light to a good incoupling position (clear effacement in the reflected light) and guide the reflected light to the CCD sensor. The measured signal should be similar to Fig. 4. 4. Flush the system with PBS buffer until a stable baseline is achieved (see Note 9).

3.5.2. Serum Sample Measurement and Regeneration

See Fig. 5 for an exemplary serodiagnosis measurement for the detection of tuberculosis specific antibodies in blood serum.

198

Nagel, Ehrentreich-Förster, and Bier

1. Adjust flow rate to 25 µL/min (see Note 10). 2. Inject the sample into the flow system. To increase the contact time to the surface, stop pumping and incubate for some minutes. Serum samples can be used both undiluted and diluted (see Notes 11–14). 3. Start flow again and change back to PBS buffer and observe the amount of binding to the surface. 4. Regenerate the surface with 50 mM sodium hydroxide (1 min) (see Note 15). 5. Flush the system with PBS buffer and inject the next sample when the base line is stable. 6. Determine the changes of the refractive index (∆Neff) caused by the different samples and evaluate these values (see Note 16).

4. Notes 1. Use only freshly prepared piranha solution. The chips should be hydrophilic after the treatment. 2. Do not incubate the chips longer than 30 s in the 5 M sodium hydroxide solution, otherwise the Ta2O5 waveguide will be corroded and destroyed. 3. To bridge a short period of time (in the range of minutes), transfer the chips into a container with water to avoid air contact. 4. Use only freshly prepared APTES solution with a well adjusted pH-value. Cover the reaction chamber to avoid evaporation. 5. The PDITC solution should not dry up during the two hours incubation time. Increase the volume if necessary. 6. Measurements with complex solutions like blood serum with optical label-free devices are only possible if the unspecific attachment to the surface can be minimized. In the literature, different approaches dealing with this topic are described (5, 6). High surface coverage with the ligand is the crucial point; either by the formation of multilayered constructions or just by coupling high concentrations of the ligand. Furthermore, the appropriate dilution of the sample or the addition of detergents and high salt concentrations in the running buffer can significantly minimize the unspecific attachment. 7. Perform the measurements always with well-tempered samples and buffers. Keep in mind that a change of one degree Celsius results in a refractive index change of 1 × 10−4.

Label-Free Serodiagnosis on a Grating Coupler

199

8. Air bubbles in the flow system have a negative influence on the measurement if they get into the measurement/flow cell. Degas the buffers in a vacuum chamber (pressure lower than 10 mbar) under gentle agitation. A set-up which sucks the buffer through the flow system is more susceptible for air bubble contamination than a set-up which applies positive pressure. A strong influence is also given by the flow rate and by the addition of detergents to the buffer. 9. Given that the measured effect is temperature dependent, use only solutions at room temperature. If the baseline is not stable, try to wash unbound material off with regeneration solution. 10. A flow rate between 20–100 µL/min is typical. Use only pumps with linear flow. For kinetic measurements it is essential to ensure that the reaction is mass transport-limited. 11. Always use standard solutions (e.g. monoclonal antibodies specific and unspecific to the coupled biomolecule) to check the success of the coupling procedure and the specificity of the binding. 12. Measurements of solutions with a large difference in their refractive index (e.g. buffer and serum) result in a huge signal shift (index jump). The determination of the amount of bound material on the surface is only possible after changing back to the running buffer. 13. Due to the fact that there is only one recognition area in the flow cell, it is necessary to check for unspecific binding to distinguish between a specific and an unspecific binding. The essential test is done with an unspecific antibody and the obtained signal is compared to the specific antibody binding. Furthermore, the amount of unspecific attachment of the sample fluid (e.g. blood serum, milk, etc.) is an important information about the passivation effect of the manufactured surface. In addition to that, the efficiency of the coupling procedure for each chip should be checked as quality control, using a standard solution of a specific monoclonal antibody. 14. Blood products are potentially extremely hazardous and all steps should be performed under adequate protection. In many countries authorizations are needed and specialized knowledge is a prerequisite for working with blood products. Blood serum should be stored in aliquots at −80 °C. Avoid repeated freezing and thawing. 15. To reuse the sensor chip, it is essential to empirically find the mildest regeneration condition to remove the analyte without impairment of the coupled ligand. A wide range of different regeneration solutions can be tested. Regeneration

200

Nagel, Ehrentreich-Förster, and Bier

with low pH solutions (e.g. 10–100 mM hydrochloric acid) is successful for protein surface regeneration. But also solutions with high pH values (10–100 mM sodium hydroxide), high ionic strength or with low concentrations of sodium dodecyl sulfate are potential regeneration solutions and have to be optimized for each assay. 16. The determined ∆Neff values of known tuberculosis negative serum samples are used to define a cut-off value (sum of the mean signal intensities plus twice the standard deviation). Consequently, samples with higher signals are positive while samples with smaller signals are negative.

References 1. Gauglitz, G. (2005) Direct optical sensors: principles and selected applications. Anal Bioanal Chem. 381, 141–155 2. Tiefenthaler, K. and Lukosz, W. (1989) Sensitivity of grating couplers as integratedoptical chemical sensors. J Opt Soc Am B. 6, 209–220 3. Comley, J. (2005) Label-free detection – new biosensors facilitate broader range of drug discovery applications. Drug Discovery World, Winter 2004/5, 63–74 4. Cooper, M. A. (2006) Current biosensor technologies in drug discovery. Drug Discovery World, Summer 2006, 68–82 5. Brynda, E., Houska, M., Brandenburg, A. and Wikerstal, A. (2002) Optical biosensors for real-time measurement of analytes in blood plasma. Biosens Bioelectron. 17, 665–675 6. Nagel, T., Ehrentreich-Förster, E., Singh, M., Schmitt, K., Brandenburg, A., Berka, A., Bier, F. F. (2008) Direct detection of tuberculosis infection in blood serum using three optical label-free approaches. Sens Actuators B: Chem. 129, 934–940 7. Ehrentreich-Förster, E., Scheller, F. W. and Bier, F. F. (2003) Detection of progesterone in whole blood samples. Biosens Bioelectron. 18, 375–380 8. Nellen, P. M. and Lukosz, W. (1990) Integrated optical input grating couplers as chemo- and immunosensors. Sens Actuators B. 1, 592–596 9. Nellen, P. M. and Lukosz, W. (1991) Model experiments with integrated optical input grating couplers as direct immunosensors. Biosens Bioelectron. 6, 517–525

10. Lukosz, W., Nellen, P. M., Stamm, C. and Weiss, P. (1990) Output grating couplers on planar waveguides as integrated optical chemical sensors. Sens Actuators B. 1, 585–588 11. Lukosz, W., Clerc, D., Nellen, P. M., Stamm, C. and Weiss, P. (1991) Output grating couplers on planar optical waveguides as direct immunosensors. Biosens Bioelectron. 6, 227–232 12. Clerc, D. and Lukosz, W. (1994) Integrated optical output grating coupler as biochemical sensor. Sens Actuators B. 19, 581–586 13. Brandenburg, A., Polzius, R., Bier, F. F., Bilitewski, U. and Wagner, E. (1996) Direct observation of affinity reactions by reflectedmode operation of integrated optical grating coupler. Sens Actuators B. 30, 55–59 14. Billitewski, U., Bier, F. F. and Brandenburg, A. (1998) Immunobiosensors based on grating couplers, in (Rogers, K. R. and Mulchandani, A., ed.) Methods in Biotechnology Vol. 7, Humana Press, Totowa, NJ, pp. 121–134 15. Clerc, D. and Lukosz, W. (1997) Direct immunosensing with an integrated-optical output grating coupler. Sens Actuators B. 40, 53–58 16. Hermanson, G. T. (ed.) (1996) Bioconjugate Techniques. Academic Press, San Diego, CA 17. Aslan, M. (ed.) (1998) Bioconjugation: Protein Coupling Techniques for the Biomedical Sciences. Stockton Press, London 18. Heise, C. and Bier, F. F. (2005) Immobilization of DNA on microarrays, in (Wittmann, C., ed.) Topics in Current Chemistry: Immobilisation of DNA on Chips II, Springer, Berlin Heidelberg, pp. 1–25

Chapter 11 CCD Camera Detection of HIV Infection1 John R. Day Summary Rapid and precise quantification of the infectivity of HIV is important for molecular virologic studies, as well as for measuring the activities of antiviral drugs and neutralizing antibodies. An indicator cell line, a CCD camera, and image-analysis software were used to quantify HIV infectivity. The cells of the P4R5 line, which express the receptors for HIV infection as well as b-galactosidase under the control of the HIV-1 long terminal repeat, were infected with HIV and then stained 2 days later with X-gal to turn the infected cells blue. Digital images of monolayers of the infected cells were captured using a high resolution CCD video camera and a macro video zoom lens. A software program was developed to process the images and to count the blue-stained foci of infection. The described method allows for the rapid quantification of the infected cells over a wide range of viral inocula with reproducibility, accuracy and at relatively low cost. Key words: Charge-coupled device, CCD; Imaging, HIV, Infectivity, Quantify, Rapid.

1. Introduction The quantification of a biological process can sometimes be cumbersome or labor-intensive. In the field of virology, several assays have been developed to measure the infectivity of the human immunodeficiency virus (HIV) during a single round of replication:

1 Portions reprinted from the Journal of Virological Methods, Vol. 137, J.R. Day, L.E. Martínez, R. Šášik, D.L. Hitchin, M.E. Dueck, D.D. Richman and J.C. Guatelli, A computer-based, image-analysis method to quantify HIV-1 infection in a single-cycle infectious center assay, Pages 125–133, Copyright 2006, with permission from Elsevier.

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_11

203

204

Day

plaque assays (1), syncytium formation assays (2, 3), fluorescent and colorimetric focal immunoassays (4) and assays using indicator cell lines (5–10). The quantification of infection in these assays requires counting the number of plaques, syncytia, foci, or cells within an infected population of cells. The use of indicator cell lines, such as those containing an integrated b-galactosidase gene under the control of the HIV long terminal repeat (LTR), has enabled the straightforward determination of viral infectivity by visualizing infected cells that are stained blue with 5-bromo-4chloro-3-indolyl-b-D-galactopyranoside (X-gal). The Magi-CCR5 (6) and P4R5 (7) indicator cell lines are frequently used as targets of infection. These are HeLa-based adherent cells that have been engineered to express CD4, the HIV receptor, as well as CCR5, one of the two predominantly used coreceptors; HeLa cells naturally express CXCR4, the other major coreceptor used by HIV. To quantify the number of infected cells by using these b-galactosidase-based indicator cell lines, one approach is to manually count blue cells by eye through a microscope. This laborious method is subject to observer error and is difficult when the density of infected cells is high, potentially limiting the dynamic range of the assay. To facilitate the counting of infected cells with speed, accuracy and optimal dynamic range, a chargecoupled device (CCD) camera was used and software was developed to analyze digital images of infected cell monolayers (11). Accurate and reproducible cell counts over a wide range of viral inocula were obtained with this method. CCD camera technology has been available for years, but the recent explosion in the use of digital cameras for personal use has accelerated technology development and facilitated the incorporation of affordable CCD detection in the laboratory.

2. Materials 2.1. Imaging Apparatus

1. 5 megapixel CCD color camera with real-time viewing, C-mount optical lens interface, and FireWire IEEE 1394 digital interface: MicroPublisher 5.0 RTV (Model # MP5.0-RTV-CLR-10, QImaging, Burnaby, BC, Canada) (see Note 1). 2. Macro video zoom lens: Optem 18–108 mm, f/2.5, C-mount (Qioptiq Imaging Solutions, Rochester, NY, formerly ThalesOptem, Inc.) (see Note 2). 3. Copy stand with a 1/4″-20 threaded mounting screw (model # CS-3, Testrite Instrument Company, Inc.). Something similar would suffice.

CCD Camera Detection of HIV Infection

205

4. Fluorescent light box (typically used for viewing X-ray films) (see Note 3). 5. Camera mount adapter, custom made by our university machine shop, consisting of a tripod mounting screw (1/4″-20 thread) and a rectangular piece of aluminum (2.25 × 1 × 0.25 in.) with two drilled holes (see Note 4). Use of the adapter requires that the copy stand have a mounting rod that can extend away from the stand. 6. Personal computer for capturing and analyzing images: Windows 2000 or XP operating system, 1 available FireWire port (IEEE 1394) or a FireWire expansion card. 2.2. Image Acquisition and Analysis Software

1. Image acquisition software, QCapture Pro, v.5.0.0.16, with USB key (dongle) and drivers for PC (QImaging, Burnaby, BC, Canada). The Micropublisher 5.0 camera was bundled with QCapture Suite software, however, the upgrade to QCapture Pro was purchased to increase control over the image capture process. 2. Image analysis software: application (5MGL.exe) and parameter file (5MGL.par), collectively termed the “Romanizer.” The software was developed in-house in the Fortran 95 programming language for the specific purpose of counting HIV-infected cells. Use of the software for other purposes may not be suitable. Alternative image-analysis applications exist, both freeware and for purchase (see Note 5).

2.3. HIV Infectivity Assay

1. Indicator cells: P4R5 HeLa cells (P4.R5 MAGI cells, AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH contributed by Dr. Nathaniel Landau). The P4R5 cell line contains the b-galactosidase gene under the control of the HIV-1 promoter region (LTR). The derivation of the parental P4 line has been reported (7). P4R5 cells express on their surface the HIV entry receptors, CD4, CXCR4 and CCR5. 2. Complete media: Dulbecco’s Modified Eagle Medium (DMEM; Gibco, Grand Island, NY) supplemented with 10% fetal bovine serum (FBS; GemCell, Woodland, CA), 100 U/ mL penicillin, 100 µg/mL streptomycin (pen/strep; Gibco, Grand Island, NY), 2 mM L-glutamine (Gibco, Grand Island, NY) and 1 µg/mL puromycin. Store at 4°C. 3. 48-well tissue culture plate (Cat. # 353078, BD Falcon, San Jose, CA). 4. Phosphate buffered saline, 1× (PBS; Invitrogen, Carlsbad, CA). 5. Fix solution: 1% formaldehyde, 0.2% glutaraldehyde in PBS. Can be made in advance and stored in the dark at 4°C for 1–2 months.

206

Day

6. X-gal stain solution made fresh: For each 1.0 mL of solution, combine 949 µL PBS, 20 µL 0.2 M potassium ferrocyanide, 20 µL 0.2 M potassium ferricyanide, 1.0 µL 2.0 M Mg2CL, and 10 µL 40 mg/mL X-gal (5-bromo-4-chloro-3-indolyl-b-Dgalactopyranoside). Stock solutions of potassium ferrocyanide, potassium ferricyanide, Mg2Cl, and X-gal can be prepared and stored in aliquots at −20°C. X-gal is dissolved in dimethyl sulfoxide (DMSO) and should be stored in the dark. The X-gal solution will turn yellow over time, but this does not affect the activity. Discard the stock if it becomes green/brown. 7. The HIV virus is either isolated and cultured from blood, or produced in vitro by transient transfection of proviral plasmid DNA into cultured cells. Viruses do not need a complete HIV genome, but they must be competent for entry into P4R5 cells and they must be capable of expressing the HIV Tat protein. Tat transactivates the HIV LTR promoter region and will enable expression of b-galactosidase in the P4R5 indicator cells.

3. Methods In order to facilitate the counting of blue-stained HIV-infected cells, a CCD camera, macro lens, and a computer program were set up to capture and process images of plated cells. The setup does not require the exact equipment listed in Subheading 2. A camera, lens, light source, computer, and a method to mount the camera are the basic required elements. The ideas described here can be adapted to different configurations to suit the needs of the application. The method for analyzing the images will vary depending on the final goal of the analysis. Although the software we developed to count HIV-infected cells may not be suitable for other applications, alternative software programs may provide the necessary tools for your particular application. 3.1. Setup of Imaging Apparatus

1. Place the copy stand on a table and the light box (or light source of choice) on the base of the stand. 2. Assemble the camera by screwing the lens onto the camera. 3. Mount the camera to the stand using the 1/4″-20 threaded mount. The most straightforward way to mount the camera is illustrated in Fig. 1a. An alternative method to mount the camera requires a custom-built adapter (see Note 6). The alternative method is illustrated in Fig. 1b and a photograph of the setup is shown in Fig. 2. Screw the mount adapter onto the copy stand, then attach the camera to the adapter using the tripod mount screw (Fig. 3).

CCD Camera Detection of HIV Infection

207

Fig. 1. Schematic diagrams of the CCD camera setup. (a) The simplest setup using direct attachment of the camera to the copy stand. The FireWire ports on the camera are on the backside in the perspective shown. (b) Use of an aluminum mount adapter and tripod screw to mount the camera rotated at 180° (see Notes 4 and 6).

Fig. 2. Photograph of the actual camera setup showing the light box, copy stand, CCD camera, macro lens and computer.

208

Day

Fig. 3. Close-up photographs of the camera mounted to the copy stand using the mount adapter.

4. The height of the camera above the light box is adjustable and depends on the zoom setting of the lens. In our setup, the distance between the surface of the light box and the mounting screw hole of the camera was 14.25 in. (36.2 cm). The tip of the lens was 6.25 in. (15.9 cm) above the light box. Level the light box and camera using a standard bubble level. 5. Attach a FireWire cable to either one of the ports on the camera. Do not connect the cable to the computer until after loading the software and drivers. 3.2. Capture and Analysis of Images 3.2.1. Capturing Images

1. These instructions assume the use of the Micropublisher 5.0 RTV camera, Optem 18–108 mm macro video zoom lens, and QCapture Pro software. Other cameras, capture software, and analysis software may be used. Please follow the directions included with your specific components. 2. Follow the setup instructions from QImaging to load the camera drivers and QCapture Pro software. You will connect the FireWire cable to the computer during this process. 3. To define the capture settings, launch QCapture Pro and click the camera icon on the toolbar (or select Acquire/Video-Digital from the menu) to open the acquisition dialog box. Use the Basic Dialog (rather than the Advanced Dialog) by clicking the “Basic Dialog” button. If the button in the lower left says “Advanced” then you are already looking at the Basic Dialog. 4. Check the settings by clicking the “More >>” button at the bottom of the window. This expands the window. Using version 5.0.0.16 of QCapture Pro and with the specific light source we used, the settings were as follows (Fig. 4) (see Note 7 for possible variations): ● Exp Acq: 00.300.00, Adjust Exp for Binning (checked)

CCD Camera Detection of HIV Infection

209

Fig. 4. Screen shot of the capture settings within QCapture Pro capture software.

●

●

Binning: Pvw (2 × 2 for focusing, 4 × 4 for moving the plate); Acq (1 × 1) Capture area dimensions: Left (320), Top (0), Right (2239), Bottom (1919) [final resolution, 1920 × 1920 pixels

●

Capture depth: 24-bit color

●

Gain: 1100; Gamma: 1.90; Offset: 1120

●

White balance: R (2.327), G (1.0), B (1.375).

5. Click “Less <<” to close the settings side of the window, then click “Preview” to see a live image. 6. Turn on the light source and place a 48-well plate (or object to be imaged) under the camera. 7. Adjust the aperture (top ring on lens) to change the amount of light passing through the lens. The amount of this adjustment will depend on the desired result. We adjusted the aperture so that a slightly gray image appeared in the preview window (see Note 8).

210

Day

8. Adjust zoom (middle ring) and focus (bottom collar) to fill the preview window with a single well of the 48-well plate. An initial height adjustment of the camera above the light box may be necessary as well, to achieve the desired result. Use 2 × 2 Preview (Pvw) binning to focus, then switch to 4 × 4 for moving the plate. As long as Acq binning remains at 1 × 1 the full resolution image will be acquired. The binning status of Pvw does not affect the acquired image, but only the image seen in the preview window. Binning at 4 × 4 allows for faster on-screen response of plate movement (see Note 9). 9. Capture images by clicking the “Snap” button. Move the 48-well plate to view the next well, click “Snap,” and repeat for the entire plate. 10. After acquiring all the images, close the live preview window by clicking the “Stop” button, then the “X” in the upper right corner of the acquisition window. 11. Save the images to a new folder. A quick way to save all the images is to select the menu Window/Close All, then click “Yes All.” When prompted, type a name for each file. Save the files as TIFF for analysis by the Romanizer software. 3.2.2. Analyzing Images

1. The Romanizer program will automatically count bluestained HIV-infected cells in 1920 × 1920 pixel, 24-bit color images. Note that in addition to counting blue-stained cells, the software may also count overly-clumped cells, dust, or fibers that appear dark in the image (see Note 10). 2. Open the file folder where the Romanizer program is stored. Copy 5MGL.exe and 5MGL.PAR into the folder where you saved your images. The program must reside in the same folder as your TIFF images, and it will analyze all images within that folder. 3. Double-click 5MGL.exe to launch the counting program. You should see one of your images with a red circle superimposed upon it (Fig. 5). The circle defines the region to be analyzed. Using the computer mouse, click and drag the circle to the center of your well to check its size. The circle should be a little smaller than the size of the well to prevent analyzing the edge of the well. If you need to adjust the size follow these steps: (a) Close the program by clicking Exit on the menu. (b) Open the file 5MGL.PAR file (see Note 11) using a simple text editor such as the Notepad. (c) Increase or decrease the number next to “circle radius in tiff pixels” depending on the desired change in the circle radius. (d) Save your changes and close the file.

CCD Camera Detection of HIV Infection

211

(e) Launch 5MGL.exe again to see the result of the size change. (f) Repeat steps a–e if needed to find the appropriate size. 4. Position the circle in the center of the well, then click “Next.” The next image in the folder will appear. Repeat the procedure for the remaining images. Due to variation in the placement of the well by hand within the frame of the digital photograph, the location of the analysis circle needs to be defined separately for each image analyzed. 5. When all circles have been defined, the program will process the images and start counting. Press “OK” when it has finished. Fig. 6a shows a digital photograph taken with the Micropublisher 5.0. The resulting image after being processed by the software is shown in Fig. 6b. 6. After background subtraction and processing, the images are placed in the file folder with an “f” preceding each filename. The results of the analysis are exported to a Microsoft Excel file called “results.xls.” “Simple count” is a raw count of spots found. “Smart count” takes into consideration that some spots may in fact be several adjacent cells, as it counts the adjacent cells separately. In most cases, “Smart count” is the preferred result to use (see Note 12).

Fig. 5. Screen shot of the Romanizer program. The red circle indicates the area that will be analyzed by the program. The position of the analysis circle was defined for each image because of potential variations in the placement of the well within the digital image.

212

Day

Fig. 6. Image-analysis of HIV-infected P4R5 cells. (a) Digital photograph of a monolayer of P4R5 cells infected with HIV and stained with X-gal within one well of a 48-well tissue culture plate using a MicroPublisher 5.0 camera and macro video zoom lens. (b) The same well after processing by the analysis software. The boxed region is magnified in Fig. 7.

3.3. HIV Infectivity Assay

1. These instructions describe an assay to quantify the infectivity of live HIV virus. Sterile cell culture technique and safe laboratory practices for handling live HIV are beyond the scope of this description. We performed these assays in a Biosafety Level 3 (BL3) facility approved by the Environmental Health and Safety department of our institution. 2. The P4R5 cell line is an adherent cell line derived from HeLa cells. The cells were engineered to express b-galactosidase after infection with HIV. Maintain them according to standard cell culture procedures for adherent cell lines. 3. Plate cells at a density of 2 × 104 cells per well in a 48-well tissue culture plate and place them in a humidified, 37°C, 5% CO2 incubator overnight. 4. The next day, carefully aspirate the media, avoiding contact with the cell monolayer. Infect the cells in duplicate wells with 100 µL of serial dilutions of virus in complete media. After a 2 h incubation at 37°C, add 400 µL of complete media. Incubate for 2 days at 37°C. 5. Carefully aspirate the media from the wells and add 1 mL of fix solution to each well. Incubate for 5 min at room temperature. The fix solution renders the virus non-infectious. 6. Aspirate the fix solution and wash the wells twice with PBS. 7. Add 250 µL of X-gal stain solution and incubate overnight at 37°C in a non-CO2 incubator. 8. Aspirate the stain solution and wash once with PBS. Wash again with deionized water. Aspirate the water from the well.

CCD Camera Detection of HIV Infection

213

After wiping down the outside of the plate with 70% ethanol, the plate can be removed from the BL3 and handled at the lab bench. 9. Invert the plate over a paper towel and allow the wells to dry. 10. The plate is now ready for imaging and analysis. Fig. 7 illustrates the final result in a magnified region of Fig. 6. Syncytia (fused neighboring HIV-infected cells) of various sizes and single cells are accurately distinguished within the monolayer of P4R5 cells. The magnified image illustrates the amount of detail seen in the original digital photograph (Fig. 7a), and for comparison, the same region was photographed using the 20× objective of a Leitz Labovert microscope (Fig. 7b). In the processed image, every object that is counted is marked with a small red dot to allow the user to examine more closely how the software is performing and exactly what objects in the image were counted (Fig. 7c). To illustrate this more clearly, the outlines of infected foci are shown encircling each red dot (Fig. 7d). Two or more red dots within a group of cells that come in contact in the image are counted as one infection, using “Simple” count and as multiple foci using “Smart” count (see Note 12).

Fig. 7. Magnified view of HIV-infected P4R5 cells. Magnified images of the boxed region of Fig. 6 depicting (a) the original digital photograph, (b) the same region viewed through the 20× objective of a microscope, (c) the processed image including small red dots as indicators of the objects counted and (d) outlines of infected cells surrounding red dot markers. The arrows indicate cell groups counted singly using Simple count and as two cells using Smart count.

214

Day

4. Notes 1. With today’s advances in digital camera technology, there are many types of affordable options. The first parameter to consider when choosing a camera is the spatial resolution that the camera provides. The resolution corresponds to the ability to discriminate between points that are closely spaced in the image (12). We initially used an analog CCD camera with a standard VGA resolution of 640 × 480 pixels (approximately 0.3 megapixels). The camera was able to resolve the larger blue-stained foci, but smaller ones remained undetected. Furthermore, the analysis software was not able to distinguish between small foci in close proximity. The resolution of the camera chosen depends largely on the object to be imaged. Smaller objects require larger arrays of pixels (higher resolution) so that they can be discriminated from one another. We upgraded to a 5 megapixel digital CCD camera (2560 × 1920 pixels) when they became more readily available and reasonably-priced. In addition to resolution, a second consideration in choosing a camera is the sensitivity requirement. The sensitivity of a camera refers to the lowest signal that can be detected (13). If the object to be imaged is fluorescent or chemiluminescent, a camera with increased sensitivity may be required. For example, QImaging makes a camera that will detect a single photon of light (Rolera-MGi); however the resolution of the camera is only 512 × 512. An increase in camera sensitivity is often accomplished at the expense of resolution. Higher sensitivity applications can also be subjected to an increase in noise; therefore, the signal-to-noise ratio becomes important, as well. Because heat generated in a camera can cause an increase in noise (dark noise), cooled cameras, such as the cooled version of the Micropublisher (MP5.0-RTV-CLR-10-C), can maintain a lower temperature and reduce the amount of noise in the system (14). Choosing the right camera for your application requires careful consideration of these factors. Consult camera manufacturers for more information and advice. 2. Like the choice of camera, the choice of lens depends on the specific object to be imaged. Macro lenses are designed for close-up photography and are capable of focusing on objects within a short distance. Using a lens with zoom capability enables greater control of the size of the object in the frame. The Optem lens we used was recommended to us and it met our needs well. 3. The choice of a light source also depends on what is to be imaged. Illumination from above or from below depends on

CCD Camera Detection of HIV Infection

215

the transparency of the subject. For our purposes, a simple fluorescent light box provided ample light to illuminate the wells of a 48-well plate from below. The light source does not necessarily need to emit a specific intensity of light because there are camera (aperture) and software (exposure) settings that can be adjusted to change the brightness of the resulting image. The color spectra emitted by the light source is also not a large factor because white balance settings can be adjusted within the capture software. A commonly used light box for viewing X-ray films was more than adequate. 4. A tripod mounting screw and a rectangular piece of aluminum were the materials used for the mount adapter (Fig. 8). The dimensions of the aluminum rectangle were 2.25 × 1 × 0.25 in. (5.7 × 2.5 × 0.6 cm). Two holes were made in the aluminum, one large enough for the tripod mounting screw to pass through easily, and the other threaded so that it would screw onto the copy stand mount (1/4″-20 thread) (Fig. 8). The holes were 1.25 in. (3.2 cm) apart. The adapter works only if the copy stand mounting rod can extend far enough and away from the stand to leave at least 3.2 in. (8.1 cm) of space for the camera body (Fig. 3). This custom-built adapter allowed the camera to be rotated 180° and mounted on the copy stand. 5. The Romanizer program is freely available for use in the type of analysis described here. Its use for counting other kinds of objects of different sizes may not produce acceptable results. Processed images should be scrutinized carefully to determine the program’s performance. You may download a copy of the software and setup instructions here: http://cfar.ucsd.edu/ romanizer/. QCapture Pro performs some basic analyses and may be one alternative. Another freely available program that

Fig. 8. Photographs of the mount adapter. One hole was drilled wide enough for the tripod mounting screw to pass through easily. The second hole was drilled with a 1/4˝-20 thread for mounting onto the copy stand.

216

Day

performs image analysis is the NIH Image/ImageJ program developed by the National Institutes of Health (15). 6. When the camera was mounted directly on the camera stand, the image on-screen was inverted, i.e., moving a microplate to the left caused the on-screen live preview image to move to the right, and moving the plate up caused the preview to move down. One can become accustomed to this type of motion, but it was not ideal. The capture software did not have the option to reverse the orientation of the image, but future versions may. Our solution to the problem was to rotate the camera 180°. Unfortunately there was no threaded mounting hole on the opposite side of the camera. Instead, we designed a mount adapter and asked our university machine shop to construct it (Fig. 8). This custom-built adapter allowed the camera to be rotated and mounted on the copy stand. 7. For consistency between experiments, the same settings were used each time. Some changes have been made in newer versions of QCapture Pro such that the parameters may not be exactly the same as shown here. We used the default settings for gain, gamma and offset, and since version 5.0.0.16 they have changed the value ranges for gain and offset. The default settings are recommended; alternately, consult QImaging for advice. Different light boxes will most likely require different settings for red, green and blue than those shown here. White balance should be set using the Auto White Balance feature of the software. The exposure of 00.300.000 (0.3 s) can be adjusted to change the brightness of the image. Differences in light box intensity may require a change to the exposure time and/or the aperture (see Note 8). Finally, setting the capture dimensions to a square area (1920 × 1920 pixels) rather than using the full rectangular capture area (2560 × 1920) reduced TIFF file size and eliminated unnecessary image data on either side of the circular well. 8. One way to adjust the brightness of the image is to adjust the aperture. Turning the aperture ring on the lens changes the diameter of a circular opening. A larger aperture will allow more light to reach the CCD array. For consistency in plate analysis, the same settings should be used for all experiments, including lens aperture. In our experiments, the aperture of the lens was empirically adjusted so that the images were not too bright to obscure faint blue cells, and not too dark to increase background counts of uninfected wells. Once an aperture setting was selected, the aperture ring was marked so that the same setting was used consistently. 9. Binning is a function in which the charges from adjacent pixels on the CCD array are combined. This is accomplished at the

CCD Camera Detection of HIV Infection

217

expense of spatial resolution (13). For example, 2 × 2 binning combines the charge from a 2 × 2 grid of pixels (4 pixels total) into one signal, effectively reducing the resolution 2 times in both the x and y axes. Similarly, 4 × 4 binning combines the charge from 16 adjacent pixels into one signal. The greatest benefit to binning is an increase in signal output and dynamic range (13). The increase in signal-to-noise ratio allows for detection of lower light levels. In QCapture Pro we checked the box “Adjust Exp for Binning.” When binning is increased from 1 × 1 (no binning) to 2 × 2 or 4 × 4, the exposure time is reduced to compensate for the increase in signal. Another benefit of binning is that the resolution of the image is reduced and it is more easily processed by the computer, effectively speeding up the on-screen response time when moving the object. We used binning solely for that purpose. Increased signal was not necessary because there was ample light emitted from the fluorescent light box. 10. An evenly plated monolayer of target cells is critical to preventing cells from overgrowing and forming dark clumps that may be detected by the software. Using clean reagents that will not leave dust or fibers in the wells is also important. 11. The parameter file, 5MGL.PAR, consists of the following default settings: ● Circle radius in tiff pixels: 860 ●

Window size in screen pixels: 640

●

Noise cutoff: 45

●

Spot area cutoff in pixels: 10

●

Minimum distance of spot maxima: 3

● Size of the structuring element: 10 Other than the circle radius, these settings should not be changed. The window size may need adjustment depending on the screen resolution of the computer. If you feel you need to adjust some of the other parameters, do so in a careful manner, comparing the results before and after changes to assess the modified performance of the program.

12. To count HIV infection, blue-stained cells were identified as dark spots in all three channels of the RGB spectrum even though they appeared predominantly blue to the naked eye. 5MGL.exe therefore analyzed a gray-scale image created by adding all three channels of the original 24-bit color TIFF image. To remove the impulse noise from the image, gray scale thinning and thickening morphological filters were applied with a point structuring element (16). Because the illumination of the image was not perfectly uniform,

218

Day

straightforward thresholding of the image could not be used to count the spots. Instead, using a square structuring element whose size was larger than the typical spot size, the software calculated the morphological opening of the image and then subtracted it from the image. The resulting image has a uniform, fluctuating background. Signal and noise were then separated by thresholding. The remaining islands (defined as unions of all side-to-side adjacent pixels with non-zero intensity) represented the actual blue-stained cells, plus a few small ghost spots that emerge by random agglomeration of noise after thresholding. The latter were typically much smaller than the islands representing actual blue cells and were removed by applying an area opening to the image (ibid). The remaining islands were considered as signals and were counted; this analysis was termed “Simple count.” However, sometimes a single island represented several adjacent cells that should have been counted separately; this was especially the case under conditions of high inoculum. Because a stained cell is darker on the inside than near the edge, adjacent cells have separate dark intensity maxima, which could be counted individually. This analysis was termed “Smart count.” Using Smart count, in terms of the intensity landscape, the hills on the islands rather than the islands themselves were counted. Both Simple and Smart counts were reported in the exported results file. For user verification of software performance, a processed version of each image containing a small red dot on every object counted was saved into the file folder containing the original images.

Acknowledgements We thank the coauthors of the original publication, Laura Martínez, Roman Šášik, Douglas Hitchin, Megan Dueck, Douglas Richman and John Guatelli for their contributions to this work. The system we developed was conceived with ideas from Chris Aiken. We are also grateful to Prentice Higley, Sherry Rostami and Nanette Van Damme for assistance during setup and evaluation. This work was supported by grants AI27670, AI043638, AI038201, the UCSD Center for AIDS Research (AI 36214), AI29164, AI047745, from the National Institutes of Health and the Research Center for AIDS and HIV Infection of the San Diego Veterans Affairs Healthcare System.

CCD Camera Detection of HIV Infection

219

References 1. Harada, S., Koyanagi, Y., and Yamamoto, N. (1985) Infection of HTLV-III/LAV in HTLVI-carrying cells MT-2 and MT-4 and application in a plaque assay. Science 229, 563–566 2. McKeating, J. A., McKnight, A., McIntosh, K., Clapham, P. R., Mulder, C., and Weiss, R. A. (1989) Evaluation of human and simian immunodeficiency virus plaque and neutralization assays. J. Gen. Virol. 70, 3327–3333 3. Nara, P. L. and Fischinger, P. J. (1988) Quantitative infectivity assay for HIV-1 and2. Nature 332, 469–470 4. Chesebro, B. and Wehrly, K. (1988) Development of a sensitive quantitative focal assay for human immunodeficiency virus infectivity. J. Virol. 62, 3779–3788 5. Akrigg, A., Wilkinson, G. W., Angliss, S., and Greenaway, P. J. (1991) HIV-1 indicator cell lines. AIDS 5, 153–158 6. Chackerian, B., Long, E. M., Luciw, P. A., and Overbaugh, J. (1997) Human immunodeficiency virus type 1 coreceptors participate in postentry stages in the virus replication cycle and function in simian immunodeficiency virus infection. J. Virol. 71, 3932–3939 7. Charneau, P., Mirambeau, G., Roux, P., Paulous, S., Buc, H., and Clavel, F. (1994) HIV-1 reverse transcription. A termination step at the center of the genome. J. Mol. Biol. 241, 651–662 8. Deng, H., Liu, R., Ellmeier, W., Choe, S., Unutmaz, D., Burkhart, M., Di Marzio, P., Marmon, S., Sutton, R. E., Hill, C. M., Davis, C. B., Peiper, S. C., Schall, T. J., Littman, D. R., and Landau, N. R. (1996) Identification of a major co-receptor for primary isolates of HIV-1. Nature 381, 661–666

9. Kimpton, J. and Emerman, M. (1992) Detection of replication-competent and pseudotyped human immunodeficiency virus with a sensitive cell line on the basis of activation of an integrated beta-galactosidase gene. J. Virol. 66, 2232–2239 10. Vodicka, M. A., Goh, W. C., Wu, L. I., Rogel, M. E., Bartz, S. R., Schweickart, V. L., Raport, C. J., and Emerman, M. (1997) Indicator cell lines for detection of primary strains of human and simian immunodeficiency viruses. Virology 233, 193–198 11. Day, J. R., Martínez, L. E., Šášik, R., Hitchin, D. L., Dueck, M. E., Richman, D. D., and Guatelli, J. C. (2006) A computer-based, image-analysis method to quantify HIV-1 infection in a single-cycle infectious center assay. J. Virol. Methods 137, 125–133 12. Beyon, J. D. E. and Lamb, D. R. (eds.) (1980) Charge-Coupled Devices and Their Applications. McGraw Hill, UK 13. Holst, G. C. (1998) CCD Arrays, Cameras, and Displays, 2nd ed. JCD Publishing, Winter Park, FL and SPIE Optical Engineering Press, Bellingham, WA 14. Howes, M. J. and Morgan, D. V. (eds.) (1979) Charge-Couple Devices and Systems. John Wiley & Sons, New York, NY 15. NIH. (2005) NIH Image/ImageJ, a public domain program developed at the US. National Institutes of Health. Available at http://rsb.info.nih.gov/nih-image/ 16. Soille, P. (2003) Morphological Image Analysis: Principles and Applications. Springer-Verlag, New York, NY

Chapter 12 Simple Luminescence Detector for Capillary Electrophoresis Antonio Segura-Carretero, Jorge F. Fernández-Sánchez, and Alberto Fernández-Gutiérrez Summary The performance of a homemade, simple, fluorescence-induced capillary electrophoresis (CE) detector is described here. It is based on LED as excitation source, a bifurcated optical fibre as a waveguide and a CCD as a photodetector. The connection of all the components is fairly easy, even for non-experts. This detector provides a low cost and rapid system for the determination of high-quantum-yield native fluorescence compounds and fluorescence derivatised compounds by CE with direct fluorescence determination. R-phycoerythrin and B-phycoerythrin were used as models for native fluorescence compounds and amine labelled with FITC were set as models for the fluorescence derivatised ones. Detection limits of 0.50 and 0.64 µg/mL for R-phycoerythrin and B-phycoerythrin and 1.6 × 10−7 M for FITC-labelled 1,6-diaminohexane were achieved. The homemade LED-IF detector is not expected to displace the LIF-IF one, but offers another possibility and a cheaper way to solve simple analytical problems for determining biomolecules. Key words: Light-emitting diode-induced fluorescence, CCD detector, Capillary electrophoresis, Protein analysis, Native fluorescence detection, Phycobiliproteins, Fluorescence-derivatised compounds, FTIC-labelled amines.

1. Introduction Capillary electrophoresis (CE) has many advantages and it has been proved to be one of the most powerful techniques for analysing biological samples (1), and more precisely for the analysis of biochemical compounds (2, 3). Since the introduction of automated CE instrumentation in the late 1980s, it has become a mature technique for the analysis of biomolecules (3). Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_12

221

222

Segura-Carretero et al.

For CE, many detection systems can be used, such as UVVisible absorption, fluorimetry, phosphorimetry, mass spectrometry, electrochemistry, conductivity and nuclear magnetic resonance, among others (4, 5). UV-absorbance is the primary detection method in CE analysis of biochemical compounds (3;) however, this detection suffers from high background noise and interference from other species (6). Laser-induced-fluorescence (LIF) is the most sensitive detection scheme in CE (3) and this detector has been widely applied for the analysis of biomolecules (3). In particular, it has been used to determine native fluorescence proteins such as phycobiliproteins (PBP) (7, 8). However, for inherently fluorescent analytes, it is necessary to choose readily available lasers that can be matched to their spectral properties. In addition, conventional lasers are generally expensive, relatively bulky, and have short lifetimes. In contrast, light-emitting diodes (LEDs) are reasonably priced and can be operated with battery power, so the output stability is significantly better than that of currently used laser and they have reasonably high intensity, small size, and a longer lifetime. The concentrated, small, cool emitter they provide is ideal for miniature analytical devices (9–11). In addition, LEDs at a variety of wavelengths in the UV-visible-near infrared are commercially available and one can easily choose suitable excitation wavelengths according to the properties of the samples. Despite their multiple advantages, the emission from LED sources is spatially incoherent and not monochromatic. When using LEDs as the excitation source for fluorimetric detection, considerable power loss might be expected due to failure in achieving proper focussing and obtaining the required spectral output. LED-induced fluorescence-CE (LED-IF-CE) detectors have been used for determining peptides (12), riboflavin (13, 14), dopamine and reserpine mixtures (15), agmatine (16), glutamic and aspartic acids (17), and D-alanine (18). However, all these LED-IF-CE detectors use plastic lenses or microscopes to focus the light from the LED to the capillary and/or to collect the fluorescence emission from the capillary to the detector (which is usually a photomultiplier). In addition, only one paper has been published on the use of a LED-IF detector for protein analysis by CE (6). It uses a deep-UV LED (280 nm) as an excitation source, which is expensive, needs a refrigeration system to be cooled down, and has only 100 h lifetime. To focus the light, it uses a set of three fused-silica ball lenses, and a photomultiplier as a detector. In addition, it uses a pin-hole filter and a bandpass filter. The several components were mounted for the alignment of all the optical parts and an A/D converter with a home-made Labview program was used to record the signals. Although the sensitivity of this device is comparable to those of pulsed UV lasers, it is expensive and cannot be used by non-experts. No

Simple Luminescence Detector for Capillary Electrophoresis

223

experimental setups, to our knowledge, have been developed using a commercial fluorescence spectrometer as fluorescence detector for detecting biomolecules. It is not easy to focus the light into the capillary or transport the emitted light to the detector but the use of an optical fibre for transporting the excitation light and collecting the emission fluorescence makes the optical system compact, simple and flexible (19). Thus, our research group has recently designed a detection cell for CE determinations with LED as light source and a CCD as detection system (20–22). All the other components of the proposed LED-IF-CE detector are commercially available and reasonably priced. The connection of all the components is fairly easy and can be performed by non-experts.

2. Materials 2.1. Reagents for the Alignment Optical Fibres and Capillary

Fluorescein isothiocyanate isomer I (FITC) was obtained from Scharlau (Barcelona, Spain). A 1 mM stock solution was prepared in acetone (Scharlau, Spain). It was stored protected from light at 4°C for no more than 2 months. 10−6 M FITC solution, which was used to adjust the distance between the tip of the probe and the capillary, was freshly prepared by diluting the FITC stock solution with bidistilled water (Millipore, Bedford, MA, USA).

2.2. Reagents for Protein Analysis

R-phycoerythrin (RPE) and B-phycoerythrin (BPE) were bought from Sigma (St. Louis, MO, USA) at 10 mg/mL concentration (see Note 1). They were supplied as suspensions in 150 mM sodium phosphate (pH 7.0), 60% ammonium sulphate, 1 mM EDTA and 1 mM sodium azide. They were stored while being protected from light at 4°C. Stock solutions of these proteins were freshly prepared (see Note 2) by diluting the appropriate amount of each protein in bidistilled water (see Note 3) and stored away from light at 4°C. Buffers were prepared by dissolving sodium borate (Sigma) in bidistilled water and adjusting the pH to the desired value by the addition of 0.5 M of HCl and/or NaOH (Sigma). The sodium salt of phytic acid (Sigma) was used as a buffer additive at 10 mM concentration. All buffers were stored in topaz bottles at room temperature for up to 6 months. All solutions were filtered through a 0.20 µm Millipore (Bedford, MA, USA) membrane filters before injection into the capillary.

2.3. Reagents for Amine Analysis

Amines (glycine, alanine, methylamine, dimethylamine, ethylamine, histamine, tyramine, hexylamine, dibutylamine, putrescine,

224

Segura-Carretero et al.

cadaverine, 1,6-hexanodiamine and agmatine) were obtained from Sigma. Stock standard solutions of different concentrations of a mixture of all analytes were prepared in bidistilled water. All the solutions were kept in a refrigerator at 4°C (see Note 4). 2.3.1. Reagents for the Derivatisation Procedure

1. Fluorescein isothiocyanate isomer I (FITC) 10−2 M in acetone (HPLC grade from Panreac). 2. Sodium carbonate/bicarbonate buffer 0.2 M at pH 9. 3. Acetone (Panreac) HPLC grade.

2.3.2. Reagents for the CE Separation

The running buffer used for electrophoretic separation of the FITC-labelled amines is: 1. Sodium borate solution (Sigma) 50 mM at pH 9.3. 2. Hexadimethrin bromide, HDB (Sigma; 0.001%, w/v). 3. 2-propanol (Panreac; 20%, v/v).

2.4. CE System

A Prince CE system (Prince Technologies, Emmen, the Netherlands) was used as CE instrument (see Note 5). The capillaries (fused silica) were purchased from Beckman with inner diameters (i.d.) of 50 µm for the analysis of proteins and 75 µm for the analysis or amines. In both cases, the total length is 57 cm and an effective separation length is 50 cm (see Note 6).

2.5. LED Induced Fluorescence Detection Apparatus

Figure 1 shows the LED-induced fluorescence CE system. The optical components are: 1. LS-450 Blue-LED (Ocean Optics) 2. 1/2″ Industrial Fluorescence Probe (Avantes BV) 3. Home-made detection cell 4. Fibre Optic Spectrometer AvaSpec-2048 (Avantes BV).

2.5.1. Ls-450 Blue-Led

A LS-450 Blue-LED pulsed light source which produces either pulsed or continuous output at 380, 395, 470, 518, 590 or 640 nm (depending on what LED was placed into the holder) is proposed as excitation source. The illumination system was also equipped with an SMA connector to attach to one of the branches of the optical fibre probe. The LED light used was the LED-470 which emits light at a maximum of 470 nm (see Note 7).

2.5.2. 1/2˝ Industrial Fluorescence Probe

A bifurcated optical fibre probe, 1/2″ Industrial Fluorescence Probe has to be used to focus the light emitted by the LED on the capillary and also to collect the fluorescence emission from the capillary to the CCD spectrometer. Figure 2 shows the scheme of this optical fibre probe. A branch of the probe which has 12 UV/VIS fibres of 200 µm-external diameter (called the illumination branch) was connected via SMA-905 with the LS-450 Blue-LED. The other branch, which is equipped with

Simple Luminescence Detector for Capillary Electrophoresis

225

½” Industrial Fluorescence Probe (Avantes) Illumination branch

LS-450 Blue-LED (Ocean Optics)

Fused-Silica capillary

CCD spectrometer (Avantes) Detection branch −

+ Detection cell CE system

Fig. 1. Picture and scheme of the experimental LED-IF-CE detector and its connection with a modular CE instrument. This figure was published in (22), Copyright Elsevier (2007).

a 600 µm-external diameter fibre and is called the detection branch, was also connected via SMA-905 with the CCD spectrometer. The end of the probe consists of a stainless steel cylinder, 1/2″ diameter, containing a 10 mm diameter × 1 mm thick sapphire windows with anti-reflection coating and the 12 illuminating 200 µm fibres surrounding the single 600 µm-fibre which is in the middle of the tip (see Note 8). 2.5.3. Home-Made Detection Cell

A home-made detection cell has to be used to align the capillary with the optical fibre probe. This cell consists of a T-style detection

226

Segura-Carretero et al.

cell similar to the cell proposed by Yang and Guan (19). One 1/2″ diameter channel was perforated at 90° in a 90 × 50 × 32 mm piece of methacrylate and painted black. Two septums were fixed to maintain the correct position of the capillary in the middle of the 1/2″ channel. A detection window (about 1 cm length) was formed in the capillary by burning off the polyimide coating and this window was placed in front of the probe which was placed at 90° to the capillary. Figure 3 shows a scheme of the home-made detection cell (see Note 9). The Fibre Optic Spectrometer AvaSpec-2048 is based on the AvaBench-75 symmetrical Czerny-Turner design with 2,048 pixel CCD Detector Array. The spectrometer has an SMA fibre-optic

1 Read fiber (φ=600µm) 12 Illumination fibers (φ=200µm)

SMA-12 UV/ VIS fiber 2m

Stainless steel tip (φext=1/2”) SMA-1 UV/ VIS fiber

15 .5

m m

Fig. 2. Scheme of the 1/2″ Industrial Fluorescence Probe supplied by Avantes et al. (20), Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

Methacrylate cube

50 mm

Detection windows

½” Septum Separation capillary

30 mm

2.5.4. Fibre Optic Spectrometer Avaspec-2048

Screws to fix the tip Bifucarted optical fiber probe

Fig. 3. Cross-sectional diagram (a) and picture (b) of the homemade detection cell used in the LED-IF-CE system. Arraez-Roman et al. (20), Copyright Wiley-VCH Verlag GmbH & Co. KGaA. Reproduced with permission.

Simple Luminescence Detector for Capillary Electrophoresis

Fused-Silica Capillary

Methacrylate cube

227

End of the probe

Septum

Fig. 3 (continued)

entrance connector, a collimating and focusing mirror and a diffractional grating code “UB” (UV/VIS; 200–800 nm) with a blaze at 300 nm and a slit of 500 µm (see Note 10). Avasoft-Full software controls the spectrometer and saves the emission spectra. The software also permits the recording of eight different timedrive functions simultaneously (see Note 11), thus permitting the recording of the emission intensity at eight different wavelengths versus time, simultaneously, or seven wavelengths and a complete emission spectrum versus time, etc. (see Note 12).

3. Methods 3.1. The Alignment of the Probe and Capillary

An optimum alignment of the probe and the capillary is necessary in the homemade detection cell to improve the sensitivity. The most important parameters to take into account in the design of a detection cell are: 1. The relative horizontal position of the probe in respect to the capillary 2. The distance between the tip of the probe and the capillary. We propose one methodology to optimise the horizontal position of the probe and two different methodologies to set the optimum distance between the probe and the capillary

3.1.1. Optimisation of the Relative Position Capillary/Probe

1. Fix either the capillary or the probe. It depends on the detection of cell use to focus them. 2. Turn on the LED light (475 nm) and set an integration time of 200 ms and an average of five on the CCD spectrometer.

228

Segura-Carretero et al.

3. Launch the measurement software, collecting the emission at 520 nm and registering the signal/noise ratio simultaneously. 4. Run a solution of 10−6 M FITC through the capillary by pressure flow. 5. When FITC crosses the detection window, increases in the signals must be registered. Wait until the signals are stable. 6. Move the probe horizontally until maximal intensity and/or signal/noise ratio. 7. Fix the probe with a screw. 3.1.2. Optimisation of the Distance between Capillary and Probe: Flow Method

1. Fix either the capillary or the probe. It depends on the detection of cell use to focus them. 2. Turn on the LED light (475 nm) and set an integration time of 200 ms and an average of five on the CCD spectrometer. 3. Launch the measurement software, collecting the emission at 520 nm and registering the signal/noise ratio simultaneously. 4. Run a solution of 10−6 M FITC through the capillary by pressure flow. 5. When FITC crosses the detection window, increases in the signals must be registered. Wait until the signals are stable. 6. Move the probe vertically until maximal intensity and/or signal/noise ratio. 7. Fit the probe with a screw (see Note 13).

3.1.3. Optimisation of the Distance between Capillary and Probe: Separation Method

1. Fix both the capillary and the probe. 2. Condition the capillary. 3. Turn on the LED light (475 nm) and set an integration time of 200 ms and an average of five on the CCD spectrometer. 4. Launch the measurement software collecting the emission at 520 nm. 5. Run a separation of 10−6 M FITC by following the same protocol as for amine-labelled compounds and obtain the electropherogram. 6. Move the probe horizontally. 7. Start a new separation (step 2) until you get the best electropherogram. This protocol is quite tedious but can be used for checking the system periodically by comparing the results (intensities, areas and signal/noise ratios).

3.2. Electrophoretic Method for Protein Analysis

First, the capillary has to be conditioned and washed following these protocols:

Simple Luminescence Detector for Capillary Electrophoresis 3.2.1. Before First Use

229

1. 0.2 M NaOH for 15 min. 2. Bidistilled water for 5 min. 3. Running buffer (50 mM borate – 10 mM phytic acid (pH 8.4) ) for 30 min. 1. 0.2 M NaOH for 3 min. 2. Bidistilled water for 2 min. 3. Running buffer for 5 min (see Note 14). The experimental conditions for the separation of proteins were: 1. The running buffer was formed by 50 mM borate – 10 mM phytic acid (pH 8.4). 2. Hydrodynamic injection was achieved for a period of 20 s at 55 mbar. 3. 20 kV was applied to power the CE separation. 4. The detection was carried out with a 470 nm LED. 5. The detection was recorded by integrating the emitted light between 550 and 650 nm with an integration time of 200 ms and an average of ten scans (see Note 15). Typical electropherograms in the LED-IF-CE device using the optimised conditions for RPE and BPE are shown in Fig. 4. The analytical performance of the proposed methodology is shown in Table 1 (see Note 16). Although the LOD was worse than that usually achieved by LIF-CE (0.1 µg/mL) (7), it has to be pointed out that those achieved with the homemade LED-IF system (0.5 µg/mL) are

10000 8000 Counts

3.2.2. Between Consecutive Injections

6000 4000 2000 0 0

1

2 Time (min)

3

4

Fig. 4. Optima electropherograms obtained for RPE (50 µg/mL) and BPE (50 µg/mL) by using the LED-IF-CE (excitation: LED-470, emission: integration 550–650 nm; integration time 200 ms, average ten scans). CE conditions: 50 mM borate buffer – 10 mM phytic acid (pH 8.4), 20 kV and 20 s (55 mbar) of hydrodynamic injection. This figure was published in (22), Copyright Elsevier (2007).

230

Segura-Carretero et al.

Table 1 Analytical performance of the determination of RPE and BPE by LED-IF-CE RPE

BPE

0.64

0.50

Calibration equation

A = 1,070c + 282

A = 1,315c + 2748

r

0.9991

0.9993

Linear dynamic range (µg/mL)b

0.64–100

0.50–100

Intra-day repeatabilityc

RSDt (%)d

0.6

0.8

RSDA (%)e

1.5

1.6

RSDt (%)d

1.0

1.0

RSDA (%)e

1.6

1.5

RSDt (%)d

1.6

1.7

RSDA (%)e

1.9

2.1

LOD (µg/mL) a

Inter-day repeatabilityc

Reproducibilityf

A represents peak area; c means analyte concentration in µg/mL for five points (n = 3) From the detection limits c Measured for six consecutives injections or each analyte within the same day (intra-) and on three different days (inter-) d For migration times e For peak areas f Measured from three consecutive injections (n = 6) with two different capillaries a

b

very competitive if we take into account the lower power of the LED compared to the laser (see Note 17). 3.3. Derivatisation Procedure of Non-fluorescence, Aminated, Compounds

The derivatisation procedure for FITC-labelled amines and amino acids has been described by our research group (23) and the parameters that affect the FITC-labelled aminated compounds are described and optimised. It consists on: 1. 495 µL of 0.2 M carbonate buffer at pH 9. 2. 1,000 µL of 10−2 M FITC solution. 3. 1,000 µL of acetone. 4. 800 µL of a mixture of 30 amines and amino acids at 1 mg/ mL each one. 5. Bidistilled water until 5 mL in a test tube. 6. This solution has to be put into a thermostatic bath for 2 h at 50°C. 7. Finally, 2 mL of the resulting solution has to be diluted up to a final volume of 10 mL with bidistilled water before the CE analysis (see Note 18).

Simple Luminescence Detector for Capillary Electrophoresis

231

3.4. Electrophoretic Method for FTICLabelled Amine Analysis

First, the capillary has to be conditioned and washed, following these protocols. 1. 0.2 M NaOH for 15 min.

3.4.1. Before First Use

3. Running buffer (50 mM sodium borate (pH 9.3), HDB (0.001%, w/v) and 20% 2-propanol) for 30 min.

3.4.2. Between Consecutive Injections

1. 0.2 M NaOH for 3 min.

2. Bidistilled water for 5 min.

2. Bidistilled water for 2 min. 3. Running buffer for 5 min (see Notes 14 and 19). The experimental conditions for the separation of proteins were as follows: 1. The running buffer was formed by 50 mM sodium borate (pH 9.3), HDB (0.001%, w/v) and 20% 2-propanol (see Note 20). 2. Hydrodynamic injection has to be achieved for a period of 12 s at 0.5 psi (34 mbar). 3. −20 kV has to be applied to power the CE separation. 4. The detection has to be carried out with a 470 nm LED. 5. The detection has to be recorded at 520 nm with an integration time of 200 ms and an average of five scans. Typical electropherograms in the LED-IF-CE device using the optimised conditions for the mixture of 13 aminated compounds is shown in Fig. 5.

4. Notes 1. Phycobiliproteins, produced by cyanobacteria, are highly fluorescent proteins. Phycobiliproteins mainly include three classes: phycoerythrin (PE), phycocyanin (PC), and allophycocyanin (APC). Phycobiliproteins which constitute up to 60% of the soluble proteins in cyanobacterial cells are important analytes to be considered, due to the importance of cyanobacteria in ocean sciences. Because of their excellent fluorescence quantum yields (0.51–0.98) they have been widely used as fluorescent labels in life sciences. 2. The protein solutions are very instable. They can be stored in topaz vials at 4°C for a maximum of 1 week. 3. Unless stated otherwise, all solutions should be prepared in water that has a resistivity of 18.2 MΩ cm and a total organic content of less than five parts per billion. This standard is referred to as “bidistilled water” in this text. It was obtained from Milli-Q water purification system.

Segura-Carretero et al.

4,5 4000

Fluorescence @ 520 nm (RFU)

232

3000

6-9

2000

B 10-12

1000 1,2

3

13

0 5

7

9

11

Time (min)

Fig. 5. LED-IF-CE optimum analysis of the mixture of aminated compounds: (1) glycine (0.64 µg/mL); (2) alanine (0.64 µg/mL); (3) methylamine (2.04 µg/mL); (4) dimethylamine (2.32 µg/mL); (5) ethylamine (3 µg/mL); (6) histamine (0.64 µg/mL); (7) tyramine (0.64 µg/mL); (8) hexylamine (0.64 µg/mL); (9) dibutylamine (1.6 µg/mL); (10) putrescine (10.24 µg/mL); (11) cadaverine (3.2 µg/mL); (12) 1,6-hexanodiamine (5.44 µg/mL); (13) agmatine (1.72 µg/mL); B blank. Experimental conditions: 50 mM sodium borate pH 9.3, 0.001% of HDB; fused silica capillary (50 cm × 75 µm i.d.); running voltage, −20 kV; injection, 12 s (0.5 psi); temperature is not controlled; excitation, 470 nm; detection, 520 nm with integration time of 200 ms and average of five. Casado-Terrones et al. (21), Copyright Springer-Verlag. With kind permission of Springer Science and Business Media.

4. The amine solutions are also instable. They can be stored in topaz vials at 4°C for a maximum of 3 weeks. 5. We propose the use of this instrument due to its low cost, simplicity and modularity. Any other instrument could be used whenever the capillary is accessible and is located outside the instrument or can be taken out. In this case, the fact that the temperature cannot be controlled in the capillary is a huge disadvantage in achieving better resolution (compared with the commercial fluorescence detector for CE). 6. The optimum inner diameter of the capillary was chosen taking into account a compromise solution between resolution and sensitivity.

Simple Luminescence Detector for Capillary Electrophoresis

233

7. More technical information about this light source can be found on http://www.oceanoptics.com/products/ls450. asp. We recommend this source because it is quite easy to replace the LED lamp and obtain several emission wavelengthswith the same lamp holder. Any other LED or lamp should be used wherever it can be assembled via SMA or other connectors with the fibre optical probe. For example, we used an Xe-lamp as excitation lamp, but it needs the use of filters or monochromators for selecting the excitation wavelength and avoiding interference of the lamp in the fluorescence determination (22). In contrast, the Xe lamp offers the possibility to excite with UV wavelengths. 8. We strongly recommend the use of this fibre optical probe. Several were tested and the best results were always obtained with the Avantes’ probe. It is especially designed to measure fluorescence. For further information you can visit www. avantes.com. 9. The detection cell was fabricated by the mechanical factory of the Centre of Scientific Instrumentation at the University of Granada: the channels for probe and capillary were perforated and aligned by using a computer numerical control (CNC) turning centre and the septums were fixed by using a two-component glue. Other alternative detection cells can be used. One of the handicaps of the proposed detection cell is that only the distance between the probe and the capillary can be varied. The horizontal relative position probe/ capillary is fixed; therefore, any mistake in the perforation of the 1/2″ channel makes it useless. The best option consists on the development of a detection cell which allows horizontal and vertical movements for the probe or the capillary or even both probe and capillary. The scheme of this potential detection cell is shown in Fig. 6. This detection cell allows the perfect horizontal and vertical alignment of both probe and capillary and could improve the sensitivity of the LED-IF-CE detector proposed. 10. One of the most important parts of the spectrometer is the diffractional grating. It has to be carefully chosen according to the application which is going to be followed. Several diffractional gratings are provided by Avantes to be used in the fibre optic spectrometer; further information can be found on the Internet (www.avantes.com). On the other hand, both the number of lines per millimetre of the grating and slit width of the spectrometer have to be chosen because they affect the resolution of the instrument. We recommend using the highest slit width (500 nm) to collect the greatest amount of light from the capillary and improve the sensitivity.

234

Segura-Carretero et al. Screw for Y-axe movement

z

Separation capillary Detection window

x

y

Screw for X-axe movement Screw for Z-axe movement

Bifurcated optical fiber probe

Fig. 6. Novel detection cell proposed to improve the sensitivity of the LED-IF-CE detector for recognising less intensive molecules that phycobiliproteins or amine-labelled compounds.

11. The functions which can be programmed are: intensity at a fixed wavelength, logical mathematical operations with the intensities at several wavelengths, the integral between two wavelengths and the spectrum between two wavelengths. 12. The use of a CCD camera improves the applicability of this detector compared to the others using PMTs because it permits the recording of several optical parameters at the same time and even allows one to record two-dimensional electropherograms whenever a home-made software is developed. Until now, no software is commercialy available to record 2D electropherograms with this instrument. 13. The best signal/noise rate achieved by using our home-made detection cell was 0.13 mm between probe and capillary. 14. We recommend using fresh buffer after six consecutive analyses. 15. The integration time is the CCD readout frequency and therefore the exposure time of the CCD detector. The average is the number of scans to average. Both parameters have to be carefully optimised. The effect of the integration time is related to the sensitivity. Normally, an increase of the integration time improves the LOD up to a certain value, and higher integration times would provide a plateau. With regard to average, it has to be mentioned that it does not affect the sensitivity. Increasing the average decreases the sharpness of the peaks, providing a worse resolution. In this case, integration time 200 ms × 5 averages means 1 point by 1 s, so approximately 26 points to draw an electrophoretic peak. 16. The limit of detection (LOD) was calculated in two ways: from the peak area on the basis of 3s0/b, where s0 is the

Simple Luminescence Detector for Capillary Electrophoresis

235

standard deviation of the blank (largest deviation of detector signal from baseline measured in a section of about 300 points in the absence of analyte) and b is the slope of every calibration equation; LODs were also calculated from the peak height based on a signal-to-noise ratio. of 3. Practically the same results were achieved in both instances, so in Table 1 only the LODs calculated from the area are shown. 17. The main cause of low LOD may be the low power of the LED (6, 20). Two different aspects should be modified to improve the LOD. From an instrumental point of view, more powerful LEDs, which can be easily implemented on commercially available fibre-optic couplers, more sensitive CCD detectors and optical fibre probe, which can collect more efficiently the fluorescence emission, improve the sensitivity of the proposed detector. In addition, the use of capillaries or detection cells which increase the optical path-length such as axial illumination cells and “bubble cell” capillaries increase the sensitivity of the proposed device. On the other hand, from an experimental point of view, an increase of the injection times and the use of on-line pre-concentration techniques, such as stacking techniques, should also increase it. 18. Figure 7 shows the derivatisation reaction between FTIC and amines. This protocol has been optimised for the derivatisation of a mixture of 30 amines (1 ppm of each one). Therefore, for real applications this protocol may be re-optimised to take into account the number and the concentration of the aminated compounds that should be labelled. 19. After 5 min of running the buffer we conditioned it but, in addition, a semipermanent polication of HDB was generated into the capillary. Therefore, it is crucial to condition the capillary with running buffer just before the injection of the sample and the electrophoretic separation. 20. Without this organic solvent addition no resolution was achieved.

O

OH

O

COOH

N

C

S

O

O

+ R NH2

OH

COOH

R HN

NH C S

Fig. 7. Derivatisation reaction between FITC and amine compounds.

236

Segura-Carretero et al.

Acknowledgements This work was supported by the projects CTQ2005-01914/ BQU of the Ministry of Education and Science and RMN-666 from the Andalusian Government. References 1. Lagu, A.L. (1999) Applications of capillary electrophoresis in biotechnology. Electrophoresis 20, 3145–3155 2. Jenkins, M.A., Guerin, M.D. (1997) Capillary electrophoresis procedures for serum protein analysis: comparison with established techniques. J. Chromatogr. B 699, 57–268 3. Quigley, W.C., Dovichi, N.J. (2004) Capillary electrophoresis for the analysis of biopolymers. Anal. Chem. 76, 4645–4658 4. Cazes, J. (Ed.) (1997) Analytical instrumentation handbook, Marcel Dekker, UK 5. Fernandez-Gutierrez, A., Segura-Carretero, A. (Eds.) (2005) Electroforesis capilar: Aproximación según la técnica de detección, Editorial Universidad de Granada, Granada 6. Sluszny, C., He, Y., Yeung, E.S. (2005) Light-emitting diode-induced fluorescence detection of native proteins in capillary electrophoresis. Electrophoresis 26, 4197–4203 7. Viskari, P.J., Kinkade, C.S., Colyer, C.L. (2001) Determination of phycobiliproteins by capillary electrophoresis with laser-induced fluorescence detection. Electrophoresis 22, 2327–2335 8. Viskari, P.J., Colyer, C.L. (2002) Separation and quantification of phycobiliproteins using phytic acid in capillary electrophoresis with laser-induced fluorescence detection. J. Chromatogr. A 972, 269–276 9. Johns, C., Macka, M., Haddad, P.R. (2004) Design and performance of a light-emitting diode detector compatible with a commercial capillary electrophoresis instrument. Electrophoresis 25, 3145–3152 10. Kuo, J.S., Kuyper, C.L., Allen, P.B., Fiorini, G.S., Chiu, D.T. (2004) High-power blue/ UV light-emitting diodes as excitation sources for sensitive detection. Electrophoresis 25, 3796–3804 11. Webster, J.R., Burns, M.A., Burke, D.T., Mastrangelo, C.H. (2001) Monolithic capillary electrophoresis device with integrated fluorescence detector. Anal. Chem. 73, 1622–1626

12. Hillebrand, S., Schoffen, J.L., Mandaji, M. (2002) Performance of an ultraviolet lightemitting diode-induced fluorescence detector in capillary electrophoresis. Electrophoresis 23, 2445–2448 13. Su, A.K., Lin, C.H. (2003) Determination of riboflavin in urine by capillary electrophoresis– blue light emitting diode-induced fluorescence detection combined with a stacking technique. J. Chromatogr. B 785, 39–46 14. Su, A.K., Chang, Y.S., Lin, C.H. (2004) Analysis of riboflavin in beer by capillary electrophoresis/blue light emitting diode (LED)-induced fluorescence detection combined with a dynamic pH junction technique. Talanta 64, 970–974 15. Tsai, C.H., Huang, H.M., Lin, C.H. (2003) Violet light emitting diode-induced fluorescence detection combined with on-line sample concentration techniques for use in capillary electrophoresis. Electrophoresis 24, 3083–3088 16. Zhao, S., Wang, B., Yuan, H., Xiao, D.J. (2006) Determination of agmatine in biological samples by capillary electrophoresis with optical fiber light-emitting-diode-induced fluorescence detection. J. Chormatogr. A 1123, 138–141 17. Wang, C., Zhao, S., Yuan, H., Xiao, D. (2006) Determination of excitatory amino acids in biological fluids by capillary electrophoresis with optical fiber light-emitting diode induced fluorescence detection. J. Chromatogr. B 833, 129–134 18. Zhao, S., Wang, B., He, M., Bai, W., Chen, L. (2006) Determination of free D-alanine in the human plasma by capillary electrophoresis with optical fiber light-emitting diodeinduced fluorescence detection. Anal. Chim. Acta 569, 182–187 19. Yang, B., Guan, Y. (2003) Light-emittingdiode-induced fluorescence detector for capillary electrophoresis using optical fiber with spherical end. Talanta 59, 509–514 20. Arraez-Roman, D., Fernandez-Sanchez, J.F., Cortacero-Ramirez, S., Segura-Carretero,

Simple Luminescence Detector for Capillary Electrophoresis A., Fernandez-Gutierrez, A. (2006) A simple light-emitted diode-induced fluorescence detector using optical fibers and a charged coupled device for direct and indirect capillary electrophoresis methods. Electrophoresis 27, 1776–1783 21. Casado-Terrones, S., Cortacero-Ramirez, S., Carrasco-Pancorbo, A., Segura-Carretero, A., Fernández-Gutiérrez, A. (2006) Comparative study between a commercial and a homemade capillary electrophoresis instrument for the simultaneous determination of aminated compounds by induced fluorescence detection. Anal. Bioanal. Chem. 386, 1835–1847

237

22. Casado Terrones, S., Fernández Sánchez, J.F., Segura Carretero, A., Fernández Gutiérrez, A. (2007) Simple luminescence detectors using a LED or a Xe lamp, optical fibre and CCD device or photomultiplier for determining proteins in capillary electrophoresis: a critical comparison. Anal. Biochem. 365, 82–90 23. Arráez-Román, D., Segura-Carretero, A., Cruces-Blanco, C., Fernández-Gutiérrez, A. (2004) Subminute and sensitive determination of the neurotransmitter serotonin in urine by capillary electrophoresis with laser fluorescence detection. Biomed. Chromatogr. 18, 422–426

Chapter 13 Optical System Design for Biosensors Based on CCD Detection Douglas A. Christensen and James N. Herron Summary The use of Charge Coupled Device (CCD) detectors as an integral part of a biosensing system has become widespread in recent years due to several advantages of this type of detection, such as the ability to image multiple zones on the sensor, the flexibility of defining the sensing configuration and the lownoise performance of the detectors. The specification of the CCD as well as the selection of the other components in this system – including the source and the filters – is driven by the particular transduction mechanism, but all parts must be matched. Particular attention must be paid to reducing the various noise components of the CCD to obtain the lowest detection level, and it is shown that cooling the CCD is often a wise choice. Key words: CCD, Fluorescent dye, Laser diode, Lens; Filter, Imaging, Noise, Thermoelectric cooler, Planar waveguide, Resolution, Detection level, Autofluorescence.

1. Introduction Biosensors have become more complex in the past few years, both in the variety of analytes detected (1) and in the density of the sensing zones, particularly for molecular arrays (2) and labon-a-chip devices (3). Quantitative sensitivities have also been pushed to the lower levels required for clinical diagnostics (4) and biohazards detection (5). Optical methods play a significant role in many of these biosensors (6). These trends have strengthened the role of CCD (Charge Coupled Device) detectors as part of the overall optical biosensor configuration. CCD detectors have historically been attractive

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI:10.1007/978-1-60327-567-5_13

239

240

Christensen and Herron

compared to photomultiplier tubes or photodiode arrays from the standpoint of size, convenience, and/or signal-to-noise ratio, and they benefit from ongoing advances in semiconductor fabrication technology. Numerous consumer applications have provided a wide variety of CCD chips that have helped keep their cost reasonable, a necessary requisite for point-of-care applications, even when packaged as a camera (7). Although each optical biosensor configuration is unique, there are common elements and similar engineering design decisions found in every application. The CCD chip itself may contain many integrated electro-optical components, such as shift registers, preamplifiers and charge-to-voltage stages (8), but it is only a part (albeit a critical one) of the overall biosensing system. The choice of the other optical components is driven by their unique application and the CCD must be wedded to these components. This chapter is organized along the lines of a decision sequence that applies to most optical biosensing schemes, starting with the most fundamental and including several considerations specific to the CCD. As an example of a typical optical biosensing design (whose components will be described in detail in Subheading 2), a fluorescent immunoassay based on CCD imaging is illustrated (9–11). Its components are shown in the block diagram of Fig. 1 and their embodiment in a prototype lab instrument is pictured in Fig. 2. In this example, capture antibodies (Abs) to the analyte of interest – for instance cTnI or hCG – are pre-immobilized

Source isolation filter

Collimating lens

Biosensor Source beam stop

Beam shaping lens(es) Fluorescence

Collection/imaging lenses Source rejection filter

CCD camera

Digital computer

Fig. 1. General components employed in a CCD-based optical biosensor system. Specific elements may vary depending upon the particular application.

Optical System Design for Biosensors Based on CCD Detection Beam directing Beam directing mirrors mirrors Flowcell Flowcell

Waveguide Waveguide sensor sensor

241

Semiconductor laser Semiconductor laser Beam shaping lenses Beam shaping lenses

Collimating Collimating lens (inside plate) lens (inside plate) Collection and imaging lens Collection and imaging lens Fluorescence Fluorescence

45-degree bending mirror 45-degree bending mirror

Laser rejection filter (inside) Laser rejection filter (inside)

CCD camera CCD camera

Fig. 2. Prototypical example of a fluorescence biosensor reader. This particular biosensing technique – a solid-phase sandwich immunoassay – employs a disposable planar waveguide (inside the flowcell) with analyte-specific antibodies immobilized on its surface. Fluorescence is excited within the evanescent zone of waveguided light from the laser and is detected by the CCD camera with attached imaging lens. The imaging CCD configuration gives this device the capability of detecting individual zones on the sensor surface.

on the surface of a thin transparent substrate acting as a planar waveguide, such that it can be molded from optical plastic. The light from a laser diode source is broadened into a sheet beam by beam-shaping lenses and coupled by an integral lens into the end of the waveguide where it propagates by total internal reflection down the length of the guide. Importantly, a portion of the light energy, called the evanescent tail, reaches in an exponential manner a small distance (about 100 nm) beyond the waveguide surface into the regions on either side, where it can function as excitation for any bound fluorescent molecules. After first being mixed with free “reporter” Abs labeled with fluorescent dye molecules, the solution containing the unknown amount of analyte is introduced into a compartment of the disposable sensor cartridge. There it comes into contact with one surface of the waveguide. Then units composed of sandwiches of [reporter Ab + analyte + capture Ab] begin forming on the waveguide surface; their density is proportional to the amount of analyte present. They are small enough (∼10–20 nm) to fall within the evanescent zone of excitation and emit fluorescence proportional to the concentration of bound analyte. Unbound reporter Abs in the bulk fluid are not present to any great extent

242

Christensen and Herron

within this evanescent zone and are therefore dark – no wash step is necessary to remove unbound material. The fluorescence is emitted more or less isotropically; some of it passes through the transparent waveguide to be collected by an imaging lens and directed to a cooled CCD detector. A laser rejection filter is placed near or within the collection/imaging lens(es) to block any source light inadvertently scattered by the waveguide sensor while still passing the fluorescent signal. A major feature of this optical arrangement is its imaging capability, by which different zones on the sensor surface can be separately detected in different regions of the CCD chip. This gives the biosensor the ability to detect multiple analytes [e.g., cTnI, myoglobin and ckMB in a cardiac panel (12)] and allows the possibility of on-board calibration zones. This particular fluorescent immunoassay will serve as an example of the choices that are typically made of the optical components of a CCD biosensor system. 1.1. Optical Transduction Means (e.g., Fluorescent Dye)

The mechanism of analyte detection – whether it be surface plasmon resonance (13), Raman scattering (14), fluorescence, colorometric absorption, or other means – normally sets the wavelength range of the system and therefore limits the specifications of the other components. Photons in the ultraviolet (UV) range (wavelengths of 280– 400 nm) possess high energy (several eV), leading to a desirably low value of thermally generated dark-current noise in the corresponding detector, as a UV detector has a band-gap energy much larger than thermal energy kT (0.026 eV) at room temperature. But the choice of available UV detectors is limited (especially for CCDs); common glass optical elements will not pass UV radiation, which means that fused silica or other materials like CaF2 or MgF2 must be used; and proteins in biological specimens, such as whole blood or plasma, exhibit significant autofluorescence when excited by UV light (see Note 1). At the other end of the scale, components in the infrared (IR) region (700 nm to 1.5 µm, corresponding to photon energy of about 1 eV) are readily available, thanks to their wide use in the telecommunications industry. But thermally generated dark current is larger; detectors must be sensitive to infrared (e.g., InGaAs or Ge); and fluorescent dyes are limited in this range. In the visible range (400–700 nm), components are plentiful, especially Si photodetectors; thermally generated dark current is moderate; and fluorescent dyes are numerous (see Notes 2 and 3).

1.2. Source (e.g., Semiconductor Laser)

Possibilities for a source include LEDs with single or multiple wavelengths in the visible and near IR; gas or solid-state lasers, such as the compact and popular 532-nm frequency-doubled laser; or semiconductor lasers (a.k.a. laser diodes). Being coherent,

Optical System Design for Biosensors Based on CCD Detection

243

the beams from lasers can be much better collimated and focused to a smaller point than those from LEDs, resulting in higher intensities at a target. Gas and solid-state lasers have better wavefront uniformity than laser diodes and therefore can be tightly focused more easily; if high focus intensity is important for a particular sensor performance, the output of a laser diode must be circularized with special lenses. Red and IR laser diodes offer appreciable power (10s of mW), small size, and high electrical efficiency (∼70%). Laser diodes can be driven in either of the two modes: constant current or constant power (assuming the laser has a built-in monitor photodiode, which is common). The constant-power mode is preferred for most quantitative biosensor applications since the signal is often proportional to the source power as well as analyte concentration. Cooling the laser diode offers three advantages, at the cost of complexity (1) the diode can be driven at higher current; (2) the output power is more stable if driven in constant current mode; and (3) regardless of the driving mode, the output wavelength is more stable. This last point is important because a drift in wavelength often translates to a change in the efficiency of the transduction means (for example, a variation in the excitation efficiency of a fluorescent dye). Figure 3 plots the excitation and emission spectra of a typical visible dye, Alexa Fluor® 660 (Invitrogen), along with a laser line near the peak of absorption. It can be seen that if the laser wavelength shifts by more than a fraction

100 Alexa 660 absorption

Normalized Values

80 Alexa 660 emission 60

40 658-nm laser line 20

0 500

550

600

650

700

750

800

Wavelength (nm)

Fig. 3. Excitation (absorption) and emission spectra of a typical red fluorescent dye, Alexa Fluor® 660 (Invitrogen). Also shown is the narrow wavelength of a diode laser positioned near optimal excitation of the dye.

244

Christensen and Herron

of a nm, the excitation efficiency will change correspondingly. A visible semiconductor laser has a wavelength vs. temperature coefficient of about + 0.2 nm/°C. 1.2. Source (e.g., Semiconductor Laser) 1.3.1. Beam Forming Lenses

1.3.2. Collection/ Imaging Lens

All of the sources mentioned above, except for gas and solid-state lasers, need a lens at their output to collect and collimate the emitted light. For example, the light emitted from the junction of a laser diode is spread into an oval-shaped cone typically 10° × 30° in extent. If the beam is to be propagated over even a few centimeters on its way to the sensor, it must be collimated (and perhaps circularized). Many off-the-shelf laser diode modules already contain lenses to perform these functions. Another separate lens may be needed to direct and shape the beam as it enters the sensor. Since the beam has been collimated and is therefore narrow, this lens can have a small diameter (or equivalently, a large f#, where f# is defined as the ratio of lens focal length f to its diameter D; thus f# = f/D). The optical signal is gathered by the collection lens and if the signal is emitted into a wide angle (as with fluorescence), this lens should have a large diameter (or a low f#, approaching f/1.0) to collect as much signal as possible. To take advantage of the capability of the CCD detector to image multiple zones, the optical arrangement of the imaging path must be configured in the geometry shown in Fig. 4 . The

lo Object Sensor area

do

D Collection/imaging lens with focal length f and f# = f/D

di

CCD area Image

li Fig. 4. Relative spacing of sensor, lens and CCD that allows imaging of individual zones on the CCD.

Optical System Design for Biosensors Based on CCD Detection

245

first step is to match the CCD size (corresponding to the maximum image size li) to the size of the sensor area (the length lo of the emitting object). This ratio defines the desired magnification M of the system (15), given by M =

li d = i, l o do

(1)

where the quantities are indicated in Fig. 4. Typical magnifications range from 0.05 to 10.0. Since l i and lo are both known, the ratio d i/do is set by Eq. 1. The focal length f of the lens then enters the calculation through the lens law for imaging: 1 1 1 + = . d i do f

(2)

The value of f is adjusted until both Eqs. 1 and 2 are satisfied with values of di and do that are practical for the allowed physical size of the optical reader. Once f is determined, the diameter D of the lens is selected to be as large as practical (f# as small as possible) to collect the maximum amount of signal. At low f#, a good quality lens, such as a compound or camera lens, is required to avoid blurring the image by spherical aberration. 1.4.Filters 1.4.1. Source Isolation Filter

1.4.2. Source Rejection Filter

This filter may be needed if the wavelength of the measured signal is shifted from the source excitation wavelength (as with fluorescence) and if the source is broad enough to spill over and pollute the signal. LEDs and many laser diodes have broad spectra and may require a source isolation filter to prevent direct spillover from source to signal. On the other hand, gas and solid-state lasers have very narrow bandwidths (much less than 1 nm) in their major output spectra and therefore are unlikely to contain wavelengths that overlap the signal. (However, even these lasers may occasionally have other weaker lasing lines, necessitating a source isolation filter.) If used, the upper cutoff wavelength of the source isolation filter must be lower than the passband of the source rejection filter, which is described next. Many optical biosensor arrangements rely on detecting a weak signal produced by a strong source at a different wavelength, and the lowest limit of detection is determined by the signal-to-noise ratio (SNR) of the signal. If a portion of the source power is detected by the CCD, it will add to the noise and worsen the sensor sensitivity. A source rejection filter whose passband matches the signal spectrum but blocks the source wavelength will significantly increase the SNR (see Note 4). Such filters are normally dielectric-layer filters and should be placed between the imaging lenses or on the input side of a single imaging lens where the

246

Christensen and Herron

fluorescence is most collimated (see Note 5). Figure 5 shows the transmission spectra of a typical source isolation filter and two choices of source rejection filters used with Alexa Fluor 660 dye. 1.5. CCD Camera 1.5.1. CCD vs. CMOS

Due to recent technological advances in the fabrication of CMOS imagers, they have moved closer to the low-noise specifications of CCD devices (16), but for lowest noise cooled applications, CCD chips still provide the best performance.

1.5.2. Resolution (Pixel Number and Size)

The appropriate CCD pixel size and number (and therefore cost) are determined by the desired size and number of discrete zones on the biosensor surface. It is wise to choose a CCD chip with several times (10–100) more pixels than the number of sensor zones because (1) the edges of the zones will not need to line up exactly with the pixel edges since several pixels are assigned to each zone; and (2) a large number of pixels allow flexibility in a possible redesign of the layout of the zones. The overall dimension of the CCD is matched with the sensor size by the magnification M of the imaging optics as given in Eq. 1; each pixel is then responsive to a region on the sensor plane 1/M as large in each dimension as the pixel.

1.5.3. Noise Sources

The ultimate sensitivity of a biosensor is often measured by its minimum detectable concentration, MDC, which is defined as the lowest analyte concentration distinguishable from zero with a 100 Alexa 660 emission

80

690ALP filter

Transmisson (%)

655WB20 filter 60

710AF40 filter 40

658-nm laser line

20

0 600

650 Separation

700

750

800

Wavelength (nm)

Fig. 5. Example spectra of a source isolation filter (here Omega 655WB20 bandpass) and two choices of source rejection filters (690ALP longpass and 710AF40 bandpass) that are appropriate to use with the Alexa Fluor 660 dye.

Optical System Design for Biosensors Based on CCD Detection

247

95% confidence level. If σc is the standard deviation (in concentration units) of the signal at zero concentration, then MDC = 2σ c .

(3)

Several noise sources contribute to the zero-concentration signal, and their variation in time – as given by their standard deviation – should be as low as possible, as discussed next. Readout Noise

The photons absorbed by each CCD pixel during the exposure period (integration period) result in a proportional number of electrons stored in the pixel’s well. When these electrons are read out by the charge-to-voltage amplifier, a small amount of variability, called readout noise, is introduced in the voltage, independent of the strength of the signal and integration time. Readout noise is specified in units of equivalent electrons; a typical value is 3–10 electrons. This noise component is usually small compared to other noise sources and can be reduced in importance by binning (i.e., combining the charges of groups of pixels into “superpixels” before reading them out) since the fixed readout noise contribution will then be a smaller percentage of the signal.

Thermal Noise

This voltage variation, also called Johnson noise, is due to the random nature of the electron passage through various conductive components as influenced by thermal vibrations. Its variance (the square of standard deviation) is proportional to the absolute temperature T of the conducting components as follows (17): s T2 = 4kT Df / R,

(4)

where k is the Boltzmann constant, ∆f is the electronic bandwidth of the sensing system, and R is the resistance of the particular component. The impact of thermal noise can be lessened by preamplification in the CCD chip before subsequent amplification. Its contribution is usually less than dark-current shot noise. Dark Current

If no photons are incident on the CCD, some electrons will still be generated in each pixel by thermal energy – the dark current. Variations in this dark current lead to noise in two ways:

Dark Current Drift

If the temperature of the CCD changes in time, the subsequent signal component from the dark current will also change, leading to uncertainty in the sensor measurement. However, if the drift is relatively slow, a CCD dark frame can be obtained immediately before or after the measurement frame by turning off the source or shuttering the camera; this frame can then be subtracted from the measurement frame canceling the contribution of the dark current. It is important that the integration times are the same for the two frames.

248

Christensen and Herron

Dark-Current Shot Noise

Even if the temperature of the CCD were to remain constant, there are still unavoidable statistical variations in time – called shot noise – in the signal that the dark current produces. The time variation is rapid, with Gaussian statistical behavior (for moderate dark current levels) and is broadband (“white noise”). The variation due to this noise contribution is proportional to the amount of dark current Id according to s d2 = 2qI d Df ,

(5)

where q is the charge of an electron. The contribution from the dark-current shot noise can often be the major source of noise if the CCD is operated at room temperature. If the optical receiver allows wavelengths longer than the source to pass (as with sensors based on fluorescence), any fluorescent signal emitted by non-analyte components in the excitation light path – such as lenses, substrates or waveguides – will be combined and confused with the true signal. Such autofluorescence is proportional to the strength of the source light and is generally greater for shorter wavelength excitation. Pure materials (fused silica, undoped glass, and plastics without additives) usually exhibit less autofluorescence than substances with fillers and dopants. Figure 6 shows the strength of autofluorescence in some examples of optical materials excited at 658 nm. Most glasses and polystyrene exhibit moderately low autofluorescence; 12000 Fluorescence Intensity (AU)

^ 4x

10000

^ 4x

8000 6000 4000 2000

Polycarbonate

Polystyrene 2

Fused Silica

Glass 8

SF-11

Glass 7

Glass 6

Glass 4

Glass 2

0 Dow 612 Polystyrene

Autofluorescence of Optical Components

Fig. 6. Relative autofluorescence intensities for several samples of optical materials used in biosensing applications. The autofluorescence was excited by a 658-nm source and detected through a 40-nm wide bandpass filter centered at 690 nm.

Optical System Design for Biosensors Based on CCD Detection

249

fused silica material has the lowest, but SF-11 high-index glass and polycarbonate show considerably more. If not carefully controlled, autofluorescence can be a significant noise source in two ways, as explained here. Photobleaching of Autofluorescence

Immediately upon exposure of the offending component to the source beam, its autofluorescence may start to decay due to photobleaching, reversibly or irreversibly. Time constants in the order of a few seconds to several minutes are common. This represents a drift in the total optical signal and can be confused with the signal change due to analyte interaction.

Autofluorescence Shot Noise

As in the case of dark current, fluctuations in the autofluorescence signal will contribute to the noise. The variance of this shot noise component is given by s a2 = 2qI a Df ,

(6)

where Ia is the portion of the output current caused by autofluorescence. Obviously, the larger the background autofluorescence, the stronger is this noise component. Total Noise

Under the typical assumption that the noise sources described above are statistically independent and Gaussian in nature, the variance of the total noise (neglecting readout noise) is the sum of the individual contributions, or s n2 = s T2 + s d2 + s a2 .

(7)

To determine the influence of total noise on sensor sensitivity, the standard deviation of the total noise in units of amperes (the square root of Eq. 7) must be divided by a scale factor K that relates the signal in amperes to the analyte concentration in units of, say, µg/cm3. K is found by calibrating the sensor. Then sc = sn / K,

(8)

and the MDC is found by Eq. 3. 1.5.4. Cooling the CCD

If autofluorescence is minimized and filters are employed to block the out-of-band light, usually the major noise source of a room-temperature system is the shot noise due to the dark current. Then it is expedient to cool the camera to reduce this noise. For CCDs fabricated from silicon (by far the most widely used in the visible range), the dark current is cut approximately in half for every 6°C reduction in chip temperature. Cooling even to a modest amount, say to −20°C, significantly reduces the dark current and its corresponding shot noise (given by Eq. 5). At

250

Christensen and Herron

that point, the other noise sources usually become the limiting factor (see Notes 6 and 7). Although cooling the CCD requires additional componentssuch as a thermoelectric (TE) cooler (a.k.a. Peltier module) inside the CCD camera along with accessories to avoid water vapor condensation on the camera window in humid environments (done with a dry-gas spacer) and a larger power supply to provide a cooling current (see Note 8), the reduction in noise and the improvement in MDC are often worth the effort. For example, Fig. 7 shows how cooling the CCD reduces the statistical uncertainty in the signal caused by shot noise from the dark current. At room temperature (upper row) the spread due to shot noise of the dark current is large enough to prevent the separation of a very dark background signal (seen on the right half of each image) from a neighboring very weak fluorescence signal (seen on the left half of each image), as evidenced by the histogram in the top row. But when the CCD is cooled to −20°C (bottom row), the fluorescence signal can be clearly distinguished from the dark background signal, as evidenced by the histogram in the bottom row. The result is a significant improvement in MDC.

a)

weak signal

dark background

weak signal

dark background

dark background

weak signal

dark background

weak signal

b)

Fig. 7. The effect on dark-current shot noise by cooling the CCD. (a) Top row: Results for CCD camera at room temperature (24°C); (b) Bottom row: Results for CCD camera cooled to −20°C. On the left are images taken by an SBIG ST-6A camera with 5-s exposure. Within each image, the right half is a dark background region containing no signal and the left half is a region of weak fluorescence (autofluorescence from a polystyrene waveguide). On the right are histograms of the pixel intensity of each entire image. Only when the camera is cooled can the weak signal be statistically distinguished from the background.

Optical System Design for Biosensors Based on CCD Detection

251

2. Materials 2.1. Fluorescent Dye Labeling

1. Alexa Fluor® 660 (AF660) dye (Invitrogen). Absorption and emission spectra are shown in Fig. 3. 2. Antibody solution. 0.1 M carbonate-bicarbonate buffer, pH 9.3. 3. Dye dissolving solution. dimethylformamide at room temperature. 4. Separation column. PD-10 desalting column (GE Healthcare).

2.2. Source

1. Semiconductor laser diode. Hitachi HL6501MG (5.6-mm diameter can) emitting at 658 nm, closely matching the peak absorption of the dye. Output power is 35 mW (CW) at 45 mA typical. Imax is 70 mA. (See Note 9). 2. Laser driver. Wavelength Electronics LDD200-2 M. 3. TE cooler. Melcor CP1.0-31-08L-1 (15 × 15-mm). 4. TE PID controller. Wavelength Electronics PID-1500. 5. TE controller thermistor. Wavelength Electronics TCS-610 (10 kΩ at 25°C).

2.3. Lenses

1. Collimating lens. Optima Precision 307-8040-780 (8.0-mm focal length, 4.80-mm clear aperture). 2. First beam shaping lens. Optima Precision line-generating optic 417-9030-660. 3. First beam shaping lens holder. Optima Precision 210-0009-000. 4. Second beam shaping lens. Edmund Scientific 32730 (50-mm focal length, 25-mm wide) cut to 30-mm length with a diamond saw, in custom-made aluminum lens holder. 5. Lens holder posts. Newport MSP-1 and MPH-05C. 6. Sliding stages. Thorlabs RC1. 7. Mounting rail. Thorlabs RLA0600 (low-profile).

2.4. Beam Directing Mirrors

1. First surface mirrors (2 ea). Edmund Scientific 43873 (28 × 28-mm). 2. Mirror mounts. Newport MM-1 (1-in.). 3. Mirror mount posts and holder. Newport SP-2 and Thorlabs PH1-ST.

2.5. Sensor Module

1. Planar Waveguide – Molded from Dow 612 polystyrene (thickness 500 µm, planar area approximately 25 × 30 mm). An integral cylindrical lens (focal length 8.5 mm, tilted at 20°) is molded onto one end of the waveguide. 2. Flowcell. Custom-made black anodized aluminum.

252

Christensen and Herron

2.6. Fluorescence Bending Mirror

1. First surface mirror. Newport 10 SD510 (flat 50 × 50-mm). 2. Mirror mount. Thorlabs KCl (2-in.). 3. Mirror mount posts and holder. Newport SP-1 and Thorlabs PH 1-ST.

2.7. Collection/ Imaging Lens

1. Camera lens. Nikon Micro-Nikkor 55 mm (f/2.8).

2.8. Laser Rejection Filter

1. Filter. Omega 700DF35 (35-nm wide bandpass filter centered at 700 nm, 1 in. diameter).

2.9. CCD Camera

1. Camera. SBIG ST-6A (with TC-241 CCD chip incorporating 375 × 242 pixels, each 23 × 27 µm. Overall array dimensions 6.8 × 8.8 mm). The CCD chip has readout noise of 30 electrons rms and a full well capacity of 400,000 electrons. A/D resolution is 16 bits.

3. Methods The design tradeoffs discussed in the Introduction are illustrated by the sample fluorescent immunoassay system shown in Fig. 2 and described below. The steps for fabricating, assembling and operating this system are as follows. 3.1. Fluorescent Dye Labeling Protocol

To avoid background autofluorescence – especially from specimens of whole blood – and to take advantage of high-power red laser diodes, Alexa Fluor 660 (AF660) is chosen as the dye for labeling the reporter antibody. AF660 contains an NHS-ester group that reacts at basic pH with the antibody’s primary amino groups including N-terminal amino and lysine ε-amino groups. 1. One milligram of antibody is dissolved in 0.3–0.5 mL of the carbonate–bicarbonate buffer. 2. A tenfold molar excess of AF660 (dissolved in 30–50 µL of dimethylformamide) is added to the antibody solution, mixed thoroughly, and allowed to react for 60 min at room temperature with agitation. 3. The labeled antibody (“conjugate”) is separated from unreacted dye using a PD-10 desalting column and equilibrated in phosphate buffered saline. 4. Both the concentration and degree of the conjugate labeling are determined from its absorption spectra using the Beer– Lambert law. AF660 absorbs maximally near 663 nm, but also exhibits some absorption at 280 nm. Thus, in calculating antibody concentrations, the absorbance at 280 nm is corrected

Optical System Design for Biosensors Based on CCD Detection

253

for the dye’s absorption at this wavelength as described in the package insert for AF660: [Ab]=

A280 nm - (0.1 × A663nm ) (9)

M e Ab M

using a value of 202,500 M−1 cm−1 for Ab, the antibody’s molar extinction coefficient. The molar concentration of AF660 is calculated from [AF660 ] =

A663nm

(10)

M e AF 660 M

using a value of 110,000 M−1 cm−1 for AF660 , AF660’s molar extinction coefficient. Degree of labeling (DL) is then determined from the ratio DL =

[AF660 ] [Ab]

(11)

This protocol typically gives DL values between 1 and 4 depending on the particular antibody being labeled. 1. A Hitachi laser diode is mounted on a custom-made machined aluminum housing shown in close-up view in Fig. 8. The laser

3.2. Source

Electrical Connector Laser diode can Collimating lens (inside plate) Thermistor Thermoelectric cooler Back plate (heat sink) Cooled plate

Line-generating lens in cylindrical housing

Fig. 8. Close-up photograph of a custom-made laser diode housing, including cooling apparatus, collimating lens, and the first (a line-generating lens) of the two beam shaping lenses.

254

Christensen and Herron

is driven in constant-power mode by a Wavelength Electronics power supply mounted on the rear of the system’s backplate. The +12 VDC input power for this module is tapped from the CCD camera’s electronics. 2. Cooling for the laser is provided by a 15 × 15-mm Melcor TE cooler sandwiched between the thin cooled plate to which the laser can is attached (held in a recessed hole by two brass 2–56 screws around its edges) and the plate forming the back of the housing, which acts as a heat sink. Silicone-free heat sink compound (Tech Spray 1978-DP) is spread thinly on both sides of the TE cooler before the plates are bolted together with 4–40 screws passing through insulating bushings in the thin laser plate. The TE module is driven by a closed-loop PID controller, also mounted on the rear of the system’s backplate, at a nominal steady-state current of about 700 mA to keep the laser temperature near +20°C. A small thermistor is inserted into a hole on the edge of the laser-mounting plate to provide temperature feedback control. A 9-pin D-sub connector serves as the electrical interface to the laser. 3.3. Lenses

1. Collimating lens. An Optima Precision molded glass aspheric lens with a fine-pitch threaded collar (not visible in the photo of Fig. 8) is threaded approximately midway into the lasermounting plate with its axis lined up with the laser and adjusted until the output beam remains collimated (after the following optics have been temporarily removed) when observed on a distant plane. 2. Beam shaping lenses. Two lenses (see Fig. 2) shape the resulting laser beam into a sheet beam approximately 2 × 20 mm in cross-section as required for coupling into the planar waveguide sensor. The first lens is an Optima Precision linegenerating optic mounted by hot glue in a threaded cylindrical holder that is screwed into an adjustable bar attached to the front plate of the laser housing. The bar is swung left or right slightly until the line beam that is formed is centered on the next lens; this adjusts for the slight mismatch between the center axis of the laser and the line-generating lens. The second lens is an Edmund Scientific recollimating cylindrical lens (to form the output beam into a sheet beam) mounted in a black anodized housing. Both the laser housing and this cylindrical lens assembly are mounted on posts on sliding stages clamped to a low-profile rail, and are adjusted in spacing until the resulting sheet beam appears collimated as it enters the location of the waveguide sensor.

3.4. Beam Directing Mirrors

1. Two 28 × 28-mm first surface mirrors are attached by doublestick tape to 1-in. mirror mounts post-mounted at appropriate locations on the backplane to steer the sheet beam into the

Optical System Design for Biosensors Based on CCD Detection

255

waveguide, and are adjusted until the beam enters the sensor module at a 20° angle with respect to the waveguide planar surface (horizontal). 3.5. Sensor Module

1. Planar waveguide. The waveguide is molded in the form of a thin rectangular sheet of 500 µm thickness from Dow 612 polystyrene. To focus the incoming sheet beam reliably into this thin waveguide (with some forgiveness in exact sensor positioning), an integral cylindrical lens is molded into one end of the waveguide. 2. Flowcell. A custom-made black anodized aluminum flowcell is clamped around the waveguide. Three fluid channels, each with an inlet port on one end and an outlet port on the other, are spatially defined by machined recessions in the upper plate and by openings in a gasket between the plate and the waveguide; the compliant gasket is lined with a low-index Teflon® FEP film (n = 1.35, close to n = 1.33 of water) to avoid outcoupling of the waveguided light by the gasket material.

3.6. Fluorescence Bending Mirror

1. A 50 × 50-mm first surface mirror is attached with doublestick tape to a 2-in. mirror mount and post-mounted to the backplane; it is adjusted to 45° for the purpose of folding the optical collection path and making the layout more compact.

3.7. Collection/ Imaging Lens

1. A high-quality Nikon Micro-Nikkor camera lens is mounted in front of the CCD camera. As it is a macrolens, its focus is adjusted to fall on the surface of the waveguide located 198 mm in front of the lens housing. Its iris is adjusted to full opening (f/2.8).

3.8. Laser Rejection Filter

1. A 35-nm wide bandpass filter centered at 700 nm is mounted with hot glue to a black cardboard ring and slipped inside the barrel of the collection/imaging lens, where it is held by black tape (see Notes 10–12).

3.9. CCD Camera

The amateur astronomy marketplace provides an economically attractive source of CCD cameras, many of them cooled to allow long (several minutes to hours) integration times. 1. An SBIG ST-6A camera head with attached lens is mounted in front of the reader’s plate; its electronics box is mounted on the rear (see Note 13). 2. The CCD chip inside the camera is cooled internally with a TE cooler and fan, and is run continuously at −20°C to reduce dark current. The camera housing has a mechanical shutter under computer control such that 5-s dark frames are captured in close time sequence to 5-s signal frames for darkframe subtraction and correction. The integration time and binning modes are controlled by computer software. As the

256

Christensen and Herron

Fig. 9. Image of fluorescence emitted (after 280 s of incubation) by sandwiches of [dyelabeled anti-hCG reporter Ab + targeted hCG protein + anti-hCG Ab] immobilized on the surface of a polystyrene planar waveguide, as imaged through the bottom of the threechannel flowcell shown in Fig. 2. Excitation laser light enters the waveguide through a molded-on lens on the right side. A CCD camera with attached lens and source rejection filter images the fluorescence that is emitted at approximately right angles through the waveguide bottom. The concentration of hCG increases stepwise from the bottom channel to the top channel; the signal associated with each channel is summed inside the dotted areas. Voids in the liquid are visible near the inlet and outlet ports but outside the summing areas.

waveguide sensor has three relatively large sensing zones, a 3 × 2 binning mode is selected to reduce the impact of readout noise and to increase readout speed. The antiblooming feature is turned off as it degrades the linearity of the photonto-output current response. Figure 9 shows a typical image obtained by this CCD camera of the fluorescence emission during a test of the sensitivity of the planar waveguide biosensor for detecting hCG, where each channel of the three-channel flowcell contains a different concentration of the analyte.

4. Notes 1. The autofluorescence (a.k.a. native or self-fluorescence) exhibited by proteins under UV excitation may eliminate the need for a separate labeling step, simplifying the biosensor. But the disadvantages of using UV excitation, as listed in the text, counterbalance this advantage. 2. When the fluorescent dye is selected, it is important to choose one with a large Stokes shift such that the peak wavelength of

Optical System Design for Biosensors Based on CCD Detection

257

emission is as far removed from the excitation wavelength as is possible. This moves the source wavelength away from the short-wavelength edge of the source rejection filter, reducing the tendency for source leakage through the filter. 3. The wavelength choice also affects the amount of autofluorescence from the optical components in the path of the excitation beam, for instance from the lenses or the waveguide in the sample in Fig. 2. Longer wavelengths (red through IR) often result in less autofluorescence and less corresponding noise. 4. The source rejection filter will also block other unwanted light (e.g., room lighting) and may be useful for that purpose alone. 5. Dielectric-layer filters are designed for collimated light, that is, light whose rays are confined in a path perpendicular to the face of the filter. For rays that pass at angles off of the perpendicular, the passband will broaden out and move to shorter wavelengths (the effect is proportional to the cosine of the angle). While there is no place in a low f# imaging system where the light rays are strictly perpendicular to the filter plane, the rays are most closely collimated in the region between the two imaging lenses, as Fig. 1 shows, or if only one lens is used, on the input side of this lens (assuming M < 1). 6. The lowest possible limit on noise is that of the shot noise attributed to the actual signal itself. This is the irreducible noise floor and it is the goal of all sensors to reduce all other noise sources until only signal shot noise remains (“shot noise limited”). 7. Non-specific binding (NSB) is also a “noise” source for immunoassays and molecular recognition assays since it confounds the true analyte measurement. But this is a surface chemistry issue outside the realm of purely optical design. 8. It is difficult to provide this relatively large amount of a current, even at a few volts, with a practical-sized battery over a reasonable time. Thus battery-operated instruments with cooled CCDs remain problematic. 9. There are several choices of lasers in this wavelength range at various powers. Thorlabs Inc. is a widely used source of laser diodes in cans. 10. It is not physically possible to put this filter between the elements of the Nikon lens where the fluorescence light rays are best collimated for optimum filter performance, so it was affixed to the input side of the lens. 11. A newer filter with similar wavelength specifications, Omega 3RD690-740, may provide better laser rejection. 12. A source isolation filter was not needed in this particular application since very little laser light, being trapped by

258

Christensen and Herron

total internal reflection inside the waveguide, finds its way to the CCD. Only a minimal amount of scattering from the waveguide was observed, and that was largely rejected by the laser rejection filter. 13. Newer amateur astronomy CCD cameras contain the electronics in the head, reducing the required footprint, and most communicate with the computer via USB 2.0 or Firewire. References 1. Halasey, S. and Park, R. (Sept. 2002) Making sense of the global IVD market. IVD Technology, 29 2. See, for example, http://www.affymetrix.com 3. Rowe, C. A. et al. (1999) Array biosensor for simultaneous identification of bacterial, viral and protein analytes. Anal. Chem. 71, 3846–3852 4. Woo, J. and Henry, J. B. (1994) The advance of technology as a prelude to the laboratory of the twenty-first century. Clin. Lab. Med. 14, 459–471 5. Sapsford, K. E. et al. (2004) Fluorescencebased array biosensors for the detection of biohazards. J. Appl. Microbiol. 96, 47–58 6. Wolfbeis, O. S. (2002) Fiber-optic chemical sensors and biosensors. Anal. Chem. 74, 2663–2677 7. See, for example, http://www.astrovid.com 8. Howell, S. B. (2006) Handbook of CCD Astronomy, Second Edition. Cambridge University Press, Cambridge 9. Herron, J. N. et al. (1998) Rapid clinical diagnostics arrays using injection-molded planar waveguides. SPIE 3259, 54–64

10. Tolley, S. et al. (2003) Single-chain polymorphism analysis of long QT syndrome using planar waveguide fluorescent biosensors. Anal. Biochem. 315, 223–237 11. Herron, J. N. et al. (Oct. 14, 1997) Apparatus and methods for multi-analyte homogeneous fluoroimmunoassays. US Patent No. 5,677,196 12. Males, R. G. et al. (2001) Cardiac markers and point-of-care testing: a perfect fit. Crit. Care Nurs. 24, 54–61 13. Fivash, M. et al. (1998) BIAcore for macromolecular interaction. Curr. Opin. Biotechnol. 9, 97–101 14. Zhang, X. et al. (2005) Surface-enhanced Raman spectroscopy biosensors: excitation spectroscopy for optimization of substrates fabricated by nanosphere lithography. IEEE Proc. Nanobiotechnol. 152, 195–206 15. Williams, C. S. and Becklund, O. A. (1972) Optics: A Short Course for Engineers and Scientists. Wiley-Interscience, New York. Chap. 6 16. Litwiller, D. (Jan. 2001) CCD vs. CMOS: facts and fiction. Photonics Spectra, 154–158 17. Agrawal, G. P. (1992) Fiber-Optic Communications Systems. Wiley-Interscience, New York. Chap. 4

Chapter 14 A Simple Portable Electroluminescence Illumination-Based CCD Detector Yordan Kostov, Nikolay Sergeev, Sean Wilson, Keith E. Herold, and Avraham Rasooly Summary In this chapter we describe a simple and relatively inexpensive Electroluminescence (EL) illumination and charged-coupled device (CCD) camera (EL-CCD) based detector for monitoring florescence and colorimetric assays. The portable battery-operated florescence detector includes an EL panel for fluorogenic excitation at 490 nm, a cooled CCD digital camera to monitor emission at 523 nm, filters and a close up lens. The detector system is controlled by a laptop computer for camera operation, image acquisition and analysis. The system was tested using a fluorogenic peptide substrate (SNAP-25) for botulinum neurotoxin serotype A (BoNT-A) labeled with FITC. The level of detection of the system was found to be 1.25 nM of the peptide, similar to the detection level of a commercial photomultiplerbased plate fluorometer. The multichannel EL-CCD was used with an assay plate capable of testing nine samples simultaneously in 1 min at this detection level. The portable system is small and is operated by a 12 V source. The modular detector was designed with easily interchangeable ELs, filters and lenses and can be used and adapted for a wide variety of florescence and colorimetric assays, florescence labels and assay formats. Key words: Electroluminescence, CCD, Charge coupled device, Fluorescence, Fluorometer, Botulinum neurotoxin, Multichannel detector, Florescence assays, SNAP-25.

1. Introduction In recent years, various approaches for fluorescent detection have been developed, including the use of photodiodes (1–4), photomultipliers (5–10) and charged-coupled device (CCD) (10–18) based detection. Unlike most photodiodes, and photomultiplier detectors, which are inherently spot detectors for analysis Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_14

259

260

Kostov et al.

of limited areas, CCD-based detectors typically have a wider field of view. This makes them ideal for multichannel detectors. One notable example of such an approach is the array biosensor, which used a CCD detector for multiplex detection of toxins and microbial pathogens and their toxins (11, 16, 19, 20). Sensitivity is a critical issue with biodetection, and although CCDs in general are less sensitive than photomultipliers, low source intensity can be compensated by long exposure time. Low light sensitivity can be further enhanced by minimizing thermal noise by cooling the CCD detector. 1.1. Light Sources for Fluorescence Detection

The intensity and spectrum of the excitation light source are critical in establishing high sensitivity for fluorescence detection. Most current fluorescence detectors utilize high intensity, high power and bulky bench-top excitation sources, including tungsten, mercury, or xenon lamps. These light sources are usually expensive, nondurable, require high voltage and limit the portability of the device. Diode lasers (11, 16, 19, 20) and light-emitting diodes (LED) (21–24) have been used as excitation sources especially for portable real time PCR detection (25), since they are low cost, have a small footprint, are simple to operate, durable and consume relatively little power. These characteristics make diode lasers and LEDs appealing excitation light sources for portable detectors. For an array detection system, involving multiple sensing points on a surface, spatially uniform lighting is required. Spatially uniform incident intensity on the array is difficult to achieve because such light sources are highly directional. Options include use of a line generator with a laser, the use of an optical waveguide (11, 16, 19, 20), and other optical systems such as mirrors, diffusers and lenses. All these methods introduce complexity into the design and many have alignment issues that impact the excitation uniformity in the sensing plane. On the detection side, photodiodes and photomultiplier detectors are effective for applications where a discrete fluorescent source is to be detected. For array sensors, multiple detectors can be mounted together, such that one sensor detects the light from each spot on the array. Coverage of a broad area requires either optical systems or a mechanical moving stage. A CCD detector is a packaged alternative that is available with many flexible optical options for imaging a surface area.

1.2. Electroluminescence

Electroluminescence (EL) is produced by a semiconductor surface that emits light in response to an alternating electric current. In EL, electrons that are excited into a higher energy state leave “holes” and when dropping back to the ground state, the excited electrons release their energy as photons upon recombining with their “holes”. The spectrum of the light emitted by EL depends on the electroluminescent materials. Although EL materials

A Simple Portable Electroluminescence Illumination-Based CCD Detector

261

are used widely for signs and instrument dial illumination (e.g. watches, clock radio, personal organizers, cockpit instrumentation), they are not currently used for fluorescence excitation and biodetection. Nevertheless, they have great potential for biodetection applications given that EL features include uniform spatial illumination, wide variety of wavelength, low cost, high efficiency, small footprint, simple operation, durability and low power consumption. However, EL generally emits far less light than other light sources used for florescence excitation such as lamps, lasers or LEDs and therefore the use of EL typically requires longer exposure time to achieve the same sensitivity. In this chapter we combine EL excitation with CCD (EL-CCD) based fluorescence detection to obtain a simple portable detector.

2. Materials

2.1. Electroluminescence-Based CCD Detector

1. Cooled high sensitivity black and white CCD camera Atik 16 (Adirondack Video Astronomy, Hudson Falls, NY). 2. 5 mm Pentax extension tube (Spytown, Utopia, NY). 3. Pentax 12 mm f1.2 lens (Spytown, Utopia, NY). 4. Emission filter A, 535/50/75 (Intor, Soccorro, NM). 5. Emission filter B (EmF-B) cat # D480/30x (Chroma Technology Corp Rockingham, VT). 6. Excitation filter A, (optional) BG 12 (Edmund Optics Inc, Barrington, NJ). 7. Excitation filter B (ExF-B) cat# D535/40 m, (Chroma Technology Corp Rockingham, VT). 8. Blue Electroluminescence panel (Being Seen Technologies, Bridgewater, MA). 9. 110 V AC inverter (Being Seen Technologies, Bridgewater, MA).

2.2. Computer Control and Data Analysis

1. PC-computer (laptop or desktop) with USB port. 2. Image analysis software ImageJ-free NIH software http:// rsb.info.nih.gov/ij/download.html 3. Data analysis software Excel (Microsoft, Redmond, WA) and Sigma plot (Sigma plot, Ashburn, VA).

2.3. Chemicals and Reagents

1. Fluorogenic peptide substrate for botulinum neurotoxin serotype A SNAP-25 labeled with FITC (List Biological Laboratories, Campbell, CA). 2. HEPES, BSA and TWEEN (Sigma-Aldrich, St. Louis, MO).

262

Kostov et al.

2.4. Assay Plate Fabrication

1. Black 3.2 mm acrylic (Piedmont Plastics, Beltsville, MD). 2. Clear 0.25 mm polycarbonate film (Piedmont Plastics, Beltsville, MD). 3. 3 M 9770 adhesive transfer Tape (Piedmont Plastics, Beltsville, MD). 4. Epilog Legend CO2 65 W cutter (Epilog, Golden, CO). 5. ModuLam 130 laminator (Think & Tinker Ltd, Palmer Lake, CO).

2.5. Flurocence Measurement

1. Microplate fluorometer (MolecularDevices, Sunnyvale, CA).

3. Methods 3.1. Selection of CCD Detector, Optical Elements, Power Source and Fluidics

1. Camera selection: A black and white astronomy camera with cooled high sensitivity CCD (Atik 16) was used as a CCD detector (Fig. 1). The camera employs a Sony ICX-429ALL CCD with 752 × 582 pixel resolution. The camera is equipped with a 16-bit analog to digital converter, enabling a dynamic range of 65,500 levels of grayscale. Peltier thermo-electric cooling is used to maintain the CCD at 25 K below ambient temperature so that a much lower thermal noise is generated on the CCD. The camera is equipped with a magnetic levitating fan, to minimize any residual vibration in long exposures. 2. Optical and illumination system: The optical system includes a Pentax 12 mm f1.2 CCD lens connected to the camera with a C adapter, and a 5 mm extension tube that enables the camera to be focused at a short distance (30 mm) and to cover a small field (15 × 20 mm). The large aperture lens permits low-light exposure with the tradeoff that the depth of the field is very small, making the camera focus on the sensing plane critical. A blue electroluminescence panel was used for excitation of the fluorogenic dye FITC that fluoresces at a wavelength of 523 nm when excited with 490 nm light. To minimize green light from the EL panel, which would limit the detection sensitivity, two excitation band pass filters were used (placed between the EL panel and the test section). The first is a wide band pass filter with a center wavelength at 440 nm and full bandwidth ±40 nm (BG12, Schott), and the second is a narrow band pass filter (D480/30x) with center wavelength at 480 nm and full bandwidth ±15 nm. The combined effect of these filters is to block frequencies in the emission spectrum range while allowing frequencies in the excitation range. Similarly, filters were placed between the test section and the CCD camera to block all but the fluorescent signal from

A Simple Portable Electroluminescence Illumination-Based CCD Detector

A

263

B 1

2 3 4 5 6 7 8 9 10 11

C

7

10/11

9

8

Fig. 1. EL-CCD based fluorometer: (a) Schematic configuration of the detector, (b) The actual fluorescence detection system, (c) EL filter enclosure. (1) CCD-EL enclosure, (2) Atic-16 CCD camera, (3) Excitation green filter A, (4) 5 mm extension tube, (5) 110 mm Pentax f1.2 lens, (6) Excitation green filter B, (7) EL-enclosure, (8) Detection plate, (9) Emission blue filter A, (10) Emission blue filter B and (11) EL panel.

the FITC. Two band pass filters were used. The first (Intor 535/50/75) has a center wavelength of 535 nm and band width ±25 nm and the second (D535/40 m) has a center wavelength of 535 nm with full bandwidth of ±20 nm. 3. Power system: As a portable system, the power requirements have been designed to be minimal, with only 800 mA of consumption, and 12 V DC input (power supply or battery) used for both the EL and the CCD, making the system mobile and practical for use in the field. The EL requires 120 VAC which is supplied by an inverter which converts 12 VDC input to 120 VAC output. The power supply is housed separately from the camera, thus minimizing electrical interference.

264

Kostov et al.

4. Assay plates: For the assay plates (Fig. 1c-8), a black acrylic is used to minimize light reflection. The polycarbonate material used did not show detectable autofluorescence and all materials used to fabricate the device did not appear to inhibit enzymatic reactions. 3.2. Assembling EL-CCD Detector

1. General detector assembly: The constructed EL-CCD is shown in Fig. 1. The CCD-based system consists of: a black plastic enclosure (Fig. 1a-1) which houses the system; the Atic-16 cooled CCD camera (Fig. 1a-2) is mounted in the enclosure with holes drilled on the top of the enclosure to enable air circulation for camera cooling. An emission (EmF-A) filter (Fig. 1a-3) is mounted on the CCD camera body with a 5 mm extension tube connected to the camera via a C mount adaptor (Fig. 1a-4). A 110 mm Pentax f1.2 lens is mounted on the extension (Fig. 1a-5) and a second filter, EmF-B is mounted on the lens (Fig. 1a-6). The EL-enclosure (Fig. 1a-7) includes the detection plate, excitation filters ExF-A and ExF-B, and an EL panel (as discussed in the following paragraph (2)). In general, the system was designed to be modular and versatile, so that various lenses, filters, light sources and fluidics systems can be used based on the detector application. The only fixed element in the detector enclosure is the CCD camera on which the emission filters are attached. Other elements of the system, including the detection plate, excitation filter ExF-A, ExF-B and EL panel are housed in an interchangeable EL filter enclosure. 2. The EL filter enclosure: The EL filter enclosure (Fig. 1c-7) is made of black acrylic with rails glued on each side, which enable the installation of the illumination card, filters and the assay plates. All acrylic parts of the EL filter enclosure were machined with the laser cutter (see Note 1). The illumination card contains the EL panel taped into an acrylic card with a rectangular excitation filter ExF-B taped directly on the top surface (Fig. 1c-10/11). The illumination card is mounted on the bottom rail of the EL filter enclosure. The circular excitation filter ExF-A is mounted on a round hole machined in an acrylic card. This filter card is mounted in the EL filter enclosure (Fig. 1c-9) above the illumination card. The assay plates (Fig. 1c-8) are mounted on the top rail of the enclosure (see Note 2). The EL filter enclosure was designed to be modular and to allow rapid changes of EL panels, filters and assay plates. The compact design of the EL filter enclosure minimizes light loss and internal light reflection (see Notes 3–4) and enables the components to slide easily into the rails, each in a fixed location. For detecting low signals, it is critical to reduce optical interference (low level light directly from the

A Simple Portable Electroluminescence Illumination-Based CCD Detector

265

EL excitation source, and light reflection). Although we used a blue EL, it emitted low level light in the green spectrum, creating interference that could limit the sensitivity of detection. To minimize this problem, two excitation filters were placed in the EL-enclosure (Fig. 1c-9, c-10) to eliminate light in the FITC emission spectrum from the EL excitation source. 3. Background light level and light uniformity: To block the EL’s blue light from reaching the camera directly (see Note 5), two green emission filters were used. One was placed on the CCD (Fig. 1a-3) and one on the lens (Fig. 1a-6). The background level, without any fluorescent sample, was measured with a 3 min exposure and was found to be very low (see Note 6) demonstrating the effectiveness of the filter system as shown in Fig. 2. The highlight around the rim of each well was caused by scattered light from the surface (see Note 7). Quantitative analysis of this background signal was done using ImageJ software. Excitation light uniformity is essential for accurate florescence measurements (see Note 8). Quantitative analysis by Image J software (Fig. 3) demonstrates various analysis methods for light measurements (see Image and data analysis below). 3.3. Assay Plate Fabrication

1. Micromachining: The assay plate was made of 3.2 or 1.0 mm thick black acrylic. The sample wells (2 mm diameter) were machined with a computer controlled 65 W Epilog Legend CO2 laser system (see Notes 1 and 9). The wells allow analysis of 20 µL samples. Smaller holes make the loading of the sample less reproducible and thus the signal is less reproducible. 2. Bonding: The polycarbonate bottom was bonded to the acrylic with double sided pressure-sensitive adhesive transfer tape (see Note 10). The transfer tape was bonded to the acrylic prior to machining so that the adhesive was removed in the well region. Thus, the sample has minimal contact with the adhesive. Device layers were aligned and then bonded by

Fig. 2. Background light measurement of the EL-CCD based fluorometer. An image of filtered EL light measured by the CCD without any sample.

266

Kostov et al.

passing the assembly through the laminating machine to eliminate air bubbles and to promote uniform bonding (Fig. 6). 3.4. Image and Data Analysis

The exposure time required for low signals was found to be in the range 30–120 s (see Note 11). The camera images can be analyzed by ImageJ software (see Note 12). Using the CCD camera enables sophisticated data analysis; Fig. 3 demonstrates such capability, using the same concentration of peptide (100 nM of SNAP-25) in all the spots (Fig. 3a). In addition to calculating the intensity of every spot, the camera enables a 2D image analysis of the intensity of each spot in a row or column (Fig. 3c) and 3D presentation (2D representing the position within the spot and the third dimension representing the intensity at each point) shown in Fig. 3b. For detailed spot analysis, the CCD image was exported to an Excel spreadsheet and to Sigma Plot to plot and analyze the data. Several analyses were conducted, including subtracting background level, calculating signal to noise ratio and performing linear regression.

Fig. 3. Light measurement and analysis. An image of fluorescence measured by the EL-CCD system with 100 nM of SNAP-25. (a) actual image, (b) 3D ImageJ data analysis, (c) 2D ImageJ data analysis.

A Simple Portable Electroluminescence Illumination-Based CCD Detector

1. Dynamic range and linearity of EL-CCD: The dynamic range of the EL-CCD measurement is dependent on the exposure time. For long exposures (e.g. 2 min) used to detect low signals, high signals tend to saturate the detector and cannot be quantitated accurately. In general, the dynamic range is 2logs at each exposure setting, so a wide dynamic range of at least 4log units can be achieved using multiple exposure times. We performed a linear regression analysis (Fig. 4) of SNAP-25 labeled with FITC (see below) which suggests that in the detected range of 1.6–200 nM of SNAP-25, the signal to noise ratio seems to be linear for the EL-CCD (Fig. 4a) similar to the linearity of the fluorometer (Fig. 4b). 2. The sensitivity of EL-CCD: To assess the detector’s performance, we used a fluorogenic peptide substrate of BoNT-A (SNAP-25) labeled with FITC that fluoresces at a wavelength of 523 nm when excited with 490 nm light. Serial dilutions of the peptide were analyzed by a commercial fluorometer and by our detector. Figure 5a shows an actual image for such an assay after a 1 min exposure. For the quantitation of the data, we used ImageJ software to calculate the intensity of every spot and subtracted the background value from the measurement. The limit of detection of EL-CCD is 1.6 nM of the substrate. A plot of the result is shown in Fig. 5b. The solid line is from the PMT and the dashed line is from the EL-CCD. The EL-CCD results are similar to the commercial fluorometer measurements. The signal/noise (S/N) ratio plot (the ratio of the sample well signal to the signal of a well with water) is shown in Fig. 5c, the solid line PMT and the dashed line is 6

A

5 4 S/N

3.5. EL-CCD Detection of Fluorogenic Peptide Substrate for Botulinum Neurotoxin Serotype a (Bont-A)

267

B 3 2 1 0 0

50

100

150

200

250

SNAP25 (nM)

Fig. 4. Dynamic range and linearity of EL-CCD: Signal to noise ratio for detection of various concentrations of FITC labeled SNAP25 peptide (a) EL-CCD measurement, exposure of 2 min and analyzed with ImageJ, (b) fluorometer measurement.

Kostov et al.

C B A 1

2

3

1600

Signal (above background)

1400 1200 1000 800 600 400 200 0 e2

S / N Ratio

268

e1

e0 0.1

1

10 SNAP25 (nM)

100

Fig. 5. The sensitivity of EL-CCD: EL-CD detection of various concentrations of FITC labeled SNAP25 peptide when excited with blue EL and detected with green filers measured for 1 min and analyzed by ImageJ. (a) CCD image of the assay, (b) Intensity plot of every spot after subtraction of the background value, (c) Signal/noise (S/N) ratio plot. The concentration of SNAP-25 peptide used in each well: A1 0 nM, A2 1.6 nM, A3 200 nM, B1 0.064 nM, B2 8 nM, B3 Empty, C1 0.32 nM, C2 40 nM, and C3 Empty. The solid line PMT based detection and the dashed line is EL-CCD.

A Simple Portable Electroluminescence Illumination-Based CCD Detector

269

EL-CCD. The S/N ratio of the EL-CCD device is similar to the fluorometer in measurements taken on the same samples (Fig. 4c). Moreover, the volume needed for the fluorometer is 100 µL while the EL detector requires only 20 µL. Thus in absolute terms, the EL-CCD was found to be ∼5 times more sensitive than a PMT-based fluorometer. 3. Factors contributing to system sensitivity: Although the CCD and EL technologies each have inherent issues of sensitivity when used for optical detection, sensitivity was maximized in our system by long exposure times and optimization of the system. Several factors were used in combination to increase the sensitivity of the EL-CCD detector to the level of photomultiplier-based fluorometer. A sensitive Sony ICX-429ALL CCD-based camera is the key element of the system. The thermo-electric cooling enables the CCD to operate at 25 K below ambient temperature, so that a much lower thermal noise is generated and much longer exposures are possible. An external power supply was used to reduce potential electrical noise. The f1.2 lens maximizes the light transmitted to the CCD. Finally, as described above, the type of EL used, the filter arrangement that reduced light interference, the light absorbing properties of the material used, and the compact configuration of the EL-enclosure (which minimizes light losses) all contributed to overall system sensitivity.

4. Notes 1. The laser’s power and speed for cutting polymers have to be determined empirically; it is recommended to use the minimum laser power to reduce overheating or burning the material. 2. The measurement of excitation light uniformity involves two steps. First, the dark signal is measured with the lens fully covered. This results in a dark signal from each pixel of the CCD array. Next, the wells are loaded with identical samples and the system is illuminated and imaged. The dark signal is subtracted from the image and the result is a measure of the excitation light uniformity as each sample should fluoresce uniformly. Figure 3 gives an example of the uniformity obtained by the EL system. 3. To measure the effectiveness of the filters, it is recommended to perform two long exposures (e.g. 3 min) without the assay plate, one with the EL on and one with the EL off. Ideally the results of these two measurements should be very similar (the blue filters pass only blue light which is blocked by the

270

Kostov et al.

C

A

B

Fig. 6. Assay’s plate fabrication; double sided adhesive roll bonding. (a) Double sided adhesive roll tape, (b) acrylic surface (c) ruler.

green filters). The difference between the measurements may suggest that the blue filters do not block all green light and/ or the green filters do not block all blue light. 4. The filter holder should be level and designed at a height which allows proper focusing of the lens system. Also, the filter holder must slide freely into the EL enclosure and must land in a reproducible position. 5. For strong bonding, remove ∼1 cm of the cover from one side of the double sided adhesive tape (Fig. 6a) and align and attach the tape with acrylic surface (Fig. 6b). With a ruler (Fig. 6c) press the tape against the surface and slowly unwind the tape and press it against the acrylic surface with the ruler to prevent air bubbles. After initial installation of the tape on the acrylic, run the acrylic (with tape) through the laminating machine. For assembling the assay plate, remove the protective cover from the exposed side of the double side adhesive (taped to the acrylic) and remove the protective cover from the polycarbonate. Align the two pieces, apply pressure and run the assay plate through the laminating machine. 6. The well “holes” on the assay plate are 2 mm in diameter, and allow analysis of 20 µL samples. Smaller holes make the loading of the sample less reproducible because of the fluid meniscus within the wells. Light scattering from the outer surface of the wells complicates quantitation. 7. The scattered light from the well walls depends on the well diameter, alignment of the optical system and the volume of fluid in the well. This scattered light is useful to mark the sample wells on the assay plates for analysis of low level signals. However, the scattered light can also interfere with low level signals. 8. Make sure that the lens, filters and the sample plate are vertically centered and aligned and the box used for the LE-CCD is sealed to cut out the light.

A Simple Portable Electroluminescence Illumination-Based CCD Detector

271

9. The camera enclosure must be sealed against external light leakage and the internal surfaces must be constructed with light absorbing material. Long exposure (e.g. 3 min) can be used to detect light leaks. 10. Make sure that the arrows on the coated filters are facing the CCD. 11. Open aperture (e.g. f1.2) will enable shorter exposure but focusing will be more limited and the image less sharp. 12. For data analysis, use the ImageJ high contrast visualization, which will not affect the values but will enable easy selection of the spots.

Acknowledgements We thank Mr. Steven Sun and Mr. Jesse Frances for their technical assistance. This work was supported in part by the Office of Public Health emergency Preparedness (OPHEP) IAG 224-05-655 (to A. Rasooly and by FDA contract HHSF223200610765P (to Dr. Yordan Kostov)

References 1. Capitan-Vallvey LF, Asensio LJ, LopezGonzalez J, Fernandez-Ramos MD, Palma AJ. Oxygen-sensing film coated photodetectors for portable instrumentation. Anal Chim Acta 2007;583:166–173 2. Mac Sweeney MM, Bertolino C, Berney H, Sheehan M. Characterization and optimization of an optical DNA hybridization sensor for the detection of multi-drug resistant tuberculosis. Conf Proc IEEE Eng Med Biol Soc 2004;3:1960–1963 3. Claycomb RW, Delwiche MJ. Biosensor for on-line measurement of bovine progesterone during milking. Biosens Bioelectron 1998;13:1173–1180 4. Bruno AE, Barnard S, Rouilly M, Waldner A, Berger J, Ehrat M. All-solid-state miniaturized fluorescence sensor array for the determination of critical gases and electrolytes in blood. Anal Chem 1997;69:507–513 5. Moehrs S, Del Guerra A, Herbert DJ, Mandelkern MA. A detector head design for small-animal PET with silicon photomultipli-

6.

7.

8.

9.

ers (SiPM). Physics in medicine and biology 2006;51:1113–1127 Takei M, Kida T, Suzuki K. Sensitive measurement of positron emitters eluted from HPLC. Appl Radiat Isot 2001;55:229–234 Ruiz-Martinez MC, Berka J, Belenkii A, Foret F, Miller AW, Karger BL. DNA sequencing by capillary electrophoresis with replaceable linear polyacrylamide and laserinduced fluorescence detection. Anal Chem 1993;65:2851–2858 Tibbe AG, de Grooth BG, Greve J, Liberti PA, Dolan GJ, Terstappen LW. Cell analysis system based on immunomagnetic cell selection and alignment followed by immunofluorescent analysis using compact disk technologies. Cytometry 2001;43:31–37 Tsukagoshi K, Jinno N, Nakajima R. Development of a micro total analysis system incorporating chemiluminescence detection and application to detection of cancer markers. Anal Chem 2005;77:1684–1688

272

Kostov et al.

10. Roda A, Manetta AC, Portanti O, et al. A rapid and sensitive 384-well microtitre format chemiluminescent enzyme immunoassay for 19-nortestosterone. Luminescence 2003;18:72–78 11. Ligler FS, Taitt CR, Shriver-Lake LC, Sapsford KE, Shubin Y, Golden JP. Array biosensor for detection of toxins. Analytical and bioanalytical chemistry 2003;377:469–477 12. Svitel J, Surugiu I, Dzgoev A, Ramanathan K, Danielsson B. Functionalized surfaces for optical biosensors: applications to in vitro pesticide residual analysis. Journal of materials science 2001;12:1075–1078 13. Liu Y, Danielsson B. Rapid high throughput assay for fluorimetric detection of doxorubicin–application of nucleic acid-dye bioprobe. Anal Chim Acta 2007;587:47–51 14. Burkert K, Neumann T, Wang J, Jonas U, Knoll W, Ottleben H. Automated preparation method for colloidal crystal arrays of monodisperse and binary colloid mixtures by contact printing with a pintool plotter. Langmuir 2007;23:3478–3484 15. Tohda K, Gratzl M. Micro-miniature autonomous optical sensor array for monitoring ions and metabolites 2: color responses to pH, K+ and glucose. Anal Sci 2006;22:937–941 16. Feldstein MJ, Golden JP, Rowe CA, Maccraith BD, Ligler FS. Array biosensor: optical and fluidics systems. Biomedical Microdevices 1999;1:139–153 17. Sohn YS, Goodey A, Anslyn EV, McDevitt JT, Shear JB, Neikirk DP. A microbead array chemical sensor using capillary-based sample introduction: toward the development of an

18.

19.

20.

21.

22.

23.

24.

25.

“electronic tongue”. Biosens Bioelectron 2005;21:303–312 Knecht BG, Strasser A, Dietrich R, Martlbauer E, Niessner R, Weller MG. Automated microarray system for the simultaneous detection of antibiotics in milk. Anal Chem 2004;76:646–654 Sapsford KE, Taitt CR, Loo N, Ligler FS. Biosensor detection of botulinum toxoid A and staphylococcal enterotoxin B in food. Appl Environ Microbiol 2005;71:5590–5592 Golden JP, Floyd-Smith TM, Mott DR, Ligler FS. Target delivery in a microfluidic immunosensor. Biosens Bioelectron 2007;22:2763–2767 Okamoto K, Niki I, Shvartser A, Narukawa Y, Mukai T, Scherer A. Surface-plasmonenhanced light emitters based on InGaN quantum wells. Nat Mater 2004;3:601–605 Schaferling M, Wu M, Enderlein J, Bauer H, Wolfbeis OS. Time-resolved luminescence imaging of hydrogen peroxide using sensor membranes in a microwell format. Appl Spectroscopy 2003;57:1386–1392 D’Auria S, Lakowicz JR. Enzyme fluorescence as a sensing tool: new perspectives in biotechnology. Curr Opin Biotechnol 2001;12:99–104 Vo-Dinh T, Alarie JP, Isola N, Landis D, Wintenberg AL, Ericson MN. DNA biochip using a phototransistor integrated circuit. Anal Chem 1999;71:358–363 Higgins JA, Nasarabadi S, Karns JS, et al. A handheld real time thermal cycler for bacterial pathogen detection. Biosens Bioelectron 2003;18:1115–1123

Chapter 15 Fluoroimmunoassays Using the NRL Array Biosensor Joel P. Golden and Kim E. Sapsford Summary Array-based biosensor technology offers the user the ability to detect and quantify multiple targets in multiple samples simultaneously (Analytical Sciences 23:5–10, 2007). The NRL Array Biosensor has been developed with the aim of creating a system for sensitive, rapid, on-site screening for multiple targets of interest. This system is fluorescence-based, using evanescent illumination of a waveguide, and has demonstrated the use of both sandwich and competitive immunoassays for the detection of both high and low molecular weight targets, respectively. The current portable, automated system has demonstrated detection of a wide variety of analytes ranging from simple chemical compounds to entire bacterial cells, with applications in food safety, disease diagnosis, homeland security and environmental monitoring. Key words: Fluorescence, Immunoassay, NRL array biosensor, Total internal reflection Fluorescence, Automated.

1. Introduction Biosensor technology, such as protein and DNA-based microarrays, provide a powerful analytical tool to researchers with the potential to address both fundamental scientific questions and to develop sensing systems for a variety of practical applications (1–8). For a biosensor to be commercially viable, however, the developed product must demonstrate a number of advantages over the existing technology, such as the ability to perform faster, be more sensitive, and have multi-analyte real-time measurements with a portable, easy-to-use, stand-alone device. The goal of the NRL Array Biosensor development has always been the sensitive, rapid,

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_15

273

274

Golden and Sapsford

on-site screening of multiple targets of interest and, as such, the system has demonstrated its utility for a variety of applications (9–11). The NRL Array Biosensor is currently being transitioned to the commercial sector for manufacture. This chapter describes the optimized protocols for both our manual and automated NRL Array Biosensor platforms. Typical immunoassay protocols are described here, as well as the preparation of spiked “real world” samples used to evaluate the matrix effects on the biosensor performance. Immunoassay development for a particular target is generally optimized using the manual NRL Array Biosensor which allows us to run dose response curves, both in the buffer and complex matrices, using 12 channel patterning and assay PDMS flow cells. As part of the immunoassay optimization, polyclonal and/or monoclonal antibodies from different commercial sources, where available, are screened to determine the best capture-tracer combination for the target of interest. The antibodies are also screened for cross-reactivity towards non-specific targets to minimize false positive rates on multi-target array waveguides. Once the immunoassays are optimized, they are then fully characterized on the automated NRL Array Biosensor. Table 1 lists the detection limits for assays developed on the NRL Array Biosensor.

Table 1 Immunoassay detection limits using NRL array biosensor (9) LODa

References

TNT

1–20 ng/mLb,c

(33)

Deoxynivalenol

0.2 ng/mLb (1–10 ng/g)d

(31)

Ochratoxin A

0.8 ng/mLb (3.8–100 ng/g)d

(29)

Aflatoxin B1

0.3 ng/mLb (0.6–5.1 ng/g)d

(32)

SEB

0.1 ng/mLb (0.1–0.5 ng/mL)d (1 ng/mL)e (100 ng/mL)f

(13, 22, 24, 25)

Cholera toxin

1.6 ng/mLb (100 ng/mL)f

(11, 22)

Botulinum toxoid A

20 ng/mLb (20–500 ng/mL)d

(24)

Botulinum toxoid B

200 ng/mLb

(11)

Target Small molecules

Protein targets Toxins

(continued)

Fluoroimmunoassays Using the NRL Array Biosensor

275

Table 1 (continued) LODa

References

8 ng/mLb

(11)

25 pg/mLb (1.3 ng/mL)d

(26, 28)

Yersinia pestis F1 antigen

25 ng/mLb,e

(13)

D-dimer

50 ng/mLb,e

(13)

Brucella abortus

3×103 cfu/mLb (5 × 105 cfu/mL)f

(11, 22)

Francisella tularensis

105 cfu/mLb (7 × 106 cfu/mL)f

(11, 22)

Salmonella typhimurium

8×104 cfu/mLb (8 × 104–4 × 105 cfu/mL)d

(21)

Shigella dysenteriae

5×104 cfu/mLb (5 × 104–8 × 105 cfu/mL)d

(23)

Campylobacter jejuni

103 cfu/mLb (2 × 103–3 × 103 cfu/mL)d

(23)

Target Ricin Allergens Ovalbumin Physiological markers

Bacterial targets

5

b

Bacillus globigii

10 cfu/mL

B. anthracis

624 cfu/mLb (7 × 104 cfu/mL)f

Escherichia coli O157:H7 Staphylococcus aureus

3

(12)

b

4

(11, 22) 4

5 × 10 cfu/mL (1 × 10 –5 × 10 cfu/mL) ∼106 cfu/mLb

d

(27) Unpublished

Viral targets MS2 bacteriophage Vaccinia

107 pfu/mLb 7

8

(12) b

10 –10 pfu/mL

Unpublished

a

cfu/mL (colony forming units/mL); pfu/mL (platform forming units/mL) Buffer LOD c Buffer LOD depends on immunoassay format d LOD range in food matrices e LOD range in clinical matrices f Fixed concentration measured in environmental matrices (not LOD) b

2. Materials 2.1. Develoment of the Manual NRL Array Biosensor Platform

1. Large format (~1” cooled CCD (Spectrasource, Teleris II with a Kodak KAF-1000 CCD chip). 2. A 2-dimensional GRIN lens array (Nippon Sheet Glass). 3. Emission filters: bandpass filter (Corion 700R70) and longpass filter (Schott RG665).

276

Golden and Sapsford

4. Custom scaffolding to hold GRIN lens array and filters. 5. Slide holder: machined to hold slide by edges and allow viewing of fluorescent assays. 6. A 635 nm diode laser (LAS-635–15, Lasermax, Rochester, NY) equipped with a line generator to spread the laser beam into a fan pattern. 7. Computer with image capture system. 8. Software for analyzing images and extracting assay data (programmed in LabWindows, National Instruments). 2.2. Development of the Automated NRL Array Biosensor Platform

1. Cooled CCD camera (Retiga 1300, Q-Imaging, Burnaby, BC, Canada). 2. A 17 mm C-mount lens (Schneider Optics, Hauppauge, NY). 3. Emission filters: bandpass filter (Corion 700R70) and longpass filter (Schott RG665). 4. Portable carrying case (Zero Enclosures, Salt Lake, UT). 5. Custom slide holder: machined to hold slide by edges and allow viewing of fluorescent assays while pressing PDMS against slide to form the channels. 6. A 635 nm diode laser (LAS-635–15, Lasermax, Rochester, NY) equipped with a line generator to spread the laser beam into a fan pattern. 7. Computer with image capture system. 8. Software for analyzing images and extracting assay data (programmed in LabWindows, National Instruments). 9. A 50 W switching power supply (Sunpower, Taipei, Taiwan). 10. Peristaltic pumps (P625/66.143, Instec, Germany). 11. Fluid valves (2-Way valve, LFAA0503418H, Lee Company, CT). 12. RS-232 interface (ADR-2000). 13. Mirror to guide laser light into microscope slide. 14. Custom mounts to hold camera, laser, circuit boards, pumps and valves.

2.3. Production of PDMS Flow Cells

1. Plexiglass patterning and assay flow cell molds and flow cell gasket mold (base and top plates). 2. NuSil MED-4011 Part A and Part B (Nusil Silicone Technology, Carpintera, CA). 3. Vacuum oven. 4. Oven. 5. Large (500–1,000 mL) plastic disposable beaker. 6. Carbon Black, finely ground.

Fluoroimmunoassays Using the NRL Array Biosensor

2.4. Biotin-Labeling of Capture Molecules

277

1. Bicarbonate buffer, 50 mM pH 8.0 (Sigma-Aldrich). 2. Dimethyl sulfoxide, DMSO (Sigma-Aldrich). 3. Phosphate buffered saline, PBS, 10 mM PB + 138 mM NaCl + 2.7 mM KCl (Sigma-Aldrich). 4. Bio-Gel P-10 column material and ~20 mL (1.5 × 14 cm) column (BioRad, Richmond, CA). 5. EZ-link biotin-LC-NHS ester (Pierce, Rockford, IL). 6. Column pump and fraction collector (optional). 7. UV-visible spectrometer and cuvette. 8. Purified mono- or poly-clonal antibody stocks; source is dependent on species, but typical companies include: Jackson ImmunoResearch (West Grove, PA), Biodesign International (Saco, ME), KPL (Gaithersburg, MD) and Toxin Technology, Inc. (Sarasota, FL).

2.5. Fluorescent Labeling of Tracer Molecules

1. Borate buffer, 50 mM pH 8.5 (Sigma-Aldrich). 2. Dimethyl sulfoxide, DMSO (Sigma-Aldrich). 3. PBS (Sigma-Aldrich). 4. Bio-Gel P-10 column material and ~20 mL (1.5 × 14 cm) column (BioRad, Richmond, CA). 5. AlexaFluor 647-NHS ester (Invitrogen/Molecular Probes, Carlsbad, CA) or Cy5-bisfunctional-NHS ester (Amersham Biosciences/GE Healthcare, Piscataway, NJ). 6. Column pump and fraction collector (optional). 7. UV-visible spectrometer and cuvette. 8. Purified mono- or poly-clonal antibody stocks; source is dependent on species (see Subheading 2.4).

2.6. Preparation of Spiked Complex Matrices

1. Food matrices were purchased from local grocery stores. 2. Pollen (Sigma-Aldrich). 3. Sand, clay, smoke extracts (from burned brush, jet propulsion fuel and signal flares) provided by US Army Dugway Proving Grounds (DPG) (Dugway, UT). 4. River water collected directly from a local river using sterile 50 mL tubes (Potomac, MD). 5. Purified analyte of interest, for spiking. As with the antibodies the source of the analyte is species- specific. For example, the mycotoxins were purchased from Sigma-Aldrich, while Staphylococcus aureus enterotoxin B (SEB) was obtained from Toxin Technology, Inc. (Sarasota, FL). 6. Methanol (Sigma-Aldrich). 7. PBS containing 1 mg/mL Bovine Serum Albumin (BSA) and 0.05% Tween-20 (PBSTB) (Sigma-Aldrich).

278

Golden and Sapsford

8. Polyvinylpyrrolidone (PVP) (Sigma-Aldrich). 9. Poly (ethylene glycol)-8000 (PEG-8000) (Sigma-Aldrich). 10. Sodium Hydroxide (NaOH) (Sigma-Aldrich). 11. Centrifuge (capable of 4,000 rpm) and 15 mL or 50 mL sterile plastic disposable centrifuge tubes. 12. Sample rocker and vortexer. 13. Waring two speed commercial food blender (Fisher Scientific, Pittsburgh, PA). 2.7. Immunoassays

1. PBS containing 0.05% Tween-20 (PBST) (Sigma-Aldrich) and PBSTB. 2. Blocking solution: PBST containing 1% Gelatin (SigmaAldrich). 3. Blocking solution: PBST containing 10 mg/mL BSA (SigmaAldrich). 4. ISMATEC® multi-channel pump (Cole-Parmer Instruments Company, Vernon Hills, IL). 5. Syringe barrels (1 mL) and 21 G or 25 G PrecisionGlide® needles (Becton Dickinson and Company, Franklin Lakes, NJ). 6. PDMS assay flow cell and chuck holder for the manual NRL Array Biosensor. 7. Analyte samples and corresponding fluorescently labeled tracer antibodies.

3. Methods 3.1. Development of Manual NRL Array Biosensor Platform

The first version of the NRL Array Biosensor was designed to interrogate the slides after the assays were performed in the PDMS flow chambers (9–17). 1. A diode laser is mounted such that the beam, expanded using a line generator, launches into the entire end edge of the microscope slide. 2. The glass slide acts as a waveguide and the resulting evanescent field at the surface excites the fluorophores. 3. The slide is mounted at the focal plane of the CCD imaging system, which is shown in Fig. 1. 4. The GRIN lens array focuses the slide surface onto the CCD array in a very short working distance (17). A standard large diameter convex lens can be used in place of the GRIN lens array, but will require more room.

Fluoroimmunoassays Using the NRL Array Biosensor

279

Fig. 1. Schematic of the manual NRL Array Biosensor detection optics. The diode laser beam is expanded to fill the slide edge using a line generator. The resulting surface fluorescence is focused onto a CCD camera using a GRIN lens array.

5. Emission filters placed between the imaging lens and the CCD array reject the laser light and pass the fluorescent emission. This detection system is a manual sensor system (i.e. it has no pumps, valves or flow chambers) and is designed to interrogate the slides for results after the assays are completed. For time-response information, a mount can be designed to secure the flow channels to the slide while it is being imaged. Still images can be made during the assays, or, depending on the image capture system, a video of the assay response can be obtained. 3.2. Development of Automated NRL Array Biosensor Platform

A portable version of the array sensor also uses a diode laser for excitation and a microscope slide as a waveguide and support for performing the assays, but in addition it has automated fluidics handling features (9, 10, 16). To make the system portable, a small cooled CCD camera and lens for imaging the slide surface are incorporated into a carrying case along with the pumps, valves and fluidics reservoirs. Figure 2 A and B show a schematic of the optics and fluidics control system and a photograph of the portable array sensor. 1. The pumps and valves used to introduce the sample and fluorescent antibody solution are controlled via an RS-232 interface board and custom signal conditioning circuitry. 2. A laptop with timing software written in LabWindows (National Instruments) sends control information via an RS-232 link. 3. A custom slide holder presses the slide against the PDMS flow channels while holding it in position to receive the excitation light. As the PDMS in contact with the slide surface

280

Golden and Sapsford

Fig. 2. (a) Optics and fluidics diagram for automated NRL Array Biosensor. The system has six tracer and sample reservoirs connected to six channels of peristaltic pumps via a bank of six valves. The valves control the flow of either sample or tracer through the assay chambers. (b) Photograph of the portable array biosensor showing the diode laser, CCD camera, slide holder and reservoirs.

can scatter the excitation light, slides are prepared with a silver coating in patterns that match where the PDMS makes contact with the slide. The silver coating maintains total internal reflection while allowing imaging of the assay in the flow channels (10).

Fluoroimmunoassays Using the NRL Array Biosensor

3.3. Production of PDMS Flow Cells 3.3.1. Production of PDMS Flow Cells for the Manual NRL Array Biosensor

3.3.2. Production of PDMS Gaskets for the Automated NRL Array Biosensor

281

Production of the PDMS patterning flow cell is described in detail in a chapter by our coworkers, Taitt, Charles and Shriver-Lake. The assay PDMS flow cell is prepared using the same procedure as the patterning PDMS flow cell, except that the channels in the Plexiglass assay mold are orientated perpendicular to those of the patterning mold. For the automated NRL Array Biosensor a 6-channel PDMS molded black flow cell gasket is prepared and inserted into the fluidics portion of the system. The black PDMS gasket is found to reduce light scatter relative to a clear PDMS gasket, improving the limit of detection (18). 1. Mix NuSil MED-4011 Part A and Part B at a 10:1 w/w ratio in a large plastic disposable beaker. 2. Mix finely ground carbon black into the uncured PDMS at a 1:25 w/w ratio. 3. Degas the mixture in a vacuum oven at room temperature until all the bubbles are removed (~30 min). 4. The black PDMS is then poured into the Plexiglass base plate mold and the mold placed in the vacuum oven for further degassing until the bubbles are removed. 5. Firmly press the Plexiglass top plate, which creates the channels, into position on the base plate mold and leave the assembly to cure at room temperature overnight. If, after 12–14 h, the PDMS requires further curing, it can be placed in a 60°C oven for 30 min. 6. After curing, the PDMS gasket is carefully removed from the mold assembly by first releasing the top plate and then the base plate mold; the resulting PDMS gasket is then trimmed and washed with soapy water to remove any residue. 7. The PDMS gasket is then incubated in 1% bovine serum albumin (BSA) prior to assembly in the fluidics system to prevent nonspecific absorption of molecules to the walls of the channels.

3.4. Biotinylation of Capture Molecules

Most of the immunoassays developed with the NRL Array Biosensor to date have used sandwich immunoassay formats for analyte detection, in which antibodies are used as capture agents. Typically, the capture antibody is biotinylated and immobilized on an avidin-coated surface. Some antigens, like mycotoxins, are not suitable for sandwich immunoassay detection due to their small size, therefore competitive immunoassays have been developed. The competitive format requires the synthesis of biotin-labeled analogs of the mycotoxins under investigation. Biotin-labeling of ochratoxin A, fumonisin B, aflatoxin B1 and deoxynivalenol is dependent on the mycotoxin used and has been described in detail in reference (19).

282

Golden and Sapsford

1. Prepare 1 mg/mL antibody in 50 mM bicarbonate buffer pH 8.0 (see Note 1). 2. Dissolve the EZ-link biotin-LC-NHS ester in DMSO to a final concentration of 1 mg/mL. 3. Add the biotin/DMSO mix to the 1 mg/mL antibody solution such that the final ratio of biotin-to-antibody is 5:1. 4. Incubate at room temperature for between 0.5 and 1 h, either on a rocking platform or mixing every 5–10 min. 5. Pre-equilibrate a Bio-Gel P-10 column (∼20–25 mL column) with PBS. 6. Load the biotin/antibody mix onto the column then elute the antibody using PBS either under gravity flow and collecting 1 mL fractions (~10 mL total), or connecting the column to a pump (0.75 mL/min) equipped with a fraction collector. Monitor the absorbance at 280 nm of the collected fractions to determine where the antibody elutes and the final concentration. 3.5. Fluorescent Labeling of Tracer Antibodies

Both Cy5 and AlexaFluor 647 have successfully been used to fluorescently label tracer antibodies for detection using the NRL Array Biosensor. AlexaFluor 647 allows the user to label the antibody at greater dye-to-antibody ratios (such as 6:1) without the risk of fluorophore self-quenching. Cy5 fluorescent dye, however, seems to label monoclonal antibodies more effectively than AlexaFluor 647. 1. Prepare 1 mg/mL antibody in 50 mM borate buffer pH 8.5 (see Note 2). 2. In the case of Cy5, 1 vial of bis-functional-NHS ester-Cy5 is dissolved in 50 µL DMSO and 25–45 µL added to the antibody (see Note 3). 3. For the AlexaFluor 647-NHS ester a 1 mg vial is first dissolved in 500 µL of anhydrous acetonitrile and divided into 20 µL aliquots, then dried using an Eppendorf spin vacuum, sealed under nitrogen and stored at −20°C (20). Three aliquots of the AlexaFluor 647 are each dissolved in 10 µL DMSO, combined and then added to the 1 mg/mL antibody solution. 4. Incubate the dye/antibody solution at room temperature for between 0.5 and 1 h, either on a rocking platform or mixing every 5–10 min. 5. Pre-equilibrate a Bio-Gel P-10 column (∼20–25 mL column) with PBS. 6. Load the dye/antibody mix onto the column; elute the antibody using PBS either under gravity flow and collecting the blue colored fractions (∼1 mL), or connecting the column to a pump (0.75 mL/min) equipped with a fraction collector. The first blue band to elute contains the dye-labeled antibody, while the second blue band consists of free dye. The

Fluoroimmunoassays Using the NRL Array Biosensor

283

dye-to-antibody ratio is determined using equations provided in the manufacturer’s instructions, by measuring the UV-visible absorbance at 280 and 650 nm (see Note 4). 3.6. Preparation of Spiked Complex Matrices

Many of the immunoassays developed using the NRL Array Biosensor have been screened in complex sample matrices to determine the effect of the matrix on various assay performance parameters such as limit-of-detection (LOD), false positive and false negative rates. The aim is to limit the amount of sample pretreatment prior to the assay in order to maintain a rapid analysis time. Sample matrices are typically spiked with the analyte of interest and then allowed to incubate for a minimum of 2 h before appropriate sample pretreatment is preformed prior to the assay. The extent of sample pretreatment is dependent on the matrix under investigation, but typical procedures for clinical, environmental and food matrices are outlined in Table 2.

Table 2 Preparation of spiked “real world” samples and their typical sample matrix pretreatment Matrix type

Examples

Typical sample preparation and pretreatment

Clinical

Saliva, nasal swabs, urine, serum, plasma

Samples used as received, spiked, with no further treatment prior to the assay

Whole blood

Blood diluted 50% with PBSTB before spiking. No further treatment prior to the assay

Chicken fecal matter

1 g of fecal matter is added to 2 mL PBSTB and sample vortexed before spiking. Samples filtered to remove large particulates prior to the assay

Environmental River water, smoke extract (from burned brush, jet propulsion fuel and signal flares) Sand, clay, pollen

Food (bacteria and protein toxins)

Samples used as received, spiked with no further treatment prior to the assay

References (13, 21)

(22)

Samples suspended in PBSTB to a final concentration of 1% (w/v) before spiking. No further treatment prior to the assay

Samples diluted 1:1 (w/v) with PBSTB and (19, 21, Pork sausage, ground 23–28) homogenized with a blender. Samples beef, ground turkey, spiked. Prior to assay samples centrifuged ground turkey sausage, at 4,000 rpm, 10 min and resulting superham, cantaloupe, eggs, natant used in the assay. Note, tomato canned corn, green samples were first neutralized using beans, tuna, mushNaOH prior to spiking rooms, tomatoes (continued)

284

Golden and Sapsford

Table 2 (continued) Matrix type

Food (mycotoxins)

Typical sample preparation and pretreatment

Examples Carcass wash

10 mL PBS + 1 mg/mL BSA was added to a sealable plastic bag containing a 3.2 lb chicken carcass, and placed on a rocker for 2 h at room temperature. The carcass wash was removed from the plastic bag and spiked. No further treatment prior to the assay

Milk

Add 10× PBSTB to the milk and spike. No further treatment prior to the assay

Lettuce leaves, Alfalfa sprouts

1 g of solid is added to 1 mL PBSTB and sample vortexed before spiking. Remove liquid and use for the assay

Yogurt

Yogurt diluted 1:1 (v/v) with PBSTB and vortexed before spiking. No further treatment prior to the assay

References

(19, Barley, wheat, oats, corn- The barley, wheat, oats, cornmeal, corn29–32) meal, cornflakes, wheat flakes, wheat pasta and roasted coffee pasta, roasted coffee samples are first blended into a fine texture. Then 0.5 g (2 mL for wine) aliquots, for all samples, are spiked with the appropriate mycotoxin in 100% methanol. The methanol is allowed to evaporate overnight (except wine). For the solid samples 2 mL of 75% methanol/water was added, the sample first vortexed for 3 min then placed on a rocker for 2 h at room temp. The samples were then centrifuged at 3,000 rpm for 10 min and supernatant collected and diluted threefold in PBSTB prior to the assay Wine, un-popped popcorn, peanuts, peanut butter, pecan nuts

3.7. Immunoassays

Note: Extra additions to diluted supernatant include; wine and coffee extracts were neutralized using NaOH. Polyvinylpyrrolidone (PVP) was added to wine samples at 0.4 mL per 2 mL of wine. Poly (ethylene glycol) [PEG 8000] was added to all nut products at a final concentration of 0.66%

The NeutrAvidin-coated microscope slides, used as waveguides by the NRL Array Biosensor, are patterned with biotin-labeled capture species using PDMS patterning flow cells as described in a chapter by our coworkers, Taitt, Charles and Shriver-Lake. The

Fluoroimmunoassays Using the NRL Array Biosensor

285

slide typically contains one or two positive control lanes consisting of biotin-rabbit-anti-chicken IgY, which help the user to align the slide for data analysis. Following patterning, the waveguides are blocked in a PBST solution containing 10 mg/mL BSA for 30 min prior to the assay. This helps to reduce non-specific binding of both the target and the tracer antibody. In the case of the mycotoxin-modified waveguides, used in competitive immunoassays, PBST containing 1% gelatin is used as a blocking agent (see Note 5). The NRL Array Biosensor has successfully demonstrated direct, sandwich, displacement and competitive immunoassay formats; however, sandwich and competitive formats, shown in Fig. 3, are more routinely used for target analyte detection and quantification (33). 3.7.1. Competitive Immunoassays Using the Manual NRL Array Biosensor

Competitive immunoassays are used with the NRL Array Biosensor for mycotoxin detection (19, 29–32). The mycotoxins ochratoxin A, fumonisin B, aflatoxin B1 and deoxynivalenol are too small to contain the two distinct recognition sites required by the more common sandwich immunoassay format. Therefore, a competitive format, using a biotin-analog of the mycotoxin patterned on the waveguide, is used to quantify mycotoxins free in solution. 1. The mycotoxin sample prepared, as described in Subheading 3.6, is diluted threefold with PBST containing BSA, Cy5-antimycotoxin antibody and Cy5-chicken IgY (positive control) such that the final concentrations are 0.66 µg/mL, 2–5 µg/ mL (depending on the mycotoxin) and 50 ng/mL, respectively, and incubated for 20 min at room temperature.

Fig. 3. Schematic representation of the two main solid-phase immunoassays used for analyte detection. (a) Sandwich assay: the amount of immobilized antigen is determined by passing a second fluorescently labeled antibody over the antibody-bound antigen. The fluorescent signal, measured from the surface, increases with increasing antigen concentration. (b) Competitive assays: competition for binding sites on the solution based fluorescently labeled antibody occurs between the unlabeled antigen in solution and immobilized antigen-analog. The fluorescent signal, measured from the surface, decreases with increasing concentration of the solution antigen.

286

Golden and Sapsford

2. The NeutrAvidin waveguide patterned with the biotin-labeled analog of the mycotoxin under investigation is first assembled with the assay PDMS flow cell using the chuck to apply even pressure between the microscope slide and PDMS flow cell. 3. Each channel is connected to the ISMATEC® multi-channel pump at one end (outlet) and syringe barrels (1 mL) attached at the opposite end (inlet). 4. First, 1 mL PBSTB at 0.5 mL/min is pumped through the channels to check for leaks. 5. Next, 0.8 mL of the sample, containing both the “free” mycotoxin and the tracer antibody, is applied to each channel at a flow rate if 0.15 mL/min. 6. The channels are then washed with 1 mL PBSTB at 0.5 mL/min. 7. The PDMS flow cell is then removed by releasing the chuck, and the slides washed in 18 mΩ Milli-Q water, dried with nitrogen and imaged on the manual NRL Array Biosensor platform. 3.7.2. Sandwich Immunoassays Using the Manual NRL Array Biosensor

Sandwich immunoassays are used with the NRL Array Biosensor for the detection and quantification of a number of protein toxins and bacteria (9). Here the slide is patterned with biotin-labeled antibodies which act as the capture species for the immunoassays. 1. The NeutrAvidin waveguide patterned with the biotin-labeled antibodies is first assembled with the assay PDMS flow cell using the chuck to apply even pressure between the microscope slide and PDMS flow cell. 2. Each channel is hooked up to the ISMATEC® multi-channel pump at one end (outlet) and syringe barrels (1 mL) attached at the opposite end (inlet). 3. First 1 mL PBSTB at 0.5 mL/min is pumped through the channels to check for leaks. 4. Next, 0.8 mL of the sample is applied to each channel at a flow rate if 0.15 mL/min (see Note 6). 5. The channels are then washed with 1 mL PBSTB at 0.5 mL/min. 6. Following the PBSTB rinse, 0.4 mL of the fluorescent antibody cocktail mix, containing Cy5- or AlexaFluor 647-labeled antibodies to each of the analyte targets of interest (typically 10 µg/mL) and Cy5-chicken IgY (positive control; 100 ng/ mL) in PBSTB, is applied to each channel at a flow rate of 0.15 mL/min. 7. The channels are washed with 1 mL PBSTB at 0.5 mL/min. 8. The PDMS flow cell is then removed by releasing the chuck, and the slide washed in 18 mΩ Milli-Q water, dried with nitrogen and imaged on the manual NRL Array Biosensor platform.

Fluoroimmunoassays Using the NRL Array Biosensor 3.7.3. Sandwich Immunoassays Using the Automated NRL Array Biosensor

287

Unlike the manual NRL Array Biosensor the automated platform makes use of custom-manufactured silver-clad microscope slides (34). The silver-clad slides are functionalized using the same procedure as the standard microscope slides. The purpose of the silver-cladding is to reduce the stripping of the evanescent field that occurs when the PDMS flow cell comes into contact with the waveguide. The fluidics of the Array Biosensor is fully automated and controlled from a computer using a program written in LabWindows/CVI (National Instruments). 1. The antibody functionalized sliver-clad waveguide is inserted into the vertical slide mount, consisting of a spring loaded plate which presses the waveguide against a 6-channel PDMS molded black flow cell gasket. 2. The sample and tracer reservoirs are mounted onto the fluidics platform where they connect, via blunt needles, through a rubber septum to the fluidics system. 3. Once the reservoirs and waveguide are mounted in the automated NRL Array Biosensor, the samples (1 mL) and tracer antibody cocktail (0.4 mL) are loaded into the appropriate reservoirs. 4. The assay program, which controls the fluidics, is then initiated. The 1.0 mL samples is flowed through each channel for 900 s before the inlet valve is automatically switched to pull the 0.4 mL tracer antibody cocktail over the waveguide surface for 300 s. After completion of the tracer cycle the pumps are reversed and the outlet valve switched to buffer, allowing a wash cycle to remove any residual fluorescent agent. 5. After completion of the assay, the laser is switched on and the slide imaged with the cooled CCD, using the camera control software.

3.8. Data Analysis

Examples of typical CCD images obtained from both the manual and automated versions of the NRL Array Biosensor are shown in Fig. 4. An image analysis program written in LabWindows (National Instruments) extracts the assay data from the images. In the software, the user aligns a virtual grid with the fluorescent array and the pixel values for each array spot is extracted and compared with the local background. A positive or negative result is calculated by comparing the fluorescent signal with a threshold value determined from the local background level plus three times the standard deviation of the background noise. The location of the spot in the array determines which analyte was detected. Throughout our studies it has been apparent that while the NRL Array Biosensor is not as sensitive as the typical “gold standard” analysis for the particular analyte of interest (see references listed in Table 2) it is similar and in some instances better than equivalent real-time, rapid immunoassay type detection systems, such as the ELISA. In most cases the NRL Array Biosensor is quicker

288

Golden and Sapsford

Fig. 4. Images taken using the manual and automated versions of the NRL Array Biosensor used for the simultaneous multi-analyte quantification of botulinum toxoid A (Bot. toxoid A), SEB and Campylobacter jejuni (Camp. J22), in buffer. Note that the CCD image of the automated system is rotated 90° to the manual system. For the manual system the 12-channel patterning flow cells were exposed left to right to [1 + 12] rabbit anti-chick IgY, [2–5] rabbit anti-Bot. tox., [6–8] rabbit anti-SEB, and [9–11] rabbit anti Camp. J. In the case of the automated system the patterning flow cells were exposed top to bottom to [1 + 12] rabbit anti-chick IgY, [2–4] rabbit anti Camp. J, [5–7] rabbit anti-SEB, and [8–11] rabbit anti-Bot. tox. The resulting assay flow cell, in both systems, was exposed first to varying concentrations of Camp. J22 (0–3 × 104 cfu/mL), SEB (0–7.5 ng/mL) and Bot. toxoid A (0–500 ng/mL), followed by a fluorescent antibody tracer cocktail.

(0.5–1 h vs. 3–4 h) and requires less sample preparation than the standard ELISA, while maintaining LOD. Figure 5 illustrates some typical data obtained using the automated NRL Array Biosensor. Here, tomato or corn samples were spiked with varying concentrations of SEB, Campylobacter jejuni (Camp. J22) and Botulium toxoid A (B. toxoid A), as described in Table 2. Following the immunoassay the image was analyzed and the dose response curves plotted for each toxin as shown in Fig. 5. For the toxin analytes the automated NRL Array Biosensor buffer LODs were found to be the same as those of the manual Array Biosensor, 50 ng/mL for B. toxoid A and 0.5 ng/mL for SEB (using optimized exposure times), using equivalent assays. The bacterial analyte, Camp. J22, LOD was slightly improved on the automated versus the manual NRL Array Biosensor platform, measuring 500 cfu/mL versus 1,000 cfu/ mL, respectively. Clearly from the dose response curves (Fig. 5) the corn matrix significantly dampens the response for the toxins B. toxoid A and SEB, resulting in increased LODs of 100 ng/ mL and 3 ng/mL respectively. However, neither of the food matrices tested appear to affect the Camp. J22 immunoassay. This study highlights one of the important conclusions obtained from a number of our studies: that it is important to calibrate the response of the NRL Array Biosensor using a dose response in

Fluoroimmunoassays Using the NRL Array Biosensor

289

Fig. 5. Dose-response assays in vegetable products. (a) A representative CCD image taken of the slide following completion of the on-line assay. (b) Dose-response curves for Camp. J22 (0–3 × 104 cfu/mL) in PBSTB (circles), tomatoes (squares) and corn (triangles), using a 10 second exposed image of the slide. (c) Dose-response curves for SEB (0–7.5 ng/mL) in PBSTB (circles), tomatoes (squares) and corn (triangles), using a 5 second exposed image of the slide (NB the 7.5 ng/mL gave a saturating signal on the CCD therefore the data was not plotted). (d) Dose-response curves for Bot. toxoid A (0–500 ng/mL) in PBSTB (circles), tomatoes (squares) and corn (triangles), using a 10 second exposed image of the slide. Error bars are the standard deviation of the mean (n = 6–8).

the matrix of interest and not simply in buffer. This minimizes the number of potential false negative responses as it allows the user to adjust the LOD thresholds appropriately for a positive response.

290

Golden and Sapsford

4. Notes 1. 50 mM Borate buffer pH 8.5–9.0 can also be used in this procedure. 2. 50 mM bicarbonate buffer pH 8.0 can also be used in this procedure. 3. The amount of Cy5 added is dependent on the particular antibody being labeled. For some antibodies, addition of 25 µL resulted in the desired final dye-to-antibody ratio, whereas other antibodies required larger volumes. 4. Typical dye-to-antibody ratios range from 2 to 4 for Cy5 and 2 to 6 for AlexaFluor 647. 5. Due to the small size of the biotin-labeled mycotoxins used in the competitive immunoassay format, waveguides are blocked with a gelatin blocker (MW 20–25 KDa) rather than the larger protein BSA (~66 KDa), to reduce the risk of hindering the active sites. 6. For typical immunoassays the surface is exposed to the sample for ~6 min, which is the time the 0.8 mL sample takes to pass through the channels. This time can be extended, however, by recycling the sample over the waveguide surface. Recycling is achieved by connecting the outlet from the pump to the inlet syringe using tubing. Previous experiments have demonstrated that recycling the sample over the waveguide for 1 h versus 6 min improves the LOD by a factor of ~10.

Acknowledgments The views, opinions, and/or findings described in this chapter are those of the authors and should not be construed as official Department of the Navy positions, policies, or decisions. References 1. Bally, M., Halter, M., Voros, J., and Grandin, H. M. (2006) Optical microarray biosensing techniques. Surface and Interface Analysis 38, 1442–1458 2. Rogers, K. R. (2006) Recent advances in biosensor techniques for environmental monitoring. Analytica Chimica Acta 568, 222–231 3. Lim, D. V., Simpson, J. M., Kearns, E. A., and Kramer, M. F. (2005) Current and developing technologies for monitoring agents of bioter-

rorism and biowarfare. Clinical Microbiology Reviews 18, 583–607 4. Kumble, K. D. (2003) Protein microarrays: new tools for pharmaceutical development. Analytical and Bioanalytical Chemistry 377, 812–819 5. Schaferling, M., and Nagl, S. (2006) Optical technologies for the read out and quality control of DNA and protein microarrays. Analytical and Bioanalytical Chemistry 385, 500–517

Fluoroimmunoassays Using the NRL Array Biosensor 6. Weller, M. G. (2005) Optical microarray biosensors. Analytical and Bioanalytical Chemistry 381, 41–43 7. Wulfkuhle, J., Espina, V., Liotta, L., and Petricoin, E. (2004) Genomic and proteomic technologies for individualisation and improvement of cancer treatment. European Journal of Cancer 40, 2623–2632 8. Wulfkuhle, J. D., Edmiston, K. H., Liotta, L. A., and Petricoin, E. F. (2006) Technology insight: pharmacoproteomics for cancer promises of patient-tailored medicine using protein microarrays. Nature Clinical Practice Oncology 3, 256–268 9. Ligler, F. S., Sapsford, K. E., Golden, J. P., Shriver-Lake, L. C., Taitt, C. R., Dyer, M. A., Barone, S., and Myatt, C. J. (2007) The array biosensor: Portable, automated systems. Analytical Sciences 23, 5–10 10. Golden, J. P., Taitt, C. R., Shriver-Lake, L. C., Shubin, Y. S., and Ligler, F. S. (2005) A portable automated multianalyte biosensor. Talanta 65, 1078–1085 11. Rowe-Taitt, C. A., Golden, J. P., Feldstein, M. J., Cras, J. J., Hoffman, K. E., and Ligler, F. S. (2000) Array biosensor for detection of biohazards. Biosensors & Bioelectronics 14, 785–794 12. Rowe, C. A., Tender, L. M., Feldstein, M. J., Golden, J. P., Scruggs, S. B., MacCraith, B. D., Cras, J. J., and Ligler, F. S. (1999) Array biosensor for simultaneous identification of bacterial, viral, and protein analytes. Analytical Chemistry 71, 3846–3852 13. Rowe, C. A., Scruggs, S. B., Feldstein, M. J., Golden, J. P., and Ligler, F. S. (1999) An array immunosensor for simultaneous detection of clinical analytes. Analytical Chemistry 71, 433–439 14. Rowe-Taitt, C. A., Cras, J. J., Patterson, C. H., Golden, J. P., and Ligler, F. S. (2000) A ganglioside-based assay for cholera toxin using an array biosensor. Analytical Biochemistry 281, 123–133 15. Golden, J., Shriver-Lake, L., Sapsford, K., and Ligler, F. (2005) A “do-it-yourself” array biosensor. Methods 37, 65–72 16. Taitt, C. R., Golden, J. P., Shubin, Y. S., Shriver-Lake, L. C., Sapsford, K. E., Rasooly, A., and Ligler, F. S. (2004) A portable array biosensor for detecting multiple analytes in complex samples. Microbial Ecology 47, 175–185 17. Golden, J. (1998), US Patent 5,827,748 18. Johnson-White, B., and Golden, J. (2005) Reduction of background signal in automated array biosensors. Measurement Science & Technology 16, N29–N31

291

19. Sapsford, K. E., Ngundi, M. M., Moore, M. H., Lassman, M. E., Shriver-Lake, L. C., Taitt, C. R., and Ligler, F. S. (2006) Rapid detection of foodborne contaminants using an Array Biosensor. Sensors and Actuators B-Chemical 113, 599–607 20. Anderson, G. P., and Nerurkar, N. L. (2002) Improved fluoroimmunoassays using the dye Alexa Fluor 647 with the RAPTOR, a fiber optic biosensor. Journal of Immunological Methods 271, 17–24 21. Taitt, C. R., Shubin, Y. S., Angel, R., and Ligler, F. S. (2004) Detection of Salmonella enterica serovar typhimurium by using a rapid, array-based immunosensor. Applied and Environmental Microbiology 70, 152–158 22. Rowe-Taitt, C. A., Hazzard, J. W., Hoffman, K. E., Cras, J. J., Golden, J. P., and Ligler, F. S. (2000) Simultaneous detection of six biohazardous agents using a planar waveguide array biosensor. Biosensors & Bioelectronics 15, 579–589 23. Sapsford, K. E., Rasooly, A., Taitt, C. R., and Ligler, F. S. (2004) Detection of Campylobacter and Shigella species in food samples using an array biosensor. Analytical Chemistry 76, 433–440 24. Sapsford, K. E., Taitt, C. R., Loo, N., and Ligler, F. S. (2005) Biosensor detection of botulinum toxoid A and staphylococcal enterotoxin B in food. Applied and Environmental Microbiology 71, 5590–5592 25. Shriver-Lake, L. C., Shubin, Y. S., and Ligler, F. S. (2003) Detection of staphylococcal enterotoxin B in spiked food samples. Journal of Food Protection 66, 1851–1856 26. Shriver-Lake, L. C., Taitt, C. R., and Ligler, F. S. (2004) Applications of array biosensor for detection of food allergens. Journal of AOAC International 87, 1498–1502 27. Shriver-Lake, L. C., Turner, S., and Taitt, C. R. (2007) Rapid detection of Escherichia coli O157: H7 spiked into food matrices. Analytica Chimica Acta 584, 66–71 28. Williams, K. M., Westphal, C. D., and ShriverLake, L. C. (2004) Determination of egg proteins in snack food and noodles. Journal of AOAC International 87, 1485–1491 29. Ngundi, M. M., Shriver-Lake, L. C., Moore, M. H., Lassman, M. E., Ligler, F. S., and Taitt, C. R. (2005) Array biosensor for detection of ochratoxin A in cereals and beverages. Analytical Chemistry 77, 148–154 30. Ngundi, M. M., Shriver-Lake, L. C., Moore, M. H., Ligler, F. S., and Taitt, C. R. (2006) Multiplexed detection of mycotoxins in foods with a regenerable array. Journal of Food Protection 69, 3047–3051

292

Golden and Sapsford

31. Ngundi, M. M., Qadri, S. A., Wallace, E. V., Moore, M. H., Lassman, M. E., Shriver-Lake, L. C., Ligler, F. S., and Taitt, C. R. (2006) Detection of deoxynivalenol in foods and indoor air using an array biosensor. Environmental Science & Technology 40, 2352–2356 32. Sapsford, K. E., Taitt, C. R., Fertig, S., Moore, M. H., Lassman, M. E., Maragos, C. A., and Shriver-Lake, L. C. (2006) Indirect competitive immunoassay for detection of aflatoxin B-1 in corn and nut products using

the array biosensor. Biosensors & Bioelectronics 21, 2298–2305 33. Sapsford, K. E., Charles, P. T., Patterson, C. H., and Ligler, F. S. (2002) Demonstration of four immunoassay formats using the array biosensor. Analytical Chemistry 74, 1061–1068 34. Feldstein, M. J., Golden, J. P., Rowe, C. A., MacCraith, B. D., and Ligler, F. S. (1999) Array Biosensor: Optical and Fluidics Systems. Journal of Biomedical Devices 1, 139–153

Chapter 16 Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral and Polarization Imagers for Fluorescence and Spectroscopic Imaging Neelam Gupta Summary Filters are a critical element in fluorescence detection used by many biosensors. One of the main limitations of the conventional optical filters used in biosensors is that they are limited to a single wavelength operation while numerous wavelengths are used in a typical fluorescence detection used for biosensing. Acousto-optic tunable filters (AOTFs) have the potential to overcome this limitation and provide both spectral and polarization information because they are wavelength agile and polarization sensitive. Such filters can be used to develop compact hyperspectral/polarization imagers. Such an imager can be readily used for real-time twodimensional spectral imaging applications. These imagers are small, vibration-insensitive, robust, remotely controlled, and programmable and can be used in the spectral region from the ultraviolet (UV) to the near infrared (NIR). A minimal amount of data processing is required for AOTF imagers because they can acquire images at only select wavelengths of interest, and the selected wavelengths can be changed based on the sensing requirements. We use AOTFs made of KDP, MgF2, and TeO2, with a Si-based CCD camera to cover different spectral regions from the UV to the NIR. A liquid crystal variable retarder (LCVR) is used to obtain two orthogonally polarized images at each wavelength The user can write software to control the operation and image acquisition for an AOTF imager for a fully computer controlled operation. Key words: Acousto-optic tunable filter, AOTF, Hyperspectral, Polarization, Imager, Liquid crystal variable retarder, LCVR; Spectroscopy, Fluorescence.

1. Introduction In carrying out absorption, transmission or fluorescence spectroscopy measurements using traditional methods, a monochromator is often used to obtain either the excitation wavelength from a wideband light source or to examine the emission wavelength Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_16

293

294

Gupta

spectrum with a suitable detector or for both. Diffraction gratings or prisms are used in monochromators to cause dispersion of light and provide light at desired wavelengths by using entrance and exit slits. A bandpass filter corresponding to a center wavelength with a narrow bandwidth can also be used in front of a xenon lamp source to obtain the excitation wavelength. A light emitting diode (LED) could also be used to obtain the excitation wavelength. In general, a laser is the best source for excitation of the sample. To examine the emitted wavelengths from the excited sample, a color filter wheel, or an emission bandpass filter corresponding to the emitted wavelength region with a sharp cutoff for the excitation wavelength, can also be used in front of a detector. Such a filter can be an interference filter. In a traditional fluorescence microscope (1) a dichroic mirror is a key element, but it is not able to perform all of the required optical functions on its own. Typically, about 90% of the light at wavelengths below the transition wavelength is reflected and about 90% of the light at wavelengths above this value is transmitted by the dichroic mirror. To select the excitation wavelength, an excitation filter is placed in the excitation path just before the dichroic mirror. Also, to more specifically select the emission wavelength of the light emitted by the sample and to block the excitation light, an emission filter is placed behind the dichroic mirror. In this position, the filter selects the emission wavelength and blocks all excitation light. Recently, a new class of agile vibrationless tunable optical filters have become available that can be electronically tuned to obtain light at a center wavelength with a narrow bandwidth (1, 2). Such filters, when placed in front of a broadband light source, can provide the desired excitation wavelength; when placed in front of a two dimensional detector or a camera it can register an image cube when scanned through a range of wavelength. The two most commonly used filters in this category are the liquid crystal tunable filter (LCTF) and the acousto-optic tunable filter (AOTF). The advantage of using such filters are that there are no moving parts involved because the tuning operation is electronically done and they are quite compact and can be designed to operate over the desired wavelength intervals in the visible and near infrared spectral regions. Filters are the critical element for fluorescence detection used by many biosensors. One of the main limitations of the conventional optical filters used in biosensors is that they are limited to a single wavelength operation while numerous wavelengths are used in fluorescence detection used in biosensors. AOTF or LCTF have the potential to overcome this limitation because they are wavelength agile. An AOTF can also provide polarization signatures due to its polarization selectivity. An LCTF consists of a stack of polarizers and tunable retardation liquid crystal plates in a Lyot-type birefringent filter design

Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral

295

to select a transmitted wavelength range while blocking all others electronically controlled liquid crystal elements. A single Lyot stage of an LCTF system is a sandwich of a quartz plate and a birefringent liquid crystal between two glass plates with a sheet polarizer on either side. To achieve monochromatic throughput, a number of these sandwiches are stacked in a series in order of increased retardance. The bandwidth of an LCTF is constant in frequency space. The passband is electronically tunable over a wide spectral range. The time to change wavelength is of the order of tens of milliseconds and ranges from 50 to 500 ms. An AOTF is a much faster device as the wavelength can be changed in tens of microseconds. We will now focus on AOTFs. An AOTF is a completely solid-state tunable filtering device with no moving parts (3–6). It consists of a specially fabricated prism in a birefringent crystal with a piezoelectric transducer mounted on one of its facets and an acoustic absorber coating applied to the opposite facet. (The thickness of the transducer is usually equal to about one half wavelength of ultrasound in the piezoelectric material). When a radio frequency signal (rf) is applied to the transducer, it generates ultrasonic acoustic waves in the crystal, which traverse the crystal creating a moving three-dimensional grating structure within the period given by the applied rf. When white light is incident on the input facet of the prism, it is diffracted due to the acoustooptic interaction in the birefringent (there are two different values of refractive indices for such material) crystal and there are two orthogonally polarized diffracted beams that propagate at some angle on either side of the incident light and correspond to a single wavelength that is inversely proportional to the applied frequency of the applied rf signal. There are also two orthogonally polarized undiffracted beams in the zero order propagating in the same direction as the incident beam and containing all incident wavelengths except the one that was diffracted. When the incident light beam is polarized, there is only one diffracted beam with polarization orthogonal to the incident beam polarization. The wavelength of the diffracted light can be tuned by changing the frequency of the applied rf signal. The bandwidth of such a filter depends upon the birefringence or difference in the values of two refractive indices as well as the design of the crystal prism and is constant in the frequency space. The passband can be tuned electronically over one octave in wavelength if only one single transducer is used. TeO2 is the most commonly used birefringent crystal that is transparent from 350 nm to 5 µm. An AOTF can be used instead of a color filter wheel or a traditional grating or prism to get narrow band light from a white light beam to excite a sample or in front of a camera or a single detector to detect the emitted light from the sample. In general, only one diffracted beam from an AOTF is used by blocking the other beams.

296

Gupta

A number of AOTF based imaging experiments have been carried out for a variety of applications (7–15). To obtain polarization information an electronically-controlled, commercial spectrally tunable nematic liquid crystal variable retarder (LCVR) is used in the path of the incident beam. The applied voltage to the LCVR is changed corresponding to the horizontal and vertical polarizations for each diffracted wavelength. The diffracted beam from the AOTF is imaged on the camera. As the wavelength filtered by the AOTF can be tuned to any value within the tuning range, we can acquire a spectropolarimetric image cube by tuning over the wavelength range (12–15). Since both the retarder and the AOTF are tuned electronically, no moving parts are involved. This makes the imager adaptive and robust as compared to other traditional hyperspectral imagers. Here we describe how an imaging experiment is set up and how it can be used for fluorescence spectroscopic measurements

2. Materials 2.1. Imager

1. An imaging TeO2 AOTF (Brimrose Corp., Baltimore, MD, http://www.brimrose.com/), specify wavelength range of operation, bandwidth, input aperture (see Notes 1–9). 2. Two irises (Edmund Optics, http://www.edmundoptics. com/US/) (see Notes 10 and 11). 3. Two convex lenses (Edmund Optics, http://www.edmundoptics.com/US/) (see Note 10). 4. Iris and lens holders (Edmund Optics, http://www.edmundoptics.com/US/). 5. Two plane mirrors each mounted on a tilt plate (Edmund Optics, http://www.edmundoptics.com/US) (see Note 10). 6. One spectrally tunable LCVR (model LRC 100, Meadowlark Optics, Frederick, CO, http://www.meadowlarkoptics. com/) without antireflection coating to cover wide spectral range (400–1,800 nm) (see Notes 10 and 11). 7. CCD camera (Watec model 902, Edmund Optics, http:// www.edmundoptics.com/US/) (see Note 12). 8. Camera lens (Edmund Optics, http://www.edmundoptics. com/US/) (see Note 13). 9. Frame grabber (PIXCI® SV4 Imaging Board, EPIX, http:// www.epixinc.com/) (see Notes 14 and 15). 10. Image capture software (XCAP-LITE, EPIX, http://www. epixinc.com/).

Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral

297

11. Computer controlled RF generator (Brimrose Corp., Baltimore, MD, http://www.brimrose.com/), specify rf frequency range of operation and output power (see Note 14). 12. Software to control rf generator (Brimrose Corp., Baltimore, MD, http://www.brimrose.com/). 13. Commercial hyperspectral imaging software ENVI4.3 (ITT Visual Information Solutions, Boulder, CO, http://www. ittvis.com) (see Note 16). 14. Computer. 15. One LCVR controller (model B2010, Meadowlark Optics, Frederick, CO, http://www.meadowlarkoptics.com/) (see Note 14). 16. Light source can be a xenon or mercury lamp or an LED or a laser emitting at the desired wavelength. Use a He–Ne laser for alignment (see Note 17). 17. RF, computer and electrical cables. 2.2. Sample Preparation

1. Sample should be mounted on a microscope slide or placed in a quartz cuvette. 2. We prepared a sample of 1 µg of fluorescein (Eastman Chemical) in one cc of acetone by stirring the powder in the liquid to dissolve it and then put it in a four sided transparent quartz cuvette with a lid.

3. Method The AOTF (see Notes 1–9) used in an imaging experiment needs to be first set up and characterized to make sure it meets the specifications provided by the vendor. The best way to do this is to set it up on an optical table or plate. The most important part of setting up any optical experiment is proper alignment. The optical axis of each component used in this experiment needs to be at the same height. This can be done by making sure that the center of each component is in a horizontal line. First test the AOTF spectral filtering operation with the computer controlled rf generator using a broadband light source to understand its tuning characteristic, i.e., how the diffracted wavelength depends upon the applied rf based on the data provided by the vendor. As mentioned earlier the diffracted wavelength is inversely proportionate to the applied rf. Next, generate two curves for the LCVR based on the manufacturer supplied data to obtain the relationship between the retardance as a function of wavelength and applied voltage: first for the voltage applied to obtain zero

298

Gupta

retardance and second for the applied voltage to obtain quarter wave retardance. Now, test the LCVR retardation operation with its controller. The two convex lenses should be bought together such that their aperture is slightly larger than the input aperture of the AOTF with f/# as close to unity as possible. This will allow most light to be transmitted through the imager. The imager should be set up as shown in Fig. 1. The first iris is used to define the incident field of view for the AOTF and the second one is used to block the unwanted beams from the AOTF and allow only a single diffracted beam onto the camera through the second lens (see Notes 10 and 11). The distance between Iris1 and Lens 1 is equal to the focal length f of the lens. Similarly, the spacing between Lens 2 and Iris 2 is also f. Each lens should be placed a focal length away from the AOTF center to obtain a confocal configuration (12, 16). The AOTF is placed symmetrically between the two lenses and the two mirrors. Each of the two mirrors is placed at 45° angle to the incident beam to fold the optical path and make the imager more space efficient (see Note 10). Each of the mirrors is mounted on a tilt plate with screws on the back. This is very helpful in optical alignment. LCVR is installed anywhere between Iris 1 and Lens1. A coaxial rf cable is used to connect the rf generator and the AOTF. The manufacturer supplied software for the rf generator

LCVR controller Iris1

Lens1

Sample

Mirror1 Light source

LCVR AOTF

RF generator

Frame grabber Iris2

RF signal

Mirror2 Lens2

Computer CCD camera with lens

Fig. 1. Shown is the optical layout for fluorescence or other spectropolarimetric imaging experiments. The components in the left hand side larger box form the optical package for the imager and the one in the smaller box on the right hand side form the computer control. The rf generator is also connected to the computer as shown here.

Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral

299

is installed in the computer and the rf generator is connected to the computer. The LCVR is connected to the LCVR controller. The frame grabber is installed in the computer as well as the image acquisition software and the video output of the camera should be connected to frame grabber (see Notes 14 and 15). After that the power should be turned on and the operation of each component should be checked. To start the optical alignment, the AOTF should be installed last in the setup. First, all the components should be installed as described above except the AOTF. A He-Ne laser should be installed such that its light enters the entrance aperture. Trace the path of light ray using a sheet of paper at each component to make sure that optical axis is in a horizontal straight line. After this, install the AOTF between the two lenses as discussed above. Make sure that the incident light is normal to the input facet of the AOTF (see Note 6). Apply the required rf signal (see Note 4) and put the paper sheet after the AOTF to see the output beams from the AOTF. Adjusting the first iris will help in finding these. After this the second iris should be adjusted to see only the diffracted beam that has been corrected for scene shift on the camera (see Notes 11 and 18). Now the imaging experiment is ready to be used (see Notes 19 and 20). A sample should be placed in front of the light source (see Note 17) such that either transmitted or scattered light is entering the input aperture and an image of the object at the excitation wavelength is visible on the camera. Each spectral image (see Note 20) should be collected with the corresponding radio frequency setting on the rf generator with two retarder voltage settings – first for zero retardance and second for quarter wave retardance for that wavelength. This will provide two ortho-gonally polarized images for each wavelength. Once an image cube is collected it can be analyzed using commercial hyperspectral image processing software ENVI 4.3. Using this software, spectral curves for each pixel can be obtained (see Note 20).

4. Notes 1. The AOTF should be bought with the largest possible input angular and linear apertures for imaging application. For non imaging application small input aperture is acceptable. 2. An AOTF used in the spectral imaging applications is known as a noncollinear AOTF because in this configuration the incident and diffracted light beams, as well as the acoustic beam, propagate in different directions as shown in Fig. 2.

300

Gupta

Absorber Absorber

Upshifted diffracted Beam (vertically polarized) Zero order undiffracted Beams (orthogonally polarized)

Unpolarized incident White light beam

Traveling Acoustic wave

Birefringent crystal

Downshifted diffracted Beam (horizontally polarized) Transducer rf signal

Fig. 2. Operation of a noncollinear AOTF is shown here with the transducer and absorber. When unpolarized white light is incident on the input facet, it gets diffracted by the traveling grating set up in the crystal by the acoustic wave (light blue lines). There two orthogonally polarized diffracted light beams at a wavelength inversely proportional to the applied rf, are coming out at an angle to the incident beam. The zero order beam contains all wavelengths except the one that was diffracted by the traveling grating. The period of the traveling grating is given by the wavelength of the acoustic wave in the crystal and can be changed by changing the applied rf. Only one of the diffracted beams is used for imaging by blocking the rest of the beams.

3. It is important to get the AOTF with a wedge at the output facet to compensate for the spectral scene shift. This is done for only one of the diffracted beams and that is the beam that should be allowed to pass through the second iris. Without such a wedge, different colors will be shifted from each other in the image plane something like what happens when light is dispersed from a prism. 4. It is important not to exceed the rf power above the specification otherwise the crystal and transducer will be damaged. 5. So far TeO2 is the best birefringent material to fabricate an AOTF with the highest diffraction efficiency (the ratio of diffracted light and incident polarized light) (17). TeO 2 AOTFs are generally designed to operate over one octave or less in wavelength and cover the wavelength region from 0.35 to 5 µm. To fabricate filters operating at shorter UV wavelengths KDP (18, 19) or MgF2 (20) can be used. KDP can be used from 220 to 480 nm spectral range and MgF 2 can be used from 190 to 490 nm. The diffraction efficiencies of AOTFs in both these materials are much lower than in TeO2 (17). 6. In the optical arrangement shown in Fig. 1, it is also possible to arrange the two convex lenses in front of the AOTF behind the first iris and the LCVR such that a colli-mated beam of light is incident on the AOTF. This is the classical optical layout in an AOTF imaging experiment (5). The camera lens in this case should be adjusted to receive light from infinite distance.

Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral

301

7. To get a high signal-to-noise-ratio with an AOTF, both the linear and angular input apertures of the AOFT should be as large as possible. Also, the bandwidth should be as wide as possible to meet the experimental requirements. The diffraction efficiency of the AOTF should be as large as possible based on the material and design properties. To improve the spectral image quality, the background image when no rf is applied should be subtracted from the image when rf is applied. The difference gives the best spectral image and a much better S/N is obtained. 8. In our measurements, the TeO2 AOTF was custom designed. Its linear input aperture is 15 × 15 mm2 with an angular aperture of 4°. A wedge was used at the output facet and the filter was designed to be used with vertically polarized output beam. The filter operates from 400 to 800 nm with applied rf from 120 to 50 MHz. The bandwidth is 10 nm at 600 nm or 278 cm−1 in frequency space. The diffraction efficiency of the filter with polarized light is 95% at 0.9 W applied power. The maximum applied rf power used was 1 W. 9. Since AOTFs are sold in small quantities, the cost is high (around $10 K). 10. Two 30 × 30 mm2 plane mirrors are used to fold the light path. These mirrors were mounted on tilt plates for ease in optical alignment. We used two 25 mm diameter irises. The distance between Iris 1 and Lens1 is 25 mm and this is also the distance between Lens 2 and Iris 2. Two single convex lenses with focal length 25 mm each and aperture also equal to 25 mm were used. The first lens was kept 25 mm away from the center of the AOTF in the path of the incident light such that it imaged the scene from infinity in the center of the AOTF. The second lens was kept 25 mm away from the center of the AOTF in the path of the diffracted beams. A commercial LCVR was placed between Iris 1 and Lens1. Its position does not need to be exactly defined. In our optical train after the LCVR, we have used two lenses in a confocal configuration (12, 16), such that the light from a distant object is first imaged in the center of the AOTF crystal and then re-imaged on a commercial 640 × 480 pixel Si-CCD camera with an objective lens. Iris 2 aperture is adjusted before the objective lens to block the downshifted diffracted and the zero order undiffracted beams. 11. Light first enters Iris 1 that defines the angular field of view of the AOTF. Next it enters a the LCVR that is controlled by an LCVR controller to produce two retardances at each wavelength such that the incident light with 0°, and 90° polarization orientations all have 0° polarization after passing through the LCVR. The reason for doing this is that

302

Gupta

in our system only the upshifted diffracted light from the AOTF with a 90° polarization is corrected for the spectral scene shift and imaged on the camera. 12. A low-cost over-the-counter 640 × 480-pixel Si-camera was used. 13. A 25 mm focal length objective camera lens was used in front of the CCD camera. 14. A frame grabber was installed in the computer as well as the image acquisition software. Both the rf generator and the LCVR controller were interfaced with the computer. A graphical user interface was developed for a seamless operation of the imager. The computer controls the applied rf signal’s frequency and amplitude as well as the applied voltage for the LCVR. 15. The CCD output was captured using a frame grabber and stored on a hard drive. 16. Hyperspectral image processing can be carried out by using a commercial software – ENVI 4.3. This software can be used to create an image cube by combining spectral images collected over a range of wavelengths for each polarization and then spectral profiles for any pixel in the image can be extracted quite easily. There are other ways of doing this by using Matlab or Mathematica, but ENVI 4.3 is specifically written to process hyperspectral images. 17. If a laser at suitable wavelength is chosen as the light source, higher quality fluorescence spectral images can be obtained. 18. Every effort should be made to cut off the stray light from entering the camera or better still the whole arrangement should be installed in a box that only lets the light enter the input side of the imager. This will also improve the image quality. Our optical package and the imager are shown in Fig. 3 and a detailed optical layout is shown in Fig. 4. 19. In Fig. 1 the sample is placed for transmission spectroscopic measurements. This is just for illustration purpose. The sample can be configured for a transmission, scattering or reflectance measurement by suitably configuring the light source and sample. 20. In Fig. 5 we show the fluorescence imaging results obtained in our laboratory using a common chemical powder fluorescein dissolved in acetone and placed inside a quarts cuvette. We used a 532-nm neodymium–yttrium laser as the excitation source. The laser was coupled to a multimode fiber. We obtained a collimated light beam from the laser by using a 10 mm focal length convex lens with 10 mm diameter. A sheet

Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral

Iris 2

303

Lens 2

CCD camera with objective lens

Mirror 2 Mirror 1

Iris 1 Lens 1

Imager on tripod

AOTF

LCVR

Fig. 3. An AOTF imager and its details are shown. The black box (15 × 20 × 10 cm3 and weighs around 2 kg) on the right contains the optical package shown in the left. An AOTF and an LCVR are shown below the optical package.

er

nm

las

2-

53

Lens 1

LCVR

Mirror 1

fluorescein sample in quartz cuvette Input Aperture Iris 1

f AOTF f

CCD Camera with lens

Output Aperture Iris 2

Mirror 2

Lens 2

Fig. 4. Shown are the details of the optical layout for the imager shown in Fig. 3. The AOTF is used with two irises, two convex lenses, two plane mirrors mounted on tilt plates, one LCVR, and one CCD camera with an objective lens. The first Iris is used to define the field of view of the AOTF and the second iris is used to pass through only one diffracted beam and block the remaining beams coming out of the AOTF as shown in Fig. 2. The distance between Iris 1 and Lens 1 is equal to the focal length f, this also is the distance between Lens 2 and Iris 2. Each of the two lenses is located f distance away from the AOTF on either side of it. Also, shown is the sample being measured for fluorescence imaging using a 532 nm laser as an excitation source. The imager is detecting a part of light scattered from the sample.

304

Gupta

520 nm_0

530 nm_0

540 nm_0

550 nm_0

560 nm_0

Fluorescence spectrum of fluorescein using 532 nm laser 100

Incident pol 45 deg_detected pol horizontal

90

Incident pol 45 deg_detected pol vertical

Intensity (arb. unit)

80 70 60 50 40 30 20 10 0 500

520

540

560 580 600 Wavelength (nm)

620

640

Fig. 5. Five of the 16 spectral images with horizontal polarization obtained from fluorescence imaging of fluorescein in acetone in a quartz cuvette. The excitation source used was a fiber-coupled frequency-doubled Nd-YAG laser emitting at 532 nm. The light was collimated by using a 10 mm focal length 10 mm aperture lens. A linear polarizer sheet with 45° polarization orientation was placed in front of the collimating lens. The emitted light due to fluorescence was imaged from 500 to 650 nm at 16 spectral bands each with two values of retardance. No filters were used to reduce the light intensity or block the excitation wavelength. The two curves correspond to the horizontal and vertical polarizations of scattered light from the sample.

polarizer was placed with 45° polarization orientation behind the lens. The sample was illuminated and the scattered light was imaged using the imager shown in Fig. 4 from 500 to 650 nm. We collected 16 spectral images to form the image cube. Each spectral band was collected with two polarization orientations – horizontal with zero retardance and vertical with 90° retardance. Five spectral images are shown in Fig. 5 along with the fluorescence plots for the two polarization orientations of the scattered light. Each point was obtained by averaging over 5 × 5 pixels in the center of the image. No blocking filter was used to block the excitation wavelength.

Biosensors Technologies: Acousto-Optic Tunable Filter-Based Hyperspectral

305

References 1. Riechman, J., Handbook of filters, http:// www.chroma.com/index.php?option=com_ content&task=view&id = 27&Itemid=28 2. Gat, N. (2000) Imaging spectroscopy using tunable filters: an overview. Proc. SPIE 4056, 50–64 3. Chang, I. C. (1976) Tunable acousto-optic filters: an overview, in Acousto-Optics: Device Development/Instrumentation/Applications, J. B. Houston, ed. Proc. SPIE 90, 12–22 4. Xu, J. and Stroud, R., Acousto-Optic Devices (Wiley, New York, 1992) 5. Goutzoulis, A. and Pape, D., Design and Fabrication of Acousto-Optic Devices (Marcel Dekker, New York, 1994) 6. Gupta, N., (2003)Acousto-optics, in Optical Engineer’s Desk Reference, W. Wolfe, ed. Optical Society of America, Washington, DC 7. Glenar, D. A., Hillman, J. J., Saif, B. and Bergstralh, J. (1994) Acousto-optic imaging spectropolarimetry for remote sensing. Appl. Opt. 33, 7412–7424 8. Cheng, L. J., Mahone, J. C., Reyes, G. F. and Suiter, H. R. (1994) Target detection using an AOTF hyperspectral imager, in Optical Pattern Recognition V, D. P. Casasent and D. Chao, eds. Proc. SPIE 2237, 251–258 9. Tang, G. C., Chen, J. T., Katz, A., Celmer, E. J., Krumm, R. W. and Alfano, R. R. (1998) Ultraviolet visible acousto-optic tunable spectroscopic imager for medical diagnostics. J. Biomed. Opt. 3, 80–84 10. Gupta,N.,Denes,L.J.,Gottlieb,M.,Suhre,D. R., Kaminski, B. and Metes, P. (2001) Object detection with a field-portable spectropolarimetric imager. Appl. Opt. 40, 6626–6632 11. Gupta, N. (2001) Remote sensing using hyperspectral and polarization images. Proc. SPIE 4574, 184–192

12. Gupta, N., Dahmani, R., and Choy, S. (2002) Acousto-optic tunable filter based visible-to near-infrared spectropolarimetric imager. Opt. Eng. 41, 1033–1038 13. Gupta, N. (2003) Hyperspectral and polarization imaging applications of acoustooptic tunable filters. Proceedings of the World Congress in Ultrasonics, 345–348. http:// www.sfa.asso.fr/wcu2003/procs/website/ articles/000518.pdf 14. Gupta, N. (2003) Hyperspectral and polarization imaging applications of acousto-optic tunable filters. Proceedings of the 32nd AIPR Workshop, 21–26 15. Gupta, N. (2005) Acousto-optic tunable filter-based spectropolarimetric imagers for medical diagnostic applications- instrument design point of view. J. Biomed. Opt. 10, 051802-1-6 16. Suhre, D. R., Denes, L. J. and Gupta, N. (2004) Telecentric confocal optics for aberration correction of acousto-optic tunable filters. Appl. Opt. 43, 1255–1260 17. Gupta, N. and Voloshinov, V. B. (2005) Hyperspectral imaging performance of a TeO2 imaging acousto-optic tunable filter in the ultraviolet region. Opt. Lett. 30, 985–987 18. Gupta, N. and Voloshinov, V. B. (2004) Hyperspectral imager, from ultraviolet to visible, with a KDP acousto-optic tunable filter. Appl. Opt. 43, 2752–2759 19. Voloshinov, V. B. and Gupta, N. (2004) Ultraviolet-visible imaging acousto-optic tunable filters in KDP. Appl. Opt. 43, 3901–3909 20. Voloshinov, V. B. and Gupta, N. (2006) Investigation of magnesium fluoride crystals for imaging acousto-optic tunable filter applications. Appl. Opt. 45, 3127–3135

Chapter 17 Photodiode-Based Detection System for Biosensors Yordan Kostov Summary Many sensors and biosensors are based on the detection of optical changes in the sensing phase. In order to build a stand-alone sensing device, a miniature and low-cost detection system is critical. Here, the method for manufacturing the most critical part (the photodetector) is described in detail. The receiver is based on a photodiode. The practical design of such device is presented here. By using it, it is possible to construct an optical sensor for fluorescence or absorption-based measurement. Discussed here are suitable methods for generation and modulation of probing light, as well as the possible optical configurations. Practical advice is given on the selection of the bandwidth, noise, amplification, etc. The described system was used for detection of green red fluorescent protein in E. coli during fermentation. Key words: Optical sensor, Absorption sensor, Fluorescence sensor, Photodiode, Optical receiver.

1. Introduction Measurements of light intensity are a necessary part of any optical (bio)chemical sensor. Regardless of whether the sensor is based on changes of absorption or fluorescence, whether it measures fluorescence intensity, spectrum, lifetime, or polarization, the main value measured will be intensity (after the light has passed some optical component – i.e. filter, monochromators, polarizer, etc). Measurements of light intensity usually rely on its conversion into electricity. In sensors, the device of choice for the task is often a photodiode. In response to light, it generates current. It is a solid state device (virtually unbreakable). It operates at relatively low voltages, produces linear output and is capable of achieving high speeds. Some of its drawbacks (in comparison with

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_17

307

308

Kostov

photomultipliers) are the lack of internal amplification and higher dark current. The internal amplification is partially addressed in the avalanche photodiode, and the dark current can be decreased by cooling the device. 1.1. Photodiode: Structure and Types

A diode is a semiconductor device which has regions with p and n-type conductivity, separated by p–n junction (Fig. 1). This junction is practically non-conductive because the charge carriers recombine in it. As a result, a region depleted of charge carriers forms around the p–n junction. A photodiode is a diode with the p–n junction exposed to light. If a photon with sufficient energy strikes the lattice, it generates a free (mobile) electron and electron hole. If this event happens in the depletion region (or close to it), these charge carriers are transported away by the field of the junction, effectively resulting in a photocurrent. The p–n junction exhibits significant capacitance. The reason for that is the presence of charges on the both side of very thin layer (the depletion region is several hundred nanometers). This capacitance limits the maximum rate of the light changes which can be observed. In other words, if the light intensity is modulated, the capacitance of the diode determines the maximum frequency at which the photodiode can be used for accurate detection. Another type of photodiode is the PIN photodiode (Fig. 2). Here, between the p- and n-type semiconductor there is a fairly wide region of undoped, intrinsic semiconductor (hence the name PIN). As it is much bigger than the regular p-n junction, the depletion region exists almost entirely inside the i-region. As a result, the capacitance of the PIN diode is significantly smaller. Furthermore,

Fig. 1. Structure of a photodiode.

Fig. 2. Structure of a PIN photodiode.

Photodiode-Based Detection System for Biosensors

309

its response time depends mainly on the time which it takes the charge carrier to drift through the depletion region. As a result, PIN diodes are much faster than the standard photodiode. Another added benefit is that the volume of the depletion region, in which the electron-hole pairs can be generated, becomes larger, thus increasing the chances for the absorbed photon to generate current. However, the PIN photodiodes are usually more expensive than the regular diodes. Still, there are low cost PIN photodiodes, like SFH203 or BPW34, which are used to monitor light modulated above 100 MHz. 1.2. Photodiode Parameters

The main parameters of any photodiode are its responsivity (sometimes called sensitivity), dark current, noise equivalent power (NEP) and response time (1). The responsivity rΦ is given in amperes per watt of the illumination radiation. Knowing it, it is possible to directly calculate the current ip that will be generated by the diode when it is exposed to certain radiant flux Fe: ip = rΦ⋅Fe.

(1)

The dark current is the current that will flow through reverse biased diode in absence of any light. Its value is specific for the diode type. It is one of the major noise sources (the other is the shot noise) and contributes significantly to the errors in the measurement. NEP is the minimum input optical power to generate photocurrent, equal to the rms noise current in a 1 Hz bandwidth. Roughly, it is equal to the minimum detectable power by a photodiode. NEP embodies all sources of noise in the photodiode: shot noise, thermal noise, current noise, etc. While there are many noise sources, all of them degrade the performance of the photodiode in the same way. The response time is the time for which the output current reaches 90% of the steady state when the input light makes a step change. The response time determines the maximum frequency of the incoming light that can be detected with the device. 1.3. Sensing Setups with PD: Absorbance, Fluorescence, Refraction Index Measurements

The optical sensors in general consist of a detector, light source and sensing phase. As stated above, PD will be used as detector. As light source, a lamp, a light emitting diode (LED) or laser diode (LD) can be used. In sensors, the use of LED is more common. The reason for this is the high speed and the low power consumption of the LED. The high speed allows the light from the LED to be modulated electronically up to several hundreds of MHz. This is impossible with lamps (their fastest modulation is several tens of Hz). The modulation is important for separation of the sensor light from the ambient light which can be done by locking in on the modulation frequency. Furthermore, the modulation is important for some measurement methods (i.e. frequency domain

310

Kostov

fluorometry). Another advantage of the LEDs over lamps is the relatively narrow spectrum of the emitted light – its half maximum is approximately 40 nm. In addition, the LED control is extremely simple – it can be turned on or off using a semiconductor switch (Fig. 3). The same switch can be used for modulation (in other words, quick turning it on and off). A drawback of the LEDs is the fact that a portion of their emission can be observed up to infrared, so they need optical filtering. If a narrow-spectrum of light emission is desirable, the LD can be used. In addition, the LD can be modulated to significantly higher frequencies. However, it requires substantially more current to be driven, which can negatively impact the power budget of the sensor. Depending on the sensing principles, optical chemical sensors can be broadly divided into 3 groups – absorbance, fluorescence and refractive index-based (2). In every group, a variety of different optical setups are possible. We describe briefly some of these setups. 1.3.1. Absorbance

In absorbance sensor, the absorbance or the reflectivity can be measured (Fig. 4). In absorbance mode (Fig. 4a), the light

Fig. 3. LED modulation switch. The current through the LED is set by the resistor R and the voltage of the power supply +V. The switch S (SN74LVC1G66 from Texas Instruments) modulates the light by switching the current on and off. The switching is controlled by the voltage level on switch control terminal C.

Fig. 4. Absorbance based setups. (a) Absorbance mode: I0 – incoming light intensity, I – out going light intensity, l – thickness of the sample (b) reflectivity mode: Reflectance R is defined as I/I0.

Photodiode-Based Detection System for Biosensors

311

passed through the sensing phase is attenuated. According to Beer’s law, the absorbance is defined as A = lg

I0 = e ×l ×C , I

(2)

where I0 and I are the intensities of the incoming and the outgoing light, ε is molar absorptivity of the sensing dye, l is the thickness of the sample and C is the concentration of the absorbing species. An absorbing sample could be probed in diffuse-reflectance mode (Fig. 4b). The relationship between the reflection and the concentration C is given by the Kubelka–Munk model: (1 - R)2 e × C (3) = 2R S where R is the reflectivity of the sensing phase and S is a scattering coefficient. The described optical arrangements are also suitable for implementation with fiber-optic waveguides. A special arrangement of the fiber (i.e. bending with small diameter of the fiber or positioning of a mirror) for absorbance measurements might be required. The light source and the receiver are usually attached to an individual fiber that goes to the sensing phase. Orientation of the light source and the detector is not so critical, as the intensity of the detected light is usually high. Often, absorption on specific wavelength(s) is measured. In sensors, a narrow-bandpass interference filter is typically used. It can be positioned either in front of the LED or the detector. 1.3.2. Fluorescence

The major difference between the absorbance and fluorescence sensor is the fact that the emission is at a different wavelength as compared with the excitation light. Using wavelength-selection devices (filters, monochromators) it is possible to a great extent to exclude the excitation light from entering the detector. This results in better signal-to-noise ratio and better sensitivity. However, the fluorescence is isotropic, and under typical conditions, only 1–10% of the total emission is captured. Furthermore, the quantum yield of the fluorophore in use may be low. All that amounts to significantly lower light intensities that has to be detected. The fluorescence based sensors usually require significant amplification by the multi-stage amplifier, more complex optics, and even optimization of the placement of a photodetector with regard to the light source. The right angle configuration (as shown on Fig. 5) is the most favorable one, as it not only avoids the direct illumination of the photodetector but also minimizes the scattered light that travels toward the photodiode.

1.3.3. Refraction Index

Measurements of refraction index in optical sensors are common. They are used for either bulk measurements (i.e. changes in

312

Kostov

Fig. 5. Fluorescence setup. Ex – excitation light, Em – emission light, F1 – excitation filter, F2 – emission filter, LS – light source, S – sample, D – detector.

density), or for studies of surface interactions (through the use of surface plasmons). However, most of the measurements require to detect the change in the direction at which the light travels (angular measurement). With the regular photodiodes this is a difficult task, as their sensing surface is quite big and not sensitive to the position of the light beam. Significantly more often arrays of photodiodes or cameras are used. Their small pixel size makes the beam deflection measurement highly accurate. 1.4. Amplification

Usually, after the conversion of the light into electricity, the signal level is too small for immediate measurements, and it needs to be amplified. This is quite often the case with photodiodes, as they lack internal amplification. For this reason, the amplifier is an integral part of the photodiode sensors (3). In this case the bandwidth of the amplified signal has to be considered. First, it is desirable to work in some way with modulated light, so it could be distinguished from the ambient light. There are two main contributors to the ambient light – the sunlight and the room lights. The sunlight can be considered as constant or very slowly changing during the measurement (frequency range below 1 Hz). The room lights pulse at a variety of fixed frequencies, from 120 Hz for the incandescent lamps to 40 kHz for some fluorescent lamps with electronic ballast. The 120 Hz frequency from the incandescent lamps is a result of the current passing twice through the wire during the 60 Hz period; in both directions, it heats the wire. This results in 120 Hz light pulses (this was a popular approach in the beginning of the twentieth century for frequency doubling). It is highly desirable to avoid direct exposure of the sensor to the ambient light, as this might saturate the photodiode. As it is difficult to completely exclude the ambient light in sensor configurations, it is recommended to use a series of AC amplifiers which block the DC component. They should also be significantly attenuating 120 Hz and the lower frequencies.

Photodiode-Based Detection System for Biosensors

313

On the other hand, it is desirable to limit the upper end of the amplification bandwidth to the light modulation frequency – this will decrease the overall noise and avoid the saturation or self oscillation of the amplifier by the noise peaking. 1.5. Applications

Photodiode-based systems are ideal for applications, where it is difficult to exclude the external light. They are especially well suited for small, hand-held or manually operated systems. One such application is on-line monitoring of the fluorescence from fluorescent proteins in living systems – i.e. during the a fermentation that produces green fluorescent protein (GFP) (5), or the production of DsRed-coupled protein in larvae (6). Below, a photodiode setup with four subsequent amplification stages is described. If high amplification is not needed, any number of stages can be skipped. The amplifier is expected to be used together with modulated light source. The modulation frequency is in the range 1–10 kHz. Measurement of the signal amplitude at specific frequency requires the use of a lock-in amplifier as a user interface. Another approach is to digitize the signal by a fast A/D converter and analyze it in the frequency domain using Fast Fourier Transform (FFT). The use of the system for GFP detection is described.

2. Materials 2.1. Parts

1. PIN Photodiode – BPW34s (Osram). 2. Op Amps. IC1 and IC2: 2xOPA2355 (Texas Instruments) or 2xOPA2354 (Texas Instruments). Any other dual CMOS operational amplifier capable of 5 V single supply operation can be also used. 3. Capacitors: C1 – 0.5 pF, C2, C3, C4 – 1 nF, C5, C6, C7, C11, C13 – 1 µF, C8, C9 – 0.1 µF, C10 – 10 nF, C12 – 33 µF. Package: SMD, size – 0603, accuracy: ±10%, voltage rating – 6.3 V, type – ceramic. 4. Resistors. R1 – 3 MOhm, R2, R3, R4 –10 kOhm, R5, R6, R7, R8 – 1 kOhm, R9 – 100 kOhm, R10 – 375 kOhm, R11 – 5 Ohm. Package: SMD, size – 0603, accuracy: ±10%, type – thin film. 5. Power supply. Low-dropout regulator TPS793475 (4.75 V, Texas Instruments) and voltage reference TLV431(Texas Instruments). The power can be supplied from a wall transformer (6 V from Radioshack) or derived from the USB bus of a computer.

314

Kostov

2.2. Practical Schematics and Board Layout

The schematic of a practical high amplification photodiode detector is given on Fig. 6. IC1A and IC1B are respectively one-half of IC1, and IC2A and IC2B – 1/2 of IC2 from the list of parts. (OPA 2355 or OPA2354 are dual op amps). The op amps are CMOS amplifiers with very low noise, high bandwidth and low polarization current. The decoupling capacitors (C9, C10) are needed one per integrated circuit. Their primary purpose is to block the DC and attenuate the low frequency interference. TLV431 is needed to generate a voltage reference for the positive input of the op amp. Its output voltage is adjusted to the middle of the power supply, 2.37 V. The power supply itself is based on a regulator that can operate with minimum difference between the input and output (∼100 mV). For better stability, its input and output are additionally filtered with 1 µF ceramic capacitor and low pass filter comprised of R11 and C12.

3. Methods 3.1. Board Preparation

When making a prototype for the photodiode-based sensor, it is a good idea not to use breadboards. While breadboarding is intended to speed-up the development process, the fact that the photodiode has very high output impedance (> 10 MOhm) makes the sensor very susceptible to electromagnetic interference. The

Fig. 6. Schematics of high-amplification AC photodetector. Using the elements in the text, the circuit delivers 3.109 V/A transimpedance gain, bandwidth 1 – 10 kHz and noise 200 µV/Hz.

Photodiode-Based Detection System for Biosensors

315

typical breadboard has large parasitic capacitance, which makes it unsuitable for this type of prototypes because of the possible capacitive coupling of the interference. It is also a good idea to use surface-mount elements from the beginning. While they are more difficult to work with as compared with the throughhole devices, most of the terminals can be on one side of the board, leaving the second side to serve as a shield provided it is grounded. The board layout is shown on Fig. 7a, and the assembly drawing (positions of the electronics parts and their orientation) – on Fig. 7b. The board layout can be photocopied, scaled appropriately and printed on computer transparency to serve as a mask. Alternatively, it can be recreated using a computer CAD program for PCB design. If such program is not readily available, the free version of Eagle 4.1 (Cadsoft, www.cadsoftusa.com) can be used. The board layout is double sided with the bottom side being solid copper, with minimum number of vias (metalized through holes that connect the top and the bottom of the board). If only one board is to be prepared, it is possible to buy photosensitized boards (Radioshack), attach the printed on the computer transparency mask and expose the board to sunlight (3 min) or 100 W UV lamp (365 nm) from a distance of 6" for approximately 3 min. Using a developer (water solution of 0.1 mM NaOH or commercially available one) the exposed photoresist is removed; after several washes with water, the board is etched with solution of FeCl3 (pH adjusted to 2), or with commercial copper etching solution (Radioshack). Finally, the rest of the resist is removed with acetone or isopropyl alcohol. The electrical connection between the top and the bottom layers is achieved using wire bridges. If a higher quality board is desirable, the CAD program can be used to generate the files for board production (drill and Gerber files). There are many board houses that will produce the boards (usually, the minimum number of boards they would produce is 3).

Fig. 7. (A) Board layout of a 10 kHz photodiode amplifier (B) Mounting drawing.

316

Kostov

The turnaround time is around 5 days (the shorter the production time, the higher the price), with the price per board as low as $15 in this minimum quantities (i.e. www.pcbfabexpress.com). A simple search on Intenet under “PCB boardhouse” gives a choice of manufacturers. 3.2. Board Assembly

The parts need to be soldered on the board. This can be done one by one using soldering iron, solder wire or soldering paste and tweezes. Down to size 0603, resistors and capacitors can be mounted by hand. The assembly of the integrated circuits is a bit more challenging (see Note 1). A magnifying glass is of great help when working with surface mount devices. Alternatively, the soldering can be done in batches by heat reflow. An in-house batch soldering can be done using a toaster oven (Caution: do not use this oven for food preparation!). First, the contact pads are covered with soldering paste (either dispensing it using a syringe and a needle, or by applying the paste through a stencil – low-cost plastic stencils can be manufactured for as low as $30, i.e. www.pololu.com). Then, the parts are positioned on the mounting pads. Minor adjustments are easy, and the surface tension of the paste will keep them in place until heating. Finally, the assembled board is placed in a temperature controlled toaster oven, and the following temperature profile is followed: 90°C – 4 min, 165°C – 2 min, 220°C – 30 s. The long warm-up period removes the moisture from the parts, during the second interval the resin is melted to cover the pads with protective antioxidant layer, and during the last one the solder melts and wets the pins. It might be desirable to tap (lightly) the oven just before taking out the board – this helps the solder surface tension to adjust the orientation of the parts. The oven is than turned off and opened to cool down. For more detailed photo description of the toaster method, the reader is referred to www.sparcfun.com. Alternatively, most of the boardhouses can also assemble the board. There are also independent assemblers. However, this is recommendable when there is high degree of confidence in the design – the assembly is significantly a more expensive procedure than board preparation.

3.3. Board Testing

Once the board is assembled, it must be visually inspected for shorts (bridges of solder, unetched portions of the copper layer). Keep in mind that bad electrical connections can happen even with commercial boards! Then, it must be light proofed – the material of the board is a glass filled composite, which transmits just enough light to distort or even compromise the measurements. To make things worse, the copper reflects fairly well long-wavelength light (above 550 nm). As a result, the sandwiched between two copper layers board works as an unintended light guide, which can deliver interfering light from LED at significant distances. Hence, is a

Photodiode-Based Detection System for Biosensors

317

good practice to cover with black paint everything on the board except for the photodiode. Before the actual measurements can start, the photodetector should be tested. First, before switching it on, a visual inspection is required. One needs to confirm that there are no missing parts, misplaced parts, shorts, etc. If there is a problem, it needs to be corrected before the power-up. When testing the board, a modulated light should be provided. The easiest way is to drive a LED from the output of a function generator. Any shape of the LED excitation voltage is acceptable. As the transimpedance gain is significant (1 nA on the output produces 3 V on the output), it is highly desirable to perform the testing at a reduced level of the ambient light. This will help to avoid the saturation of the output stages of the amplifiers. Illumination of the photoreceiver by the LED from various distances should produce various amplitudes of the output signal. This should be verified on the output of every stage (on the traces connecting C5 and R5, C6 and R6, C7 and R7) with an oscilloscope. Testing of the board should be done thoroughly – with the light source on and off, with the room lights on and off. A proper characterization of the photodetector helps to avoid complications down the road to a fully assembled sensor. 3.4. Measurement

Prior to measurements, the optics (filters, lenses, mirrors) should be mounted in front of the photodiode. The outside rim of the filter or other optics should be covered with black rubber coating (quite acceptable is the “liquid rubber” black cable insulation available from Home Depot) to avoid unwanted light leakage. If high amplification is used, it is also necessary to electrically shield the photodetector (see Note 2). Only after that the photodetector can be positioned with the optical path directed toward the sensing phase (i.e. cell with the absorptive or fluorescent material). The board must be firmly attached to the same base that holds the cell. The optical path (i.e. the positions of the LED, cell and photodetector) should be held constant during the measurement. The output can be measured directly with an oscilloscope to verify the shape of the signals and for some crude measurements. For accurate measurements, the signal should be routed to a lock-in amplifier or A/D converter. The actual measurements of light intensity should be performed by determination of the AC amplitude of the output signal at the modulation frequency for the light source. In the analog domain, this is done using lock-in amplifier. Such an amplifier can be standalone (high-quality units are available for Stanford Research Systems or Perkin-Elmer) or it can be constructed on board using the single chip AD630 from Analog Devices (the reader is referred to the data sheet of the integrated circuit, which shows how to implement it as lock-in detector). Alternatively, the measurements

318

Kostov

can be done in the digital domain by digitizing the output voltage using fast A/D converter. Currently, there is a variety of USBpowered ADCs that are easy to operate. The sampling frequency must be at least twice the frequency of the signal (Nyquist limit) to accurately reconstruct the signal. Digital filtering, digital lock-in techniques, or Fourier transform can be implemented to extract the amplitude of the signal. This approach can be combined with visualization and storage software (i.e. using platforms like Labview or Matlab) to facilitate the readout and archiving. 3.5. Fluorescent Protein Concentration Measurements

Photodiode-based detection was used in several devices for measurements of the concentration of fluorescent proteins in vivo in E. coli and larvae (5, 6). Here we will present an application of the photodiodes for on-line measurement of the concentration of green fluorescent protein (GFP) during bacterial fermentation in shake flasks. The overall setup is shown in Fig. 8. The system is designed as a thin “coaster”, which is placed under the shake flask in the shaker (Fig. 8a). The black plastic case serves simultaneously as a stand for the flask and protects the electronics from leaks. It also has a liquid-tight window made of polycarbonate to allow for passage of the light. Block-schematic of the device is presented in Fig. 8b. The system carries on-board the photodiode amplifier, a singleIC clock (LTC6900, Linear technology) for LED modulation and a lock-in amplifier. The accuracy of the clock is ∼0.5%; if a higher frequency stability is needed, a quartz-stabilized clock should be employed. In this particular case, due to space- saving requirements, the lock-in amplifier is designed from discrete switches and op amps to emulate the internal architecture of AD630. If a lock-in amplifier with better noise parameters is needed, additional filtering can be implemented directly on the analog switches (7), or a commercial device can be used. The excitation of the GFP fluorescence was performed using high brightness (6,000 mcd) blue LED (emission maximum 450 nm). Bright blue LEDs are available from several vendors, i.e. Digikey or Newark. Its light was filtered using blue glass (BG12, Schott). The filter rejects practically all green and red emission from the LED and is non-fluorescent. The LED was modulated using the approach depicted at Fig. 3. The optical port of the photoreceiver (see Note 3) was filtered with an interference filter centered at 535 nm and bandwidth 50 nm from Intor (Soccorro, NM). The filter transmits 75% of the light in the pass band. Outside of the band, the absorption of the filter is > 4 (attenuation higher than 104).The photodetector output was connected to the lock-in amplifier. It was important to orient the excitation light at 45° so its reflection does not directly strike the detector filter. This was achieved by bending the LED pins at the desired direction (Fig. 8c).

Photodiode-Based Detection System for Biosensors

319

Fig. 8. Photodiode-based fluorometer. (a) fluorometer under the shake flask (b) Block schematics of the fluorometer. ExF, EmF – excitation and emission filters, PD – photodiode, BF – bottom of the flask (c) Board equipped with the photodiode amplifier (the photodiode and the first two stages are under the shied). Board features also power supply, LED driver, lock-in detector and optics.

The power for the system was provided from the USB port of a personal computer. In order to remove the possible spikes and interferences the +5 V from the USB was down regulated using low-dropout regulator. The used op amps can operate without any problems from +4 V; we have chosen to regulate the supply at 4.75 V, using TPS793475 (Texas Instruments). The system output was measured using USB powered 12 bit ADC (U12, LabJack Inc., Lakewood, CO). The readout from the ADC was done using LabView 7.1. The system was capable to measure GFP concentrations over a range of 3 orders (Fig. 9). The sensitivity was 12.4 V.L/g. The background from fermentation media (0.048 V) was about twice the signal from deionized water (0.023 V). The limit of detection was 0.52 mg/L, or 1.9 × 10−8 M, with an absolute noise ±0.002 V (see Note 4). The system was used successfully for monitoring of GFP production. In a similar setup, the system was used for mon-

320

Kostov

Fig. 9. Output voltage as a function of the GFP concentration.

itoring of infection of larvae (6), where the larvae were attacked by a baculovirus expressing GFP.

4. Notes 1. Op Amp mounting. The board layout allows using either OPA 2355 or OPA 2354 as amplifiers. OPA2355 features additional shutdown pins, which do not affect the work of the amplifier if left unconnected. If OPA2354 is used, care should be taken to fit it on the pads so that the spots for pin 5 and 6 (control pins for OPA2355) should remain unoccupied. When soldering by hand small-pitch surface mount integrated circuits, the easiest way is to flood solder all the pins and then remove the shorts with desoldering braid. If a solder paste is used without stencil, it tends to form microminiature solder balls. They can cause very annoying intermittent shorts; the remedy for that is to thoroughly wash the board with isopropyl alcohol using fine brush. 2. Shielding. Electrical shielding is of immense importance for systems where the light source is closely positioned to the photodetector, as the unshielded photodiode can pick up the modulation of the LED through capacitive coupling. Typically, the best solution is to put the board in a metal box that is electrically connected to the common terminal of the circuit. This approach requires accurate positioning of the board against the box opening (aperture) which allows the incoming light in. An alternative is to use cap made of tinned iron sheet which is soldered directly to the perimeter of the board. In this approach, it is necessary to ensure good electrical contact between the grounds on the top and the bottom layers.

Photodiode-Based Detection System for Biosensors

321

A notch in the shield over the output trace would prevent short-circuiting the amplified signal. It is not necessary to flood solder the shield; tacking it every 0.5" as it achieves the same level of shielding. This practice allows for much easier removal of the shield in case of necessity. 3. Photodiode vs. photomultiplies. The presented system was an all-solid-state solution that was designed as a replacement for a photomultiplier based system. The correlation between the results measured with either of the systems was 0.999. This poses the question whether it is reasonable to use photodiodes as a detector. It depends on a combination of factors: whether the sensor should be rugged, whether it can tolerate the presence of high voltage and possibly breakable parts, whether there is a chance of (accidental) exposure to strong ambient light, etc. Usually, the most compelling reason not to use photodiodes is the low level of the measured light. At low intensities (>3 × 107 photons per ms at the photodetector) the signal to noise ratio of the photodiodes begins to deteriorate – above that, it is close to the theoretical limit (4). Below 107 photons per ms the SNR of a photodiode becomes worse than the one of a photomultiplier.A major reason for that is the dark noise of the device. Cooling down the detector can significantly improve the noise, but adds complexity to the sensor and increases the power budget. In general, the rule of the thumb is: if the intensity is above 107 photons per ms, use photodiode. However, only head-to-head comparison with the specific devices in place can show which detector is better. 4. Bandwidth vs. noise. The noise of an amplifier is an important feature when measuring low-level light. In this case, the low-noise level comes with a price – severe limitation of the bandwidth. The described photodetector acts as a band pass filter – its maximum amplification is between 1 and 10 kHz. The lower cut-off frequency is set by the RC chain R5C5 (R6C6 and R7C7, respectively). It is responsible for DC-blocking and attenuating the ever-present 60 and 120 Hz. The high cut-off frequency is set by R2C2 (R3C3, R4C4). and its variation has multiple consequences. First, the change of the resistor value varies the overall amplification. The feedback capacitor values influence both, the bandwidth of the amplifier (in other words, the maximum modulation frequency of the light source that will be amplified) and simultaneously the noise (Fig. 10). It could be seen that without feedback capacitors the network exhibits a peak in the amplification at 500 kHz (Fig. 10b) and noise levels of ∼0.8 V (Fig. 10a)! Introduction of the capacitor C1 (i.e. 0.2 pF) decreases the noise ∼20 times, but the bandwidth drops to ∼400 kHz. Further introduction of C2 (Fig. 10c, d) reduces the noise another ten times, but at the price of

322

Kostov

Fig. 10. Effect on the values of C1, C2 and C3 on the noise and bandwidth of the photodetector.

bandwidth being shrunk down to 10 kHz. Adding C3 finally reduces the noise to the target 3 mV without more changes in the bandwidth (Fig. 10e, f). Addition of C4 has very small effect, the effect on the values of C1, C2 and C3 on the noise and bandwidth of the photodetector is shown in Fig. 10. The 10 kHz bandwidth is sufficient for amplitude measurements of absorbance or fluorescence. However, if a higher bandwidth is

Photodiode-Based Detection System for Biosensors

323

needed, the capacitor values can be changed to accommodate that. It is a good idea to model the frequency response of the new network using a PSpice program (i.e. TINA-TI, available free of charge from Texas Instruments) when changing any of the passive elements values.

References 1. Gowar, J. (1993) Optical Communication Systems, 2 ed., Prentice-Hall, Hempstead, UK 2. Kostov Y., Rao G. (2000) Low-cost optical instrumentation for biomedical measurements. Rev. Sci. Instr., 70, 4361–4374 3. Graeme, J. (1995) Photodiode Amplifiers: Op Amp Solutions, McGraw-Hill, New York, NY, 4. Wu, J.-Y., Cohen, L.B. (1993) Fast Multisite Optical Measurement of Membrane Potential. In Fluorescent and Luminescent Probes for Biological activity. (Mason, W.T., ed.) 1st ed., Academic Press, London, UK

5. Kostov Y., Albano C.R., Rao G. (2000) All-solidstate GFP sensor. Biotech. Bioeng. 70, 473–477 6. Märtens, O. (2000) Precise synchronous detectors with improved dynamic reserve, IEEE Trans. Instr. Meas. 49, 1046–1049 7. Dalal, N.G., Cha, H.J., Kramer, S.F., Kostov, Y., Rao, G., Bentley, W.E. (2006) Rapid noninvasive monitoring of baculovirus infection for insect larvae using green fluorescent protein reporter under early-to-late promoter and a GFP-specific optical probe. Process Biochem. 41, 947–950

Chapter 18 Photodiode Array On-chip Biosensor for the Detection of E. coli O157:H7 Pathogenic Bacteria Joon Myong Song and Ho Taik Kwon Summary An integrated circuit (IC) of photodiode array (PDA) microchip system was used for the on-chip detection of E. coli O157:H7 based on an enzymatic bioassay and light absorption property of the reaction product. The PDA microchip consisting of an array of 12 × 12 photodiode detection elements served as a photosensor as well as a protein-immobilizing sample platform. As a result, E. coli O157:H7 could be detected directly on the surface of PDA detection elements. E. coli O157:H7 was detected by forming a “sandwich-type” enzymatic immunocomplex on the PDA detection elements using an on-chip bioassay. The quantitative analysis of E. coli O157:H7 immunocomplex was carried out based on the light absorption property of the enzymatic reaction products of E. coli O157:H7 immunocomplexes with respect to a red beam produced by light emitting diodes (LEDs) installed right above the PDA microchip. During the on-chip bioassay, the wet photodiode detection elements exposed to a lot of biological materials or buffer solutions were capable of maintaining their photosensing capabilities. The portable PDA on-chip biosensor permits direct optical detection of E. coli O157:H7 and eliminates the necessity of the conventional expensive microplate reader that is incompatible with the size of the protein microarray. Key words: Bipolar photodiode array microchip, On-chip bioassay, Biosensor, E. coli O157:H7 immunocomplex, Protein microarray, Light absorption.

1. Introduction Rapid and simultaneous monitoring of protein or DNA microarray has been performed using commercially available detectors such as microplate reader or microscope (1, 2). However, these detectors are bulky and not compatible with the biological microarray in size although their detections are highly sensitive, accurate, and reproducible. The construction of a miniature integrated biochip device containing a detector such as the complementary metal oxide semiconductor (CMOS) microchip can be Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_18

325

326

Song and Kwon

suggested as a solution to solve such incompatibility (3). In order to achieve further, a miniature biochip system, PDA on-chip detection system is an attractive tool (4). In this system the bipolar PDA microchip plays the dual role of a DNA/protein immobilizing sampling platform and of a phototransducer. Consequently, a complicated optical alignment in the detection system could be completely eliminated and further a miniature and simple integrated field-usable optical biochip system could be produced. In this work, the scheme of on-chip detection of the pathogenic bacteria E. coli O157:H7 involves three major steps: (1) immobilization of E. coli O157:H7 on the PDA surface as sandwich-type immunocomplexes, (2) carrying out an enzymatic bioassay involving the immunocomplexes that covers the PDA detection elements with enzymatic reaction products, and (3) employing the light absorption property of the enzymatic reaction products against the red beam irradiation from the LEDs. Immobilization of anti-E. coli O157:H7 antibody on the PDA detection elements is achieved by sequential treatment of the PDA surface with silanization reagent (3-aminopropyl) triethoxy silane or APTES), glutaraldehyde, and anti-E. coli O157:H7 antibody. The sandwich type of E. coli O157:H7 immunocomplexes on the surface of PDA are composed of anti-E. coli O157:H7 antibody, E. coli O157:H7 bacteria, and alkaline phosphatase-labeled anti- E. coli O157:H7 antibody. Reaction between alkaline phosphatase in the immunocomplex and nitroblue tetrazolium/5bromo-4-chloro-3-indolyl-phosphate (NBT/BCIP) formed the blue precipitates. The intensity of the irradiated red beam reaching the PDA detection elements is inversely proportional to the amount of blue precipitates produced by the enzymatic reaction on the surface of PDA. Based on this light absorption property, quantitative analysis of E. coli O157:H7 bacteria was achieved.

2. Materials 2.1. Surface Chemistry

1. 10% (v/v) 3-Aminopropyltriethoxysilane (APTES) (Catalog No. 440140, Sigma-Aldrich, St. Louis, MO) was prepared in dry toluene. 2. Phosphate buffer solution of pH 7.0 was prepared by dissolving 0.3402 g of KH2PO4 (Catalog No. P5379, Sigma-Aldrich, St. Louis, MO) and 0.4180 g of K2HPO4 (Catalog No. P2222) in 500 mL of water. 3. 5% (v/v) glutaraldehyde (Catalog No. 16310, Electron Microscopy Science, Fort Washington, PA) solution was made in phosphate buffer solution.

Photodiode Array On-chip Biosensor for the Detection of E. coli

2.2. Formation of E. coli Immunocomplex

327

1. Affinity purified antibody to E. coli O157:H7 (Catalog No. 01-95-90, Kirkegaard & Perry Labs., Gaithersburg, MD) stock solution: The lyophilized product was rehydrated per supplier’s protocol by dissolving the pellet in 1 mL of aqueous glycerol (50%) (antibody concentration = 1.0 mg/mL). The stock solution was stored at −20°C. Working solutions were prepared by diluting the stock solution with PBS immediately before use. 2. E. coli O157:H7 positive control (Catalog No. 50-95-90, Kirkegaard & Perry Labs., Gaithersburg, MD) stock solution (3 × 109 cfu/mL): Dextran stabilized product was rehydrated per supplier’s protocol by dissolving the lyophilized pellet in 1 mL of aqueous glycerol (50%). The stock solution was stored at −20°C. Working solutions were prepared by diluting the stock solution with PBS immediately before use. 3. Alkaline phosphatase-labeled affinity purified antibody to E. coli O157:H7 (Catalog No. 05-95-90, Kirkegaard & Perry Labs., Gaithersburg, MD) stock solution: The lyophilized product was rehydrated per supplier’s protocol by dissolving the pellet in 1 mL of aqueous glycerol (50%) (antibody concentration = 0.1 mg/mL). The stock solution was stored at −20°C. Working solutions were prepared by diluting the stock solution with PBS immediately before use. 4. BCIP/NBT enzymatic substrate solution: One part each of the BCIP concentrate and the NBT concentrate of the BCIP/ NBT phosphatase substrate system (Catalog No. 50-81-00, Kirkegaard & Perry Labs., Gaithersburg, MD) were mixed with the ten parts of the supplied 0.1 M Tris-HCl buffer solution before use. 5. Ethylenediamine tetraaceticacid (EDTA) solution: EDTA solution (9 mM) was prepared in 10 mM Tris-HCl medium.

3. Methods 3.1. PDA Microchip System

1. The PDA microchip system was fabricated using conventional bipolar semiconductor technology. As shown in Fig. 1a, the PDA microchip system consists of a PDA microchip containing a 12 × 12 array of photodiodes, LEDs, and a test board. The individual photodiode detection element is composed of a photodiode with 300 µm diameter, current amplifiers, and a current-to-voltage converter (Fig. 1b). The pixel-to-pixel distance is 100 µm. 2. The photodiode detection element is fabricated by the incorporation of the P-type implant layer on the N-type epitaxial

328

Test board

Song and Kwon

LED

PDA microchip

Fig. 1. (a) Photograph of the bipolar PDA microchip system. (b) Schematic diagram of the PDA microchip. The photocurrent of the individual photodiode is amplified by the current amplifier and converted to voltage by the current-to-voltage converter.

layer. A vertical structure of a photodiode detection element is shown in Fig. 2. Highly concentrated P-type silicon barriers were placed at both bottom ends of P-type silicon substrate for isolation of the individual photodiode detection element. The concentrated N-type buried layer was formed on the P-type substrate in order to reduce the resistance of the N-type epitaxial layer. Then, the N-type epitaxial layer was formed on the buried layer. The PN junction was achieved by the incorporation of the P-type implant into the N-type epitaxial layer. The N-type implant was incorporated to connect the N-epitaxial layer and Al metal. The surface of the fabricated PN junction layer was then oxidized. The contact layer was obtained by partial removal of the oxidized layer using a buffered oxide etchant, where the first metal wiring was performed. Al metal was deposited on the contact layer. The metal wiring process was completed on the deposited Al metal layer using photolithography. An Al striper was used as

Photodiode Array On-chip Biosensor for the Detection of E. coli

329

2’nd Metal 1’st Metal Oxide Layer (SiO2) Emitter Implant Layer (N-type)

Contact

Base Implant Layer (P-type)

Isolation (P-type) Epitaxial Layer (N-type) Bottom Isolation (P-type) Buried Layer (N-type) Substrate (P-type)

Fig. 2. A vertical structure of a photodiode detection element in the PDA microchip.

an etchant. The first metal wiring layer was coated with an insulation layer of silicon oxide to perform a second metal wiring on it. The second metal wiring was also completed using photolithography. Current amplifiers and current-to-voltage converters (4) in individual photodiode detection elements were fabricated by this metal wiring based on photolithography. As a result, the individual pixel consists of photodiode, current amplifier, and current-to-voltage converter. These components are integrated in an unit pixel and the PDA microchip contains these 144 pixels. The test board transfers photosignals from the PDA to a laptop computer. An array of LEDs was placed right above the PDA microchip to irradiate the PDA microchip with the red beam. 3. The individual photodiode can be addressed and read using digital I/O lines and an analog-to-digital conversion channel supplied by an RS232 interface to the laptop computer. Custom software written in the C language was used for data acquisition process. Figure 3 shows an algorithm of the custom software. There is an AT89S52 Micom (Micro-computer) in the test board. The Micom is operated by the custom software. Communication between the test board and the software in a laptop computer was achieved via the RS232 interface. The communication is initiated with a start signal in the software. When the Micom receives the start signal, the Micom sends a command to the LEDs installed right above the test board so that the LEDs beams are emitted. The produced LEDs beam is irradiated on the PDA microchip. After the emission of the LEDs beams, the Micom tests all the pixels in order to read output signals from the PDA. The Micom can access all the 12 column pixels in the first row one by one using row and column address decoders. Identical tests are performed with respect to all the 12 column pixels in

330

Song and Kwon

Test F/W Flow Chart Initialization

N Test Start?

Y PWM Pulse generation : -. 2 usec x PWM level timer running -. Interrupt Enable -. Delay 600 msec

Set Test Column and Row cell

Ready to read ADC0802

N

Does ADC interrupt generate?

Y Read ADC data Send ADC data to PC

Y Is any test column and Row cell ?

N Fig. 3. An algorithm of the custom software that operates the entire PDA microchip system. The software was written in the C language.

the remaining 11 rows. The output signals from the PDA are converted to digital values by the 8-bit ADC (analog to digital converter) in the test board (5). Then Micom reads output signals and transfers those signals to the laptop computer via the RS232 interface. Figure 4 shows a block diagram to represent an operation process of the test board. 4. As shown in Fig. 5, E. coli O157:H7 immunocomplex was immobilized directly on the surface of photodiode detection elements.

Photodiode Array On-chip Biosensor for the Detection of E. coli

Row[1..12] Col[1..12] Analog S / W_1

Analog S / W_13

OUT

331

8bit_ADC

Col[1..12] Analog S / W_2 Light Source Col[1..12] Analog S / W_3

Row_1 Row_2 Row_3 12 by 12 Row_4 Photosensor Row_5 Array Row_6

Row Decoder

Col[1..12] Analog S / W_4

Col[1..12]

Analog S / W_5 u-COM

PC

… Row_12

Col[1..12] Analog S / W_6

PWM

… Col[1..12] Analog S / W_12 Column Decoder

LPF

To Light Source

Fig. 4. A block diagram that shows an operation process of the test board containing the PDA microchip. The analog-todigital converter and column and row address decoders are contained in the test board.

Fig. 5. A schematic diagram of the protein microarray directly immobilized on the surface of PDA detection elements. The E. coli O157:H7 immunocomplex was formed by sequential on-chip antibody–antigen reactions. Blue precipitates were produced as a result of enzymatic reaction between alkaline phosphatase-labeled anti-E. coli O157:H7 antibodies and NBT/BCIP.

332

Song and Kwon

3.2. Test of PDA Microchip System

1. The blank signal of the PDA microchip system was measured by irradiating a bare (virgin) chip with the red beam from LEDs in a dark room. The PDA microchip system was manipulated so that the signal becomes 2.5 when the red beam is irradiated onto the bare PDA microchip.

3.3. Surface Chemistry

1. First a set of six PDA microchips was treated with 10% 3-APTES solution at 115°C in an oven for 2 h. After 2 h, each chip was rinsed thoroughly with toluene and dried at room temperature. 2. The silanized chips were then treated with 2.5% glutaraldehyde solution for 1 h at room temperature. Each chip was washed thoroughly with potassium phosphate buffer (pH 7.0) to remove any unbound glutaraldehyde. 3. For immobilization of anti-E. coli O157:H7 antibody on the PDA microchip, the anti-E. coli O157:H7 antibody stock solution (1 mg/mL) was diluted with PBS to a final concentration of 20 µg/mL. Eighty microliter of the diluted solution was spotted onto the photodiode detection elements of each chip, and the chips were incubated overnight at 4°C. The unbound anti-E. coli O157:H7 antibodies were rinsed off with PBS.

3.4. Formation of E. coli Immunocomplex

1. Anti-E. coli O157:H7 antibody-immobilized chips were then subjected to 10% BSA blocking solution (Catalog No. 50-61-01, Kirkegaard & Perry Labs., Gaithersburg, MD) for 1 h at room temperature. 2. The E. coli O157:H7 stock solution (3 × 109 cfu/mL) was serially diluted with PBS to get six E. coli O157:H7 sample solutions of concentrations 5.0 × 104, 2.0 × 105, 7.5 × 105, 2.0 × 106, 5.0 × 106, 1.0 × 107 cfu/mL. 3. Eighty microliter of each of E. coli O157:H7 sample solutions were then spotted separately on six BSA-treated PDA microchips. The chips were incubated for 1 h at room temperature. Unbound E. coli O157:H7 was rinsed off with PBS. 4. The alkaline phosphatase-labeled anti-E. coli O157:H7 stock solution was diluted with PBS (1:50). Eighty microliter of the diluted solution was spotted onto the PDA microchip to react with E. coli O157:H7 captured by the anti-E. coli O157:H7 antibody immobilized on the photodiode detection elements. 5. After 1 h incubation at room temperature, the unbound alkaline phosphatase-labeled anti-E. coli O157:H7 antibodies were rinsed off with PBS.

3.5. Enzymatic Reaction for Producing Blue Precipitate

1. 40 µL of the BCIP/NBT working solution was added onto the PDA microchip containing the E. coli immunocomplexes. The optimal enzymatic reaction between the BCIP/NBT

Photodiode Array On-chip Biosensor for the Detection of E. coli

333

and the alkaline phosphatase of the E. coli immunocomplex was achieved by incubating the chips at 37°C for 1 h. The amount of blue precipitates produced by the enzymatic reaction is proportional to the number of E. coli O157:H7 attached to the surface. 2. The enzymatic reaction was ceased by adding 40 µL of Tris-EDTA buffer solution (final EDTA concentration 1 mM) onto the PDA microchip. 3.6. Detection on the PDA Microchip

1. The quantitative detection of E. coli O157:H7 was performed by the irradiation of the red beam onto the PDA microchip covered with blue precipitates. 2. Control signal is the average digital value (Sc) obtained from all the pixels when the red beam is irradiated onto an entire PDA microchip whose on-chip bioassay is performed without E. coli O157:H7. 3. Each PDA chip was irradiated with the red beam and the average signal was obtained as a function of the number of E. coli O157:H7. The average digital value obtained from 12 photodiode detection elements at this time is Ss and the magnitude of Ss is dependent on the number of E. coli O157:H7. The magnitude of Ss is reduced compared to Sc because the intensity of red beam reaching the photodiode detection elements decreases due to absorption by blue precipitates. 4. The final signal at a given E. coli concentration is Sc–Ss. Unknown concentration of bacteria can be detected from a plot of (Sc–Ss) values as a function of known E. coli concentration. 5. The background noise level is the standard deviation of control signals obtained from control experiments. The noise level was determined to be 35 mV. Based on the background noise level of 35 mV, a detection limit of 4.5 × 104 E. coli. O157:H7 (at S/N = 3) can be obtained. The entire output signals from 144 photodiode detection elements are displayed as the digital value in the custom software.

4. Notes 1. On-chip detection of E. coli O157:H7 pathogenic bacteria using the PDA microchip may lead to erroneous result if the pHs of reaction solutions deviate from the neutral range (pH 7.0 to 7.5). The PDA detection elements may get damaged, e.g., corroded when the pH of the reaction solution is out of neutral range.

334

Song and Kwon

4.500 4.000 3.500

Signal (V)

3.000 2.500 2.000 1.500 1.000 0.500 0.000 After APTES

After glutaraldehyde After Goat antiE.coli O157:H7

After BSA

After E.coli O157:H7

After phosphataselabeled goat antiE.coli O157:H7

Fig. 6. Photosignals as measured at every reaction step during the on-chip bioassay. The pHs of solutions for BSA treatment and E. coli O157:H7 capturing were 8.0 and 8.2, respectively. As a result, large deviation of photosignals from the average value of 2.37 V was observed. When the pH of the reaction solution for on-chip bioassay was maintained at 7.0, photosignals were stabilized around the average value of 2.37 V, as measured after APTES, glutaraldehydes, anti- E. coli O157:H7 antibody, and alkaline phosphatase-labeled anti-E. coli O157:H7 antibody treatments.

2. To confirm normal operation of the PDA microchip during the on-chip reactions, the signal at every on-chip reaction step has to be monitored. As shown in Fig. 3, a large deviation of the signal (from 2.37 (average) ± 0.053 V) reveals abnormality of the photodiode detection element as a photosensor. 3. Figure 6 is a typical example that represents the variation of photosignals at every reaction step due to the variation of pH of reaction solutions. The pHs of BSA and E. coli O157:H7 solution were 8.0 and 8.2, respectively. As a result, large deviations of the signals could be observed. 4. A normal operation of the PDA microchip is assured as long as pH range of the on-chip reaction solution is maintained between 7.0 and 7.5.

Acknowledgments This research is sponsored by the Korean Ministry of Commerce, Industry and Energy under contract 10023590-2006-02 program with Seoul National University.

Photodiode Array On-chip Biosensor for the Detection of E. coli

335

References 1. Nakauchi, G., Inaki, Y., Kitaoka, S., Yokoyama, C., and Tanabe, T. (2002) Application of L-cystein derivative to DNA microarray. Nucleic Acids Res. Suppl. 2, 257–258 2. Allain, L. R., Askari, M., Stokes, D. L., and VoDinh, T. (2001) Microarray sampling-platform fabrication using bubble-jet technology for a biochip system. Freenius J. Anal. Chem. 371, 146–150 3. Song, J. M., Culha, M., Kasili, P. M., Griffin, G. D., and Vo-Dinh, T. (2005) Detection

of single bacteria using a compact CMOSbased immunosensor. Biosens. Bioelectron. 20, 2203–2209 4. Song, J. M., Yang, M., and Kwan, H. T. (2007) Development of a novel DNA chip based on a bipolar semiconductor microchip system. Biosens. Bioelectron. 22, 1447–1453 5. Kwon, H. T., Yang, M. S. (2003) Apparatus for analyzing for a disease using photodetector cell circuit. KR 0400202

Chapter 19 DNA Analysis with a Photo-Diode Array Sensor Hideki Kambara and Guohua Zhou Summary A simple instrument or device for easy DNA analysis is required. A combination of bioluminometric assay and photo-detection with an inexpensive photo-diode array provides a simple instrument for various DNA analyses. Its characteristics and applications for DNA analyses are described. Key words: DNA analysis, DNA sequencing, Pyrosequencing, SNPs, Photodiode, Photo-diode array, Luminescence, Step-by-step sequencing.

1. Introduction Gel electrophoresis coupled with laser induced fluorescence detection is generally used for analyzing DNA (1). A capillary array DNA sequencer has been successfully used to sequence the human genome (2, 3). With the completion of the human genome project in 2003, massive amounts of DNA data are available to be used in solving medical, environmental, and food problems. In the post genome-sequencing era, simple and inexpensive DNA analysis devices or instruments are required in addition to a high throughput instrument. Simple and inexpensive DNA analyzers can be made by using inexpensive photo-sensors. Here the characteristics of a small DNA analyzer with a photo-diode array and its application to DNA analysis are described. 1.1. Pyrosequencing

The small DNA analyzer is based on bioluminometric detection with a photo-diode and step-by-step nucleotide incorporation reactions (pyrosequencing) (4, 5). Figure 1 shows the principle of

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_19

337

Kambara and Zhou primer 5’ A C G ssDNA

3

+

T G

T A C G

C A T G

dATP

degradation

dCTP

degradation

dTTP

degradation

dGTP

PPi

+ + +

5’

A C G

3’

T G +

5’ 3’

ATP T A

PPi

luminescence

C G G

ATP

C C T T

A A

luminescence

A C G T A C G G A A T G C A T G C C T T A A 1 min

CCC AA TTGG TT G

C C T T A A

luciferase reaction

C A T G

dATP

fluorescence

338

GG CC AA TT A T G

AA C

3 bases 2 bases 1 base noise

time

Fig. 1. Principle of pyrosequencing technology and pyrogram: Four enzymatic reactions are carried out simultaneously. As the reaction solution is diluted by the dNTP additions, peak heights decrease gradually as indicated by the dashed lines. It is partly because the reaction speed slows down with the step while the peak areas stay almost constant.

pyrosequencing. The target DNA is amplified by PCR to prepare a single stranded template DNA. The reaction chamber contains reagents for ATP production as well as luminescence reaction. After the template DNA is purified, it is put into the reaction chamber together with the DNA polymerase and a primer. The primer hybridizes to the template DNA, and this is followed by nucleotide incorporation reactions obtained by sequentially adding four deoxynucleotide species (dNTPs) one-by-one. When an added nucleotide species is complementary to the target, nucleotide incorporation occurs to extend the complementary DNA strand and to produce an inorganic pyrophosphate (PPi) as a by-product. The pyrophosphate reacts with adenosine 5´ phosphosulfate (APS) to produce ATP by the enzymatic reaction of ATP sulfurylase. ATP reacts with luciferin to emit luminescence through a luciferase reaction. The luminescence is detected with an array sensor or a photo-diode. The base sequence is determined by observing the luminescence and the injected nucleotide species. The enzymatic reactions are summarized in Fig. 2. The original method of pyrosequencing uses APS as a substrate for the reaction to produce ATP. This is the substrate for luciferase reaction as well. Although the reaction speed of

DNA Analysis with a Photo-Diode Array Sensor

339

dNTP DNA polymerase

Primer

ssDNA

Extension

PPi AMP + oxyluciferin + CO2

APS

ATP sulfurylase

PPi

Luciferase

ATP

SO2 2-

hv Luciferin + O2

Fig. 2. Enzymatic reactions used in pyrosequencing. Nucleotide incorporation reaction, ATP production reaction, nucleotide degradation reaction and luminometric reaction occur simultaneously in one chamber.

luciferase-APS reaction is slow, the reaction gives a large background because the amount of APS is large compared to that of ATP. Another enzymatic reaction for producing ATP from PPi can be used. This reaction uses another enzyme, pyruvate orthophosphate dikinase (PPDK), to produce ATP with the substrates of AMP and phosphoenolpyruvate (PEP). DNA polymerase (ssDNA - primer)n + dNTP ¾¾¾¾¾¾ ®(ssDNA - primer)n + 1 + PPi 2+

PPDK, Mg AMP + PPi + PEP ¾¾¾¾¾ ® ATP + pyruvate + Pi 2+

luciferase, Mg ATP + luciferin + O2 ¾¾¾¾¾¾ ® AMP + Oxyluciferin + CO2 + PPi + light

As neither AMP nor PEP can be a substrate for the luciferase reaction, almost no background signal is produced even if a large amount of AMP or luciferase is used in the reaction (6). 1.2. Instrumentation for Pyrosequencing 1.2.1. Outline of the System

1.2.2. Device

A schematic diagram of a pyrosequencing system is shown in Fig. 3. It consists of eight sets of four reagent reservoirs and fluidics for injecting small amounts of reagents into eight reaction chambers, respectively. Each reaction chamber has a photo detector attached at the bottom. The system was made in house. Photographs of the prototype system are shown in Figs. 4 and 5. All movements are controlled with a microcomputer (Renesas Technology (Tokyo), HD64F3052BF25) according to the time chart as shown in Fig. 6. The required photo-detection sensitivity is estimated as follows. 1. Estimation of luminescence emitted from a reaction chamber: Usually 0.5–1 pmol of DNA sample is used for DNA sequencing.

340

Kambara and Zhou

dNTP reagent reservoirs dispensing nozzles

gas cylinder PC

valves loading

up and down valve control motor dispenser Position control

H8 microcomputer

reaction chamber vibration control

reaction chambers vibrator

reaction chamber holder

AD converter photo diode

Amp1

x1 x10 x100

Amp8

Amp2

multiplexer

Amp

Fig. 3. Schematic diagram of the instrument.

air pressure dispensers reaction chamber

computer small DNA analyzer

Fig. 4. Photograph of a small DNA analyzer.

DNA Analysis with a Photo-Diode Array Sensor

341

reagent reservoirs with dispensers

reaction chamber

Fig. 5. Photograph of the inside of the instrument.

reagent reservoir

air pressure

up

down dispensers reaction chamber reaction mixture injected dNTP

mixing by vibration

0.1s

0.2s

Dispenser position 2s Nucleotide pressure valve

0.2s 0.1s

Negative pressure valve 4.5s Reaction chamber vibration

115.5s

Fig. 6. Time chart for operating the system.

When one base extension reaction is carried out with 1 pmol of template DNA, one pmol of PPi is produced. If all of them are consumed for producing ATP and then luminescence, the total number of photons caused by the extension reaction is

342

Kambara and Zhou

about 4 × 1011 assuming that the quantum yield of luciferase reaction is 0.7. 2. Effect of ATP degradation by apyrase and ATP reproduction by a cycle reaction: The by-product of luciferase reaction, PPi, reacts with AMP and PEP again to produce ATP, and this causes a cycle reaction that can last until all AMP and PEP are consumed to produce ATP provided apyrase, which degrades ATP, does not exist in the reaction chamber. As ATP degradation and luciferase reactions are carried out simultaneously in one chamber, the average photon number caused by a nucleotide incorporation reaction is estimated to be as large as the number of incorporated nucleotides. 3. Estimation of photo-electron current detected with a photo-diode: Assuming that the luminescence lasts for 40 s, the photon flux emitted from the reaction chamber is estimated to be 1010/s. Assuming that 1% of luminescence is collected with the detector, the collected photo-flux is 108photons/s.(Generally, a large % of luminescence can be collected with a photo-detector.) The photoelectron current is determined by the characteristics of the detector device as shown in Fig. 7. The detection sensitivity of 0.3 A/W is obtained with a photodiode S1133 (from Hamamatsu Photonics, Shizuoka, Japan) for 560 nm light. As the energy of 560 nm light is 2.23 eV/photon, a photon flux of 108/s corresponds to an energy flux of 3.5 × 10−11 W, which gives a photo-current of 10−11A on a S1133 photo-diode. As

Si photodiode S1133

Package

ceramics

detection area

2.8X2.4 mm

detectable wavelength

320-730nm

most efficient wavelength (lp)

560nm

photosensitivity at lp

0.3A / W

dark current max.

0.01nA

Fig. 7. Photosensitivity curves of photo-diode S1133.

DNA Analysis with a Photo-Diode Array Sensor

343

the photo-current corresponds to a flux of 0.6 × 108 photoelectrons/s, the quantum yield is about 0.6. 4. Electric circuit for a photo-diode array: An example of the electric circuit used for S1133 is shown in Fig. 8. An output voltage is determined by the resistor Rf (1 GΩ) and the amplification factor of the buffer amplifier (×2, ×20, and ×200 scales can be selected). A signal of 2 V is obtained by receiving 10−11 A of photocurrent (×200 scale). This is easily detected with a photo-diode S1133. 5. Detection limit of the system: The detection limit is determined by the fluctuation of dark current of a photo-diode. A photodiode S1133 can detect 10−14A of photo-current at S/N = 3, which corresponds to 2 mV signals (×200 scale). It is possible to detect DNAs at a f-mole level with the photo-diode. The photo-diode is covered with a conductive film coated with ITO (Indium Tin Oxide) to reduce the electric noise. R1 C

R2

Rf

R3 − 5VA

− 5VA

C0

C0

Rs

_ PD1

+

_

A/D +

opA129UB

op07CS C0

C0 +5VA

+5VA

+5VA C0 C0

C − 5VA

Rf

1

8

Multiplexer Max4051ACSE

− 5VA C0

PD: photodiode S1130-1 opA129UB : Operational Amplifier op07CS : Operational Amplifier Rf: 1GΩ R1: 180kΩ R2: 18 kΩ R3: 1.8kΩ Rs: 1 kΩ C : 100pF C0: 100,000pF

_ PD8

+

opA129UB C0

+5VA

Multiplexer (MAX40551ACSE, Maxim lntegrated Products, Sunnyvale, CA) Operational Amplifier Operational Amplifier AD converter ( ADS1271PW, Texas Instruments, Dallas, TX)

Fig. 8. Electric circuit used for luminescence detection with S1133 photodiode.

344

Kambara and Zhou

6. Signal to noise ratio: In the typical pyrosequencing condition, a half picomole of template DNA is used. It gives a signal of about 0.5 V. The background luminescence due to a side reaction of the luciferase with the APS in the conventional condition is in a range of 0.5–1 V. The background becomes negligible in the new enzymatic reaction using PPDK and AMP. In the latter case, the background signal as well as the noise level is determined by the luminescence and its fluctuation due to the luciferase reaction of dNTP as well as residual ATP or PPi in the reagents. Usually it is in a range of 5–10 mV. Therefore the signal-to-noise ratio for sequencing 0.5 pmol of DNA in the latter condition is about 50 or more. 1.2.3. Reaction Chamber

A part of an immuno-plate (eight micro-wells of 350 µL) from NUNC (Nunc-Immuno™ BreakApart™ Modules; CAT No473539) is used as reaction chamber. A micro-well plate is fixed on a holder. The holder is attached with a small vibrator to mix reagents. The bottoms of the micro-wells are flat and placed 0.5 mm above the photodiodes. Usually the reaction chambers are filled with 30 µL or more of reaction solution for luminometric assays. If a reduced amount of reaction solution is required, handmade small chambers (as shown in Fig. 9) can be used.

1.2.4. Dispenser

A plastic dispenser chip that is made up of four reagent reservoirs with four dispensing nozzles (capillary tubes) attached to the reservoirs (Fig. 10) is used. The capacity of each reservoir is 40 µL. Usually 20 µL of reagent is supplied to each reservoir, which is enough for 50 such injections. The capillary tubes are 20 mm long and their inner diameter is 50 mm. The reservoirs contain four different dNTPs dissolved in buffer for pyrosequencing. Four different ddNTPs in buffer are used instead of dNTPs for SNPs

Reaction chambers (NUNC)

Reaction chambers in a holder

Small reaction chambers (homemade type)

Fig. 9. Photograph of reaction chambers. A part of micro-titer plate from NUNC is used routinely.

DNA Analysis with a Photo-Diode Array Sensor

345

capillary nozzle

reagent reservoir

Fig. 10. Photograph of a dispenser with four reservoirs each containing one of four nucleotide species. Each nucleotide is injected into a reaction chamber from the capillary by applying air pressure on the reservoir.

detection. Dispenser chips are filled with dNTPs or ddNTPs just before the experiment to avoid the decomposition of the reagents.

2. Material 2.1. Material for PCR Amplification

1. High Fidelity PCR Buffer Invitorogen, (Carlsbad, CA), Cat. No. 11304-011. 2. dNTP mixture (10 mM), Invitorogen, (Carlsbad, CA), Cat. No. 18427-013. 3. MgSO4 (50 mM), Invitorogen, (Carlsbad, CA), Cat. No. 11304-011. 4. Autoclaved, distilled water, Invitorogen, (Carlsbad, CA), Cat. No. 10977-023. 5. PCR Primers, Sigma Genosys (Hokkaido, Japan). 6. Platinum Taq High Fidelity, AB: Applied Biosystems (Foster City, CA) Cat. No.11304-01.

2.2. Reagents for Pyrosequencing

1. PPDK-E, Kikkoman (Chiba, Japan). 2. Luciferase(LUC-H), Kikkoman (Chiba, Japan). 3. Apyrase, grade VI, Sigma (ST. Louis, MO). 4. Luciferin, Kikkoman (Chiba, Japan).

346

Kambara and Zhou

5. PEP, Kikkoman (Chiba, Japan). 6. BSA, Sigma (St Louis, MO). 7. DTT, Sigma (St Louis, MO). 8. Nucleotides Sp-dATP-α-S Biolog Life Science Institute (Bremen, Germany). 9. dTTP, dCTP, and dGTP, GE Healthcare (Piscataway, NJ). 10. Tris (hydroxymethyl) aminomethane, GIBCO Industries (Langley, OK). 11. EDTA and Mg(Ac)2, Sigma (St. Louis, MO). 2.3. Preparation of PCR Reaction Reagent

The components of the PCR reaction mixture are listed in Table 1. 1. The primer sequences used to amplify the 181-bp fragments of TPMT gene from human genomic DNA (NCBI accession No AB045146) are as follows. Forward primer: 5´-TGTTGAAGTACCAGCATGCAC-3´ Reverse primers: 5´-biotin-AAATTACTTACCATTTGCGATCA-3´.

Table 1 PCR reaction mixture Volume (mL)

Final concentration

Supplier

Cat. No.

Store at

10× High Fidelity PCR Buffer

5

1×

I

11304–011

−20°C

dNTP mixture (10 mM)

1

0.2 mM each

I

18427–013

−20°C

MgSO4 (50 mM)

2

2 mM

I

11304–011

−20°C

Primer forward, 5´-biotin* (original concentration: 10 µM)

1

0.2 µM

SG

–

−20°C

Primer reverse (original concentration:10 µM)

1

0.2 µM

SG

–

−20°C

Template DNA 50 ng/µL (extracted from genomic DNA)

1

1 ng/µL

–

–

4°C

Platinum Taq high fidelity

0.2

1.0 U

AB

11304–011

−20°C

Autoclaved, distilled water

39

I

10977–023

Room temperature

Components

I Invitorogen(Carlsbad, CA); SG Sigma Genosys (Hokkaido, Japan); AB Applied Biosystems (Foster City, CA)

DNA Analysis with a Photo-Diode Array Sensor

347

2. Buffer solution for primers: 10 mM Tris–HCl (pH 7.5). 3. Buffer solution for template DNA: TE buffer (10 mM Tris– HCl (pH 7.5), 1 mM EDTA). 2.4. Reagents Used for Preparing SingleStranded DNA Template

1. Beads for magnetic separation: Dynabeads® M-280 Streptavidin (2.8 µm ID) (Dynal A.C., Oslo, Norway).

2.5. Reagents and Buffer Preparation for Pyrosequencing

The reagents used in pyrosequencing are listed in Table 2 together with the suppliers. 1. Preparation of buffer for luminometric assay: A half liter of 2 × Buffer solution (pH 7.75) is prepared by mixing 200 mM Tris-HCl, 4 mM EDTA and 40 mM Mg(Ac)2 in ultra pure water. The solution is stored at 4°C overnight followed by filtering with a filter (450 Filter Units-500 mL Capacity, MF75-series; Nalgene labware, Cat No 450-0020). The filtered solution is fractioned in falcon tubes (15 mL each) to be stored at 4°C. This is to reduce the risk of contamination.

2. Binding & Wash buffer: 10 mM Tris–HCl (pH 7.5), 1 mM EDTA, 2 M NaCl.

2. Preparation of reaction mixture for luminometric assay with PPDK-AMP system: The reaction mixture for luminometric assay and pyrosequencing consists of the reagents listed in Table 2. The reaction mixture is prepared in a 2 mL tube at first. Then

Table 2 Luminometric assay mixture (2 mL) Components

Volume (µL)

Final concentration

Original concentration

Supplier (state)

2 × Buffer

1,000

–

–

–

Ultra pure water

385.6

–

–

G

PPDK-E

62.8

33.8 U/mL

1076 U/mL

K (liquid)

Luciferase (LUC-H)

245.6

523.0 GLU/mL

4258.1 GLU/ mL

K (liquid)

Apyrase, grade VI

36

1.8 U/mL

100 U/mLl

S (solid)

Luciferin

200

0.4 mM

4 mM

K (liquid)

PEP

2

0.04 mM

40 mM

K (liquid)

AMP

8

0.2 mM

50 mM

S (solid)

BSA

20

0.10%

10%

S (liquid)

DTT

40

2 mM

100 mM

S (solid)

K Kikkoman (Chiba, Japan); S Sigma (St Louis, MO); G GIBCO Industries(Langley, OK)

348

Kambara and Zhou

it is subdivided into four 500-µL tubes to store at −20°C. For preventing the decomposition of the reagents, repetitive freezing and thawing should be avoided. Therefore, the reaction mixture should be divided into small bottles. The frozen reagent should be kept at room temperature for 30 min before use. 3. Preparation of nucleotide reagents: The concentrations of dNTPs are listed in Table 3. Each dNTP is dissolved in the buffer (pH 7.75) containing the components in Table 4. Each dNTP is stored at −20°C. As dNTP solution frequently contains PPi which produces a background signal, inorganic pyrophosphatase (PPase; USB Corp. (Cleaveland, OH) ) is added to the dNTP solution, which is incubated at room temperature for 15 min. The concentration of PPase in the dNTP solution is 20 mU/µL and 3.2 U of PPase should be added to 160 µL of each dNTP solution. The enzymatic reaction is stopped by placing the dNTP tube on ice. It is recommended to use the dNTP solutions within 24 h. 4. Primer: The primer (sequence: 5´-TGTTGAAGTACCAGCATGCAC-3´) of concentration 10 µM is produced by Sigma Genosys (Hokkaido, Japan). It is stored at −20°C before use. 5. Template DNA: The fragment of TPMT gene (181 bp) from human genomic DNA is used. 6. DNA polymerase: The original concentration of DNA polymerase I, Klenow Fragment, Exo-cloned from Ambion (Austin, TX) is 5 U/µL. It is stored at −20°C before use. In the case of 8-chamber pyrosequencing, the total amount of reaction mixture for analyzing eight samples in parallel is 240 µL (30 µL × 8). Add 1.5 µL of DNA polymerase to 240 µL of luminometric assay reaction mixture. 7. Preparation of terminator (ddNTPs) reagents: Four ddNTPs are from GE Healthcare. Each of them is dissolved in the buffer listed in Table 4 with the concentration of 125 µM.

Table 3 Preparation of dNTPs Nucleotides

Supplier (state)

Final concentration (µM)

Sp-dATP-α-S

BLSI (liquid)

250

dCTP

GEH (solid)

125

dGTP

GEH (solid)

125

dTTP

GEH (solid)

125

BLSI Biolog Life Science Institute (Bremen, Germany); GEH GE Healthcare (Piscataway, NJ)

DNA Analysis with a Photo-Diode Array Sensor

349

Table 4 Buffer composition Components

Supplier (state)

Concentration (mM)

Tris (hydroxymethyl) aminomethane

G (solid)

100

EDTA

W (solid)

0.5

Mg(Ac)2

S (solid)

5

DTT

S (solid)

1

G GIBCO Industries (Langley, OK); W WAKO Pure Chemical Industries (Osaka, Japan); S SIGMA (St.Louis, MO)

3. Methods 3.1. PCR Amplification and Amplicon Purification

The principle of PCR is shown in Fig. 11. The reverse primer is labeled with biotin at 5′-end for separating single-stranded template DNA. 1. Take 39 µL of autoclaved distilled water into a PCR tube (0.2 µL, Applied Biosystems Cat. No 8010540). 2. Add 5 µL of 10× High Fidelity PCR Buffer to the solution. 3. Add 2 µL of MgSO4 (50 mM) to the solution. 4. Add 1 µL of dNTP mixture (10 mM) to the solution. 5. Add 1 µL of forward primer and 1 µL of reverse primer (5′-biotin) to the solution. 6. Add 1 µL of template DNA (50 ng/µL: extracted from genome DNA) into the solution. 7. Add 0.2 µL of Platinum Taq High Fidelity 8. Mix the reagent gently by pipetting the solution several times. 9. Set the tube on a Gene Amp PCR System 9700 (Applied Biosystems, Foster City, Ca). 10. Denature the reagent at 94°C for 2 min and follow that with 40 cycle reactions (94°C for 15 s; 55°C for 30 s; 70°C for 60 s). 11. Remove the tube and place it on ice. 12. Take 1 µL of the product for confirming the length of the product by the 2100 Bioanalyzer (Agilent Technologies, Santa Clara, Ca) (This step can be skipped.).

350

Kambara and Zhou 3’ template DNA 5’

5’ 3’

denaturing and primer annealing 5’ reverse primer 3’

3’ forward primer 5’

1st cycle

extension 5’ 3’

3’ 5’

denaturing and primer annealing 5’

3’

3’

5’

2nd cycle

extension 5’

3’

3’

5’

~ 40cycles 5’

3’

amplified DNA fragments

3’

5’

Fig. 11. Outline of the PCR amplification process.

13. Purify PCR products with QIAquick PCR Purification Kit (Qiagen, Cat. No. 28104) according to the instructions attached to the product. Next, dissolve the purified product in 25 µL of TE buffer. 3.2. The Preparation of Single-Stranded Template DNA

Figure 12 is a schematic of the single-stranded DNA production process. 1. Take 25 µL of Dynabeads (from Dynabeads® M-280, Cat 112-062, 10 mL) solution into a 1.5 mL tube. 2. Separate the Dynabeads from the original buffer with a magnet (Dynal, Cat No 120.20). 3. Add 50 µL of binding & wash buffer (10 mM Tris–HCl (pH7.5), 1 mM EDTA, 2 M NaCl into the solution. Wash the beads dissolved in the solution by vibrating the tube with a Voltex mixer. After washing the beads, remove the buffer solution. Repeat the washing process three times.

DNA Analysis with a Photo-Diode Array Sensor

double stranded DNA

biotin

351

bead streptavidin

+

Dynabead M-280 PCR solution

immobilization DNA on the bead

alkaline denaturation 0.1N NaOH

separation bead with a magnet

magnet

neutralization

washing

binding / wash buffer

single stranded DNA

DNA samples immobilized on bead

DNA samples in a solution

Fig. 12. Outline of the DNA purification process.

4. Add 25 µL of PCR product together with 25 µL of binding & wash buffer to the tube and mix with the beads by stirring at room temperature for 30 min (PCR product, double-stranded DNA, is immobilized onto the bead surfaces.). 5. Separate the beads from the solution with a magnet. Add 50 µL of NaOH (0.1 M) to the beads and keep the mixture at room temperature for 5 min. The complementary DNA strands are separated in the solution and the target DNA stays on the beads. 6. Separate the beads from the solution with a magnet and the target DNA is on the bead surface while the complementary DNA strands are in the solution. 7. Neutralize the solution containing the complementary DNA strands to pH7.5–8.0 by adding 50 µL of 0.1 M HCl to obtain free template DNA in 100 µL of solution.

352

Kambara and Zhou

8. Wash the beads in the tube using the same procedure described in step 3. Then wash them with 125 µL of ultra pure water three times. 9. After removing the water, add 50 µL of binding & wash buffer and vibrate the tube with a Vortex mixer. 10. Now you have single-stranded DNA samples immobilized on bead surfaces as well as in a solution. 3.3. DNA Sequencing by Pyrosequencing

Figures 4 and 5 are photographs of a small DNA analyzer based on luminometric assay. The measurement is carried out according to the following process. 1. Open the cover of the instrument. 2. Set eight reagent reservoirs in the upper holders. 3. Put reaction chambers in the reaction chamber holders. 4. Close the cover of the instrument. 5. Turn on the switch of the instrument and the computer. 6. Various operations are carried out according to the time chart as shown in Fig. 6. (a) Move down the reagent reservoirs with four capillary nozzles to dip into the reaction solutions. (b) Inject one nucleotide species at a time from a nozzle by applying a small pressure onto one of the four reservoirs for 0.2 s. The injection of four nucleotide species is carried out by turns. The amount of reagent dispensed per shot is 0.4 µL. Then the reservoirs with capillary nozzles are lifted up to be placed in the air. As the four capillary nozzles are dipped in the reaction solution together, small amounts of reagents besides the injected reagent species are frequently leaked out from the nozzles. To prevent the leakage of reagents, a small negative pressure is applied to the reservoirs to make air gaps in the capillary nozzles which are useful to prevent the undesirable leakage of reagent during the injection. (c) The reaction chamber holder is vibrated with a small motor (Tsukasa Electric Co. (Tokyo), TG-87A-GU) during the measurement after dispensing dNTP into the solution. (d) The detection of photo signals starts with the injection of dNTP and lasts for 120 s. The detection is carried out every 0.2 s. Then the measurement cycle is repeated 100 times for sequencing about 50 bases. 7. All the data obtained in a cycle are transferred to a computer at a time.

DNA Analysis with a Photo-Diode Array Sensor

353

8. A set of data per cycle contains 600 data points which make one peak in a pyrogram. The 600 data points are summed up to give the signal strength per nucleotide incorporation reaction which is used for sequence analysis. The time chart of the operation is shown in Fig. 6. The interval of every injection is 120 s, which is selected so that the degradation of dNTP by Apyrase is completed before the next dNTP injection. The main steps needed for measurements: 1. Turn on the main switch and confirm that the background signal is less than 0.005 V. The background signal may be caused by stray lights or an induction current that can be removed by shielding the photo-diodes with a conductive plastic film. 2. Take 30 µL of the pyrosequencing mixture into a reaction chamber. 3. Put the reaction chamber onto the instrument to check the background signal again. If the background signal is less than 0.01 V, you can proceed. If you have a large background signal, your sample may be contaminated with PPi or ATP. If so, wait for a while until apyrase degrades most of the ATP. It is also recommended to add a small amount of PPase into the reaction mixture prior to the addition of APS for degrading any residual PPi (When the ATP sulfurylase-APS system is used to produce ATP from PPi, the background due to APS becomes around 0.5 V due to the APS. (see Notes 1-3). 4. After confirming that the background signals are negligibly small, add 1 µL of DNA polymerase to the mixture in a reaction chamber. It is recommended to add a ddNTP mixture containing four ddNTPs of 3 pmol each to terminate the strand extension of contaminated DNA in the mixture that may be introduced in the preparation process. Keep the mixture at 30°C for 2 min so that the added ddNTPs are degraded by apyrase. 5. Take 1 µL of primer and 5 µL of template DNA into one tube (0.5 mL). Put the tube in a water bath to increase the temperature to 96°C for 5 s and cool it down to room temperature for primer hybridization to the template DNA. 6. Take 1 µL of the template DNA hybridized with the primer into the pyrosequencing mixture in a reaction chamber. (Now all the extendable DNA termini except for the primers are terminated and signals due to the primer extension are obtained.) 7. Put dispensers, filled with 20 µL of dNTPs in each of the four reservoirs, on the instrument. The dispenser units move down to dispense reagents. The capillary nozzle ends of the dispensers are placed 0.5 mm below the reaction mixture surfaces in reaction chambers while dNTPs are being dispensed (see Notes 1-3).

354

Kambara and Zhou

8. Start the sequencing measurement. At first, the dispenser unit moves down to dip the capillary nozzles into the reaction solution and then air pressure is applied to the dispensers for 0.2 s to inject 0.4 µL of dNTP into the reaction chambers. The reaction chambers are vibrated with a vibrating motor attached to the chamber holder for 30 s after the injection. Luminescence measurement starts from 5 s after the injection. 9. The order of dNTP injection is dATPαS, dTTP, dGTP, and dCTP. Observed signals together with the injected nucleotide species from a program are shown in Fig. 13. The DNA sequence can be determined by observing the signal for each nucleotide injection. Although no signal appears for the injections of dATP, dTTP, and dGTP, a signal obtained by injecting dCTP can be observed in the figure. This indicates the addition of base “C” to the complementary DNA strand. Here, for convenience, we call the signal intensity corresponding to one base extension a signal-unit here for the convenience. As two “C”s were incorporated, the signal intensity was two signal-units in Fig. 13. By injecting dTTP, a signal as large as one signal-unit appeared, and this indicated that the extended DNA strand had the sequence of 5´-primer-CCT-3´. Then, we observed a photo-signal as large as two signal-units that appeared after dGTP injection indicating that two G were added to the complementary DNA strand. The sequence was determined to be 5´-primer-CCTGGATT-3´. The next signal for G was a half signal-unit and the following “A” signal was about one and a half signal-units indicating the sample included a hetero-sequence of G/A at the indicated region. The sequence was determined as 5´-primer-CCTGGATT(G/A)ATGGCAACT-3´. As described here, it becomes frequently difficult to accurately determine the sequence of genomic samples when SNPs especially heterozygous sequences appear because it requires accurate and quantitative signal intensities for sequencing analysis. 3.4. SNPs Detection by One Base Extension with Ddntps

The procedure for SNP typing with ddNTPs is basically the same as that for pyrosequencing except that four terminators (ddNTPs) are used instead of dNTPs for nucleotide incorporation reactions (7). Each terminator is injected once. Obtaining pyrograms for hetero-samples is complex as has been demonstrated previously, however, the new method gives a simple spectra even when a sample has a plural of SNP sites. A schematic of the method is shown in Fig. 12a. A primer is constructed to hybridize to the target DNA as the nucleotide incorporation occurs at the SNPs site. Nucleotide incorporation reactions are conducted by injecting four ddNTPs one at a time. Only one base extension occurs when the injected ddNTP

Signal intensity (V)

0

CC

T

GG

500

A

TT

G

0.5

A

1.5

T C

1000

GG AA C T

TGC

1500

AA T

2000

T

C

Time (sec)

CC T

2500

A T

3000

T

CCC AAA

C A

3500

T G

T

C

4000

AAA

Fig. 13. Pyrogram obtained with a fragment of human TPMT gene. The dNTP dispensing order is A → T → G → C → A. The base species on the peaks indicate the sequence of DNA complementary to tvhe template.

0.0

0.4

0.8

1.2

DNA Analysis with a Photo-Diode Array Sensor 355

356

Kambara and Zhou

is complementary to the target DNA, which makes the spectrum much simpler than that for pyrosequencing. Obtaining a spectrum from one-base-extension reaction with ddNTPs is simpler than that with dNTPs as is shown in Fig. 14. As the use of ddNTPs makes spectral patterns very simple, plural SNPs can be determined simultaneously. Figure 15 shows the nine possible spectral patterns for dual SNPs. The dual SNPs can be determined by observing the spectral pattern. Even triple SNPs can be determined simultaneously with the method. The new method uses ddATP although dATPαS is usually used in pyrosequencing to reduce background luminescence. Although dATP can be a substrate for a luciferase reaction, ddATP is not a substrate for a luciferase reaction, but is a good substrate for nucleotide incorporation reaction. This is an advantage of the new SNP typing method.

AAATC

+

GAATC

ddCTP AAATC

+

AAATC

+

dCTP C

AAATC

+

C GAATC

T

TTT AAATC

C GAATC

ddATP

No reaction

T

+

CTT GAATC

C

dTTP decomposition C

T

dATPαS

T A

No reaction

ddGTP

C

dTTP C

+

C GAATC

dCTP decomposition

ddTTP T AAATC

GAATC

TTTA AAATC

+

CTTA GAATC

A C

dATP αS decomposition T C

dGTP

T A G

TTTAG AAATC

+

A

CTTAG GAATC

G

C

dGTP decomposition

(a)

(b)

Fig. 14. Principle of SNP detection by ddNTP addition (a) and pyrosequencing (b): Four ddNTPs or dNTPs are injected into a reaction chamber in turn. Signals appear only when the nucleotide incorporations occur at the SNP site in case (a), which makes the spectrum simple. However, a pyrogram for a SNP sample becomes complex because nucleotide incorporation reactions can occur at a position other than the SNP site.

DNA Analysis with a Photo-Diode Array Sensor 4

4

(AA/TT)

3

4

(AC/TT)

3 2

2

1

1

1

0 A

T

G

C

4

0 A

T

G

C

4

(AA/TC)

3

A

(AC/TC)

3

2

1

1

1

T

G

C

4

0

A

T

G

C

4

(AA/CC)

3

0

(AC/CC)

2

2

1

1

1

0 T

G

C

A

T

G

C

(CC/CC)

3

2

A

C

4

3

0

G

(CC/TC)

3

2

A

T

4

2

0

(CC/TT)

3

2

0

357

0 A

T

G

C

A

T

G

C

Fig. 15. Spectral patterns for two SNPs obtained with ddNTP addition. As the patterns are limited to nine for two SNP sites, two SNPs are easily determined by observing the pattern.

4. Notes 1. ATP production system using ATP sulfurylase and APS: The pyrosequencing reaction mixture used for ATP sulfurylaseAPS system is shown in Table 5. As APS is a substrate of luciferase reaction, the background signal due to APS is as large as 0.5 V. The background signal decreases with time because APS is consumed for the reactions. 2. The four nozzles of a dispensing unit are dipped into the reaction solution at a time. The pressure is applied on one of the nozzles (working nozzle) to inject the corresponding dNTP. The other three nozzles (resting nozzles) are also dipped in the solution, however, the leakage from the other dNTPs has to be avoided. To prevent dNTP leakage from non-working nozzles, an air gap is formed in the nozzles by applying a small negative pressure on all the nozzles when they are at the waiting position in the air. 3. Phase shift in pyrosequencing.

358

Kambara and Zhou

Table 5 Luminometric assay mixture based on APS and ATP sulfurylase (2 mL) Components

Volume (µL)

Final concentration

Original concentration

Supplier (state)

2 × Buffer

1,000

–

–

–

Ultra pure water

338

–

–

G

ATP sulfurylase

20

200 mU/mL

0.02 U/µL

S (solid)

Luciferase (LUC-H)

246

523 GLU/mL

4,258 GLU/mL

K (liquid)

Apyrase, Grade VI

36

1.8 U/mL

100 U/mL

S (solid)

Luciferin

200

0.4 mM

4 mM

K (liquid)

Adenosine 5′ phosphosulfate (APS)

100

5 µM

0.1 mM

S (solid)

Bovine serum albumin (BSA)

20

0.10%

10%

S (liquid)

DTT

40

2 mM

100 mM

S (solid)

K Kikkoman (Chiba, Japan); S Sigma (St Louis, MO); G GIBCO Industries (Langley, OK)

In pyrosequencing, the uniformity of signals is important for accurate base reading. The signal intensity is generally proportional to the number of extended DNA strands and influenced by the reaction yield for nucleotide incorporation reactions. Sometimes the nucleotide incorporation reactions do not occur uniformly in a reaction chamber which causes positive and negative phase shifts in the reaction (6, 8). A positive phase shift occurs when the degradation of injected nucleotide species has not been completed before the injection of the next nucleotide species and two nucleotide species are incorporated at the same time for some DNA strands. This produces DNA strands in different lengths and in advanced reaction steps. The minus phase shift occurs when the incorporation reactions for all DNA strands are not completed because the degradation of nucleotides is too fast or the mixing of the injected nucleotides in a chamber is insufficient. Then the not-extended DNA strands coexist together with the extended DNA strands in a reaction chamber. Examples of the phase shifts are shown in Fig. 16. The accumulation of both phase shifts prevents an accurate base calling because signals are always observed for any nucleotide species injection. DNA of up to 40 bases can be rather easily sequenced and the readable length can be extended to more than 70 bases by optimizing the conditions.

DNA Analysis with a Photo-Diode Array Sensor dNTP injection order : A

Sequence of complementary DNA strand : AGTTGTCCTA

A

A

T

A

G

AG

C

AG

A

AG A

G C

G

C A

T

100

0.0

G

C

400

A

0

T

G C

100

A

T

G

200

C

A

3.3mU/µl (b)

300

T

G

C

400

500

T

G

G

A

2.2mU/µl (c)

c

0.0

A

0

T

G

C A

100

T

G

200

1.2

C

A

300

T

G

C

400

500

TT

0.8

A

0.4

G

A

T G

1.1mU/µl (d)

T

G

0

C

A

200

T

G

C

400

1.6

600

CC

c

0.0

600

CC

TT

0.3

500

T

G

0.6

AGTTG AGTTGT AGTTG AGTTGTCC

T

300

CC

G

A

0.2

AGTT

G C A

200 TT

0.4

relative intensity

T

A T

0

0.6

C A

6.7mU/µl (a)

g

A

0.0

insufficient dNTP degradation

T

G

G

A

0.1

G

CC

TT

0.2

imperfect strand extension

T

359

A

T

600

G C

800

1000

TT A

0.8 0.0

A

0

G

G T

G

200

C

A

400

T

G

C C

600

CC

T A

800

T

G

0.6mU/µl (e)

C

1000 1200

time, s

Fig. 16. Pyrograms showing plus and minus phase shift. (a) apyrase:6.7 mU/µL, (b) apyrase:3.3 mU/µL, (c) apyrase:2.2 mU/µL, (d) apyrase:1.1 mU/µL, (e) apyrase:0.6 mU/µL. The spectral patterns change with the amount of apyrase in a reaction chamber. When the amount of injected dNTP is not sufficient or the apyrase amount is large, dNTP is degraded before the DNA strand extension is completed. The extended DNA strands together with non extended DNA strands are produced in the chamber. The inhomogeneous reaction produces minus phase shift in a spectrum as shown in (a). When the amount of apyrase is very small, insufficient dNTP degradation occurs where injected dNTP stays for a long period in the chamber. They are incorporated together with the newly injected dNTP species and positive phase shift is thus produced as shown in (e).

Acknowledgments The author would like to thank Dr Tomoharu Kajiyama and Dr Akihiko Kishimoto for the development of technologies and instruments, and Ms Sumiyo Takiguchi and Ms Mari Goto for their help in experiments.

References 1. Kheterpal, I. and Mathies, R.A. (1999) Capillary array electrophoresis DNA sequencing. Anal Chem, 71, 31A–37A 2. Zubritsky, E. (2002) How analytical chemists saved the human genome project. Anal Chem, 74, 23A–26A 3. Kambara, H. and Takahashi, S. (1993) Multiple-sheathflow capillary array DNA analyser. Nature, 361, 565–566

4. Ronaghi, M., Uhlen, M. and Nyren, P. (1998) A sequencing method based on real-time pyrophosphate. Science, 281, 363–365 5. Zhou, G.-H., Kamahori, M., Okano, K., Harada, K. and Kambara, H. (2001) Miniaturized pyrosequencer for DNA analysis with capillaries to deliver deoxynucleotides. Electrophoresis, 22, 3497–3504 6. Zhou, G.-H., Kajiyama, T., Gotou, M., Kishimoto, A., Suzuki, S. and Kambara, H. (2006)

360

Kambara and Zhou

Enzyme system for improving the detection limit in pyrosequencing. Anal Chem, 78, 4482–4489 7. Zhou, G.-H., Gotou, M., Kajiyama, T. and Kambara, H. (2005) Multiplex SNP typing by bioluminometric assay coupled with terminator incorporation (BATI). Nucleic Acids Res, 33, e133

8. Gharizadeh, B., Nordstrom, T., Ahmadian, A., Ronaghi, M. and Nyren, P. (2002) Long-read pyrosequencing using pure 2´-deoxyadenosine-5´-O’- (1-thiotriphosphate) Sp-isomer. Anal Biochem, 301, 82–90

Chapter 20 Miniaturized and Integrated Fluorescence Detectors for Microfluidic Capillary Electrophoresis Devices Toshihiro Kamei Summary Microfluidic devices are revolutionary in their ability to use very small quantities of liquid samples and to perform biochemical analyses with unprecedented speed. Toward the goal of a lab-on-a-chip that integrates a series of analysis steps and analytical components into a single microfluidic device, one of the most critical aspects of size reduction is miniaturizing and integrating the fluorescence detection system. We present here details of a new integrated fluorescence detection system. A microfluidic biochemical analysis device is mounted on a compact detection platform that comprises a fluorescence-collecting microlens and micromachined fluorescence detector in which a multilayer optical interference filter is monolithically integrated and patterned on a hydrogenated amorphous silicon (a-Si:H) photodiode. A central aperture in the micromachined a-Si:H fluorescence detector allows semiconductor laser light to pass up through the detector and to irradiate a microchannel of the microfluidic analysis device. Such an optical configuration enables a detachable, reusable, compact module to be constructed for the excitation source and detector. The micromachined a-Si:H fluorescence detector exhibits high sensitivity for practical fluorescent labeling dye, making it ideal for application to portable point-of-care microfluidic biochemical analysis devices. Key words: Microfluidic, Lab-on-a-chip, Hydrogenated amorphous silicon, Photodiode, Fluorescence, Electrophoresis.

1. Introduction The manipulation of a minute quantity of fluid (pl–nl) in a microchannel, termed microfluidics, has emerged in the last decade as an interdisciplinary field between molecular biology and electronics. Similar to the scaling law for a metal-oxide semiconductor field-effect transistor (MOSFET) in an integrated circuit that, as transistors gets smaller, they can switch faster Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_20

361

362

Kamei

and use less power, the polymerase chain reaction (PCR) in a nanolitter reactor has been dramatically speeded up from the conventional microlitter scale. In addition, an unprecedentedly small sample plug of approximately 100 µm that can be electrokinetically formed in a microchannel network has achieved capillary electrophoresis (CE) with high-speed and high-separation efficiency (1). Microfluidic plumbing technology based on a silicone elastomeric membrane has been used to perform cell sorting and combinatorial screening of protein crystallization conditions (2–4). Despite this rapid progress in microfluidic biochemical assays, a high-sensitivity microfluidic analysis still typically requires a bulky fluorescence detection system comprising a laser, optical device, and detector. In order to realize the potential portability of a microfluidic analysis system, the fluorescence detection system must also be miniaturized and integrated. Hydrogenated amorphous silicon (a-Si:H) is an ideal choice of material for integrating a fluorescence detector for a variety of reasons (a) it exhibits high sensitivity at the emission wavelength of most practical labeling dyes such as green fluorescence protein, DNA intercalators, ethidium bromide, and fluorescein; (b) it exhibits a low-dark current suited for low-noise measurement; (c) it can be monolithically integrated on a laser diode due to its disordered structure and low-temperature fabrication process; and (d) its manufacture is inexpensive (5). Although an avalanche Si photodiode (APD) shows much higher sensitivity, the limit of detection (LOD) of an integrated fluorescence detector is determined by the efficacy of the integrated optics rather than by the detector sensitivity itself; there has been no need to use a higher sensitivity detector than a photodiode at this point. Our systematic study has shown that an a-Si:H PIN photodiode exhibited LOD of <1 nM in terms of a fluorescein solution under confocal optics where no background photocurrent due to laser light scattering was observed (5). On the other hand, the fairly high background photocurrent present in the integrated a-Si:H fluorescence detector leads to LOD of 7 nM for fluorescein concentration (6). Nevertheless, the integrated a-Si:H fluorescence detector is capable of performing various assays including DNA fragment sizing, detection and identification of pathogens, and enantiomer detection of amino acids (5–7). We present here details of the fabrication process for a micromachined a-Si:H fluorescence detector in which a SiO2/ Ta2O5 multilayer optical interference filter is monolithically integrated on an a-Si:H photodiode with an integrated detection system comprising a fluorescence-collecting microlens and micromachined fluorescence detector that acts as a platform for a microfluidic electrophoresis device.

Miniaturized and Integrated Fluorescence Detectors

363

2. Materials 2.1. A-Si:H Photodiode Fabrication

1. A glass wafer with low concentrations of alkaline metals such as K and N (e.g., Corning 1737 glass). 2. A solution (Semico Clean 56, Furuuchi Chemical Co., Japan) for cleaning glass wafers. 3. A high-purity Cr target (>99.9%) for bottom electrode sputtering. 4. A solution (stripper 106, TOK, Japan) for stripping photoresist. 5. A solution containing 6–16% of cerium ammonium nitrate and 4–11% of nitric acid for Cr etching. 6. A spin coater (e.g., Mikasa 1H-DX2) to spin photoresist on a wafer. We typically use OFPR-8600LB 33cp photoresist (TOK, Japan), except for the Al lift-off process for which SIPR-96843.0 photoresist (Shinetsu Chemical Co., Japan) is used. We use a stepper (e.g., Ultratech) for wafer exposure, but a mask aligner (e.g., MA6, Suss Microtech) works as well. 7. High-purity SiH4 gas (>99.999%) for a-Si:H deposition. 8. A high-purity and high-density indium tin oxide (ITO) target (purity >99.99%; relative density >95%) for top electrode sputtering. 9. A high-purity Al target (>99.999%) for Al electrode sputtering.

2.2. Integration and Patterning of the Optical Interference Filter

1. A mixed solution of phosphoric acid (78.9%), nitric acid (2.8%), and acetic acid (3%) for Al etching.

2.3. Integration Detection Platform

1. A 2-mm diameter half-ball lens made of BK7 (Edmund Optics, USA) for the fluorescence-collecting microlens.

2. The SiO2/Ta2O5 multilayer filter is designed and deposited by such companies as Optoquest Co. (Japan) and Barr associates (USA), using ion-assisted deposition (IAD) or ion beam sputtering (IBS).

2. A semiconductor laser emitting light at 488 nm (e.g., Sapphire, Coherent, USA). 3. Black-anodize an Al platform to suppress light reflection on the surface. 2.4. Microfluidic Electrophoresis Device Fabrication

1. Borofloat glass wafers (76-mm diameter, 1.1-mm thick, Schott, NY) for microfluidic electrophoresis devices. 2. A mask aligner (e.g., MA6, Suss Microtech) for a wafer exposure. 3. HF solution (50%, Semiconductor grade, Morita Chemical, Japan) for channel etching of the glass wafer.

364

Kamei

4. Piranha is heated mixture of 80% concentrated sulfuric acid and 20% hydrogen peroxide. 5. A furnace (KDF-90, Denken, Japan) for thermally boding glass wafers. 2.5. Microfluidic Electrophoresis Analysis

1. Microfluidic electrophoresis devices can be fabricated by such companies as Institute of Microchemical Technology (IMT, Japan) or Micronit Microfludics (Netherlands). 2. A high-voltage power supply designed for microfluidic electrophoresis (e.g., HVS448, LabSmith, USA).

3. Methods 3.1. a-Si:H Photodiode Fabrication

The details of fabrication procedures of a-Si:H PIN photodiode are described in this section. Brief fabrication steps are as follows: Cr deposition on a glass substrate/resist patterning/etching; deposition of nip a-Si:H and then ITO/resist patterning/etching; SiN deposition/resist patterning/etching; resist patterning/Al deposition/lift-off. An a-Si:H PIN photodiode typically exhibits excellent sensitivity to visible light and its quantum efficiency exceeds 80% at a wavelength between 500 and 600 nm. Dark current of an a-Si:H photodiode should also be very low: several pA or less at room temperature when its outer diameter is 2 mm. These features are suited for high sensitivity visible fluorescence detection: 1. These instructions assume the use of multichamber plasma-enhanced chemical vapor deposition apparatus (e.g., PD-2203LS, SAMCO, Japan) such that each layer of the a-Si:H PIN photodiode is deposited in a separate chamber, suppressing cross contamination by impurities such as B and P. 2. Rinse glass wafers with water (see Note 1), before ultrasonically cleaning for 5 min. Immerse them in acetone, before ultrasonically cleaning again for 10 min. Rinse them with water and immerse them in the cleaning solution (Semico Clean 56), before ultrasonically cleaning for a further 20 min. Thoroughly rinse them with water and dry. 3. Sputter chromium on a glass wafer to a thickness of approximately 200 nm. Pattern the photoresist on the Cr-deposited glass wafer by photolithography. Wet etching is performed by dipping it in a Cr etchant until the prescribed region becomes transparent. Remove the photoresist by immersing in the resist-stripping solution at 80°C, dip the wafer in isopropyl alcohol for 1 min, rinse thoroughly with water and dry.

Miniaturized and Integrated Fluorescence Detectors

365

4. Deposit P-doped a-Si:H (n layer) at a temperature of 250°C to a thickness of approximately 20 nm by plasma decomposition of a source gas mixture of SiH4 and PH3 (e.g., PH3/SiH4 = 1,000 ppm). 5. Deposit nondoped a-Si:H (i layer) at a temperature of 250°C to a thickness of approximately 500 nm by plasma decomposition of an SiH4 source gas. 6. Deposit B-doped a-Si:H (p layer) at a temperature of 250°C to a thickness of approximately 20 nm by plasma decomposition of a source gas mixture of SiH4 and B2H6 (e.g., B2H6/SiH4 = 5,000 ppm). 7. Sputter ITO to a thickness of approximately 70 nm. Pattern photoresist on the wafer by photolithography. Etch ITO with a mixture of CH4 and H2 (e.g., CH4/H2 = 1/5) and then a-Si:H with SF6 in reactive ion etching (RIE) apparatus. Remove the photoresist by immersing in the resist stripping solution at 80°C, dip the wafer in isopropyl alcohol for 1 min, rinse thoroughly with water and dry. 8. Deposit SiN by plasma decomposition of a mixture of SiH4, NH3, and N2 at 300°C. The thickness should be approximately 300 nm. Pattern photoresist on the wafer by photolithography. Etch with SF6 in RIE. Remove the photoresist by immersing in the resist stripping solution at 80°C, dip the wafer in isopropyl alcohol for 1 min, rinse thoroughly with water and dry. 9. Pattern photoresist on the wafer by photolithography. Then sputter aluminum on it, lifting off the unwanted Al by dipping it in the resist stripping solution at 80°C. Dip the wafer in isopropyl alcohol for 1 min, rinse thoroughly with water and dry. 3.2. Integration and Patterning of the Optical Interference Filter

The details of SiO2/Ta2O5 multilayer optical interference filter patterning are described in this section. Both materials such as SiO2 and Ta2O5 are difficult to etch by RIE, so it is virtually impossible to directly etch such a thick SiO2/Ta2O5 optical filter. We have also found no photoresist appropriate to lift off such a thick optical filter and to be tolerant to filter coating temperature (230°C). Therefore, a lift off process using an Al/Si bilayer as a sacrificial layer has been adopted to pattern the optical filter. 1. Sputter aluminum to a thickness of approximately 10 µm, and then deposit Si to a thickness of 200 nm. Pattern photoresist on the wafer by photolithography. Anisotropically etch the Si film by SF6 in RIE, before isotropically wet etching with Al to form an overhang structure. 2. Send the sample to the company to deposit the SiO2/Ta2O5 multilayer optical interference filter on it (see Note 2). The spectroscopic properties of the filter we typically use are shown in Fig. 1.

366

Kamei

3. Immerse the wafer in an aluminum etchant at approximately 60°C, lifting off the unwanted region of the optical interference filter (see Note 3). Rinse thoroughly with water and dry. 4. Deposit SiO2 by plasma decomposition of a mixture of tetraethoxysilane (TEOS) and O2 at 300°C. The thickness should be approximately 400 nm. 5. Pattern photoresist on the wafer by photolithography. Etch SiO2 layer to make contact with Al as well as the Cr electrode. A plan view of the micromachined a-Si:H fluorescence detector is shown in Fig. 2a together with a schematic illustration of its cross-sectional view in Fig. 2b. The integrated a-Si:H fluorescence detector that can be coupled with a microfluidic electrophoresis device is shown in Fig. 3, and a more extended system diagram is shown in Fig. 4. A half-ball lens (2-mm diameter) and the micromachined a-Si:H fluores-

(a) 100

T (%)

80 60 40 20 0

(b) 102 101

Log T (%)

3.3. Integrated Detection System

100 10−1 10−2 10−3 10−4 400

450

500

550

600

650

700

Wavelength (nm) Fig. 1. Simulated spectroscopic properties of the sio2/Ta2O5 multilayer optical interference filter we typically use. a Transmission (%) vs. wavelength (nm); b Log (transmission (%) ) vs. wavelength (nm).

Miniaturized and Integrated Fluorescence Detectors

367

(a) µFD

Cr

Al

Aperture

500µm

(b) SiO2 / Ta2O5 Optical filter

SiN

ITO p i a-Si:H n

µFD~7µm

SiO Al Glass substrate

Cr

Fig. 2 a. Plan view of the optical micrograph of a micromachined a-Si:H fluorescence detector (µFD). b Schematic cross-sectional view of the detector. (c) 1 mm Microchannel 2 mm-dia. Half-Ball Lens

Fluorescence

Semiconductor Laser (488 nm)

Fig. 3 a. A microfluidic electrophoresis device mounted on a black-anodized Al platform that comprises a half-ball lens and a micromachined a-Si:H fluorescence detector. b A micromachined fluorescence detector placed underneath the Al platform. c A schematic cross-sectional view of the microfluidic electrophoresis device, half-ball lens, and fluorescence detector. The excitation light from a semiconductor laser passes up through the fluorescence detector and half-ball lens to irradiate a microchannel of the microfluidic electrophoresis device.

368

Kamei

HV

Computer

microfluidic electrophoresis devices

micromachined fluorescence detector

LIA f

R

Θ

LASER Chopper

Lens

Mirror

Fig. 4. A system diagram of the integrated detection system for a microfluidic electrophoresis device. The excitation light passes through an optical chopper, lens, mirror, and a micromachined a-Si:H fluorescence detector, and is loosely focused in the microchannel of a microfluidic electrophoresis device. The photocurrent under a reverse-bias voltage of 1 V synchronized to the chopped laser light is detected by a lock-in amplifier (LIA). Data acquisition and high-voltage (HV) control for microfluidic separation are accomplished by a personal computer.

cence detector were assembled and supported by black-anodized aluminum, forming a compact platform for attaching to a microfluidic electrophoresis device. The annular micromachined a-Si:H fluorescence detector and transparent glass substrate allow vertical laser excitation while avoiding direct incidence of the excitation light on the detector. Fluorescence is collected by the half-ball lens and transmitted by the optical interference filter, which simultaneously eliminates the excitation light. Ray trace simulation (ZEMAX, Focus Software, CA, USA) indicated that a 2-mm-diameter half-ball lens would approximately collimate the fluorescence emitted from a microchannel located behind the 1-mm-thick Borofloat glass substrate. The procedure to construct the integrated detection system is as follows: 1. Mount the half-ball lens on the Al platform by using epoxy resin to make their surface flat. 2. Loosely focus laser light (approximately 30 µm) through a convex lens (NA ~0.01). The focal point may be close to the microchannel. 3. Adjust a mirror underneath the detector (see Fig. 4) so that laser light passes upward normal to the half-ball lens.

Miniaturized and Integrated Fluorescence Detectors

369

4. Place an iris above the center of the Al platform (half-ball lens) and then adjust the position of the Al platform (half-ball lens) by two orthogonal micrometers so that the laser light goes through the iris, making sure the laser light passes through the center of the half-ball lens. 5. Place and fix the micromachined fluorescence detector on the Al platform to coaxially align the detector with the half ball lens. 6. Apply a reverse-bias voltage of 1 V to the a-Si:H photodiode and connect its output to a lock-in amplifier (LIA; e.g., SR840, Stanford Research Systems, CA, USA). Direct the laser light through an optical chopper (e.g., SR540, Stanford Research Systems, CA, USA) and detect the photocurrent synchronized with the chopped laser light to reduce background interference due to environmental light. 7. Digitize and acquire an LIA output at 20 Hz by using LabView (e.g., NI-DAQ E6024, National Instruments, TX, USA). 3.4. Microfluidic Electrophoresis Device Fabrication

The procedure to fabricate microfluidic electrophoretic devices is described in this section. Plasma-deposited Si film is used as a hard mask for HF wet etching and glass wafers are thermally bonded. The mask design of the microfluidic CE, which we typically use is shown in Fig. 5. The width of microchannel in the mask is 20 µm. 1. Rinse glass wafers with water, before ultrasonically cleaning for 5 min. Immerse them in acetone, before ultrasonically cleaning again for 10 min. Rinse them with water and immerse them in the cleaning solution, before ultrasonically cleaning for a further 20 min. Thoroughly rinse them with water and dry. 2. Deposit a-Si:H at a temperature of 250°C to a thickness of approximately 250 nm by plasma decomposition of an SiH4 source gas. 3. Pattern the photoresist on the Si-deposited glass wafer by photolithography. 4. Etch a-Si:H with SF6 in RIE. Remove the photoresist by immersing in the resist stripping solution at 80°C, dip the wafer in isopropyl alcohol for 1 min, rinse thoroughly with water, and dry. 5. Dip the glass wafer in HF solution for 7 min, producing a 120-µm wide, 50-µm deep microchannel. Rinse thoroughly with water and dry. 6. Drill access holes with diamond bits for sample, waste, cathode, and anode reservoirs. Immerse the drilled wafers in acetone, before ultrasonically cleaning for 15 min. Thoroughly rinse them with water and dry.

370

Kamei

5

4

3

1

2

S C

A

W

5

4

5

4

3 3

1

2 2

1

2

3

4

5

1

1

2

3

4

5

1

2

3

4

5

Fig. 5. Mask design of the microfluidic electrophoresis device, which we typically use. S, W, C, and A, respectively, stand for the sample, waste, cathode, and anode reservoirs.

7. Strip off both sides of a-Si:H layers with SF6 in RIE. 8. Immerse them in the cleaning solution (Semico Clean 56), before ultrasonically cleaning for 15 min. Thoroughly rinse them with water. 9. Immerse them in piranha at 120°C for 15 min. Thoroughly rinse them with water and dry. 10. Align the drilled and etched wafer with a blank wafer by hand. Sandwich them by macor blocks and insert them into a furnace. 11. Raise temperature in the furnace until 660°C at a rate of 5–10°C/min. Keep temperature at 660°C for 12 h and naturally lower temperature. 3.5. Microfluidic Electrophoresis Analysis

The procedure to perform microfluidic electrophoretic separation of DNA restriction fragment digests is described in this section. A so-called cross-injection method has been adopted. A slight modification of the mask design or sample preparation enables detection and identification of pathogens, enantiomer detection of amino acids, and detection of glucose (5, 7). A further reduction of background photocurrent due to laser light scattering

Miniaturized and Integrated Fluorescence Detectors

371

should improve LOD, making the micromachined fluorescence detector more applicable to biochemical assays including DNA sequencing. 1. Dissolve hydroxyethylcellulose (HEC, Hercules, DE, USA) in a 1×Tris-acetate/EDTA (TAE) buffer to 1% w/v for the sieving matrix. Dilute oxyazole yellow (YO, Invitrogen, OR, USA), an intercalating dye used for fluorescent labeling of DNA fragments, to 1 µM in the sieving matrix. Degas this mixture under vacuum while stirring for 5 min and then centrifuge it to remove any bubbles. 2. Apply labeled HEC to the anode with a syringe and force it through the channel. 3. Dilute a HaeIII digest of φX174 bacteriophage DNA (Takara, Japan) to 100 ng/µL in water (see Note 4) and then put the diluted DNA sample (5 µL) in the sample reservoir. Put 10 µL of the 1×TAE buffer in each of the other reservoirs. 4. Drop index-matching fluid (series A, n = 1.474, Cargille Laboratory, NJ, USA) on the half-ball lens to optically couple with the microfluidic CE. 5. Adjust the position of the microfluidic CE chip with the two orthogonal micrometers so that the laser irradiates the microchannel of the microfluidic electrophoresis device. This irradiated point corresponds to a detection point several cm away (approximately 4 cm in our case) from the injection cross point of the channels. 6. Insert Pt electrodes into all reservoirs to make electrical contact. 7. Inject the DNA sample at 830 V/cm for 20–40 s from the sample reservoir to waste reservoir while the potentials of the other reservoirs are close to the floating potential. 8. Switch the electric field to electrophorese the sample to the separation channel at 230 V/cm while applying a back-biased electric field of 160 V/cm between the injection cross point and both the sample and waste reservoirs to avoid any bleeding of the sample. 9. Start to measure the LIA output after the electrophoresis starts. A plot of the LIA output as a function of time gives an electropherogram. Figure 6 presents results obtained according to the above procedure. Despite neither data processing nor baseline subtraction, all 11 peaks of the DNA fragments could be detected with a good S/N ratio and well resolved, including 271 and 281 bp peaks, in approximately 2-min separation. Provided that the S/N ratio for the peak of each fragment is determined as the ratio of the peak height to noise level of the baseline, the LOD values

Kamei

872 1078 1353

372

C

271

281

194 234

310

W

S

118

603

Fluorescence

~3.5 cm

72

A

0

20

40

60

80

100

120

140

160

Time (s) Fig. 6. Electropherogram of the Hae III digest of φX174 bacteriophage DNA (100 ng/µL) using the micromachined fluorescence detector. The DNA fragment length is indicated in bp for each peak. The channel geometry for the microfluidic CE is schematically shown top left, where S, W, C, and A, respectively, stand for the sample, waste, cathode, and anode reservoirs.

(S/N = 3) for each fragment vary from 150 to 430 pg/µL with an average figure of 250 pg/µL. Theoretical plates for each fragment range from 26,000 to 128,000, with an average figure of 68,000. A conventional confocal fluorescence detection system typically exhibits an LOD value of 10–100 pM in fluorescein concentration, still much lower than the micromachined fluorescence detector, but such high sensitivity is not needed for many applications. When high-sensitivity detection is required, it may be more effective and easier to compensate the detector sensitivity by coupling with the DNA preamplification and/or cell preconcentration method (8), because the optical interference filter we adopted in this work is already superior and suited for visible fluorescence detection from fluorescein and YO.

4. Notes 1. Unless stated otherwise, all the solutions should be prepared in water that has a resistivity of more than 18 MΩ cm and total organic content of less than ten parts per billion. This is referred to as “water” in this text. The glass wafers should also be rinsed with this water. 2. Design of an optical interference filter is proprietary but its structure we have adopted here is essentially multiple Fabry–Perot cavity. For example, in the case of design: Air

Miniaturized and Integrated Fluorescence Detectors

373

100

T (%)

80 60 40 20 0 400

450

500

550

600

650

700

Wavelength (nm) Fig. 7. Computed transmittance of the multiple cavity filter against wavelength (nm). Design: Air|LHLL( (HLH)4L)10HLHHLH|glass. L = SiO2 (n = 1.45980), H = Ta2O5 (n = 2.14064) and reference wavelength is 553 nm.

|LHLL( (HLH)4L)10HLHHLH|glass where L and H stand for a quarter-wave layer of high and low refractive index, respectively; L = SiO2, H = Ta2O5, and reference wavelength is 553 nm, computed transmittance of such a filter is shown in Fig. 7. Although more ripples appear in Fig. 7 compared to commercial filter shown in Fig. 1a, similar features could be reproduced. 3. The lift-off process to pattern the optical interference filter is the most difficult; it might take 1 week and the Al etchant often damages the Al electrode of the a-Si:H photodiode. To mitigate against this, we make many scratches on the unwanted region (optical filter/Al/Si) with a pitch of approximately 2 mm, prior to the lift-off process, reducing the lift-off time to 4–5 h and preventing the Al electrode from being damaged. 4. The DNA sample is normally diluted with the running buffer, in this case, the 1×TAE buffer. The DNA sample is diluted with water, so that stacking takes place to concentrate the sample plug in the separation channel and enhance the peak intensity.

Acknowledgments This study was carried out partly in collaboration with Prof. Richard Mathies’ group at University of California in Berkeley. I would like to particularly thank Dr. James Scherrer for his contribution to construct the optical setup and Dr. Brian Paegel for his consultation concerning microfluidic electrophoresis. This work was supported in part by New Energy and Industrial Technology Development Organization (NEDO) of Japan.

374

Kamei

References 1. Harrison, D. J., Fluri, K., Seiler, K., Fan, Z. H., Effenhauser, C. S., and Manz, A. (1993) Micromachining a miniaturized capillary electrophoresis-based chemical-analysis system on a chip. Science 261, 895–97 2. Unger,M.A.,Chou,H.P.,Thorsen,T.,Scherer,A., and Quake, S. R. (2000) Monolithic microfabricated valves and pumps by multilayer soft lithography. Science 288, 113–16 3. Fu, A. Y., Chou, H. P., Spence, C., Arnold, F. H., and Quake, S. R. (2002) An integrated microfabricated cell sorter. Analytical Chemistry 74, 2451–57 4. Hansen, C. L., Skordalakes, E., Berger, J. M., and Quake, S. R. (2002) A robust and scalable microfluidic metering method that allows protein crystal growth by free interface diffusion. Proceedings of the National Academy of Sciences of the United States of America 99, 16531–36

5. Kamei, T., Paegel, B. M., Scherer, J. R., Skelley, A. M., Street, R. A., and Mathies, R. A. (2003) Integrated hydrogenated amorphous Si photodiode detector for microfluidic bioanalytical devices. Analytical Chemistry 75, 5300–05 6. Kamei, T., and Wada, T. (2006) Contact-lens type of micromachined hydrogenated amorphous Si fluorescence detector coupled with microfluidic electrophoresis devices. Applied Physics Letters 89, 114101 7. Kamei, T., Toriello, N. M., Lagally, E. T., Blazej, R. G., Scherer, J. R., Street, R. A., and Mathies, R. A. (2005) Microfluidic genetic analysis with an integrated a-Si: H detector. Biomedical Microdevices 7, 147–52 8. Lagally, E. T., Lee, S. H., and Soh, H. T. (2005) Integrated microsystem for dielectrophoretic cell concentration and genetic detection. Lab on a Chip 5, 1053–58

Chapter 21 Photomultiplier Tubes in Biosensors Yafeng Guan Summary Photomultiplier tubes (PMT) are widely used for the weak light detection in some types of biosensors. A light detection system for biosensors based on PMT generally contains optic fibers, PMT, and filters. Basic principles of those accessories were provided in this chapter. The guides to selecting fibers, filters, PMT, and power suppliers in practical applications were presented. Major problems that may occur with the instruments were listed and discussed. Key words: Fluorescence detection, LIF, LED-IF, PMT.

1. Introduction Light detection technology is a powerful tool that provides deeper understanding of more sophisticated phenomena. Measurement using light detectors offers unique advantages such as nondestructive analysis of a substance, high-speed properties, and extremely high detectability. By coupling optical principles and methods with the biological specificity, optical biosensors have been a particular field of biosensors. In these cases, photomultiplier tubes (PMT) is often the optimal optical detector for biosensors due to its superior properties of high sensitivity to low-light-level, remarkable stability, fast response to wide wavelength spectral, and wide linearity. PMT in biosensors is generally used to measure the light intensity of fluorescence, absorption light, and chemiluminescence, produced during the process of the specific combination between the biological recognition element and substrate. Because of the wide wavelength response

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_21

375

376

Guan

of PMT, filters or monochromators should be used to enhance the wavelength resolution of sensors when applied to single wavelength detection. In this chapter, basic principles and structures of PMT will not be introduced in detail. We will briefly describe an optical detection system using PMT to detect the slight changes of the weak light signal produced in biosensors. In addition, some important aspects related to the correct use of PMT, such as optic fibers, filters, and power suppliers, will be mentioned in proper place.

2. Materials A light diode emitting at 470 nm (50 mW, Shifeng Optics, China) was chosen as the light source. The lens of f10 and f15 (GCL010131, GCL-010132, Daheng Optics, China) was used to focus the excitation light. Filter for the light source was BP470 (Huibo Optics, China). The sample flowed in a fused-silica capillary (100 µm i. d., Yongnian, Inc., China). The optical fiber (core 0.5 mm, cladding 0.6 mm, Chunhui, Inc., China) was used to collect the fluorescence. The emitting filter was chosen to be BP530 (Huibo Optics, China). PMT (H5784) was obtained from Hamamatsu, Japan. A weak current amplifier TJ-110 was from Taiji Computer cooperation limited, China. The data acquisition station KF-98 (Taiji co., China) was used for signal acquisition and data processing.

3. Methods 3.1. Filter

Filters are used to pass a band of wavelengths (bandpass filters) or to block wavelengths longer or shorter than some desired value (cutoff filters). The use of filter before the PMT can reduce the background light and enhance the sensitivity (1). The characteristics of the filters are illustrated by plots of their spectral transmittance (T) vs. wavelength, as shown in Fig. 1. A filter with higher spectral transmittance can be beneficial to the sensitivity of the biosensor constructed. The choice of the filters should be made by considering wavelength of the excitation light and the emission light of the system used in the biosensors. Figure 2 shows a typical fluorescence system and the corresponding filter used before the PMT. To get high sensitivity, the lm should be consistent with the wavelength

Photomultiplier Tubes in Biosensors

377

(a) 100

T (%)

80 60 40 20 0

(b) 102

Log T (%)

101 100 10−1 10−2 10−3 10−4 400

450

500

550

600

650

700

Wavelength (nm) Fig. 1. Transmittance of bandpass (a) and short-wavelength cutoff (b) filters. a lm, the wavelength of maximum transmittance; Tm, the maximum transmittance; and ∆l, FWHM. b lm, the wavelength at which the transmittance over 90% of its maximum value; lh, the wavelength at which T = 0.5Tm; and lc, the cutoff wavelength.

of the emission light, and kept away from the wavelength of the excitation light used in the system. 3.2. Optical Fiber

Optical fibers are experiencing greater use in biosensors for several reasons. Because they are mechanically flexible, light can be transmitted over curved paths. Thus the optical fiber can replace several mirrors in directing light between two points in a biosensor. A single optical fiber cannot transmit an image; bundles of fine glass, quartz, or plastic fibers can be used for image transmission if the fibers are small enough in diameter that each fiber transmits rays from a small area of the object. The user can choose the optical fiber according to the biosensor constructed.

3.3. Photomultiplier Tube

When light enters a PMT, it is detected and it produces a signal through the following processes:

378

Guan

(a) µFD

Cr

Al

500µm

Aperture

(b) SiO2 / Ta2O5 Optical filter

SiN

SiO Al Glass substrate

ITO p i a-Si:H n

µFD~7µm

Cr

Fig. 2. The fluorescence system and the corresponding filter used before the PMT. a A typical fluorescence system. ex is the excitation light and em is the emission light. b The bandpass filter chosen for the system.

1. Passes through the input window. 2. Electrons in the photocathode are excited by the light and emitted into the vacuum. 3. Photoelectrons are accelerated and focused by the focusing electrode onto the first dynode where they are multiplied by means of secondary electron emission. The secondary emission is repeated at the successive dynodes. 4. The multiplied secondary electrons emitted from the last dynode are finally collected by the anode. It is important to know the conditions of the light to be detected before we choose a proper PMT. The parameters listed in Table 1 should be taken into account when making a selection. Spectral response characteristics of PMT are mainly related to photocathodes and window materials. Here we introduce these materials and the corresponding detection wavelength range. 3.3.1. Photocathodes

Most photocathodes (2–4) are made of compound semiconductors, which consist of alkali metals with a low work function. There are approximately ten kinds of photocathodes currently

Photomultiplier Tubes in Biosensors

379

Table 1 The parameters should be taken into account when selecting a PMT Incident light conditions

Selection reference

Light wavelength

Window material Photocathode spectral response

Light intensity

Number of dynodes Dynode type Voltage applied to dynodes

Light beam size

Effective diameter Viewing configuration (side-on of head-on)

Speed of the optical phenomenon

Response time

employed in practical applications. Each photocathode is available as a transmission (semitransparent) type or a reflection (opaque) type, with different device characteristics. The photocathode materials commonly used in PMT are as follows: 1. Ag–O–Cs. Transmission type photocathodes using this material are sensitive from the visible light through the near infrared region, i.e., from 300 to 1,200 nm, which the reflection type exhibits a slightly narrower spectral response region from 300 to 1,100 nm. Compared with other photocathodes, this photocathode has lower sensitivity in the visible region, but good sensitivity at longer wavelengths in the near infrared region. So both transmission and reflection type Ag–O–Cs photocathodes are mainly used for near infrared detection. 2. GaAs (Cs). A GaAs crystal activated with cesium is used for both reflection type and transmission type photocathodes. The reflection type GaAs (Cs) photocathode has sensitivity across a wide range from the ultraviolet through the near infrared region around 900 nm. It demonstrates a nearly flat, high-sensitive spectral response curve from 300 to 850 nm. The transmission type has a narrower spectral response range because shorter wavelengths are absorbed. It should be noted that if exposed to incident light with high intensity, these photocathodes tend to suffer sensitivity degradation when compared with other photocathodes primarily composed of alkali metals.

380

Guan

3. InGaAs. This photocathode provides a spectral response extending further into the infrared region than the GaAs photocathode. Additionally, it offers a superior signal-to-noise ratio in the neighborhood of 900–1,000 nm in comparison with the Ag–O–Cs photocathode. 4. Cs–I and Cs–Te. Cs–I and Cs–Te are not sensitive to solar radiation therefore often called “solar blind.” The sensitivity of Cs–I sharply falls off at wavelengths longer than 200 and it is exclusively used for vacuum ultraviolet detection. Cs–Te is not sensitive to wavelengths longer than 300 nm. 5. Multialkali (Sb–Na–K–Cs). This photocathode uses three or more kinds of alkali metals. Due to high sensitivity over a wide spectral response range from the ultraviolet through the near infrared region around 850 nm, this photocathode is widely used in broadband spectrophotometers. 6. High-temperature, low-noise bialkali (Sb–Na–K). As with bialkali photocathodes, two kinds of alkali metals are used in this photocathode type. The spectral response range is almost identical to that of bialkali photocathodes, but the sensitivity is somewhat lower. This photocathode can withstand operating temperatures up to 175°C while other normal photocathodes are guaranteed to no higher than 50°C. For this reason, it is ideally suited for use in oil well logging where PMT are often subjected to high temperatures. In addition, when used at room temperatures, this photocathode exhibits very low dark, which makes it very useful in low-level light measurement such as photon counting applications where low noise is a prerequisite. 3.3.2. Window Materials

Most photocathodes have high sensitivity down to the ultraviolet region. However, because ultraviolet radiation tends to be absorbed by the window material, the short wavelength limit is determined by the ultraviolet transmittance of the window material (5, 6). The window materials commonly used in PMT are as follows: 1. Borosilicate glass. This is the most commonly used window material. Because the borosilicate glass has a thermal expansion coefficient very close to that of Kovar alloy, which is used for the leads of PMT, it is often called “Kovar glass.” The borosilicate glass does not transmit ultraviolet radiation shorter than 300 nm. It is not suited for ultraviolet detection shorter than this wavelength. Moreover, some types of head-on PMT using a bialkali photocathode employ a special borosilicate glass (so-called “K-free glass”) containing a very small amount of potassium (K40), which may cause unwanted background counts. The K-free glass is mainly used for PMT designed for scintillation counting where low background counts are desirable.

Photomultiplier Tubes in Biosensors

381

2. UV glass (UV-transmitting glass). As the name implies, this transmits ultraviolet radiation well. The short wavelength cutoff of the UV glass extends to 185 nm. 3. Synthetic silica. Synthetic silica transmits ultraviolet radiation down to 160 nm and in comparison with fused silica, offers lower levels of absorption in the ultraviolet region. Since silica has a thermal expansion coefficient greatly different from that of the Kovar alloy used for the stem pins (leads) of PMT, it is not suited for used as the bulb stem. As a result, a borosilicate glass is used for the bulb stem and then a graded seal, using glasses with gradually changing thermal silica bulb. Because of this structure, the graded seal is very fragile and proper care should be taken when handing the tube. In addition, helium gas may permeate through the silica bulb and cause an increase in noise. Avoid operating or storing such tubes in environments where helium is present.

3.3.3. Dynode Types and Features

4. MgF2 crystal. The crystals of alkali halide are superior in transmitting ultraviolet radiation, but have disadvantage of deliquescence. A magnesium fluoride (MgF2) crystal is used as a practical window material because it offers very low deliquescence and allows transmission of vacuum ultraviolet radiation down to 115 nm. There are a variety of dynode types available and each type exhibits different gain, time response, uniformity, and secondaryelectron collection efficiency depending upon the structure and the number of stages. The optimum dynode type must be selected according to application: 1. Circular-cage type. The circular-cage type has an advantage of compactness and is used in all side-on PMT and in some head-on PMT. The circular-cage type also features fast time response. 2. Box-and-grid type. This type, widely used in head-on PMT, is superior in photoelectron collection efficiency. Accordingly, PMT using this dynode offer high detection efficiency and good uniformity. 3. Linear-focused type. As with the box-and-grid type, the linearfocused type is widely used in head-on PMT. Its prime features include fast time response, good time resolution, and excellent pulse linearity. 4. Venetian blind type. The Venetian blind type creates an electric field that easily collects electrons, and is mainly used for head-on PMT with a large photocathode diameter. 5. Mesh type. This type of dynode uses mesh electrodes stacked in close proximity to each other. There are two types: coarse mesh type and fine mesh type. Both are excellent in output linearity and have high immunity to magnetic fields. When used with a cross wire anode or multianode, the position of

382

Guan

incident light can be detected. Fine mesh types are developed primarily for PMT, which are used in high magnetic fields. 6. MCP (microchannel plate). A microchannel plate (MCP) with 1-mm thickness is used as the base for this dynode structure. This structure exhibits dramatically improved time resolution as compared with other discrete dynode structure. It also assures stable gain in high magnetic field and provides position-sensitive capabilities when combined with a special anode. 7. Metal channel dynode. This dynode structure consists of extremely thin electrodes fabricated by advanced micromachining technology and precisely stacked according to computer simulation of electron trajectories. Since each dynode is in close proximity to one another, the electron path length is very short ensuring excellent time characteristics and stable gain even in magnetic fields. 8. Electron bombardment type. In this type, photoelectrons are accelerated by a high voltage and strike a semiconductor so that the photoelectron energy is transferred to the semiconductor, producing a gain. This simple structure features a small noise figure, excellent uniformity, and high linearity. Figure 3 shows a typical application example in which a PMT is used in a biosensor. A blue LED driven by a 5 V constant voltage source through a 100 Ω current-limiting resistor was used as the excitation source. LED light was collimated and focused with two quartz achromatic lenses into the capillary. To

7

6

3

5

4

1 2

2

Fig. 3. The structure of a biosensor (1) light source; (2) lens; (3) filter for the light source; (4) sample; (5) optical fiber; (6) filter; and (7) PMT.

Photomultiplier Tubes in Biosensors

383

reduce the scattering light from the capillary wall, an aperture of 0.2 mm was used to restrict the beam size. An interference filter was inserted between two lenses to eliminate the interference of long wavelength from LED. A detection window on the capillary was formed by burning off the polyimide coating (5 mm in length) with an electrical coiled resistance. Fluorescence was collected with a right-angle geometry by an optical fiber (see Note 1) and passed through two blocks of interference filters. The distance between the fiber and the capillary, as well as the distance between the fiber and the filter (see Notes 2 & 3), was set to be 0.5 mm. The fluorescence signal was then detected by a metal package miniaturized PMT (see Notes 4–8). The signal from the PMT was acquired by chromatographic workstation. Figure 4 illustrated a solid schematic view of the sensor. The application of the biosensor to detect the riboflavin in watermelon sample was demonstrated. The limit of detection for riboflavin was 10 µg/L. The sensor exhibited an excellent linear behavior over the concentration range of 10–1,000 µg/L (R = 0.9996). The riboflavin in the watermelon sample was determined to be 100 µg/L (S/N = 3). Figure 5 showed the chromatogram obtained from an analysis of riboflavin solution with several concentrations. The sensor can also be applied to electrophoresis to detect several biomolecules. Figure 6 demonstrated the analysis of three amino acids labeled by FITC. The

Fig. 4. The solid schematic view of the biosensor (1) light source; (2) lens; (3) filter for the light source; (4) sample; (5) optical fiber; (6) filter; and (7) PMT.

384

Guan

Fig. 5. Chromatogram of riboflavin solutions with several concentrations.

Fig. 6. Electropherogram of FITC-labeled amino acids (1) FITC-labeled lysine; (2) excess FITC; (3) FITC-labeled Tryptophan; and (4) FITC-labeled phenylalanine.

detection limits of 10, 9, and 4.8 nM (S/N = 3) were achieved for lysine (Lys), tryptophan (Trp), and phenylalanine (Phe), respectively. Potential applications of the sensor are routine analysis of protein, peptide, amino acids, and others compounds. 3.4. Selecting a Power Supply For PMT

The operation stability of a PMT depends on the total stability of the power supply characteristics including drift, ripple, temperature dependence, input regulation, and load regulation. The power supply must provide high stability that is at least ten times as stable as the output stability required for the PMT. Series-regulator type high-voltage power supplies have been widely used with PMT. Recently, a variety of switching-regulator types have been put on the market and are becoming widely

Photomultiplier Tubes in Biosensors

385

used. Most of the switching-regulator type power supplies offer compactness and light weight, yet provide high voltage and high current. However, with some models, the switching noise is superimposed on the AC input and high-voltage output or the noise is radiated. Thus, sufficient care is required when selecting this type of power supply, especially in low-light-level detection, measurement involving fast signal processing, and photon-counting applications. The high-voltage power supply should have sufficient capacity to supply a maximum output current which is at least two times the current actually flowing through the voltage-divider circuit used with the PMT. Table 2 shows the guide for selecting the correct high-voltage power supply. Recently, commercial PMT modules comprised of a PMT, a high-voltage power supply (HV) circuit, and a voltage-divider circuit are available on the market. Using this type of PMT modules eliminate an external HV from an external power supply. All that need is simple wiring and low-voltage input. Supply approximately 15 V to the low-voltage input, ground the GND terminal, and connect the control voltage and reference voltage input according to the gain adjustment method as described by manufactures (see Notes 9 & 10). 3.5. Connecting the PMT Output to Data Acquisition System

The output of a PMT is a current, while the expected signal of signal conditioning circuit is a voltage. Therefore, the current output should be converted into a voltage signal by some means. One simple method for the current output of a PMT into a voltage output is to use a load resistance. Since the PMT may be thought of as an ideal constant current source, the voltage output of I × RL (I is output current and RL is the load resistance) can be obtained. Another method is to use a current-to-voltage conversion circuit using an operational amplifier. A basic circuit using an operational amplifier is shown in Fig. 7. TJ-110 we used is a

Table 2 The characteristics of the high-voltage power supply for PMT Line regulation

±0.01% or less

Load regulation

±0.2% or less

Ripple noise

0.05% or less

Temperature coefficient

±0.05%/K or less

386

Guan

Fig. 7. Current-to-voltage conversion circuit using an operational amplifier.

typical current amplifier module using an operational amplifier current-to-voltage conversion circuit. With this circuit, the output voltage Vo is given by Vo = –Ip· zRf · Some PMT modules contain an integrated current-to-voltage conversion circuit and the output of PMT is a voltage. It is not necessary to use an additional conversion circuit. After conversion, the voltage signal is measured with a high sensitive ammeter or connected to the input terminals of the data acquisition system KF-98.

4. Notes 1. For optical fibers, make sure that the end should be flattened and mirror polished. Pay attention to the minimum bending radius of the fiber used; do not exceed the value. 2. Do not touch the filter with bare hands or expose it to the dust. If not used, store it in dry environments. 3. Make sure that the optical fiber placed to the filter as near as possible, but had no physical contact with the filter, since the fiber may cause damage to the film of the filter. 4. PMT is a very high sensitive photodetector; the users have to read-through the guide for users before use, and handle/operate the module carefully. 5. The PMT is fragile by shock and vibration, so handle it carefully not to drop or add excess shock. 6. Dust and fingerprints in the window will cause loss of signal light transmittance, so do not touch the window portion of the PMT with bare hands or expose it to the dust. Should it be dirty, wipe it with alcohol.

Photomultiplier Tubes in Biosensors

387

7. Do not expose the PMT to strong light, even when it is not operated. Direct sunlight and other strong light illumination may cause damage to it. When the PMT is not used, keep it in dark storage. 8. The PMT should not be stored or operated in an environment of high-pressure Helium gas over partial pressure in air. Helium gas can penetrate through the window of the PMT, and increase the noise level or damage the performances. 9. Check the wiring before turning on the power supply. If the voltage is turned on at incorrect cable wiring, for example, when the voltage polarity is supplied incorrectly, the PMT can be damaged. 10. Do not supply the voltage over the maximum of the guide for users said. It will cause the damage to the PMT.

References 1. D. James, J.R. Ingel, R.C. Stanley, Spectrochemical Analysis, Prentice-Hall, Englewood Cliffs, NJ (1988) 2. Hamamatsu Photonics Catalog: Photomultiplier Tubes 3. T. Hiruma, SAMPE Journal 24, 35 (1988)

4. A.H. Sommer, Photoemissive Materials, Robert E. Krieger Publishing Company, Huntington, NY (1980) 5. Handbook of Optics, McGraw-Hill, New York, NY (1987) 6. J.A.R. Samson: “Techniques of Vacuum Ultraviolet Spectroscopy”: Wiley, New York, NY (1967)

Chapter 22 Integrating Waveguide Biosensor Shuhong Li, Platte Amstutz III, Cha-Mei Tang, Jun Hang, Peixuan Zhu, Yunqi Zhang, Daniel R. Shelton, Jeffrey S. Karns Summary The Integrating Waveguide Biosensor was developed for rapid and sensitive detection of bacterial cells, spores, and toxins. A sandwich format of immunoassay was employed using Salmonella as model. The analyte was immunocaptured on the inner surface of the waveguide and then detected by the antibody conjugated with fluorescent dye. The waveguide was illuminated by an excitation light at a 90° angle. The emitted light from fluorescent labels on the surface of the waveguide was efficiently collected and channeled to a detector at the end of the waveguide, while minimizing interference from the excitation light. Utilizing fluorescent dye Cy5, a 635-nm diode laser for excitation, and a photomultiplier tube detector, the Integrating Waveguide Sensor System was able to detect approximately ten captured cells of Salmonella. Key words: Biosensor, Integrating waveguide sensor, Integrating waveguide biosensor, Fluorescence detection, Salmonella.

1. Introduction When fluorescent labels emit light, the emission is typically in all directions such that only a small fraction of the light is collected by the detector as signal. Simultaneously, light from the excitation source, auto fluorescence from the sample and the sample container, Raman emission from water, and other light collected by the detector contribute to background noise. The ability to maximize the signal while minimizing the background noise, i.e., high signal-to-noise ratio results in a lower limit of detection and improved instrument sensitivity.

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_22

389

390

Li et al.

A platform to detect analyte on the surface of a waveguide is shown in Fig. 1, demonstrating a captured analyte on the inside surface of the capillary waveguide binding a fluorescent detection antibody. The signal is obtained by illuminating the capillary tube at a 90° angle relative to its length and subsequent collection of the emitted fluorescence from one end of the waveguide. The emitted fluorescent light is efficiently gathered and guided to the end of the capillary tube waveguide and goes through a set of lenses and optical filters to the optical detector. The signal is maximized by integrating the emitted light from all of the fluorescent labels on the capillary tube surface. At the same time, the background noise is reduced as a consequence of the excitation light being directed at a 90° angle to the waveguide surface. The Integrating Waveguide Biosensor (IWB) has an inherently high signal-to-noise ratio, because it gathers a high percentage of the fluorescence signal, and because the noise contributed by the excitation is reduced. The IWB was originally developed by Ligler et al. at the Naval Research Laboratory (NRL) in Washington, DC (1, 2). Ligler et al. (2) reported a detection limit of 40 pg/mL for mouse IgG and 30 pg/mL for staphylococcal enterotoxin B (SEB) in a sandwich assay format, which is about 100-fold more sensitive than the evanescent-wave fiber-optic and array biosensor technologies previously developed (3–8). We have constructed a compact and portable instrument using a low-cost capillary tube as the waveguide. When the IWB was applied to the detection of Salmonella, the limit of detection was ten captured Salmonella cells. The IWB utilizing solid phase assay of analyte on the waveguide surface belongs to a family of integrating waveguide sensors detection principles. The implementation of the liquid-phase

Fig. 1. Basic principle of the integrating waveguide sensor where a capillary is used as the waveguide. The analyte is captured on the inner surface of the capillary waveguide. Fig. 2. Schematic of Creatv’s Integrating Waveguide Sensor.

Integrating Waveguide Biosensor

391

detection of the Integrating Waveguide Sensor is described in Chap. 24 (9).

2. Materials 2.1. Instrument Components

1. Illumination light source. Diode laser of 635 nm, 30 mW (LaserMax, Rochester, NY). 2. Signal cleanup filters. A RG665 color glass filter (Edmund Industrial Optics, Barrington, NJ) and an interference band-pass 670-nm filter (Omega Optical, Brattleboro, VT). 3. Detector. Photomultiplier tube (PMT) H7827-11 (Hamamatsu, Bridgewater, NJ). 4. Signal digitalizing. DT9802 Data acquisition card (Data Translation, Inc., Marlboro, MA). 5. Laser line generator. A rod lens from LaserMax, Rochester, NY, combined with a cylinder lens from Edmund Industrial Optics, Barrington, NJ. 6. Signal collimating lens. An aspheric lens combined with a double convex lens. 7. Signal focusing lens. Double convex lens from Edmund Industrial Optics, Barrington, NJ. 8. User interface. LabVIEW™ software (National Instruments Corp., Austin, TX).

2.2. Simulation and Analysis

1. Optical ray tracing. TracePro software (Lambda Research Corp., Littleton, MA).

2.3. Capillary Waveguide

1. Capillary waveguides. 50 mm long borosilicate glass capillary tube with 1.66 mm outer diameter and 1.23 mm inner diameter (Drummond Scientific Company, Broomall, PA).

2.4. Chemicals and Reagents for Assay

1. NeutrAvidin™ (Pierce Biotechnology, Rockford, IL). 2. GMBS. 4-maleimidobutyric acid N-hydroxysuccinimide ester (Sigma-Aldrich, St. Louis, MO). 3. Buffers. PBST (10 µg/mL in PBS containing 0.05% Tween) and PBSTB (PBST containing 2% BSA). 4. Analyte. Salmonella typhimurium (ATCC 53648) (ATCC, Manassas, VA). 5. Capture antibody. Biotinylated rabbit polyclonal antibody (Biodesign International, Saco, Maine). 6. Detector antibody. Cy5-conjugated goat polyclonal antibody (KPL, Gaithersburg, MD).

392

Li et al.

3. Methods The IWB consists of an illumination optical system, capillary tube waveguide, detection optical system, and data acquisition system (Fig. 2); the details of each component part will be discussed in the following sections. One of the design changes from NRL’s experimental instrument was the elimination of the laser chopper and lock-in amplifier, which were necessary to minimize ambient background light. In contrast, the light-tight construction of Creatv’s instrument eliminates interference from ambient light, obviating the chopper and lock-in amplifier, thereby reducing both the size and the cost. Other design changes are automation of data acquisition and packaging into a portable bench-top format. Laser beam is expanded and collimated to fit the dimension of the capillary tube, so that the capillary tube can be evenly illuminated. Light exiting from the end of the capillary tube is collimated before passing through optical filters, in order to block excitation light and pass emission light. After passing the optical filters, the signal is focused down to the detection area of the PMT. 3.1. Illumination Optical System

The illumination system provides a collimated beam to the analytesensing surface. The system consists of a laser, a cylinder lens to generate a line pattern, and a collimating cylinder lens, as shown in Fig. 2. The features of the system include: ● Stable laser illumination ●

Collimated beam

●

Low noise in the fluorescence emission range

●

Illumination of the whole length of capillary waveguide

Fig. 3. Results of tracepro simulation of NRL’s fused-silica capillary tube (0.70 mm inner

Integrating Waveguide Biosensor

393

The optical system was optimized using Cy5 as the fluorescent dye. The peak absorption wavelength of Cy5 is 647 nm, while the peak emission wavelength is 670 nm, with a range of 650– 750 nm. A diode laser (LaserMax, Rochester, NY) at 635 nm was chosen to provide maximum separation of the excitation and emission wavelengths. Capillary tube is 1.66 mm in diameter and 50 mm long. In order to illuminate the whole tube, the laser beam needs to be about 2 mm × 50 mm in size. The dimensions of the laser beam was expanded to 2 mm × 50 mm to cover the capillary tube waveguide and collimated to provide illumination at a 90° angle to the capillary surface. 3.2. Waveguide

Borosilicate capillary tube is chosen as waveguide and they have the following properties: ● Good transmission for both laser excitation and fluorescent emission wavelengths ●

Low autofluorescence

●

Convenient and inexpensive

Compatible with bioassays In NRL’s initial development, the capillary tubes they used is fused-silica capillary pieces 38 mm long (0.70 mm inner diameter and 0.85 mm outer diameter) with optically polished ends and coated on the outer surface with PTFE were obtained from Polymicro Technologies (Phoenix, AZ). These capillary tubes provided a low limit of detection, but their high cost and fragile nature made them unsuitable for routine use. In order to select the optimum capillary tube waveguide geometry, simulations of the optical properties of various geometries and materials were conducted using the optical ray tracing software TracePro (Lambda Research Corp., Littleton, MA) to perform analysis of illumination of the fluorescent labels on the surface of the capillaries. TracePro uses Monte Carlo simulations to set up the rays for ray tracing and to compute optical flux. TracePro accounts for absorption, reflection, refraction, and scattering of light according to the material and surface properties of the capillary tube. Simulations were performed using incident rays of collimated 635-nm laser light. Initial simulations of the NRL capillary tube waveguide indicated that the NRL’s capillary geometry did not provide uniform illumination of the analyte sensing surface on the inner surface of the capillary. TracePro ray tracing results for the NRL capillaries is shown in Fig. 3a. The trajectories of the incident rays from the excitation laser light (entering from the top of the figure), are bent in the silica capillary (index of refraction = 1.46), and are reflected at the capillary tube surfaces due to change of the refraction index. The capillary is filled with buffer, because ●

394

Li et al.

diameter and 0.85 mm outer diameter) with liquid inside ray trajectories (a) and intensity plot (b).

NRL’s experiments obtained better results with fluid inside the capillary. The rays shown in Fig. 3a are displayed in shades of gray: black rays denote the rays with flux equal or greater than 67% of the incident laser light; medium gray rays denote the rays with flux between 33 and 67% of the incident laser light; while

Integrating Waveguide Biosensor

395

light rays denote the rays with flux less than 33% of the incident laser light. Figure 3a shows that bottom portions of the capillary surface received concentrated excitation light, while some areas to either sides of the capillary received no excitation light. TracePro can also display the results of the excitation laser intensity for a cross section of the capillary tube, where the grayscale from black to white indicates no laser light to brightest laser light, respectively. The dotted white circle in Fig. 3b denotes the location of the inner surface of the capillary tube used by NRL and shows that this capillary geometry does not provide uniform illumination of the analyte sensing surface. We subsequently analyzed several borosilicate glass capillary tubes of various sizes (index of refraction n = 1.52) commercially available from Drummond Scientific. Representative results are shown for the capillary tube with outer diameter of 1.661 mm and inner diameter of 1.226 mm, which was easy to handle and had superior performance (Fig. 4a, b). The TracePro simulation was performed with empty capillaries. Ray tracing trajectories (Fig. 4a) and grayscale cross-sectional view of laser excitation intensity (Fig. 4b) show nearly uniform illumination of the analyte sensing surface, indicated by the dotted white circle. Based on these simulations, this capillary geometry was chosen for the experiments.

Fig. 4. Results of tracepro simulation of Drummond borosilicate capillary tube (1.226 mm inner diameter and 1.661 mm outer diameter) with air inside ray trajectories (a) and intensity plot (b).

396

Li et al.

Fig. 4. (continued)

In order to design the detector optical system, we needed to estimate the properties of the signal that exits from the end of the capillary tube. A different TracePro simulation of the propagation of fluorescent emission signal from Cy5 dyes on the inner surface of the capillary was gathered and guided to the end of tube. The result is presented in Fig. 5 as a candela plot of luminous intensity plot or emission flux distribution vs. angle at the exit of the capillary tube from one dye with 10,000 emission rays. Darker gray indicates more flux. Simulation with large number of dyes would provide uniform angular distribution. The software TracePro, however, can only simulate one dye at a time. Figure 5 indicates that the emission signal exits the tube end in a cone shape with wide angles ranging from 30 to 60°. Thus, it is very important for the detecting optics to collect all the signals efficiently from the large emission angles from the end of the capillary tube. 3.3. Detection Optical System

The purpose of the detector optical system is to: Efficiently collect the emission light from the end of the waveguide

●

●

Filter out any laser light

Integrating Waveguide Biosensor

397

Fig. 5. Candela plot of emission flux distribution from Drummond capillary tube vs. angle at the exit of the capillary tube. Darker gray indicates more flux.

●

Select the relevant portion of the emission light for detection

Focus the transmitted signal to the detector The light from the end of the waveguide includes both excitation and emission light. The excitation light has to be separated out before testing the signal. A 670-nm band-pass filter combined with a 665-nm long-pass filter are used to block excitation light. In order for band-pass filter to work properly, the light has to be collimated. The light from the end of the capillary has large angles (about 60° half angle). An aspheric lens plus a double convex lens are incorporated to collimate the light, as shown in Fig. 2. After passing through the collimating lenses and filters, the excitation light is removed, and only the emission light passes through. Finally, a focusing lens focuses the signal onto the PMT. It is important that the capillary tube should be aligned with the illumination light and the detection system. The alignment of optics is guaranteed by mounting the illumination optics, signal detection optics, and optics holder in a precision machined chamber. ●

398

Li et al.

3.4. Data Acquisition and User Interface

The output form PMT is voltage signal ranging from 0 to 10 V. A data acquisition card is used to digitalize the voltage signal and send it to a computer. The gain of PMT is adjustable. If the signal is too low, the PMT gain can be raised; and if the signal is too high and saturates the PMT, the gain can be lowered. The signal has to be normalized accordingly. LabVIEW software is used to control the IWB and collect the data. The user interface, shown in Fig. 6, consists of three panels: SET-UP PANEL, TEST PANEL, and RESULT PANEL. User fills in the information on the SET-UP PANEL. User starts the test by clicking the TEST button. For each testing, the chart at the right side of the front panel will show all the readings collected by PMT as a function of time. An average value will be calculated and displayed on top of the chart. The test data can be saved in EXCEL format. The file name and path will be shown on the RESULT PANEL. Figure 7 shows a schematic of the interior of the instrument. This model uses a H7827-11 Hamamatsu PMT, a DT9802 data acquisition card (Data Translation, Inc.), and a computercontrolled user interface using LabVIEW™ software (National Instruments Corp.).

3.5. Testing Results

IWB was applied to the detection of Salmonella using sandwich immunoassay. Glass capillary tubes were prepared as previously

Fig. 6. Computer interface of Creatv’s prototype Integrating Waveguide Biosensor.

Integrating Waveguide Biosensor

399

Fig. 7. Schematic of Creatv’s portable prototype IWB using capillary tubes as waveguides for Cy5 fluorescent dye.

described (2). Briefly, capillary tubes were cleaned with methanol/ HCl and sulfuric acid, dried with nitrogen, silanized with 3-mercaptopropyl trimethoxysilane in anhydrous toluene under nitrogen atmosphere, incubated with 4-maleimidobutyric acid N-hydroxysuccinimide ester (GMBS) (Sigma-Aldrich, St. Louis, MO), and then treated with NeutrAvidin (Pierce Biotechnology, Rockford, IL). After conjugation of NeutrAvidin inside the capillary tubes, they were washed with PBST, and stored at 4°C until use. The capillary tubes were incubated with PBSTB containing 2% BSA, and then a biotinylated capture antibody in PBSTB was immobilized inside the capillary. After rinsing with PBST, sample was introduced into the capillary tubes and incubated for 1 h at room temperature. The capillary tubes were subsequently washed with PBSTB, and then filled with Cy5 conjugated detector antibody in PBSTB for 1 h. After removal of the unbound detector antibody and thorough washing with PBST, the capillary tubes were tested using the IWB instrument. A serially diluted culture of S. typhimurium (ATCC strain 53648) in PBST buffer (phosphate buffered saline + Tween + Triton) was tested using the Creatv’s IWB. Biotinylated rabbit polyclonal antibody from Biodesign International (Saco, Maine) was used for capture and Cy5 conjugated goat polyclonal antibody from KPL (Gaithersburg, MD) was used for detection. The fluorescence signals obtained using 40–4 × 106 Salmonella cells per capillary are shown in Fig. 8a. The detection threshold (Mean + 3 SD (Standard Deviation) of zero concentration) was 23.4 mV. These data indicate that as few as 40 Salmonella cells per capillary resulted in a detectable signal of 26.9 mV. Based on an estimated capture efficiency of 18–29% (data not shown), the absolute detection sensitivity of the IWB is approximately ten cells. This is consistent with our previous assay to detect E.

400

Li et al.

Fig. 8. Sandwich immunoassay results for detection of Salmonella using the IWB using Cy5 fluorescent label. a Fluorescence signal response to total cell number of Salmonella in a capillary. Threshold detection limit (mean value plus three times of standard deviation) is 23.4 mV, as indicated by the dashed line. b Linear correlation between log10 of net fluorescence signal and log10 of total cells per capillary.

coli (10). For high cell concentrations, the R2 value is 0.998 (Fig. 8b). When the cells per sample are low, the signal readings are accurate, but cells per sample and percentage of cells captured by the waveguide are susceptible to sampling error. We do not have means to provide accurate count of the number of cells captured on the capillary waveguide. The linear relationship between log10 of fluorescence signal and log10 of input cell concentration (Fig. 8b) allows for quantitative detection for high concentrations and only semiquantitative detection for lower concentrations.

Integrating Waveguide Biosensor

401

4. Notes 1. Collimating the laser light or signal through the interference filter is important to reduce background noise. 2. In capillary tube preparation, all the steps from cleaning until up to the immobilization of NeutrAvidin should be done on the same day. After conjugation of NeutrAvidin inside capillary tubes, the capillary tubes were washed with PBST, and may be stored at 4°C until use. 3. The instrument requires a warm up time of about 20 min. This is for the PMT to stabilize.

Acknowledgments We would like to thank Dr. Francis Ligler (Naval Research Laboratory, Washington DC) for helpful advice and comments. This work was supported by NIH Small Business Innovation Research grants R43 CA094430 and R43 AI052684 and a grant from the Maryland Technology Development Corporation (TEDCO).

References 1. Feldstein, M. J., MacCraith, B. D., Ligler, F. S. (2000) Integrating multi-waveguide sensor, US Patent No. 6,137,117 issued on October 24, 2000 2. Ligler, F. S., Breimer, M., Golden, J. P., Nivens, D. A., Dodson, J. P., Green, T. M., Haders, D. P., Sadik, O. A. (2002) Integrating waveguide biosensor. Anal. Chem. 74, 713–719 3. Anderson, G. P., Golden, J. P., Ligler, F. S. (1994) An evanescent wave biosensor – Part I: Fluorescent signal acquisition from step-etched fiber optic probes. IEEE Trans. Biomed. Eng. 41, 578–584 4. Feldstein, M. J. et al. (1999) J. Biomed. Microdevices 1, 139–153 5. Golden, J. P., Anderson, G. P., Ogert, R. A., Breslin, K. A., Liger, F. S. (1993) Evanescent-wave fiber optic biosensor: challenges for real-world sensing, SPIE Proceedings Series, Vol. 1976 (Meeting 8–9 Sep. 1992, in Boston, MA), pp. 2–8

6. Golden, J. P., Anderson, G. P., Rabbany, S. Y., Ligler, F. S. (1994) An evanescent wave biosensor – Part II: Fluorescent signal acquisition from tapered fiber optic probes. IEEE Trans. Biomed. Eng. 41, 585–591 7. Rowe, C. A., Scruggs, S. B., Feldstein, M. J., Golden, J. P., Ligler, F. S. (1999) An array immunosensor for simultaneous detection of clinical analytes. Anal. Chem. 71, 433–439 8. Rowe-Taitt, C. A., Golden, J. P., Feldstein, M. J., Cras, J. J., Hoffman, K. E., Ligler, F. S. (2000) Array biosensor for detection of biohazards. Biosens. Bioelectron. 14, 785–794 9. Li, S., Zhang, Y., Amstutz, P., Tang, C.-M., Multiplex integrating Waveguide Sensor – Signalyte™-II, Chapter 24 in this book 10. Zhu, P., Shelton, D. R., Karns, J. S., Sundaram, A., Li, S., Amstutz, P., and Tang, C.-M. (2005) Detection of water-borne E. coli O157 using the integrated waveguide biosensor. Biosens. Bioelectron. 21, 678–683

Chapter 23 Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves and In-Fiber Microchannel Sensors Rudi Irawan and Swee Chuan Tjin Summary In life sciences, the problem of very small volume of sample, analytes, and reagents is often faced. Microfluidic technology is ideal for handling costly and difficult-to-obtain samples, analytes, and reagents, because it requires very small volume of samples, in order of µL or even nL. Among many types of optical techniques commonly used for biosensing in microfluidic chip, fluorescence detection technique is the most common. The standard free-space detection techniques used to detect fluorescence emission from microfluidic channel often suffer issues like scattering noise, crosstalks, misalignment, autofluorescence of substrate, and low collection efficiency. This chapter describes two fluorescence detection methods, based on in-fiber microchannels and in-fiber grooves, which can solve those problems, as the techniques integrate the excitation and emission light paths, and the sensing part. Utilizing an optical fiber as a sensing component makes these detection techniques suitable for lab-on-a-chip or µTAS applications. Key words: Optical fiber sensor, In-fiber microchannel, In-fiber grooves, Fluorescence, Microfluidic chip.

1. Introduction Among the optical techniques commonly used for biosensing applications, fluorescence-based sensing is the most common and the most highly developed due to their high sensitivity, versatility, accuracy, and fairly good selectivity. Currently, the fluorescence sensors incorporated with microfluidic devices mainly use free space configurations (1, 2), which can experience issues like noise and crosstalks due to back scattering, as well as low fluorescence collection efficiency (3–5) that can degrade the sensitivity of the Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_23

403

404

Irawan and Tjin

system. The detection system in free space configurations may also collect the autofluorescence signals from microfluidic substrates, which may be relatively high and time dependent (6, 7). The integration of the excitation and emission waveguide, and the sensing part may be able to solve these matters (8, 9). In this chapter, we propose a fiber optic as a waveguide and sensor. A fiber optic sensor can be embedded into a microfluidic chip easily due to its size and flexibility (9, 10), and also be used to transport the excitation light to the fluorophores immobilized inside the microchannel, or to collect the generated fluorescence emission, or both (3, 8, 9, 11). Because the optical fiber can transmit the excitation light to the sample and transmit the emitted fluorescence to the detector, it makes the system simpler and more compact for sample excitation and emission collection than a free space configuration method where the light source and the detector may be situated next to each other (3, 8–10). Optical fiber sensors are usually etched (to thin the cladding), tapered, or side polished to improve light coupling from within the fiber to the external environment, measurand. In the experiments explained in this chapter, we construct microstructures on fibers using a direct write CO2 laser machine instead of etching or polishing method, so that the optical fiber sensor can be constructed cheaply and rapidly. Two types of optical fiber sensor microstructures are explored here, i.e., in-fiber microchannel and in-fiber grooves. The in-fiber microchannel and in-fiber grooves enhance the interaction between the transmitted light through the fiber and the environment, increase the collection efficiency of fluorescence emission to a great degree, and enables the sensor to be able to detect very low light intensities emitted by low concentrations of fluorophores, so that the sensitivity of a fluorescence sensor is improved significantly. The investigations are conducted to find out how the length of the channels and the number of the grooves affect the sensitivity of the sensor as well as the smallest concentration of fluorescein solutions that can be detected. A comparison is also made into how a PMMA optical fiber sensor performs against a sensor fabricated in a silica fiber. Another advantage of in-fiber microchannel structure is that it also simplifies the fabrication of a microfluidic channel for fluid transport, because it can also function as a microfluidic channel. This chapter describes an optical fiber fluorescence sensor based on in-fiber microchannel or in-fiber grooves that can be cheaply and rapidly fabricated using a direct write CO2 laser system, and embedded in a microfluidic card suitable for lab-on-a-chip and µTAS applications. The general schematic diagram of the system, including fluid delivery, microfluidic card, optical fiber sensor, and detection system, is illustrated in Fig. 1.

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

405

2. Materials 2.1. Fabrication of In-Fiber-Microchannel Sensor

1. PMMA optical fiber (PMMA fiber), 1 mm core diameter with 10 µm fluorinated polymer cladding, or Plastic clad silica fiber (PCS fiber), 600 µm/750 µm (core/clad diameter) from Ceram Optec Industries, Inc. 2. Computer-controlled continuous wave CO2 laser direct writing machine, from Epilog Laser (Legend 24TT) as shown in Fig. 2, with two-dimensional robotic arms and z-stage for

Injection valve and sample loop with disposable syringe Buffer (PBS)

Waste

Voltmeter

550 nm Highpass filter PMT

470nm LED Syringe pump

Microfluidic card with in-fiber microstructures

Lens system to focus light and 470nm bandpass filter

Power Supply for PMT

Fig. 1. General schematic diagram of the system.

Fig. 2. Direct write CO2 laser machine used to fabricate microfluidic cards and microstructures in optical fibers.

406

Irawan and Tjin

laser focusing adjustment as sketched in Fig. 3. This CO2 laser machine is originally designed for engraving polymeric materials. 3. Computer-aided design software, such as CorelDRAW. 4. Ultrasonic cleaner (Bransonics, B1510). 5. Flame source, propane flame. 2.2. Fabrication of InFiber-Grooves Sensor

1. PMMA optical fiber (PMMA fiber), 500-µm core diameter with 10-µm fluorinated polymer cladding. 2. Computer-controlled continuous wave CO2 laser direct writing machine (Epilog Laser, Legend 24TT) with two-dimensional robotic arms and z-stage for laser focusing adjustment (see Figs. 2 and 3). 3. Computer-aided design software, such as CorelDRAW. 4. Ultrasonic cleaner (Bransonics, B1510).

2.3. Fabrication of Microfluidic Card with Microchannel

1. PMMA sheet, 1 mm thick. 2. Mylar sheet from Duponts, 0.002˝ thick, manufactured with double-layer 3M501 adhesives. 3. Mylar sheet from Duponts, 0.002˝ thick, with no adhesive. 4. Computer-controlled continuous wave CO2 laser direct writing machine (Epilog Laser, Legend 24TT) with two-dimensional robotic arms (see Figs. 2 and 3). 5. Computer-aided design software, such as CorelDRAW. 6. Nanoport assemblies (PN: N-126H) and drug delivery tubes for inlets and outlets tubing (Capillary PEEK Tubing, PN: 1569) from Upchurch Scientific. 7. Pressure device, a modified mechanical vice grip/clamp (to press and clamp Mylar and PMMA sheets together), and heater or oven. 8. Epoxy Adhesive (PN: N-008) from Upchurch Scientific.

Y-stage Mirror

Y-stage Mirror

Focusing Lens Z-stage Mirror

CO2 laser

Fig. 3. Layout of the direct write CO2 laser machine.

Working piece

X-stage

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

1. Light source, blue LED (NSPB300A, Nichia Chemical Industries) with wavelength centered at 470 nm (FWHM: 28 nm), that covers the excitation wavelength of fluorescein (see Fig. 4). 2. Power supply, current source (Phihong Dual Tracking DC Power Supply, PP-30-6-2B) for the LED. 3. Band-pass filter (SP Corion Filters, 470 ± 5 nm) to limit the spectral width of the excitation light. 4. Microfluidic chip with microchannel and optical fiber sensor embedded into it. 5. High-pass filter (Andover Corp, 550 nm) to filter out the excitation light entering the detector. 6. One-millimeter diameter pin hole made from a black anodized thin aluminium plate, installed at the front side of the detector to prevent the stray light entering the detector. 7. Two 40× objective lenses (infinity corrected objectives) from Olympus to focus the excitation light into the fiber and to collect the fluorescence emission at the other end of the fiber. 8. Convex lens (25 mm focal length) to focus the fluorescence light into the detector. 9. X–Y–Z micropositioners (Newport Corporation) to position the focusing and collection lens systems accordingly. 10. Optical detector, a minicompact module of photon multiplier tube (PMT, Hamamatsu H5784-02), which has spectral response of 300–880 nm. This PMT module, which includes a high-gain built-in amplifier and a built-in DC–DC highvoltage converter, is compact (2 cm × 2 cm × 6 cm) and

Intensity (a.u.)

2.4. Fluoresence Detection System

407

300

350

400

450

500

550

Wavelength (nm)

Fig. 4. Spectrum of the light source.

600

650

700

408

Irawan and Tjin

requires only 15 V DC (see Fig. 5), so that it is very convenient to develop a compact and sensitive detection system. 11. Power supply (Hamamatsu, C7169) for the minicompact module of PMT H5784-02 (see Fig. 6). This power supply is specially designed for the minicompact module of PMT. 12. A digital multimeter, Agilent 34401A Multimeter equipped with automatic data acquisition system and computer. 13. A graphical software Origin for calculating data statistical analysis and plotting graphs. 2.5. Fluorophore Solutions Used for Sensor Tests

1. Fluorescein powders purchased from Sigma-Aldrich. 2. Phosphate-buffer saline (PBS) at pH 7.4.

Fig. 5. PMT module used as the photodetector.

Fig. 6. Power supply for the PMT module.

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

409

3. Magnetic stirrer, micropipette, and beakers. 4. Microfluidic pump system (Precision Syringe Pump from Kloehn Co. Ltd, Las Vegas, Versa Pump 6 P/N 54022, see Fig. 7, and injection valve P/N V-451 and sample loop P/N 54022 from Upchurch Scientific, Oak Harbor, USA, see Figs. 8 and 9). 5. Syringe for the microfluidic pump, purchased from Kloehn Co. Ltd, Las Vegas. 6. Disposable syringe for the injection valve and sample loop.

(A) Buffer (PBS) Injection valve and sample loop Syringe pump Microfluidic card

Waste

(B) Fig. 7. a Precision syringe pump. b Connection of syringe pump, injection valve, sample loop, microfluidic card, and waste.

410

Irawan and Tjin

Injection valve

Sample loop Syringe

Fig. 8. Injection valve and sample loop.

1

1

Pump

6

2

Sample in

Pump

6

2

Sample in

Column

5

3

Sample waste

Column

5

3

Sample waste

4

4

(A)

(B)

Fig. 9. Schematic diagram of injection valve and sample loop as viewed from rotary knob. a Connections when samples are loaded into the sample loop. b Connections when samples are pumped through the microfluidic channel.

3. Methods The methods described below outline (a) fabrications of both types of optical fiber sensors, in-fiber-microchannel type and in-fibergrooves type, (b) fabrication of microfluidic card and embedding the optical fiber sensor in the microfluidic card, and (c) testing the sensors. 3.1. Fabrications of In-Fiber-Microchannel and In-Fiber-Grooves Sensors

Because the PMMA fiber has a very thin cladding, a few microns, removing the cladding before constructing the microchannels or grooves in PMMA fibers is not required. In contrast, removing the jacket and cladding of silica fibers is necessary, so that it is recommended to use type silica fibers that have jackets and claddings that can be removed easily, like the silica fibers suggested here, PCS fibers.

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

411

1. The microchannels and grooves that are going to be fabricated in the fibers are drawn precisely using CorelDRAW, then the drawings are transferred to the CO2 laser direct writing machine. In our tests, we made various lengths of microchannels, 5, 10, 15, 20, and 25 mm, and the width of all chanels is 100 µm. The numbers of grooves were made 5, 10, 20, 30, 40, and 50. Examples of CorelDRAW drawings for constructions of in-fiber microchannel and in-fiber grooves are shown in Fig. 10. 2. The cladding and jacket of silica fibers are removed using flame, then they are cleaned using Kimwipes and ethanol. 3. Before the fabrications of microstructures, the fibers are cleaned thorougly to get rid of from dust and finger prints that may stick on the surfaces of fibers. PMMA fibers should not be cleaned using acetone or ethanol. Wiping the PMMA fibers using wet Kimwipes should be enough. 4. The CO2 laser machine stage has no readily built optical fiber holder. We made the optical fiber holder from 2-mm thick of Acrylic sheet. A V-groove was constructed in the Acrylic sheet, and clamps were placed at two ends of the V-groove. Hence, for fabrications of in-fiber-microchannel and in-fibergrooves sensors, the optical fibers were placed and clamped inside the V-groove of Acrylic sheet to hold the optical fibers securely and firmly. 5. The setting of the CO2 laser machine, particularly the laser power and scanning speed, must be adjusted before printing microchannels or grooves on the fibers. The laser setting needs to be optimized and it depends on the characteristics of material and the quality of the CO2 laser system used. Ideally, low laser power should be used, but the laser beam should be often multiply scanned over the same area of the optical fiber to create well-defined structures. The microstructure depth increases linearly with the laser power set and the number of laser passes (12). It was found that because the glass transition temperature of silica is higher than PMMA, to obtain the same depth of structures, the silica core fibers require higher laser

Direction of laser scan

Written pattern

(A)

Direction of laser scan Written pattern

(B)

Fig. 10. Examples of CorelDRAW drawings for constructions of microstructures in optical fibers. a For in-fiber microchannel. b For in-fiber grooves.

412

Irawan and Tjin

power and/or a greater number of laser passes than the PMMA fibers do. In our experiments, to get approximately 200-µm depth microchannels in the PCS fiber, the power of CO2 laser machine was set at 32 W, 15 laser passes, and 3.25 mm/s scanning speed. On the other hand, to get the same depth in the PMMA fibers, the power of CO2 laser machine was set at 20 W and single laser pass. However, the laser setting may vary from machine to machine and from time to time. There are many factors affecting the laser power hitting the optical fiber, such as the conditions and the qualities of the lenses and the mirrors installed in the laser machine. Before the microstructure fabrications are started, the vertical position of optical fiber is adjusted to the position of the focal point of laser beam by adjusting the X-stage of CO2 laser machine. 6. The ablation process on the PMMA fibers forms a thin layer of white residue in the microstructures. This white layer can be partly eliminated by annealing the microstructures of fibers at 80°C for an hour. On the other hand, the ablation process on the silica core fibers still maintains the microstructures with clear surfaces. 7. Any debris or particles left by microfabrication processes in the fibers are cleaned by sonicating the microstructures in water. 8. The sizes of in-fiber-microchannel and in-fiber-grooves were estimated using Scanning Electron Microscope and/or surface profiler, actual SEM images of microstructures constructed in optical fibers using a direct write CO2 laser machine shown in Fig. 11. 3.2. Fabrications of Microfluidic Cards with Microchannels and Embedding the Optical Fiber Sensors in the Cards

1. Thin Mylars manufactured with double-layer adhesive are placed at the top and bottom of 1 mm PMMA (see Fig. 12a). 2. The drawings of microfluidic card and microchannel that are going to be fabricated on PMMA and Mylar sheets are precisely drawn using CorelDRAW, then the drawings are transferred to the CO2 laser direct writing machine. The width of microchannels should not be bigger than the diameters of the optical fibers used for sensor and the length of the microchannels is at least the same as that of the sensing area of the optical fibers. Examples of CorelDRAW drawings for fabrication of microfluidic card are sketched in Fig. 13. 3. The setting of the CO2 laser machine, particularly the laser power and scanning speed, must be adjusted before printing or cutting microfluidic cards and microchannels on PMMA and Mylar sheets. The laser setting needs to be optimized depending on the characteristics of material and laser machine. Since the fabrications of microfluidic cards and microchannels are based on cutting the polymers, their qualities are less sensitive to the setting of the CO2 laser machine as compared to

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

413

Fig. 11. SEM images of microstructures constructed in optical fibers using a direct write CO2 laser machine. a 600-µm core diameter PCS fiber with an in-fiber microchannel. b 1-mm core diameter PMMA fiber with an in-fiber microchannel. c Cross section of microchannel in 1 mm diameter PMMA fiber. d 500-µm core diameter PMMA fiber with in-fiber grooves.

Microfluidic Channel

PMMA

(A)

Mylar with double adhesives

(B)

Microfluidic channel Optical Fiber Sensor embedded into microfluidic channel

Mylar

(C) Fig. 12. Cross section of microfluidic cards. a Microfluidic card substrate before writing microfluidic channel. b After writing microfluidic channel. c After the optical fiber sensor imbedded in the microfluidic card.

the fabrications of the microstructures of optical fiber sensors. However, low scanning speed results in smoother microchannel wall, and too much laser power can cause the microchannel size bigger than the specified size due to the burning effect. After cutting, the cross section of microfluidic card and microchannel are as depicted in Fig. 12b.

414

Irawan and Tjin

(A)

(B)

(C) Fig. 13. Examples of CorelDRAW drawings for microfluidic card fabrications. a Middle layer. b Top layer. c Bottom layer.

4. The optical fiber sensor is placed at the bottom of the microchannel. The gaps between the optical fiber sensor and the wall of microchannel, if any, must be eliminated to avoid leak. It can be achieved by fitting the optial fiber sensor against the microchannel walls tightly (see Figs. 12c and 14a). 5. Then, the microchannels are closed using nonadhesive Mylar sheets pressed against the adhesive layer of Mylar attached to PMMA. To get good seal, they are pressed between two plates at temperature 40°C for few hours (see Figs. 12c and 15). 6. The inlet and outlet of microchannels are made at the top Mylar sheet (see Figs. 15 and 16). The inlet port is connected to a syringe pump through a Nanoport assembly and drug delivery tube, from Upchurch Scientific, Inc., Washington and the outlet is connected to a waste container. To prevent any leak, the nanoport assemblies are glued to the inlet and the outlet of microchannels using Epoxy Adhesive and the drug delivery tubes are tightened to the nanoport assemblies by finger tightening. 3.3. Testing the Sensors

1. The optical fiber sensor under the test, installed inside a microchannel of a microfluidic card, is placed in the experimental setup as shown in Fig. 14b. 2. The excitation beam, from a blue LED, is filtered by a bandpass interference filter (470 ± 5 nm), then it is focused into the one end of the optical fiber sensor by an objective lens. The circuit of the LED is shown in Fig. 17, the power supply, Phihong Dual DC Power Supply, is operated at constant current mode.

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

415

Fluorescence Light

Microchannel Fiber in microfluidic card Sensing part of fiber

Excitation Light

(A)

LSFF

Microfluidic card with in-fiber microstructures

550nm high pass filter PMT

470nm LED

LSCF 5.0V ?

V V

OFF

A A V?

A

COM

(B) Fig. 14. a Illustration of microfluidic card, microchannel, and embedded optical fiber sensor. b Experimental setup. LSFF: lens system to focus the excitation light into the fiber. It consists of a 470-nm LED, an objective lens, and 470-nm band-pass filter. The 470-nm LED is attached to the objective lens and the 470-nm band-pass filter is placed inside the objective lens. The distance between the objective lens and the fiber is adjusted using a micropositioner, so that the excitation light is coupled into the fiber efficiently. LSCF: lens system to couple the fluorescence emission from the fiber into the detector. It consists of an objective lens, a convex lens, and 550-nm high-pass filter. High-pass filter is attached to the PMT detector, and the convex lens is 25 mm from the PMT. The distance between the objective lens and the fiber is adjusted using a micropositioner, so that it reaches maximum fluorescence light coupling.

Inlet

A

Microchannel

Outlet

Groove

Optical fiber and grooves A (A) Microchannel

In-fiber channel

A

Optical fiber and channel A

(B)

Fig. 15. Optical fiber sensor embedded in a microfluidic channel. a Side-view and crosssection in-fiber grooves embedded in a microfluidic channel. b Side-view and crosssection in-fiber channel embedded in a microfluidic channel.

416

Irawan and Tjin

(A)

Drug delivery tubes

(B)

Nanoport assemblies Optical fiber Microfluidic channel and embedded optical fiber sensor

Fig. 16. Microfluidic card devices. a The substrate is clear PMMA. b The substrate is black PMMA.

Fig. 17. Circuit of the LED. The power supply, Phihong Dual DC Power Supply, is operated at constant current mode.

20 mA

LED

3. A collection lens system, an objective lens, to collect the fluorescence emission transmitted through the optical fiber sensor is installed at the other end of the fiber. The output of the collection lens system is focused into the detector, a minicompact PMT module, by a convex lens. Between the detector and the convex lens, a high-pass filter (550 nm) is installed to avoid the excitation beam from entering the photodetector that can cause noise of detection. A 1-mm diameter pin hole is also placed at the detector to prevent the stray light entering the detector. Since the photodetector used is a compact PMT module, which includes a high-gain built-in amplifier, the outputs of the photodetector are readily measured using a digital multimeter connected to a computer through a data acquisition system. This compact PMT module is also accompanied by a specifically dedicated power supply, so that the electrical connections of the detection system are simple (see Fig. 18). The digital multimeter used is equipped with built-in and convenient data acquisition interface and software. What we need to do is just to connect the multimeter to the one of the external ports of a computer and run the provided software. The data will be downloaded and stored automatically into a designated file. The time interval and frequency of data downloading can be determined accordingly. The details of the computer interface and the algorithms of the digital multimeter data acquisition system can be found in the manual of Agilent 34401A multimeter. 4. Various concentrations of fluorescein solutions are prepared using fluorescein powder and phosphate-buffer saline (PBS)

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

417

Cables to connect the PMT and power supply PMT

Cable to connect the output of PMT and a digital multimeter

Power Supply

Fig. 18. Electrical connections of the PMT module and power supply.

to stabilize its pH at 7.4. The solutions are stored inside syringes and protected from light exposure using a black cover to avoid photobleaching before measurements. The syringes are placed into the injection valve and the solutions are loaded into the sample loop (see Figs. 8 and 9). Then, the microfluidic syringe pump pumps PBS to drive different concentrations of the fluorescein solutions previously loaded inside the sample loop into the microchannels (see Fig. 7). This method is useful to avoid cross contaminations of sample. The connection of microfluidic syringe pump, injection valve, sample loop, and microfluidic card can be illustrated as shown in Figs. 1 and 7b. 5. The sensitivity of the optical fiber sensors is evaluated by filling the microchannels with known concentrations of fluorescein solutions. The tests are started from the lowest concentration of prepared fluorescein solutions, and then progress to consecutively higher concentrations. During the tests, the samples are continuously pumped through the microchannel at a constant speed, so that the photobleaching effect is minimized. 6. Between the measurements, the microchannel and the optical fiber sensors are cleaned thoroughly by flushing with buffer. 7. The fluorescence intensities of various fluorescein concentrations are corrected with respect to the background fluorescence, the fluorescence signals when the microfluidic channel is loaded by buffer only. Therefore, all the fluorescence intensities shown in Figs. 19–23 are corrected signals, which are

Irawan and Tjin 10000

Fluorescence Intensity (a.u)

9500 9000 8500 8000 7500 7000 6500 6000 5500 5000 0

5

10

15

20 25 30 35 40 Number of grooves

45

50

55

60

Fig. 19. Effect of the numbers of grooves to the sensitivity of sensor.

65000 60000

Fluorescence intensity (a.u)

418

55000 50000 45000 40000 35000 30000 25000 20000 15000 10000 5000 0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Fluorescein concentration (µg / l)

Fig. 20. Fluorescein concentrations vs. fluorescence intensities. The fiber is made from PMMA and the number of grooves is 20.

the background fluorescence signals when the buffer only is inside the microchannels are subtracted from the fluorescence signals measured when the fluorescein solutions are inside the microchannels. 8. The data are taken at least ten times. The means, the errors, and the graphs are calculated and plotted using a graphical

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

419

4400

Fluorescence Intensity (a.u)

4200 4000 3800 3600 3400 3200 3000 2800 2600 0.0

0.5

1.0 1.5 Channel length (cm)

2.0

2.5

Fluorescence intensity (a.u.)

Fig. 21. Effect of microchannel lengths to the sensitivity of sensor.

15000 14000 13000 12000 11000 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 0.0

0.1

0.2

0.3 0.4 0.5 0.6 0.7 0.8 Fluorescein concentration (µg / l)

0.9

1.0

Fig. 22. Fluorescein concentrations vs. fluorescence intensities for in-fiber microchannel sensor made from PCS fiber.

software, Origin. The results show the sensors have good signal to noise ratio. For example, the optical fiber fluorescence sensor based on in-fiber microchannel is able to detect 0.005 µg/L of fluorescein in PBS solution with the signal to noise ratio better than 5.

Irawan and Tjin

Fluorescence intensity (a.u.)

420

15000 14000 13000 12000 11000 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Fluorescein concentration (µg /l)

Fig. 23. Fluorescein concentrations vs. fluorescence intensities for in-fiber microchannel sensor made from PMMA fiber.

4. Notes 1. The length of in-fiber microchannel or number of in-fiber grooves in the optical fiber sensor determines the sensitivity of the sensor. Therefore, it is important to check what is the optimum length of in-fiber microchannel or optimum number of in-fiber grooves for each designed optical fiber sensor. 2. The procedures of sensor tests are the same for both optical fiber sensors based on in-fiber microchannel and in-fiber grooves. 3. Fluorescein solutions diluted in PBS are used for sensor sensitivity tests and Microfluidic card substrates, PMMA and Mylar, may have auto background fluorescence. Therefore, all the fluorescence intensities shown in Figs. 19–23 are previously corrected by the background fluorescence, the fluorescence signal when only buffer is inside the channel. The readings are taken few times, at least ten times, to get statistical average data and error bars. 4. Fluorescein solutions are very sensitive to strong light exposure. Exposing fluorescein solutions to strong intensity of light can cause photobleaching. Therefore, before measurements, fluorescein solutions must be protected against light exposure. Fluorescence intensity and spectrum of fluorescein are also affected by pH. Hence, fluorescein powder must be

Detection of Fluorescence Generated in Microfluidic Channel Using In-Fiber Grooves

421

diluted using buffer with consistent pH. Fluorescein solutions are slightly hazardous in case of skin contact (irritant), eye contact, and ingestion. Therefore, it is recommended to use glove, safety glass, and laboratory coat while handling fluorescein solutions, and waste must be disposed according to local environmental control regulations. 5. The microfluidic pump causes vibrations, and the alignment of optical system, light source, detector, and microfluidic card is rather sensitive to vibrations. Therefore, it is recommended that the microfluidic pump and the sensing system are placed on different tables or placed on vibration absorbers. 6. The PMT used is a compact PMT module, which includes a high-gain built-in amplifier and a built-in DC–DC highvoltage converter, and is equipped with a special power supply. The outputs of PMT are readily measured using a digital multimeter. Cable connections from the PMT to the power supply and a digital multimeter can be refered to the user manual of H5784-02 PMT from Hamamatsu. 7. Mylar sheets from Duponts can be purchased with adhesive layers or without adhesive layer depending on our needs or orders.

Acknowledgments We thank to the Biomedical Research Council of Singapore for the financial support under the Singapore-University of Washington Alliance Programme, and Republic Polytechnic, Singapore for providing us facilities and supports to write the manuscript. We also thank to Chia Meng Tay for his helps to fabricate microstructures in the fibers.

References 1 Irawan, R., Tjin, S. C., and Fu, C. Y. (2005) Integration of a fluorescence detection system and a laminate-based disposable microfluidic chip. Microwave and Optical Technology Letters 45(5), 456–460 2 Yao, B., Luo, G., Wang, L., Gao, Y., Lei, G., Ren, K., Chen, L., Wang, Y., Hu, Y., and Qiu, Y. (2005) A microfluidic device using a green organic light emitting diode as an integrated excitation source. Lab on a chip 5, 1041–1047 3 Irawan, R., Tjin, S. C., Zhang, D., and Fang, X.-Q. (2005) Fluorescence detection system

and laminate-based disposable microfluidic chip. Chinese Optics Letters 3, S173–175 4 Irawan, R., Tjin, S. C., Yager, P., and Zhang, D. (2005) Cross-talk problem on a fluorescence multi-channel microfluidic chip system. Biomedical Microdevices 7(3), 205–211 5 Van Orden, A., Machara, N. P., Goodwin, P. M., and Keller, R. A. (1998) Single-molecule identification in flowing sample streams by fluorescence burst size and intraburst fluorescence decay rate. Analytical Chemistry 70, 1444–1451

422

Irawan and Tjin

6. Piruska, A., Nikcevic, I., Lee, S. H., Ahn, C., Heineman, W. R., Limbach, P. A., and Seliskar, C. J. (2005) The autofluorescence of plastic materials and chips measured under laser irradiation. Lab on a Chip 5, 1348–1354 7. Hawkins, K. R. and Yager, P. (2003) Nonlinear decrease of background fluorescence in polymer thin-films – a survey of materials and how they can complicate fluorescence detection in µTAS. Lab on a Chip 3, 248–252 8. Hubner, J., Mogensen, K. B., Jorgensen, A. M., Friis, P., Telleman, P., and Kutter, J. P. (2001) Integrated optical measurement system for fluorescence spectroscopy in microfluidic channels. Review of Scientific Instruments 72(1), 229–233

9. Irawan, R., Tay, C. M., Tjin, S. W., and Fu, C. Y. (2006) Compact fluorescence detection using in-fiber microchannels – its potential for lab-on-a-chip applications. Lab on a Chip 6, 1095–1098 10. Polynkin, P., Polynkin, A., Peyghambarian, N., and Mansuripur, M. (2005) Evanescent fieldbased optical fiber sensing device for measuring the refractive index of liquids in microfluidic channels. Optics Letters 30(11), 1273–1275 11. Tjin, S. C. and Irawan, R. (2006) Microfluidic immunoassay device. Patent Reference No. PCT/SG2006/000044. Pending patent 12. Klank, H., Kutter, J. P., and Geschke, O. (2002) CO2-laser micromachining and backend processing for rapid production of PMMA-based microfluidic systems. Lab on a Chip 2, 242–246

Chapter 24 Multiplex Integrating Waveguide Sensor: Signalyte™-II Shuhong Li, Yunqi Zhang, Platte Amstutz III, and Cha-Mei Tang Summary A platform to detect multiplex fluorescent labels was developed based on liquid phase implementation of the Integrating Waveguide Sensor detection principles. The liquid sample is held in a capillary cuvette with a lens at one end. The excitation light incident on the cuvette at 90° angle. The emitted fluorescence is efficiently gathered and propagated to the end of the waveguide cuvette, exiting via the lens to the detector. The capillary cuvette acts as a waveguide to efficiently gather the emission signal, providing high detection sensitivity for small sample sizes. Excitation sources ranging from 470 to 635 nm are four high-powered LEDs, allowing for multiplex fluorescence assays and a spectrometer is used to collect the signal from 390 to 790 nm. The cuvette can hold 1–35 µL samples. This technology can be used for a wide variety of assays and detection needs, such as FRET, end point PCR reading, immunoassays, chemiluminescence detection, multiplex quantum dots assays, polarization assays, etc. Key words: Biosensor, Spectrofluorometer, Fluorometer, Integrating waveguide sensor.

1. Introduction The Integrating Waveguide Biosensor (IWB) technology, which detects bound analytes on the inner surface of a capillary tube, was shown to be a very sensitive detection platform in another chapter in this book (1–3). Many assays, however, are more naturally performed in solution and this chapter describes an implementation of the Integrating Waveguide Sensor for liquid-phase assays, called Signalyte™-II. The basic principle of the liquid-phase integrating waveguide sensor (4) is shown in Fig. 1. Detection and quantitation are achieved by illuminating the cuvette (i.e., optical waveguide) at a 90° angle relative to the length of the waveguide and subsequent Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_24

423

424

Li et al.

Fig. 1. Schematic showing of the principle of liquid-phase integrating waveguide sensor detection, where the fluorescent labels are in the sample solution inside a cuvette. The excitation light impinges perpendicular to the cuvette. The cuvette and the sample together acts as a waveguide sending the signal out of the end of the cuvette. The lens at the end of the cuvette focuses the emission signal.

collection of emitted fluorescence from the end of the waveguide. There is a half-ball lens at the end of the cuvette, which focuses the signal down to the detection optical system. The principle of the detection is similar to the solid-phase integrating waveguide sensor described by Li et al. in Chap. 22 (3). Signalyte™-II provides high sensitivity for small sample sizes. Noise from excitation light is reduced because the excitation light is perpendicular to the detector. Signal from fluorescence is increased because the cuvette and the sample inside together act as a waveguide for the fluorescent signal from the whole sample to the end and exiting via the half-ball lens. Signalyte™-II provides four high-power LEDs for illumination. Wavelength of the LED can be varied. In this paper, they are 470, 530, 590 and

Multiplex Integrating Waveguide Sensor: Signalyte™-II

425

635 nm. Four excitation sources enable multiplex assays utilizing a wide range of fluorescent dyes. Creatv also have a single excitation source model called Signalyte™. Sensitivity results for Cy5 and FITC fluorescent labels using Signalyte™-II are described below. Signalyte™-II enables assays that the fluorescent labels are in the liquid such as FRET-based assays, chemiluminescence detection, end point PCR, and DNA and RNA detections without using thermal cycling amplification, immunoassays, multiplex quantum dots assays, polarization assays, etc.

2. Materials 2.1. Instrument Components

1. Excitation light source: LEDs of 470, 530, 590, and 635 nm (Lumileds, http://www.lumiledsfuture.com). 2. Band-pass filters for cleaning up LED light sources: 3. Dichroic filters: Chroma, Rockingham, VT. 4. LED collimation lens: aspheric lens. 5. Lens for illuminating cuvette: cylinder lens. 6. Lens for signal detecting: aspheric lens. 7. Long-pass filters: color glass filters. 8. Filter wheel (Thorlabs, Newton, NJ). 9. BTC111 Spectrometer (B&W Tek, Newark, DE). 10. SmartMotor: SM2315D (ONExia, West Chester, PA)

2.2. Software

1. Optical ray tracing: TracePro software (Lambda Research Corp., Littleton, MA). 2. Instrument control and user interface: LabVIEW (National Instruments, Austin, TX).

2.3. Cuvette, Chemicals, and Reagents

1. Cuvette (Roche Diagnostics, Indianapolis, IN). 2. Cy5 labeling kit PA35000 (GE Healthcare). 3. FITC dye: AC11925 from (Fisher, Pittsburgh, PA).

3. Methods Signalyte™-II consists of LED illumination system, cuvette, spectrometer detection system, sample movement system, and computer control system (Fig. 2); the details of each component part will be discussed in the following sessions.

426

Li et al.

Fig. 2. Schematic of a liquid-phase Integrating Waveguide Sensor, Signalyte™-II, with four excitation light sources, spectrometer as detector and computer control.

3.1. LED Illumination System

Currently, the illumination system uses four LEDs with central wavelength of 470 nm, 530 nm, 590 nm, and 635 nm respectively. The four LEDs provide the excitation for the most commonly used fluorescence dyes in market. Different choices of LED wavelength can be easily implemented to suite the user need. LED has more wavelength options and costs less than laser. High-power 3-W LED was chosen to compensate the fact that LED is less powerful in a narrow band of wavelength than laser. A band-pass filter is needed to clean up the spectrum in order to reduce noise in the signal wavelength regime. The beam from LED was collimated and expanded to about 2 mm × 30 mm using a collimating lens and a cylindrical lens in order to illuminate the cuvette. Compare to laser, LED has large emission angle. In order to collimate as much light as possible, a collimating lens with short focal distance and large NA is used. Each LED has its own collimating lens and band-path filter. All four LEDs and their optics are mounted onto a common housing so that the four LEDs share one light path to illuminate the cuvette. There are three dichroic filters in the housing. The dichroic filters are arranged in a way so that the light from each LED can go to the cuvette. A cylinder lens is attached at the end of the housing in order to focus the beam into 2 mm thick in one direction.

3.2. Simulation of Cuvette

To help selecting the cuvette, we conducted simulations of various geometries of cuvette using the optical ray tracing software TracePro (Lambda Research Corp., Littleton, MA).

Multiplex Integrating Waveguide Sensor: Signalyte™-II

427

The cuvette is equivalent of a regular capillary tube and a lens at the end, which focuses the light from the capillary waveguide and seals the cuvette. A variety of lens shapes was simulated, including: 1. Flat end 2. Round end 3. A half-ball lens at the end 4. A lens that is thinner than half-ball lens 5. A lens that is thicker than half-ball lens TracePro is used to simulate the cuvette. TracePro uses Monte Carlo simulations to set up the rays for ray tracing and to compute optical flux. TracePro accounts for absorption, reflection, refraction, and scattering of light according to the material and the surface properties of the capillary tube. To analysis the illumination of the flurescent labels in the liquid sample inside the cuvette, simulations were performed using incident rays of collimated 635-nm light and the result is shown in Fig. 3. In Fig. 3a, the trajectories of the incident rays from the excitation light (entering from the top of the figure), are bent by the borosilicate glass cuvette (index of refraction of n = 1.52), and reflected at the capillary tube surfaces due to change of index of refraction. The cuvette is filled with buffer. The rays shown in Fig. 3a are displayed in shades of gray: black rays denote the rays with flux equal or greater than 67% of the incident light; medium gray rays denote the rays with flux between 33 and 67% of the incident light; while light rays denote the rays with flux less than 33% of the incident light. Figure 3b is the irradiance map on the inner surface of the glass cuvette, which shows that the sample inside the cuvette is fairly evenly illuminated. We simulate the transmission of fluorescent signal in cuvette. The cuvette is filled with solution with fluorescent dyes inside. The fluorescent dyes are considered point source in TracePro simulation. Light emits from the fluorescent dyes is guided by the cuvette to the lens at its end, and then focused. The size of the focal point and angular distribution are compared. Among the different lens shapes, the cuvette with a half-ball lens provides the best result with the smallest focusing point as well as a moderate angular distribution. Figure 4 shows the power density at the focal plane. The cuvette has a small focal point of less than 2 mm in diameter. The cuvette with a half-ball lens is commercially available. A 35-µL cuvette made with borosilicate glass with a 1.2 mm ID (inner diameter) and 1.55 mm OD (outer diameter) is available through Roche Diagnostics (Indianapolis, IN) and the borosilicate glass cuvette is produced by Drummond Scientific Company (Broomall, PA).

428

Li et al.

Fig. 3. a Results of TracePro ray tracing simulation of excitation light trajectories on borosilicate cuvettes (1.2 mm inner diameter and 1.55 mm outer diameter) with water inside. The liquid sample is well illuminated. b Irradiance map of cuvette inner surface generated from Fig. 3a.

Multiplex Integrating Waveguide Sensor: Signalyte™-II

429

Fig. 4. Results of TracePro simulation intensity plot at the focal spot (0.3 mm from the end of the half-ball lens) from the 35-µL borosilicate cuvettes with water inside.

3.3. Signal Detection System

Signal detection system consists of a set of lenses that focus the signal from the cuvette, through a long-pass color glass filter to the spectrometer. A long-pass filter for each excitation LED wavelength is needed to block the excitation LED light, so that only the emission signal is detected by the spectrometer. All four long-pass filters are incorporated into a filter wheel and controlled by computer. The input to the spectrometer is a 0.29 NA slit. The spectrometer is sensitive from 390 to 790 nm with a cooled CCD inside. To achieve high sensitivity in Signalyte™-II, noise has to be reduced. Noise can be classified into two sources: ● Background noise from the instrument Nonspecific binding of fluorescent dyes in sample The background noise exists even when there is no fluorescence in cuvette. The background noise is from the instrument itself. It is caused by LED, lenses and filter, and electronic noise from spectrometer. In illumination system, LED has wide spectrum that extends to signal range, although band-pass filter is used to cut off the unwanted spectrum, because of the imperfection of optics, there are still some noise coming through the optics and goes to the detection system. Long-pass filter in detection

●

430

Li et al.

system blocks most of the excitation light, and the part that leaks into spectrometer forms the background noise. Electronic noise of spectrometer also contributes to background noise. In order to minimize background noise, collimation of LED source, choice of filters, and optimizations of lenses are all important. Electronic noise is reduced by choosing spectrometer with cooled CCD. Nonspecific binding comes from fluorescent dyes that are not captured by purpose. This can be reduced by optimizing the assay. 3.4. Sample Movement

Samples are loaded into a holder that can hold nine cuvettes. For the eight cuvettes on the right side of the holder, the distance between the adjacent cuvettes is compatible with a multichannel pipetter. After finishing the assay in a 96 well plate, samples can be transferred to the holder using a multichannel pipetter. The cuvette at the “0” position on the left is for reference sample (for example, negative control). In order to differentiate the reference sample from other testing samples, the distance between the eighth cuvette and the ninth cuvette is larger than the distance between other adjacent cuvettes. The illumination and signal detection systems are at fixed positions. A Smartmotor™ is incorporated to move each cuvette to the testing position. After a sample is moved to the testing position, a LED is turned on to illuminate the cuvette, and the signal is detected by a spectrometer. It is important that the cuvette should be aligned with the illumination light, and the lens at the bottom of the cuvette should be aligned with the detection system. The alignment of optics is guaranteed by mounting the illumination optics, signal detection optics, and tube holder in a main mounting piece machined with precision. Sample holder movements are programmed and calibrated so that every cuvette is moved to the same testing position.

3.5. Computer Control System

●

LED control

●

Motion control

●

Spectrometer control

Filter wheel control When Signalyte™-II runs a test, one tube is moved to the testing position, stopped; one LED is turned on, spectrometer takes the data, and the data are transferred to computer, processed, and displayed on the screen. A computer control system programmed in LabVIEW handles LED on/off, sample motion control, spectrometer data acquisition, filter wheel motion control, data processing and display, etc. All four LEDs are connected to a digital I/O control device. The digital I/O control device, motor, spectrometer, and filter wheel are connected to computer through a USB HUB. They are converted to USB port first if they are not originally USB. ●

Multiplex Integrating Waveguide Sensor: Signalyte™-II

431

Fig. 5. User interface print screen. The full spectrum is the plot at the bottom. The bar chart at the upper left shows the peak value chosen by the user and the spectrum on the top right shows the user-selected section.

A user interface is also needed for the user to interact with the instrument. Figure 5 shows the user interface. The user can input a note for each sample, choose the exposure time, select the spectrum range to display, and pick the peak emission wavelength to generate the result. Up to eight samples can be tested with one load. A zero concentration sample cuvette will also be tested to provide the background reference. The user can enter notes for each tube. A motion control system is incorporated to move each cuvette to the testing position. A progress bar shows the number of the sample that is under testing. After all the eight samples are tested, a full spectrum of all the samples is shown on the bottom of panel. A bar chart that shows the signal at the emission peak designated by the user, and a subset of the spectra selected by the user is shown on the top right panel. Both the bar chart and the subset of the spectrum can be saved as EXCEL and/or JPEG files. 3.6. Results

A model of the exterior of the Signalyte™-II is shown in Fig. 6. On the upper right side, there is a door that can be opened when user

432

Li et al.

Fig. 6. Photography of a Signalyte™-II instrument.

loads samples. Eight samples, plus one reference, can be loaded in the holder and can be tested in about 1 min. In order to test the limit of detection for different LED light source, we selected two typical fluorescence dyes: Cy5 was tested by using red LED (635 nm) and FITC was tested by using blue LED (470 nm). Cy5 dye was diluted to serial dilutions using PBS buffer. Background noise is obtained by testing PBS buffer when illuminated by red LED. This background noise is subtracted from the signals. The results of the spectrum are shown in Fig. 7. X-axis is wavelength in nanometer unit. Y-axis is relative intensity without unit. The limit of detection of Cy5 is 2.5 pM (0.088 fmol in 35 µL). FITC dye was diluted to serial dilutions using PBS buffer. Background noise is obtained by testing PBS buffer when illuminated by blue LED. This background noise is subtracted from the signals. The results of FITC testing are shown in Fig. 8. The limit of detection of FITC is 25 pM (0.88 fmol in 35 µL). The difference in limit of detection for Cy5 and FITC is mainly because the difference in excitation efficiency and difference of dye’s photon efficiency. The excitation efficiency for Cy5 excited at 635 nm is only 65%, while the excitation efficiency for FITC excited at 470 nm is only 45%. On the other hand, Cy5 has higher photon efficiency than FITC that means with the same excitation power, Cy5 convert more energy into emission.

Multiplex Integrating Waveguide Sensor: Signalyte™-II

433

Fig. 7. Testing result with the background removed for Cy5 using 635-nm LED showing the limit of detection of 2.5 pM for a 35-µL sample size.

Fig. 8. Testing result with the background removed for FITC using 470-nm LED showing the limit of detection of 25 pM for a 35-µL sample size.

If a user wants to test a sample with unknown concentration, a standard curve needs to be established first. A standard curve is a plot of the signal vs. concentration. First, a serial dilution of assay is prepared with the same protocol of the testing sample, except with known concentrations. Second, the serial dilution samples are tested on Signalyte™-II. Then, after subtracting background from all the signals, a plot of signal vs. concentration is generated. That is usually a linear curve. The relationship between concentration and signal can be expressed in a formula. Then, the unknown sample is tested. The signal for the unknown sample is obtained by subtracting background. Concentration of the

434

Li et al.

unknown sample can then be calculated according to the standard curve. The standard curve only needs to be tested once for the same assay. The generation of the curve and the calculation of concentration can also be programmed in the instrument.

4. Notes 1. IWB was originally developed by Ligler et al. at the Naval Research Laboratory (NRL) in Washington, DC. Creatv Microtech licensed the technology and developed Signalyte™-II to use the technology on liquid phase platform. 2. High-power LED generates heat during operation. Temperature rise will cause color shift, which means change in emission wavelength. To avoid this, a heat sink is needed to be mounted on the back of LED. 3. System alignment is important for obtaining high sensitivity. Cuvette is only 1.55 mm outer diameter, and it needs to be aligned with both the excitation and detection optics. When user loads the sample into the holder, make sure every cuvette is in proper position, not tilted. 4. Signalyte™-II requires a warm up time of about 20 min. This is for the CCD in spectrometer to be fully cooled in order to get a low and stable background.

References 1. Feldstein, M. J., MacCraith, B. D., Ligler, F. S., US Patent No. 6,137,117 issued on 10/24/2000 2. Ligler, F. S., Breimer, M., Golden, J. P., Nivens, D. A., Dodson, J. P., Green, T. M., Haders, D. P., Sadik, O. A., Integrating waveguide sensor. Anal. Chem. 74, 713–719, 2002

3. Li, S., Amstutz, P., Tang, C. M., Hang, J., Zhu, P., Zhang, Y., Shelton, D. R., and Karns, J. S., Integrating Waveguide Sensor, Chapter 22 in this book 4. Tang, C.-M. and Amstutz, P., III, Sensitive Emission Light Gathering and Detection System, US Patent Application, filed on October 3, 2006

Chapter 25 CCD Based Fiber-Optic Spectrometer Detection Rakesh Kapoor Summary Highly sensitive and cost effective measurement tools are required in biotechnology research and applications. Fluorescence provides very simple, cost effective, and sensitive methods in most of the biosensor techniques. Spectrometer is an essential tool for any kind of spectroscopic measurements. A charged coupled device (CCD)-based fiber optic spectrometer is highly compact, light weight, and an extremely easy to use tool. In this chapter, we have described the use of CCD-based fiber-optic spectrometers in detection of fluorescence signal from a fiber-optic-based sensor. The method can easily be extended to fluorescence detection in any other application. Key words: Spectrometer, CCD, Charge coupled device, Fiber-optic, Fluorescence, Fiber-optic sensor.

1. Introduction Biotechnology research and applications require highly sensitive and cost effective measurement tools, thus fluorescence provides very simple, cost effective, and sensitive methods in most of the biosensor techniques. These biosensors can be used to detect various kinds of bioactive compounds like an enzyme, an antibody, or a nucleic acid (1). The fluorescence is generated either by the attached fluorescence molecules to the analytes (2, 3) or it could be autofluorescence generated by the analyte. To achieve better results and faster measurements, new technologies are emerging. A charged coupled device (CCD)-based fiber-optic spectrometer is one such device. These spectrometers are highly compact, light weight, and extremely easy to use. Here we have described the use of CCD-based fiber-optic spectrometers

Avraham Rasooly and Keith E. Herold (eds.), Methods in Molecular Biology: Biosensors and Biodetection, Vol. 503 © Humana Press, a part of Springer Science + Business Media, LLC 2009 DOI: 10.1007/978-1-60327-567-5_25

435

436

Kapoor

in detection of fluorescence. To illustrate various aspects of its use we have chosen its application in a fiber-optic-based sensor (4).

2. Materials 2.1. CCD-Based Fiber-Optic Spectrometer

1. High-resolution Spectrometer (HR2000+, Ocean Optics, Inc., Dunedin, FL). 2. HR Grating 400–800 nm range; Installed (H9, Ocean Optics, Inc., Dunedin, FL). 3. Order-sorting detector filter; Installed (OFLV-400800, Ocean Optics, Inc., Dunedin, FL). 4. Detector Collection lens; Installed (L2, Ocean Optics, Inc., Dunedin, FL). 5. Optical bench entrance aperture, 200 µm width; Installed (SLIT-200, Ocean Optics, Inc., Dunedin, FL). 6. Laptop computer with USB2 port and Window XP operating system (Latitude D620, Dell, Inc.). 7. 600-µm core silica clad, collection fiber with SMA 905 terminators on both ends (P6002-VIS/NIR, Ocean Optics, Inc., Dunedin, FL).

2.2. Laser Diode

1. Laser diode 405 nm and 4 mW Power (LDM405, Thorlabs, Inc., Newton NJ). 2. 405-nm narrow band filter (NT43104, Edmund Optics, Inc, Barrington, NJ). 3. Lens tube (SM1L05, Thorlabs, Inc., Newton NJ). 4. Lens tube Spanner Wrench (SPW602, Thorlabs, Inc., Newton NJ).

2.3. Collection/ Excitation Chamber

1. 30-mm Cage Cube (C4W, Thorlabs, Inc., Newton NJ). 2. Two packaged collimation lenses of f = 11.0 mm and NA = 0.25 (F220SMA-A, Thorlabs, Inc., Newton NJ). 3. Two collimation lens Mounting Adapters (AD11F, Thorlabs, Inc., Newton NJ). 4. Kinematic Mounting Plate (B4C, Thorlabs, Inc., Newton NJ). 5. Cage Cover Plate (B1C, Thorlabs, Inc., Newton NJ). 6. Short-pass filter (SP500R/25, Maier Photonics, Inc., Manchester Center, VT). 7. 1˝ Optic Mount for mounting Beam splitter (B5C, Thorlabs, Inc., Newton, NJ). 8. 1˝ End cap (SM1CP1, Thorlabs, Inc., Newton NJ).

CCD Based Fiber-Optic Spectrometer Detection

437

9. Laser Line Long-pass Filter 457.9NM (NT47501, Edmund Optics, Inc, Barrington, NJ). 10. Lens tube (SM1L10, Thorlabs, Inc., Newton NJ). 11. Fiber adapter (SM1SMA, Thorlabs, Inc., Newton NJ). 2.4. Fiber Probe

1. 600-µm core silica clad fiber (Fiber-600-VIS/NIR, Ocean Optics, Inc., Dunedin, FL). 2. Bare Fiber Terminator (BFTU, Thorlabs, Inc., Newton NJ). 3. SMA 905 Fiber Connector (10640A, Thorlabs, Inc., Newton NJ). 4. Diamond Wedge Scribe (S90W, Thorlabs, Inc., Newton NJ). 5. Bunsen Burner (17928027, VWR Scientific Products). 6. Ultrasonic Cleaner (1533722C, Fisher Scientific International). 7. Centrifuge (13100510, Fisher Scientific International). 8. Hydrofluoric acid 50% (A1461LB, Fisher Scientific International). 9. PBS (BP24384, Fisher Scientific International). 10. Dry Acetone (AC32680, Fisher Scientific International). 11. Sodium Hydrochloride (Fisher Scientific International). 12. Sodium Bicarbonate (Fisher Scientific International). 13. Aminosilane Reagent (3-Aminopropyltriethoxysilane) (80370, Pierce Biotechnology, Inc., Rockford, IL). 14. Cysteamine hydrochloride (MEA) (30078, Sigma-Aldrich Co., St. Louis, MO). 15. EDTA (17892, Pierce Biotechnology, Inc., Rockford, IL). 16. Desalting Column (CS-800, Princeton Separations, Adelphia NJ). 17. Antibodies to be immobilized. 18. Sulfo-SMCC cross linker (22322, Pierce Biotechnology, Inc., Rockford, IL).

3. Methods 3.1. Selection of Spectrometer and Its Accessories

1. For the biosensor applications it is assumed that real time signal monitoring is the preferred method. This requires that the spectrometer should have a high speed of spectra recording. HR2000+ from Ocean Optics has the capability of transferring spectra continuously at a rate of 1 ms per spectra. Such a high speed is ideal for real time recording of the signal.

438

Kapoor

2. Grating is the main component of any spectrometer. Choice of grating depends on the required spectral range. For most of biosensor experiments, the spectral range is generally from 400 to 800 nm (4). For this kind of spectral range, a preinstalled H9 grating should be ordered with the spectrometer. 3. Most of the time higher order spectra interfere with required first order spectrum therefore it is always safe to get a spectrometer with preinstalled long-pass filter. HR2000+ should be obtained with a preinstalled OFLV (Variable Long-pass Order-sorting Filters) filter to the detector’s window to eliminate second- and third-order effects. This will give a clean first-order spectrum. 4. Another parameter to decide while getting a spectrometer is the slit width as most of the CCD-based miniature spectrometers come with a preinstalled slits. The width of the slit determines the amount of light entering the spectrometer. Higher the width more light will enter the instrument but smaller slit width gives better optical resolution. If no slit is used, the diameter of the fiber connected to the spectrometer determines the size of the entrance aperture. In most of the biosensor experiments, the recorded spectra are either from a molecular dye or from the autofluorescence of the sample. The spectral width of these spectra is typically from 30 to 100 nm (4). A highest available slit width of 200 µm is reasonable to get maximum light collection efficiency and optimum resolution of 5 nm. 5. Generally slit height is larger than the CCD array height, to utilize maximum input intensity a cylindrical lens is needed to focus the light from the tall slit onto the shorter detector elements. The spectrometer should be ordered with a preinstalled L4 lens. This cylindrical lens is fixed to the detector’s window and it increases light-collection efficiency and reduces stray light. It is also useful in a configuration with a largediameter fiber for low light-level applications. 3.2. Assembling Collection/Excitation Chamber

A photograph of assembled collection chamber is shown in Fig. 1: 1. Laser diodes produce significant emission in the red spectral range besides lasing at the laser wavelength. This red light can interfere with the fluorescence signal. To avoid this unnecessary emission (see Note 1) a laser line filter NT43104 (Edmund Optics, Inc.) should be installed in front of the Laser diode LDM405 (Thorlabs, Inc.). Filter should be first mounted into the lens tube SM1L05 (Thorlabs, Inc.) as this lens tube fits well on the laser diode head. 2. Mount the laser assembly (see Note 2) on one of the four side holes of the cage cube C4W (Thorlabs, Inc.).

CCD Based Fiber-Optic Spectrometer Detection

439

Fig. 1. a Photograph of the collection chamber along with a fiber probe. b Photograph showing interior of the collection chamber.

3. Intensity of small amount of scattered or back-reflected laser light is generally comparable with the fluorescence signal and thus can interfere with its detection. To avoid this interference, direct entry of the scattered and back-reflected laser light should be prevented from entering the spectrometer. This can be achieved by using laser line long-pass filter NT47501 (Edmund Optics, Inc.). Transmission of this filter for fluorescence signal wavelengths (457–630 nm) is around 98% but for laser wavelength (405 nm) its transmission is only 10−6%. First mount the filter NT47-501 (see Note 3) in a SM1L10 lens tube and then mount one end of this tube on one of the side holes (right angle to the laser mounted hole) of the cage cube C4W. On the other end of the lens tube, mount the collimation lens F220SMA-A (Thorlabs, Inc.) with the help of lens Mounting Adapters AD11F (Thorlabs, Inc.). Collimation lens should face toward the cage cube. This is the fluorescence collection port (Fig. 1) of our experimental setup. Connect one end of a 2-m fiber P600-2-VIS/NIR (Ocean Optics, Inc.) to this port (SMA connector side of the collimation lens assembly) and other end to the spectrometer input port. 4. Third hole of the cage cube opposite (180°) to the laser mounted hole is used as the output port or fiber probe port.

440

Kapoor

Second collimation lens F220SMA-A (Thorlabs, Inc.) should be mounted with the help of lens Mounting Adapters AD11F (Thorlabs, Inc.) on this hole. Collimation lens should face toward the cage cube. 5. Block the fourth hole with the help of 1″ end cap SM1CP1 (Thorlabs, Inc.). 6. Short-pass filter SP500R/25 (Maier Photonics, Inc.) should be mounted on a 1˝ Optic Mount B5C (Thorlabs, Inc). This filter has about 98% reflectivity from 530 to 689 nm and about 90% transmission for 405 nm. The mounted short-pass filter should further be mounted on the kinematic mounting plate B4C (Thorlabs, Inc.). 7. Kinematic mounting plate with short-pass filter SP500R/25 should be fitted on bottom of the cage cube in such a way (see Note 4) that filter is inside the cage and makes 45° angle with the incoming laser light (Fig. 1). The angle should be such that the reflected fraction of the laser should hit the blind port with end cap SM1CP1. For optimum alignment of the short-pass filter replace the SM1CP1 end cap (see Note 5) with fiber adapter SM1SMA (Thorlabs, Inc.). Connect one end of the P600-025-VIS/NIR fiber to the fiber adapter and point other end toward a white paper. Turn the laser switch on and make sure the shutter in front of laser is on. Now slowly rotate kinematic mounting plate for maximum laser output intensity through the P600-025-VIS/NIR fiber. At the maximum output intensity position tighten the four screws of the kinematic mounting plate. Once the alignment is over, replace the fiber adapter with end cap. 8. Now the setup is ready to record fluorescence signal. A photograph of the experimental setup is shown in Fig. 2 and a schematic of the setup is shown in Fig. 3. 3.3. Preparation of Fluorescence Fiber Probe

1. Use a diamond wedge scribe S90W (Thorlabs, Inc.) to cut a 10-cm long fiber piece from Fiber-600-VIS/NIR (Ocean Optics, Inc.)

3.3.1. Fabrication of Silica Fiber Probe

2. With the help of a Bunsen Burner, burn off 2 cm buffer coating from one end of the 10-cm fiber probe. 3. Rinse the probe with distilled water in ultrasonic cleaner for 4 min. 4. Prepare a 15% Hydrofluoric acid solution by mixing about 2.2 parts of distilled water with one part of 50% hydrofluoric acid. 5. Dip buffer less part (2 cm) of probe in to the 15% hydrofluoric acid until etch to the desired diameter of 270 µm (for 15% HF, the etching rate is about 18.5 µm/h) (see Note 6). 6. Again rinse the etched probe with distilled water in ultrasonic cleaner for 4 min.

CCD Based Fiber-Optic Spectrometer Detection

441

Fig. 2. Photograph of the experimental setup. The laser used in this setup has a separate laser diode driver but LDM405 from Thorlabs, Inc. has a built-in driver.

Spectrometer

LD

Collection fiber

ND

BF

CL

Computer Collection Chamber

LF

CL SF

Fiber Probe

Fig. 3. Schematic of evanescent wave excited fluorescence-based fiber-optic sensor. LD Laser diode, CL Collimation lens, BF Band-pass filter, ND Neutral density filter, LF Long-pass filter, and SF Short-pass filter.

7. After etching burn off the buffer coating from the whole 10 cm long probe on a Bunsen burner. 8. Rinse the probe with distilled water in ultrasonic cleaner for 4 min.

442

Kapoor

9. Rinse the probe with 1 N NaOH in ultrasonic cleaner for 4 min. 10. Rinse the probe with distilled water in ultrasonic cleaner for 10 min. 11. Rinse the probe with acetone in ultrasonic cleaner for 10 min. 3.3.2. Immobilization of Antibodies on Silica Fiber Probe Aminosilylate the Fiber Probe Surface

1. Thoroughly wash and dry the glass or quartz fiber probe surface to be coated (see Note 7). 2. Prepare a 2% solution of the Aminosilane Reagent (3-Aminopropyltriethoxysilane) in acetone. For example, mix 1 part Aminosilane Reagent with 49 parts dry (i.e., water-free) acetone. Prepare a volume sufficient to immerse required probe length. (2 cm probe length will require 400 µl of solution in 1.0 mL without-cap plastic tube or 400 µl in 0.6 mL centrifuge tube.) 3. Immerse probes in the diluted reagent for 60 s. 4. Rinse surface with dry acetone.

Partially Reduce Antibody to Produce Sulfhydryls for Coupling

1. Prepare 0.5 M MEA stock solution by dissolving 6 mg Cysteamine hydrochloride (MEA) in 100 µl Coupling Buffer. (For future use, store this solution at 4°C). 2. Prepare 0.1 M EDTA stock solution by dissolving 3.72 g EDTA power in 100 mL PBS buffer. (This solution could be stored at room temperature.) 3. Prepare Reducing Agent by mixing 50 mM MEA and 10 mM EDTA into Antibody solution. (For total 20 µL of Antibody before dilution, mix 2.5 µL 0.5 M MEA, 2.5 µL 0.1 M EDTA, and 20 µL 0.5 mg/mL Antibody in a 0.2 mL tube.) 4. Incubate the Reducing Agent for 90 min at 37°C (after this step start working on Subheading “Maleimide-Activate the Amino-Modified Surface,” after completing Subheading “Maleimide-Activate the Amino-Modified Surface,” come back to step 5 of this section). 5. Prepare a Desalting Column CS-800 (Princeton Separations) by adding 650 µL EDTA buffer in Desalting Column powder and incubate at least 30 min. Then spin it for 2 min (the maximum amount of reagent for one column could desalt is 100 µL). 6. Purify the reduced antibody from the Reducing Agent using the Desalting Column (proceed to Subheading “Cross-Link Sulfhydryl-Containing Antibody to Activated Surface” as you must have already completed section 3).

Maleimide-Activate the Amino-Modified Surface

1. Prepare a 4.2 mM Sulfo-SMCC cross linker solution in PBS. (For total 400 µL of cross linker solution, add 0.8 mg Sulfo-SMCC

CCD Based Fiber-Optic Spectrometer Detection

443

powder to 400 µL PBS buffer). This solution must be used immediately to avoid hydrolysis (see Notes 8 and 9). 2. Incubate probes in cross linker solution for 1 h at room temperature. 3. Rinse the modified surface with Coupling Buffer (see Note 10). Cross-Link SulfhydrylContaining Antibody to Activated Surface

1. Dilute the reduced antibody solution to about 5–10 µg/mL. 2. Submerge the maleimide-activated fiber probe with the antibody solution. The antibody solution may be diluted in Coupling Buffer to a volume sufficient to submerge the required probe length (generally 2 cm). 3. Incubate it for 4 h at room temperature. 4. Thoroughly rinse the surface with Coupling Buffer to ensure that only covalently attached antibody molecules remain. 5. The surface is now ready to use for detection assays and other applications. Depending on stability of the particular antibody, the surface material may be dried for storage or kept covered in buffer containing 0.02% sodium azide.

3.4. Recording Spectrum/Signal

When immobilized antibodies come in contact with the specific analyte, the analyte get attached to the probe. Now either the analyte itself produce some kind of autofluorescence, or fluorescence can be generated by attaching another labeled antibody to this analyte (sandwich method). 1. Mount the prepared fiber probe on a bare fiber terminator BFTU (Thorlabs, Inc.) by inserting its unetched side into SMA 905 fiber connector attached to BFTU. A photograph of a mounted fiber is shown in Fig. 4. 2. Connect mounted fiber-optic sensor probe to the output port (see Note 11) of the collection chamber. 3. First install the spectrometer software (see Note 12) on the laptop Latitude D620 (Dell, Inc.) and then connect the

Fig. 4. A photograph of a mounted fiber probe.

444

Kapoor

spectrometer HR2000+ to the USB2 port of the laptop with the provided USB cable. Once the HR2000+ is installed, it must be configured by operating software’s Configure Spectrometer options so that operating software (OOIBase32) recognizes the HR2000+ spectrometer. Details can be obtained from the operation manual of the software. 4. For power stabilization, it is always a good idea to start the laser few minutes before starting the experiment. 5. Start the spectrometer operating software and open the laser shutter. Record the fluorescence signal by clicking the appropriate button. Try to adjust the integration time of the spectrometer to a value, such that the maximum spectral amplitude is less than 90% of the full scale. This will ensure that recorded signal is not saturated. 6. For maximizing the signal adjust the three hex adjuster fine alignment screws (see Note 13) on the back (Fig. 1) of kinematic mounting plate B4C (Thorlabs, Inc.) 7. Once the maximum signal value is known and integration time is adjusted, close the laser shutter and record the dark back ground signal (see Note 14) and store it. 8. Save this signal and turn the background deduction button on. (Details about these various operations can be read in the software manual of the spectrometer). 9. Now record the spectrum of the collected (see Note 15) fluorescence and save it to the file.

4. Notes 1. Red tail emission from the laser diode can be further reduced if the laser is operated at much higher power than the required power and neutral density filters are placed in the laser path along with the band-pass filter to reduce the power level to the required value. 2. All the SM1 series lens tubes and cage cube (Thorlabs, Inc.) holes have compatible threading therefore mounting of these tubes on the cage cube is straightforward. 3. Make sure that the coated part of the laser line filter NT43-104 faces the laser side. Generally there is a mark on the ring of the filter which indicates the coated side. 4. Make sure that the coated part of short-pass filter faces the spectrometer port or collection port.

CCD Based Fiber-Optic Spectrometer Detection

445

5. Replacement of end cap SM1CP1 with fiber adapter SM1SMA and its connection with the fiber P600-025-VIS/ NIR is only done for alignment purpose. After completing the short-pass filter alignment, fiber adapter should be replaced back with end cap. 6. Make sure the probe stand vertically in the acid solution. 7. Perform steps 2 and 3 in a fume hood. 8. The Reducing Agent (Subheading “Partially Reduce Antibody to Produce Sulfhydryls for Coupling”) should be made first because it needs 90 min incubating time. 9. The better reaction of NHS-ester is in the environment of pH = 8.3. However, maleimide group will slowly hydrolyze and loses its reaction specificity for sulfhydryls at pH values >7.5. For these reasons, conjugations with these cross linkers are usually performed at pH 7.2–7.5. 10. The maleimide-activated surface may be dried and stored desiccated at 4°C for later use. 11. This connector is for 630-µm outer diameter fiber. For other diameter fibers appropriate SMA 905 fiber connector should be used. 12. This port is called output port as the excitation laser comes out of this port although it also acts as an input port for collected fluorescence. 13. Never connect the spectrometer to the computer, without first installing the required software and drivers. This is a onetime step if you are using the same computer for future experiments. 14. This is one time adjustment for improving the signal to noise ratio. 15. Always keep the integration time of the back ground signal and actual signal same.

References 1. Mehrvar, M., Bis, C., Scharer, J.M., Young, M.M., Luong, J.H., 2000. Fiber-optic biosensors – trends and advances. Anal. Sci. 16, 677–692 2. Hirschfeld, T.E., Block, M.J., 1984. Fluorescent immunoassay employing optical fiber in capillary tube, US Patent No. 4447546

3. Andrade, J.D., VanWagenen, R.A., Gregonis, D.E., Newby, K., Lin, J.N., 1985. Remote fiberoptic biosensors based on evanescent-excited fluoroimmunoassay: Concept and progress. IEEE Trans. Electron. Devices ED-32(7), 1175–1179 4. Kapoor, R., Kaur, N., Nishanth, E.T., Bergey, E.J. Halvorsen, S.W., Prasad, P.N., 2004. Biosens. Bioelectron. 20, 345–349

INDEX A Absorption sensor.......................................................... 311 Achromatic lenses.........................................38, 42, 46, 382 Acousto-optic tunable filter ................................... 293–304 A/D converter ............................ 38, 43, 222, 313, 317, 318 Adherent cell ......................................................... 204, 212 Affinity selection ............................................................. 73 Alkaline denaturation .................................................... 351 Alkaline-phosphatase .............................326, 327, 331–334 Alkanethiols ...........................................................5, 18, 75 Alkylthiols ................................................................. 74, 75 Amine groups ......................... 108, 111–113, 158, 193, 194 Amine-modified surfaces ...............................110, 422–443 Amine-thiol........................................................... 108, 110 Aminodextran........................................................ 171, 173 Amorphous silicon ................................................ 361, 362 Amplification .........................3, 5, 6, 13, 21, 247, 307, 308, 311–314, 317, 321, 343, 345, 349–350, 372, 425 Analog-to-digital conversion ......................................... 329 Angular scan ................................. 71, 90–92, 102, 103, 133 Antibody................................ 3–7, 9, 12–14, 16–19, 21–24, 28, 31, 33, 34, 39, 44–46, 73, 77, 81, 82, 84, 85 Antibody purification ............................................ 277, 327 Anti-streptavidin antibody .......................................... 6, 13 ATP................................................ 338, 339, 341, 342, 344, 346, 348, 353, 354, 356–358 Autofluorescence ........................... 180, 239, 242, 248–250, 252, 256, 257, 264, 393, 403, 404, 435, 438, 443

B Bacteriophage .................................................275, 371, 372 Bacteriophage DNA .............................................. 371, 372 Bandpass filter ............................... 222, 248, 252, 255, 275, 276, 294, 311, 376, 378, 405, 414 Bandwidth ......................133, 182, 185, 245, 247, 262, 263, 294–296, 301, 307, 309, 312–314, 318, 321, 322 Base implant layer ......................................................... 329 Basic fibroblast growth factor (bFGF) ...100, 123, 124, 135 BCIP/NBT ........................................................... 327, 332 Beads ....................................................6, 13, 347, 350–352 Beam directing .......................................241, 251, 254–255 Beam splitter ....................................................4, 7, 10, 436 Beamsteering ................................................................... 38 Bending .................................. 241, 252, 255, 311, 318, 386 Binding inhibition assay .............................84, 85, 167, 175

Binding peptide ............................................................. 124 Biotin........... 6, 13, 49, 52, 54, 57–63, 74, 75, 78–81, 95, 97 Bipolar photodiode .........................................325, 326, 328 Bipolar semiconductor ................................................... 327 Blocking .............................9, 12, 13, 16–18, 78, 81, 82, 96, 101, 133, 135, 278, 285, 295, 300, 304, 321, 332 Blood 3, 5, 9, 13, 189, 190, 192–194, 197–199, 206, 242, 252, 283 Board assembly .............................................................. 316 Board preparation .................................................. 314–316 Board testing ......................................................... 316–317 Bonding ............................................ 12, 185, 265–266, 270 BoNT-A .........................................................259, 267–269 Botulinum neurotoxin ............................259, 261, 267–269 BSA blocking .......................................................... 12, 332 Buried layer ........................................................... 328, 329

C Campylobacter jejuni.................................................275, 288 Capillary ........................................ 139–141, 143–146, 151, 155, 156, 161, 163 Capillary electrophoresis.........................221–235, 361–373 Capillary waveguides ..............................390–392, 400, 427 Carcinoembryonic antigen (CEA) .............5, 102, 128, 129 Casein 76, 78, 82, 96, 110, 134 CCD.........................................8, 14, 15, 17, 25, 26, 28–31, 34, 73, 85, 103, 172, 190–192, 196, 197, 203–218, 221, 223–228, 234 Cell culture .................................................................... 212 Cell-targeting ................................................................ 217 Chopper .........................................................368, 369, 392 Chromium ....................................................11, 18, 23, 364 Cleaning .................................................. 39, 43, 44, 46, 86, 116, 173, 177, 186, 194, 196, 363, 364, 369, 370, 401, 425 CMOS ...................................................246, 313, 314, 325 CO2 laser ....................................................................... 406 Collimated beam ................................. 28, 29, 46, 103, 183, 244, 254, 302, 392, 426 Collimated light .......................8, 14, 19, 28, 40, 42, 43, 46, 244, 257, 302, 304, 382, 393, 397, 401, 426, 427 Collimating lens ..............................................92, 102, 171, 240, 241, 244, 251, 253, 254, 257, 304, 391, 392, 397, 425, 426, 436, 439–441 Colloidal gold .............................. 94, 95, 99, 101, 118–119, 127, 128, 135

447

BIOSENSORS AND BIODETECTION 448 Index Colorimetric .......................................................... 204, 259 Combinatorial ............................................................... 362 Competitive ....................................... 18, 84, 190, 230, 273, 281, 285–286, 290 Competitive immunoassay.............. 273, 281, 285–286, 290 Conjugated ................................. 3, 6, 14, 16, 389, 391, 399 Continuous flow immobilization Convex lens ................................. 42, 46, 95, 103, 278, 296, 298, 300–303, 368, 391, 397, 415, 416 Cooled CCD ......................................... 242, 257, 259, 264, 275, 276, 279, 287, 429, 430 Cooling.......................................... 214, 239, 243, 249–250, 253, 254, 260, 262, 264, 269, 308, 321 CorelDRAW ..........................................406, 411, 412, 414 Coupling optics ..................................................72, 73, 171 Critical angle ..........................................................4, 40, 90 Crosslinking .................................................................. 112 Current amplifiers ..................................327–329, 376, 386 Current-to-voltage conversion .............................. 385, 386 Current-to-voltage converters ............................... 327–329 Cy5...... .......................................... 277, 282, 285, 286, 290, 389, 391, 393, 396, 399, 400, 425, 432, 433

D Dark current .................................. 172, 242, 247–250, 255, 308, 309, 342, 343, 362, 364 Data acquisition............................................................. 386 Deoxynivalenol ...............................................274, 281, 285 Detection cell .........................................223–227, 233–235 Dextran..........................74, 95–97, 104–108, 116, 124, 327 Dielectric layer...................................... 50, 54, 90, 245, 257 DMAP ...................................................................... 76–78 DNA...........................................35, 74, 111, 124, 135, 139, 141, 158, 159, 161, 173, 190 analysis............................................................. 337–359 sequencing ................................ 337, 339, 352–354, 371 dNTPs... ................................. 338, 344, 345, 348, 353–357 Double-layer adhesive ................................................... 412 Double-sided adhesive................................................... 270 Double sided adhesive tape ........................................... 270 Dual-waveguide interferometric ...................................... 50 Dynode ...................................................378, 379, 381–384

E E.coli..... ................................... 101, 126, 307, 318, 325–334 EDC..... ...................... 12, 13, 17, 19, 23, 33, 76, 77, 79, 96, 97, 99, 100, 102, 105–108, 110, 111, 116, 117, 122, 130, 133, 142, 157, 184, 193–195, 197 EDC-NHS.................................. 12, 13, 19, 105, 106, 108, 110, 111, 116, 117, 122, 133 EL-CCD ................................................259, 261, 263–269 Electrode .........................363, 366, 371, 373, 378, 381, 382 Electroluminescence .............................................. 259–271 Electrophoresis ....................... 221–235, 337, 361–373, 383

ELISA... ..................................... 12, 18, 129, 193, 287, 288 Ellipsometry .....................................................22, 167, 169 Emission filter ....................................... 261, 264, 265, 275, 276, 279, 294, 312, 319 Epitaxial layer ........................................................ 328, 329 Epoxy.... ..........................................................268, 406, 414 Etching ...........................152, 153, 163, 315, 363–365, 369, 404, 440, 441 Evanescent wave ..................4, 37, 53, 61, 90, 190, 390, 441 Excitation filter............................... 261, 264, 265, 294, 312

F α-Fetoprotein ................................................................ 5, 9 Fiber optic ..................................... 37, 38, 40–43, 141, 144, 146, 147, 171, 311, 390, 404, 435–445 Fiber probe ..................................... 226, 234, 237, 439–443 Flow cell ........................................ 4, 10–12, 15–18, 24, 25, 30, 31, 50, 52, 54, 57, 60, 62 Flow through ................................................77, 79–81, 309 Fluidics ............................................ 76, 77, 79, 81, 85, 139, 262, 264, 279–281, 287, 339, 361 Fluorescein .................................... 223–224, 297, 302–304, 362, 372, 404, 407, 408, 416–421 Fluorescein isothiocyanate isomer I (FITC)................. 221, 223–224, 228, 230, 235, 259, 261–263, 265, 267–268, 383–384, 425, 432–433 Fluorescence ...........................3–19, 46, 129, 139, 180, 190, 221–226, 230–233, 235, 239–242, 244–246, 248–250, 252, 255–257, 259–261, 263–264, 273, 279, 293–304, 307, 309–313, 318, 322, 337, Fluorescence activated cell sorting................................. 362 Fluorescence detector ..................... 223, 232, 260, 361–373 Fluorogenic peptide substrate .................259, 261, 267–269 Fluoroimmunoassays ............................................. 273–290 Fluorometer .....................259, 262, 265, 267, 269, 319, 423 Fluorophore ....... 3, 6, 7, 14–15, 17, 278, 282, 311, 404, 408 Flurocence ..................................................................... 262 Folding .............. 94, 135, 252, 255, 284–285, 298, 301, 390 Food..........................................65, 74, 76, 78–79, 273, 275, 277–278, 283–284, 288, 316, 337 Frame grabber.........................................296, 298–299, 302 Frequency variation ....................................................... 321 FTIC.... ..........................................................221, 231, 235 Fumonisin.............................................................. 281, 285 Functionalization ...................... 76, 156, 158–160, 173, 193

G Gelatin............................................................278, 285, 290 Glucose .................................................................. 104, 370 Gold thin film ..............................................4, 5, 10, 14, 19 Gold wire......................................................................... 11 Grating ........................... 69, 91, 92, 95, 103, 162, 163, 186, 189–200, 227, 229, 233, 294–295, 300, 436, 438 Grating coupler ............................................... 69, 189–200

BIOSENSORS AND BIODETECTION 449 Index H

L

Half-ball lens ...................363, 366–369, 371, 424, 427, 429 Hapten .................................................................. 129, 175 Hemicylindrical prism ......................................7, 10–11, 16 His-tagged..................................................98, 99, 116–117 Human immunodeficiency virus (HIV) .................................................203–218, 318 Hybridization .................................... 74, 89, 100, 124, 125, 135, 177, 186, 353 Hybridoma ...................................................................... 39 Hydrogel.....................................................74, 95, 109, 173 Hydrophobic ......................................... 43, 74, 95, 98, 107, 115, 116, 193

Label-free ...............................4, 40, 49, 58, 63, 65, 89–135, 139–164, 167, 168, 179, 189–200 Labeling .............................................. 16, 37, 49, 139, 179, 251–253, 256, 277, 281, 282, 361, 362, 371, 425 LabVIEW ................................... 40, 42, 46, 145, 147, 149, 163, 222, 318, 319, 369, 391, 398, 425, 430 Laminating ............................................................ 266, 270 Laminator...................................................................... 262 Laser.... ............................. 4, 7, 8, 10, 14, 15, 24, 28, 53–55, 61, 63, 72, 91, 92, 94, 95, 102, 141, 145, 147–150, 162, 163, 170, 190–191, 195, 197, 222, 230, 239, 241–246, 251–255, 257, 258, 260, 261, 264, 265, 269, 276, 278 Laser-induced-fluorescence (LIF) ..........221, 222, 229, 375 Laser line generator ....................................................... 391 Leaky waveguide ............................................................. 89 Lens..... .............................................. 4, 7, 8, 10, 14–15, 24, 25, 28, 29, 46, 92, 95, 102, 103, 171, 182, 183, 191, 203, 204, 206–209, 212, 214, 216, 239–242, 244–245, 251–257, 259, 261–262, 264 Light.................................................. 4–6, 8, 10, 11, 14–16, 18, 19, 21, 22, 25–29, 31, 34, 37, 40–43, 46, 50, 53, 54, 59, 66, 67, 69, 70, 72, 73, 85, 90–92, 94, 95, 102, 103, 118, 133, 140, 142, 146, 147, 152, 162, 163, 167–168 Light box ........................................205–208, 210, 215–217 Light-emitting diodes (LED) .........................92, 102, 121, 221–226, 228–235, 260, 294, 297, 309–311, 316–320, 328, 375, 382–383, 405, 407, 414–416, 424–426, 429, 430, 432,–434 Light scatter ........................... 133, 270, 281, 303, 362, 370 Line scanning ...........................................37, 38, 40, 43–45 Liposome ............................................................... 109, 115 Liquid core optical ring resonator (LCORR) .................................................. 139–164 Liquid crystal variable retarder (LCVR) ..............................................293, 296–303 Luciferase ...................................... 338–339, 342, 344, 345, 347, 356–358 Luminescence ........................ 221–235, 259–271, 337–339, 341–344, 354, 356, 375, 423, 425 Luminescence detector .......................................... 221–235 Luminometric .................337, 339, 344, 347, 348, 352, 358 Lysozyme .............................................................. 101, 126

I IC clock ......................................................................... 318 Illumination...................... 4, 19, 28, 67, 72, 85, 91, 92, 133, 172, 214, 215, 217, 224–226, 235, 259–271, 273, 304, 309, 311, 317, 387, 390–393, 395, 397, 423–430, 432 Immobilization .................................... 6, 12–13, 16, 17, 58, 73–81, 89, 90, 95–97, 104–106, 108–113, 115–117, 122, 134, 158–160, 164, 173–174, 326, 332, 351, 401, 442 Immunoassay ..................................... 3, 5, 7, 10, 12–14, 17, 204, 240–242, 252, 257, 273–290, 389, 398, 400, 423, 425 Impedance ............................................................. 314, 317 Incident angle ......................4, 5, 10, 14, 16, 19, 31, 50, 168 In-fiber .................................................................. 403–421 In-fiber grooves ..................................................... 403–421 Injecting system ......................................................... 30, 34 Injection valve.........................................405, 409, 410, 417 Integrated circuit (IC) .................... 314, 317, 318, 325, 361 Integrating Waveguide Biosensor (IWB) .........................................389–401, 423, 434 Integration time...............................................15, 227–229, 231, 232, 234, 247, 255, 444, 445 Interference filter ......................................... 8, 15, 102, 133, 294, 311, 318, 361–363, 365–366, 368, 372, 373, 383, 401, 414 Interferometer ................................................180–182, 184 Interferometry ................................................. 22, 179–186 Iris......................................... 4, 7, 10, 38, 41, 42, 255, 296, 298–301, 303, 369 Isoelectric point (PI)...............................104, 108, 119, 133

K Kinetic.. ..................................21, 28, 33, 49, 60–62, 71, 89, 90, 93, 102, 104, 119, 120, 124, 129, 139, 158, 161, 175, 176, 178, 180, 185, 186, 199, 362 Kinetics analysis .............................................................. 33 Kretchmann ....................................................................... 4

M Magnesium fluoride (MgF2) crystal ............................. 381 Magnet ........................................ 6, 39, 41, 66–68, 72, 141, 190, 222, 262, 314, 347, 351, 381–382, 409 Magnetic beads.................................................................. 6 Mask................................................. 39, 315, 363, 369–370 Mask aligner .................................................................. 363

BIOSENSORS AND BIODETECTION 450 Index Media..... ..........................................................55, 67, 76, 78, 79, 81, 205, 212, 232, 319 Medium ..............................51, 53, 56, 59, 66, 82, 101, 168, 181, 190, 205, 327, 394, 427 Membrane ..............................................109, 128, 223, 362 Metal film...............................................................4, 66, 69 Microchannel ..........361–362, 367–369, 371, 382, 403–421 Microfluidic .........76, 90, 139, 143, 171, 361–373, 403–421 Monochromatic light.....................................21, 28, 70, 72, 85, 90–91, 102, 172 Monochromators ...................................186, 233, 293–294, 307, 311, 376 Monte Carlo simulations ....................................... 393, 427 Motion controller ............................... 42, 43, 141, 145, 147 MUA solution ........................................................... 22, 23 Multiplexer .....................................................170, 340, 343 Mumps Virus ............................................................ 37–46 Mycotoxin aflatoxin ............................................... 281, 285 Mylar.... .......................................... 406, 412–414, 420, 421

N Nanobeads ............................................................... 3, 6, 13 Nanoparticles................................................................... 22 Neutral density filter................8, 14, 19, 191, 195, 441, 444 NHS/EDC ................................................................ 12, 17 N-hydroxylsuccinimide (NHS) .....................12–13, 19, 23, 52, 54, 76, 77, 79, 96, 97, 99, 100, 102, 105–108, 110, 111, 114, 116–118, 122, 130, 133, 142, 157, 171, 173, 193–195, 197, 252, 277, 282, 445 Noise sources .......................... 246, 247, 249, 250, 257, 309

O Objective ............................8, 15, 24, 28, 54, 182–183, 213, 301–303, 407, 414–416 Ochratoxin .....................................................274, 281, 285 Oligo probe immobilization .................................. 158–159 One base extension .........................................341, 354, 356 Op Amp .................................................313, 314, 318–320 Optical fiber .........................41, 43, 44, 171, 226, 234, 376, 377, 382–383, 386, 403–407, 410–417, 419, 420 Optical path length (OPL)............. 140, 168, 179–181, 235 Optical ring resonator............................................ 139–164 Optical table ............................................. 7, 10, 38, 42, 297 Oriented immobilization ................................................. 77 Oxide layer .............................................................. 54, 329

P Pathogen.........................................................260, 325–334 PDMS. ........................... 274, 276, 278–281, 284, 286–287 Peltier................................................................93, 250, 262 Peptide purification ....................................................... 100 Peristaltic pump ............................. 7, 10, 12, 141, 150, 163, 170, 171, 196, 276, 280 Pesticide........................................................................... 84

Phase detection.......................................................... 21–35 Phase difference....................................................25, 26, 91 Phase distribution ......................... 21, 22, 26–28, 31–32, 34 1,4-Phenylene diisothiocyanate (PDITC) ............................................193, 195–198 Phospholipid ....................................................98, 115, 116 Photobleaching.........................................19, 249, 417, 420 Photodetector .................................. 70, 142, 147, 149, 150, 163, 221, 242, 307, 311, 314, 317, 318, 320–322, 386, 408, 416 Photodiode .........................................4, 7, 8, 10, 14, 15, 55, 56, 71, 172, 240, 243, 259, 260, 307–323, 325–334, 337, 342–344, 361–364, 369, 373 Photo-diode array.................................................. 337–359 Photodiode detector ............................. 7, 8, 10, 14, 15, 314 Photolithography............................328–329, 364–366, 369 Photomultiplier ....................... 17, 222, 240, 259–260, 269, 308, 321, 375–385, 389, 391 Photonic crystal ......................................................... 49–63 photoreceiver ............................................21, 142, 317, 318 Photoresist ..............................................315, 363–366, 395 Phycobiliprotein ............................. 221, 222, 231, 233, 234 PID........................................................................ 251, 254 Piezoelectric .................................................................. 295 Pinholes lens............................................................ 24, 102 Piranha solution ...................................... 11, 18, 22, 23, 54, 86, 172, 194, 196, 198, 364 Planar waveguide ............................ 238, 241, 251, 254–256 Plano-convex lens .......................................92, 95, 102, 103 Plasma.................................9, 11–13, 16–19, 86, 242, 283, 364–366, 369 Plasmid .......................................................................... 206 Platinum electrode......................................................... 371 Polarized light ............................... 4, 10, 25–26, 28, 40–42, 91, 167, 300, 301 Polarizer.......................... 25, 31, 40, 41, 72, 90–92, 95, 103, 294, 295, 302, 304, 307 Polarizing filter .......................................................8, 14, 15 Polarizing prism ................................. 25, 26, 28, 29, 38, 40 Poly(diallyldimethylammonium chloride) (PDDA)..............................................37, 39, 44, 45 Polyethylene glycols (PEG) ....................171, 173, 278, 284 Poly L-lysine ............................................................. 39, 44 Polymerase chain reaction (PCR) ................................. 160, 260, 338, 345–346, 349–351, 362, 423, 425 Polymethyl methacrylate (PMMA).......................... 16, 30, 171, 404–406, 410–414, 416, 418, 420 Prism..................................... 4, 7, 8, 10, 11, 15, 16, 19, 23, 25, 26, 28, 29, 31, 34, 38, 40, 42, 46, 52, 69, 71, 72, 85, 90–92, 95, 103, 162, 295, 300 Prostate-specific antigen ................................................ 105 Protease ................................................................. 139, 141 Protein A ........................................ 102, 126, 129–131, 133 Protein array ....................................... 37, 38, 40, 41, 43–46 Protein coupling ............................................................ 189

BIOSENSORS AND BIODETECTION 451 Index PVC.... ......................................................... 7, 10, 16, 196 Pyrosequencing....................... 337–349, 352–354, 356–358

Q Quartz tungsten......................................................... 38, 46

R Raman spectroscopy ...................................................... 180 Reactive surface ................................................76, 177, 193 Reflectometric Interference Spectroscopy ............. 167–178 Reflectometry ........................................................ 167–178 Refractive index ......................................... 3, 4, 6, 8, 11, 16, 18, 19, 21–23, 25–28, 32, 35, 37, 43, 50, 54, 62, 65–70, 72, 73, 85, 89–93, 95, 103, 118, 139–141, 152–155, 163, 169–171, 177, 180, 181, 185, 189–192, 197–199, 310, 372 Resonance............................................ 3–19, 21–35, 37, 40, 42–45, 54–56, 58, 62, 65–86, 91, 92, 103, 125, 133, 134, 150, 157, 180, 189, 222, 242 Resonance wavelength ....................................37, 40, 42–45 Resonant Mirror (RM)...............................49, 89–135, 189 RF generator ..................................................297–299, 302 RNA.........................................................................425 Rotation stage................................... 4, 8, 10, 14, 15, 38, 42

S Salmonella ........................................275, 289–291, 398–400 Sample loop .................................... 172, 405, 409, 410, 417 Sandwich ..................................................... 3, 5, 12, 14, 17, 83, 84, 94, 127, 128, 174, 177, 241, 254, 256, 273, 281, 285–287, 295, 316, 325, 326, 370, 389, 390, 398, 400, 443 Sandwich assay .............................. 12, 14, 93–94, 127, 128, 273, 281, 285, 390, 400 Screening ................................ 172, 189, 190, 273, 274, 362 Secondary antibody .......3, 5–7, 9, 12–14, 16, 17, 19, 84, 95 Self-assembled monolayers (SAMS) .............................. 5, 14, 18, 23, 74, 76–79 Semiconductor laser ....................... 241–244, 251, 363, 367 Serodiagnosis ..............................................37–40, 189–200 Serum...............................................9, 13, 23, 37, 44, 45, 76, 78, 79, 82, 98–102, 121, 122, 142, 177, 184, 189, 190, 192–195, 197–200, 205, 277, 281, 283, 358 Signal amplification ......................................3, 5, 6, 13, 312 Signal digitalizing .......................................................... 391 Silanization .................................... 104, 133, 172, 173, 195, 196, 326, 332, 399 Silanized chips ................................ 104, 172, 173, 196, 332 Silicon 7, 10, 11, 54, 119, 249, 254, 328, 329, 361, 362 Silicon oxide ............................................................ 54, 329 single-stranded DNA (ssDNA)............................. 338, 339 SNPs 337, 344, 354–357 Spatial modulation .................................................... 21–35 Spectral filtering ............................................................ 297

Spectral SPR ............................................................. 37–46 Spectrofluorometer ........................................................ 423 Spectrometer ..................................... 37, 38, 40–43, 45, 70, 73, 85, 170–172, 182, 183, 186, 223–228, 233, 277, 423, 425, 426, 429–430, 434, 444, 445 Spectroscopy ........................................ 3–19, 167–178, 180, 189, 190, 293 Spin coater..................................................................... 363 SPR array ................................................ 22, 27, 37, 38, 40, 41, 44–46, 71, 73, 76 Sputter. .......................................................39, 52, 363–365 Stability .18, 67, 73, 74, 78, 79, 85, 185, 194, 222, 314, 318, 375, 384, 443 Stabilization of baseline ................................................... 12 Staining ..... 46, 203, 204, 206, 210, 212, 214, 217, 218, 255 Staphylococcus aureus enterotoxin B (SEB) ..........................................274, 277, 288, 390 Streptavidin ........................................... 3, 6, 13, 49, 51, 52, 57–63, 76, 78–80, 97, 98, 100, 106, 111–112, 123–125, 134, 193, 347, 351 Striper 328 Surface functionalization ....................... 76, 78, 81, 96, 106, 108–111, 113, 156, 158–160, 173, 179, 193 Surface passivation .................................189, 190, 193, 199 Surface plasmon field-enhanced fluorescence.............. 3–19 Surface Plasmon Resonance (SPR) .................3–19, 21–35, 37, 49, 51, 65–86, 167, 180, 189, 242, 312 Syringe pump ............75, 141, 150, 151, 405, 409, 414, 417

T Temperature control ......................... 62, 149, 150, 186, 316 Thermal noise......................... 242, 247, 260, 262, 269, 309 Thermister ............................................................. 142, 150 Thermoelectric cooler.....................................239, 250, 253 Thrombin .............................................................. 165, 187 Titering ................................................................... 46, 344 Total internal reflection (TIR).........................4, 14, 40, 50, 51, 53, 140, 241, 257, 273, 280 Toxin........................................ 84, 259–261, 267, 274, 277, 281, 283–286, 288, 290, 387, 390 TracePro simulation.......................... 392, 395, 396, 427429 Translation stages ................................................ 39, 41–44 Triton X-100 ........................................................39, 44, 46 Trypsin .......................................................................... 156 Tumor.. ................................................................ 3–19, 128 Tunable filter ......................................................... 293–304 Tunable laser................................... 141, 147, 149, 150, 162 Tween... ....................................9, 12, 13, 17, 18, 44, 76, 81, 96–101, 261, 391, 399 Tween-20 .................................... 13, 39, 44, 75, 76, 95–99, 101, 102, 125, 135, 194, 197, 277, 278

U UV glass ........................................................................ 381

BIOSENSORS AND BIODETECTION 452 Index V Valve.... ...............................30, 34, 170, 171, 276, 279, 280, 287, 340, 341, 405, 409, 410, 417 Vector....................................................................... 67, 181

W Wafer... ........................................... 363–366, 369, 370, 372 Waveguide ...........................................................49–51, 54, 89–91, 102, 140, 162, 190, 193, 195, 198, 221, 239, 241, 242, 248, 250, 251, 254–258, 260, 273

Waveguide sensor .................................. 241, 242, 254, 256, 389–391, 423–434 Wavelength interrogation .................................... 37–38, 50 Wavelength scan ................................................91, 92, 102, 103, 109, 133 Wollaston prism .......................................25–26, 28, 29, 34

Z Zoom lens .............................................................203, 204, 208, 212, 214