Protein Microarray for Disease Analysis: Methods and Protocols

Methods in Molecular Biology™ Series Editor John M. Walker School of Life Sciences University of Hertfordshire Hat...

Author: Catherine J. Wu

195 downloads 991 Views 8MB 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

Methods

in

Molecular Biology™

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

wwwwwwwwwwwwwwww

Protein Microarray for Disease Analysis Methods and Protocols Edited by

Catherine J. Wu Division of Hematologic Neoplasia, Department of Medical Oncology, Cancer Vaccine Center, Dana-Farber Cancer Institute, Boston, MA, USA

Editor Catherine J. Wu Division of Hematologic Neoplasia Department of Medical Oncology Cancer Vaccine Center Dana-Farber Cancer Institute Boston, MA 02115 USA [email protected]

ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-61779-042-3 e-ISBN 978-1-61779-043-0 DOI 10.1007/978-1-61779-043-0 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2011921931 © Springer Science+Business Media, LLC 2011 All rights reserved. This work may not be translated or copied in whole or in part without the written permission of the publisher (Humana Press, c/o Springer Science+Business Media, LLC, 233 Spring Street, New York, NY 10013, USA), except for brief excerpts in connection with reviews or scholarly analysis. Use in connection with any form of information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed is forbidden. The use in this publication of trade names, trademarks, service marks, and similar terms, even if they are not identified as such, is not to be taken as an expression of opinion as to whether or not they are subject to proprietary rights. While the advice and information in this book are believed to be true and accurate at the date of going to press, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein. Printed on acid-free paper Humana Press is part of Springer Science+Business Media (www.springer.com)

Preface Protein microarrays are a rapidly growing segment of proteomics that enable highthroughput discovery-driven research through direct measurement of the molecular endpoints of various physiological and pathological states. The human genome has some 30,000 protein-coding genes, while the human proteome is estimated to have at least 90,000 proteins. By now, protein microarrays have been used for identifying protein– protein interactions, discovering disease biomarkers, identifying DNA-binding specificity by protein variants, and for characterization of the humoral immune response. In this volume, we provide concise descriptions of the methodologies to fabricate microarrays for comprehensive analysis of proteins or the response to proteins that can be used to dissect human disease. These methodologies are the toolbox for revolutionizing drug development and cell-level biochemical understanding of human disease processes. Three general categories of arrays have been developed, which we describe in detail in this volume. The first and most commonly used are the protein-detecting analytical microarrays, described in Part I. Conventionally, the design of these arrays is based on the principle of a sandwich immunoassay. Thus, these capture protein on an array surface from biologic samples and quantify presence of those specific analytes using a detection reagent. Arrays may be coated with antigen-specific antibodies to detect specific proteins from body fluids (Chap. 1), whose identity can be confirmed using label-free detection based on mass spectrometry (Chap. 2). An alternative to detection on solid phase uses newly available bead-based strategies (Chap. 3). Antibody-based detection can be also implemented in a high-throughput fashion on reverse-phase protein arrays. Here, cell lysates are printed to a solid support, followed by quantitative immunodetection, as described in Chap. 4. These general designs have been further modified by other investigators to optimize exploration of specific biologic problems. For example, aptamer (Chap. 5) and recombinant lectin (Chap. 6) arrays have been successfully developed. A second category of protein microarray is antigen microarrays that seek to detect antigen-specific antibody from biologic samples (primarily serum and plasma), covered in Part II. Here, arrays are coated with tens to thousands of proteins in order to detect specific reactive antibodies. These have proven valuable for biomarker discovery and detection. Many possible formats of antigen expression on microarrays are now available. Both commercial high-density protein microarrays that express recombinant protein for serum profiling, as well as technology for custom production of arrays to express a tailored collection of proteins, are now available (Chap. 7). Technology to synthesize comprehensive arrays of peptides has also been established (Chap. 8). Finally, high-throughput protein fractionation strategies have been developed that enable array spotting of antigens in their native format (Chap. 9). Production and isolation of proteins can be cost- and laborintensive. As an alternative, programmable arrays, in which cDNA-containing plasmids are spotted on solid support and protein is freshly translated in situ, offer a versatile solution to the problem of recombinant protein production (Chap. 10).

v

vi

Preface

The final category of protein microarray is protein function microarrays to interrogate direct biochemical and physical interactions among biomolecules (Part III). These include profiling of protein–protein, protein–lipid, protein–DNA/RNA, and small molecule interactions. In Chap. 11, we provide protocols for high-throughput mammalian-based detection of protein–protein interactions, operating on the principle of two-hybrid screening techniques. Programmable arrays have been also developed for this purpose (Chap. 12). Among the many specific applications of protein function arrays are the detection of kinase– substrates interactions (Chap. 13) and the characterization of posttranslational modifications that can serve important regulatory functions in eukaryotic cells (Chap. 14). In most cases, discovery by protein microarray screening requires validation of candidate targets, in order to focus subsequent biologic studies. Part IV of this volume offers two separate approaches to candidate target validation. Both require independent production of the protein analyte to confirm specific reactivity. Both the generation of protein microarrays and the implementation of validation steps have been greatly accelerated by the recent availability of large insect and mammalian proteome libraries. Within these libraries, numerous open reading frames have been cloned and deposited in vector formats that are amenable to protein expression (Part V). The two final sections of the volume are devoted to signal detection strategies (Part VI) as well as data analysis techniques (Part VII). The most conventional and widely used methods are based on fluorometric or colorimetric methods (Chap. 18), while newer label-free detection systems, such as using FRET (Chap. 19) or surface plasmon resonance (SPR) (Chap. 20), will likely be increasingly employed in the future. Validated software for analysis of protein microarrays is only developing now and is obviously critically important for data analysis (Chap. 21). Finally, knowledge of the publicly available databases that are relevant to proteomics studies can enable more efficient data analysis (Chap. 22). We hope that this volume provides a solid framework for understanding how protein microarray technology is developing and how it can be applied to transform our analysis of human disease. I am grateful to all the authors for their outstanding contributions to this edition. Boston, MA

Catherine J. Wu

Acknowledgments I want to thank my family for their support for all my academic endeavors. I want to also acknowledge the excellent assistance from Diana Ng in preparing this volume.

vii

wwwwwwwwwwwwwwww

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

v xi

Part I Protein-Detecting Analytical Microarrays 1 Detecting and Quantifying Multiple Proteins in Clinical Samples in High-Throughput Using Antibody Microarrays . . . . . . . . . . . . . . . . . . . . . . . . Tanya Knickerbocker and Gavin MacBeath 2 Analysis of Serum Protein Glycosylation with Antibody–Lectin Microarray for High-Throughput Biomarker Screening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chen Li and David M. Lubman 3 Antibody Suspension Bead Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Jochen M. Schwenk and Peter Nilsson 4 Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies . . . . . . . . . Víctor J. Cid, Ekkehard Kauffmann, and María Molina 5 Identification and Optimization of DNA Aptamer Binding Regions Using DNA Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nicholas O. Fischer and Theodore M. Tarasow 6 Recombinant Lectin Microarrays for Glycomic Analysis . . . . . . . . . . . . . . . . . . . . Daniel C. Propheter, Ku-Lung Hsu, and Lara K. Mahal

3

15 29 37

57 67

Part II Antigen Microarrays for Immunoprofiling 7 Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection . . . . . 81 Persis P. Wadia, Bita Sahaf, and David B. Miklos 8 SPOT Synthesis as a Tool to Study Protein–Protein Interactions . . . . . . . . . . . . . 105 Dirk F.H. Winkler, Heiko Andresen, and Kai Hilpert 9 Native Antigen Fractionation Protein Microarrays for Biomarker Discovery . . . . . 129 Robert J. Caiazzo, Jr., Dennis J. O’Rourke, Timothy J. Barder, Bryce P. Nelson, and Brian C.-S. Liu 10 Immunoprofiling Using NAPPA Protein Microarrays . . . . . . . . . . . . . . . . . . . . . . 149 Sahar Sibani and Joshua LaBaer

Part III Protein Function Microarrays 11 High-Throughput Mammalian Two-Hybrid Screening for Protein–Protein Interactions Using Transfected Cell Arrays (CAPPIA) . . . . . . . . . . . . . . . . . . . . . 165 Andrea Fiebitz and Dominique Vanhecke 12 Protein–Protein Interactions: An Application of Tus-Ter Mediated Protein Microarray System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Kalavathy Sitaraman and Deb K. Chatterjee 13 Kinase Substrate Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 Michael G. Smith, Jason Ptacek, and Michael Snyder

ix

x

Contents

14 A Functional Protein Microarray Approach to Characterizing Posttranslational Modifications on Lysine Residues . . . . . . . . . . . . . . . . . . . . . . . 213 Jun Seop Jeong, Hee-Sool Rho, and Heng Zhu

Part IV Strategies for Validation of Candidate Targets 15 Multiplexed Detection of Antibodies Using Programmable Bead Arrays . . . . . . . . 227 Karen S. Anderson 16 A Coprecipitation-Based Validation Methodology for Interactions Identified Using Protein Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 Ovidiu Marina, Jonathan S. Duke-Cohan, and Catherine J. Wu

Part V Generation of Proteomic Libraries 17 Development of Expression-Ready Constructs for Generation of Proteomic Libraries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 Charles Yu, Kenneth H. Wan, Ann S. Hammonds, Mark Stapleton, Joseph W. Carlson, and Susan E. Celniker

Part VI Detection Methods 18 Reverse Phase Protein Microarrays: Fluorometric and Colorimetric Detection . . . 275 Rosa I. Gallagher, Alessandra Silvestri, Emanuel F. Petricoin III, Lance A. Liotta, and Virginia Espina 19 Förster Resonance Energy Transfer Methods for Quantification of Protein–Protein Interactions on Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . 303 Michael Schäferling and Stefan Nagl 20 Label-Free Detection with Surface Plasmon Resonance Imaging . . . . . . . . . . . . . 321 Christopher Lausted, Zhiyuan Hu, and Leroy Hood

Part VII Data Analysis Techniques for Protein Function Microarrays 21 Data Processing and Analysis for Protein Microarrays . . . . . . . . . . . . . . . . . . . . . . 337 David S. DeLuca, Ovidiu Marina, Surajit Ray, Guang Lan Zhang, Catherine J. Wu, and Vladimir Brusic 22 Database Resources for Proteomics-Based Analysis of Cancer . . . . . . . . . . . . . . . . 349 Guang Lan Zhang, David S. DeLuca, and Vladimir Brusic Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365

Contributors Karen S. Anderson • Cancer Vaccine Center, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA Heiko Andresen • Karlsruhe Institute of Technology, Karlsruhe, Germany Timothy J. Barder • Eprogen, Darien, IL, USA Vladimir Brusic • Cancer Vaccine Center, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA Robert J. Caiazzo, Jr. • Molecular Urology Laboratory, Division of Urology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA Joseph W. Carlson • Department of Genome Dynamics, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Susan E. Celniker • Department of Genome Dynamics, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Deb K. Chatterjee • Protein Expression Laboratory, Advanced Technology Program, SAIC-Frederick, Inc., NCI-Frederick, Frederick, MD, USA Víctor J. Cid • Departamento de Microbiología II, Facultad de Farmacia, Universidad Complutense de Madrid, Madrid, Spain David S. DeLuca • Cancer Vaccine Center, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA Jonathan S. Duke-Cohan • Immunobiology Laboratory, Department of Medical Oncology, Dana-Farber Cancer Institute, Boston, MA, USA Virginia Espina • Center for Applied Proteomics and Molecular Medicine, George Mason University, Manassas, VA, USA Andrea Fiebitz • Campus Benjamin Franklin, Charité, Berlin, Germany Nicholas O. Fischer • Physical and Life Sciences Directorate, Lawrence Livermore National Laboratory, Livermore, CA, USA Rosa I. Gallagher • George Mason University, Manassas, VA, USA Ann S. Hammonds • Department of Genome Dynamics, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Kai Hilpert • Karlsruhe Institute of Technology, Karlsruhe, Germany Leroy Hood • Institute for Systems Biology, Seattle, WA, USA Ku-Lung Hsu • Department of Chemistry and Biochemistry, University of Texas at Austin, Austin, TX, USA Zhiyuan Hu • Institute for Systems Biology, Seattle, WA, USA Jun Seop Jeong • Department of Pharmacology and Molecular Sciences, High Throughput Biology Center, Johns Hopkins School of Medicine, Baltimore, MD, USA Ekkehard Kauffmann • Zeptosens – A Division of Bayer (Schweiz) AG-, Witterswil, Switzerland Tanya Knickerbocker • Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA, USA Joshua LaBaer • Virginia G. Piper Center for Personalized Medicine, Biodesign Institute, Arizona State University, Tempe, AZ, USA Lance A. Liotta • George Mason University, Manassas, VA, USA xi

xii

Contributors

Christopher Lausted • Institute for Systems Biology, Seattle, WA, USA Chen Li • Department of Chemistry, The University of Michigan, Ann Arbor, MI, USA Brian C.-S. Liu • Molecular Urology Laboratory, Division of Urology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA David M. Lubman • Department of Chemistry, Comprehensive Cancer Center, The University of Michigan, Ann Arbor, MI, USA; Department of Surgery, The University of Michigan Medical Center, Ann Arbor, MI, USA Gavin MacBeath • Department of Chemistry and Chemical Biology, Harvard University, Cambridge, MA, USA Lara K. Mahal • Department of Chemistry and Biochemistry, University of Texas at Austin, Austin, TX, USA; Department of Chemistry, New York University, New York, NY, USA Ovidiu Marina • Department of Radiation Oncology, William Beaumont Hospital, Royal Oak, MI, USA David B. Miklos • Department of Medicine, Blood and Marrow Transplantation Division, Stanford University, Stanford, CA, USA María Molina • Departamento de Microbiología II, Facultad de Farmacia, Universidad Complutense de Madrid, Madrid, Spain Stefan Nagl • Institute of Analytical Chemistry, University of Leipzig, Leipzig, Germany Bryce P. Nelson • Gentel Biosciences, Inc., Madison, WI, USA Peter Nilsson • Science for Life Laboratory, Department of Proteomics, School of Biotechnology, KTH – Royal Institute of Technology, 10691 Stockholm, Sweden Dennis J. O’Rourke • Molecular Urology Laboratory, Division of Urology, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA, USA Emanuel F. Petricoin III • George Mason University, Manassas, VA, USA Daniel C. Propheter • Department of Chemistry and Biochemistry, University of Texas at Austin, Austin, TX, USA Jason Ptacek • The Baxter Laboratory for Stem Cell Biology, Department of Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA, USA Surajit Ray • Department of Mathematics and Statistics, Boston University, Boston, MA, USA Hee-Sool Rho • Department of Pharmacology and Molecular Sciences, High Throughput Biology Center, Johns Hopkins School of Medicine, Baltimore, MD, USA Bita Sahaf • Department of Medicine, Blood and Marrow Transplantation Division, Stanford University, Stanford, CA, USA Michael Schäferling • Institute of Analytical Chemistry, Chemo- and Biosensors, University of Regensburg, Regensburg, Germany Jochen M. Schwenk • Science for Life Laboratory, Department of Proteomics, School of Biotechnology, KTH – Royal Institute of Technology, 10691 Stockholm, Sweden Sahar Sibani • Virginia G. Piper Center for Personalized Medicine, Biodesign Institute, Arizona State University, Tempe, AZ, USA

Contributors

xiii

Alessandra Silvestri • George Mason University, Manassas, VA, USA; CRO-IRCCS, National Cancer Institute, Aviano, Italy Kalavathy Sitaraman • Protein Expression Laboratory, Advanced Technology Program, SAIC-Frederick, Inc., NCI-Frederick, Frederick, MD, USA Michael G. Smith • Illumina, Inc., San Diego, CA, USA Michael Snyder • Department of Genetics, Stanford University School of Medicine, Stanford, CA, USA Mark Stapleton • NuGEN Technologies, Inc., San Carlos, CA, USA Theodore M. Tarasow • Tethys Bioscience, Inc., Emeryville, CA, USA Dominique Vanhecke • Center for Biomedicine, University Basel, Basel, Switzerland Persis P. Wadia • Department of Medicine, Blood and Marrow Transplantation Division, Stanford University, Stanford, CA, USA Kenneth H. Wan • Department of Genome Dynamics, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Dirk F.H. Winkler • Peptide Facility, Kinexus Bioinformatics Corporation, Vancouver, BC, Canada Catherine J. Wu • Division of Hematologic Neoplasia, Department of Medical Oncology, Cancer Vaccine Center, Dana-Farber Cancer Institute, Boston, MA, USA Charles Yu • Department of Genome Dynamics, Lawrence Berkeley National Laboratory, Berkeley, CA, USA Guang Lan Zhang • Cancer Vaccine Center, Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA, USA Heng Zhu • Departments of Pharmacology and Molecular Sciences and Oncology, High Throughput Biology Center, Johns Hopkins School of Medicine, Baltimore, MD, USA

wwwwwwwww

Part I Protein-Detecting Analytical Microarrays

wwwwwwwwwwwwwwww

Chapter 1 Detecting and Quantifying Multiple Proteins in Clinical Samples in High-Throughput Using Antibody Microarrays Tanya Knickerbocker and Gavin MacBeath Abstract Many diagnostic and prognostic tests performed in the clinic today rely on the sensitive detection and quantification of a single protein, usually by means of an immunoassay. Even in the case of monogenic diseases, however, single markers are often insufficient to provide highly reliable predictions of disease onset, and the accuracy of these predictions only decreases for polygenic diseases and for very early detection or prediction. Recent studies have shown that predictive reliability increases dramatically when multiple markers are analyzed simultaneously. Antibody microarrays provide a powerful way to quantify the abundance of many different proteins simultaneously in a variety of sample types, including serum, urine, and tissue explants. Because the assay is highly miniaturized, very little sample is required and the assay can be performed in high-throughput. Using antibody microarrays, we have been able to identify prognostic markers of early mortality in patients with end-stage renal disease and have built multivariate models based on these markers. We anticipate that antibody microarrays will prove similarly useful in other discovery-based efforts and may ultimately enjoy routine use in clinical labs. Key words: Antibody microarray, Prognosis, Diagnosis, ELISA, Sandwich immunoassay, Highthroughput

1. Introduction Although some diseases can be accurately diagnosed by detecting a single mutation in a gene or by observing elevated serum levels of a single protein marker, most disease states are much more complex. For example, conditions such as high blood pressure, heart disease, or renal failure have both a genetic and environmental component and even diseases such as cancer, which are largely genetic in origin, are often difficult to diagnose using a simple, univariate test. Several recent studies have shown that the accuracy of cancer diagnoses can be enhanced substantially using multivariate approaches based on gene expression profiles (1–3). In addition, Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_1, © Springer Science+Business Media, LLC 2011

3

4

Knickerbocker and MacBeath

multivariate signatures based on DNA polymorphisms (4) or protein levels (5, 6) are proving useful in predicting how patients respond to targeted therapies. To usher in this era of personalized medicine, we need tools that can accurately, sensitively, and simultaneously measure the levels of many different proteins in a variety of clinical samples (serum, urine, and tissue explants). In addition, to enable the discovery of new diagnostic or prognostic signatures, we need methods that are relatively inexpensive and are compatible with high-throughput investigations. Antibody microarrays offer all of these features. They mimic an enzyme-linked immunosorbant assay (ELISA), but in a miniaturized and multiplexed format (Fig. 1). In a typical antibody microarray experiment, a panel of “capture antibodies” is spotted at high spatial density onto a solid support, typically a chemically derivatized glass substrate (Fig. 1a). A clinical sample (e.g., serum) is then applied to the array, and the immobilized antibodies capture

a

b

Cocktail of detection antibodies

Clinical sample (e.g., serum) Cytokines

1. Incubate 2. Wash 3. Add detection antibodies

Glass substrate

Capture antibody

Antibody microarray

d

c

4. Incubate 5. Wash 6. Add labeled secondary antibody

7. Incubate 8. Wash 9. Scan for fluorescence

Fig. 1. Detecting and quantifying multiple proteins in clinical samples using antibody microarrays. (a) Capture antibodies are spotted at high spatial density onto a chemically derivatized glass substrate, where they become immobilized. When a clinical sample (e.g., serum) is applied to the array, each immobilized antibody captures its cognate antigen. (b) After a brief washing step, a cocktail of detection antibodies is applied to the array. Each detection antibody recognizes and binds to its cognate antigen. (c) After a brief washing step, the arrays are incubated with a labeled secondary antibody, which recognizes and binds to all of the detection antibodies. For convenience, the secondary antibody is best labeled with a bright fluorophore, such as PBLX-3. (d) After a final washing step, the arrays are dried and scanned for fluorescence.

Detecting and Quantifying Multiple Proteins

5

their cognate antigens. After a brief washing step, the captured proteins are detected by applying a cocktail of “detection antibodies” (Fig. 1b). To visualize and quantify the detection antibodies, the arrays are again washed and probed with a labeled secondary antibody (Fig. 1c). In a standard ELISA, highly sensitive detection is achieved using an enzyme label, such as horseradish peroxidase, which amplifies the signal by catalytically converting a soluble substrate into a chromophoric product. In an antibody microarray experiment, the final signal must be localized to each spot. A variety of strategies have been developed to achieve highly sensitive detection in a spatially localized fashion. For example, the process of rolling circle replication has been exploited to achieve enzymemediated signal amplification (7, 8). This method enables the detection of many proteins at concentrations as low as 1 pg/mL. We have found, however, that equally sensitive detection can be achieved in a more straightforward fashion without enzymemediated signal amplification using a secondary antibody that has been coupled directly to an extremely bright fluorophore (9). (PBXL-3, a phycobilisome protein complex isolated from red algae and cyanobacteria.) The biggest limitation of antibody microarrays, as well as other multiplexed technologies such as the Luminex® bead-based immunoassay, is the availability of suitable antibodies. Sandwichstyle immunoassays require two highly specific antibodies that recognize distinct, nonoverlapping epitopes on their target proteins. For this reason, most studies using antibody microarray technology have focused on cytokines, chemokines, and other frequently studied serum protein for which high quality, matched pairs of antibodies are commercially available (10). To date, antibody microarrays have been used to discover multivariate signatures for diagnostic purposes. For example, antibody microarrays were recently used to detect differential glycosylation patterns on a variety of serum proteins, which may prove useful for the early detection of pancreatic cancer (11). Similarly, antibody microarrays directed at a large panel of cluster of differentiation (CD) antigens on leukemias and lymphomas from peripheral blood and bone marrow aspirates showed high levels of consistency with diagnoses obtained using conventional clinical and laboratory criteria (12). In our own lab, we have used antibody microarrays to identify prognostic markers of early mortality in patients with endstage renal disease (ESRD) (9). This study serves as an example for how antibody microarrays can be used for discovery purposes. Approximately, 10% of patients with ESRD die within the first 3–4 months of initiating hemodialysis and, to date, no single marker has been found that accurately predicts outcome. We set out to develop a multivariate model that predicts which patients

6

Knickerbocker and MacBeath

are most at risk of dying within the first 15 weeks of initiating treatment. To do this, we collected serum samples from 468 patients initiating dialysis (13). We then assembled a panel of 14 matched pairs of antibodies directed at cytokines and other serum proteins that had previously been associated with ESRD, hypertension, or diabetes (14). To facilitate the rapid and accurate measurement of all 14 proteins in all 468 patient samples, we developed a high-throughput assay in which the capture antibodies were microarrayed in individual wells of 96-well microtiter plates (Fig. 2a). Serum samples were applied to each array and the captured cytokines were detected using a cocktail of biotinylated detection antibodies. The detection antibodies were subsequently visualized and quantified using PBXL-3-labeled streptavidin. Using this simple procedure, we were able to achieve exquisite sensitivity: most cytokines could be detected at a concentration of 1 pg/mL (Fig. 2b). The absolute concentration of each cytokine in each sample was determined by relating the fluorescence intensity of the microarray spots to a standard curve, generated for each cytokine in a multiplexed fashion using one column of each microtiter plate (Fig. 2a, b). For redundancy, each array contained five replicate spots of the capture antibodies and every sample was analyzed on two arrays. Overall, the average coefficient of variation was 6.6% for replicate spots within an array and 11% for replicate samples on separate arrays. Using these microarrays, cytokine levels were measured in all 468 patient samples (Fig. 2c). To develop a multivariate prognostic test, we started by building linear, additive models using logistic regression (9). To avoid overfitting and to construct a model that incorporates only as many variables as are necessary, we adopted the following strategy. If n is the number of variables in the model, we started with n = 1 and, in an incremental fashion, performed an exhaustive search for the best n-variable model. We continued to increment n until no n-variable model could be found in which all of the parameters were statistically significant (P < 0.05 for each cytokine). Based on this criterion, the best model was obtained using three cytokines: angiogenin (Ang), interleukin-12 (IL-12), and vascular cell adhesion molecule-1 (VCAM-1). We then refined our efforts by building generalized additive models (15). As anticipated, the nonparametric models picked up fine features in the relationship between death risk and each cytokine (Fig. 2d). We found that high levels of IL-12 and Ang are associated with low risk of early mortality, whereas increased levels of VCAM-1 are associated with increased risk of death. Interestingly, the three molecular markers are produced by and act on different cell populations. This may explain why a simple additive model is sufficient to capture their associations with early mortality.

a

b

4.5

Log10 { fluorescence }

Detecting and Quantifying Multiple Proteins

4.0

Ang EGF Fet-A ICAM IL-12 IL-1α IL-8 MIP-1β RANTES TNFβ TNFR2 TNFR1 VCAM-1 VEGF

3.5 3.0 2.5 2.0 1.5 1.0 0.5

7

1.0

1.5

2.0

2.5

3.0

3.5

4.0

Log10 { [cytokine] (pg/mL) }

c Outcome

0 −2 −4 200

500 1000 2000 [Ang] (pg/ml)

1 0 −1 −2 1 10 50 [IL-12] (pg/ml)

log-odds of death

2

log-odds of death

d

log-odds of death

Ang EGF Fet-A ICAM IL-12 IL-1α IL-8 MIP-1β RANTES TNF-β TNFR2 TNFR1 VCAM-1 VEGF

2 0 −2

100 1000 10000 [VCAM-1] (pg/ml)

Fig. 2. Serum cytokine levels measured using antibody microarrays. (a) 14 anticytokine capture antibodies were spotted in quintuplicate in each well of a 96-well microtiter plate. Serum samples were applied to each well in columns 1–11 and twofold serial dilutions of a mixture of the 14 cognate cytokines were applied to the wells in column 12. (b) Standard curves generated from the purified cytokines in column 12 of the microtiter plate. (c) Serum cytokine levels of 468 patients initiating hemodialysis. For visualization only, each cytokine was normalized relative to its mean over all the samples and the patients were ordered according to the first principle component of the cytokine profiles. The outcome of each patient is shown at the top (red died with 15 weeks of initiating dialysis; black survived more than 15 weeks). (d) Model built using the cytokine levels that represent the best three-variable model. The solid red lines are the mean of 100 bootstrap samples and the dashed black lines show the variance.

Cytokines acting on the same cell often exhibit synergistic or antagonistic effects (16), but IL-12, Ang, and VCAM-1 are, to a first approximation, independent. We also found in this study that molecular markers are not uniformly prognostic, but instead vary in their value depending on a combination of clinical variables (age, diastolic blood pressure, serum albumin, and method of vascular access) (9). This may explain why previous reports aiming to

8

Knickerbocker and MacBeath

identify prognostic markers without taking into account clinical variables were either conflicting or showed that markers have marginal prognostic value. Just as treatments are now being tailored to specific subsets of patients, our results show that prognosis can also benefit from a personalized approach. We anticipate that antibody microarray technology will play an increasingly important role in biomarker discovery and may ultimately be used on a routine basis in a clinical setting for the purposes of diagnosis, prognosis, patient selection in clinical trials, and theragnostics.

2. Materials 2.1. Buffers (see Note 1)

1. 10× HBS: 100 mM HEPES, 100 mM NaCl, 0.4% NaN3, pH 7.4. 2. Cy3-BSA: Bovine serum albumin (BSA) can be labeled according to manufacturer’s protocol (Amersham CyDye™ Antibody Labeling Kit, Piscataway, NJ). The Cy3-labeled BSA may be stored at 4°C wrapped in aluminum foil for approximately 3 months. 3. Printing Buffer: 1× HBS, 20% glycerol and 0.005 mg/mL Cy3-labeled BSA. This buffer should be freshly prepared from stock (1) for each print. 4. Dilution Buffer: The dilution buffer should be made so that the final concentration of the solution in each well after the addition of both the dilution buffer and the sample is 10 mM in HEPES, 10 mM in NaCl, 0.04% NaN3, pH 7.4. The actual concentrations of reagents in the dilution buffer will vary with the amount of sample added to each well. 5. Wash buffer I: 1× HBS with 1% BSA (w/v). 6. Wash buffer II: 1× HBS with 0.1% Tween-20.

2.2. Additional Equipment and Reagents

1. Aldehyde-displaying glass substrates (112.5 × 74.5 × 1 mm) (Erie Scientific Company, Portsmouth, NH). 2. A piezoelectric or contact microarrayer. 3. Bottomless 96-well microtiter plates (Greiner BioOne, Kremsmünster, Austria). 4. Silicone gaskets (Grace Bio-Labs, Bend, OR). 5. Streptavidin-conjugated Columbia, MD).

PBXL-3

6. A small-orbit orbital shaker. 7. A microarray plate scanner.

(Martek

Biosciences,

Detecting and Quantifying Multiple Proteins

9

3. Methods 3.1. Printing and Storage of the Microarray Plates

1. Reconstitute all monoclonal capture antibodies according to manufacturer’s instructions and then dilute to a final concentration of 0.5 mg/mL in Printing Buffer. 2. Microarray the antibodies on 112.5 × 74.5 × 1 mm aldehydedisplaying glass substrates using a piezoelectric or contact microarrayer (see Notes 2–6). Ninety-six identical microarrays should be fabricated in a 12 × 8 pattern on the glass, with an interarray pitch of 9 mm (to match the spacing of a 96-well microtiter plate). Each array should consist of a regular pattern of spots, with a center-to-center spacing determined empirically for each arrayer. A pitch of 250–350 mm is typical. Between three and five, spots should be printed for each antibody to provide redundant measurements. 3. Attach the glass to the bottom of a bottomless 96-well microtiter plate using an intervening silicone gasket (see Note 7). 4. Seal the arrays with foil and store at −80°C for at least 4 h, but no longer than 6 weeks (see Notes 8 and 9).

3.2. Preparation of the Mixing Plate

1. Prepare a serial dilution series of recombinant antigen in the first row of a low-binding, 96-well microtiter plate. Use Dilution Buffer as the diluent (see Note 10). (a) For samples with a variable composition and low protein concentration (such as urine), add 15% fetal bovine serum (FBS) to both the standard curve and the samples (see Note 11). (b) For samples such as serum or tissue culture supernatant, FBS may be added to the standard curves to ensure a complex environment similar to that of the samples. (c) Appropriate standard curve concentrations will vary with each antibody, but we typically use a 12-point, twofold serial dilution series ranging from 1 ng/mL of each antigen down to 0.5 pg/mL. For particularly abundant antigens, up to 200 ng/mL may be appropriate although many biologically significant antigens are present at lower concentrations in body fluids and cell culture supernatants. 2. In the remainder of the plate, dilute the clinical samples using Dilution Buffer. (a) For most biological samples, start with a 1:1 or 1:4 dilution, depending on the sample volume available. (b) Additional sample dilution(s) may be required for samples with antigens present at high concentrations. All antigen concentrations should be within the range of the standard curve.

10

Knickerbocker and MacBeath

3.3. Preparation of the Microarray Plate

1. In this section, perform all incubations at 4°C on a small-orbit orbital shaker. The shaker should be set at the maximum possible speed so that it does not cause cross-contamination (~400 rpm). Following each incubation, decant the solution by inverting the plate and shaking by hand. 2. Remove the microarray plate from the −80°C freezer and immediately add 300 mL of Wash Buffer I to each well. Incubate for 5 min (see Note 12) then decant the wash solution by inverting the plate and shaking by hand. 3. Repeat twice for a total of three washes. 4. Add 300 mL of Wash Buffer I to each well and incubate for an additional 1 h to block any remaining aldehydes. 5. Remove Wash Buffer I by decanting; then transfer at least 40 mL from each well of the mixing plate to the corresponding well of the microarray plate. Cover the microarray plate with a foil seal and incubate for up to 24 h (see Note 13). 6. Wash the plate three times for 5 min each with Wash Buffer I. 7. After decanting the final wash, add 40 mL of a mixture of biotinylated detection antibodies (0.5 mg/mL in Wash Buffer I) to each well and incubate for 1 h. 8. Decant the solution then wash the plate three times for 5 min each with Wash Buffer I. 9. Add 100 mL of a 4-mg/ml solution of streptavidin-conjugated PBXL-3, prepared in Wash Buffer I, to each well. Incubate for 1 h in the dark. From this point on, minimize exposure to light. 10. After decanting the PBXL-3 solution, wash the plate two times for 5 min each with Wash Buffer I. 11. Wash the plate once with Wash Buffer II. 12. Rinse the plate twice with ddH2O. 13. Centrifuge upside down for 1 min at 1,000 × g to remove residual water.

3.4. Scanning, Image Analysis, and Data Analysis

1. Scan the microarray plates using a scanner that accommodates microtiter plates (e.g., an LS400 scanner, Tecan, Salzburg, Austria) (see Note 14). 2. Using microarray analysis software, quantify the intensity of each spot. Do not use local background correction. Instead, generate a row of phantom spots within each well and subtract the mean intensity of the phantom spots from each microarray spot. 3. To generate the standard curve, plot the log of the mean fluorescence intensity of replicate microarray spots as a function of the log of the cytokine concentration. This should yield a straight line.

Detecting and Quantifying Multiple Proteins

11

4. Relate the mean intensity of replicate spots for each antibody and each clinical sample back to the standard curve to obtain values for the concentration of each cytokine in each clinical sample. 5. Calculate the mean concentration of replicate measurements of each sample (replicate arrays).

4. Notes 1. Unless otherwise noted, all buffers may be stored at room temperature for up to 1 year or until visible signs of contamination appear. 2. Not all antibodies that work for Western blots and other techniques will work on antibody microarrays. Be sure to validate each pair of antibodies using purified antigens. Mix all antigens together and use detection antibodies one at a time and in combination to ensure that detection antibodies do not cross-react with any of the other analytes under investigation. R & D Systems (Minneapolis, MN) is an excellent source of matched pairs of antibodies and their cognate antigens, particularly for the study of cytokines and chemokines. 3. In general, monoclonal antibodies are used as the capture antibody while biotinylated polyclonal antibodies are used for detection. If no monoclonal antibody is available, two polyclonal antibodies may be used. Ideally, these two polyclonal antibodies will have been raised against distinct and nonoverlapping epitopes. 4. When preparing the source plate, mix the antibody in Printing Buffer in an eppendorf tube and then transfer it to the source plate (microtiter plate). This ensures adequate mixing of the solution and increases reproducibility between wells when multiple pins are used to print the same sample. 5. In general, we have found aldehyde-displaying glass surfaces to be more robust and reproducible than epoxide- or aminedisplaying glass, nitrocellulose-coated glass, or hydragels. 6. Pay close attention to the liquid level in the source plate while fabricating microarrays. Even when using a 384-well microtiter plate as a source plate, substantial evaporation can occur during extended print runs. To minimize evaporation, use a cooling block set to between 4 and 10°C, if available, and set the relative humidity at 70–80%. If these options are not available, use an aluminum foil seal with a small hole over each well to allow tip/pin access. Check the liquid level after each print run (or more often, if necessary) and add ddH2O

12

Knickerbocker and MacBeath

as needed to ensure that the antibody concentration remains constant throughout the print run(s). 7. After printing or after assembly of the microarray plate, the arrays may sit at room temperature for several hours without appreciable loss of antibody reactivity. 8. To protect the arrays from freezer burn, they should be sealed, either in a plastic bag or with a foil cover. If using a cover, be sure that it will remain sealed when stored at −80°C. If using a cover that projects into the wells (such as a rubber Storage Mat), attach the cover to the bottomless 96-well microtiter plate before attaching the glass substrate. This ensures that the glass substrate is not pushed off the silicone gasket when the cover is applied to the wells (due to positive pressure). 9. Arrays can be stored for up to 1 year at −80°C, although some loss in activity will occur. Plates that have been stored for several months are best used for assay development purposes. For data collection, plates should be stored for no longer than 6 weeks. 10. For samples with low concentrations of total protein (e.g., urine), preblock the mixing plate with BSA to minimize protein loss. To do this, add enough Dilution Buffer containing 1% BSA (w/v) to completely fill each well, incubate for 1 h at room temperature, and then decant the blocking solution. For samples with very low protein content, all plastics should be rinsed with Dilution Buffer containing 1% BSA (w/v) before contacting the samples. 11. For samples with low concentrations of total protein, add 15% FBS (v/v) to the samples to minimize loss of target proteins. Test each new pair of antibodies to ensure than they do not cross-react with bovine proteins. 12. When removing antibody microarrays from the −80°C freezer, be sure to add Blocking Buffer immediately (within seconds). The buffer usually freezes when added to the wells and then thaws within minutes. Allowing the arrays to warm up, even slightly, results in poor spot morphology (“comet tails,” “coffee rings,” etc.). 13. The best length of time to incubate the antibody microarrays with the samples should be determined empirically. It depends on the concentration of antigens in the sample, the affinities of the capture antibodies for their antigens, and the efficiency of agitation. Incubation times generally range from 1 to 24 h. 14. Antibody microarrays should be scanned at several scanner settings (PMT voltages). For each antigen, use the scan with the highest possible setting that does not include any saturated pixels.

Detecting and Quantifying Multiple Proteins

13

Acknowledgment We thank Ravi Thadhani for directing ArMORR (Accelerated Mortality on Renal Replacement), a prospective study of ESRD patients, and Jiunn-Ren Chen for data analysis and interpretation. This work was supported by awards from the WM Keck Foundation and the Arnold and Mabel Beckman Foundation, and by grants from the National Institutes of Health (DK071674 and DK068465). T.K. is the recipient of an Eli Lilly Graduate Student Fellowship. References 1. Alizadeh AA, Eisen MB, Davis RE et al (2000) Distinct types of diffuse large B-cell lymphoma identified by gene expression profiling. Nature 403:503–511 2. Liang Y, Diehn M, Watson N et al (2005) Gene expression profiling reveals molecularly and clinically distinct subtypes of glioblastoma multiforme. Proc Natl Acad Sci U S A 102:5814–5819 3. Ramaswamy S, Tamayo P, Rifkin R et al (2001) Multiclass cancer diagnosis using tumor gene expression signatures. Proc Natl Acad Sci U S A 98:15149–15154 4. Zhou SF, Di YM, Chan E et al (2008) Clinical pharmacogenetics and potential application in personalized medicine. Curr Drug Metab 9:738–784 5. Duffy MJ, Crown J (2008) A personalized approach to cancer treatment: how biomarkers can help. Clin Chem 54:1770–1779 6. Hanash S (2003) Disease proteomics. Nature 422:226–232 7. Schweitzer B, Roberts S, Grimwade B et al (2002) Multiplexed protein profiling on microarrays by rolling-circle amplification. Nat Biotechnol 20:359–365 8. Shao W, Zhou Z, Laroche I et al (2003) Optimization of rolling-circle amplified protein microarrays for multiplexed protein profiling. J Biomed Biotechnol 5:299–307

9. Knickerbocker T, Chen JR, Thadhani R, MacBeath G (2007) An integrated approach to prognosis using protein microarrays and nonparametric methods. Mol Syst Biol 3(123):1–8 10. MacBeath G (2002) Protein microarrays and proteomics. Nat Genet 32:526–532 11. Li C, Simeone DM, Brenner DE et al (2009) Pancreatic cancer serum detection using a lectin/glyco-antibody array method. J Proteome Res 8:483–492 12. Belov L, Mulligan SP, Barber N et al (2006) Analysis of human leukaemias and lymphomas using extensive immunophenotypes from an antibody microarray. Br J Haematol 135: 184–197 13. Thadhani R, Tonelli M (2006) Cohort studies: marching forward. Clin J Am Soc Nephrol 1:1117–1123 14. USRSD (2005) National Institutes of Health. National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda 15. Buja A, Hastie T, Tibshirani R (1989) Linear smoothers and additive models (with discussion). Ann Statist 17:453–555 16. Natarajan M, Lin KM, Hsueh RC et al (2006) A global analysis of cross-talk in a mammalian cellular signalling network. Nat Cell Biol 8: 571–580

wwwwwwwwwwwwwwww

Chapter 2 Analysis of Serum Protein Glycosylation with Antibody–Lectin Microarray for High-Throughput Biomarker Screening Chen Li and David M. Lubman Abstract The complexity of carbohydrate structures and their derivatives makes the study of the glycome a challenging subset of proteomic research. The microarray platform has become an essential tool to characterize glycan structure and to study glycosylation-related biological interactions, by using probes as a means to interrogate the spotted or captured glycosylated molecules on the arrays. The highthroughput and reproducible nature of microarray platforms have been highlighted by their extensive applications in the field of biomarker validation, where a large number of samples must be analyzed multiple times. This chapter presents an antibody–lectin microarray approach, which allows the efficient, multiplexed study of the glycosylation of multiple individual proteins from complex mixtures with both fluorescence labeling detection and label-free detection based on mass spectrometry. Key words: Microarray, Antibody, Glycoprotein, Biomarker, Serum, Lectin, MALDI, Mass spectrometry

1. Introduction Glycosylation is the most commonly occurring posttranslational modification on proteins involved in numerous biological processes, such as protein–protein interactions, protein folding, immune recognition, cell adhesion, and intercellular signaling. The function of glycoproteins is highly dependent on their carbohydrate structure. The alteration on the glycans is associated with multiple biological events and has been reported in a variety of diseases, especially cancer (1–4). In the search for effective glycosylated biomarkers for targeted diseases, there has been a great deal of effort invested in profiling and characterization of

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_2, © Springer Science+Business Media, LLC 2011

15

16

Li and Lubman

lycoproteins in complex samples. Cell lines, tissue, and other g types of biofluids have been studied by mass spectrometry, fractionation techniques, and microarrays (5–9). Although a microarray assay does not usually provide in-depth structural information on the glycans compared to mass spectrometry, it is able to identify and quantify numerous glycosylation patterns and simultaneously analyze hundreds of samples in a high-throughput manner with excellent reproducibility (9–13). We herein describe an antibody–glycoprotein sandwich assay for high-throughput glycoprotein biomarker screening, where a fluorescent lectin and MALDI-MS are used to quantitatively measure glycosylation levels and identify analytes captured on the antibody arrays, respectively. The scheme for this procedure is illustrated in Fig. 1. The antibodies are first printed on nitrocellulose coated glass slides to generate identical arrays. Printed slides are processed to chemically block the glycans on the antibodies, which are otherwise reactive with lectins used for detection (10). After properly diluted human serum samples are deposited onto the separated antibody arrays, the captured antigens are probed with different lectins with a wide spectrum of binding specificity. The binding of the lectin is measured with a secondary fluorescent dye through a biotin–streptavidin reaction. To verify the effectiveness of previously discovered glycoprotein biomarkers, hundreds of serum samples collected from patients with different disease states are examined in parallel with healthy controls for altered glycosylation patterns. The technical error and bias in the analysis is minimized in several ways, including introducing a control slide to assess spatial variation on a slide and balancing samples from different groups on each slide to reduce experimental bias. MALDI-MS detection has only recently been used to detect peptides on a modified gold surface

Fig. 1. Experimental scheme using lectin and MALDI-MS detection with antibody microarray to analyze glycosylation of serum glycoproteins.

Analysis of Serum Protein Glycosylation

17

coated with antibodies (14). An on-slide digestion method, developed in our previous work (15), exploited the utility of MALDI-MS to identify antibody-captured proteins. The whole digestion, including automatic trypsin spotting and incubation, requires less than 10 min. While the antibody–lectin sandwich microarray provides a means to measure glycosylation changes on specific proteins captured from complex samples using lectin probes in a high-throughput array format, fluorescence-based detection provides limited structural information and cannot distinguish some glycoforms that have similar affinity with lectins, such as (GlcNAc)2(Man)8 and (GlcNAc)2(Man)9. Therefore, the MALDI-MS detection of the tryptic products of the captured protein on the antibody array serves as a complementary technique to verify the identity of the target of the antibody and a means to monitor the nonspecific binding so as to optimize the dilution fold for the experiment. As such, mass spectrometry is a powerful alternative to fluorescent detection, as it confirms the identity of the captured analyte and detects any undesired binding.

2. Materials 2.1. Antibody–Lectin Microarray with Fluorescence Detection

1. Monoclonal antibodies, for serum amyloid P component (SAP; Abcam), Alpha-1-beta glycoprotein (A1BG; Abnova), Antithrombin III (Abcam).

2.1.1. Printing

3. Nova nitrocellulose slides (GraceBio), PATH nitrocellulose slides (Gentel).

2. Nanoplotter 2.0 (GeSiM).

4. Printing buffer: 30% phosphate buffering saline (PBS), concentration of antibody diluted by water to 0.3 mg/mL. 5. 96-Well sample plate (BioRad). 2.1.2. Antibody Blocking

1. Washing buffer: PBS-T 0.1 (0.1% Tween-20). 2. Coupling buffer: 0.02 M sodium acetate, pH 5.5. 3. Oxidation buffer: 0.2 M sodium peroidate in coupling buffer. 4. 4-(4-N-maleimidophenyl)butyric acid hydrazide hydrochloride (MPBH) (Pierce), Cys–Gly (Sigma).

2.1.3. Hybridization of Slides

1. Blocking buffer: 1% w/v BSA in PBS-T 0.5 (0.5% Tween-20). 2. Sample buffer: 0.1% Brij-45, 0.1% Tween-20 in PBS. 3. Primary detection solution: For Aleuria aurentia (AAL), Maackia amurensis (MAL), Lens culinaris agglutinin (LCA),

18

Li and Lubman

and Sambuccus Nigra (SNA) – 10 mg/mL biotinylated lectin solution in PBS-T 0.1; and for Concanavalin A (ConA), 1 mg/mL lectin in PBS-T 0.1. All biotinylated lectins were purchased from Vector Laboratories (Burlingame, CA). 4. Washing buffer: PBS-T 0.1 (0.1% Tween-20). 5. Secondary detection solution: 1:1,000 solution of 1 mg/mL Streptavidin conjugated to Alexafluor555 (Invitrogen) in PBS-T 0.1. 6. Speedvac. 7. SIMplex Multiplexing system (Gentel). 2.1.4. Slide Scanning

1. Axon 4000A scanner (Molecular Devices, Sunnyvale, CA).

2.1.5. On-Slide Digestion

1. Sequencing grade modified trypsin (Sigma). 2. Acetonitrile (ACN). 3. Ammonia bicarbonate. 4. Oven. 5. Nanoplotter 2.0 (GeSiM). 6. Wetted paper box.

2.1.6. MALDI-QIT-TOF

1. MALDI-QIT-TOF (Shimadzu Biotech, Manchester, UK). 2. Trifluoroacetic acid (TFA). 3. 2,5-Dihydroxybenzoic acid (DHB), prepare 10 mg/mL solution in 50% ACN, 0.1% TFA. 4. Stainless steel plate adaptor.

3. Methods 3.1. Antibody Array Printing

The number of antibody arrays that can be printed on each slide is determined by the size of the arrays. The most popular format involves 16 coated pads on a standard 1 × 3 in. slide. Each pad is able to contain more than 9 × 9 spots with 0.6 mm spacing. For MALDI-MS detection, the sensitivity is much lower than fluorescence. Therefore to generate a spectrum with good S/N, additional sample needs to be printed on each spot. 1. Antibodies are diluted to 0.5 mg/mL in printing buffer and transferred to a 96-well sample plate. 2. Edit the spot layout in the NanoPlotter program to produce a 2 × 7 format of identical arrays with a 9 mm row and column distance from each other. The spacing between the spots is 0.6 mm. Each antibody is printed in triplicate. For MALDI-MS detection, the spacing between the spots is 1.5 mm.

Analysis of Serum Protein Glycosylation

19

3. Antibody solution is spotted onto nine ultrathin nitrocellulose coated slides. The first slide is discarded because of high variation of printing; the other eight are used for the experiment. Each spotting event results in 500 pL of sample being deposited and is programmed to occur 5 times/spot to ensure that 2.5 nL is being spotted per sample. The spot diameter is around 250 mm. For MALDI-MS detection, the amount of antibody on each of the spots in the antibody array is increased from 5 to 100 droplets. The spot diameter is around 700 mm (see Note 1). 3.2. Antibody–Lectin Array with Fluorescence Detection 3.2.1. Antibody Array Blocking

The IgG antibodies are usually glycosylated (15). The antibody glycans are reactive to detection lectins, thus need to be modified. To prevent the reaction between the antibody glycan and lectin, the antibodies on the slides are chemically derivatized with a modified method described in the previous work of Haab (10). 1. The printed slides are dried at room temperature overnight before gently being washed with PBS-T 0.1 and incubated in coupling buffer with 0.1% Tween 20 for 10 min. The slides are washed again with coupling buffer without Tween 20 before oxidation (see Note 2). 2. The slides are incubated in freshly made oxidation solution at 4°C in the dark. After 3 h the slides are removed from the oxidizing solution and rinsed with coupling buffer with 0.1% Tween 20 until the white precipitation disappears. The washing usually takes 30–60 min (see Note 3). 3. The slides are immersed in fresh 1 mM MPBH (in coupling buffer) at room temperature for 2 h to derivatize the carbonyl groups, then incubated with 1 mM Cys-Gly (in PBS-T 0.1) at 4°C overnight to stabilize the −SH group on MPBH. The slides are subsequently blocked with blocking buffer for 1 h and dried by spinning the slide at 1,000 rpm in a centrifuge (see Note 4).

3.2.2. Optimizing Conditions

Before screening a large number of samples, the optimum concentration of the serum is determined by a serial dilution test. In the dilution test, serum is diluted with sample buffer by 2–600 folds and incubated with different blocks of antibody arrays on a single slide (details of the experiment in Subheading 3.2.4). The signal is detected by lectin SNA (or any lectin) and plotted in Fig. 2. The figure depicts how the intensity of the signal changes for three antibodies (against Serum Amyloid P component, A1BG, and Antithrombin III) with decreasing dilution fold. A rising trend was noted from the 600× dilution to the 50× dilution for the three glycoproteins shown. In the 50× dilution to the 20× dilution, the signal was relatively unchanged

20

Li and Lubman

Fig. 2. Saturation curve showing how the antibodies (against serum Amyloid C component, A1BG, Antithrombin III) respond to different dilution of serum with SNA lectin detection. X-axis shows fold serum dilution before hybridization on the antibody array. The y-axis is the intensity of the signal. Reprinted with permission from Li et al. (15).

except for Antithrombin III, where the signal increased 20% from the 50× dilution to the 20×. The signal remained the same from the 20× dilution until it reached the 5× dilution, where a saturation of the signal has occurred. A decrease of signal for all three glycoproteins from the 5× dilution to the 2× dilution of serum sample can be seen in Fig. 3, likely due to competing nonspecific binding on the antibodies. The result of the dilution test demonstrates that the antibodies were saturated by their target protein at 20× dilution or above in the process of hybridization. Below 50× dilution, the antibodies were not completely occupied so the signal decreased with additional dilution. The nonlinear relation between the concentration of the serum and the intensity of the signal could be attributed to various factors that may affect the antibody–antigen reaction, including accessibility of the antibodies, diffusion rate, and solubility of the antigen in the hybridization buffer. Nonspecific binding on the antibodies was also considered as a possibility, but was further investigated and excluded by on-target digestion and MALDI-MS analysis. To analyze the difference of the glycosylation on potential biomarker proteins, protein expression levels must be normalized. Under saturation conditions, the amount of target biomarkers captured on the antibody spots was equal to the capacity of the printed antibody which should be the same in all the replicate blocks. As a result, protein assay is no longer needed and the intensity of the signal on the microarray directly represents the level of glycosylation.

Analysis of Serum Protein Glycosylation

21

a G1 G2 G4 G1 G3 G4 G2 G3 C1 C1 G1 G2 G4 G1 G3 G4 G3 G4 G1 G2 C1 C2 G3 G4 G1 G2 G3 B Slide 1

G2 G3 G4 G1 C1 C2 G2 G3 G4 G1 G2 B Slide 2

G1 G2 G3 G4 C1 C2 G1 G2 G3 G4 G1 B Slide 3

G4 G1 G2 G3 C1 C2 G4 G1 G2 G3 G4 B Slide 4

C1 C1 C1 C1 C1 C1 C1 C1 C1 C1 C1 C1 Slide 5

G3 G4 G1 G2 C1 C2 G3 G4 G1 G2 G3 B Slide 6

G2 G3 G4 G1 C1 C2 G2 G3 G4 G1 G2 B Slide 7

G1 G2 G3 G4 C1 C2 G1 G2 G3 G4 G1 B Slide 8

b

Fig. 3. Parallel processing of 77 samples on eight slides. (a) Sample arrangement on eight slides. G1, G2, G3, and G4 are four different groups of samples. Control samples are C1 and C2. B is blank. (b) A picture of SIMplex multiwell device.

3.2.3. Experimental Design

In the high-throughput biomarker screening, we usually parallel print and process eight slides which contain 112 identical blocks of antibody array. To minimize the technical error and bias on these blocks, serum samples are arranged to balance different disease/healthy groups and reference blocks are also introduced to adjust to signals of different blocks and slides. We provide an example of how to arrange samples on slides to minimize experimental biases in Fig. 3. 1. The slides are labeled from 1 to 8 in their printing order (see Note 2). 2. Slide 5 is used as a control slide; all the blocks on the control slide are incubated with a control serum sample C1. 3. Block 7 and block 8 on each slide except slide 5 are used as control blocks; they are incubated with control samples C1 and C2, respectively. 4. Block 14 is used as blank and incubated with sample buffer only. 5. The other 77 blocks are incubated with 19 samples from each of the four disease groups and 1 extra sample from a random group in a designated order to balance the number of samples from each group on any particular block (Fig. 3a).

22

Li and Lubman

3.2.4. Hybridization of Slides

1. The slides are placed into the SIMplex (Gentel) Multiplexing device which has 16 wells for each slide (the bottom two wells are not used) to separate the antibody arrays and prevent cross contamination between adjacent wells (Fig. 3b). 2. Serum samples are aliquoted into a volume of 10 mL in each vial and diluted 10× with 90 mL sample buffer. Diluted samples are added into the wells of the SIMplex Multiplexing device and incubated for 1 h with gentle shaking at room temperature. The wells must be sealed to prevent evaporation of samples (see Notes 5 and 6). 3. After completion of serum hybridization, slides are rinsed with PBS-T 0.1 three times to remove unbound proteins. The slides are incubated with biotinylated lectin solution in a plastic box with gentle shaking for an hour at room temperature. 4. The slides are washed 3 times with PBS-T 0.1 and incubated with secondary detection solution with gentle shaking for an hour at room temperature. 5. The slides are again washed 3 times with PBS-T 0.1, dried by centrifuge and kept at 4°C before scanning.

3.2.5. Slide Scanning

1. The dried slides are scanned with an Axon 4000A scanner. 2. Alexa555 labeled slides are scanned in the green channel (wavelength 545 nm). The photomultiplier tube (PMT) gain should be adjusted to obtain the best S/N without saturation. The size of the pixel of the image is 10 mm. 3. The program Genepix Pro 6.0 is used to extract the numerical data.

3.2.6. Data Analysis

The nonbiological variation between blocks on the same slide is termed as on-slide variation. This variation is mainly generated by antibody printing and slide scanning and its feature is that every slide follows the same pattern (i.e., the blocks at the top of the slides are brighter than the bottom ones). The blocks on the control slide incubated with the same control sample are thus used to estimate the on-slide variation and calculate adjustment index for all the blocks. The slide-to-slide variation is considered as specific changes of the signal that effect all the blocks on a single slide. This variation is estimated by control blocks on each of the slides. The data is adjusted by a second index calculated by comparing the signal of the control blocks to exclude the slide-to-slide variation. An example of assaying glycosylation expression of A1BG is shown in Fig. 4. Lectin SNA is used to probe the sialic acid present at the termini of the glycans of this protein. As shown in this figure, the mean value of the cancer samples is significantly higher than the other three groups (p < 0.05).

Analysis of Serum Protein Glycosylation

23

Fig. 4. Distribution of sialylation levels detected by lectin SNA on A1BG. The spots present the signal of the glycan on captured antigen for individual samples from different classes. The long and short lines give the mean value and the standard error of the mean, respectively.

1. A threshold of signal-to-background ratio is set at 3 and spots that are under this threshold are excluded. 2. The background-subtracted median of the intensity for the triplicates of each antibody is averaged and taken as a single data point into analysis. 3. On-slide variation index for antibody 1 in block 1 equals to the average signal of antibody 1 over all the blocks on the control slide divided by the signal of antibody 1 in block 1.

I Ab1.B1 = Avg Ab1.CS / SAb1.B1 . 4. Slide-to-slide variation index for antibody 1 on slide 1 is calculated as follows: AvgAb1.S1 is the average signal of antibody 1 on slide 1. AvgAb1.AS is the average signal of antibody 1 on all the slides.

I Ab1.S1 S = Avg Ab1.A / Avg Ab1.S1 . 5. The final adjusted signal is calculated by the following formula: SAb1.B1.S1 is the raw signal, SAb1.B1.S1.ad is the adjusted signal.

SAb1.B1.S1.ad = SAb1.B1.S1 * I Ab1.B1 * I Ab1.S1 . 6. For each antibody the signal can be normalized to one for easy comparison, SAb1.B1.S1.n is the normalized signal.

SAb1.B1.S1.n = SAb1.B1.S1.ad /Avg Ab1.

24

Li and Lubman

3.3. On-Slide Digestion and MALDI Sample Preparation

Nonspecific binding on antibodies may occur when the microarray is exposed to a concentrated and complex protein mixture such as serum. A commonly used method to study the specificity of an antibody is to digest and identify the protein released from antibody-conjugated medium, whereas eluting the captured protein is not very efficient and the procedure includes four or more steps. Thin layers of a conductive metal oxide and nitrocellulose make the surface of PATH slide perfect for MALDI. We developed an on-slide digestion and MALDI sample preparation protocol using the NanoPlotter to precisely spot enzyme and matrix to antibody arrays on the slide after the serum hybridization. Antibody arrays exposed to differently diluted sera are analyzed by this method to see if nonspecific binding occurs. Trypsin spotted on the antibody array usually simultaneously digests both the captured protein and the antibody; hence the tryptic peptides of the antibody must be excluded from the mass spectra for us to choose the peaks of interest. In an example, we prepared three identical spots of SAP antibody in separated blocks, which were then incubated with sample buffer (as control), 10× diluted serum, and 2× diluted serum and subjected to on-slide digestion and MALDI-MS. The MALDI-MS spectra of the three spots are shown in Fig. 5. The peaks that appear in the spectrum of the control spot are considered to be peptides of the antibody. The three highest peaks between 1,150 and 1,250 were identified by MS/MS as peptides from the Fc region of mouse IgG. In spectrum b where the antibody spot was hybridized with 10× diluted serum, the peaks at 1,166 and 1,407 m/z, are identified by MS/MS as the peptides digested from the target antigen, and the peak 993 matches the mass of a tryptic peptide of SAP. In the spectrum c there are two additional peaks. One of these was identified as human albumin, while the other one could not be identified or matched with a peptide mass of the target antigen. The additional peaks indicate that nonspecific binding might have occurred to the antibody spot. The serum was further diluted to assess the detection limit of the MALDI-MS technique. At 500× dilution (data not shown), the peak at 1,166 m/z disappeared while the 1,407 m/z still showed a signal-to-noise ratio of 2–3. Thus, the 500× dilution is considered as the detection limit of SAP, which is present in human serum with a concentration of around 30 mg/mL (16). The introduction of mass spectrometry based label-free detection has the potential to further characterize the glycan structure. However, due to the presence of the tryptic peptides of the antibody and the lack of a glycopeptide enrichment step, only a limited number of the nonglycosylated peptides of the antigen could be seen in the spectra. To improve the MALDI-MS detection of the targeted antigen and its glycopeptides, we are searching for other chemical strategies to block the tryptic digestion of the antibody and enrichment methods to selectively ionize the glycopeptides.

Analysis of Serum Protein Glycosylation

25

Fig. 5. The MALDI-MS spectra generated on the microarray spots of Amyloid p component antibody after on-target digestion. The peaks identified as Amyloid p component were marked with bold arrows where the extra peaks appearing in (c) were marked with regular arrows. (a) Control spot, without incubation of serum; (b) incubated with10× diluted serum; (c) incubated with 2× diluted serum. Reprinted with permission from Li et al. (15).

1. Antibody slide is printed and hybridized with diluted serum as described above. 2. Trypsin is diluted with 50 mM ammonium bicarbonate in 20% ACN and kept on ice before use. 3. Keep the humidity of the Nanoplotter chamber higher than 70% (use a humidifier or lay a wet paper towel on the deck).

26

Li and Lubman

4. In the program, set the same spot layout on the slide, print 100 droplets (0.5 nL per droplet) of trypsin on each spot (see Note 7). 5. Move the printed slide to a wet paper box and incubate them in an oven at 37°C for 5 min. Make sure the trypsin solution does not dry out on the spots. 6. Take the slide out from the oven, print the DHB solution on the slide with the same spot layout (50 droplets per spot). 3.4. M ALDI-MS

1. Tape the slide onto a stainless steel MALDI plate adaptor, insert it into the MALDI-MS instrument. 2. Mass spectrometric analysis of the microarray slides was performed using the Axima quadrupole ion trap-TOF. Acquisition and data processing were controlled by Launchpad software (Kratos, Manchester, UK). A pulsed N2 laser light (337 nm) with a pulse rate of 5 Hz was used for ionization. Each profile resulted from two laser shots. Argon was used as the collision gas for CID and helium was used for cooling the trapped ions. 3. TOF was externally calibrated using 500 fmol/mL of bradykinin fragment 1–7 (757.40 m/z), angiotensin II (1046.54 m/z), P14R (1533.86 m/z), and ACTH (2465.20 m/z) (SigmaAldrich). The mass accuracy of the measurement under these conditions was 50 ppm. 4. The power of the laser is set at 80 to ionize the spots on the microarray. The focus of the laser can be moved from spot to spot manually under the camera or by using the Raster function to set up an automatic scan for all the spots.

4. Notes 1. When the pin on the Nanoplotter is in poor condition or the instrument is not set up correctly, the quality of antibody printing may fluctuate or gradually worsen as the printing continues. Sticky components, such as glycerol, in the antibody printing solution may also cause unstable printing. A simple test can be done in advance to assess the performance of the pin. Print 1,000 spots with a random antibody on a transparent slide. Observe the residue after the spots are dried. If the residues are in an intact round shape and their sizes and colors do not vary significantly, then the printing is acceptable, otherwise the printer needs to be checked or the printing solution must be changed.

Analysis of Serum Protein Glycosylation

27

2. Many types of chemicals can contaminate the nitrocellulose coating on the slide, resulting in increased background. The slides should not be labeled with any kind of marker. A disposable plastic box is a very good container for slide washing. 3. In the glycan blocking procedure, after the antibody is oxidized by NaIO4, white precipitation forms on the slides. This precipitation must be completely washed away before moving on to the next step. 4. Blocked slides should not be kept in solution for too long, while dried ones can be stored at 4°C for a long period of time. 5. Serum sample must be aliquoted immediately upon arrival and stored at −80°C. Serum frozen and thawed more than twice should not be used. When the sample set consists of multiple groups, all the samples must be in the same frozen and thaw cycle for bias-free comparison. 6. All the incubation should be done with gentle shaking to prevent uneven binding. 7. The higher number of droplets of antibody solution printed on the slides does not result in a higher density of antibody on the spot because the coating of the PATH slides is so thin that a few droplets are able to saturate the surface. The concern for the minimum amount of antibody solution printed on each spot is position variation, i.e., repeated printings on the same spot do not perfectly overlap. Printing 100 droplets of antibody solution produces a larger spot size which guarantees a certain area of overlap between the antibody spot and the printing of trypsin and matrix.

Acknowledgements Our work on microarray development described herein has been supported in part under grants from the National Cancer Institute under grant NCI R21 12441, R01 CA106402. This work has also received partial support from the National Institutes of Health under R01GM49500. We would like to thank Dr. Brian Haab and Dr. Chen Songming of the Van Andel Institute for sharing with us the procedures of preparing the antibody arrays. We would also like to thank Stephanie Laurinec, Jes Pedroza, and Missy Tuck for collection of the samples used in this work.

28

Li and Lubman

References 1. Rudd PM, Elliott T, Cresswell P, Wilson IA, Dwek RA (2001) Glycosylation and the immune system. Science 291:2370–2376 2. Kobata A, Amano J (2005) Altered glycosylation of proteins produced by malignant cells, and application for the diagnosis and immunotherapy of tumours. Immunol Cell Biol 83:429–439 3. Gessner P, Riedl S, Quentmaier A et al (1993) (1993) Enhanced activity of cmp-newac-galbeta-1–4glcnac-alpha-2, 6-sialyltransferase in metastasizing human colorectal tumor-tissue and serum of tumor patients. Cancer Lett 75:143–149 4. Gorelik E, Galili U, Raz A (2001) On the role of cell surface carbohydrates and their binding proteins (lectins) in tumor metastasis. Cancer Metastasis Rev 20:245–277 5. Zhao J, Simeone DM, Heidt D, Anderson MA, Lubman DM (2006) Comparative serum glycoproteomics using lectin selected sialic acid glycoproteins with mass spectrometric analysis: application to pancreatic cancer serum. J Proteome Res 5:1792–1802 6. Ressom HW, Varghese RS, Goldman L et al (2008) Analysis of MALDI-TOF mass spectrometry data for discovery of peptide and glycan biomarkers of hepatocellular carcinoma. J Proteome Res 7:603–610 7. An HJ, Peavy TR, Hedrick JL et al (2003) (2003) Determination of N-glycosylation sites and site heterogeneity in glycoproteins. Anal Chem 75:5628–5637 8. Block TM, Comunale MA, Lowman M et al (2005) Use of targeted glycoproteomics to identify serum glycoproteins that correlate with liver cancer in woodchucks and humans. Proc Natl Acad Sci U S A 102:779–784

9. Patwa TH, Zhao J, Anderson MA, Simone DM et al (2006) Screening of glycosylation patterns in serum using natural glycoprotein microarrays and multi-lectin fluorescence detection. Anal Chem 78:6411–6421 10. Chen SM, LaRoche T, Hamelinck D et al (2007) Multiplexed analysis of glycan variation on native proteins captured by antibody microarrays. Nat Methods 5:437–444 11. Zhao J, Patwa TH, Qiu WL et al (2007) Glycoprotein microarray with multi-lectin detection: unique lectin binding patterns as tools for classifying normal, chronic pancreatitis, and pancreatic cancer sera. J Proteome Res 5:1864–1874 12. Wu YM, Nowack DD, Omenn GS et al (2008) Mucin glycosylation is altered by pro-inflammatory signaling in pancreatic-cancer cells. Pancreas 37:502 13. Yue TT, Goldstein IJ, Hollingsworth MA et al (2009) The prevalence and nature of glycan alterations on specific proteins in pancreatic cancer patients revealed using antibody-lectin sandwich arrays. Mol Cell Proteomics 7:1697–1707 14. Evans-Nguyen KM, Tao SC, Zhu H et al (2008) Protein arrays on patterned porous gold substrates interrogated with mass spectrometry: detection of peptides in plasma. Anal Chem 5:1448–1458 15. Li C, Simeone DM, Brenner DE et al (2009) Pancreatic cancer serum detection using a lectin/glyco-antibody array method. J Proteome Res 8:483–492 16. Nyboa M, Olsenb H, Jeuneb B et al (1998) Increased plasma concentration of serum amyloid P component in centenarians with impaired cognitive performance. Dement Geriatr Cogn 9:126–129

Chapter 3 Antibody Suspension Bead Arrays Jochen M. Schwenk and Peter Nilsson Abstract Alongside the increasing availability of affinity reagents, antibody microarrays have been developed to become a powerful tool to screen for target proteins in complex samples. Besides multiplexed sandwich immunoassays, the application of directly applying labeled sample onto arrays with immobilized capture reagents offers an approach to facilitate a systematic, high-throughput analysis of body fluids such as serum or plasma. An alternative to commonly used planar arrays has become available in form of a system based on color-coded beads for the creation of antibody arrays in suspension. The assay procedure offers an uncomplicated option to screen larger numbers of serum or plasma samples with variable sets of capture reagents. In addition, the established procedure of whole sample biotinylation circumvents the purification steps, which are generally required to remove excess labeling substance. We have shown that this assay system allows detecting proteins down into lower pico-molar and higher picogram per milliliter levels with dynamic ranges over three orders of magnitude. Presently, this workflow enables the profiling of 384 clinical samples for up to 100 proteins per assay. Key words: Suspension bead array, Antibody array, Serum, Plasma, Labeling

1. Introduction The exploration of the human proteome is one of the major challenges of the postgenomics era, focusing on a better understanding of disease-related processes (1). Recent developments of miniaturized and parallelized technology platforms now offer affinity-based alternatives to widely used mass spectrometric analysis. Among these methods, various protein microarrays have been implemented into proteomic profiling approaches demonstrating their applicability in high-throughput screening for marker proteins in patient samples (2). Two alternative formats have been developed; reverse-phase microarrays, where large numbers of lysates from tissues and cells or serum samples are spotted onto array surfaces for the parallel analysis of a single

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_3, © Springer Science+Business Media, LLC 2011

29

30

Schwenk and Nilsson

parameter, and the forward-phase setting, such as multiplexed sandwich immunoassays or antibody arrays, which both utilize immobilized capture reagents to analyze many parameters (3). While dedicated robotic devices, which arrange molecules on microscopic slides with functionalized surfaces, are needed produce planar protein microarrays, alternative platforms have been employed for a parallelized and miniaturized analysis. One of these is based on a flow cytometeric system that currently allows to determine the identity of up to 500 color-coded micrometer sized beads in cooccurrence to protein interaction dependent reporter fluorescence (4). Arrays are thereby created in suspension by mixing beads with different codes, denoted here as bead IDs, and immobilized capturing reagents. This platform has recently been utilized to adapt the concept of antibody arrays from previously described planar arrays (5). The described workflow, summarized in Fig. 1, offers a microtiter plate-based alternative to methods based on planar microarrays for the analysis of labeled serum and plasma protein profiling and can be used for highly multiplexing in both the dimension of parameters measured per sample as well as samples studied per analysis. An example of a protein profile obtained from this approach is given in Fig. 2. Here, intensity levels over more that two orders of magnitude and a low intensity variability of £20% are observed.

2. Materials 2.1. B ead Coupling

1. Beads: MagPlex or MicroPlex microspheres (Luminex Corp). 2. Activation buffer (1×): 100 mM Monobasic Sodium Phosphate (Sigma), pH 6.2, stored at +4°C for up to 3 months and at −20°C for long term. 3. EDC solution: 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride (EDC, Pierce), aliquoted in screwcapped tubes and stored at +4°C. Dissolve in activation buffer to 50 mg/ml directly prior usage. 4. S-NHS solution: 50 mg/ml Sulfo-N-Hydroxysuccinimide (NHS, Pierce), prepared as aliquots in screw-capped tubes and stored at −20°C. Dissolve in activation buffer to final concentration directly prior usage. 5. Coupling buffer: 100 mM 2-(N-morpholino)ethanesulfonic acid (MES) pH 5.0, stored at +4°C for up to 3 months and at −20°C for long term. 6. Wash buffer: 0.05% (v/v) Tween20 in 1× PBS pH 7.4 (PBST). 7. Antibody detection solution: 0.25 mg/ml R-Phycoerythrin modified antispecies antibodies (e.g., Jackson), diluted to this concentration in PBST (see Note 1).

Antibody Suspension Bead Arrays

31

Fig. 1. Workflow overview.

2.2. S ample Labeling

1. Sample dilution buffer: 1× PBS pH 7.4. 2. Labeling solution: 10 mg/ml Sulfo-N-Hydroxysuccinimidepolyethylene oxide biotin (NHS-PEO4-Biotin, Pierce), dissolved in dimethyl sulfoxide (DMSO, Sigma) directly before use. 3. Stop solution: 1 M Tris–HCl pH 8.0, stored at +4°C and added cold.

2.3. A ssay Procedure

1. Assay buffer (1×): 0.1% (w/v) casein, 0.5% (w/v) polyvinylalcohol, and 0.8% (w/v) polyvinylpyrrolidone (all Sigma), prepared in PBST and stored at +4°C for up to 3 months and at −20°C for the long term. Supplement before use with 0.5 mg/ml rabbit IgG (Bethyl).

32

Schwenk and Nilsson

MFI [AU]

10000 1000 100 10

Ab -65

Ab -60

Ab -55

Ab -50

Ab -45

Ab -40

Ab -35

Ab -30

Ab -25

Ab -20

Ab -15

Ab -10

Ab -05

Ab -01

1

Antibodies

Fig. 2. Intensity profile of a plasma sample. A bead mixture composed of 68 antibodies was employed to determine intensity levels for the targeted proteins in a plasma sample. Such profiles typically cover intensity range over more than two orders of magnitude (50–20,000 AU). Standard deviations of £20% can be commonly obtained from replicates.

2. Stop solution (4×): 4% paraformaldehyde (PFA) solution, to store at +4°C. Dilute 1:4 in PBS prior to usage. 3. Detection solution: R-Phycoerythrin modified streptavidin (Invitrogen) diluted to 0.5 mg/ml in PBST directly before use and protected from light.

3. Methods 3.1. B ead Coupling

In the following, a method for antibody coupling is described, for which magnetic and nonmagnetic beads can be utilized. The main difference between these two bead types is the handling of the beads during an exchange of surrounding liquid solution. For coupling quantities not exceeding the amount of positions found in bench top microcentrifuges, we suggest using microcentrifuge tubes or tubes with filter inserts to pellet the beads via centrifugation, while magnetic beads can additionally be manipulated by magnetic forces without centrifugation. For more than 24 couplings in parallel, microtiter plate based protocols are preferred. Hereby, proteins can be immobilized on nonmagnetic beads in filter bottomed microtiter plates (Millipore) with a filter pore sizes below bead diameter and vacuum devices (Millipore) accommodate these plates to remove liquid. For magnetic bead coupling in plates, dedicated plate magnets are available (LifeSept, Dexter Magnetic Technologies) to facilitate bead sedimentation and fixation. 1. Prepare antibodies at the desired concentration (e.g., 3 mg or a solution with antibody concentration of 30 mg/ml per 1 × 106 beads) in coupling buffer (see Note 2).

Antibody Suspension Bead Arrays

33

2. The beads are to be distributed in desired portions (e.g., 80 ml = 1 × 106 beads) into the wells of a half-area plate and the beads are washed with 3× 100 ml activation buffer (see Note 3). 3. Prepare fresh solutions of NHS and EDC, both at 50 mg/ml in activation buffer. Prepare 0.5 mg of each chemical per bead ID and coupling, and add 10 ml NHS, 10 ml EDC, and 80 ml activation buffer to each bead ID. 4. Incubate 20 min under continuous, gentle shaking, and wash thereafter with 3× 100 ml coupling buffer. 5. Continue without interruption (see Note 4) by adding the antibody solution to the activated beads and incubate for 2 h under continuous, gentle shaking. 6. The beads are washed 3× with 100 ml wash buffer. 7. The beads are then recovered from the wells into microcentrifuge tubes with 3× 100 ml wash buffer. The liquid is removed and 100 ml storage buffer is added prior to the bead storage at +4°C in the dark for at least 1 h. 3.2. Bead Mixture Preparation

The yield of antibodies immobilized on beads should be judged after the coupling. To allow a balanced and economic amount of beads to be applied and counted during the measurements, equal numbers of beads should be combined in a bead mixture. To facilitate this, the beads can be counted and an initial bead concentration can be determined which allows calculating the required volumes to be added in a common stock solution. During this bead counting procedure, the rate of antibody immobilization can be additionally approximated via fluorescently labeled antispecies specific antibodies. 1. The tubes with antibody-coupled beads are to be vortexed and sonicated for 5 min. 2. Each bead solution is diluted 1/100 in antibody detection solution (see Notes 1 and 5) in a microtiter plate. 3. The plates are incubated for 20 min and measured. 4. The number of counts per bead ID is multiplied by a correction factor of 3.3 for a 1/100 dilution to obtain a first estimation of beads per microliter storage solution. From this number the volumes of beads in storage solution can be calculated which are to be applied into the bead mixture. The required number of beads supplied should be adjusted for each assay procedure and be based on the quantity of beads being counted by the instruments. We suggest to always obtain ³32 counts per bead ID. 5. After each measurement and for the preparation of new bead mixtures, the count average is to be calculated for each bead ID and new volumes can be determined. We suggest adjusting

34

Schwenk and Nilsson

these volumes to a theoretical bead count, which is 20% above the estimate: For 100 beads to be counted from the new bead mixture, the previously obtained volumes should be calculated for 120 beads per assay and bead ID. 3.3. Sample Labeling

1. The serum or plasma samples are to be thawed according to the preferred protocol (see Note 6). 2. The samples are vortexed and centrifuged for 10 min at 10,000 × g to pellet insoluble components. 3. A previously designed plate layout, in which samples should be located randomly, is followed by transfer of 30 ml of serum/ plasma into the respective wells of a PCR plate, which is then sealed and centrifuged for 2 min at 1,500 × g. 4. As an option, the samples are incubated for 30 min at elevated temperatures such as 56°C (see Note 7) followed by 15 min at 20°C using in a thermo cycler. Using the heated lid function of the cycler helps to prevent the samples to evaporate into the lid/seal. 5. Transfer 3 ml into a second PCR plate containing 27 ml PBS, seal the plate, vortex, and centrifuge for 2 min at 1,500 × g. 6. Add 2.5 ml of NHS-Biotin to each well (see Note 8), then seal the plate, vortex and centrifuge for 2 min at 1,500 × g, and incubate for 2 h at 4°C under continuous shaking in a microtiter plate mixer. 7. Add 25 ml of 1 M Tris–HCl pH 8.0 to each well, seal the plate, vortex, and centrifuge for 2 min at 1,500 × g. 8. Store the plates at −20°C until usage or use directly.

3.4. Assay Procedure

1. The labeled samples are thawed and diluted 1/50 in assay buffer, which had been prepared in a PCR plate. Seal the plate, vortex, and centrifuge for 2 min at 1,500 × g. 2. The samples incubated for 30 min. As an option, the samples are treated for 30 min at elevated temperatures such as 56°C (see Note 7), followed by 15 min at 20°C using the heated lid function of the thermo cycler. Thereafter, the plate is vortexed and centrifuged for 2 min at 1,500 × g. 3. The previously prepared bead mixture is distributed into the wells of a half-area plate and protected from light. Then 45 ml of the diluted, labeled samples are added to the wells (see Note 5) and incubated at 23°C over night under continuous shaking on a microtiter plate mixer. 4. The plates are then washed 3× with 75 ml wash buffer, incubated with stop solution for 10 min and washed 1× with 75 ml wash buffer again.

Antibody Suspension Bead Arrays

35

5. R-PE labeled streptavidin is then added to each well at 0.5 mg/ml and 30 ml and the plates are incubated for 20 min under continuous shaking. 6. The plates are then finally washed 3× with 75 ml wash buffer and 100 ml of wash buffer are added before the plates are measured with the Luminex instrumentation. 7. Set the instrumentation setting according to the bead IDs included in the mixture and count at least 50 beads per bead ID. We suggest using the “median fluorescence intensity” to further process your data. An example of a plasma protein profile is shown in Fig. 2.

4. Notes 1. Other fluorescent dyes than R-Phycoerythrin such as Alexa546, Alexa532, or Cy3 can be utilized as well, but Luminex Corp. has indicated that lower reporter signal intensities are to be observed. 2. Employ solutions of purified proteins and avoid other stabilizing proteins, Tris or other amine-based buffers as they reduce the coupling efficiency. 3. At all times, try to minimize the light exposure, especially to direct sunlight, as the internal fluorescence of the beads as well as reporter fluorophores could be bleached. During incubation, protect the plates with an opaque cover or place plate into a light-tight box. 4. Do not interrupt the activation process after dissolving EDC and NHS, as these active substances are susceptible to hydrolysis resulting in a loss in activity. 5. When combining beads with solutions for counting and assay procedure, always distribute small volume bead solution (e.g., 5 ml) into the well first, then add larger volume buffer portion (e.g., 45 ml) to allow an instant distribution of the beads. 6. We have found that thawing overnight at +4°C was most practical if a larger number of samples were to be processed. Otherwise, place tube(s) into a 42°C water bath until a minor fraction of ice was still visible. 7. We have observed that heat treatment of labeled samples in combination with the applied multiplexed assay procedure affects antibody performance (5). This can lead to improved protein detectability by changing the accessibility of the epitopes in a complex sample solution but should be tested and balanced with the tendency of proteins to precipitate at higher temperatures.

36

Schwenk and Nilsson

8. Do not interrupt the experimental flow after dissolving NHS-Biotin, as this active substance is susceptible to hydrolysis resulting in a loss in activity. Add NHS-Biotin to the side of each well using single- or multichannel dispensers so that the labeling reactions for all samples are started contemporaneously.

Acknowledgments We like to thank the entire staff of the Human Proteome Resource (HPR) initiative for their tremendous efforts within the Human Protein Atlas project. This work is supported by the PRONOVA project (VINNOVA, Swedish Governmental Agency for Innovation Systems), and by grants from the Knut and Alice Wallenberg Foundation. References 1. Hanash S (2003) Disease proteomics. Nature 422:226–232 2. Templin MF, Stoll D, Schwenk JM, Pötz O, Kramer S, Joos TO (2003) Protein micro arrays: promising tools for proteomic research. Proteomics 3:2155–2166 3. Kingsmore SF (2006) Multiplexed protein measurement: technologies and applications of protein and antibody arrays. Nat Rev Drug Discov 5:310–320

4. Fulton RJ, McDade RL, Smith PL, Kienker LJ, Kettman JR Jr (1997) Advanced multiplexed analysis with the FlowMetrix system. Clin Chem 43:1749–1756 5. Schwenk JM, Gry M, Rimini R, Uhlen M, Nilsson P (2008) Antibody suspension bead arrays within serum proteomics. J Proteome Res 7:3168–3179

Chapter 4 Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies Víctor J. Cid, Ekkehard Kauffmann, and María Molina Abstract Infection of cells and tissues by pathogenic microorganisms often involves severe reprogramming of host cell signaling. Typically, invasive microorganisms manipulate host cellular pathways seeking advantage for replication and survival within the host, or to evade the immune response. Understanding such subversion of the host cell by intracellular pathogens at a molecular level is the key to possible preventive and therapeutic interventions on infectious diseases. Reverse Protein Arrays (RPAs) have been exploited in other fields, especially in molecular oncology. However, this technology has not been applied yet to the study of infectious diseases. Coupling classic in vitro infection techniques used by cellular microbiologists to proteomic approaches such as RPA analysis should provide a wealth of information about how host cell pathways are manipulated by pathogens. The increasing availability of antibodies specific for phosphorylated epitopes in signaling proteins allows monitoring global changes in phosphorylation through the infection process by utilizing RPA analyses. In our lab, we have shown the potential of RPA technology in this field by devising a microarray consisting of lysates from cell cultures infected by Salmonella typhimurium. The protocols used are described here. Key words: Reverse protein arrays, Cell lysate arrays, In vitro infection, Virulence factors, Salmonella, Type III secretion, Host cell signaling, Protein kinase, Protein phosphorylation

1. Introduction Reverse protein arrays (RPAs) technology is an antibody-based proteomic approach based on high-throughput dot blots performed on cell lysates printed to a solid support, followed by quantitative immunodetection. RPA assays have been used to monitor cell signaling in different contexts, especially in the field of oncology (1–4). Nevertheless, its application to the study of host–pathogen interactions, as proposed here, has not been exploited to date. Obligate intracellular parasites such as viruses

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_4, © Springer Science+Business Media, LLC 2011

37

38

Cid, Kauffmann, and Molina

and some bacteria (typically chlamydiae and rickettsiae), as well as facultative intracellular pathogens, such as invasive bacteria and fungi, are able to subvert host cell signaling to promote their internalization into target cells, their intracellular survival, or their replication (5). Interaction of infectious agents with host cell pathways is also responsible for cell and tissue damage. Some pathogens have been also described to modulate signaling events aimed to evade the immune response. The pathogens achieve these goals by expressing specific virulence factors that directly interfere with the function of cell signaling proteins. One paradigmatic example is the injection into the host cell cytoplasm of bacterial proteins (“effectors”) by specialized secretory systems (for example, type III and type IV secretion systems) (6–8). This is a common mechanism for both plant and animal bacterial pathogens. Such effectors are regulators of cytoskeletal components, such as actin and tubulin, regulators of GTPases, kinases, or phosphatases of proteins or phosphoinositides, or regulators of ubiquitin ligases (9, 10). Thus, they are able to reprogram the host cell to generate a comfortable environment for their colonization. RPA technology opens the possibility of analyzing by immunoblotting the presence of particular proteins or their posttranslational modifications, such as phosphorylation, in cell lines after exposure to a particular pathogen in a given variety of experimental conditions. To set up the working conditions and demonstrate the potential of this technique, we have recently used RPAs to assess the involvement of Salmonella typhimurium type III secretion system (T3SS) effectors (11). These proteins are translocated from the bacterium to the host-cell cytoplasm by the T3SS encoded by the Salmonella pathogenicity island I (SPI-1), which is specifically involved in remodeling actin and promoting bacterial internalization during infection (12). A general scheme on the experimental design followed in our approach, which could be applied to other host–pathogen systems, is presented in Fig. 1. Specifically, we infected in vitro HeLa cells, widely used as a model

Fig. 1. General scheme of RPA hybridization coupled to in vitro infection as a tool to monitor signaling events through host–pathogen interaction. Alternative possibilities for experimental design are noted. The desired pathogen is grown in a variety of experimental conditions, as desired. The choice of different mutants or isolates will provide information on the contribution of mutated factors or particular isolates to discrete signaling events in the host model. Choice of the host cell line for in vitro infection (epithelial cells vs. lymphoid cell lines; primary vs. immortal cultures) should also be determined by the nature of the biological question under study. At the desired times after infection, control and problem cells are detached from culture plates and a collection of lysates is prepared, calibrated, and four different dilutions of each lysate are spotted in duplicate in a format of six arrays per chip. Zeptosens ZeptoMARK chips and the corresponding analytical technology were used in our assays (see Note 12). Hybridization of these chips with a collection of prevalidated antibodies specific to particular proteins or their posttranslational modifications (i.e., phosphorylation) considered as markers for the activation of signaling pathways, and subsequent quantification of the data, will reveal a wealth of information on how such pathways respond to contact with the pathogen.

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies EXPERIMENTAL DESIGN - Mutants in virulence-related genes - Different clinical isolates - Various infection times - Different infection doses - Pre-treatment with antibiotics etc.

Pathogen strains

in vitro infection EXPERIMENTAL DESIGN - Immortal lines - Primary cultures - siRNA directed silencing etc.

Host cell line

Lysate collection representing all infection conditions Array printing Hybridization with a collection of validated antibodies Antibody 1 Antibody 2

... Antibody n

Comparative quantitative readout of - protein levels - postranslational modifications (phosphorylation...) Data analysis

39

40

Cid, Kauffmann, and Molina

for bacterial infection on epithelial cells (13), with either wild-type Salmonella or different mutants impaired in particular aspects of T3SS function, including a mutant unable to assemble the T3SS (invA), as well as combinations of particular mutations in wellknown T3SS effectors, such as SopE and SopB. SopE and its paralog SopE2 are activators of the small GTPase Cdc42 that elicits signals aimed to recruit actin to the area of bacterial contact to promote membrane ruffling and eventual internalization of the bacterium (14, 15). Besides, SopE and SopE2 have been involved in the activation of mitogen-activated protein kinase (MAPK)mediated signaling pathways in the host cell (16). SopB, also known as SigD, is a phosphatidylinositol-phosphate phosphatase involved in different steps of the formation of the Salmonellacontaining vacuole (17–19) that also interacts with Cdc42 (20) and specifically triggers activation of protein kinase B (Akt) in the host cell (21). The method detailed here proved useful to confirm previously described signaling events that depend on SopB and SopE effectors, to detect novel changes in phosphorylation, and to assess the contribution of those particular bacterial effectors to such changes (11). We believe that the same approach could be used for other Salmonella effectors (more than 30 T3SS-secreted proteins have been detected, many of them of yet unknown function), other invasive bacteria (Shigella, Listeria, Mycobacterium, Chlamydia, etc.) or fungi (Candida, Cryptococcus), or even noninvasive pathogens that are known to severely subvert host cell signaling (enteropathogenic E. coli, etc.).

2. Materials 2.1. Preparation of Lysates 2.1.1. In Vitro Infection

1. HeLa cells (human adenocarcinoma cervix epithelial cell line, obtained from the American Type Culture Collection CCL-2, Manassas, VA) (see Note 1). 2. Salmonella strain SL1344 (22) and genetically manipulated derivatives (11). 3. Growth medium for HeLa: RPMI 1640 (Biological Industries, Israel) supplemented with 10% fetal calf serum, penicillin (100 units/mL), streptomycin (100 mg/mL), and 2 mM l-glutamine. 4. Growth medium for bacteria: Luria Bertani (LB) broth: 10 g/L bacto tryptone, 5 g/L bacto yeast extract, and 10 g/L NaCl, sterilized by autoclaving. If plasmid maintenance was required, LB was supplemented with 12 mg/mL chloramphenicol. 5. P100 plates (10 cm tissue culture dishes, Cellstar).

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies 2.1.2. Cell Lysis

41

1. Phosphate-buffered saline (PBS) (Fluka). 2. Cell lysis buffer CLB1 (Zeptosens, Switzerland) (see Note 2).

2.2. Sample Preparation for Spotting 2.2.1. Determination of Protein Concentration in Lysates

1. Cell lysis buffer CLB1 (Zeptosens, Switzerland). 2. Phosphate-buffered saline (Fluka). 3. Coomassie Plus – The Better Bradford assay reagent (Thermo Fisher). 4. Bovine serum albumin (BSA) ampoules 2 mg/mL (Thermo Fisher). 5. 96-Well microtiter plates, flat bottom, suitable for optical readout (Greiner bio-one, Germany). 6. Microtiter plate reader suitable for absorbance measurement at 595 nm (e.g., SpectraMax Plus, Molecular Devices, CA). 7. Low volume liquid handling robot, 96 channel (e.g., Zephyr, Twister II, Caliper Life Sciences, MA).

2.2.2. Preparation of Spotting Microplate

1. ZeptoMARK CSBL1 – Spotting Buffer (Zeptosens, Switzerland). 2. ZeptoMARK CLB1 Lysis Buffer (Zeptosens, Switzerland). 3. Dilution buffer: 1 part CLB1, 9 parts CSBL1. 4. Fluorophore conjugated with albumin from bovine serum (BSA) [e.g., Cy5 (GE Healthcare)]. 5. Alexa Fluor 647 (Invitrogen), HiLyte Fluor 647 (Anaspec), or Dylight 649 (Thermo Fisher). 6. ZeptoMARK Spotting Buffer for References CSBR1, (Zeptosens, Switzerland). 7. ZeptoMARK Reference Dilution Buffer RDB1 (Zeptosens, Switzerland). 8. PBS tablets (Fluka). 9. Sodium azide (NaN3) (Sigma). 10. Microcentrifuge filtration unit, 1.5 mL tube 0.22 mm (e.g., Millipore UFC30GV00). 11. Low volume liquid handling robot, 96 channel (e.g., Zephyr in combination with Twister II, Caliper Life Sciences, MA).

2.3. Microarray Spotting

1. Noncontact microarray spotter, modified NanoPlotter (Zeptosens/GeSiM, Switzerland).

2.4. Chip Blocking

1. ZeptoFOG ultrasonic nebulizer blocking station (Zeptosens, Switzerland). 2. ZeptoCHIP blocking racks, including lids and handles (Zeptosens, Switzerland). 3. ZeptoMARK blocking buffer BB1 (Zeptosens, Switzerland) (see Note 3).

42

Cid, Kauffmann, and Molina

2.5. Reverse Array Assay

1. ZeptoCARRIER chip holder with fluidic cells (Zeptosens, Switzerland). 2. Eight-channel aspiration tool (e.g., VacuSafe, IBS Integra Biosciences, Switzerland). 3. Specific primary antibodies, directed against the proteins and posttranslationally modified proteins of interest (see Subheading 3.8). 4. Antispecies secondary antibodies (full IgGs or Fab fragments), carrying fluorescence labels in the red spectral range, e.g., Cy5 (GE Healthcare), Alexa Fluor 647 (Invitrogen), HiLyte Fluor 647 (Anaspec), or Dylight 649 (Thermo Fisher).

2.6. Readout and Data Analysis

1. ZeptoREADER planar waveguide fluorescence reader for ZeptoMARK chips, assembled in microplate-footprint ZeptoCARRIERs (Zeptosens, Switzerland). 2. ZeptoVIEW 3 reverse array analysis software (Zeptosens, Switzerland).

3. Methods The choice of a particular set of different infection conditions will depend on the experimental problem to tackle. Thus, RPA experimental design is the key to obtain a maximum amount of valuable and comprehensive information. Among the diverse RPA printing and analysis systems available in the market, we have used the technology developed by Zeptosens. As shown in Fig. 2, the standard configuration, this system allows working simultaneously with multiples of 32 samples. It should be considered that controls of noninfected cells must be included for each of the conditions tested. Thus, for example, if a collection of 30 mutants lacking particular virulence genes is available for the pathogen under study, they should be compared with a wild-type control sample at a fixed time point after infection, as well as with a noninfected cells sample to complete the array. Also, RPA analysis is a powerful tool to investigate the timing of signaling dynamics. So an alternative experimental design could be to study eight time points after infection for only two given mutants or strains of interest as compared to the same points of a wild-type or control strain and the equivalent set of samples from control uninfected cells. The method presented is based on our experience with Salmonella typhimurium infection of HeLa cells, but we believe it could be adapted to any host–pathogen system of in vitro infection. Experiments on cultures of fibroblasts or lymphoid cells might reveal novel aspects of the modulation of signaling by the pathogen, especially for bacteria that are able to survive within professional phagocytes, such as Salmonella typhi or Mycobacterium

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies

D

Zeptosens

E

F

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

Buffer Standards

relative concentration 0.25

1

reference

Ref

0.5

C

Ref

0.75

B

Ref

1.0

A

Ref

reference

6 Arrays /Chip

43

Fig. 2. Layout of ZeptoMARK chip with six arrays. Each array accommodates four columns of reference spots and duplicate spots of four dilutions of 32 lysate samples. Control samples for the secondary antibody are spotted at the right bottom of each array. Reference spotting solution containing a BSA-fluorophore conjugate is spotted as columns of reference spots on each array.

tuberculosis. Different host cell lines, even primary cultures, or the influence of the presence of compounds or particular environmental conditions for both the pathogen and the host could also be compared (see Fig. 1). The most important prerequisite of this method is the quality of the antibodies selected for the experiment. Any antibody should be very specific for the epitope to be detected because any cross-reaction with other proteins in the lysates will give rise to background noise and failure to detect variations of the target epitope in the samples. All antibodies used must be previously validated for both their specificity and their sensitivity (see Subheading 3.8). To evaluate posttranslational modifications, such as phosphorylation, an antiphosphoprotein and antiprotein pair of antibodies must be used. 3.1. Preparation of Lysates 3.1.1. In Vitro Infection (see Note 4)

1. To prepare HeLa cells for the infection, grow them in a humidified 5% CO2 tissue culture incubator at 37°C for 24 h on P100 plates. 2. Before infection, wash HeLa cells twice with the same growth medium lacking any antibiotic supplement, and add fresh medium. 3. For the preparation of Salmonella inocula, grow bacterial cells overnight at 37°C in LB medium in a shaking orbital incubator (200 rpm). 4. Use 150 mL of these saturated cultures to inoculate 5 mL of the same fresh medium and incubate further until an OD600 = 0.5 (see Note 5).

44

Cid, Kauffmann, and Molina

5. At the time of infection (t0), add 25–30 mL of the bacterial inoculum per milliliter of RPMI 1640 medium. This should yield a multiplicity of infection (MOI or bacteria:cell ratio) of about 80:1 (see Note 6). In our experiments, to study bacterial invasion, infected cells were incubated for 10, 30, or 60 min (time points t10, t30 and t60) (see Note 7). 6. If it is intended to study modulation of host cell signaling by internalized bacteria in the absence of further invasion, it is necessary to remove extracellular bacteria. In this case, HeLa cells incubated for 60 min should be washed twice with fresh medium, then medium supplemented with 100 mg/mL gentamycin should be added and incubated for 1 h to kill extracellular bacteria. Then, incubate cells in medium with a lower gentamycin concentration (10 mg/mL) for the desired extra time to obtain further time points. In our experiments, we incubated cells for one extra hour (tis; “is” stands for “intracellular survival”). Intracellular survival can be monitored by counting colony-forming units (CFUs) on LB agar plates. 3.1.2. Cell Lysis (see Note 8)

1. Wash HeLa cells once with PBS. 2. Add CLB1 lysis buffer (0.2 mL per 10 cm-diameter culture dish), scrape and transfer into 1.5-mL reaction tubes. 3. Incubate 30 min at room temperature. 4. Centrifuge cell lysates for 5 min at 15,000 × g in order to remove debris. 5. Collect supernatants and freeze in liquid nitrogen. Cell lysate samples should be stored frozen at −20 or −80°C. They should be thawed only immediately prior to use.

3.2. Sample Preparation for Spotting

3.2.1. Determination of Protein Concentration in Lysates

Samples might be prepared for spotting by manual pipetting. However, in routine application of RPA, a liquid handling robot for protein quantification and the subsequent dilution steps will be beneficial in various ways. The sample preparation will be faster, avoiding degradation of samples and evaporation. Additionally, it will become more reproducible and the risk of erroneously exchanging samples is greatly reduced. The protocols for protein quantification and the preparation of the spotting described here can efficiently be integrated on a low-volume 96-channel Zephyr liquid handler (Caliper Life Science). An attached Twister II plate handler provides the interface to the microplate reader and serves as microplate storage. 1. Freshly prepare the BSA solutions shown in Table 1 (final CLB1 concentration 5%) to elaborate a standard curve. 2. Equilibrate the Coomassie solution and PBS buffer, and lysate samples to room temperature. After equilibration, gently shake Coomassie solution preventing generation of foam.

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies

45

Table 1 BSA standard solutions BSA (2 mg/mL ampoules) (mL)

CLB1 (mL)

PBS (mL)

Final BSA-concentration (mg/mL)

0

50

950

0

25

50

925

0.05

50

50

900

0.1

75

50

875

0.15

100

50

850

0.2

150

50

800

0.3

200

50

750

0.4

300

50

650

0.6

3. Vortex the lysate samples and centrifuge them at 10,000 × g for 1 min. Collect supernatants. 4. Dilute each sample 1:20 with PBS: Add 2.5 mL of the sample to 47.5 mL PBS at room temperature and mix thoroughly. 5. For Bradford Assay Preparation, transfer two times 10 mL of each of the diluted lysate samples and of each BSA standard solution (see Table 1) into duplicate wells of an empty 96well plate. 6. Add 240 mL Coomassie solution to each well. Work quickly, e.g., by using a multichannel pipette. 7. Place the 96-well plate on an appropriate plate mixing device for careful mixing of the reagents without generating foam or air bubbles. Incubate the solutions at room temperature. 8. Mix the plate again for 30 s right before measuring the absorbance at 595 nm in the MTP-Reader. Measure absorbance exactly 10 min after addition of the Coomassie solution. Since the development of the assay signal is time-dependent, it is advisable to include a calibration curve in every microtiter plate. 9. Generate a calibration curve by plotting the measured absorbance values versus the BSA concentrations of the standards and perform a linear regression. 10. Calculate back the protein concentrations in the lysates from the measured absorbance values, taking into account the dilution factor of 20.

46

Cid, Kauffmann, and Molina

3.2.2. Preparation of Reference Spotting Solution

For referencing of the fluorescence intensities on the array to the excitation light intensity, a reference spotting solution (RSS) containing a BSA-fluorophore conjugate is spotted as columns of reference spots on each array (see Fig. 2). RSS is prepared as follows: 1. Prepare a stock solution of the BSA-fluorophore in PBS and 0.1% sodium azide at a concentration of 5 mg/mL. It can be aliquoted and stored at −20°C for at least 1 year. 2. Thaw 10 mL aliquot of the stock solution. Centrifuge solution for 20 s at 10,000 × g and use the supernatant only. 3. Thaw a 500 mL aliquot of buffer RDB1, vortex well. 4. Thaw a 500 mL aliquot of buffer CSBR1, vortex well. 5. Prepare solution A by adding 2 mL of stock solution to 198 mL of RDB1 (dilution 1:100). This dilution may be varied in order to obtain the desired reference spot intensity of approximately 30,000 counts in an emission image with 10 s integration time. 6. Prepare solution B by adding 2 mL of Solution A to 198 mL RDB1 (dilution 1:100). 7. Prepare solution C by pipetting 180 mL of CSBR1 and 15 mL RDB1 into a microcentrifuge filter unit and filter it at 10,000 × g. 8. Add 5 mL of Solution B to filtered Solution C and mix well. The reference solution is ready for use.

3.2.3. Preparation of Spotting Microplate

The spotting solutions for the samples are prepared in 384-well polypropylene microwell plates which are used as source plates by the spotting robot. The well position of each solution has to be adapted according to the transfer scheme of the spotting robot and the desired spotting layout. 20 mL of spotting solution per well in a 384-well microplate is sufficient for the NanoPlotter described in this protocol. It is advisable to spot a series of at least four dilutions of each sample to allow for verification of linearity of dose response in the assays. Prepare the spotting solutions as follows: 1. Normalize samples to 2 mg/mL total protein concentration using buffer CLB1. 2. Dilute normalized samples tenfold with buffer CSBL1 (e.g., 20 mL sample + 180 mL CSBL1). This solution (0.2 mg/mL total protein concentration) is the first dilution spotted. 3. For the second spotting solution, dilute the first spotting solution with the dilution buffer (1 part CLB1, 9 parts CSBL1) to 0.15 mg/mL total protein concentration (e.g., 60 mL first spotting solution + 20 mL buffer).

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies

47

4. For the third spotting solution, dilute the first spotting solution with the dilution buffer (1 part CLB1, 9 parts CSBL1) to 0.1 mg/mL total protein concentration (e.g., 40 mL first spotting solution + 40 mL buffer). 5. For the forth spotting solution, dilute the first spotting solution with the dilution buffer (1 part CLB1, 9 parts CSBL1) to 0.05 mg/mL total protein concentration (e.g., 20 mL first spotting solution + 60 mL buffer). 3.3. Microarray Spotting

In the spotting process, droplets of 400 pL of each lysate solution, the reference solution, and control samples are deposited as arrays on ZeptoMARK hydrophobic chips. The best reproducibility and robustness of the spotting process have been reached using piezoelectric dispensing systems as, for example, provided on the NanoPlotter (Zeptosens/GeSiM, Germany). The array layout described here accommodates 32 lysate samples in four dilutions as duplicate spots, plus a negative control (spotting buffer) and a positive control [rabbit IgG and mouse IgG (a mix of mouse IgG1, IgG2a and IgG2b)], both as duplicate spots (see Fig. 2). 1. Start up the NanoPlotter and load/create a suitable spotting program (see Note 9). 2. Place source plate(s) on the spotter. The plate holder should be chilled to a temperature at which no evaporation of spotting solution can be measured over 12 h. At 21°C/50% relative humidity, the temperature of the coolant will typically be at 14°C (noncondensing conditions). 3. Place the chips on the slide deck. ZeptoMARK chips bind the proteins in the lysate droplets by hydrophobic interaction. It is advisable to spot more chips than at least needed for the envisioned number of antibodies (see Note 10). 4. Flush the piezo-electric dispensers (NanoJets, GeSiM) at least 5 min using the backfill water system. 5. Check the performance of the piezo jets with one of the samples using the stroboscope camera which is installed on the spotter. A single droplet with stable position should be generated. To optimize droplet formation, tune pulse width and voltage for the piezo jets in an alternating manner. Once optimized, pulse parameters are rather stable over the lifetime of a piezo jet. 6. Load the spotting program and start the spotting run. 7. After spotting, dry chips for 1 h at 37°C.

3.4. Chip Blocking

1. Place the ZeptoFOG blocking station in a laboratory hood. 2. Equilibrate blocking buffer BB1 to room temperature and filter it through a 0.5-mm cellulose acetate syringe filter into the blocking station.

48

Cid, Kauffmann, and Molina

3. Place upto two blocking racks with ZeptoMARK chips in the blocking station. The carrier must be tilted in a way that the spotted side of the chips is facing upward. The chips should be tilted towards the air inlet. 4. Close the ZeptoFOG blocking station with the lid. Connect the air inlet to the membrane pump. 5. Switch on the ultrasound generator. Fog is generated now. After 30 s, switch on the membrane pump. Nebulized blocking buffer will now rise. If optimally adjusted, the dense blocking fog is distributed homogeneously in the ZeptoFOG blocking station and a weak stream of fog leaves the system through the outlet. 6. Block the chips for 30 min. The blocking buffer should not warm up above 40°C. 7. Switch off membrane pump and ultrasound generator. Install lids and handles to the racks. Move to first wash bath filled with ca. 1.5 L water immediately. 8. Rinse the blocking rack holding the chips thoroughly in a sequence of four water baths with 1.5 L of water. 9. To dry the blocked chips, place the blocking racks (without lids and handles) on top of a 96-well plate in a microtiter plate centrifuge. The spotted side of the chips should face the center of rotation. Spin dry the chips at a maximum of 197 × g for 3 min applying maximum centrifuge acceleration. 10. Thorough cleaning of equipment is essential for low number of particle artifacts on blocked chips. Rinse the ZeptoFOG blocking station intensively with water and clean the blocking racks with water in an ultrasonic bath. 11. Store the blocked ZeptoMARK chips at 4°C until use. 3.5. Reverse Protein Array Assay

Six ZeptoCHIPs at a time are assembled in microplate-format ZeptoCARRIERs with fluidic cells. The micro flow system allows addressing each of the six arrays of a chip individually with 50 mL of antibody solution. The detection of the specific proteins follows a two-step sequential assay: 1. Prime all arrays with 100 mL of ZeptoMARK assay buffer CAB1, aspirate buffer again with the 8-channel aspiration tool. 2. Inject 50 mL of primary antibody solution ZeptoMARK assay buffer CAB1 or CAB2 in each cell. 3. Incubate for 20 h at 25°C in the dark. 4. Wash three times with ZeptoMARK buffer CAB1. 5. Inject 80 mL of appropriate fluorescent-labeled secondary antibody solution (at 500-fold dilution in buffer CAB1) in each cell.

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies

49

6. Incubate for 2.5 h at 25°C in the dark. 7. Wash the arrays again three times with ZeptoMARK buffer CAB1. After the last wash, leave the buffer in the cells for readout. In this two-step assay, the secondary antibody itself might generate unspecific signal on the spots (“Blank Signal”). Separate arrays, incubated with buffer only in the first part of the assay, and secondary antibody solution in the second part of the assay should therefore be prepared. The ZeptoMARK chips are imaged in a ZeptoREADER using planar excitation (see Note 11 and Fig. 4).

3.6. Readout and Data Analysis

1. Acquire fluorescence images of the arrays in the ZeptoREADER, exciting at lex = 635 nm and collecting fluorescence at lem = 670 nm, with exposure times of 1, 5 and 10 s. These are typical exposure times for measurement of medium to low abundant signaling pathway markers. 2. Determine the mean net fluorescence intensities of all spots (Image analysis module in ZeptoVIEW 3). Typically, a spot diameter setting of 100 mm will be optimal. The background should be determined for each spot individually and subtracted from the sample spot signal leading to the net signal. 3. Reference the mean net intensities of the sample spot to the excitation light intensity, represented by the reference spots (spotted fluorescence-labeled BSA, see Fig. 3). This results in

a

b

1.5

1.5

RFI

2

RFI

2

1

0.5

0 0

1

0.5

0.25

0.5

0.75

relative dilution of total protein

1

0

0

0.25

0.5

0.75

1

relative dilution of total protein

Fig. 3. (a) Dilution plot of two lysate samples with differing mean RFI under ideal assay conditions. Normalization with the mean RFI results in lines with slope equal to 1. (b) Dilution plot of a lysate sample under assay conditions with significant unspecific binding. The slope of the normalized dilution plot differs from 1. Note that the normalized dilution plot is only used for quality control. The mean RFI for each lysate sample is calculated from the nonnormalized data.

50

Cid, Kauffmann, and Molina

“referenced fluorescence intensity” (RFI) for each spot (Result module of ZeptoVIEW 3). 4. Calculate the mean RFI for each sample with the corresponding standard deviation based on an error-weighted linear fit through the RFI values of the eight sample spots (duplicate spots of the four sample dilutions) (23). The center among all spotted sample dilutions should be used as mean RFI value (Reverse array module of ZeptoVIEW 3). 5. Subtract the mean RFI of the blank assays with the corresponding secondary antibody and assay buffer from the mean RFI values of each marker assay. 3.7. Data Interpretation and Quality Control

RPAs allow quantifying the relative amounts of each (posttranslationally modified) marker protein in the set of lysate samples of an experiment. However, a number of quality control metrics have to be fulfilled. The immunoassay has to be performed in the linear range, i.e., excess of antibody. This can be verified using a dilution plot, in which the RFI values of single spots of a lysate sample are plotted against the sample dilution (see Fig. 4a). The signal of the dilution series has to correlate linearly with the spotted sample concentration. The correlation coefficient of a linear regression of RFI value with sample dilution can be used as quality parameter. The contribution of unspecific signal to the RFI should be small, i.e., when extrapolated to infinite dilution, the signal should approach zero. For an entire sample set, this can efficiently be verified using the slope of the dilution plot after normalization with the mean RFI, bNORM. In an ideal assay, bNORM equals 1. The effect of unspecific signal is exemplified in Fig. 3b.

planar waveguide principle

laser beam

planar waveguide excitation

confocal excitation

CCD camera

Fig. 4. Schematic of the planar waveguide excitation and image acquisition principle incorporated in the ZeptoREADER. Unlike confocal excitation, the evanescent field of the waveguide excites only surface-confined fluorophores and the direction of excitation is perpendicular to the direction of detection. This results in significant reduction of the background.

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies

51

9 30 min post-infection data

Ratio phosphoprotein/protein (nomalized to non-infected cells)

8 WT invA sopE/E2 sopB sopE/E2/B

4

3

2

1

0

Akt1 GSK-3β FoxO1 p70 S6K p42/44 S473 S9 T421 S256 (ERK) S424 T202 Y204

p38 T180 Y182

C-Jun S73

IkBα S32

CREB S133

PLCγ1 Y783

Fig. 5. Evaluation of the activation of several signaling pathways after infection of HeLa cells with wild-type or mutants of Salmonella typhimurium lacking different subsets of the invasion-related proteins invA, sopE, sopE2, and sopB, as noted. Data represented in ordinates are the ratio of the signal obtained from the analysis of the phospho-epitopes indicated relative to that of the nonphosphorylated total protein. These data were previously normalized with respect to those obtained for the noninfected control. Thus, values above 1 imply an increase in phosphorylation in the conditions studied over background, whereas values below 1 imply a decrease in phosphorylation for that particular marker. Original raw data and further information for these experiments can be found in ref. (11).

The signal and the immunoassay should not reach saturation. On the other hand, spots at high sample dilution should only be taken into account if they provide measurable signal above the background with a signal-to-noise ratio above 3. Both events can be verified by determining the curvature of the dilution plot. In host–pathogen interaction studies, the mean RFI values obtained for a particular lysate of infected cells with each antibody should be normalized to the mean RFI value of the corresponding noninfected control. To assess differences in posttranslational modifications, such as phosphorylation, on particular signaling proteins, a phosphorylation ratio should be calculated from normalized mean RFI data obtained using the antiphosphoproteinspecific antibody and the antiprotein antibody for each particular lysate. A graphic representation of the phosphorylation ratio for different signaling proteins after 30 min of infection with different mutant strains of Salmonella on HeLa cells is provided in Fig. 5. 3.8. Antibody Validation

Like in a classical dot blot, the primary antibodies used on RPAs have to be highly specific. As only specific antibodies enable relative quantification of markers, significant effort is required to ensure the antibody validation before they are applied (24).

52

Cid, Kauffmann, and Molina

Antibodies for total, unmodified proteins should be tested for specificity in Western blots from non-overexpressing cell lines. Antibodies against modified proteins can be validated for specificity in a Western blot of broadly stimulated cultured cells. For example, an antibody against the phosphorylated form of AKT at position serine 473 could be tested on cultured cells stimulated with the unspecific Ser/Thr phosphatase inhibitor calyculin. Additionally, the antibodies have to be validated for linearity of dose–response in the reverse array experiment. The antibody concentration and buffer should be optimized for signal intensity and for the quality parameters described in Subheading 3.7. Typically, the optimum concentration for a reverse array is about double the concentration used in a Western blot.

4. Notes 1. It is important to obtain mammalian cell lines from a reliable source and to keep good frozen stocks and records. Low passage number (<15) cells should be used. The choice of an adequate cell line is crucial for interpretation: Some immortal cell lines have high basal phosphorylation levels for certain pathways, and this would lead to high background levels. The line chosen should be as similar to the tissue infected in vivo as possible. HeLa cells are easy to handle and may express most qualities of epithelial cells, but caution should be taken with the results obtained, because differences with the target tissue of the pathogen (in the case of Salmonella, intestinal epithelial cells) might be considerable. 2. The CLB1 buffer contains a high concentration of denaturants. This puts all biological activity and enzymes such as phosphatases to a halt the moment the buffer is added to the cultured cells and preserves the status of posttranslational modifications. CLB1 lysis buffer has to be stored at −20°C (due to the presence of DTT). Thaw lysis buffer prior to use and refreeze it immediately. Repeated freeze–thaw cycles do not have any impact on its quality or solubilization efficiency. Do not use thawed lysis buffer longer than a day. 3. Blocking and assay buffers should always be filtered using a 0.45-mm cellulose acetate syringe filter. 4. For a recent specific technical article on in vitro infection with Salmonella, see ref. (13). 5. Bacteria should be freshly grown and kept at mid- to late-log phase at the time of infection. 6. MOI is important. It refers to the ratio of bacteria per host cell used in the infection. MOIs from 10:1 to 100:1 are used

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies

53

for most purposes regarding in vitro infection with Salmonella. The most adequate MOI to tackle your scientific question should be chosen. Infection conditions should be previously set up in small-scale experiments to determine experimental RPA layout. 7. In order to monitor the samples that will be used to print the array and meet the required experimental quality, invasion tests should be made through the process. Invasion could be monitored either by microscopic visualization of actin, membrane ruffling, or fluorescent staining of bacteria, or by quantification of intracellular bacteria using the gentamycin protection assay on parallel samples [see refs. (11, 13, 25) for specific methods]. 8. The protocol described in Subheading 3.1.2 is meant for adherent cell lines. For nonadherent cell lines, such as lymphoid cells, spin down cell suspensions for 2 min at 300 × g, discard supernatants, and resuspend pellets in PBS (use 1/10 of starting volume). Then spin down cell suspensions again for 2 min at 300 × g, discard supernatants, resuspend pellets in 50 mL lysis buffer, and proceed as for adherent cells. For 50 mL of cell lysate with a concentration of around 3 mg/mL, a density of 106 cells is required. Please note that these values can vary with different cell types or cell lines. 9. Only the steps of the protocol that are specific to RPA spotting are described here. It is assumed that the NanoPlotter microarray spotter is set up according to the manufacturer’s instructions. An environment with reduced dust and controlled temperature (21–23°C) as well as controlled humidity (50%) improves the spotting reproducibility. Relative humidity above 60% might impair spot quality. A versatile spotting layout for 32 samples in four dilutions and duplicate spots is displayed in Fig. 3. This layout also has proved to be very robust in terms of the data quality generated. The array consists of 16 columns × 22 rows of spots at a pitch of 300 mm. Six arrays of this dimension are printed on one ZeptoMARK chip. 10. Spotted and blocked ZeptoMARK chips may be stored for up to 1 year at 4°C and will still provide consistent protein expression and activation profiles. 11. The ZeptoREADER couples a laser beam into the ZeptoMARK chip’s waveguide and acquires fluorescence images of the arrays using a cooled CCD camera (see Fig. 4). The planar waveguide technology incorporated in the ZeptoMARK chips and the ZeptoREADER provides an enhanced detection sensitivity compared to confocal scanners (26, 27). This allows for the detection of low abundant markers in spotted lysate samples. 12. Trademarks used are the property of their respective owners.

54

Cid, Kauffmann, and Molina

Acknowledgments We thank C. Molero, I. Rodríguez-Escudero, A. Alemán, R. Rotger, and other members of our lab for their help and support. M. Ehrat, G.M. Kresbach, and J. van Oostrum are acknowledged for their critical comments on the manuscript. This work was possible thanks to Grants BIO2007-67299 from Ministerio Educación y Ciencia and S-SAL-0246-2006 from Comunidad Autónoma de Madrid (Spain) to M. M. References 1. Paweletz CP, Charboneau L, Bichsel VE, Simone NL, Chen T, Gillespie JW et al (2001) Reverse phase protein microarrays which capture disease progression show activation of pro-survival pathways at the cancer invasion front. Oncogene 20:1981–1989 2. Wulfkuhle JD, Aquino JA, Calvert VS, Fishman DA, Coukos G, Liotta LA et al (2003) Signal pathway profiling of ovarian cancer from human tissuespecimens using reverse-phase protein microarrays. Proteomics 3:2085–2090 3. Grubb RL, Calvert VS, Wulkuhle JD, Paweletz CP, Linehan WM, Phillips JL et al (2003) Signal pathway profiling of prostate cancer using reverse phase protein arrays. Proteomics 3:2142–2146 4. Spurrier B, Honkanen P, Holway A, Kumamoto K, Terashima M, Takenoshita S et al (2008) Protein and lysate array technologies in cancer research. Biotechnol Adv 26:361–369 5. Bhavsar AP, Guttman JA, Finlay BB (2007) Manipulation of host-cell pathways by bacterial pathogens. Nature 449:827–834 6. Hueck CJ (1998) Type III protein secretion systems in bacterial pathogens of animals and plants. Microbiol Mol Biol Rev 62:379–433 7. Coburn B, Sekirov I, Finlay BB (2007) Type III secretion systems and disease. Clin Microbiol Rev 20:535–549 8. Galan JE, Wolf-Watz H (2006) Protein delivery into eukaryotic cells by type III secretion machines. Nature 444:567–573 9. Stavrinides J, McCann HC, Guttman DS (2008) Host-pathogen interplay and the evolution of bacterial effectors. Cell Microbiol 10:285–292 10. Angot A, Vergunst A, Genin S, Peeters N (2007) Exploitation of eukaryotic ubiquitin signaling pathways by effectors translocated by bacterial type III and type IV secretion systems. PLoS Pathog 3:e3

11. Molero C, Rodríguez-Escudero I, Alemán A, Rotger R, Molina M, Cid VJ (2009) Addressing the effects of Salmonella internalization in host cell signaling on a reverse-phase protein array. Proteomics 9:3652–3665 12. Galan JE (1999) Interaction of Salmonella with host cells through the centisome 63 type III secretion system. Curr Opin Microbiol 2:46–50 13. Steele-Mortimer O (2008) Infection of epithelial cells with Salmonella enterica. Methods Mol Biol 431:201–211 14. Hardt WD, Chen LM, Schuebel KE, Bustelo XR, Galan JE (1998) S. typhimurium encodes an activator of Rho GTPases that induces membrane ruffling and nuclear responses in host cells. Cell 93:815–826 15. Stender S, Friebel A, Linder S, Rohde M, Mirold S, Hardt WD (2000) Identification of SopE2 from Salmonella typhimurium, a conserved guanine nucleotide exchange factor for Cdc42 of the host cell. Mol Microbiol 36:1206–1221 16. Zhou D, Chen LM, Hernandez L, Shears SB, Galan JE (2001) A Salmonella inositol polyphosphatase acts in conjunction with other bacterial effectors to promote host cell actin cytoskeleton rearrangements and bacterial internalization. Mol Microbiol 39:248–259 17. Dukes JD, Lee H, Hagen R, Reaves BJ, Layton AN, Galyov EE et al (2006) The secreted Salmonella dublin phosphoinositide phosphatase, SopB, localizes to PtdIns(3) P-containing endosomes and perturbs normal endosome to lysosome trafficking. Biochem J 395:239–247 18. Hernandez LD, Hueffer K, Wenk MR, Galan JE (2004) Salmonella modulates vesicular traffic by altering phosphoinositide metabolism. Science 304:1805–1807 19. Terebiznik MR, Vieira OV, Marcus SL, Slade A, Yip CM, Trimble WS, Meyer T et al (2002) Elimination of host cell PtdIns(4, 5)P(2) by

Reverse Protein Arrays Applied to Host–Pathogen Interaction Studies bacterial SigD promotes membrane fission during invasion by Salmonella. Nature Cell Biol 4:766–773 20. Rodríguez-Escudero I, Rotger R, Cid VJ, Molina M (2006) Inhibition of Cdc42dependent signaling in Saccharomyces cerevi siae by phosphatase-dead SigD/SopB from Salmonella typhimurium. Microbiology 152:3437–3452 2 1. Steele-Mortimer O, Knodler LA, Marcus SL, Scheid MP, Goh B, Pfeifer CG et al (2000) Activation of Akt/protein kinase B in epithelial cells by the Salmonella typh imurium effector sigD. J Biol Chem 275:37718–37724 22. Hoiseth SK, Stocker BA (1981) Aromaticdependent Salmonella typhimurium are nonvirulent and effective as live vaccines. Nature 291:238–239 23. Bevington PR, Robinson DK (2002) Data Reduction and Error Analysis for the Physical Sciences. Mc Graw-Hill, New York

55

24. Van Oostrum J, Calonder C, Rechsteiner D, Ehrat M, Mestan J, Fabbro D et al (2009) Tracing pathway activities with kinase inhibitors and reverse phase protein arrays. Proteomics Clin Appl 3:412–422 25. Alemán A, Rodríguez-Escudero I, Mallo GV, Cid VJ, Molina M, Rotger R (2005) The amino-terminal non-catalytic region of Salmo nella typhimurium SigD affects actin organization in yeast and mammalian cells. Cell Microbiol 7:1432–1446 26. Ghatnekar-Nilsson S, Dexlin L, Wingren C, Montelius L, Borrebaeck CAK (2007) Design of atto-vial based recombinant antibody arrays combined with a planar waveguide detection system. Proteomics 7: 540–547 27. Duveneck GL, Abel AP, Bopp MA, Kresbach GM, Ehrat M (2002) Planar waveguides for ultra-high sensitivity of the analysis of nucleic acids. Analytica Chimica Acta 469:49–61

wwwwwwwwwwwwwwww

Chapter 5 Identification and Optimization of DNA Aptamer Binding Regions Using DNA Microarrays Nicholas O. Fischer and Theodore M. Tarasow Abstract DNA aptamers are versatile recognition elements for pharmaceutical, diagnostic, and life science applications. Identification and optimization of the minimal functional sequence after aptamer selection is a bottleneck for developing aptamer applications. DNA microarray technology proved a facile means for screening thousands of aptamer sequence permutations to identify functional aptamer domains. This chapter describes the detailed methodology for designing aptamer arrays to identify minimal aptamer binding domains as well as elucidating the relationship between aptamer structure and function, using immunoglobulin E as a model protein. Key words: Aptamers, DNA arrays, Aptamer arrays, Aptamer characterization, Aptamer optimization

1. Introduction The discovery and optimization of high-affinity binding molecules are extremely important for medical and life sciences. Myriad technological approaches have been employed over the past decades to develop and tune the affinity reagents for pharmaceutical, diagnostic, and life science applications. Tailoring affinity reagents for particular applications, however, requires optimization of the reagent properties, including size, specificity, affinity, and chemical properties. Traditional approaches to reagent optimization have been onerous, being both labor and time intensive. Any approach that alleviates this bottleneck will greatly simplify the task of tailoring affinity reagents for any given application. While many types of affinity reagents are in use today, nucleic acid aptamers provide a number of characteristics that make them attractive for an array of applications. Of particular interest are single-stranded DNA aptamers, as they are traditionally small in size (15–30 nucleotides in length), can bind their protein targets

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_5, © Springer Science+Business Media, LLC 2011

57

58

Fischer and Tarasow

with dissociation constants in the picomolar to nanomolar range (1), can discriminate against closely related targets (2), can be chemically synthesized and modified, can be routinely denatured and renatured, and are relatively stable under a variety of conditions (3). Furthermore, they are typically derived using a proven and scalable in vitro process termed systematic evolution of ligands by exponential enrichment (SELEX), by which specific aptamer sequences are isolated from extremely large pools of random oligonucleotide sequences using an iterative selection process (4, 5). Non-SELEX methods have also been used for aptamer identification in recent years (6, 7). Regardless of the approach used for aptamer identification, optimization is required to realize the utility of aptamers as biomolecular affinity platforms. In particular, two obstacles exist that preclude the efficient and rapid identification of optimized, high-affinity DNA aptamers. First, a large number of target binding sequences are typically identified at the end of an aptamer selection process, ranging from 10 to 200 individual sequences. To find the sequence with the highest binding affinity and specificity, binding measurement with each individual aptamer sequence is required. Secondly, while the length of the ssDNA libraries is typically 80 nucleotides (40 nt random region flanked by two 20 nt priming regions), only a portion of this sequence is actually required for analyte binding (8) and that minimal sequence is likely not optimal for a given set of desired properties (e.g., affinity, stability, etc.). Identifying the minimal functional sequence requires the sequential truncation of the parent aptamer sequence (9). This iterative process is both time and labor intensive, requiring months to interrogate just a few sequences. Similarly, identifying the optimal sequence requires single point and covariation mutational experiments that also are time and resource intensive. Hence, a process that can simultaneously and discretely survey thousands of individual aptamer sequence permutations would greatly accelerate this process. The application of in situ synthesized, high-density DNA microarray technology has been demonstrated to increase by orders of magnitude the number of aptamer sequences that can be surveyed for binding properties. All possible truncations and mutations of hundreds of parent aptamer sequences can be interrogated simultaneously in a single microarray experiment, quickly identifying the minimal functional aptamer sequence of highest affinity. Importantly, probing the contribution of individual nucleotides within the aptamer sequence provides a wealth of information to accelerate our understanding of aptamer sequence structure and function relationships (10, 11). Subsequent applications of arraybased aptamer synthesis and characterization have also been used to guide an in silico SELEX processes. Ultimately, array-based optimization of aptamers can be used to generate multiplexed, high-density aptamer arrays for analyte identification and classification for pharmaceutical, diagnostic, and life science applications.

Identification and Optimization of DNA Aptamer Binding Regions

59

2. Materials The materials listed below were tailored for Agilent DNA microarrays. However, similar items would be required regardless of microarray vendor. IgE was used as a model protein target, using previously identified aptamers against IgE for the aptamer array design (10, 12). 2.1. Aptamer Microarray

1. Custom designed Agilent microarray. Important parameters to be considered when designing the array are the length of the oligonucleotides and the number of subarrays (see Note 1). 2. SureHyb DNA Microarray Hybridization Chamber Kit (Agilent, Santa Clara, CA), which includes the chamber base, chamber cover, and clamp assembly. 3. 8-Chambered gasket slides (Agilent). 4. Agilent Landing Lights: Fiducial complementary ssDNA, modified with Cy3 and HPLC purified (Integrated DNA Technologies, Coralville, IA): 5¢-CCAGTGACTT TCGTCA CTGG AAAACGATCG TTTCCGATCG AAAAGCTAGC TTTCGCTAGC/3Cy3Sp/-3¢.

2.2. Fluorescent Labeling of IgE

1. Immunoglobulin E from human myeloma plasma (Athens Research & Technology) e280 = 272,000 M−1 cm−1. 2. DyLight 549 Microscale Antibody Labeling Kit (Pierce, Rockford, IL). DyLight fluorophores are activated with N-hydroxysuccinimide (NHS) esters and are amine reactive. Kit includes borate buffer, which is recommended for the conjugation reaction.

2.3. Array Blocking, Probing, and Washing

1. IgE binding buffer (PBS-M): 138 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 1.1 mM KH2PO4, 1 mM MgCl2, pH 7.4. Store at room temperature. 2. Wash buffer (PBS-MT): 0.1% (v/v) Tween-20 (SigmaAldrich, St. Louis, MO), dissolved in PBS-M. Store at room temperature. 3. Casein blocking solution: 1% (w/v) Casein (Sigma-Aldrich), dissolved in PBT-MT. Prepare fresh and store at 4°C until ready to use (see Note 2). 4. BSA blocking solution: 5% (w/v) BSA (Sigma-Aldrich), dissolved in PBT-MT. Prepare fresh and store at 4°C until ready to use. 5. Dextran sulfate blocking solution: Dextran sulfate (SigmaAldrich) is dissolved at 0.2% (w/v) in PBS-MT. Aliquot and store at −20°C.

60

Fischer and Tarasow

3. Methods 3.1. Aptamer Microarray Design

1. Sequence design: We have found that two sets of serial truncations (i.e., serial 5¢ and serial 3¢ truncations) provide significant information regarding sequence dependent binding (Fig. 1) (see Note 3). It is suggested that all aptamer sequences incorporate a 3¢ poly-T spacer, minimally ten nucleotides in length (11) (see Note 4). As DNA arrays are synthesized from the 3¢ end (i.e., the 5¢ end of oligo is solvent exposed and 3¢ end of oligo is immobilized), the poly-T spacer is appended to the 3¢ end of the aptamer sequence. In the absence of tailored software to generate the desired sequence/spacer pairings, Microsoft Excel can be readily employed. When compiled in a spreadsheet using basic Excel tools (in particular the “text and data” functions), sequences can be rapidly prepared and appended (see Notes 5 and 6). 2. Array design: The array design depends on the number of sequences, truncations, and independent conditions to be studied. We found that arrays featuring eight identical subarrays provide a facile way to interrogate blocking and incubation parameters simultaneously. Once parameters are optimized, single arrays with up to 48,000 features can then be used to maximize the number of sequences to be tested. The final aptamer array design needs to include fiducials (to enable alignment of the scanned image for feature extraction) and randomization of sequence samples in triplicate.

a 5’

3’

5’ truncations

5’

3’ truncations

3’ full aptamer sequence

b 5’

5’ truncations

3’ binding sequence

5’

3’ truncations

3’ poly-T spacer

Fig. 1. Design of the aptamer array and serial truncations of the full length DNA aptamer. As the serial truncation remove nonbinding sequences, the ability of the truncate to bind the target is increased. Subsequent truncations that remove bases comprising the binding sequence will abrogate binding. Two schemes for designing the truncations are depicted. (a) Serial truncations with a fixed length poly-T spacer will result in features with oligonucleotides of variable lengths. (b) Serial truncations with compensatory T additions to the poly-T spacer for every base removal will result in features with oligonucleotides of identical length. Maintaining the length and distance from the chip across all truncates eliminates a variable in a and should provide additional consistency when comparing the activity of truncates. Figure 1 was modified from Fischer et al. (10).

Identification and Optimization of DNA Aptamer Binding Regions

61

This is typically accomplished by vendor or by individual using vendor software. The procedure to fluorescently label IgE using the DyLight 549 Microscale Antibody Labeling Kit is outlined below. This procedure will require modification if other labeling methods are used.

3.2. Labeling of IgE

1. Dilute IgE to 1 mg/mL with PBS, and supplement with borate buffer to 50 mM to adjust pH to 8.5. 2. Add IgE solution to preweighed DyLight ester, invert gently to mix. 3. Incubate the reaction for 1 h with gentle agitation and protect from light. 4. To remove unreacted dye, add reaction mixture to Purification Resin equilibrated with PBS, and gently mix. Centrifuge for 30 s at 1,000 × g and collect the purified protein. 5. Dialyze the labeled protein against PBS-M overnight at 4°C. 6. Determine ratio of dye incorporation according to manufacturer’s instructions. 7. Aliquot labeled protein and store at −20°C. 3.3. Array Blocking/ Probing/Washing

1. Add 55 mL of blocking solution to the wells of the gasket slide (see Fig. 2) (see Note 7). b

a 1: BSA (BSA)

2: BSA (none)

3: BSA+DS (none)

350

4: Casein (Casein)

5: Casein (none)

6: Casein+DS (none)

Signal:noise ratio

300 250 200 150 100 50 0

1 2 3 4 5 6 Blocking conditions

Fig. 2. Optimization of aptamer array blocking and binding parameters to maximize signal-to-noise ratios. (a) Representative regions of identical subarrays on the same array blocked and probed under varying conditions. Individual subarrays were blocked with either bovine serum albumin (BSA) or casein in the presence or absence of dextran sulfate (DS). Constituents of the binding buffer containing IgE are indicated in parentheses (BSA, casein, or none). The combination of casein during both blocking and binding resulted in highest feature intensity and uniformity. BSA used in both blocking and binding produced features with a pronounced halo effect. All other tested parameters provided features of comparatively lower intensities. (b) The signal-to-noise ratio was determined by comparing the average fluorescence intensity of the highest intensity feature to a negative control. In the case of IgE, the highest signal-to-noise ratio was achieved with casein in both blocking and binding steps. The graph depicts averages of triplicate data points, and error bars represent 1 S.D. Figure 2 was modified from Fischer et al. (10).

62

Fischer and Tarasow

2. Place the slide onto the gasket slide so that the active microarray surface is in contact with the solution, then close and tighten the chamber. Incubate for 1 h at room temperature (see Note 8). 3. Remove the slide from the chamber, remove the gasket, and gently place slide into a basin containing PBS-M (40 mL). No extensive washing is required at this step. 4. Add 55 mL of the sample (10 nM labeled IgE and ~1 nM fiducial ssDNA) diluted in appropriate blocking solution to each well of the gasket slide (see Note 9). 5. Place slide onto gasket slide, close and tighten chamber, incubate in the dark for at least 2 h at room temperature. 6. Carefully remove slide from chamber, remove gasket, and place slide into basin with >50 mL PBS-MT on oscillating platform for 3 min. Repeat for a total of 5 washes. If desired, final wash can be conducted in PBS-M to remove residual Tween-20. 7. To dry slide, place in a clean 50 mL conical tube and centrifuge 5 min at 2,000 × g. 8. To store the slide prior to analysis, place in a clean conical tube purged with N2. 3.4. Array Scanning and Data Analysis (Depends on Array Manufacturer)

1. Aptamer array scanning: Array scanning and feature extraction details will depend on the array vendor and scanner model. For the Agilent arrays, slides were scanned using the Agilent DNA microarray scanner at a 5 m resolution. In cases where feature saturation was observed, the slides were rescanned using the extended dynamic range (XDR) feature. The fluorescence intensity at each feature was measured using Agilent’s Feature Extraction Software and compiled into a database for subsequent analysis. 2. Data analysis: Microsoft Access (or similar database software) can be used to maintain the complete array database and facilitate sorting and filtering of the array data (see Note 5). Microsoft Excel can be used for relatively rapid graphical analysis of the sequence–function relationship when the fluorescence intensity of the truncations are graphed sequentially (see Fig. 3). M-Fold online software (13) enables correlation of aptamer structure and function by predicting thermodynamic parameters of oligonucleotide secondary structure (see Note 10). By conducting the analysis to display all possible structures for each sequence, even those that are thermodynamically unfavorable, the interplay between aptamer structure can be mapped to the function (i.e., fluorescence binding data) (see Fig. 4) (see Note 11).

Identification and Optimization of DNA Aptamer Binding Regions

63

Fluorescence Intensity (a.u.)

Consensus Sequence 10000

1000 5’ truncations 3’ truncations

100 5’

T T T AT CCGT T CC T CC T AGT GG

3’

Group 1 binding structure

T A T T T

C

C G T GA

C T C T C A

1

2

Region 3

4

5

6 lowest energy structure

1.0 0.5

no 2° structure present

0.0

4000 3500 3000 2500 2000 1500 1000 500 0

Group 2 non-binding structure 1

Group 3 non-binding structure 2 ... N

G C T T A T CT C T A T A A T C G A G C C G T G A T G G G C N ... A T

-0

-5

-10

-15

-20

-25

-30

-35

-40

Fluorescence Intensity (a.u.)

C G GGGT G T A TC A T T A A GC AT GC AT GC T A AT G C T A T ... N C G N ... T T A A T T A C C G G A C A T G G T A C G T A T C A C G G A T C G G A C G G C ... N A T N ...

Fraction of lowest free energy

Fig. 3. Interrogation of aptamer sequence and function elucidates the minimized functional binding domains. Serial truncations from both 5¢ and 3¢ ends of the aptamer sequence are superimposed. A disruption of target binding, observed as a decrease in fluorescence intensity, occurs when key nucleotides are removed. By superimposing the 5¢ and 3¢ serial truncations of the parent aptamer (filled and open circle, respectively), the minimally required IgE binding sequence is represented by those truncations exhibiting nominal fluorescence. In the case of the IgE clone (D-12.0), the experimentally determined binding region corresponds to the IgE consensus binding sequence TTTATCCGTTCCTCCTAGTGG. Figure 3 was modified from Fischer et al. (10).

-45

Serial 5’ truncations

Fig. 4. The free energy of observed secondary structures (IgE clone D-66.0) can be correlated to observed fluctuations in fluorescence intensity upon IgE binding, providing insight into the relationship between structure and function. Three dominant structural groups for the truncation set were predicted by Mfold (left ): (1) a stable stem:loops structure presenting the complete binding sequence in the loop, (2) a stem–loop structure, whereby the binding sequence is involved in both stem and loop formation, (3) a collection of short, random secondary structures. Aptamer bases required for IgE binding are highlighted. The corresponding DG values of these folding groups are plotted (top) in relation to overall fluorescence intensity observed at each 5¢ truncate (bottom). Six unique regions correlating free energy of the competing structural groups with fluorescence intensity are identified: (1) increased accessibility to binding domain by removal of 5¢, nonbinding bases, (2) higher propensity for the correct consensus fold due to lower stability of Group 2, (3) stem disruption in Group 1, increased stability of Group 2, and introduction of Group 3 compete with Group 1 formation, (4) Group 2 is completely abrogated, making Group 1 the most favorable secondary structure, (5) low stability in Group 1 stem limits proper folding, but binding of IgE target can readily induce proper folding, (6) Group 1 fold is no longer possible as bases required for stem formation and IgE binding are removed. Mfold structures were calculated using experimental parameters (25°C, 137mM NaCl, 1 mM MgCl2). Error bars are 1 SD of triplicate data points and all DG values are in kcal/mol. Figure 4 was modified from Fischer et al. (10).

64

Fischer and Tarasow

4. Notes 1. Typically, the standard oligonucleotide length on DNA arrays is ca. 60 nt. Customized arrays featuring longer oligonucleotides (up to ca. 100 nt) can be prepared, although at a higher cost. Ideally, aptamer constant/primer regions should be included in the design of the aptamer arrays, as these regions may be involved in secondary or tertiary structures required for binding. However, testing the aptamer variable region may suffice in some cases. 2. The maximal solubility of casein is 1%, and requires heating of the casein solution to 60°C for ca. 30 min. The remaining insoluble material is pelleted using a tabletop centrifuge (4,000 × g for 1 min) and passed through a 0.22 mm syringe filter. 3. Alternatively, more comprehensive truncation sets can be tested, such as single base truncations from both 5¢ and 3¢ and/or single base 5¢ truncations with two-base 3¢ truncations. Other groups have used similar approaches to probe mutational impact on and optimization of target binding. 4. In our original aptamer array design, the poly-T spacers were maintained at ten bases for all serial truncations. However, based on our observations (10) and those of others (11), longer spacers may provide the target protein great accessibility to the aptamer binding region. In particular, we observed an erosion of signal in all 3¢ truncation series, which we attribute to a decrease in the target protein’s accessibility to the binding sequence due to accessibility factors. As the 3¢ truncations remove a nucleotide from the base of the aptamer, the aptamer binding region is “pulled” toward the array surface. To ameliorate this effect, we suggest compensating the removal of each base from the aptamer by adding a base to the poly-T spacer, in effect maintaining the position of the binding region constant with respect to the array surface. This is reflected in Fig. 1b. 5. The Text and Data Function tools found in Microsoft Excel can be used to rapidly generate serial truncations, from both 3¢ and 5¢ ends. Also, tools exist to stitch together multiple text strings, facilitating the incorporation of the necessary poly-T spacer sequences into the aptamer array design. The RIGHT/LEFT tools allow one to separately prepare pairings of truncated sequences and appropriately sized poly-T spacers RIGHT(text [number_chars]) = provides indicated number of characters starting from right of the text.

Identification and Optimization of DNA Aptamer Binding Regions

65

LEFT(text [number_chars]) = provides indicated number of characters starting from the left of the text. The CONCATENATE tool allows stitching the truncated aptamer sequence to the poly-T spacer. CONCATENATE(text1, [text2], [text3], …) = joins text strings in a single cell. 6. When naming individual sequences, sufficient identifying information should be included in separate columns. This will facilitate any downstream sorting and filtering of fluorescence intensity data, providing for more agile analysis. 7. The 55 mL sample solution volume is optimized for an 8-chambered gaskets slide. From this point on, avoid letting the slide dry completely. 8. We found that stationary incubation works well for our aptamer arrays. When using stationary incubation, it is necessary to use caution during slide/gasket assembly to avoid the formation of air bubbles. Any trapped air will prevent contact between the array surface and the target solution, resulting in uneven target binding. Arrays designed with triplicate features, however, may provide sufficient redundancy to salvage data collected from a subarray compromised by an air bubble. Conversely, the standard incubation of DNA arrays is facilitated by the presence of a single, large mixing bubble within the chamber. Vertical rotation using a hybridization oven will ensure proper mixing. 9. The aptamer arrays were incubated with a range of IgE concentrations from 0.1 to >300 nM. IgE binding was observed at 0.1 nM IgE with a substantial signal-to-noise ratio, suggesting that significantly less protein is required. This may prove useful for protein targets that are in limited supply. 10. The Mfold server (version 3.2) can be accessed at http://mfold. bioinfo.rpi.edu/cgi-bin/dna-form1.cgi. The Quikfold server, which facilitates simultaneous analysis of multiple sequences, can be found at http://dinamelt.bioinfo.rpi.edu/quikfold.php. 11. For Mfold analysis, ensure that conditions during target incubation are represented (e.g., 37 mM Na+, 1 mM Mg++, 25°C).

Acknowledgments This work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344 with support from Lawrence Livermore National Laboratory (LLNLJRNL-415696).

66

Fischer and Tarasow

References 1. Proske D, Blank M, Buhmann R, Resch A (2005) Aptamers – basic research, drug development, and clinical applications. Appl Microbiol Biotechnol 69:367–374 2. Green LS, Jellinek D, Jenison R, Ostman A, Heldin CH, Janjic N (1996) Inhibitory DNA ligands to platelet-derived growth factor B-chain. Biochemistry 35:14413–14424 3. Nimjee SM, Rusconi CP, Sullenger BA (2005) Aptamers: an emerging class of therapeutics. Annu Rev Med 56:555–583 4. Tuerk C, Gold L (1990) Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase. Science 249:505–510 5. Ellington AD, Szostak JW (1990) In vitro selection of RNA molecules that bind specific ligands. Nature 346:818–822 6. Berezovski M, Musheev M, Drabovich A, Krylov SN (2006) Non-SELEX selection of aptamers. J Am Chem Soc 128:1410–1411 7. Knight CG, Platt M, Rowe W, Wedge DC, Khan F, Day PJ, McShea A, Knowles J, Kell DB (2009) Array-based evolution of DNA aptamers allows modelling of an explicit

8. 9.

10.

11.

12.

13.

sequence-fitness landscape. Nucleic Acids Res 37:e6 Jayasena SD (1999) Aptamers: an emerging class of molecules that rival antibodies in diagnostics. Clin Chem 45:1628–1650 Shangguan D, Tang ZW, Mallikaratchy P, Xiao ZY, Tan WH (2007) Optimization and modifications of aptamers selected from liver cancer cell lines. Chembiochem 8:603–606 Fischer NO, Tok JBH, Tarasow TM (2008) Massively parallel interrogation of aptamer sequence, structure and function. PLoS ONE 3:e2720 Katilius E, Flores C, Woodbury NW (2007) Exploring the sequence space of a DNA aptamer using microarrays. Nucleic Acids Res 35:7626–7635 Wiegand TW, Williams PB, Dreskin SC, Jouvin MH, Kinet JP, Tasset D (1996) Highaffinity oligonucleotide ligands to human IgE inhibit binding to Fc epsilon receptor I. J Immunol 157:221–230 Zuker M (2003) Mfold web server for nucleic acid folding and hybridization prediction. Nucleic Acids Res 31:3406–3415

Chapter 6 Recombinant Lectin Microarrays for Glycomic Analysis Daniel C. Propheter, Ku-Lung Hsu, and Lara K. Mahal Abstract The cell surface is covered with a myriad of carbohydrates that form a complex matrix of oligosaccharides. Carbohydrate recognition plays critical roles in pathogenesis, trafficking, and differentiation. Lectin microarray technology presents a novel platform for the high-throughput analysis of these structurally diverse biopolymers. One drawback of this technology has been limitations imposed by the commercially available plant lectins used in the array. Not only are a majority of these plant-derived proteins glycosylated, which can complicate glycomic analysis, but they also differ in activity and availability. Our lab has recently introduced recombinant lectins to enhance the stability and scope of our lectin panel. Herein, we provide a detailed procedure for the expression of bacterially-derived lectins and their application to a recombinant lectin microarray. Key words: Recombinant lectin, Microarray, Glycomics, Glycosylation

1. Introduction Glycomics, the high-throughput analysis of carbohydrates, is an arduous task given the inherent structural and chemical properties of glycans (1, 2). The carbohydrates of the mammalian glycome are complex, consisting of linear and branched polymers of structural isomers. The development of lectin microarray technology has helped to address these issues by using carbohydratebinding proteins (lectins), which can discriminate glycan structures and linkages, to analyze the glycome in a high-throughput manner (3, 4). Although this method has proven quite useful, the number of commercially available lectins, most of which are plant-derived, limits the level of structural resolution observed with the microarray. In addition, plant lectins often show lotto-lot differences in activity and availability due to their purification from natural sources. Recombinant lectins have helped to address these issues by taking advantage of bacterial expression

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_6, © Springer Science+Business Media, LLC 2011

67

68

Propheter, Hsu, and Mahal

systems, which allow for more stringent quality control and systematic purification techniques. Since bacteria require lectins for host-pathogen interactions and few bacterial lectins have been identified, bacterial genomes may represent a voluminous source of lectins (5). Furthermore, bacterial lectins are not known to be glycosylated (6). Therefore, our lab has recently cloned and purified a small set of recombinant lectins, and analyzed their glycanbinding profiles (7). These lectins are a nice complement to the commercially available lectin panel and show great promise as a scaffold for creating an expansive and diverse set of glycan-binding proteins via directed evolution.

2. Materials 2.1. Cloning and Purification of Recombinant DNA

1. Purified genomic DNA (may be purchased or isolated from natural source). 2. pET-41 Ek/LIC vector kit (Novagen, San Diego, CA). Kit contains: 1 mg pET-41 Ek/LIC vector, 10 mL pET-41 Ek/ LIC control vector, 25 units of T4 DNA polymerase (LIC qualified), 50 mL 10× T4 DNA polymerase buffer, 100 mM DTT, 50 mL 25 mM EDTA, 40 mL 25 mM dATP, 1.5 mL nuclease-free H2O, 22 × 50 mL NovaBlue Competent Cells (DH5a), 0.2 mL BL21(DE3) competent cells, 0.2 mL BL21(DE3)pLysS cells, 5 × 2 mL SOC media, and 10 mL test plasmid. 3. MJ Mini Personal Thermal Cycler (Bio-Rad, Hercules, CA). 4. Purified PCR insert: 0.2 pmol (after PCR). 5. Taq polymerase and dNTP solutions (Novagen). 6. Electrocompetent DH5a (Invitrogen, Carlsbad, CA). 7. Agarose, ethidium bromide, and kanamycin sulfate (Thermo Fisher Scientific, Rockford, IL). 8. Agarose gel-running equipment (Bio-Rad). 9. QIAprep spin miniprep kit (Qiagen, Valenica, CA). 10. Luria Broth (LB): 10 g tryptone, 5 g yeast extract, 10 g NaCl, 1 L of H2O (autoclave). 11. LB-Agar kanamycin plates: 4 g tryptone, 2 g yeast extract, 4 g NaCl, 6 g BactoAgar, 400 mL H2O (autoclave). After autoclaving the LB-agar mixture, let cool until warm to the touch, then add kanamycin sulfate (30 mg/mL final concentration). Pour solution into 100 × 15 mm petri dishes and let stand at room temperature for 20 min. Store at 4°C for up to 3 months. 12. 50× TAE buffer: 2.0 M Tris-base, 0.1 M ethylenediamine tetraacetic acid disodium dehydrate, 57.1 mL glacial acetic acid, 1 L ddi H2O, pH 8.5.

Recombinant Lectin Microarrays for Glycomic Analysis

2.2. Protein Expression and Purification

69

1. Electrocompetent BL21(DE3) cells (Novagen). 2. Terrific broth (TB): 12 g tryptone, 24 g yeast extract, 4 mL glycerol, 900 mL total ddi H2O (autoclave). 3. Avanti J-E Centrifuge (Beckman Coulter, Fullerton, CA): Equipped with rotors (JS-5.3 and JA-25.50). 4. Smartspec 3000 (Bio-Rad). 5. IC600 Incubator (Yamato Scientific, Chuo-ku, Tokyo). 6. Micropulser (Bio-Rad). 7. Rotary shaker and incubator (ATR, Laurel, MD). 8. Phosphate Buffer Saline (PBS): 100 mM sodium phosphate, 150 mM sodium chloride, pH 7.4. 9. Lysis buffer: PBS + 0.2% Triton X-100. 10. Aqueous 1,000× protease inhibitor cocktail: 3 mg leupeptin, 5 mg antipain, 12.5 mg pefabloc, 25 mg benzamidine, 50 mg trypsin inhibitor, 2.5 mL aprotin. Store aliquots at −20°C (Thermo Fisher Scientific). 11. Dimethylsulfoxide (DMSO) 1,000× protease inhibitor cocktail: 10 mg chymostatin, 5 mg pepstatin, 2 mL DMSO. Store aliquots at −20°C (components from Thermo Fisher Scientific, Rockford, IL). 12. DNAse (New England Biolabs, Ipswich, MA). 13. BioLogic LP low-pressure gradient chromatography with fraction collected (Bio-Rad). 14. GSTrap HP, 1 mL glutathione column (Amersham, Piscataway, NJ). 15. Slide-A-Lyzer Dialysis Cassette, 3,500 MW cut-off, 0.1–0.5 mL capacity. 16. DC Protein Assay (Bio-Rad). 17. BioTek Synergy HT plate reader (Bio-Tek, Winooski, VT). 18. Lactose, reduced glutathione, Bovine Serum Albumin (BSA) fraction V, and chicken egg white lysozyme (Thermo Fisher Scientific).

2.3. ELISA Activity Assay

1. Sodium azide and o-phenylenediamine hydrochloride (OPD) (Thermo Fisher Scientific). 2. Grenier 96-well Microlon microtiter plates (Grenier Bio One, Monroe, NC). 3. ELISA wash buffer: PBS + 0.05% Tween-20. 4. ELISA blocking buffer: PBS + 5% BSA. 5. PBST++: PBST + 1% BSA, 1 mM CaCl2, 1 mM MgCl2. 6. Anti-His6-horseradish peroxidase conjugated antibody (antiHis6-HRP) (Novus Biologicals, Littleton, CO).

70

Propheter, Hsu, and Mahal

7. OPD solution: 0.4 mg/mL OPD, 0.004% H2O2, 0.1 M phosphate/citrate buffer, pH 5.0. 8. Quenching reagent: 2.5 M H2SO4. 2.4. Recombinant Lectin Microarray

1. Print buffer: PBS, 0.5 mg/mL BSA, 1 mM monosaccharide (see Table 1). 2. Nexterion H slides (Schott North America, Elmsford, NY). 3. Contact microarray printer such as the SpotBot2 (ArrayIt Corp., Sunnyvale, CA) or the Microgrid II (DigiLab, Inc., Holliston, MA). 4. SMP3 pins (ArrayIt Corp.). 5. 16-frame FAST frame hybridization chamber (Schleicher and Schuell, Keene, NH). 6. Genepix 4100A slide scanner and Genepix Pro 5.1 software (Molecular Devices Corporation, Union City, CA). 7. Slide blocking buffer: 50 mM ethanolamine in 50 mM sodium borate, pH 8.0. 8. PBST+: PBS + 0.005% Tween 20, 1 mM CaCl2, 1 mM MgCl2. 9. Slide wash buffer: PBST (PBS + 0.005% Tween 20). 10. Slide spinner (Labnet International, Edison, NJ). 11. Coplin jars. 12. 384-well plates (Whatman, Piscataway, NJ). 13. NHS-Cy3 or Cy5 (GE Healthcare Life Sciences, Piscataway, NJ).

Table 1 Lectin print list Lectin

Source

Sugar in print

Specificity

GafD

F17 fimbriae (Escherichia coli)

GlcNAc

b-GlcNAc

PA-IL

Nonfimbriae (Pseudomonas aeruginosa)

Galactose

Galactose

PA-IIL

Nonfimbriae (P. aeruginosa)

Fucose

Fucose/mannose

PapGII

P-pili (E. coli)

Galactose

GbO4

PapGIII

P-pili (E. coli)

Galactose

GbO5

RS-IIL

Nonfimbriae (Rhizoctonia solanacearum)

Mannose

Mannose/fucose

GafD-m

F17 fimbriae (E. coli)

GlcNAc

b-GlcNAc (80% reduction in binding)

Recombinant Lectin Microarrays for Glycomic Analysis

71

3. Methods Glycan recognition of cell surface carbohydrates using lectin microarrays has proven to be a valuable tool for glycomics (8). By utilizing a simple methodology to clone and express bacterial lectins from microbial genomes, we can expand the detection capabilities of the lectin microarray technology (Fig. 1). The pET-41 Ek/LIC vector system enables both rapid cloning and the incorporation of two affinity tags: A hexa-histidine (His6) and an N-terminal glutathione-S-transferase (GST), for dual modes of purification and detection. The recombinant lectins are overexpressed in BL21 Escherichia coli cells, and desired cells are selected against kanamycin sulfate. Expression and purification use standard methods. Once the recombinant protein is purified, the lectin activity (defined as the signal-to-noise ratio, S/N) is tested against glycoprotein standards. The ELISA assay presented here can be used to test lectin activity against multiple glycoproteins, thereby enabling a wide range of binding profiles to be analyzed. Once the recombinant protein has displayed significant activity (S/N >5), the lectin can be utilized in the microarray. In brief, the recombinant lectin microarray is fabricated by spotting the bacterial lectins onto an N-hydoxysuccinimide-, (NHS-), activated glass slide, and immobilization is achieved through amine-coupling of side-chain lysines.

Fig. 1. Schematic representation of the cloning, expression, and use of bacteria-derived recombinant lectins. The desired lectin is cloned out of the microbial genome and amplified by PCR. The gene is then annealed into the pET-41Ek/LIC vector and expressed into E. coli. The expressed protein is purified, and analyzed for activity using ELISA and microarray techniques. Adapted from Hsu et al. (7).

72

Propheter, Hsu, and Mahal

Fig. 2. Recombinant lectin microarray screening of tumor cell lines (ACHN, and Sk-MEL-5). Samples were prepared and analyzed as previously described (8). In brief, cell membranes were sonicated, and the resulting micellae isolated and labeled with NHS-Cy5. The samples were then incubated with the recombinant microarray (10 mg in 100 mL of buffer) and the arrays processed as previously described. A clear differential pattern can be observed between ACHN and Sk-Mel-5, which is described in the text.

A list of the recombinant bacterial lectins cloned, purified, and added to the lectin microarrays to date is given in Table 1. These lectins come from a variety of bacterial sources and include both adhesins from pili (GafD, PapGII, and PapGIII) and secreted lectins (PA-IL, PA-IIL, and RS-IIL). Incubation of the printed slides with fluorescently labeled samples provides a discernable pattern that gives insight into the extent of glycosylation of a given sample (9). Using this technology, we have shown that even the small panel of recombinant bacterial lectins utilized to date (Table 1) can distinguish tumor cell lines in the NCI-60 panel (Fig. 2). Clear differences can be observed between the renal cell carcinoma ACHN, which shows fucosylation (PA-IIL in the absence of RS-IIL), the presence of terminal b-N-acetyl-d-glucosamine (GafD) and galactosylation (PA-IL), and Sk-Mel-5, which shows an absence of both the terminal GlcNAc and galactose epitopes. Although one can obtain differences with this small of a lectin panel you cannot obtain a comprehensive snapshot of the glycome. However, the inclusion of these lectins in a larger lectin microarray format allows for a far more detailed analysis than is presented herein. 3.1. Cloning and Purification of Recombinant DNA

1. Identify microbial lectin via BLAST, the literature or other sources. 2. Prepare primers flanking the lectin encoding region (see Note 1). 3. Prepare PCRs as follows: 1× reaction buffer, 400 mM dNTP solution, 1 mM 5′ primer, 1 mM 3′ primer, 2.5 units of Taq

Recombinant Lectin Microarrays for Glycomic Analysis

73

Polymerase, and template DNA. Dilute to final volume of 50 mL with ddi H2O. (see Note 2). 4. Place the PCR tubes into the thermocycler and run on the following conditions: 95°C for 10 min, 95°C for 30 s, 45°C for 30 s (Tm), and 72°C for 1 min. Let the reaction go for 40 cycles, and cool PCRs to 4°C (see Note 3). 5. Prepare a 2% w/v agarose gel solution in 1× TAE buffer and heat until mixture is miscible (see Note 4). Add EtBr to a final concentration of 0.5 mg/mL. Pour gel into cast and allow it to solidify. Next, add 5 mL of DNA ladders and 2 mL of PCR mixture. Run gel on 90 V for 45 min, then visualize under UV irradiation. 6. To anneal the PCR insert, first determine the amount of PCR product required for the T4 treatment by using the following formula: (number of base pairs in insert) × 650 × 0.2 pmol = n pg PCR insert. 7. In a sterile 1.5 mL microfuge tube, add the amount of purified PCR product calculated in step 6 (n pg), 2 mL 10× T4 DNA polymerase buffer, 2 mL 25 mM dATP, 1 mL 100 mM DTT, and 0.4 uL 2.5 units/uL T4 DNA polymerase. Add enough ddi H2O to have 20 mL of total volume (see Note 5). 8. Mix the components by flicking the tube and then incubate at room temperature for 30 min. 9. Inactivate the enzyme by incubating at 75°C for 20 min. 10. To anneal into pET-41 Ek/LIC vector, mix 1 mL of the vector with 2 mL of the treated PCR insert in a sterile 1.5 mL microfuge tube and incubate for 5 min at room temperature. 11. Add 1 mL of 25 mM EDTA to the reaction mix, and incubate at room temperature for 5 min (see Note 6). 12. Transform competent DH5a with 1 mL of the annealing reaction, add 1 mL SOC media and allow cells to recover for 1 h, shaking at 250 rpm at 37°C (see Note 7). 13. After 1 h, plate the transformed cells on LB-Agar plates (see Note 8). Allow the cells to grow overnight (~12 h) at 37°C. 14. Pick single colonies and grow in 5 mL of LB (~15 h) with 30 mg/mL kanamycin on a rotary shaker and incubator at 250 rpm at 37 C. 15. After 15 h, take the optical density of the colonies at 600 nm (OD600). Pick the best growing colony and inoculate in 25 mL of LB with kanamycin (30 mg/mL) and place on rotary shaker at 250 rpm, 37°C, for ~15 h. 16. After overnight culture, take OD600, and purify plasmid DNA via Qiagen Miniprep Kit and Instructions. Once DNA is isolated, check the DNA sequence.

74

Propheter, Hsu, and Mahal

3.2. Protein Expression and Purification

1. Transform electrocompetent BL21(DE3) cells with recombinant DNA. Transform 1–5 mg of DNA into a 50 mL aliquot of BL21 cells (see Note 9). 2. Upon electroporation, promptly add 1 mL of LB and grow on a rotary shaker for 1 h at 250 rpm and 37°C. Next, plate cells onto LB-Agar plates and incubate at ~15 h at 37°C. 3. Next, pick single colonies and grow each in 5 mL of LB with kanamycin (30 mg/mL) for 15 h at 250 rpm and 37°C. 4. Take OD600 of colonies, choose a colony with an average rate of growth, and inoculate 5 mL culture into 25 mL culture (see Note 10). Grow the culture to an OD600 of 0.7–1.0, then induce the culture with 1% w/v lactose and grow for 3 h at 250 rpm and 37°C (see Note 11). 5. After 3 h, transfer culture into centrifuge tubes and pellet cells at 6,000 × g, 4°C, 15 min (see Note 12). Discard the supernatant. 6. Resuspend pellet in 1 mL of lysis buffer and dilute 1,000× DMSO and aqueous protease inhibitor cocktails to 1× in lysis buffer (see Note 13). Then add approximately 1 mg/mL lysate of chicken egg white lysozyme and mix at 4°C for 30 min (see Note 14). 7. Next, immediately add DNAse, 5 mg/mL of lysate final concentration, and incubate further at 4°C for 10 min (see Note 15). 8. Centrifuge the samples in the appropriate tubes at 30,000 × g for 30 min at 4°C. Keep the supernatant. 9. Purify the lysate using the BioLogic LP low-pressure gradient chromatography system (or similar system). Load the supernatant onto an equilibrated glutathione column at a flow rate of 0.5 mL/min (see Note 16). Wash the column with ~10 column volumes of PBS at a rate of 1 mL/min. Elute lectin with 10 mM of reduced, free acid glutathione in PBS collecting 1 mL fractions at a rate of 1 mL/min (see Note 17). 10. Monitor lectin purification by 10% SDS-PAGE analysis (see Note 18). Pool fractions containing lectin and dialyze against PBS at 4°C. Aliquot, flash freeze, and store at −80°C (see Note 19).

3.3. ELISA Activity Assay

1. Dilute glycoprotein to a final concentration of 1–10 mg/mL in PBS containing 0.1% NaN3. Take a 96-well plate and coat each well with 100 mL of glycoprotein and incubate for ~12 h at 4°C. 2. Wash each well with wash buffer 5×. Next, add ELISA blocking buffer to each well and incubate at room temperature for 1 h (see Note 20). 3. After blocking, wash each well with wash buffer 5×. Next, dilute lectin into PBST++ (see Note 21). Add 50 mL of each

Recombinant Lectin Microarrays for Glycomic Analysis

75

dilution to each well (see Note 22). Incubate samples at room temperature for 1 h. 4. After incubation, wash each well with wash buffer 5×, then dilute anti-His6-HRP to an optimized dilution in the wash buffer + 1% BSA (see Note 23). Next, add 50 mL of anti-His6HRP solution into each well and incubate at room temperature for 1 h. 5. Wash each well with wash buffer 5×, and freshly prepare OPD reagent buffer. Add 100 mL to each well immediately after the last wash. Incubate at room temperature for 30 min, add 50 mL of stopping reagent to each well, and read on BioTek Plate Reader at 492 nm wavelength (see Note 24). 3.4. Recombinant Lectin Microarray

1. Prepare samples as previously described (8). 2. Dilute lectins to 1 mg/mL in print buffer and 1 mM monosaccharide as specified (see Table 1). 3. Print lectins onto Nexterion H slides using the SpotBot personal microarray with an SMP3 pin. Maintain cold plate at 8°C and internal humidity at 50–60%. 4. Print 5 spots per lectin, to ensure spot quality, on a 16-subarray format (see Note 25). 5. Upon completion of the print, slides are allowed to warm to room temperature in the SpotBot arrayer for 1 h while maintaining humidity control. Slides are then placed into blocking buffer inside a Coplin jar for 1 h at room temperature. 6. After blocking, wash slides with PBST 3× for 3 min, rinse once with PBS, and dry using the slide spinner. 7. Affix a 16-well subarray FAST frame to the slide and incubate with appropriate fluorescent sample for 2 h at room temperature (see Note 26). 8. After incubation, aspirate sample from the subarrays and wash 5× with PBST, once with PBS, and then dry using a slide spinner (see Note 27). 9. Scan slides using the Genepix 4100A scanner at the appropriate wavelength. Extract data using Genepix Pro 5.1 software and analyze using Microsoft Excel and/or Graphpad Prism 4.0.

4. Notes 1. The designed primers must have the overlapping LIC extensions. The forward primer must begin with 5¢ GAC GAC GAC AAG A 3¢ and the reverse primer must begin with 5¢ GAG GAG AAG CCC GG 3¢.

76

Propheter, Hsu, and Mahal

2. You will have to titrate the amounts of template DNA used in the first PCR to obtain the desired results. 3. The melting temperature (Tm) and number of cycles may need to be changed in order to obtain the desired results. 4. Depending on the size of your DNA, you may need to augment the percent of agarose used. 5. pET-41 Ek/LIC vector kit has a control vector which should be used to gauge efficiency of system. Also, it is important to include as a negative control plasmid with no insert to evaluate the selection. T4 DNA polymerase from Novagen is specifically designed for these ligation-independent cloning reactions. Nuclease-free or ddi H2O (i.e., from a purification system) may be used. 6. The T4 treated insert can be stored at −20°C for up to 3 months. 7. SOC may be substituted for LB in this recovery although efficiency may be reduced. Transform the DNA using a Micropulser (Bio-Rad), following the Bio-Rad electroporation protocol (found at http://www.bio-rad.com). 8. For the best results, plate two dilutions of sample. Also, the negative control should be plated to ensure the integrity of the kanamycin. 9. Following the Bio-Rad electroporation protocol referred to earlier (see Note 4). 10. Colony-dependent variations in protein expression arise, so be sure to test ~3 to 5 colonies. Take the best expressing colony and move on to next step. 11. The cultures can be easily scaled up to a 4 L culture 12. This pellet can be stored at −80°C for an indefinite amount of time. 13. Typically we add 4 mL of lysis buffer per 100 mL of culture. If using a 1,000× protease inhibitor cocktail, simply add 1 mL/mL of culture. 14. Be sure to keep all reagents and solutions on ice. 15. If lysate is very viscous, the DNA can be sheared by drawing the suspension through an 18-gauge needle several times. Keep a small aliquot (~50 mL) of the crude lysate for SDS-PAGE gel analysis. 16. We have found that taking the following steps ensures that the column maintains integrity through multiple experiments: First, flush the system with ddi H2O and inspect to make sure no clogs or air are in the system. Second, flow filtered PBS for 5 min at a rate of 1 mL/min. Then load supernatant and follow the protocol. After eluting the column, wash the column

Recombinant Lectin Microarrays for Glycomic Analysis

77

with 5 column volumes of 70% ethanol, followed by 5 column volumes of PBS, and then 1 column volume of 20% ethanol. Store column at 4°C indefinitely. 17. Be sure to keep an aliquot (~50 mL) of the lysate after purification for SDS-PAGE gel analysis. 18. You may need to use a different % acrylamide gel depending on the molecular weight of the desired lectin. 19. Aliquots may be stored for up to 6 months. 20. Be sure to keep the 96-well plate covered to prevent contamination. 21. We recommend serial dilutions for the first ELISA to obtain a larger, more informative dataset. 22. For negative control wells, simply add PBST++ with no lectin. 23. We recommend testing serial dilutions of anti-His6-HRP on initial work to optimize the working dilution. 24. Read ELISA plates also at 620 nm as the reference wavelength. For data analysis, subtract the readings at 620 nm from the 492 nm data set to obtain the true values. The 620 nm value is a background value taken to correct for any imperfections in the sample plate. 25. You can print in a 24-well format and/or limit the number of spots to three, based on previous protocols. Print spots are typically 100 mm. 26. Be sure to keep slides unexposed to light, which affects fluorescence. 27. Be sure to keep slides in the dark. Slides can be kept for long term storage at −20°C. References 1. Mahal LK (2008) Glycomics: towards bioinformatic approaches to understanding glycosylation. Anticancer Agents Med Chem 8: 37–51 2. Hirabayashi J (2008) Concept, strategy and realization of lectin-based glycan profiling. J Biochem 144(2):139–147 3. Pilobello KT, Krishnamoorthy L, Slawek D, Mahal LK (2005) Development of a lectin microarray for the rapid analysis of protein glycopatterns. Chembiochem 6:985–9 4. Hirabayashi J (2004) Lectin-based structural glycomics: glycoproteomics and glycan profiling. Glycoconj J 21(1):35–40 5. Sharon N (2006) Carbohydrates as future anti-adhesion drugs for infectious diseases. Biochim Biophys Acta 1760:527–37

6. Dodson KW, Pinker JS, Rose T, Magnusson G, Hultgren SJ, Waksman G (2001) Structural basis on the interaction of the pyelonephritic E. coli adhesion to its human kidney receptor. Cell 105:733–43 7. Hsu KL, Gildersleeve JC, Mahal LK (2008) A simple strategy for the creation of a recombinant lectin microarray. Mol Biosyst 4:654–62 8. Pilobello KT, Slawek DE, Mahal LK (2007) A ratiometric lectin microarray approach to analysis of the dynamic mammalian glycome. Proc Natl Acad Sci USA 104:11534–9 9. Krishnamoorthy L, Bess JW Jr, Preston AB, Nagashima K, Mahal LK (2009) HIV-1 and microvesicles from T-cells share a common glycome, arguing for a common origin. Nat Chem Biol 5(4):244–250

wwwwwwwwwwwwwwww

Part II Antigen Microarrays for Immunoprofiling

wwwwwwwwwwwwwwww

Chapter 7 Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection Persis P. Wadia, Bita Sahaf, and David B. Miklos Abstract Recombinant antigen arrays represent a new frontier in parallel analysis of multiple immune response profiles requiring only minute blood samples. In this article, we review the benefits and pitfalls of recombinant antigen microarrays developed for multiplexed antibody quantification. In particular, we describe the development of antigen arrays presenting a set of Y chromosome-encoded antigens, called H-Y antigens. These H-Y antigens are immunologically recognized as minor histocompatibility antigens (mHA) following allogeneic blood and organ transplantation. Clinically relevant B-cell responses against H-Y antigens have been demonstrated in male patients receiving female hematopoietic stem cell grafts and are associated with chronic graft versus host development. This chapter discusses our recombinant antigen microarray methods to measure these clinically relevant allo-antibodies. Key words: H-Y proteins, Antibodies, Plasma, Recombinant antigen microarrays, Minor histocompatibility antigens

1. Introduction Identifying, understanding, and confirming complex multicellular processes, such as immunity, require a systems biology approach to integrate each component’s function and regulation within the network. Traditionally, genes and proteins were discovered and characterized in isolation as individual molecules. However, the development of DNA microarrays facilitated multiplexed gene expression pattern analysis in a variety of genomes spanning bacteria (1–3) to human (4, 5). In immunology, gene expression profiling has determined important lymphocyte gene regulation pathways and their linked biological functions (6–9). However, a more complete understanding of adaptive immune responses requires systematic target screening of proteomes isolated from Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_7, © Springer Science+Business Media, LLC 2011

81

82

Wadia, Sahaf, and Miklos

bacteria, viruses, or humans. Antibody secretion marks effective B lymphocyte immune responses, and historically, these antibodies have identified specific targets ex vivo via western blot, immunoprecipitation, or Enzyme-Linked Immunosorbant Assay (ELISA). While Western blot and immunoprecipitation provides qualitative antigen identification from complex lysates, determining the specific protein or responsible gene often requires numerous subsequent biochemical fractionations and sequencing reactions. ELISA quantifies antibody against specific antigens, but their single antigen design consumes precious samples and resources. In contrast, protein microarrays enable high-density presentation of thousands of spatially isolated candidate antigens. Following antibody incubation, specific antigen binding is detected with fluorochrome conjugation. In fact, differential flurochrome conjugation of multiple samples enables multiplexed detection using the same antigen microarray. Ideally, these protein microarrays contain highly-purified antigens (see Note 1) that maintain native protein structure and include posttranslational modifications (see Note 2). In this chapter, we discuss two critical considerations for the generation of recombinant antigen microarrays: (1) the format of the antigens to be printed (Subheading 1.1) and (2) optimization of printing the recombinant protein on printing substrates (Subheading 1.2). We will discuss commercially available microarrays followed by a detailed description of our approach to optimizing the generation of microarrays to express custom antigens (H-Y antigens) for the detection of allo-antibodies (Subheadings 1.3 and 1.4). 1.1. Considerations in Expression of Recombinant Antigens for Protein Microarrays

Posttranslational modifications vary by organisms used for recombinant protein expression. The various organisms used to produce proteins include: Escherichia coli, yeast (10), CHO cells (11), or baculovirus in insect cells (12, 13), and are listed in Fig. 1. The scientific need to preserve posttranslational modifications determines expression system requirements and is also offset by expression efficiency. Modifications such as phosphorylations, acylations, glycosylations, and carboxylations demand a eukaryotic expression system since prokaryotic expression, such as through E. coli, lacks the necessary posttranslational machinery. However, the disadvantage of decreased protein yield through eukaryotic expression is overcome by the decreased antigen requirements for the protein microarray. Nonetheless, bacterial expression will suffice for many recombinant antigen expression needs and remains ubiquitously available, inexpensive, and fast. A significant disadvantage of bacterial expression is the frequent development of inclusion bodies necessitating protein denaturation with subsequent renaturation attempts. Yeast systems and baculovirus-infected insect cells represent reasonable compromises providing proteins in large amounts with eukaryotic modifications.

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection Bacterial cells

Yeast

Insect cells

83

Mammalian cells

Transform cells with gene (with/without tag) Induce expression of protein Small scale purification of expressed protein: select the most robust protein expression colony/culture Large scale purification of expressed protein Secreted protein obtained in the supernatant of cells Native Nickel affinity chromatography

Nonsecreted protein obtained as inclusion bodies Denatured Nickel affinity chromatography Renaturation of purified protein

Quantify and concentrate expressed protein

Print expressed protein of interest

Fig. 1. Flow-sheet for protein expression and purification. A schematic flow-sheet of choosing an expression system and purifying the proteins is detailed in the figure.

However, these systems are more laborious and expensive than prokaryotic systems. Figure 1 presents a schema for the steps involved in antigen purification for recombinant antigen microarrays after the appropriate expression system is chosen (to be discussed in detail in the Subheading 3). The incorporation of epitope tags (GST, V5 or 6xHis tags) for detection and/or isolation of expressed antigens provides a major advantage for recombinant microarray development. Expression plasmids inserting open reading frames (ORF) in frame following N-terminal tags usually provide high-yield protein expression and an affinity tag for protein purification. C-terminal epitope tag recognition indicates the entire ORF has been expressed (see Note 3). One example of a commercially available high-density protein microarray that prints proteins expressed in the baculoviral expression system are Protoarrays™ marketed by Invitrogen (Carlsbad, CA). More than 9,000 human proteins with N-terminal GST epitopes expressed in baculovirus-infected insect cells are affinitypurified and printed in duplicate on nitrocellulose-coated slides. An advantage of using commercial microarrays is that there are

84

Wadia, Sahaf, and Miklos

Fig. 2. Representative figure of a subarray from a commercially available protein microarray with controls and antigen hits. Commercially available protein microarrays contain 48 subarrays with 9,000 proteins printed in duplicate (Protoarray, version 5.0). A representative subarray is shown with negative controls such as Buffer, GST tags in different concentrations, and empty spots. The subarray also contains positive controls, such as anti-human IgG and human IgG, each printed in four concentrations. We use human IgG3 (second highest concentration) and we aim to obtain an MFI of 55,000–60,000 while scanning to normalize our arrays. Alexa 647 is printed in various positions, but fixed positions, across subarrays to help distinguish subarrays while gridding the spotted antigens.

numerous controls printed on each subarray, and once a target has been identified, the protein can be purchased for further analysis or ELISA development analysis. A representative subarray with negative and positive controls is shown in Fig. 2. Negative controls include buffer, empty spots, and GST tags printed in different concentrations and positive controls include human IgG as well as anti-human IgG printed in four different concentrations. Currently, cost prevents wide use of proteome microarrays, but increased content and decreased cost are expected. Our laboratory has extensive experience in using the commercially available protein microarrays from Invitrogen. We used

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

85

Fig. 3. Donor, Pretransplant, and 12-month Posttransplant proteins detected serologically using commercially available protein microarrays. The same representative subarray (Protoarray version 3.0) is shown across three different slides which were processed using donor serum, pre, and posttransplant plasma. One of the two targets identified statistically is shown in the figure (CHAF1b). This was chosen because CHAF1b was absent in the donor and pretransplant plasma sample, but was recognized by antibodies in the posttransplant plasma sample.

protein microarrays to serologically identify Nucleolar and Spindle Associated Protein 1 (NuSAP1) and Chromatin Assembly Factor 1, subunit B (p60) (CHAF1b) as targets of new antibody responses that developed after allogeneic hematopoietic cell transplantation (HCT; Fig. 3). Western blots and ELISA validated their postHCT recognition and enabled ELISA testing of 120 other alloHCT patients with various malignancies. CHAF1b-specific antibodies were predominantly detected in AML patients, whereas NuSAP1-specific antibodies were exclusively detected in AML patients 1 year posttransplant (p < 0.0001). Expression profiles and RT-PCR showed that NuSAP1 was predominately expressed in the bone marrow CD34+CD90+ hematopoietic stem cells, leukemic cell lines, and B lymphoblasts as compared to other tissues or cells. Thus, NuSAP1 is recognized as an immunogenic antigen in 65% AML patients following allogeneic HCT and suggests a tumor antigen role. In conclusion, though protein microarrays is a nascent technology, clinically important tumor antigens can be identified as new antibody targets after allogeneic HCT using high-density protein microarrays (14). As with each new technology, there are advantages and disadvantages. The primary disadvantages of commercially available protein microarrays are that (1) there are new versions available with little bioinformatics support to compare across different versions and (2) thus far they are not normalized for protein concentration. Marina et al. (15) have recently published a concentration-dependent analysis (CDA) method to normalize for the concentration of the spotted antigen on commercially available protein microarrays. Their method is complementary to other commonly employed analyses and demonstrated experimental validation of 92% of hits identified by the intersection of

86

Wadia, Sahaf, and Miklos

CDA with other tools. However, these disadvantages are offset by the many advantages of using recombinant protein microarray antibody detection technology; these include the ability to simultaneously screen (1) thousands of antigens, (2) antigens that reflect varied tissue organs, and (3) thousands of antigens using only small amounts of sample, in a multiplexed, quantitative, reproducible, and rapid fashion. Thus, combining the advantages of antibody-binding specificity and antigen microarray technology provides global humoral immune assessment. 1.2. Considerations in Printing of Recombinant Antigens in Custom Protein Microarrays

Printing substrates vary in their array-binding efficiency, antigen orientation, and epitope availability. Printing proteins by absorption (nitrocellulose and PVDF-coated glass slides have high protein-binding capacity) (16) is a popular method since no further antigen modification is necessary and the bound proteins often retain the functional active-binding sites. However, antigen retention varies and is weaker for absorption printing as compared to affinity or covalent cross-linking of proteins on the array. Some printing substrates enable nonspecific covalent binding such as polylysine or aldehyde-coated glass slides (17, 18). In either absorption or covalent cross-linking methods, proteins will bind to the array in often unpredictable orientations. Substrates linking antigens via affinity-printing can both improve printed substrate retention and maintain antigen orientation. 6xHis epitope-tagged proteins bound onto nickel-coated glass slides are an example of affinity-printing (19–21). Affinity-printing recombinant proteins promise improved antigen retention and consistent orientation. We have used epitope-tag affinity purification of custom proteins and then directly printed the antigens onto nitrocellulosecoated microarrays. We have found that the critical variables that need to be considered to optimize custom printing of recombinant antigen microarrays are: (1) solubility of proteins (see Note 4); (2) print buffer conditions (our H-Y antigens are printed in their 250 mM imidazole elution buffer); and (3) step gradient versus a linear gradient elution step. For example, in our own experience, we have used affinity chromatography in which 6xHis epitopes specifically bind Nickel-NTA resins and are subsequently eluted by increasing concentrations of imidazole. We have found that a linear elution gradient with linearly increasing imidazole concentrations results in broad elution peaks which lead to diluted proteins in the eluate. Thus, instead, we use a step gradient of imidazole concentrations (low to high concentration of imidazole) to obtain sharp elution peaks. This results in smaller aliquots of tagged proteins, but with high protein purity. Antigen printing remains unpredictable, but depends upon printing pin size and type, print buffer, and microarray surfaces. Because heterogeneous antigens yield a spectrum of array binding,

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

87

strategies to confirm and normalize retained antigens are important. Basically, printing pins vary by size and style (solid pins vs. quill pins). Solid pins deposit the antigens while touching the coating surface of the slide, while quill pins dispense repeating aliquots via capillary action, limiting the need for pin reloading (see Note 5). Solid and quill pins vary by their spot sizes with the spot diameter ranging from 62.5 to 600 mm. In our experience, we have used a solid pin (300 mm spot diameter with 360 mm spacing between spots) and have also used a quill pin (100 mm spot diameter with 120 mm spacing between spots; also used for printing the H-Y arrays). The spot sizes can be varied to accommodate more antigens per subarray if needed. The amount of antigen printed, i.e., protein concentrations, should be uniform intra- and interarray for each antigen (see Note 6). Spots printed should also be of the same size and shape. Irregular spots may make further analysis nonreproducible (see Note 7). Also, if many spots are printed from one-time uptake of antigen in the pin, care should be taken to account for evaporation, leading to concentrated antigen spots printed in a later batch as compared to the initial batch of slides (see Note 8). After antigen printing and processing antigen microarrays, bioinformatics for these antigen microarrays printing and detection as well as final analyses need to be standardized. Bioinformatics needs to take into account the local background versus the spot background. Once a spot is gridded, either the software can contour to the shape of the spot or can remain a fixed circle. If the spot remains a fixed circle, the mean fluorescence intensity (MFI) of the spot is determined by the average of the pixels within the enclosed fixed spot (see Note 9). 1.3. Allogeneic Antibodies Against H-Y Antigens Develop After Sex-Mismatched Transplantation

Our group focuses on identifying novel minor histocompatibility antigens (mHA) responsible for graft-versus-leukemic (GVL) effects and graft-versus-host disease (GVHD) after allogeneic HCT. Historically, allogeneic immune responses have been characterized as alloreactive T-cell clones where T-cells play key roles in posttransplant alloimmune responses, but T cell epitope determination remains laborious cell culture-dependent and HLA restricted (22). Our studies have shown that allogeneic B-cell responses against mHA influence clinical outcome (23). Specifically, allogeneic antibodies (allo-Ab) against mHA encoded on the Y chromosome, called H-Y antigens, develop in patients undergoing sex-mismatched HCT in association with both chronic GVHD development and persistent disease remission (24). Female lymphocytes develop devoid of the Y-chromosome and remain naïve to H-Y antigens. After these female lymphocytes are transplanted into male recipients, they recognize these H-Y antigens as foreign and mount a coordinated T- and B-cell immune response (25). Our studies have demonstrated that allogeneic antibodies develop against five H-Y antigens, namely,

88

Wadia, Sahaf, and Miklos

EIF1AY, RPS4Y, DDX3Y, ZFY, and UTY, but not their 91–99% identical X-homologues EIF1AX, RPS4X, DDX3X, ZFX, and UTX, respectively (24). We developed an ELISA panel to characterize the frequency, intensity, and specificity of these H-Y allo-Abs (23). In order to further multiplex and more accurately quantify H-Y and H-X antibodies, we developed a custom H-Y recombinant antigen microarray. The microarray print and development method are described in this chapter. 1.4. H-Y Antigen Microarray Development

Our H-Y microarray presents five human H-Y and H-X proteins that include C-terminal 6xHis and V-5 epitopes printed in quadruplicate on nitrocellulose-coated glass slides. All proteins are expressed in E. coli and isolated using nickel affinity chromatography. Nickel-bound 6-his tagged proteins are eluted by imidazole competition (24). In addition to the custom proteins, positive and negative proteins are also printed on the slide. The positive control proteins printed are human IgG in three different concentrations (IgG1 < IgG2 < IgG3), which should give similar readings across different processing time points because printed IgG recognition depends on uniform secondary antibody application, processing, and flurochrome detection (Fig. 4). To assess and ensure the quality of the batch of slides printed, the first two, middle two, and last two slides are probed with anti-V5 antibody, which will detect all antigens (Fig. 4). In addition to the standard negative controls such as bovine serum albumin (BSA), printing buffer, and blank spots, we have selected an HIV protein (p24 subunit). HIV-p24 antigen is expressed and purified from E. coli in a similar manner to our custom proteins. Since our patients tested negative for HIV, the fluorescent reading obtained for HIV-p24 antigen is considered background nonspecific reactivity to E. coli proteins and HIV-p24 measurements are subtracted as background. In short, the methodology consists of blocking the slide with 1% BSA followed by addition of a plasma sample. After application of the secondary antibody, the slides are washed, dried, and scanned. We have observed agreement between duplicate spots within a slide (R2 = 0.96). In our studies, we identify allo-ab targets 1 year after transplantation. Thus, when a male patient undergoes hematopoietic cell transplantation (HCT) with a female donor, the male pretransplant plasma fails to detect H-Y antigens, but 1 year following HCT, 50% F→M HCT patients develop allo-Ab against at least one of the H-Y antigens (23, 24). Thus, the pretransplant plasma serves as its own internal control and one can monitor patients longitudinally for the development of an immune response against each of the H-Y antigens.

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

a

Anti-human IgG detection without patient sera

b

Anti-V5 detection alone

c

89

Anti-human IgG detection with patient sera

IgG1 IgG2 IgG3

Optimize printing and scanner detection using three different concentrations of IgG

Normalize recombinant antigen printing using V5 tag on proteins

Normalize inter-slide variation by normalizing IgG detection and then quantify specific IgG response to antigens

Fig. 4. Schematic representation of H-Y Microarray Detection. A schematic representation of an H-Y slide is shown in the figure. The H-Y slide in (a) is probed with anti-human IgG and thus only the three different concentrations of IgG are identified. This detection helps to visualize and optimize printing across various subarrays. (b) Is an H-Y slide probed with anti-V5 for antigen detection to normalize for antigens in a batch inter-slide and intra-slide (across subarrays). (c) Shows that when an H-Y slide is probed with patient plasma, all the IgG spots and a few H-Y antigens are identified as hits. Using the fluorescence intensity units of IgG one can normalize for processing and scanning the arrays in one batch.

Demonstrating this concept schematically, Fig. 4 shows three slides printed with different (H-Y) antigens and three different concentrations of IgG printed in two subarrays. Panel A shows recognition of only IgG spots when the slide is probed with antihuman IgG Ab. These spots are preferentially printed at the beginning of each subarray to visualize correct printing orientation and optimize printing. Panel B shows a slide probed with only anti-V5 Ab that recognizes all the H-Y antigens tagged with V5 epitope tag, but not the IgG spots. This data helps normalize antigen printing across batches and subarrays. Thus, when patient sera/plasma is applied, all IgG spots are recognized in addition to patient-specific H-Y antigens (Panel C). Figure 5 represents H-Y arrays that were printed and probed with anti-human IgG (Panel A) and anti-V5 Ab (Panel B). As mentioned, only IgG spots or all H-Y antigens are recognized. When a male patient plasma/sera with a female donor HCT is applied on the slide, specifically this patient recognizes DDX3Y, ZFY, and UTY (Panel C),

90

Wadia, Sahaf, and Miklos

Fig. 5. Representative H-Y Microarray Detection. (a) Shows a slide probed with anti-human IgG where three different concentrations of IgG (IgG1 < IgG2 < IgG3) in each subarray are identified. (b) Shows an H-Y slide probed with anti-V5 identifying all the H-Y antigens tagged with V5, and as expected, no immunoglobulins are detected. For our custom H-Y arrays, we also print infectious agent antigens such as VZV, Pneumococcus, and EBV and negative controls such as printing buffer and BSA, and also have some empty spots. Since these infectious antigens are not tagged with V5 (except for EBV ), these antigens and negative controls will not be recognized by the anti-V5 antibody. (c) Shows detection of all three IgG concentrations in each subarray along with DDX3Y, UTY, and ZFY as targets on the H-Y array when probed with sera from a male patient collected 1 year after undergoing HCT from a female donor. Detection of these same proteins was absent pretransplant (d).

whereas detection of these same proteins was absent pretransplant (Panel D). In addition to the H-Y antigens, viral antigens are also spotted on the slide and EBV is recognized by antibodies in the pretransplant and posttransplant plasma irrespective of the donor (Panel C and D).

2. Materials 2.1. Preparation of Sera/Plasma

1. Venous blood collection tubes (BD Vacutainer®) or (BD Vacutainer serum tubes). 2. Cryovials (Fisher Scientific). 3. Centrifuge (Beckman Coulter Inc, Allegra 6KR centrifuge).

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

91

4. 1× PBST pH 7.4 (1× PBS [Gibco, Invitrogen] + 0.1% Tween 20 [American Bioanalytical]). 5. 1 M EDTA (VWR International). 2.2. Recombinant Antigens for Printing

1. Anti-human IgG Alexa647 conjugate (Molecular probes, Invitrogen). 2. Purified Human IgG (Sigma). 3. H-Y proteins ORFs cloned into E. coli.

2.3. Bacterial Cell Culture and Purification of Recombinant Protein Expression

1. 2XYT broth (EMD Chemicals, Inc). 2. Ampicillin (Amp; BD Diagnostics). 3. IPTG (EMD Chemicals, Inc). 4. B-PER Bacterial Protein Extraction Reagent (Pierce Chemicals). 5. Imidazole (Sigma; 10 mM Imidazole). 6. Glycerol (EMD Chemicals). 7. Monobasic sodium phosphate (MP Biomedicals, LLC). 8. Dibasic sodium phosphate (Sigma). 9. Urea (Sigma). 10. NuPAGE® Sample reducing agent (10×; Invitrogen). 11. NuPAGE® MOPS SDS running buffer (20×; Invitrogen). 12. NuPAGE® LDS 4× LDS sample buffer (Invitrogen). 13. NuPAGE® Novex 4–12% Bis-Tris gel 1.0 mm, 12 well (Invitrogen). 14. Anti-V5-HRP antibody (Invitrogen). 15. TBS buffer (Invitrogen). 16. 1.5 ml Eppendorf tubes (Eppendorf). 17. Eppendorf 5415D centrifuge (Eppendorf). 18. BD Falcon™ disposable centrifuge tubes, polypropylene, conical bottom (BD Biosciences). 19. mm glass beads (BioSpec Products, Inc). 20. Milk powder. 21. RNAse A (Qiagen). 22. Beckman centrifuge (Beckman Rotor No. SW28). 23. Sorvall RC-5 centrifuge (Thermo Scientific). 24. Dri-Block® heaters (Techne, Ltd). 25. VWR® Vortex Mixer (VWR). 26. DU® 800 UV/Vis spectrophotometer (Beckman Coulter). 27. Sonicator (Fischer Scientific, Model 100). 28. Dounce Homogenizer.

92

Wadia, Sahaf, and Miklos

29. FPLC chromatographic system (GE Healthcare, Model AKTA). 30. Denatured Nickel affinity chromatography Lysis Buffer (500 ml): Make stock 10× Buffer B (500 ml): 5 M NaCl (146.1 g), 200 mM dibasic Na phosphate (26.81 g). Bring volume to 500 ml (may need heat). Add 6 M Guanidine HCl (285 g) to 50 ml of 10× Buffer B, pH 7.8. Bring volume to 500 ml (heat on stir plate to get guanidine into solution, adjust pH with 10 N NaOH). 31. Wash Buffer pH = 6.0 (500 ml): Make stock 10× Buffer A (500 ml): 5 M NaCl (146.1 g), 200 mM monobasic Na phosphate (13.7 g). Bring volume to 500 ml (may need heat). Make stock 10× Buffer B (500 ml): 5 M NaCl (146.1 g), 200 mM dibasic Na phosphate (26.81 g). Bring volume to 500 ml (may need heat). Add 36.9 ml of 10× Buffer A and 13.1 ml 10× Buffer B with 100 ml of 20% Glycerol and adjust the pH to 6.0. 32. 6 M Urea Buffer (500 ml): Make stock 10× Buffer A (500 ml): 5 M NaCl (146.1 g), 200 mM monobasic Na phosphate (13.7 g). Bring volume to 500 ml (may need heat). Make stock 10× Buffer B (500 ml): 5 M NaCl (146.1 g), 200 mM dibasic Na phosphate (26.81 g). Bring volume to 500 ml (may need heat). Add 36.9 ml of 10× Buffer A and 13.1 ml 10× Buffer B; to the mix, add 100 ml of 20% Glycerol and 180 g urea (6 M final concentration). Adjust the pH to 6.0. 33. Imidazole Elution Buffer (500 ml): Make stock 10× Buffer A (500 ml): 5 M NaCl (146.1 g), 200 mM monobasic Na phosphate (13.7 g). Bring volume to 500 ml (may need heat). Add 50 ml of 10× Buffer A to 100 ml of 20% Glycerol and Imidazole (34.04 g) (final concentration 1 M). Adjust the pH to 6.0. 34. Native Nickel affinity chromatography Lysis Buffer (500 ml): Make a stock of 10× Buffer A (500 ml): 5 M NaCl (146.1 g), 200 mM monobasic Na phosphate (13.7 g). Bring volume to 500 ml (may need heat). Make 10× Buffer B (500 ml): 5 M NaCl (146.1 g), 200 mM dibasic Na phosphate (26.81 g). Bring volume to 500 ml (may need heat). Add 50 mM Imidazole (1.7 g) to12 ml of 10× Buffer A and 38 ml of 10× Buffer B. Bring volumes to 500 ml and confirm pH 7.8.

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

93

35. Wash Buffer pH = 6.0 (500 ml): Make stock 10× Buffer A (500 ml): 5 M NaCl (146.1 g), 200 mM monobasic Na phosphate (13.7 g). Bring volume to 500 ml (may need heat). Make stock 10× Buffer B (500 ml): 5 M NaCl (146.1 g), 200 mM dibasic Na phosphate (26.81 g). Bring volume to 500 ml (may need heat). Add 36.9 ml of 10× Buffer A and 13.1 ml 10× Buffer B and adjust the pH to 6.0. 36. Imidazole Elution Buffer (500 ml): Make stock 10× Buffer A (500 ml): 5 M NaCl (146.1 g), 200 mM monobasic Na phosphate (13.7 g). Bring volume to 500 ml (may need heat). Add 50 ml of 10× Buffer A to Imidazole – 34.04 g (final concentration 1 M). Adjust pH to 6.0. 2.4. Printing the Purified Recombinant Proteins to Obtain Protein Microarrays

1. Slides: Precoated glass slides with nitrocellulose FAST Slide-1 Pad (Whatman, Inc.). 2. Proteins to be printed (1 mg/ml). 3. 384-well amplification plates (Nunc). 4. 1× PBS (Gibco, Invitrogen). 5. Contact protein printer (Bio-Rad, Model. ChipWriter Pro). 6. Stealth Micro spotting prints (Telechem International). 7. Stealth Microarray printhead for 32 pins (Telechem International).

2.5. Probing and Developing H-Y Recombinant Protein Microarrays

1. Four-well trays (quadriPERM four-chamber culture dish) (Greiner ISC Express). 2. 10× PBS (Gibco, Invitrogen). 3. Blocking Buffer: 3% Bovine Serum Albumin (BSA; Sigma), 1× PBS, 0.1% Tween 20. 4. Wash buffer: 1× PBS, 0.1% Tween 20. This buffer is made fresh. 5. Anti-human IgG Alexa647 conjugate (Molecular probes, Invitrogen) (diluted 1:1,000 in 1× PBST). 6. Anti-V5 FITC (Invitrogen) (diluted 1:100 in 1× PBST). 7. Eppendorf centrifuge (Fisher Scientific). 8. Lab rotator (Lab-Line Instruments).

2.6. D ata Analysis

1. GenePix 4000B microarray scanner (Molecular Devices Corporation). 2. GenePix Pro software (Molecular Devices Corporation).

94

Wadia, Sahaf, and Miklos

3. Methods 3.1. Preparation of Plasma Samples

1. Once blood from either patients/control subjects is obtained using the sera collection tube or plasma collection tube, the tube is inverted gently 5 times. 2. Centrifuge the tubes for 10 min at 1,430 × g at 4°C. 3. Transfer all sera/plasma obtained in 1 ml aliquots and add 0.1 mM EDTA (final concentration) in each aliquot to avoid fibrin formation. 4. Store aliquots at −80°C until further use. 5. When ready to use sera/plasma, thaw the samples on ice and vortex the samples for 5 s. 6. Centrifuge samples at 16,100 × g for 10 min at 4°C and use sera/plasma from the top of the tube. 7. For protein microarrays, dilute the plasma samples (1:150) by taking 5 ml from the top of the centrifuged plasma tube (avoid dipping tips in the centrifuged plasma tubes below half the volume) and add this to 750 ml of 1× PBST.

3.2. Recombinant Antigen Design

1. 11 H-Y antigens were selected for our H-Y recombinant protein array, namely, EIF1AY, RPS4Y, ZFY, DDX3Y, and UTY, along with their X-homologues, EIF1AX, RPS4X, ZFX, DDX3X, UTX, respectively, and HIV-p24. 2. Full-length cDNA for each gene was reversed transcribed from male peripheral blood mononuclear cells and polymerase chain reaction (PCR)-amplified with primers derived from GenBank sequences (23). 3. Each gene was Topo-cloned (Invitrogen, Carlsbad, CA) and expressed with C-terminal V5 epitope tag and 6 histidine residues in E. coli (pET-Dest42) and female-derived 293 cell line (pcDNA-Dest40). 4. Protein HIV-p24 is encoded by the second open reading frame (ORF) proteolytically processed from the Gag-pol polypeptide in vivo during HIV infection. The HIV-p24 was expressed in E. coli (pET-Dest42) and was purified in similar fashion. Since all patients/donors were previously screened for antibody to HIV-p24 and were known to be negative, reactivity with recombinant HIV-p24 was used as a negative control on our protein microarray and was subtracted from each patient’s H-Y protein measurement after probing the array with patient/donor sera. 5. These proteins are engineered such that they are tagged with 6-Histidine tag (6xHis). After cell lysis, the proteins are purified by fast protein liquid chromatography (FPLC) using a

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

95

nickel agarose column. In the FPLC column, the purification steps typically are binding, washing, and elution. The nickel agarose column only binds proteins tagged with 6xHis tag. These bound proteins are washed and eluted using imidazole. Imidazole being structurally similar to histidine competes with the 6xHis-tagged proteins for available nickel sites. By stepwise increasing the concentrations of imidazole, the proteins get displaced from the column, thus giving a pure recombinant protein solution. 6. H-Y proteins are also C-terminally tagged with a V5 epitope tag. The V5 epitope tag is a short series of amino acids (GKPIPNPLLGLDST) that is not usually cross-reactive with mammalian sera. The tag facilitates detection of proteins in cell lysates or to detect eluted protein purity after FPLC for recombinant proteins using western blots. 3.3. Bacterial Cell Culture and Purification of Recombinant Protein Expression

Day 1: 1. Streak an LB-Ampicillin (LB-Amp) plate with cells containing the vector for the corresponding H-Y protein of interest. 2. Incubate the plates for 18 h or overnight at 37°C for colony formation. Day 2: 1. Select five distinct colonies from LB-Amp plates. 2. Inoculate each colony in 2 ml cultures of 2XYT liquid broth containing 50 ml/ml Ampicillin (2XYT + Amp; see Note 10). 3. Incubate inoculated cultures for 18 h or overnight at 37°C with constant agitation (rotating shaker). Day 3: Perform a miniinduction protocol to choose the clone expressing the maximum protein from the five inoculated culture tubes: 1. Inoculate 100 ml of the incubated culture into 1 ml of fresh media (2XYT + Amp). Preserve the rest of the inoculated cultures from Day 2 at 4°C. 2. Incubate the cultures at 37°C for 1 h. 3. Induce expression of protein in cultures for 1–2 h using 1 mM IPTG (10 ml from a 100 mM IPTG stock in 1 ml media). 4. Harvest cells by pelleting in a 1.5-ml Eppendorf at 16,100 × g for 4 min. Discard the supernatant. 5. Prepare samples for an SDS-PAGE. {Add water (65 ml/ sample), 4× LDS buffer (25 ml/sample), and 10× reducing agent (10 ml/sample), along with ten 1.0 mm glass beads to lyse the cells}.

96

Wadia, Sahaf, and Miklos

6. Vortex continuously at high speed for 1 min, then place samples in a heat block at 80°C for 10 min, and again vortex for another 1 min. 7. Centrifuge tubes for 10 s at 3,060 × g and add 15 ml of prepared samples in each lane of a 4–12% Bis-Tris Gel (see Note 11). 8. Proceed with SDS-PAGE (4–12% Bis-Tris Gel in MOPS running buffer). 9. Western Blot the proteins to determine the most robust protein expressor clone. 10. Detect protein using Anti-V5 HRP antibody (1:5,000 in 5% milk in 1× TBS) and visualizing the blots with ECL. 11. The clone with the maximum level of protein expression is used for 4 L scale-up preparation on day 4. 12. Add 200 ml of the selected high protein expression clone to each four 100 ml cultures (2XYT + Amp). 13. Incubate for 18 h or overnight at 37°C. Day 4: A. Proteins obtained under Denatured Condition Protocol 1. Add each of the saturated 100 ml cultures into 1 L of fresh media (2XYT + Amp). 2. Cells are incubated at 37°C and optical density (OD) is checked after 1 h periodically until OD reading of 0.7–0.8 is achieved (see Note 12) 3. Induce each flask with a final concentration of IPTG as 1 mM. 4. Grow cultures for the next 2 h. 5. Centrifuge cultures at 2,050 × g for 20 min at 4°C in 225 ml conical BD Falcon centrifuge tubes. 6. After obtaining cell pellets, discard the supernatants and continue working on ice to prevent protein degradation (see Note 13). 7. Wash pellets with increasing concentrations of imidazole in 80 ml lysis buffer (BPER) and 20 ml of DNAse 8. From the 80 ml resuspended pellet, add 40 ml into a homogenizer. 9. Dounce (15 strokes with a tight pestle) and then split the 40 ml into two centrifuge tubes with 20 ml each. Repeat with the other remaining 40 ml of resuspended pellets. 10. Centrifuge at 31,780 × g for 15 min in Sorvall centrifuge. 11. Discard the supernatant and keep inclusion body pellets (protein is in pellets).

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

97

12. Resuspend in 80 ml 50/50 BPER and water with same concentration of imidazole as in step 7, and homogenize as above. 13. Centrifuge at 16,100 × g for 15 min. 14. Discard the supernatant and keep inclusion body pellets. 15. Resuspend in 40 ml 1:10 BPER:water with same concentration of imidazole as in step 7, homogenize, and centrifuge again (two tubes with 20 ml each). 16. Resuspend and combine pellets for final spin in 15 ml Denatured Nickel affinity chromatography 6 M Guanine HCl lysis buffer at 37°C to improve solubilization. 17. Dounce homogenize (15 strokes) the pellets and transfer the homogenized mixture to ultracentrifuge tubes. 18. Centrifuge tubes in high speed centrifuge, SW28 swinging bucket rotor, at 141,000 × g for 30 min. 19. Collect supernatant to run on an FPLC column. Day 4: B. Proteins obtained under Native Condition Protocol 1. Follow steps 1–6 as described in section Day 4: A: Proteins obtained under Denatured Condition Protocol. 2. Suspend all pellets in a total of 30 ml of native lysis buffer. 3. Add lysozyme to a final concentration of 1 mg/ml and incubate on ice for 30 min. 4. Sonicate on ice. 5. Add RNAse A (1:1,000 dilution) and DNAse I (1:5,000 dilution) and incubate on ice for 10–15 min. 6. Centrifuge in Sorvall centrifuge at 100,000 × g for 5 min. 7. Using the supernatant, run it through a high speed centrifuge, SW28 swinging bucket rotor at 141,000 × g for 30 min. 8. Collect supernatant to run on an FPLC column and extraction is performed using a native nickel extraction protocol. Day 5: 1. HIV-p24, EIF1AY, and EIF1AX proteins follow native nickel affinity chromatography protocol and all other H-Y proteins require denatured nickel affinity chromatography processing when synthesized in E. coli. All proteins were solubilized in their respective buffers and subsequently purified by nickel affinity chromatography. 2. Apply supernatants to a 10-ml Nickel affinity column that has been equilibrated in 6 M Guanidine lysis buffer or native nickel affinity chromatography lysis buffer based on the protein extraction protocol.

98

Wadia, Sahaf, and Miklos

3. For proteins using the denatured protocol, the lysis buffer is then exchanged with 6 M Urea solution over an eight-column volume gradient. For proteins using the native affinity nickel chromatography protocol, the 6 M Urea step is absent. (Instead of imidazole, a reduction in pH can also be used as an alternative to elute proteins). 4. The 6 M urea solution (denatured protocol) or the native lysis buffer (native affinity method) is washed with a wash buffer (pH = 6.0). 5. The proteins are eluted using an Imidazole elution buffer (containing 500 mM sodium chloride, 20 mM sodium phosphate, pH 6.0, and 20% glycerol) with increasing step gradients of Imidazole. 6. We use FPLC automation to run this protocol overnight and collect 48 fractions of eluted proteins of various concentrations. 7. The use of 6xHis tags and nickel affinity columns allows onestep protein purification, under either native or denaturing conditions, from dilute solutions and crude lysates, and also does not depend on the 3-D structure of the protein or the 6xHis tag. Day 6: 1. The fractions collected are tested using SDS-PAGE followed by Western blotting and proteins are detected using Anti-V5 HRP antibody. 2. The highest concentration fractions are reserved for microarray printing at 1 mg/ml. 3.4. Printing the Purified Recombinant Proteins to Obtain Protein Microarrays

1. A 384-well plate with purified proteins is prepared where 10 ml of each protein (1 mg/ml) is placed in wells separated by a single column and two rows. The moisture control in the printing chamber must be optimized to minimize solvent evaporation preventing antigen concentration flux. 2. 100 slides are placed in the correct orientation on the protein printer platform and immobilized using vacuum. 3. The printer is programmed to load 0.2 ml in the solid pin attached to the printer from the 384-well plate to spot proteins. 10 ml protein/well in the 384-well plate prints 100 slides. 4. The first two, middle two, and the last two slides after the print run are tested and processed using anti-V5 FITC to ensure all the proteins are printed and in the correct concentrations.

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

3.5. Probing and Developing H-Y Recombinant Protein Microarrays 3.5.1. Probing with Anti-V5 to Detect Printed Protein Spots

99

1. All steps are carried out at room temperature. 2. Add 5 ml of wash buffer in one chamber of the four-well tray. 3. Prewet the H-Y protein microarray for 5 min at room temperature with gentle agitation. 4. Remove wash buffer by aspiration either using vacuum suction or with the aid of a pipette. 5. Add 5 ml of blocking buffer and incubate for 2 h at room temperature with gentle agitation. 6. Remove the block buffer by aspiration. 7. Dilute FITC-labeled anti-V5 antibody 1:1000 (v/v) in 1× PBST buffer. 8. Add 5 ml of the diluted antibody on the array without touching the array. 9. Incubate at room temperature for 2 h without any agitation. 10. Place the slide in sterile 50 ml conical tube containing 25 ml wash buffer. 11. Wash the slide for 10 min at room temperature with gentle agitation with wash buffer. 12. Discard the wash buffer and repeat two more times. 13. Place the protein microarray in a sterile, dry, 50 ml conical tube and centrifuge the tube at 228 × g for 10 min at room temperature.

3.5.2. Probing with Human Serum/Plasma

1. All steps are carried out at room temperature. 2. Add 5 ml of wash buffer in one chamber of the four-well tray. 3. Prewet the H-Y protein microarray for 5 min at room temperature with gentle agitation (see Note 14). 4. Remove wash buffer by aspiration either using vacuum suction or with the aid of a pipette. 5. Add 5 ml of blocking buffer and incubate for 2 h at room temperature with gentle agitation. 6. Remove the block buffer by aspiration. 7. Dilute plasma 1:150 (v/v) in 1× PBST buffer as described in Subheading 3.1. 8. Add 150 ml of the diluted plasma on the array without touching the array. 9. Using forceps, place a Lifter Slip on the array to cover the membrane area, ensuring that there are no air bubbles trapped between the membrane and lifter slip. 10. Incubate at room temperature for 2 h without any agitation.

100

Wadia, Sahaf, and Miklos

11. Place the slide in sterile 50 ml conical tube containing 25 ml wash buffer. 12. Carefully remove the lifter slip after it detaches from the slide in the conical tube, making sure not to touch the surface of the protein microarray, and discard the lifter slip. 13. Wash the slide for 10 min at room temperature with gentle agitation. 14. Discard the wash buffer and repeat two more times. 15. Remove the H-Y protein microarray and place it in a new four-well tray chamber. 16. Add 5 ml of the secondary detection antibody diluted in wash buffer in the chamber without touching the surface of the slide. 17. Incubate for 1 h at room temperature without any agitation. 18. Place the slide in a sterile 50 ml conical tube containing 25 ml wash buffer. 19. Wash the slide for 10 min at room temperature with gentle agitation. 20. Discard the wash buffer and repeat two more times. 21. Place the protein microarray in a sterile, dry, 50 ml conical tube and centrifuge the tubes at 228 × g for 10 min at room temperature. 22. Ensure that the slides are completely dry (there should be no translucent areas) and place the slides in a slide container box in the dark at room temperature until they are read. Ideally, the slides should be read within 24 h of probing, processing, and drying the slide. 3.6. D ata Analysis

1. Place the slides in the microarray slide holder of the scanner with the side containing the protein printed spots facing the laser source. 2. Scan the slides using wavelength 488 for slides probed using anti-V5 FITC or wavelength 647 for slides probed with patient sera/donors. 3. Photomultiplier tube (PMT) gain should be 600 U, laser power should be 100%, the pixel size and focus position should be 10 and 0 mm, respectively. 4. In order to insure maximum and reproducible dynamic range and since the maximum reading value for the reader is 60,000 fluorescent units, the highest IgG concentration is set at 55,000 signal by adjusting the PMT voltage. In this way, one can avoid being off-scale for readings and should calibrate the reader on daily basis, ensuring the reproducibility of the assay between readings.

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

101

5. Using GenePix Pro software (Molecular Devices Corporation) package, a gal file is created indicating the position of each protein spotted on the array (see Note 15). 6. The protein microarray is scanned at the selected wavelength, and after gridding the obtained tiff image file with the gal file, the results for MFI readings for each spot are obtained and results of patients and controls can be compared. For each spot, a local background around the diameter of the spot is calculated and may be subtracted from the MFI of the spot accounting for the background. This background will reflect and account for the unwanted background shifts that may occur in different location on the nitrocellulose slide. 7. The Genepix reading of the mean fluorescent intensity (MFI) value for each spot is created in the GenePix Result file or GPR file. In this file the fluorescence intensity of each spot is presented and annotated in table format. Median, mean standard error, local background for each spot is also tabulated and could be used for control and calculation purposes.

4. Notes 1. Protein purity is essential to avoid false positive results. 2. Most genes are polymorphic encoding single nucleotide polymorphisms and splice variations resulting in multiple isoforms. For custom arrays, using other techniques, one can determine the isoform of the antigen to be printed on the array. Alternatively, all isoforms obtained can be printed on the array. However, one may not necessarily know the isoform printed on a commercial array. 3. Care needs to be taken to remove the native stop codon from the ORF in order for the plasmid-encoded C-terminus tag to be incorporated in the protein. 4. The antigens to be printed should be in the right buffers with optimal pH and the buffers containing preservative agents like azide or stabilizing agents like glycerol should be avoided. Maintenance of native protein conformation requires nondenaturing isolation. 5. In our experience, a quill pin is preferred because while printing, solid pins can damage the coating surface of the slide, or if inadequate printing material is present in the pin, the spots will be doughnut-shaped when the slide is processed. 6. Ideally, protein concentration should be uniform. Protein stability should be assessed to determine shelf storage conditions. Shelf life of antigens printed (stability) can be monitored

102

Wadia, Sahaf, and Miklos

using epitope detection, e.g., using anti-V5 antibody against the V5 tag bound to each antigen. 7. If irregular spots are obtained, each operator may grid the spot differently leading to various and nonreproducible results. For example, if the spots are “comet” shaped, one can grid the whole spot or part of the spot and leave the tail of the comet out and these could lead to significantly different results for the same printed antigen spot. 8. In order to avoid concentration through evaporation of antigens from the pins while printing, ensure the printing takes place in a defined humidity environment. 9. If the intensity of the processed antigen spot is unevenly distributed within the circle, the MFI of the spot will be low as compared to the MFI when analysis is done using software that fits the contour of the processed spot, and hence, we prefer using software which takes the shape of the processed spot. 10. Do not add Ampicillin until after autoclaving. Add Ampicillin when the temperature of the media has reached room temperature. 11. Take less viscous surface of supernatant and add about 15 ml of sample in a gel lane. If the sample is too viscous (won’t sit at the bottom of the well), vortex and centrifuge again. 12. Absorbance is measured at 595 nm with the blank being fresh 2XYT + Amp media. 13. Keep pellets on ice and/or at 4°C at all times until the lysis buffer step. 14. Make sure that there are no white patches of nitrocellulose surface seen and the slide is completely wet, else it will cause white patches of uneven staining for further steps. 15. A file linking the spot position with its identity is created and usually is called a gal file (GenePix Array List). The software to create this file is usually provided with the GenePix scanner, but alternatively one can create a similar file as a tab delimited file in Microsoft Excel.

Acknowledgments This work was supported by NIH R21 HL084318-01A1 and P01 CA049605. We would like to thank Mrs. Fang Wu for her help in processing the H-Y slides. We would also like to thank Dr. John Coller, Director of the Stanford Protein Array Core Facility, for his advice and printing of the H-Y recombinant antigen arrays.

Recombinant Antigen Microarrays for Serum/Plasma Antibody Detection

103

References 1. Spence RP, Wright V, Ala-Aldeen DA, Turner DP, Wooldridge KG, James R (2008) Validation of virulence and epidemiology DNA microarray for identification and characterization of Staphylococcus aureus isolates. J Clin Microbiol 46:1620–1627 2. Jin D, Qi H, Chen S, Zeng T, Liu Q, Wang S (2008) Simultaneous detection of six human diarrheal pathogens by using DNA microarray combined with tyramide signal amplification. J Microbiol Methods 75:365–368 3. Jandu N, Ho NK, Donato KA, Karmali MA, Mascarenhas M, Duffy SP, Tailor C, Sherman PM (2009) Enterohemorrhagic Escherichia coli O157:H7 gene expression profiling in response to growth in the presence of host epithelia. PLoS One 4:e4889 4. Mengual L, Burset M, Ars E, Lozano JJ, Villavicencio H, Ribal MJ, Alcaraz A (2009) DNA microarray expression profiling of bladder cancer allows identification of noninvasive diagnostic markers. J Urol 182:741–748 5. Li L, Wadia P, Chen R, Kambham N, Naesens M, Sigdel TK, Miklos DB, Sarwal MM, Butte AJ (2009) Identifying compartment-specific non-HLA targets after renal transplantation by integrating transcriptome and “antibodyome” measures. Proc Natl Acad Sci U S A 106: 4148–4153 6. Sakuishi K, Oki S, Araki M, Porcelli SA, Miyake S, Yamamura T (2007) Invariant NKT cells biased for IL-5 production act as crucial regulators of inflammation. J Immunol 179:3452–3462 7. Imamichi T, Yang J, Huang DW, Brann TW, Fullmer BA, Adelsberger JW, Lempicki RA, Baseler MW, Lane HC (2008) IL-27, a novel anti-HIV cytokine, activates multiple interferon-inducible genes in macrophages. Aids 22:39–45 8. Hwang SS, Kim YU, Lee W, Lee GR (2009) Differential expression of nuclear receptors in T helper cells. J Microbiol Biotechnol 19:208–214 9. De Vos J, Hose D, Reme T, Tarte K, Moreaux J, Mahtouk K, Jourdan M, Goldschmidt H, Rossi JF, Cremer FW, Klein B (2006) Microarray-based understanding of normal and malignant plasma cells. Immunol Rev 210:86–104 10. Yamamoto R, Sakamoto T, Nishi S, Sakai M, Morinaga T, Tamaoki T (1990) Expression of human alpha-fetoprotein in yeast. Life Sci 46:1679–1686 11. Carruthers AM, Warner AJ, Michel AD, Feniuk W, Humphrey PP (1999) Activation

of adenylate cyclase by human recombinant sst5 receptors expressed in CHO-K1 cells and involvement of Galphas proteins. Br J Pharmacol 126:1221–1229 12. Predki PF, Mattoon D, Bangham R, Schweitzer B, Michaud G (2005) Protein microarrays: a new tool for profiling antibody cross-reactivity. Hum Antibodies 14:7–15 13. Mattoon D, Michaud G, Merkel J, Schweitzer B (2005) Biomarker discovery using protein microarray technology platforms: antibodyantigen complex profiling. Expert Rev Proteomics 2:879–889 14. Wadia PP, Coram M, Armstrong RJ, Mindrinos M, Butte AJ, Miklos DB (2010) Antibodies specifically target AML antigen NuSAP1 after allogeneic bone marrow transplantation. Blood 115(10):2077–2087 15. Marina O, Biernacki MA, Brusic V, Wu CJ (2008) A concentration-dependent analysis method for high density protein microarrays. J Proteome Res 7:2059–2068 16. Stillman BA, Tonkinson JL (2000) FAST slides: a novel surface for microarrays. Biotechniques 29:630–635 17. MacBeath G, Schreiber SL (2000) Printing proteins as microarrays for high-throughput function determination. Science 289: 1760–1763 18. Angenendt P, Glokler J, Murphy D, Lehrach H, Cahill DJ (2002) Toward optimized antibody microarrays: a comparison of current microarray support materials. Anal Biochem 309:253–260 19. Kusnezov W, Pulli T, Witt O, Hoheisel JO (2005) Solid Supports for protien microarrays and related devices. Jones and Bartlett, Sudbury, pp 247–283 20. Paborsky LR, Dunn KE, Gibbs CS, Dougherty JP (1996) A nickel chelate microtiter plate assay for six histidine-containing proteins. Anal Biochem 234:60–65 21. Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P, Lan N, Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P, Dean RA, Gerstein M, Snyder M (2001) Global analysis of protein activities using proteome chips. Science 293:2101–2105 22. Goulmy E (1996) Human minor histocompatibility antigens. Curr Opin Immunol 8:75–81 23. Miklos DB, Kim HT, Zorn E, Hochberg EP, Guo L, Mattes-Ritz A, Viatte S, Soiffer RJ, Antin JH, Ritz J (2004) Antibody response to DBY minor histocompatibility antigen is

104

Wadia, Sahaf, and Miklos

induced after allogeneic stem cell transplantation and in healthy female donors. Blood 103: 353–359 24. Miklos DB, Kim HT, Miller KH, Guo L, Zorn E, Lee SJ, Hochberg EP, Wu CJ, Alyea EP, Cutler C, Ho V, Soiffer RJ, Antin JH, Ritz J (2005) Antibody responses to H-Y minor histocompatibility antigens correlate

with chronic graft-versus-host disease and disease remission. Blood 105:2973–2978 25. Zorn E, Miklos DB, Floyd BH, Mattes-Ritz A, Guo L, Soiffer RJ, Antin JH, Ritz J (2004) Minor histocompatibility antigen DBY elicits a coordinated B and T cell response after allogeneic stem cell transplantation. J Exp Med 199:1133–1142

Chapter 8 SPOT Synthesis as a Tool to Study Protein–Protein Interactions Dirk F.H. Winkler, Heiko Andresen, and Kai Hilpert Abstract Peptide arrays are a widely used tool in proteomic research for investigations of drug development and molecular interactions including protein–protein or protein–peptide interactions. Most peptide synthesis techniques are able to simultaneously synthesize only up to a few hundred single peptides. Using the SPOT™ technique, it is possible to synthesize and screen in parallel up to 8,000 peptides or peptide mixtures. In addition, such peptides can be released from the membrane and transferred onto peptide microarrays. Here we present protocols for the peptides synthesis on cellulose including the preparation of different cellulose membranes and easy-to-use detection methods on these peptide macroarrays. In addition, a protocol to produce and screen peptide microarray using the SPOT technology is provided. Key words: Spot synthesis, Peptide array, Screening, Cellulose membranes, Protein–protein interaction, Protein–peptide interaction, Microarray

1. Introduction Last year we are celebrating the 20th anniversary of the first presentation of the SPOT™ method by the team of Ronald Frank in 1990 (1). Since that presentation, the SPOT synthesis has proven to be a powerful tool for the study of protein–protein interactions (2–4). This method is inexpensive, easy to perform, and can be established in virtually any laboratory. SPOT synthesis is a special type of parallel solid-phase peptide synthesis on planar, porous surfaces – most commonly on cellulose membranes. The activated coupling solutions are positionally addressed and delivered in small drops to distinct points on the membrane forming a pattern of small spots. Over several coupling cycles, peptides are built upon these spots (5, 6). Using the automated SPOT synthesis,

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_8, © Springer Science+Business Media, LLC 2011

105

106

Winkler, Andresen, and Hilpert

it is possible to synthesize and screen up to 8,000 peptides, peptide mixtures, or other organic compounds on membranes of the size of a 96-well plate (8 × 12 cm) up to about letter size (19 × 28 cm) (7–10). The screening of peptide libraries synthesized using the SPOT technique is usually carried out while the peptides are bound to the membrane (11–13). The detection of bound proteins is similar to a Western Blot. In order to test the peptides in solution, it is also possible to release the peptides from the membrane. These free peptides can be further used either directly for types of screening (14, 15) or for the preparation of peptide microarrays (16, 17). SPOT synthesis protocols presented in this publication can be carried out manually or using automation. Manual SPOT synthesis is most convenient for rather relatively small numbers of peptides (up to 100) and large pipetting volumes (>0.5 ml). For more complex libraries, it is recommended to perform the synthesis semi- or fully-automated (18). Here we describe the standard procedures for SPOT synthesis of peptides. Protocols for the synthesis of modified peptides, for example, cyclic peptides or such with side-chain modifications, are described elsewhere in the literature (6). The following protocols are applicable not only to the use of natural amino acid building blocks, but can also be adopted for the use of unnatural amino acids and several other organic building blocks, e.g., peptide nucleic acid (PNA) monomers and peptoidic elements (19–22). The SPOT method is not restricted to the production of single membrane-bound peptide macroarrays, but can also provide soluble peptides for the production of multiple identical copies of peptide microarrays. In fact, the peptide amount yielded from a single spot of a peptide macroarray is sufficient to produce several hundred identical peptide spots in microarray formats. This is particularly attractive when many different samples have to be screened against the same set of peptides. For this reason, we also include protocols for the release of the peptides from the cellulose membranes, their attachment to chemically modified glass surfaces as well as for the analysis of protein–protein interactions with the corresponding peptide microarrays.

2. Materials Solvents necessary: 1. N,N¢-Dimethylformamide (DMF; toxic, flammable). 2. Methanol (MeOH; toxic, flammable). 3. Ethanol (EtOH; flammable). 4. N-Methylpyrrolidone (NMP).

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

107

5. Diethylether (DEE; highly flammable). 6. Dichloromethane (methylene chloride, DCM; toxic, cancer hazard). 7. 1,4-Dioxane (flammable, suspected carcinogenic). 8. Dimethylsulphoxide (DMSO). The quality of solvents for the required washing steps should be of at least ACS. Solvents for dissolving reagents for the synthesis must be amine- and water-free. Due to possible decomposition under the influence of light, organic solvents, with the exception of MeOH and EtOH, should be stored in the dark. If not noted different elsewhere, the used water is always distilled/deionized. 2.1. Preparation of Cellulose Membranes and SPOT Synthesis of Macroarrays

1. Membranes are prepared from filter paper Whatman 50 or Whatman 540 (Whatman) (20, 23). Several ready-to-use cellulose membranes are commercially available (e.g., from AIMS Scientific, Intavis, or Sigma-Genosys). 2. Amino-acid ester-linked membranes: diisopropylcarbodiimide (DIPC, DIC; Fluka; very toxic), N-Methylimidazole (NMI; Sigma; flammable, corrosive), and Fmoc-b-alanine (EMD Biosciences) or Fmoc-glycine (GL Biochem) (see Note 1). 3. Amino-alkyl ether-linked membranes: 70% perchloric acid (Alfa Aesar; oxidizing, corrosive), epibromohydrine (Fluka; toxic), 1,3-diaminopropane (Alfa Aesar; toxic, corrosive, flammable), 4,7,10-trioxa-1,13-tridecanediamine (Fluka; corrosive), and sodium methylate (sodium methoxide; Fluka; highly flammable, corrosive). 4. Staining solution: 0.002% bromophenol blue (BPB; Sigma) in MeOH (20 mg in 1 l). 5. Coupling reagents: DIC and N-hydroxybenzotriazole (HOBt; EMD Biosciences; flammable). Coupling reagents are only necessary when no preactivated amino acid derivatives are used (see Note 2). 6. Nonpreactivated amino acids with protection groups according to the Fmoc strategy (24, 25) (EMD Biosciences and GL Biochem); preactivated amino acid derivatives with protection groups according to the Fmoc strategy, e.g., pentafluorophenyl esters (OPfp ester; EMD Biosciences and Bachem) (26) (see Note 2). 7. Piperidine (Sigma; toxic, highly flammable). 8. Capping solution: 2% acetic anhydride (Sigma; flammable, corrosive) in DMF. For second consecutive treatment, 2% ethyl-diisopropylamine (DIPEA, DIEA; Sigma; corrosive) can be added to deprotonate amino groups and buffer the acetic acid generated during reaction.

108

Winkler, Andresen, and Hilpert

9. Deprotection solution A: Trifluoroacetic acid (TFA; VWR; corrosive) (v/v) containing 5% dist. water (v/v), 3% triisopropylsilane or triisobutylsilane (TIPS or TIBS; Fluka; irritant) (v/v), 1% phenol (Sigma; toxic, corrosive) (w/v) (Important! see Note 3). 10. Deprotection solution B: 60% TFA (v/v), 3% TIPS or TIBS (v/v), 2% dist. water (v/v), 1% phenol (w/v), 34% DCM (v/v) (Important! see Note 3). 2.2. Peptide Modifications for the Preparation of Microarrays and Cleavage of Peptides from the Cellulose Membrane

1. N1-(9-Fluorenylmethoxycarbonyl)-1,13-diamino-4,7, 10-trioxatridecane-succinamic acid (Fmoc-TTDS-OH; Iris Biotech).

2.3. Preparation of Microarrays

1. Spotting buffer: 0.1 M sodium acetate buffer pH 5.0 containing 10 mM sodium cyanoborohydride (very toxic!) and 10% (v/v) glycerol (see Note 4).

2. Coupling reagents: DIC, HOBt, and NMI. 3. 4-(Fmoc-hydrazino)-benzoic acid (Fmoc-HBA; Bachem). 4. Ammonia gas (Air Liquide; irritant, corrosive).

2. Microscope glass slides with aldehyde surface coating (Schott Nexterion®) or epoxide coating (Corning Life Sciences) (see Note 5). 3. Blocking Buffer: 0.5% (v/v) Nonidet P40 (Sigma-Aldrich), 5% (w/v) skim milk in 1 M Tris–HCl pH 8.5. The skim milk should be freshly added. Homogenization is facilitated by warming the buffer to 60°C and application of ultrasound. 4. Petri dishes with 145 mm diameter (Greiner Bio-One), parafilm, filter paper. 5. QuadriPERM® 4-compartment cell culture plates (Greiner Bio-One). 6. Nitrogen gas. 2.4. Detection Methods

1. 50 mM Tris-buffered saline (TBS), pH 8.0. 2. 50 mM TBS with 0.2% (v/v) Tween (TBS-T). 3. Blocking buffer: 5% (w/v) Casein or skim milk (Sigma) and 4% (w/v) Sucrose (Sigma) in TBS-T (see Note 6). 4. Probing solution (protein, antibody): Depending on the estimated affinity, 0.1–10 mg/ml protein/antibody in blocking buffer. For a lower estimated affinity, a higher concentration of the protein should be used (see Note 7). For protein mixtures (blood, plasma, cell extracts, etc.), an estimation of the concentration of the target protein is necessary.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions 2.4.1. Detection of Bound HRP-Labeled Protein Using Chemiluminescence

109

1. Horseradish peroxidase (HRP, POD)-labeled or unlabeled antibody against the probed peptides/protein: depending on the estimated affinity 0.1–1.0 mg/ml (in blocking buffer) (see Notes 7 and 8). 2. 10 ml mixture for chemiluminescence detection (see Note 9): 1 ml 1 M Tris–HCl pH 8.5; 22 ml 80 mM p-coumaric acid in DMF; 50 ml 250 mM luminol in DMF; 3 ml 30% H2O2; 9 ml water; Mix the p-coumaric acid and the luminol with the corresponding amount of Tris–HCl. Add to this mixture the corresponding amount of water. Activate this solution immediately before use by mixing with hydrogen peroxide.

2.4.2. Detection of Bound HRP-Labeled Protein Using Staining

1. For the usage of protein solutions see step 1 in Subheading 2.4.1.

2.4.3. Detection of Bound AP-Labeled Protein Using Staining

1. Alkaline phosphatase (AP)-labeled or unlabeled antibody against the probed protein: depending on the estimated affinity 0.1–1.0 mg/ml (in blocking buffer) (see Notes 7 and 8).

2. Mixture for staining (10 ml): 5 mg 4-chloro-1-naphthol dissolved in 1.7 ml methanol, 2.5 ml 200 mM Tris–HCl pH 7.4 (24.2 g/l), 100 mg NaCl, 5.8 ml H2O, 5 ml 30% H2O2. First, dissolve the NaCl in the above amount of water and Tris buffer (pH 7.4), followed by adding the methanolic chloronaphthol solution. Shortly before use, mix this solution with hydrogen peroxide.

2. NBT stock solution: 0.5% (w/v) nitrotetrazolium blue chloride (NBT; Sigma) in 70% aqueous DMF; this solution can be stored cold (<8°C) for at least 1 year. 3. BCIP stock solution: 0.5% (w/v) 5-bromo-4-chloro-3-indolyl phosphate (BCIP; Fluka) in DMF; this solution can be stored cold (<8°C) for at least 1 year. 4. Tris buffer: 100 mM NaCl, 5 mM MgCl2, 100 mM Tris; adjusted with HCl or NaOH to pH 9.5. 5. Staining solution: 660 ml NBT stock solution + 330 ml BCIP stock solution mixed in 10 ml Tris buffer. Always prepare this mixture fresh before use. 2.4.4. Regeneration (Stripping) of Used Cellulose Membrane

1. Regeneration buffer I: 8 M urea and 1% SDS in water.

2.4.5. Fluorescence Detection of Bound Ligands on Peptide Microarrays

1. Incubation buffer: 0.15 M NaCl, 0.5% (v/v) Nonidet P40, and 5% (w/v) skim milk in 12 mM PBS pH 7.4. For proteins with high disposition to unspecific binding, it is recommended to use CrossDown buffer (AppliChem) with addition of 5% (w/v) skim milk as the incubation buffer instead.

2. Regeneration buffer II: 60% TFA, 30% EtOH, and 10% water (Important! see Note 3).

110

Winkler, Andresen, and Hilpert

2. Washing buffer: 0.15 M NaCl, 0.5% (v/v) Nonidet P40 in 12 mM PBS pH 7.4. 3. Secondary antibody: fluorescently labeled polyclonal antibody against the protein of interest, preferably conjugated to Cy5 (Amersham) or Dy647 (Dyomics) (see Note 10). 4. LifterSlips™ (Erie Scientific) and Super Pap Pen (Electron Microscopy Sciences) or FAST Frames (Whatman) (see Note 11).

3. Methods 3.1. Preparation of Cellulose Membranes and SPOT Synthesis of Macroarrays

3.1.1. Preparation of Esterified Membranes

In this chapter we provide protocols for the preparation of c ellulose membranes as well as for the synthesis of peptides on such membranes. Cellulose is a polysaccharide and contains only free hydroxyl groups. However, the hydroxyl groups are not as reactive as amino groups. Therefore, the cellulose usually needs modification to present amino groups on its surface in order to allow easy coupling of amino acids. There are two different types of chemical modifications of the cellulose membranes that are commonly used (27). On membranes of the first type, the amino functionalization is achieved by esterification of the hydroxyl groups of the cellulose with amino acids (5, 28). Due to the relative weak ester bond, this membrane type is particularly useful to yield free peptides. On membranes of the second type, the amino-alkyl ether-linked membranes, the formation of an ether bond between the cellulose and a diaminoalkane leads to the amino functionalization. Here we describe CAPE (= cellulose-amino-hydroxypropyl ether) and TOTD (= trioxa-tridecanediamine) membranes that are suitable for the investigations of peptides bound to the cellulose (29, 30). For the basic cellulose matrix, cut a section of filter paper large enough to accommodate all peptide spots. For a spotting volume of 0.1 ml, the distance between the centers of two spots should be at least about 3 mm and at least about 8 mm for 1 ml spotting volume. If not noted differently elsewhere, all washing, incubation, and reaction steps are performed using a rocking shaker. All washing steps should be performed for at least 30 s each, unless mentioned differently. 1. Amino functionalization by esterification of filter paper with amino acids: In order to modify a cellulose sheet with a size of 8 × 12 cm (size of a 96-well plate), prepare a 10 ml of solution of 625 mg Fmoc-b-alanine or 600 mg Fmoc-glycine in DMF. Add 374 ml DIC and 317 ml NMI. For larger membranes, use more reagent solution accordingly. Mix the solution well.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

111

Transfer the mixture into a chemically resistant box with lid and place the filter paper into the box. Avoid air bubbles under the paper. The surface of the membrane should be slightly covered by the solution. Let the membrane react with the reaction mixture in the closed box for at least 2 h; we recommend an overnight treatment. Shaking is not necessary (see Note 12). After the treatment, wash the membrane 3 times thoroughly with DMF. The membrane can be stored at −80°C for several weeks until needed (see Note 13). 2. Fmoc-deprotection: Treat the membrane twice with 20% piperidine in DMF for at least 5 min each. Wash the membrane at least 3 times with DMF, followed by washing at least twice with MeOH or EtOH. 3. Staining (optional) (31) (see Note 14): Treat the membrane with staining solution for at least 2 min until the filter paper shows a homogeneous blue color. If the staining is insufficient, repeat the treatment with fresh staining solution. After staining, wash the membrane at least twice with MeOH or EtOH, until the washing solution remains colorless. 4. Dry the membrane in the air stream of a fume hood or with a hair dryer without heat (see Note 15). The membrane is now ready for the first coupling. 3.1.2. Preparation of Amino-Alkyl Ether-Linked Membranes

Amino-alkyl ether-modified membranes can be stored for a longer time without significant loss of activity. In order to save material and time, it is recommended to prepare a large membrane (e.g., 18 × 28 cm) and cut it into several smaller pieces later if needed. For the preparation of smaller membranes, use proportionally less reagent solutions. 1. Treat the membrane with a mixture of 50 ml MeOH containing 1 ml 70% (v/v) aqueous perchloric acid for several minutes. Wash once with MeOH and dry the membrane on air. 2. Treat the dry membrane with 50 ml of a mixture of 5 ml epibromohydrine and 500 ml 70% aqueous perchloric acid in 1,4-dioxane in a closed box. After 3 h, wash the membrane once with EtOH or MeOH for about 15 min. 3. (a) In order to prepare a CAPE membrane, let the activated membrane react overnight with about 60 ml of a 50% solution of 1,3-diaminopropane in DMF. (b) If a TOTD membrane needs to be prepared, let the activated membrane react overnight with about 60 ml of a 50% solution of 4,7,10-trioxa-1,13tridecanediamine in DMF. 4. The next morning, wash the membrane 3 times with DMF, twice with water, and then 3 times with MeOH. 5. Treat the membrane with a methanolic suspension of 5 M sodium methylate.

112

Winkler, Andresen, and Hilpert

6. Wash the membrane 3 times with MeOH, 5 times with water, and again 3 times with MeOH. After drying on air, the membrane is ready to use (see Note 15). The membrane can be stored at −20°C for several months until needed. 7. Staining (optional) (31) (see Note 14): Wash the membrane at least 3 times with MeOH or EtOH. Treat the membrane with staining solution for at least 2 min until the filter paper shows a homogeneous blue color. If the staining is not sufficient enough, renew the staining solution. After staining, wash the membrane at least twice with MeOH or EtOH, until the wash solution remains colorless and then air dry the membrane. 3.1.3. Preparation of Coupling Solutions

For the preparation of coupling solutions, we describe here two different methods that were most popularly used. For the first method, preactivated Fmoc-protected amino acid derivatives (e.g., pentafluorophenyl esters) are used (3). This method has the advantage that only one type of reagent is required. That makes the preparation of coupling solutions simple and reduces the likelihood of mistakes during the preparation. A disadvantage of this method lies in the fact that activated esters are only commercially available for standard amino acids. For nonstandard amino acids, the preactivated derivatives would need to be synthesized (26), or the second method may be used instead. The second method requires in situ-activated amino acids (32). Activation of the amino acids is carried out by adding an activator and a coupling reagent to a nonactivated Fmoc-protected amino acid derivative. This method is more time consuming, but can be applied for all building blocks available for use in Fmoc solid-phase peptide synthesis. Method 1: Prepare 0.3 M solutions of the amino acid derivatives (except serine) in NMP. Due to poor solubility, the serine derivative must be dissolved in DMF. Except for the arginine derivative, stock solutions of preactivated amino acids can be stored at or below −20°C for at least 1 week. Before starting the first synthesis cycle at each day, replace the amino acid solutions of the previous day by fresh solutions from the stock (see Note 2). Due to instability, the solution of preactivated arginine derivatives must be prepared fresh every day. Method 2: Prepare a 0.9 M solution of HOBt in NMP. Dissolve the Fmoc-amino acids or protected building blocks in that HOBt solution to a concentration of 0.45 M. Except for the arginine derivatives, these solutions can be stored at −20°C for at least a week. Use new portions of the prepared amino acid/HOBt solutions every day (see Note 2). To these solutions, add a freshly prepared mixture of 20% DIC in NMP at a ratio of 3:1 (see Note 16).

SPOT Synthesis as a Tool to Study Protein–Protein Interactions 3.1.4. SPOT Synthesis of the Peptide Array

113

1. Definition of the synthesis pattern (optional): Prepare a solution of activated Fmoc-b-alanine according to method 1 or 2 described in Subheading 3.1.3. Use DMSO as solvent (not DMF or NMP). Deliver the required volume of that solution to all spot positions (see Note 17). Allow to react for about 20 min. In order to achieve a higher loading, at least one repeat of the delivery is recommended. If the separate definition of the synthesis pattern will not be carried out, start at step 6. 2. Blocking remaining free amino groups (capping): Place the membrane face down in a box filled with an appropriate amount of capping solution. Remove all air bubbles being trapped under the membrane. Do not shake! Remove the liquid after about 5 min and add an appropriate amount of a mixture of capping solution with 2% DIPEA. Place the membrane in this solution and let it react for another about 20 min. 3. Removal of Fmoc-protecting group: Wash with DMF 4 times. Treat twice with 20% piperidine/DMF for 5 min each. Wash again 4 times with DMF and then at least twice with MeOH or EtOH. 4. Staining (optional) (31) (see Note 14): Treat the membrane with staining solution in a box while shaking. If the staining solution changes its color to blue very rapidly, renew the solution. Perform the staining until the spots are dyed sufficiently (Fig. 1) (see Note 18). Wash the membrane at least twice with MeOH or EtOH until the wash solution remains colorless. 5. Dry the membrane in the air stream of a fume hood, or by using a hair dryer without heat (see Note 15). The membrane is now ready for the amino acid coupling cycle. 6. Amino acid coupling: The stepwise build-up of peptides starts from the C-terminal amino acid. Deliver the desired volumes of activated amino acid solutions to the corresponding positions on the membrane. Use at least 20% more amino acid solution volume than used in step 1 of Subheading 3.1.4. After each delivery cycle, allow the reaction to occur for at least 20 min. To achieve a higher coupling yield, it is recommended to repeat the spotting of the activated amino acids at least once. 7. Capping (blocking unreacted free amino groups) (see Note 19): Place the membrane face down in a box filled with an appropriate amount of capping solution for about 5 min. Do not shake! Replace the solution by an appropriate amount of a mixture of capping solution with 2% DIPEA. Place the membrane in this solution for at least 5 min.

114

Winkler, Andresen, and Hilpert

Fig. 1. Images of SPOT membranes on b-Ala-modified Whatman 540 filter paper as examples of peptide macroarrays. Membrane (a) shows 55 manually prepared large peptides. The size of the membrane is about 8 × 11 cm. Each spot has a diameter of about 7 mm. Membrane (b) is a SPOT membrane with 1,120 automatically synthesized peptide spots. Each spot has a diameter of about 2 mm. The size of the array is about 8 × 12 cm. All spots were stained for visualization in accordance to step 4 of Subheading 3.1.4.

8. Stepwise building up the peptide chain: Except for the last coupling cycle, repeat steps 3–7. For the last coupling cycle, carry out only steps 3, 5, and 6! 9. Removal of last Fmoc-protecting group: Wash with DMF 4 times. Treat twice with 20% piperidine/DMF for 5 min each. Wash again 4 times with DMF followed by 3 times washing with DCM.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

115

10. Final side-chain deprotection: Treat the membrane with at least 25 ml of deprotection solution A (see Note 3). For larger membranes, use proportionally more deprotection solutions A and B. The membrane must always be well-covered by the deprotection solutions. Keep the box tightly closed. Do not shake! After 30 min, pour off the solution very carefully and treat the membrane afterwards with at least 25 ml of deprotection solution B for 3 h in the closed box without shaking. Pour off the solution very carefully (see Note 20). Wash the amino-acid ester-modified membranes first for at least 3 times with DCM and twice with methanol. All types of membranes have to be washed at least 5 times with water for at least 1 min each until the pH value of the washing solution is between 5 and 7. Wash the membrane 4 times with MeOH or EtOH for at least 1 min each. Dry the membrane in the air stream of a fume hood or with a hair dryer without heat. 3.2. Peptide Modifications for Preparation of Microarrays and Cleavage of Peptides from the Cellulose Membrane

In order to prepare peptide microarrays on glass slides, it is sually necessary to yield solutions of free peptides (see Note u 21). For that purpose, the peptides have to be cleaved from the cellulose membrane. Moreover, the preparation of peptide microarrays on glass slides from peptides initially synthesized with the SPOT technique requires a further N-terminal functionalization of the peptides. This can usually be carried out using the corresponding building blocks as additional coupling cycles (see steps 3, 5, 6, and 7, Subheading 3.1.4) prior to the final deprotection (see step 10, Subheading 3.1.4).

3.2.1. Introduction of a Spacer

In contrast to the SPOT membranes, glass slides are nonporous, impermeable substrates. Although this feature obviously benefits the achievable resolution of peptide spots in the microarray grid and reduces the required sample volume for the analysis to few microliters, it significantly impairs the binding of the protein analyte in solution to the immobilized peptide ligand due to steric hindrance and sluggish molecular arrangement kinetics. To counteract this attribute, it is imperative to incorporate a spacer in order to spatially distance the peptide from the glass surface which will improve its effective degree of freedom. We found that hydrophilic polyethylene glycol spacers are best suited to achieve this desired effect (33). The corresponding modification of the peptides is achieved by coupling the PEG-building block after the built-up of the peptide sequences. 1. Prepare 0.75 ml solution of 69 mg HOBt and 245 mg of Fmoc-TTDS-OH in NMP. This solution is stable and can be stored at −20°C (see Note 2) for at least a week. 2. Shortly before use, prepare a 20% (v/v) solution of DIC in NMP.

116

Winkler, Andresen, and Hilpert

3. Add 0.25 ml of the DIC/NMP mixture to the prepared 0.75 ml amino acid/HOBt solution (see Note 16). 4. Perform the spotting of the solution at least as a double coupling as described in step 6 in Subheading 3.1.4. 5. Carry out again the Fmoc deprotection according to step 3 in Subheading 3.1.4. 3.2.2. Coupling of a Linker Group

The immobilization of peptides on modified glass surfaces requires the peptides to be equipped with an appropriate linker function (16, 34). This is due to the fact that the mere random immobilization of fully unprotected peptides, i.e., the unspecific surface conjugation via their nucleophilic side-chain amines or thiols, often impairs the molecules’ ability to be later recognized by their binding partners. Hence, we recommend to modify the peptides with a suitable N-terminal linker function that later selectively reacts with a complementary function on the chip surface to avoid the risk of chemical modification of amino acid side chains crucial for a given biological function. The linker function of our choice is 4-hydrazinobenzoic acid (HBA), which is conveniently available as a building block for Fmoc solid-phase peptide synthesis (Fig. 2). The hydrazine moiety will chemoselectively react with a suitable electrophile on the glass surface at acidic pH. Aldehydecoated surfaces have proven to be particularly suitable for this purpose as they are conveniently hydrophilic and have a longer half-live compared to strong electrophiles (see Note 22). However, we have used this linker also to immobilize peptides on epoxide and succinimidyl ester-modified glass surfaces with good results. Alternatively to the covalent immobilization via the HBA linker, peptides can be affinity-immobilized via a biotin linker to a streptavidin surface (35, 36).

Fig. 2. HPLC of the neuropeptide, Mastoparan, with 2× b-Ala spacer and HBA linker synthesized via SPOT synthesis. The left major signal (A) corresponds with the desired peptide, while the right large signal (B) relates to the peptide without the HBA linker.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

117

The attachment of either linker molecule to the N-terminus of the membrane-bound peptides can simply be considered an additional coupling cycle. The protocol for the coupling of the linker group is similar to the coupling of the spacer building block (see Subheading 3.2.1). 1. Prepare 0.75 ml solution of 69 mg HOBt and 169 mg FmocHBA in NMP. This solution is stable and can be stored at −20°C (see Note 2) for at least a week. 2. Shortly before use, prepare a 20% solution of DIC in NMP. 3. Add 0.25 ml of the DIC/NMP mixture to the prepared 0.75 ml linker/HOBt solution (see Note 16). 4. Since the coupling of the Fmoc-HBA is more difficult, perform the spotting of that solution at least as a triple coupling. 5. Carry out the final Fmoc-deprotection according to step 9 in Subheading 3.1.4. 6. Perform the side-chain deprotection according to step 10 in Subheading 3.1.4. 3.2.3. Cleavage of the Peptides from a Membrane as Free Peptide Amides

The method described here involves the exposure of the entire dry membrane or the punched-out spots to ammonia gas. The strongly basic environment breaks the ester bond between the peptides and cellulose by forming a C-terminal amide (see Note 23). The release of peptides with ammonia gas can only be achieved on ester-linked membranes. Commercially available membranes often have other than ester bonds between the functional linker and the cellulose. These membranes as well as homemade amino-alkyl ether-linked membranes require the modification with an additional linker and corresponding cleavage protocols (e.g., thioester (37), HMB linker, (38), Rink linker (39, 40)) (see Note 24). 1. Place the dry membrane or the punched-out spots in a glass desiccator (see Note 25). 2. Set the desiccator under vacuum. 3. Fill the desiccator with ammonia gas either directly from a gas bottle or with ammonia filled in a balloon and sucked into the desiccator by the vacuum. 4. In order to replace most of the air by ammonia, repeat steps 2 and 3 at least twice. 5. Let react overnight. 6. On the following day, open the desiccator under a fume hood (Attention! Ammonia gas is highly corrosive and irritant!). Let the gas dissipate over at least 30 min while in the fume hood.

118

Winkler, Andresen, and Hilpert

7. If not done before, punch out the spots. Transfer the disks into wells of microtiter plates (MTPs) or into vials, in order to dissolve the adsorbed free peptides with the sample or transfer buffer (see Note 26). 3.3. Preparation of Microarrays

1. Design a layout for the peptide microarray. Consider about 10% of all spots for suitably defined specificity controls (e.g., unrelated or randomized peptide sequences) that should be deliberately distributed across the array. 2. Reconstitute the peptide probes in a final concentration of 250 mM in spotting buffer. The required volume of spotting solution depends on the specifications of the used microarray printer (see Note 27). The loading at one spot on a bAla-modified membrane is about 300 nmol. That would be enough for 1.2 ml sample volume. 3. Mark the glass slides with a diamond scriber and place them dust-free into the array printer. Avoid contacting the microarray surface. Set up the arrayer and print the slides. 4. Put pieces of filter paper into petri dishes and moisturize the paper with 150 ml water (see Note 28). Place microarray slides into a dish and seal the lids with parafilm. Incubate the slides overnight at room temperature. In this state, the slides can be stored for several weeks at 4°C. 5. Prior to use, freeze the slides for approximately 1 h at −80°C. Provide a slim 150 ml beaker with blocking buffer. Take each slide separately from the freezer and rapidly dip it into the blocking buffer, moving it up and down for about 1 min. This treatment prevents the commonly observed comet effects of spots as the result of excess peptide binding rapidly to the surface around the spot when washed. Afterwards, place the slides printed side up in separate chambers of a quadriPERM cell culture plate and fill the chambers with 10 ml blocking buffer. Maintain 2 h of blocking at room temperature with shaking. 6. Rinse the slides with running water and dry in a stream of nitrogen.

3.4. Detection Methods

Since the number of probing techniques is too huge to be covered here, we focus on the description of the most common type – the detection of bound proteins such as antibodies which are labeled either with horseradish peroxidase or alkaline phosphatase. This detection can only be performed if the peptides are still attached to a membrane or glass slide. If not noted different elsewhere, all washing steps should be performed for at least 5 min each.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions 3.4.1. Detection of Bound HRP-Labeled Protein Using Chemiluminescence

119

1. If the membrane is dry, wash twice for 10 min each with methanol (or ethanol). Then, wash 3 times with TBS-T. 2. Blocking: In order to suppress unspecific binding, a treatment with blocking buffer is necessary for at least 2 h at room temperature. Treatment overnight at room temperature or over the weekend at 4°C is also possible. 3. Incubation with protein sample: After washing once with TBS-T, incubate the membrane with the protein sample for at least 2 h at an appropriate temperature. That depends on the stability of the protein and at what temperature the biological function occurs. 4. Incubation with a target protein binding protein (e.g., primary antibody): Repeat step 3 using the (HRP-labeled) secondary antibody solution instead of the sample (see Note 8). 5. Detection: Wash at least 3 times with TBS. Prepare the chemiluminescence solution (see Subheading 2). Before treatment with the detection solution, remove excess washing buffer on the membrane by placing the membrane on paper towels (see Note 29). Use forceps for all manipulations of the membrane. Place the membrane on a plastic sheet (PP or PE). Pour the detection buffer over the whole membrane. In order to achieve a homogeneous distribution, shake the membrane gently on the plastic sheet. A reaction time of about 1 min is recommended. The detection of the chemiluminescence can be carried out using X-ray film (in a dark room) or a chemiluminescence imager (see Note 30).

3.4.2. Detection of Bound HRP-Labeled Protein Using Staining

1. For treatment of the membrane with the protein sample and the detection antibody, follow the instructions in steps 1–4 of Subheading 3.4.1. 2. Detection: Wash at least 3 times with TBS-T. Activate the staining solution (see Subheading 2) by adding the hydrogen peroxide shortly before use. Treat the membrane with the activated mixture while shaking. Spots containing HRPlabeled proteins develop a violet or brown color. The staining reaction can be stopped by intensive washing with water.

3.4.3. Detection of Bound AP-Labeled Protein Using Staining

1. For treatment of the membrane with the protein sample and the detection antibody, follow the instructions in steps 1–4 of Subheading 3.4.1 with the exception that the proteins/ antibodies are labeled with AP instead of HRP. 2. Detection: Wash at least 3 times with TBS-T. Prepare the detection solution shortly before use. Treat the membrane with the detection solution until staining of the spots is sufficient enough. Stop the staining reaction by washing thoroughly with water.

120

Winkler, Andresen, and Hilpert

3.4.4. Regeneration (Stripping)

1. The removal of bound proteins is two-step procedure (see Note 31). In the first step, treat the membrane overnight at 37°C with regeneration buffer I. After washing twice with water, treat the membrane for at least 1 h at room temperature with regeneration buffer II (see Note 3). After finishing the second regeneration step, wash the membrane at least 3 times with TBS, followed by washing at least twice with water. The membrane could be stored at −80°C after washing with methanol and drying on air or could be used wet for another probing (see Notes 13 and 32).

3.4.5. Fluorescence Detection on Peptide Microarrays

1. Prepare the incubation buffer by homogenizing 5% (w/v) skim milk (50 mg/ml of buffer) in PBS or CrossDown Buffer. 2. Prepare a solution of the protein of interest in the incubation buffer. The appropriate final concentration of the protein essentially depends on the expected affinity of the protein as well as its predisposition to unspecific binding. A concentration of 1 mg/ml is usually a useful starting condition. To estimate the required analysis volume, consider the injection volume required for the specific type of LifterSlips or FAST Frames and include a surplus of ca. 10–20%. 3. Cover the microarrays with LifterSlips or FAST Frames, taking care not to spoil the printed areas. If using LifterSlips with slides that comprise more than one array, use a Super Pap Pen or an equivalent liquid blocker to separate these areas with hydrophobic boundaries which prevent cross-contamination of different samples. Place the pipette tip directly at one of the open sides of the cover glass so that capillary force starts to absorb fluid from the tip. Then start to inject the sample in a slow, continuous action until the chamber is completely filled. Proceed in a similar way when using FAST Frames. Important: Never drop the sample or other fluids directly from top onto the printed areas! 4. Prepare a moist chamber by wetting the filter paper in a petri dish with approximately 1 ml water and carefully place the slides into the chamber. Incubate for 2 h at room temperature or overnight at +4°C. FAST Frames are directly sealed with the provided chamber cover and can be placed on a rocking shaker or a thermomixer with microplate holder during incubation. 5. Remove the LifterSlips and the bulk of incubation solution by rinsing the slides with running water. Place the slides in separate chambers of a quadriPERM cell culture plate and fill the chambers with 10 ml washing buffer. Wash the slides with good shaking for 5 min. Repeat this washing step two more times using fresh washing buffer after each step. Rinse the slides with water and dry in a stream of nitrogen.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

121

Fig. 3. Gray mode fluorescence image of a peptide microarray after incubation with a 1:50 dilution of patient serum, followed by sandwich-detection of captured serum proteins with a Cy5-labeled polyclonal goat antiserum. Increasing fluorescence intensity is expressed according to the translation bar in the bottom-left corner and is generally significant for higher serum abundance of the target protein. The microarray consists of four identical 15 × 15 feature subarrays. Each spot contains either a specific peptide or a control peptide, and registration spots are present at each corner of a subarray. Peptides were printed at a concentration of 250 mM in 0.1 M sodium acetate buffer (pH 5.0), containing 10 mM sodium cyanoborohydride and 10% (v/v) glycerol on aldehyde-coated glass slides. The inset shows a magnified subarray. It is advisable to create peptide arrays consisting of multiple repeats of identical subarrays to compensate for inhomogeneities, artifacts, or spoiled areas, as exemplified here by the blurred and merged peptide spots.

6. Dilute the secondary antibody to a concentration of 10 mg/ml in incubation buffer. 7. Proceed with the secondary antibody incubation in the same manner as described for the protein sample in steps 3 and 4 using an incubation time of 1 h in the dark. 8. Wash and dry the slides as described in step 5. Prevent lengthy light exposure. 9. Scan the slides with an appropriate microarray scanner in the Cy5-channel according to the manufacturer’s instructions. Extract the fluorescence data with the analysis software of choice (Fig. 3).

4. Notes 1. Due to linear structure of the molecule and the resulting flexibility, b-alanine is commonly used for amine functionalization of the filter paper by esterification. More recently, due to the higher loading on the membrane and lower stability during

122

Winkler, Andresen, and Hilpert

cleavage, glycine is used for the functionalization of membranes prepared for the synthesis of peptides which will be cleaved from the support (28). Other amino acids can be used, but there may be a higher risk of losing functionality due to the lower chemical stability of the ester bond between amino acid and cellulose. 2. Reagents must be protected from moisture. To avoid condensation from air humidity, keep frozen containers closed until they warmed up to room temperature (approx. 30 min before use). 3. Do not fill the components into the TFA! Otherwise, the mixture could heat up to a dangerous level! Mix the additives first and add then the TFA into that mixture. Wear appropriate safety protection. 4. Sodium cyanoborohydride is used for the reductive amination when coupling peptides with hydrazine-linker to aldehyde surfaces. It can be omitted for the coupling of peptides with that linker to epoxide-coated or NHS-coated surfaces. Add the sodium cyanoborohydride freshly to the buffer. Sodium cyanoborohydride is extremely toxic! Use all necessary safety measures. Prepare the buffer in a fume hood! 5. The protocol described here was initially developed for aldehydecoated surfaces, but has been successfully adapted to epoxidecoated surfaces as well. Typically, epoxide surfaces are significantly more hydrophobic than aldehyde surfaces which can be advantageous to achieve a higher resolution, i.e., higher spot density, in the microarray. 6. The use of bovine serum albumin (BSA) instead of casein or skim milk is possible, although it is not recommended due to an often higher background signal. 7. If the affinity is expected to be very low, it is possible to use plain TBS or TBS-T instead of blocking buffer. In that case, due to possible unspecific binding, the background signal might be higher compared to using the blocking buffer. 8. If the labeled antibody/protein against the probed protein is not available, an additional incubation step with a labeled antibody against the first antibody is necessary. The concentration should be between 0.1 and 1.0 mg/ml or as recommended in the manufacturers’ protocol (in blocking buffer). The treatment of the membrane follows the same procedure as for the incubation carried out with a primary antibody/ protein. The incubation time should be at least 1 h. 9. There are several ready-to-use chemiluminescence detection substrates commercially available. For the detection on cellulose membranes, we recommend the “SuperSignal West Dura Extended Duration Substrate” (Pierce).

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

123

10. The spectral properties of these fluorophores do not interfere with autofluorescence phenomena often observed in proteins. Furthermore, they possess the photostability required for microarray imaging, and their excitation and emission spectra are compatible with most fluorescence instrumentations. 11. Both LifterSlips and FAST Frames come in different sizes and geometries. A suitable version has to be chosen according to the microarray dimensions. The advantage of using FAST Frames is that the microarray slides can be agitated during the sample incubation to improve mass transport, whereas the incubation with LifterSlips relies on diffusive transport only. However, FAST Frames generally require a much larger sample volume than LifterSlips. 12. In order to achieve a constant physicochemical behavior, it is recommended to wash the membrane twice with DMF after the amino acid coupling and treat the membrane then with an appropriate amount of a mixture of capping solution with 2% DIPEA for at least 20 min. 13. For long-term storage, a reduction or loss of functionality is possible. Storage of the membranes at −80°C is recommended. If the membranes are stored at higher temperatures, there is a higher risk of losing activity. Before storage, wash the membrane thoroughly with MeOH or EtOH because any traces of amines could lead to a total loss of activity within a relatively short time! Air-dry the membrane in a fume hood or use a hair dryer without heat. In order to proceed with the synthesis after storage, treat the membrane once with DMF for about 20 min. 14. The staining has no influence in the synthesis process. Nevertheless, it is recommended as an indication of the completion of the previous coupling step and the presence of free amino groups. Free amino groups of coupled amino acids would be stained blue. In contrast to esterified membranes, the dyed background of homemade amino-alkyl ether-linked membranes may disappear only slowly over several coupling cycles. But it does not affect the quality of the synthesis. 15. For faster drying, the membrane could be additionally washed twice with DEE. 16. In order to avoid problems caused by side product formation (hardly soluble urea), shake the in situ-activated amino acid solutions for about 30 min; then centrifuge the mixtures and use the clear supernatant for the synthesis. 17. Because of its flexibility and linear structure, b-alanine is commonly used for the definition of the spot pattern. It works as a spacer to achieve a distance between the amino functions and the cellulose matrix. However, the homemade amino-alkyl

124

Winkler, Andresen, and Hilpert

ether-linked membranes and several commercially available membranes readily provide a distance between the cellulose and the amino functions that is larger than for home esterified membranes. In these cases, it is not necessary to use a spacer. The definition of spot pattern can also be performed by spotting of the activated C-terminal first amino acids. 18. The staining of the spots must not be too strong since a high amount of absorbed BPB could lead to an incorporation of some dye into the peptide. In that case, the removal of the dye is difficult and may affect detection after incubation. Due to varying acidity of the coupled amino acids and the built-up peptide chain, differences in the intensity of the staining of the spots are normal. For example, spots where aspartic acid or cysteine as last coupled amino acids may show no or very little staining, whereas alanine and lysine show usually a deep blue color. 19. The capping step after the amino acid coupling is optional. But it is recommended since the blocking of unreacted amino groups leads to lower number of side products. 20. After TFA treatment, the membrane may become very soft. Shake very gently and do not try to lift the membrane out until it becomes harder and less likely to break apart during the washing steps. 21. The CelluSpot method is the only technique for the preparation of peptide microarrays which is using solutions of peptide–cellulose conjugates (41). 22. Schiffbase-adducts of primary amines and aldehydes are reversible with the equilibrium at acidic pH in favor of the free peptide, whereas the hydrazone bond between the aromatic hydrazine-linker and the surface aldehyde is considerably stable. This further ensures the site-specific immobilization of the peptides. 23. If it is necessary to yield a free carboxyl group at the C-terminus, do not treat the membrane with ammonia gas. Instead punch out the spots, transfer them into vials, and treat them with an aqueous strong basic solution such as 50% ammonium hydroxide, 1% triethylamine, or 1 M sodium hydroxide solutions (15, 17). 24. Ready-to-use PEG membranes are available from AIMS Scientific and Intavis (42). Those membranes are more stable against high TFA concentration and may be treated only with deprotection solution A for 3.5 h. 25. Since many desiccators made from plastics are not inert to ammonia gas, it is strongly recommended to use glass desiccators only.

SPOT Synthesis as a Tool to Study Protein–Protein Interactions

125

26. In order to reduce possible contamination with side products from the edge of the spot, the diameter of the punched-out membrane disks should be smaller than the spot diameter. 27. If the spotting solutions are provided in a MTP for the printing process, it is important to use polypropylene MTPs to avoid the adsorption of peptides to the plate walls. 28. It is important to moisturize the chambers sparingly, otherwise the microspots tend to absorb too much water and merge on the surface. 29. Use of towels made of nonrecycled paper is recommended. The use of recycled towels leads often to disturbances during detection by reactions with trace materials in the paper. 30. In case the chemiluminescence signal is too high, continue the washing steps for 15–30 min and repeat the treatment with the substrate and the detection again. 31. After probing, the membrane must remain wet for regeneration. A previously dried membrane is much harder to regenerate than a wet one. 32. There is always the risk that the regeneration is incomplete. Therefore, it is recommended to use always new membranes and, if not, the membrane should be probed first with the negative control and then with the protein sample. For further probing with the same membrane, completeness of the regeneration should be checked by repeating the negative control. Stained spots are generally very hard to regenerate. To remove the dye, the membrane should be treated with DMF overnight. References 1. Frank R, Güler S, Krause S, Lindenmaier W (1991) Facile and rapid “spot-synthesis” of large numbers of peptides on membrane sheets. In: Giralt E, Andreu D (eds) Peptides 1990. Proceedings of the 21st European peptide symposium. ESCOM, Leiden, pp 151–152 2. Hilpert K, Winkler DFH, Hancock REW (2007) Cellulose-bound peptide arrays: preparation and applications. Biotechnol Genet Eng Rev 24:31–106 3. Kramer A, Schneider-Mergener J (1998) Synthesis and application of peptide libraries bound to continuous cellulose membranes. Meth Mol Biol 87:25–39 4. Reineke U, Volkmer-Engert R, SchneiderMergener J (2001) Applications of peptide arrays prepared by the SPOT-technology. Curr Opin Biotechnol 12:59–64

5. Frank R (1992) Spot-synthesis: an easy technique for the positionally addressable, parallel chemical synthesis on a membrane support. Tetrahedron 48:9217–9232 6. Hilpert K, Winkler DFH, Hancock REW (2007) Peptide arrays on cellulose support: SPOT synthesis – a time and cost efficient method for synthesis of large numbers of peptides in a parallel and addressable fashion. Nat Protoc 2:1333–1349 7. Frank R (2002) The SPOT-synthesis technique. Synthetic peptide arrays on membrane supports-principles and applications. J Immunol Meth 267:13–26 8. Bowman MD, Jacobson MM, Blackwell HE (2006) Discovery of fluorescent cyanopyridine and deazalumazine dyes using small molecule macroarrays. Org Lett 8:1645–1648

126

Winkler, Andresen, and Hilpert

9. Heine N, Ast T, Schneider-Mergener J, Reineke U, Germeroth L, Wenschuh H (2003) Synthesis and screening of peptoid arrays on cellulose membranes. Tetrahedron 59:9919–9930 10. Kramer A, Schneider-Mergener J (1995) Highly complex combinatorial cellulosebound peptide libraries for the detection of antibody epitopes. In: Maia HLS (ed) Peptides 1994. Proceedings of the 23rd European peptide symposium. ESCOM, Leiden, pp 475–476 11. Kramer A, Volkmer-Engert R, Malin R, Reineke U, Schneider-Mergener J (1993) Simultaneous synthesis of peptide libraries on single resin and continuous cellulose membrane supports: examples for the identification of protein, metal and DNA binding peptide mixtures. Pept Res 6:314–319 12. Hilpert K, Elliott M, Jenssen H, Kindrachuk J, Fjell CD, Körner J, Winkler DFH, Weaver LL, Henklein P, Ulrich AS, Chiang SH, Farmer SW, Pante N, Volkmer R, Hancock REW (2009) Screening and characterization of surface-tethered cationic peptides for antimicrobial activity. Chem Biol 16:58–69 13. Otvos L Jr, Pease AM, Bokonyi K, Giles-Davis W, Rogers ME, Hintz PA, Hoffman R, Ertl HCJ (2000) In situ stimulation of a T helper cell hybrodoma with a cellulose-bound peptide antigen. J Immunol Meth 233:95–105 14. Hilpert K, Elliott MR, Volkmer-Engert R, Henklein P, Donini O, Zhou Q, Winkler DFH, Hancock REW (2006) Sequence requirements and an optimization strategy for short antimicrobial peptides. Chem Biol 13:1101–1107 15. Bhargava S, Licha K, Knaute T, Ebert B, Becker A, Grötzinger C, Hessenius C, Wiedemann B, Schneider-Mergener J, Volkmer-Engert R (2002) A complete substitutional analysis of VIP for better tumor imaging properties. J Mol Recognit 15:145–153 16. Reimer U, Reineke U, Schneider-Mergener J (2002) Peptide arrays: from macro to micro. Curr Opin Biotechnol 13:315–320 17. Lizcano JM, Deak M, Morrice N, Kieloch A, Hastie CJ, Dong L, Schutkowski M, Reimer U, Alessi DR (2002) Molecular basis for the substrate specificity of NIMA-related kinase-6 (NEK-6). Evidence that NEK-6 does not phosphorylate the hydrophobic motif of ribosomal S6 protein kinase and serum- and glucocorticoid-induced protein kinase in vivo. J Biol Chem 277:27839–27849 18. Gausepohl H, Behn C (2002) Automated synthesis of solid-phase bound peptides. In: Koch J, Mahler M (eds) Peptide arrays on

membrane support. Springer, Berlin, pp 55–68 19. Weiler J, Gausepohl H, Hauser N, Jensen ON, Hoheisel JD (1997) Hybridisation based DNA screening on peptide nucleic acid (PNA) oligomer arrays. Nucleic Acids Res 25:2792–2799 20. Blackwell HE (2006) Hitting the SPOT: small-molecule macroarrays advance combinatorial synthesis. Curr Opin Chem Biol 10:203–212 21. Hoffman B, Ast T, Polakowski T, Reineke U, Volkmer R (2006) Transformation of biologically active peptide into peptoid analogs while retaining biological activity. Pept Prot Lett 13:829–833 22. Jobron L, Hummel G (2000) Solid-phase synthesis of unprotected N-glycopeptide building blocks for SPOT synthesis of glycopeptides. Angew Chem Int Ed 39:1621–1624 23. Wildemann D, Erdmann F, Alvarez BH, Stoller G, Zhou XZ, Fanghänel J, Schutkowski M, Lu KP, Fischer G (2006) Nanomolar inhibitors of the peptidyl prolyl cic/trans isomerase Pin1 from combinatorial peptide libraries. J Med Chem 49:2147–2150 24. Fields GB, Noble RL (1990) Solid phase synthesis utilizing 9-fluorenylmethoxycarbonyl amino acids. Int J Pept Prot Res 35: 161–214 25. Zander N, Gausepohl H (2002) Chemistry of Fmoc peptide synthesis on membranes. In: Koch J, Mahler M (eds) Peptide arrays on membrane support. Springer, Berlin, pp 23–39 26. Atherton E, Sheppard RC (1989) 7.2 Activated esters of Fmoc-amino acids. In: Atherton E, Sheppard RC (eds) Solid phase peptide synthesis – a practical approach. Oxford University Press, Oxford, pp 76–78 27. Volkmer R (2009) Synthesis and application of peptide arrays: Quo vadis SPOT technology. Chem Biol Chem 10:1431–1442 28. Kamradt T, Volkmer-Engert R (2004) Crossreactivity of T lymphocytes in infection and autoimmunity. Mol Divers 8:271–280 29. Licha K, Bhargava S, Rheinlander C, Becker A, Schneider-Mergener J, Volkmer-Engert R (2000) Highly parallel nano-synthesis of cleavable peptide-dye conjugates on cellulose membranes. Tetrahedron Lett 41:1711–1715 30. Ast T, Heine N, Germeroth L, SchneiderMergener J, Wenschuh H (1999) Efficient assembly of peptomers on continuous surfaces. Tetrahedron Lett 40:4317–4318

SPOT Synthesis as a Tool to Study Protein–Protein Interactions 31. Krchnak V, Wehland J, Plessmann U, Dodemont H, Gerke V, Weber W (1988) Noninvasive continuous monitoring of solid phase peptide synthesis by acid-base indicator. Collect Czech Chem Commun 53:2542–2548 32. Molina F, Laune D, Gougat C, Pau B, Granier C (1996) Improved performances of spot multiple peptide synthesis. Pept Res 9:151–155 33. Andresen H, Grötzinger C, Zarse K, Kreuzer OJ, Ehrentreich-Förster E, Bier FF (2006) Functional peptide microarrays for specific and sensitive antibody diagnostics. Proteomics 6:1376–1384 34. Andresen H, Grötzinger C (2009) Deciphering the antibodyome – peptide arrays for serum antibody biomarker diagnostics. Curr Proteomics 6:1–12 35. Winkler DFH, McGeer PL (2008) Protein labeling and biotinylation of peptides during spot synthesis using biotin p-nitrophenyl ester (biotin-ONp). Proteomics 8:961–967 36. Andresen H, Bier FF (2009) Peptide microarrays for serum antibody diagnostics. Meth Mol Biol 509:123–134

127

37. Boisguerin P, Leben R, Ay B, Radziwill G, Moelling K, Dong L, Volkmer-Engert R (2004) An improved method for the synthesis of cellulose membrane-bound peptides with free C termini is useful for the PDZ domain binding studies. Chem Biol 11:449–459 38. Volkmer-Engert R, Hoffmann B, SchneiderMergener J (1997) Stable attachment of the HMB-linker to continuous cellulose membranes for parallel solid phase spot synthesis. Tetrahedron Lett 38:1029–1032 39. Tapia V, Ay B, Triebus J, Wolter E, Boisguerin P, Volkmer R (2008) Evaluating the coupling efficiency of phosphorylated amino acids for SPOT synthesis. J Pept Sci 14:1309–1314 40. Rau HK, DeJonge N, Haehnel W (2000) Combinatorial synthesis of four-helix bundle hemoproteins for tuning of cofactor properties. Angew Chem Int Ed 39:250–253 41. Beutling U, Städing K, Stradal T, Frank R (2008) Large-scale analysis of protein-protein interactions using cellulose-bound peptide arrays. Adv Biochem Eng Biotechnol 110:115–152 42. Zander N (2004) New planar substrates for the in situ synthesis of peptide arrays. Mol Divers 8:189–195

wwwwwwwwwwwwwwww

Chapter 9 Native Antigen Fractionation Protein Microarrays for Biomarker Discovery Robert J. Caiazzo, Jr., Dennis J. O’Rourke, Timothy J. Barder, Bryce P. Nelson, and Brian C.-S. Liu Abstract In this protocol, we used the T24 human bladder cancer cell line as a source of native antigens to construct fractionated lysate microarrays. Subsequently, these microarrays were used to compare the autoantibody responses of individuals with interstitial cystitis/painful bladder syndrome (IC/PBS) to those of normal female controls. To accomplish this, T24 cells were lysed under nondenaturing conditions to obtain native antigens. These native antigens were then fractionated in 2D using a PF-2D liquid chromatography; the first dimension separated the proteins by their isoelectric points, and the second separated them according to hydrophobicity. The resulting protein fractions were printed onto nitrocellulose-coated glass slides (PATH slides) to create a set of fractionated lysate microarrays. To compare the autoantibody responses of IC/PBS patients with normal controls, the fractionated lysate arrays were competitively hybridized with fluorescently labeled IgG samples purified from both IC/PBS and control sera. This protocol presents a detailed description of the creation and use of native antigen fractionated lysate microarrays for autoantibody profiling. Key words: Autoantibody profiling, Protein microarrays, Cell lysate, Protein fractionation, PF-2D, Autoantibodies, Mass spectroscopy

1. Introduction An important sector of proteomic research today is the hunt for new and novel disease-specific biomarkers, especially those with excellent diagnostic and prognostic capabilities. The discovery of such markers will not only allow clinicians to better detect diseases, like cancer, but may also lead to the development of new and personalized treatments. The search for these diagnostic and prognostic markers is greatly aided by the high-throughput, multiplex capabilities of protein microarray technologies. Protein Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_9, © Springer Science+Business Media, LLC 2011

129

130

Caiazzo et al.

microarrays are typically found in two formats: The forward-phase array (FPA) and the reverse-phase array (RPA). FPAs are constructed by printing many bait molecules (such as antibodies) onto an array surface, allowing for the profiling of complex samples; such arrays are usually probed with cell lysates or patient serum samples (1). RPAs, however, are constructed by printing multiple test samples and can be probed using a single detector, such as an antibody (2, 3). While the reverse-phase format was originally used to identify biomarkers in laser capture microdissection (LCM) samples, RPAs can be constructed using a wide variety of test samples including Serum (4), bodily fluids (cerebrospinal fluids, synovial fluid, etc.) (1), cell lysate (3, 5), and protein fractions (6–8). For more information on forward and reverse phase microarrays, please refer to the following references (1, 9). In this chapter, we will focus on a specific type of RPAs – the construction and use of native antigen fractionated lysate microarrays for autoantibody profiling. While most antigen microarrays use recombinant proteins to profile autoantibodies, native antigen fractionation protein microarrays use proteins that are extracted from cell lines or tumor lysates. The use of recombinant proteins or synthetic peptides as printed antigens has limitations, as they lack proper posttranslational modifications and may not represent the unique modifications that are associated with a disease state. The native protein microarrays, therefore, include potentially unique disease-associated epitopes that are not present in recombinant proteins. To date, fractionated lysate arrays have been utilized in the study of both lung and prostate cancer (6–8). Here we use fractionated lysate arrays to compare the autoimmune profile of interstitial cystitis/painful bladder syndrome (IC/PBS) patients to those of normal female controls. IC is a debilitating, chronic bladder syndrome. Urinary frequency, urgency, and pelvic pain are the major symptoms. It is well established, however, that inflammation is associated with IC. And although autoimmunity is debated as a potential cause, certain aspects of IC suggest that it may play a role in initiating or sustaining the chronic inflammatory response evident in this disease. For example, the degeneration of bladder epithelial cells that occurs as a result of chronic inflammation in IC may stimulate the production of autoantibodies. Thus, the presence of inflammation/autoimmunity in IC may allow the use of the body’s own immune response as a means of identifying biomarkers of IC. To create the protein fractions, cellular lysate obtained from the T24 human bladder cancer cell lines was subjected to a 2D fractionation procedure. In the first dimension, proteins were separated according to differences in their isoelectric point (pI), through isoelectric focusing or chromatofocusing. In the second dimension, the collected fractions were separated according to hydrophobicity using reverse phase high-performance liquid chromatography.

Native Antigen Fractionation Protein Microarrays

131

Once the lysate was fractionated, microarrays were constructed by printing the protein fractions onto a support surface, such as nitrocellulose-coated PATH slides or APiX slides. These fractionated lysate microarrays were then used to compare the autoantibody response of individuals with IC/PBS to that of healthy controls. To do this, the microarrays were competitively incubated with fluorescently labeled IgG samples from IC/PBS and control patients using a two slide dye-swap procedure. In this procedure, a portion of each IgG sample is labeled with a different fluorophore, e.g., Cy™5 and Cy™3. This sampling method eliminates potential variability in labeling efficiency. For example, if Cy™5 reacts more efficiently with the IgG than Cy™3 does, the results will be biased in favor of the sample labeled with Cy™5. With the two slide dye-swap method, however, this potential variability is eliminated because each sample is labeled with each dye, and allows a ratio-based normalization for each spot. While this protocol focuses on the construction and use of fractionated lysate arrays to examine the autoantibody response of IC/PBS, it can be adapted for the study of autoantibody expression in a variety of conditions in which autoimmunity plays a role (e.g., cancer).

2. Materials 2.1. Construction of Fractionated Native Antigen Lysate Microarrays

1. T24 human bladder cancer cell line (see Note 1).

2.1.1. Cell Culture and Lysis

4. 0.05% Trypsin-EDTA (Gibco, Bethesda, MD).

2. Cell culture plates or flasks. 3. Cell culture medium (RPMI medium 1640 plus 10% FBS, 100 IU/mL penicillin and 100 mg/mL streptomycin). 5. 2 mL aspirating pipettes. 6. 10 mL serological pipettes. 7. 15 mL conical centrifuge tubes. 8. Phosphate buffered saline (1×) 9. Lysis buffer: 7.5 M urea, 2.5 M thiourea, 12.5% glycerol, 50 mM Tris, 2.5% n-octylglucoside, 1.25 mM of protease inhibitor. 10. PD-10 desalting column (GE Healthcare, Piscataway, NJ). 11. ProteoSep Start Buffer (SB) (Eprogen, Darien, IL).

2.1.2. Protein Fractionation

1. PD-10 desalting column (GE Healthcare). 2. ProteoSep Start Buffer (SB) (Eprogen). 3. ProteoSep Eluent Buffer (EB) (Eprogen). 4. ProteoSep Chromatofocusing (CF) column (Eprogen).

132

Caiazzo et al.

5. ProteoSep High Performance Reverse Phase (HPRP) Column (Eprogen). 6. 2.0 mL 96-well plate (Axygen, Union City, CA). 7. 650 mL 96-well plates (Orochem, Lombard, IL). 8. HPLC Grade Water. 9. HPLC Grade Acetonitrile. 10. Trifluoroacetic Acid (TFA). 11. A Beckman ProteomeLab PF2D-automated instrument for both first and second dimension fractionation and analysis or an HPLC capable of fraction collection based on pH (e.g., GE Healthcare AKTA Purifier) and an HPLC capable of >5,000 psi operation for the second dimension analysis. 12. Fraction collector capable of 2 mL fraction collection into a deep 96-well plate from the first dimension CF fractionation and a Fraction collector capable of using 96-well plates. 2.1.3. Microarray Printing

1. PATH Protein Microarray Slides (Gentel Biosciences, Madison, WI). 2. Non-Contact Microarray Printer with 150 mm Pins (Genetix, Ltd., Boston, MA). 3. SpeedVac Centrifuge. 4. 384 well low protein binding multiwell plate. 5. ddH2O or equivalent. 6. PBS (1×). 7. Glycerol (Sigma, St. Louis, MO). 8. Gentel Blocking Buffer (1×) (Gentel Biosciences).

2.2. Sample Preparation and Microarray Experiments

1. Serum samples acquired in accordance with IRB-approved protocol and stored at −80°C.

2.2.1. Purification of IgG

4. Microcentrifuge.

2. Pipettes. 3. Pipette tips. 5. Vortex. 6. Melon™ Gel IgG Spin Purification Kit (Thermo Fisher Scientific, Rockford, IL) containing Melon™ Gel IgG Purification Support, Melon™ Gel Purification Buffer, and Handee™ Mini-Spin Columns and accessories. 7. End-over-end rotator. 8. BCA™ Protein Assay Reagent Kit (Thermo Fisher Scientific) containing BCA™ Reagent A, BCA™ Reagent B, and Albumin standard ampules (2 mg/mL).

Native Antigen Fractionation Protein Microarrays

133

9. 96-well plate or other size-appropriate plate for spectrometer. 10. Spectrometer capable of reading absorbance at 562 nm. 2.2.2. Labeling of IgG with Fluorescent Dyes

1. Cy™3 monofunctional reactive dye pack (GE Healthcare). 2. Cy™5 monofunctional reactive dye pack (GE Healthcare). 3. Protein Extraction & Labeling kit containing Extraction/ Labeling Buffer, Blocking Buffer, and Desalting Buffer (Clontech, Mountain View, CA). 4. Microcentrifuge. 5. Vortex. 6. Pipettes. 7. Pipette tips. 8. Micropipette tips. 9. 1.5 mL microcentrifuge tubes.

2.2.3. Removal of Unbound Dye

1. Protein Desalting Spin Columns (Thermo Fisher Scientific). 2. Protein Extraction & Labeling kit containing Extraction/ Labeling Buffer, Blocking Buffer, and Desalting Buffer (Clontech). 3. Microcentrifuge. 4. Vortex. 5. Pipettes. 6. Pipette tips. 7. 1.5 mL microcentrifuge tubes. 8. 2.0 mL microcentrifuge tubes.

2.2.4. Microarray Incubation with Patient IgG

1. PATH Protein Microarray Slides printed with Fractionated T24 Lysate (from Subheading 2.1.4). 2. Wash Buffer: 1× PBS, 0.05% Tween-20. 3. Incubation buffer: 1× PBS, 0.05% Tween-20, 1% BSA (see Note 2). 4. 15 mL conical centrifuge tubes. 5. Incubation tray (see Note 3). 6. Forceps. 7. Pipettes. 8. Pipette tips. 9. Rocking platform. 10. Kimwipes. 11. Swinging bucket centrifuge with adaptor for 50 mL tubes. 12. Phosphate buffered saline (1×).

134

Caiazzo et al.

13. Ultrapure water. 14. Screw cap slide container (Thermo Fisher Scientific). 2.3. Results and Analysis 2.3.1. Microarray Scanning and Quantitation

1. Microarray scanner compatible with 75 × 25 × 1 mm slides and capable of dual color analysis. The scanner must be capable of measuring fluorescence in the ranges of the Cy3 and Cy5 fluorescent labels. 2. Microarray scanning software. 3. Microarray quantification software 4. Microsoft Excel.

2.3.2. Statistical Analysis

1. Microarray data analysis software.

3. Methods 3.1. Construction of Fractionated Native Antigen Lysate Microarrays 3.1.1. Cell Culture and Lysis

1. The T24 human bladder cancer cells are passaged when approaching confluence using 1× trypsin/EDTA to create maintenance cultures on 20 × 100 mm tissue culture plates. For cell lysis and subsequent protein fractionation, a cell pellet that is roughly 0.5 mL in size (or ~108 cells) is needed; prepare a suitable number of tissue culture plates to meet this need and grow them until ~90% confluent (see Note 1). 2. Add 1–2 mL of 1× trypsin/EDTA per plate of cells and place the plates into a 37°C cell culture incubator until the cells begin to detach (~5–10 min). 3. Rinse the plates with cell culture medium containing serum, and pipette the medium containing the detached cells into appropriately sized centrifuge tube. Centrifuge the tubes at ~1,000 × g for 5 min to pellet the cells. 4. Remove the supernatant from the centrifuge tubes, and rinse the cell pellet with 5 mL of 1× PBS. Centrifuge the tube again to repellet the cells. 5. Repeat step 4 to ensure that the cell pellet is thoroughly washed. Remove the supernatant (see Note 4). 6. Mix ~0.5 mL of the thoroughly washed cell pellet (or ~108 cells) in 2.0 cc of Lysis buffer and vortex vigorously. 7. Centrifuge the lysate at >6,000 × g for 15 min at 4°C. 8. Remove the supernatant from the cell debris and refrigerate at 2–8°C until needed. For long term storage (>1 week), keep frozen at −20°C. 9. Equilibrate a PD-10 Column with 25 cc of ProteoSep Start Buffer (SB) (see Note 5).

Native Antigen Fractionation Protein Microarrays

135

10. Add the 2.5 cc of the centrifuged sample to the PD-10 column and discard the eluent. 11. Elute and collect the lysis proteins with 3.5 cc of SB and dilute to 5.0 cc with SB. 3.1.2. Protein Fractionation: First Dimension – Chromatofocusing

1. Prepare the working buffers for the first dimension as follows: (a) Start buffer (SB) i. Warm to room temperature and sonicate for 5 min. ii. Using a calibrated pH meter, adjust the pH to 8.5 ± 0.1 using either a saturated solution of iminodiacetic acid (IDA) if the buffer is too basic or 1 M NH4OH if the buffer is too acidic (see Note 6). iii. Store in refrigerator (4–8°C) until needed. (b) Eluent buffer

1. Warm to room temperature and sonicate for 5 min.

2. Using a calibrated pH meter, adjust the pH to 4.0 ± 0.1 using either a saturated solution of IDA if the buffer is too basic or 1 M NH4OH if the buffer is too acidic.

3. Store in refrigerator (4–8°C) until needed.

2. Set the operating conditions for the HPLC system as follows (these settings are for an HPLC system with a single pump, injector, UV, and pH detectors): Column: HPCF 1D Column. Flow rate: 0.2 mL/min. Detection: UV 280 nm. Temperature: Ambient. Mobile phase: Start buffer: pH 8.5 ± 0.1. EB: pH 4.0 ± 0.1. 3. First dimension chromatofocusing procedure: (a) If using a Beckman ProteomeLab PF2D instrument, inject 5.0 mL of the PD-10 exchanged lysed sample using the default 2D fraction collection method preprogrammed into the PF2D computer. Loading 1–5 mg of total protein on the CF column is recommended (If using an AKTA Purifier or equivalent pH controlled HPLC instrument, before injection, flush the HPCF column with 100% HPLC grade water for 10CV (1CV = 0.87 mL), then equilibrate column with 30 CV of SB. After equilibration, inject 5.0 mL of the PD-10 exchanged lysed sample. Loading 1–5 mg of total protein on the CF column is recommended.).

136

Caiazzo et al.

(b) After injection, wash the CF column with 100% start buffer at a flow rate of 0.2 mL/min while monitoring the baseline absorbance. Collect this eluent wash as a single fraction since all proteins with pI’s >8.5 will elute from the column during this wash step. After the UV absorbance values return to baseline, stop the flow of Start Buffer eluent. This wash step will take ~10–30 min (see Note 7). (c) Before performing the pH gradient, flush the tubing and pump head with 100% EB (the flush volume depends on the system configuration) in order to facilitate starting of the pH gradient. Initiate start of 100% EB at 0.2 mL/ min for 95 min to perform the pH gradient. At ~70 min the pH of the eluent should be pH 4.0 ± 0.1. Collect fractions every 0.3 pH units using a fraction collector controlled by pH change. (d) After completing the EB run, wash the HPCF column with 10 column volumes of a 1.0 M sodium chloride 30% n-PrOH solution (filtered through 0.45 mm membrane filter). Collect the first 4 mL of this eluent as a fraction for analysis with the second dimension HPRP column. (e) Transfer the HPCF pH fractions from the fraction collector to appropriate vials for use with the HPRP HPLC autosampler for the second dimension HPRP analysis. Fractions should be stored at 2–8°C if the second dimension analysis will be delayed for more than ~8–10 h. (f) After the salt wash, wash the CF column with 10 column volumes of HPLC grade water. 3.1.3. Protein Fractionation: Second Dimension – Reversed Phase HPLC

1. If using a Beckman ProteomeLab PF2D instrument, inject 250 mL of each CF fraction collected from the first dimension fractions using the default 2D fraction collection method preprogrammed into the PF2D computer. Collect second dimension fractions in the 650 mL 96-well plates, collecting the effluent from the HPRP column with retention times between 10 and 22 min. If using another type of HPLC, use the following operating conditions (these setting are for an HPLC system with low dead volume mixing and <10 mL UV flow cell and column heater capable of >5,000 psi operation): Column: HPRP 2D Column Mobile phase: A: 0.1% TFA in water (HPLC grade) B: 0.08% TFA in acetonitrile (HPLC grade) Gradient: (a) 100% A for 2 min (b) 0–100% B in 30 min.

Native Antigen Fractionation Protein Microarrays

137

(c) Hold at 100% B for 5 min. (d) 100% B to 100% A in 2 min. (e) 100% A for 5 min. Flow rate: 0.75 mL/min Detection: UV 214 nm Temperature: 50°C Injection vol.: 250 mL injections are recommended for each HPCF fraction collected in the first dimension. Less can be injected if desired, but 250 mL is the recommended minimum injection volume for detecting low level expressed proteins present using standard HPLC equipment. 2. Save the Raw UV absorbance data or each HPRP analysis of the CF fractions for protein mapping and data analysis using the ProteoVue Software (Eprogen). 3.1.4. Microarray Printing and Blocking

The 96-well plates containing the second dimension fractions were used to fabricate the microarrays (see Note 8). 1. Add 50 mL of 40% glycerol in PBS to each well already containing a water/acetonitrile/TFA mixture. 2. Evaporate the plates using a SpeedVac centrifuge until only 20 mL of the glycerol remains. 3. Add 30 mL of PBS to each well to make a 40% glycerol/PBS print buffer and transfer into two sets of 384 well plates for printing. 4. Using a solid pin QArray2 (Genetix, Ltd.) print the arrays on a suitable support surface, in this case the arrays were printed on thin layer nitrocellulose PATH and Apix slides (Gentel Biosciences, Inc.) (see Note 9). 5. To allow the slides to dry to completion and prevent spot migrations, the printed slides should be set aside for at least 72 h prior to blocking them. To block the slides, immerse them in 1× Gentel Block Buffer for 1 h. Afterward, the slides can be air dried and stored in a sealed slide box under desiccating conditions until use.

3.2. Sample Preparation and Microarray Experiments 3.2.1. Purification of IgG Using Melon™ Gel Kit

The following steps may be carried out several days prior to the rest of the microarray experiment. Purified IgG can be stored for up to 1 week at 4°C. If IgG is to be stored for longer than 1 week, aliquots should be placed in a −20°C freezer for storage until use; avoid repeated freeze/thaw cycles. 1. Equilibrate the Melon™ Gel IgG Purification Support and Purification Buffer to room temperature (~30 min) and swirl the bottle containing the Purification Support (do not vortex) to obtain an even suspension. To ensure proper gel

138

Caiazzo et al.

slurry dispensing, use a wide bore or cut pipette tip to dispense 500 mL of gel slurry into a Handee™ Mini-Spin Column placed in a microcentrifuge tube. Swirl the bottle of gel slurry before pipetting each sample to maintain an even gel suspension. 2. Centrifuge the uncapped column/tube assemblies for 1 min at 3,000 × g, then remove the spin columns and discard the flow-through. 3. Add 300 mL of Purification Buffer to the column, pulse the centrifuge for 10 s and discard the flow-through. Repeat this wash once. Place the bottom caps on the columns. 4. Add 50 mL of each serum sample diluted 1:10 in 1× Melon™ Gel Purification Buffer to a column. Cap the columns and incubate for 5 min at room temperature with end-over-end rotation. 5. Remove the bottom caps from the columns, loosen the top caps and insert the spin columns into fresh 2 mL collection tubes. Then, centrifuge for 1 min at 3,000 × g to collect the purified antibody in the collection tubes. 6. Set up a new column corresponding to each sample that has been purified, and repeat steps 2–5 in order to further purify the collected IgG using fresh Melon™ Gel (see Note 10). 7. Measure the concentration of IgG in each purified sample using Pierce’s BCA™ Protein Assay Reagent Kit. 8. Dilute each sample to 1 mg antibody/mL by adding the appropriate volume of 1× Melon™ Gel Purification Buffer (see Note 11). The final amount of IgG should be at least 200 mg per sample. 3.2.2. Labeling of IgG with Fluorescent Dyes

The following steps may be carried out several days prior to the rest of the microarray experiment. However, Subheadings 3.2.2 and 3.2.3 must be carried out together so that unbound dye is removed before storage of the labeled IgG. Labeled IgG can be stored for a maximum of 1 week at 4°C protected from light. If labeled IgG is to be stored for longer than 1 week, aliquots should be protected from light in a −20°C freezer until use. Repeated freeze/thaw cycles should be avoided. Once the Cy™ dyes are reconstituted, they must be used immediately. 1. Set up and label one 0.5 mL microfuge tube for each sample (four tubes total: A-Cy™3, A-Cy™5, B-Cy™3, B-Cy™5). 2. Transfer 100 mg of the appropriate purified antibody (1 mg/mL) from Subheading 3.1 to the corresponding tube prepared in step 1. 3. Tap the bottom of the Cy™ Dye reagent vials against a hard surface to ensure that there is no dye in the caps and reconstitute

Native Antigen Fractionation Protein Microarrays

139

one vial of Cy™3 dye and one vial of Cy™5 dye by adding 110 mL of Clontech Labeling/Extraction Buffer to each vial. 4. Vortex the two dye vials for 20 s and briefly centrifuge to collect the reconstituted dyes at the bottom. 5. Add 4 mL of Cy™3 to each of the corresponding tubes from step 2 (see Note 12). 6. Add 2 mL of Cy™5 to each of the corresponding tubes from step 2 (see Note 12). 7. Vortex the four microfuge tubes gently and briefly centrifuge the microfuge tubes to collect the samples at the bottom. 8. Incubate the tubes for 90 min at 4°C protected from light. Mix the samples every 20 min by inverting the microfuge tubes. 9. Add 4 mL of Clontech Blocking Buffer to each tube. 10. Incubate the tubes for 30 min at 4°C protected from light. Mix the samples every 10 min by gently vortexing the microfuge tubes. 11. Proceed immediately with the removal of unbound dye. 3.2.3. Removal of Unbound Dye with Desalting Columns

As long as you work quickly, desalting can be completed at room temperature. Otherwise, if you have access to a cold room, we suggest you complete the procedure at 4°C. 1. Set up and label one 1.5 mL microcentrifuge tube and one Protein Desalting Spin Column for each dye labeled sample (see Note 13). 2. Twist off the bottom of each column and loosen the caps before placing each one in its collection tube and centrifuge each column at 1,500 × g for 2 min to remove the storage buffer. Note the side of each column where the compacted resin is slanted upward, and be sure to place the columns in the centrifuge with this area of the column facing outward in all subsequent centrifugations. 3. Transfer each desalting column to a fresh 2 mL microcentrifuge tube. 4. Add 400 mL 1× Clontech Desalting Buffer to each of desalting columns. Then, centrifuge the columns for 2 min at 1,500 × g. Discard the flow-through. Repeat this step once. 5. Blot the bottom of the columns against a laboratory tissue to remove excess liquid, and place the columns into a fresh, labeled 1.5 mL microcentrifuge tube. 6. Carefully apply the Cy™3 and Cy™5 labeled samples (~105 mL) directly onto the center of the resin bed of the corresponding column. Allow the samples to pass into the columns.

140

Caiazzo et al.

7. Centrifuge the columns for 2 min at 1,500 × g to collect the desalted, labeled antibodies. 8. At this point, purified, labeled IgG samples can be stored for future use, or used in Subheading 3.2.4 of a current experiment. 3.2.4. Antibody Array Incubation with Patient IgG

These microarray experiments utilize a dye swap procedure (see Note 14), which necessitates the use of two microarray slides. One slide will be incubated with the “Mix 1” samples (IgG A-Cy™3 and IgG B-Cy™5) and the other slide will be incubated with the “Mix 2” samples (IgG A-Cy™5 and IgG B-Cy™3). Be sure to record the barcodes printed on the microarray slides so that they can be identified as either Mix 1 or Mix 2. 1. Label one 15 mL conical tube “Mix 1” and label one 15 mL conical tube “Mix 2”; then add 5 mL of Incubation Buffer (1× PBS, 0.05% Tween-20, 1% BSA) to each of these tubes. 2. Transfer the entire sample of IgG A-Cy™3 and IgG B-Cy™5 (200 mg total; from Subheading 3.2.3) to the Mix 1 tube from step 1. Store the tube at 4°C until needed. 3. Transfer the entire sample of IgG A-Cy™5 and IgG B-Cy™3 (200 mg total; from Subheading 3.2.3) to the Mix 2 tube from step 1. Store the tube at 4°C until needed. 4. Add the PATH fractionated lysate microarrays to two of the chambers of the incubation tray. Cover each slide with 5 mL of Wash Buffer, and rock the slides at room temperature for 1 min. Remove the buffer from the incubation chambers. Repeat this wash five times, for a total of 6 washes. 5. Add the contents of the tube labeled “Mix 1” to incubation chamber 1 and “Mix 2” to incubation chamber 2. 6. Incubate the fractionated protein array slides at room temperature for 1 h with gentle rocking. Every 15 min, use a pipette tip to lift one end of the slide while gently rocking the incubation tray. 7. Add 5 mL of Wash Buffer to each wash chamber, transfer the slides to their respective wash chambers, and incubate at room temperature for 1 min with gentle rocking. 8. Remove the buffer from the wash chambers. 9. Repeat steps 7–8 five times, for a total of 6 washes. 10. Add 5 mL of PBS (1×) to each wash chamber and incubate at room temperature for 1 min with gentle rocking. 11. Remove the PBS from the wash chambers. 12. Add 5 mL of ultrapure water to each wash chamber and incubate at room temperature for 1 min with gentle rocking. 13. Remove the water from the wash chambers.

Native Antigen Fractionation Protein Microarrays

141

14. Dry the slides. It is important to remove as much moisture as possible from the surface of the slides before the liquid evaporates passively. (a) Using scissors or a knife, puncture a small, round hole in the bottom of the slide container. This will facilitate the removal of excess liquid from the slides during centrifugation. (b) Using gloved hands and touching only the edges of the slides, hold the slides so that the excess liquid drips toward the bottom of the array slides (the area containing the manufacturer’s label/barcode) and gently touch this edge to a clean Kimwipe several times. (c) Carefully place the slides in the empty slide container with the ends containing the manufacturer’s label/barcode at the bottom of the vial. Do not touch the array surface. (d) Cap the vial and centrifuge the slides at ~1,000 × g for 25 min at room temperature. 15. Proceed immediately with microarray scanning (Subhead ing 3.3.1). 3.3. Results and Analysis 3.3.1. Microarray Scanning and Quantitation (see Note 15)

Antibody microarray slides should be scanned using a laser scanner, such as the Axon GenePix 4000B or the Perkin Elmer ScanArray 4000, according to the manufacturer’s specifications. The scanner must be able to measure fluorescence in the ranges of the Cy™3 and Cy™5 fluorophores. 1. Turn on the scanner and allow the lasers to warm up. The lasers on the Perkin Elmer ScanArray 4000 require 15 min to warm-up prior to scanning your arrays. 2. Run a quick/preview scan of the entire slide in order to determine the area containing the arrayed features. 3. Create a scan protocol on the computer attached to the microarray scanner (see Note 16). (a) Set the protocol to scan for Cy™3 and Cy™5 fluorescence. (b) Determine the area containing the arrayed protein fractions and select this as the portion of the array to be scanned. (c) If possible, select an area on the array that contains no arrayed features to be scanned for background intensity. (d) Set the laser powers and PMT (photomultiplier) Gains so that the signal is high enough without being saturated (see Note 17). We suggest the following settings for scanning with the Perkin Elmer ScanArray 4000 Cy3: PMT = 62%; laser power = 90% Cy5: PMT = 50%; laser power = 90%

142

Caiazzo et al.

4. Insert the first slide, containing the Mix 1 autoantibodies, into the scanner. For the ScanArray 4000, the microarray surface is inserted face up. 5. Begin the scan and make sure that the signal is sufficiently high, but not saturated. You may need to readjust the laser and PMT Gain settings or rerun an Automatic Sensitivity Calibration if the image is saturated or the signal is too low. 6. Save the Cy™3 and Cy™5 images as separate TIFF files. The single-file TIFF format is the most useful for quantifying data from the images. 7. Scan the second microarray slide, containing the Mix 2 autoantibodies, using the same settings used to scan the first slide and save the images in the same manner. 8. Obtain the GAL file that corresponds to your fractionated lysate microarrays. This file should be available from the institution/company that printed the microarrays. If you have printed your own slides, you will need to create a GAL file; information on creating your own GAL file can be found at: http://www.moleculardevices.com/pages/software/gn_gal_ examples.html 9. Using the GenePix Pro software, use the “Alt Y” command to open the GAL file from step 8 and open the TIFF files corresponding to your first slide (Mix 1) using the “Alt O” command. 10. Automatically align the grid with the array features using “F8.” Use the Zoom-In feature to ensure the proper alignment of the grid with the features of your array. You can adjust the fit of the Grid to the entire array or to individual array features using the tools on the left-hand side of the screen (see Note 18). 11. Carry out an Automatic Analysis using the “Alt A” command (see Note 19) and save your data as a GPR (GenePix Results) file using the “Alt U” command (see Note 20). 12. Data can be exported to Excel using the “Ctrl A” command to select all data, followed by the “Ctrl C” command to copy all data to the clipboard. 13. Data from your first slide can now be pasted into a blank Excel sheet with the “Ctrl V” command. 14. Repeat steps 9–13 using the TIFF images from your second slide (see Note 21). Note, the same GAL file should be used to extract data from the second slide since slides provided as a pair always have the same lot number. Examples of autoantibody profiling of fractionated lysate microarrays are shown in Figs. 1 and 2. Figure 1 shows a scan

Native Antigen Fractionation Protein Microarrays

143

Fig. 1. An example of results obtained from a fractionated lysate microarray experiment where the dye-swap method was employed. On the left is a composite image of the “Mix 1” scans, with the IC/PBS IgG labeled with Cy™3 (green) and the control IgG labeled with Cy™5 (red ). Shown in comparison (on the right ) is an image of a fractionated lysate microarray incubated with unlabeled IC/PBS IgG; this slide was visualized using the APiX chromogenic detection system.

Fig. 2. Results obtained from a fractionated cancer lysate microarray where the dye-swap method was used. Autoantibody profile of an ovarian cancer patient was compared with an age-matched normal healthy control. Antibodies were purified from the sera and used as probes against a cancer cell lysate generated from a cell line (also can be generated from autologous tumor) that was fractionated and spotted on the array as outlined in this chapter. In this “Mix 1” scan, autoantibodies from an ovarian cancer patient were labeled with Cy™3 (green) and autoantibodies from a control sample were labeled with Cy™5 (red ). This figure shows a slide with 12 subarrays of 8 × 12 spots, with each spot representing one of the 960 fractions that were separated by our 2D fractionation protocol. In this slide, each fraction is spotted once (can be spotted in duplicates or triplicates depending on the user). Each fraction (thus each spot) contains approximately 3–5 proteins per fraction. As shown in this figure, there are 57 fractions containing antigens recognized by ovarian sera alone (green), 50 fractions that were recognized in common between cancer and control IgG (yellow ), and only two fractions that were recognized by the control patient sample (red ). The red dots at the corner of each subarray are Cy™5-labeled streptavidin orientation markers and do not represent any autoantibody reactivity.

144

Caiazzo et al.

image from a fractionated lysate microarray experiment where the dye-swap method was employed. In this experiment, test IgG was purified from the serum of an IC/PBS patient, while control IgG was purified from the sera of an age-matched normal female; IgGs were differentially labeled with fluorescent dyes. The spots that appear green represent Cy™3 reactivity, and the red spots represent the Cy™5 reactivity. Shown in comparison is an image of a fractionated lysate microarray visualized using the APiX chromogenic detection system. Figure 2 shows a scan image from a cancer fractionated lysate microarray experiment with dye-swap on an ovarian cancer sample. In this experiment, IgG was purified from a patient with ovarian cancer, while control IgG was purified from an age-matched normal female control. IgGs were differentially labeled with fluorescent dyes. 3.3.2. Biostatistical Data Analysis

Biostatistical analyses are needed to appropriately normalize, analyze, and interpret the vast amounts of data obtained from microarray studies. There are many analysis tools to extract reliable information from microarray data (commercial software packages as well as programs provided on the web at no cost to investigators). Regardless of the tools that are employed, the fractionated lysate array data must be normalized and transformed before reasonable data comparisons can be made (see Note 22). This protocol will examine analysis using the bioinformatics software package Acuity 4.0. 1. Import each of the microarrays slides (GPR files) into Acuity by clicking the “Import Microarrays” link on the Common Tasks menu located on the left hand side of the screen. 2. Once all of the microarrays have been imported, the microarrays can be normalized using the “Normalization Wizard” link on the Common Tasks menu. Select all of your microarrays located in the Project Tree menu and open the Normalization Wizard. Here researchers can normalize their data using a variety of methods, though we recommend performing a ratio-based normalization. 3. Create a data set from the normalized microarrays by selecting them in the Project Tree menu and clicking on the “Create and Open Dataset” link on the Common Tasks menu. Once a dataset has been created, researchers can manipulate their data in a number of ways, including sorting columns, removing rows, removing columns, and finding specific values. Additionally, researchers can now perform advanced statistical testing and clustering to further examine their data (see Note 23).

Native Antigen Fractionation Protein Microarrays

145

4. Notes 1. While this protocol specifically uses the T24 human bladder cancer cell line as the source of cellular lysate for protein fractionation and microarray construction, researchers should feel free to use alternate cell lines or tissue samples as the source of their cellular lysate. 2. The incubation buffer is simply the wash buffer (1× PBS + 0.05% Tween-20) with 1% BSA added. While the wash buffer can be made beforehand and stored at room temperature on the bench, we recommend making the incubation buffer fresh the day of the microarray experiments. 3. We use the rectangular quadriPerm cell culture containers as incubation trays. These containers are split into four equally sized chambers and are large enough to hold standard sized microscope slides. Each of the chambers is roughly 3.5 in. long, 1.25 in. wide, and 0.5 in. deep; these dimensions allow the microarray slides to be incubated in 5 mL of buffer. The quadriPerm containers are available from Sigma-Aldrich, though researchers should feel free to use any container of similar dimensions for the slide incubations. 4. At this point, we froze and shipped the cell pellet on dry ice to another facility. It may be necessary for researchers to send their cell or tissue samples to outside facilities, such as Eprogen, for protein fractionation. 5. For ProteoSep Start Buffer (SB) preparation, please look at Subheading 3.1.2, step 1(a). 6. If needed, IDA is available from Sigma-Aldrich. 7. Fractions from the first dimension chromatofocusing should be collected using the 2.0 mL 96-well plates. 8. Extra fractions can be stored at −80°C in the 96-well plates following fractionation for downstream MS analysis. 9. Positive controls (1–100 mg/mL Human IgG) and negative controls (print buffer only) were also printed along with the second dimension fractions. 10. The IgG purification procedure is repeated in order to ensure exclusive isolation of IgG from patient sera. We have found that with only one purification run, high-abundance proteins other than IgG may be present in the eluted solution. 11. In order to label the purified antibodies with fluorescent dyes, the antibodies must be in a buffer free of primary amines; the use of buffers that contain primary amino groups, such as

146

Caiazzo et al.

TRIS and glycine, will inhibit the labeling reaction. The Melon™ Gel Purification Buffer used in Subheading 3.2.1 is compatible with all of the Cy™ labeling reagents used in Subheading 3.2.2. 12. GE Healthcare supplies the Cy™3 and Cy™5 dyes as monofunctional N-hydroxysuccinimide (NHS)-esters in dried premeasured amounts. The NHS-ester is a functional group that cross-links to primary amines. This reaction releases the NHS group and produces a covalent amide bond that links the dye to the amine. Due to the unique affinity ratio of each dye to the antibody that it is labeling, different volumes of the two dyes are used in the labeling reactions. We recommend that researchers conduct an estimation of the final dye/ protein ratio on their own according to the manufacturer’s instructions. 13. The Protein Desalting Spin Columns are used to remove free dye. These columns contain a desalting resin and molecular weight cutoff. They perform well in desalting small sample volumes (30–120 mL), providing excellent protein recovery and ³95% retention of small molecules and salts (<7 kD). 14. We perform a competitive hybridization, which requires the use of two fluorescent dyes and a dye swap procedure. If researchers are interested in performing the microarray experiments using a single labeled sample (no competition) that is perfectly acceptable. We have also performed these slides with unlabeled IgG followed by subsequent incubations with biotin labeled anti-human IgG and streptavidinDyLight™649 (Cy5 equivalent). Additionally, we have also used the APiX chromogenic detection system from Gentel Biosciences, Inc.; for more information on the scientific background of this detection method, please view the following references (10, 11). 15. While we discuss the use of GenePix Pro 6.0 and Acuity 4.0, other data acquisition and analysis programs or methods may be better suited for use based on one’s experimental design. We highly recommend talking with biostatistical and bioinformatics experts to help researchers develop strong experimental designs and to aid in downstream data acquisition and analysis. 16. There are various software packages available for creating scan and quantitation protocols. ScanArray Express by Packard Bioscience or GenePix Pro 6.0 by Molecular Devices is recommended for the scanning of the microarray slides. GenePix Pro is recommended for the quantitation of your results. 17. The microarray may need to be scanned several times before the ideal laser power and PMT settings are determined.

Native Antigen Fractionation Protein Microarrays

147

If your scanner has the option of running an Automatic Sensitivity Calibration, this option may be used. 18. If the signal is low or there is a lot of background fluorescence, the grid may not align with the spots properly. If this is the case, you will need to manually align the grid with the image. 19. The tab labeled “Histogram” at the top of the screen allows you to check the normalization of your data. When dye and scanning intensities are perfectly normalized, the Count Ratio will be equal to 1.0. If your Count Ratio deviates significantly from 1.0, you should attempt to adjust the PMT gains used for scanning so that this value approximates 1.0. The “Scatter Plot” tab allows you to view the distribution of your data to ensure that there are a minimal number of outlying data points. 20. GPR files are used in some biostatistical analyses. For example, the microarray data analysis software, Acuity 4.0 by Molecular Devices, uses this file format in order to perform data manipulation and analyses, including Ratio-based normalization, LOWESS normalization, hierarchical clustering, nonhierarchical clustering, and t-tests for statistical significance. 21. When opening the TIFF files from the second slide, it is necessary to check the box next to “Replace current images” in the “Open Images” dialogue box. 22. Normalization and transformation of data accomplish two aims. First, log-transformation of data scales the range of your data; since the dye swap data is acquired in ratio form, it is important that the scale allows for equal differential expression by the sample represented in both the numerator and the denominator of the ratio. Second, normalization centers data around an expected mean value; log-transformed data is centered around the expected median (or mean) log-ratio of 0.0 across all data points. 23. Once researchers have identified fractions of interest from their data, proteins within the fractions can be identified through MS analysis of fractions saved after the fractionation process. Sample preparation of the fractions will depend on the type of MS analysis being performed.

Acknowledgments This work was supported in part by grants DK063665, DK066020, DK075566 from the National Institutes of Health to B.C.S. Liu. Additional funding was supported by the Interstitial Cystitis Association and the Fishbein Family Foundation.

148

Caiazzo et al.

Disclosure: Timothy J. Barder, Ph.D. is the President and owner of Eprogen. Bryce P. Nelson, Ph.D. is an employee and shareholder of Gentel Biosciences, Inc. References 1. VanMeter A, Signore M, Pierobon M, Espina V, Liotta LA, Petricoin EF 3rd (2007) Reverse-phase protein microarrays: application to biomarker discovery and translational medicine. Expert Rev Mol Diagn 7:625–633 2. Charboneau L, Tory H, Chen T, Winters M, Petricoin EF 3rd, Liotta LA, Paweletz CP (2002) Utility of reverse phase protein arrays: applications to signaling pathways and human body arrays. Brief Funct Genomic Proteomic 1:305–315 3. Paweletz CP, Charboneau L, Bichsel VE, Simone NL, Chen T, Gillespie JW, EmmertBuck MR, Roth MJ, Petricoin IE, Liotta LA (2001) Reverse phase protein microarrays which capture disease progression show activation of pro-survival pathways at the cancer invasion front. Oncogene 20:1981–1989 4. Janzi M, Odling J, Pan-Hammarstrom Q, Sundberg M, Lundeberg J, Uhlen M, Hammarstrom L, Nilsson P (2005) Serum microarrays for large scale screening of protein levels. Mol Cell Proteomics 4:1942–1947 5. Nishizuka S, Charboneau L, Young L, Major S, Reinhold WC, Waltham M, Kouros-Mehr H, Bussey KJ, Lee JK, Espina V, Munson PJ, Petricoin E 3rd, Liotta LA, Weinstein JN (2003) Proteomic profiling of the NC1-60 cancer cell lines using new high density reverse-phase lysate microarrays. Proc Natl Acad Sci USA 100:14229–14234 6. Taylor BS, Pal M, Yu J, Laxman B, KalyanaSundaram S, Zhao R, Menon A, Wei JT,

7.

8.

9.

10.

11.

Nesvizhskii AI, Ghosh D, Omenn GS, Lubman DM, Chinnaiyan AM, Sreekumar A (2008) Humoral response profiling reveals pathways to prostate cancer progression. Mol Cell Proteomics 7:600–611 Bouwman K, Qiu J, Zhou H, Schotanus M, Mangold LA, Vogt R, Erlandson E, Trenkle J, Partin AW, Misek D, Omenn GS, Haab BB, Hanash S (2003) Microarrays of tumor cell derived proteins uncover a distinct pattern of prostate cancer serum immunoreactivity. Proteomics 3:2200–2207 Qiu J, Madoz-Gurpide J, Misek DE, Kuick R, Brenner DE, Michailidis G, Haab BB, Omenn GS, Hanash S (2004) Development of natural protein microarrays for the diagnosing cancer based on an antibody response to tumor antigens. J Proteome Res 3:261–267 Mattoon D, Michaud G, Merkel J, Schweitzer B (2005) Biomarker discovery using microarray technology platforms: antibody-antigen complex profiling. Expert Rev Proteomics 2:879–889 Sia SK, Linder V, Parviz BA, Siegel A, Whitesides GM (2004) An integrated approach to a portable and low-cost immunoasay for resource-poor settings. Angew Chem Int Ed Engl 43:498–502 Huber M, Wei TF, Muller UR, Lefebvre PA, Marla SS, Bao YP (2004) Gold nanoparticle probe-based gene expression analysis with unamplified total human RNA. Nucleic Acids Res 32:e137

Chapter 10 Immunoprofiling Using NAPPA Protein Microarrays Sahar Sibani and Joshua LaBaer Abstract Protein microarrays provide an efficient method to immunoprofile patients in an effort to rapidly identify disease immunosignatures. The validity of using autoantibodies in diagnosis has been demonstrated in type 1 diabetes, rheumatoid arthritis, and systemic lupus, and is now being strongly considered in cancer. Several types of protein microarrays exist including antibody and antigen arrays. In this chapter, we describe the immunoprofiling application for one type of antigen array called NAPPA (nucleic acids programmable protein array). We provide a guideline for setting up the screening study and designing protein arrays to maximize the likelihood of obtaining quality data. Key words: NAPPA protein microarray, Immunoprofile, Immunosignature, Autoantibody, Breast cancer, Diabetes, Autoimmune, Proteomics, Serum screening, Antigen

1. Introduction Protein microarrays are powerful in their ability to test hundreds to thousands of proteins simultaneously and in parallel in a miniaturized format. Most protein microarrays fall grossly into two categories: antibody and antigen arrays. Antibody arrays, in which numerous antibodies (or other affinity reagents) are printed on a slide, were first developed by Haab et al. (1) and later utilized by Sreekumar et al. (2) to demonstrate feasibility in detecting cancer antigens in a complex cell lysate. Antigen arrays involve the display of proteins on the microarray and can be used in the identification of serological autoantibodies present in patients, but not controls. Their feasibility was first demonstrated by Joos et al. (3) for the Ro and La autoantigens for Sjogren’s syndrome. Recently, several developments in the use of antigen microarrays in immunoprofiling of patients to identify disease signatures were reported (4–10). One approach used in cancer studies is to Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_10, © Springer Science+Business Media, LLC 2011

149

150

Sibani and LaBaer

identify autoantibodies targeting self-proteins present in cancer patients but not controls (11–21). Antigen microarrays are ideal for this purpose as they provide a set of target antigens to which the autoantibodies can bind. Traditionally, antigen arrays were made by individually purifying proteins and printing them on the microarray, a long, tedious and expensive endeavor. NAPPA (nucleic acids programmable protein array) microarrays offer a platform in which proteins are made from printed cDNA-containing plasmids to produce a just-in-time, fresh protein microarray that is incubated with human serum to detect autoantibody binding (Fig. 1) (22–24). A secondary labeled antihuman antibody is added to visualize the autoantibodies bound to autoantigens. NAPPA microarrays have been successfully used in the detection of autoantibodies to p53 in breast and ovarian cancer, ML-IAP in melanoma patients, BCL2 in prostate cancer, and GAD65 and IA2 in type 1 diabetes patients (25, 26). NAPPA microarrays were shown to express over 94% of their proteins regardless of protein size or type (25). Their testing in the detection of the p53 autoantibody in breast cancer serum provided a CV of 7% for within day testing and 11% day-to-day testing (25). The remainder of this chapter details study design using NAPPA microarrays. Although the production of the microarrays is beyond the scope of this chapter, a detailed protocol was published in 2008 (24).

Fig. 1. Schematic representation of a NAPPA protein microarray. (a) NAPPA microarrays are made by printing cDNA-harboring plasmids on glass slides that are later transcribed and translated in situ to create a fresh protein microarray (top panel ). When human serum is incubated with the microarray, autoantibodies bind their target antigens and are identified using a labeled secondary antibody (lower panel ). (b) An image of an expressed NAPPA microarray. For each experiment, it is highly recommended that the expression efficiency of the microarray is assessed at the onset.

Immunoprofiling Using NAPPA Protein Microarrays

151

1.1. Study Design Considerations

In designing serum screening studies, several study design issues should be addressed prior to initiation in order to generate interpretable and meaningful results. These include the following.

1.1.1. Serum Samples

In obtaining serum samples for the study, the following need to be considered: 1. Serum samples from the patient and control groups are best when matched for gender and age to avoid confounding factors arising from differences in these variables. 2. Samples may also require pairing based on other variables depending on the experiment (e.g., same smoking status when studying lung cancer) 3. Researchers should ensure that the sera used were collected around the same period of time using the same method. This avoids the confounding effects of different serum collection SOPs that may be used at different institutions, or even different divisions of the same institution (e.g., outpatient clinic vs. hospital phlebotomy staff). 4. Study coordinators need to obtain IRB approval and informed consent from the patients and controls to be included in the study.

1.1.2. Experimental Setup

Large serum screening studies face several challenges (see Note 1) that may be best handled by dividing the project into three stages: 1. Pilot screen – A pilot screen of 20–50 patient and control samples against a test microarray is carried out to test signal variation among patients and controls. These data can then be used for a statistical power study that will identify the number of samples required for an expected frequency of an autoantibody in the population, the variability of the signal for autoantibody in the population, and the degree of certainty desired to ensure that a marker will be found if it is there. 2. Training stage – Patient and control sera are screened against the entire panel of proteins to identify antigens that can discriminate between patients and controls. The number of sera screened depends on the results of the power study conducted during the pilot screen. The sera used during the training stage can overlap those used during the pilot screen. Proteins that pass this step will form the validation set. 3. Validation stage – The aim here is to validate the antigens identified during the training set using an independent set of sera. The number of sera used here is also dictated by the power calculation from the pilot screen. Key here is that none of the samples used for validation (cases or controls) was used in any of the previous studies. The ideal validation study is

152

Sibani and LaBaer

performed in a blinded fashion. For antigens confirmed by validation, it is useful to test them by an alternate method. One method would be ELISA, commonly used in clinical laboratories, which also serves the purpose of facilitating adoption of the biomarkers in clinical tests. 1.2. Protein Microarray Considerations

The structure and content of the protein microarrays also need to be considered to ensure good quality data and interpretable results. Items to be contemplated should include the following.

1.2.1. Spot Replicates

These replicates are used to measure two different variations valuable during statistical analysis: 1. Zone variation: This type of variation, which can occur with virtually all microarray technologies, results from a microarray printing or processing method that causes one or more region(s) of the array to erroneously display different signal magnitude(s). Examples are illustrated in Fig. 2 where

Fig. 2. Examples of zone variations that can occur on protein microarrays. (a) The zone variation here shows a strong signals at the top of the slide and weaker signals at the bottom. (b) In this case, the top-to-bottom zone variation is additionally confounded with a regional oval-shaped variation observed in the lower third of the slide. Practice of microarray processing technique will alleviate these variations over time.

Immunoprofiling Using NAPPA Protein Microarrays

153

overall spots at the top of the slide are more intense than at the bottom. Such variation can be adjusted for by two different methods: (a) Printing of identical features from the same protein sample throughout the microarray as references to monitor zone variations. In the case of NAPPA arrays, this was accomplished with a grid of eight features across the width of the slide by 12 spots along its length. These 48 features typically consist of negative controls discussed below. (b) Averages of regional or neighboring features can be used in place of the identical features method described above. In this case, we assume that the majority of proteins do not show reactivity to serological autoantibodies and display similar levels of nonspecific binding. Hence averaging 16 proteins in a 4 × 4 grid, for example, throughout the microarray would identify regions of varying background signals that demonstrate zone variation. 2. Printing variability: Variation in printing efficiency and spotting chemistry contributes to this type of variability that can be monitored by printing multiple identical features. These features are best placed within a close vicinity to each other to avoid the confounding effects of zone variation. 1.2.2. Controls

Controls are of the utmost importance in monitoring proper microarray processing and technical and biological variability. There are three types of controls that should be included on the arrays: 1. Processing positive controls. In order to ensure that the arrays are working appropriately, various positive controls should be included on the arrays. To confirm that the antihuman secondary antibodies are working and to provide reference features, human IgG can be included. It is also useful to include a protein that is likely to reveal a response in most individuals, regardless of whether they are patients or controls. Examples of such proteins include the EBNA1 antigen, from the Epstein Barr virus to which approximately 90% of the adult population have antibodies, or childhood vaccines such as tetanus toxoid. 2. Negative controls. These are used to determine background or noise levels on the microarrays during the data analysis. They should be distributed throughout the microarray and are used to detect and adjust for zone variations. 3. Disease-specific controls. Whenever possible, it is best to include positive controls for a disease to test the viability of the serum screening conditions. It should be noted, though,

154

Sibani and LaBaer

that not all diseases have known controls and not all patients will be reactive to such controls, hence their availability and usefulness may be limited. 1.2.3. Technical Reproducibility Test

As with all large screening experiments that are carried out over the course of weeks or months, the degree of technical reproducibility needs to be assessed to ensure that the differences observed between test groups are real. Here are the forms of technical reproducibility that should be considered: 1. Within Day reproducibility: This tests the microarray-tomicroarray variability within one processing run. It is measured by testing each of three or four serum samples on two or three identical microarrays. It is best not to proceed to a full scale screen until the coefficient of variation of such tests is less than 10% for 80% of the features interrogated. Otherwise, the microarray processing protocol needs to be reoptimized. 2. Day-to-day reproducibility: This measures the microarray-tomicroarray variability between tests, each run on a different day. Since most large scale screening studies are processed over the course of weeks, the daily reproducibility needs to be addressed and the variability minimized. One method to minimize the likelihood of obtaining nonspecific variations between patients and controls is to process the same number of patients and controls daily (such as five patients and five controls every day).

2. Materials 2.1. Activation of cDNA-Based Microarrays

1. NAPPA microarrays (see Note 2). 2. HybriWell gaskets (Grace). 3. TNT ® T7 Quick Coupled Transcription/Translation System (Promega). 4. RNaseOUT (Invitrogen). 5. DEPC water (Ambion). 6. EchoTherm™ IN30 Bench Top, Chilling/Heating Programmable Incubator (Torrey Pines Scientific). 7. SuperBlock (Pierce). 8. Phosphate buffered saline (1× PBS): 137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM KH2PO4. Adjust pH to 7.4 with HCl if necessary. 9. 5% milk blotto: Dissolve 5 g of nonfat dry milk in 1 l of 1× PBS. Add Tween-20 to final concentration of 0.2% (see Note 3).

Immunoprofiling Using NAPPA Protein Microarrays

2.2. Detection of Protein Display on the Microarrays

155

1. Corning ® Hybridization Chamber. 2. Mouse anti-GST antibody (Cell Signaling). 3. Antimouse HRP-conjugated antibody (Jackson Laboratories). 4. TSA (tyramide signal amplification) reagent (Perkin Elmer). 5. Lifter slips, 24 × 65 mm (Erie).

2.3. Serum Antibody Profiling

1. 5% milk blotto: Dissolve 5 g of nonfat dry milk in 1 l of 1× PBS. Add Tween-20 to final concentration of 0.2% (see Note 3). 2. Corning ® Hybridization Chamber (Product). 3. Mouse antihuman IgG HRP-conjugated antibody (Jackson ImmunoResearch). 4. TSA reagent (Perkin Elmer). 5. Lifter slips, 24 × 65 mm (Erie). 6. ProScan Array Scanner (Perkin Elmer).

3. Methods Serological autoantibody screening using protein microarrays provides a rapid and efficient method to profile an individual’s humoral immune response to known or unclassified antigens. Loosely based on the broadly utilized ELISA assay, this method of serum screening requires specific optimization to microarrays to avoid artifacts and technical variations that would lead to false data. There are different types of microarrays, and each possess unique advantages and challenges. The remainder of this chapter will focus on a specific type of cDNA-based protein microarrays called NAPPA (Nucleic Acid Programmable Protein Array). NAPPA arrays are built by printing a plasmid containing the cDNA of a protein tagged with GST, along with an anti-GST capture antibody. The arrays are converted to functional protein microarrays through in situ protein production and capture using an in vitro expression system. They are then treated like any other protein microarray with attention to avoiding protein degradation and maintaining stability. Serum samples from patients or controls are diluted in a milk-based buffer and added to the microarrays in an overnight incubation at 4°C. Such long incubations allow low abundance and/or weak affinity autoantibodies adequate time to bind their target proteins on the microarray. The microarrays are then washed, incubated with a secondary antibody, and then visualized for subsequent quantification and analysis. The study size (i.e., number of serum samples) required to complete a comprehensive and meaningful investigation depends

156

Sibani and LaBaer

on numerous factors including the frequency with which any one autoantibody is expected within the patient population, the estimated clinical specificity of such an antibody in predicting disease state, the relative affinity of the autoantibodies to their targets, the antigen density on the microarrays, and the technical reproducibility of the microarrays. It is imperative for researchers to work with a biostatistician prior to starting the screen in order to determine the study size and number of replicates that will be needed to obtain statistically significant data. 3.1. Activation of NAPPA Microarrays

1. Microarrays on which the proteins will be synthesized in situ from a cDNA require activation which will lead to the synthesis and capture of each protein. On NAPPA microarrays, the microarrays are first immersed in Superblock for 30–60 min to block any nonspecific protein binding sites that may exist on the microarray. These would include nonspecific binding to the cDNA-containing spots as well as the glass surface on which the microarrays are printed. At the end of this incubation, the slides are rinsed with dH2O and dried using filtered, pressurized air. 2. Promega’s T7-based coupled transcription-translation rabbit reticulocyte lysate expression system is prepared according to the manufacturer’s instruction with the exception of the addition of both of the provided amino acid mixes so as to obtain a full complement of amino acids. For examples, the amino acid mix lacking methionine and that lacking leucine are combined together. Each microarray will consume 150 ml of the lysate. 3. A hybriwell gasket is applied to the slide with the outline sealed by the adhesive material, thus forming a thin chamber above the slides. This chamber is filled with rabbit reticulocyte lysate and incubated at 30°C for 90 min in a programmable incubator. Each spot will synthesize its target protein that gets captured through its tag by an anti-tag antibody present in the spot. To ensure proper immobilization of these proteins, the slides are incubated at 15°C for at least 30 min following the expression protocol. It is absolutely essential for this step to be completed in order to capture the greatest amount of protein on the spot and to get the highest density possible. 4. At the end of the expression and capture incubations, the hybriwell is gently removed without disrupting the proteins on the slide (see Note 4) and immersed into a 5% milk blotto solution for three 5 min washes, followed by an hour’s incubation at room temperature.

3.2. Detection of Protein Display on the Microarrays

1. For each daily experiment, at least one array should be processed to measure the amount of protein displayed (i.e. expressed and captured) per spot. This will be used for

Immunoprofiling Using NAPPA Protein Microarrays

157

future analysis (see Note 5). Each slide is incubated in 2 ml of anti-GST antibody diluted 300-fold in 5% blotto. The slides are placed in a Corning hybridization chamber and an end-over-end rotator and allowed to incubate for 1 h at room temperature (see Note 6). 2. Slides are removed from the Corning hybridization chamber and washed three 5 min washes with 5% milk blotto, followed by another incubation with the HRP-conjugated secondary antimouse antibody diluted 500-fold in 5% milk blotto. The secondary is incubated for 1 h with end-over-end mixing in the Corning hybridization chamber. 3. Slides are removed from the secondary antibody and washed three 5 min washes in 1× PBS. They are rinsed with distilled water, and 500 ml of a 50-fold dilution of Cy3 TSA is added to each slide. A lifter slip is applied to spread the TSA, and the slides are incubated for 10 min in the dark. 4. The lifter slip is removed, and the slide is rinsed with dH2O, dried using filtered pressurized air, placed in a slide box, and stored in a dark dry place until ready to scan. 3.3. Serum Antibody Profiling

1. Antibody profiling is carried out at the end of step 4 of Subheading 3.1, when the microarrays have been expressed and blocked. The serum (or plasma) is centrifuged at 14,000 rpm for 10 min at 4°C in a microcentrifuge to separate out any leftover lipids and cellular debris (see Note 7). 2. Serum is diluted 200–900-fold into 2 ml 5% milk/0.3× PBS-T buffer (see Note 8) and applied to the microarray in a Corning hybridization chamber. The slides are placed in an end-over-end rotator and incubated overnight at 4°C to allow low abundance and/or weak affinity antibodies to bind their target proteins. 3. The next morning, the slides are taken out of the Corning hybridization chamber and washed 3 times with 5% milk/ blotto for 5 min each. 2 ml of HRP-conjugated antihuman IgG antibody, diluted 500-fold in 5% milk/blotto, is applied to the slides in a Corning hybridization chamber. The slides are incubated for 1 h at room temperature with end-over-end rotation. 4. The slides are removed from the slide chamber and washed 3 times in 1× PBS, each wash of 5 min. The slides are rinsed with dH2O and 500 ml of 50-fold diluted TSA reagent is applied to each slide. A lifter slip is placed on top to spread the TSA across the entire slide and incubated for 10 min at room temperature in the dark. 5. The lifter slip is removed and the slides are rinsed with dH2O, dried with filtered pressurized air, and stored in a slide box in a dark dry place until ready to scan.

158

Sibani and LaBaer

Fig. 3. Examples of strong antigen signals for EBNA (solid arrow ) and p53 (dashed arrow ) in a breast cancer patient (a) and GAD65 (solid arrows) in a type 1 diabetes patient (b).

6. Slides are scanned using the red laser for Cy3 at an intensity that will not cause signal saturation (see Note 9). Quantify the spots using Microvigene and analyze by advanced statistical methods such as the binomial proportion, Wilcoxon, or Fisher tests. An example of strongly reactive antigens p53, EBNA, and GAD65 are shown in Fig. 3.

4. Notes 1. Screening many thousands of proteins against hundreds of patient samples may be difficult to manage physically, analytically, and economically. Challenges include the risk of overfitting the data, high false discovery rates, and dismissal of potentially legitimate biomarkers. To avoid such problems and decrease the financial costs associated with large serum screening studies, a subset of 50–100 sera can be screened

Immunoprofiling Using NAPPA Protein Microarrays

159

against the entire panel of protein antigens with the goal of removing any proteins that do not show a signal higher than background or do not show a difference between patients and controls. In this strategy, the hope is to reduce the number of antigens about tenfold in order to have a more focused set of proteins to begin the training studies. 2. NAPPA microarrays are custom designed to contain up to 2,532 features/spots of which up to 2,304 spots are proteinencoding features. They are DNA arrays that are converted into proteins arrays on the day of use, thus displaying fresh protein for each experiment. 3. The milk needs to be stirred for over an hour to ensure that it is completely dissolved. Any particulate matter that is present in the milk can adhere to the glass slide leading to either masking of real protein spots or false signals. 4. Researchers need to be careful not to place any pressure on the middle of the slide when removing the hybriwell so as not to smear proteins away from their designated spots. 5. Measurements of protein display are used to determine the success of the expression and capture on the array. Typically, over 90% of the features display their target protein. 6. When the Corning hybridization chamber is closed, the antibody/milk must cover at least half the slide to avoid drying in the middle of the slide. 7. In our experience, the source of antibody, whether plasma or serum, did not impact the efficiency of antibody binding to the antigens presented on the arrays. 8. Serum samples from different individuals display different response magnitudes. We have observed that approximately 10–15% of the sera have a high nonspecific background regardless of their patient or control designation. Moreover, for such large scale studies, samples were typically collected on different days by different persons introducing variation in serum reactivity solely due to the collection method. To adjust for both of these issues, we titer each serum prior to screening it by testing it on mini-arrays of 36 protein spots to identify an optimal dilution factor that provides an acceptable background that will not overwhelm the true signals. This is combined with the use of lower PBS concentrations to promote the binding of low affinity autoantibodies. We have found that in the majority of cases, a 600-fold dilution into 5% milk/0.3× PBS-T worked best. 9. It is important to avoid saturation of the signal in order to obtain a valid quantifiable result. Most scanners provide a method to monitor signal intensity, including a warning for signal saturation.

160

Sibani and LaBaer

Acknowledgments The authors would like to acknowledge support for this work by the NIH “Biomarker Detection Using NAPPA Tumor Antigen Arrays” U01 CA117374 and from the Juvenile Diabetes Research Foundation, “The use of protein microarrays to study autoimmunity and diabetes” 17-2007-1045. References 1. Haab BB, Dunham MJ, Brown PO (2001) Protein microarrays for highly parallel detection and quantitation of specific proteins and antibodies in complex solutions. Genome Biol 2:1–13 2. Sreekumar A, Nyati MK, Varambally S, Barrette TR, Ghosh D, Lawrence TS, Chinnaiyan AM (2001) Profiling of cancer cells using protein microarrays: discovery of novel radiation-regulated proteins. Cancer Res 61:7585–7593 3. Joos TO, Schrenk M, Hopfl P, Kroger K, Chowdhury U, Stoll D, Schorner D, Durr M, Herick K, Rupp S, Sohn K, Hammerle H (2000) A microarray enzyme-linked immunosorbent assay for autoimmune diagnostics. Electrophoresis 21:2641–2650 4. Tan HT, Low J, Lim SG, Chung MC (2009) Serum autoantibodies as biomarkers for early cancer detection. Febs J 276(23):6880–6904 5. Yu X, Schneiderhan-Marra N, Hsu HY, Bachmann J, Joos TO (2009) Protein microarrays: effective tools for the study of inflammatory diseases. Methods Mol Biol 577: 199–214 6. Song Q, Liu G, Hu S, Zhang Y, Tao Y, Han Y, Zeng H, Huang W, Li F, Chen P, Zhu J, Hu C, Zhang S, Li Y, Zhu H, Wu L (2009) Novel autoimmune hepatitis-specific auto antigens identified using protein microarray technology. J Proteome Res 9(1):30–39 7. Lorenz P, Kreutzer M, Zerweck J, Schutkowski M, Thiesen HJ (2009) Probing the epitope signatures of IgG antibodies in human serum from patients with autoimmune disease. Methods Mol Biol 524:247–258 8. Quintana FJ, Farez MF, Viglietta V, Iglesias AH, Merbl Y, Izquierdo G, Lucas M, Basso AS, Khoury SJ, Lucchinetti CF, Cohen IR, Weiner HL (2008) Antigen microarrays identify unique serum autoantibody signatures in clinical and pathologic subtypes of multiple sclerosis. Proc Natl Acad Sci U S A 105: 18889–18894

9. Auger I, Balandraud N, Rak J, Lambert N, Martin M, Roudier J (2009) New autoantigens in rheumatoid arthritis (RA): screening 8268 protein arrays with sera from patients with RA. Ann Rheum Dis 68:591–594 10. Roche S, Dauvilliers Y, Tiers L, Couderc C, Piva MT, Provansal M, Gabelle A, Lehmann S (2008) Autoantibody profiling on high-density protein microarrays for biomarker discovery in the cerebrospinal fluid. J Immunol Methods 338:75–78 11. Qiu J, Hanash S (2009) Autoantibody profiling for cancer detection. Clin Lab Med 29:31–46 12. Kijanka G, Murphy D (2009) Protein arrays as tools for serum autoantibody marker discovery in cancer. J Proteomics 72:936–944 13. Liu W, Wang P, Li Z, Xu W, Dai L, Wang K, Zhang J (2009) Evaluation of tumour-associated antigen (TAA) miniarray in immunodiagnosis of colon cancer. Scand J Immunol 69:57–63 14. Gure AO, Altorki NK, Stockert E, Scanlan MJ, Old LJ, Chen YT (1998) Human lung cancer antigens recognized by autologous antibodies: definition of a novel cDNA derived from the tumor suppressor gene locus on chromosome 3p21.3. Cancer Res 58:1034–1041 15. Gure AO, Stockert E, Scanlan MJ, Keresztes RS, Jager D, Altorki NK, Old LJ, Chen YT (2000) Serological identification of embryonic neural proteins as highly immunogenic tumor antigens in small cell lung cancer. Proc Natl Acad Sci U S A 97:4198–4203 16. Scanlan MJ, Chen YT, Williamson B, Gure AO, Stockert E, Gordan JD, Tureci O, Sahin U, Pfreundschuh M, Old LJ (1998) Characterization of human colon cancer antigens recognized by autologous antibodies. Int J Cancer 76:652–658 17. Jager D, Stockert E, Gure AO, Scanlan MJ, Karbach J, Jager E, Knuth A, Old LJ, Chen YT (2001) Identification of a tissue-specific

Immunoprofiling Using NAPPA Protein Microarrays putative transcription factor in breast tissue by serological screening of a breast cancer library. Cancer Res 61:2055–2061 18. Forti S, Scanlan MJ, Invernizzi A, Castiglioni F, Pupa S, Agresti R, Fontanelli R, Morelli D, Old LJ, Pupa SM, Menard S (2002) Identification of breast cancer-restricted antigens by antibody screening of SKBR3 cDNA library using a preselected patient’s serum. Breast Cancer Res Treat 73:245–256 19. Jager D, Unkelbach M, Frei C, Bert F, Scanlan MJ, Jager E, Old LJ, Chen YT, Knuth A (2002) Identification of tumor-restricted antigens NY-BR-1, SCP-1, and a new cancer/ testis-like antigen NW-BR-3 by serological screening of a testicular library with breast cancer serum. Cancer Immun 2:5 20. Kilic A, Schuchert MJ, Luketich JD, Landreneau RJ, Lokshin AE, Bigbee WL, El-Hefnawy T (2008) Use of novel autoantibody and cancer-related protein arrays for the detection of esophageal adenocarcinoma in serum. J Thorac Cardiovasc Surg 136: 199–204 21. Taylor BS, Pal M, Yu J, Laxman B, KalyanaSundaram S, Zhao R, Menon A, Wei JT, Nesvizhskii AI, Ghosh D, Omenn GS, Lubman DM, Chinnaiyan AM, Sreekumar A (2008) Humoral response profiling reveals

22.

23.

24.

25.

26.

161

pathways to prostate cancer progression. Mol Cell Proteomics 7:600–611 Ramachandran N, Hainsworth E, Bhullar B, Eisenstein S, Rosen B, Lau AY, Walter JC, LaBaer J (2004) Self-assembling protein microarrays. Science 305:86–90 Ramachandran N, Hainsworth E, Demirkan G, LaBaer J (2006) On-chip protein synthesis for making microarrays. Methods Mol Biol 328:1–14 Ramachandran N, Raphael JV, Hainsworth E, Demirkan G, Fuentes MG, Rolfs A, Hu Y, LaBaer J (2008) Next-generation high-density self-assembling functional protein arrays. Nat Methods 5:535–538 Ramachandran N, Anderson KS, Raphael JV, Hainsworth E, Sibani S, Montor WR, Pacek M, Wong J, Eljanne M, Sanda MG, Hu Y, Logvinenko T, LaBaer J (2008) Tracking humoral responses using self assembling protein microarrays. Proteomics Clin Appl 2:1518–1527 Anderson KS, Ramachandran N, Wong J, Raphael JV, Hainsworth E, Demirkan G, Cramer D, Aronzon D, Hodi FS, Harris L, Logvinenko T, LaBaer J (2008) Application of protein microarrays for multiplexed detection of antibodies to tumor antigens in breast cancer. J Proteome Res 7:1490–1499

wwwwwwwwwwwwwwww

Part III Protein Function Microarrays

wwwwwwwwwwwwwwww

Chapter 11 High-Throughput Mammalian Two-Hybrid Screening for Protein–Protein Interactions Using Transfected Cell Arrays (CAPPIA) Andrea Fiebitz and Dominique Vanhecke Abstract We here describe a new and cost-effective method for the high-throughput detection of protein–protein interactions in mammalian cells that combines the advantages of mammalian two-hybrid systems with those of microarrays. Nanoliters of samples containing mixtures of bait and prey expression plasmids together with an autofluorescent reporter are immobilized on glass slides in defined array formats and air-dried. Subsequently, monolayers of adherent mammalian cells are grown on these slides so that only cell clusters on top of each feature become transfected, whereas the surrounding cells remain untransfected. If the expressed proteins show any interaction, the bait and prey proteins inside the cells are functionally linked together at the promoter of the autofluorescent reporter, reconstituting transcriptional activity, and cells become fluorescent. The cluster of cells that express that particular combination of bait and prey constructs can be identified by its position in the array by simple fluorescence detection using common DNA array scanners or high-throughput microscopy. CAPPIA allows the quantitative detection of specific protein interactions in different types of mammalian cells and under the influence of different compounds. The high number of preys that can be tested per slide together with the flexibility to interrogate any bait of interest and the small amounts of reagents that are required makes this assay currently one of the most economical high-throughput detection assays for protein–protein interactions in mammalian cells. Key words: Reverse transfection, CAPPIA, Cell arrays, Mammalian two-hybrid, High-throughput, Protein–protein interaction, Hormone-dependence

1. Introduction Most biological processes require the cooperation of pairs of proteins or the formation of large functional complexes of proteins. Until now, high-throughput analyses of protein interactions were typically performed in yeast (1) and putative Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_11, © Springer Science+Business Media, LLC 2011

165

166

Fiebitz and Vanhecke

interactions were then confirmed in mammalian two-hybrid assays on a gene-by-gene scale (2, 3). In classical two-hybridbased assays, two proteins of interest are ectopically expressed as fusion proteins, one with a DNA-Binding Domain (DBD) and the other with a transcriptional Activating Domain (AD) of a transcriptional activator. If both proteins show any interaction, the DBD and AD are functionally linked together at the promoter, reconstituting transcriptional activity of the transcription factor and inducing the expression of a specific reporter (4–6). In vivo two-hybrid systems offer advantages over in vitro biophysical or biochemical methods. Indeed, some protein–protein interactions are too weak and/or transient to be detected in in vitro assays. In addition, some interactions require cell typedependent posttranslational modifications of the proteins and/or specific cofactors in the cellular microenvironment (7). High-throughput two-hybrid assays have been instrumental in the rapid screening of large numbers of protein pairs. However, the assay requires extensive liquid handling infrastructures (1, 8) making it less accessible for many research groups. In addition, the analysis of mammalian PPI within a cellular context that more closely mimics the native protein environment would be preferred in order to reduce the rate of false negative interactions, i.e., true interactions that cannot be detected due to inappropriate folding and/or posttranslational processing of the proteins. The Cell Array Protein–Protein Interaction Assay or CAPPIA described in this chapter is made for the parallel analysis of thousands of proteins for interacting partners within mammalian cells by combining cell arrays (9) with the more classical mammalian two-hybrid assay. The different steps of the assay are depicted in Fig. 1. Nanoliter volumes of different solutions, each containing a particular bait expression plasmid, a prey expression plasmid, and a reporter plasmid complexed with transfection reagent, are immobilized on glass slides in well-defined array formats. When these slides are overlaid with a monolayer of living cells, only those cells that grow on top of a particular spot of DNA will get transfected and will start to overexpress specific chimeric bait and prey proteins. If these two proteins can interact with each other, they will transactivate the autofluorescent reporter making that cluster of cells fluorescent while the surrounding cells remain negative. CAPPIA combines the advantages of mammalian twohybrid systems with those of microarrays, a combination that is expected to save considerable time and expense. Indeed, CAPPIA slides are printed with the same robotic microarray devices used to print conventional DNA microarrays. Consequently, cell arrays require far less DNA, transfection reagents, and cells as compared to assays performed in microwell plate format. The use of an autofluorescent-based reporter in

Fig. 1. Schematic overview of the CAPPIA assay. (a) Samples containing bait, prey, and reporter plasmids are prepared as described in material and methods. Typically, every sample on one array contains the same bait and the same reporter, but a different prey, shown as AD1 – ADNM. The samples are printed on glass slides with standard microarray devices used to print conventional DNA microarrays. After drying of the printed array, the DNA is transiently immobilized on the slides. Because the plasmid mixtures were treated with transfection reagent before being applied to the array, the slides are immediately ready for transfection. (b) Adherent mammalian cells are carefully seeded on top of the printed slides and are allowed to cover the whole slide. (c). After 2–3 days of culture, transfection is stopped by fixing the monolayer of cells and mounting the slide with a cover slip. (d) Fluorescent signals (green and red features) can be analyzed directly using common DNA array scanners or high-throughput microscopy. The picture shown was obtained using the BIOCCD image reader of applied biosystems. Only those cells that grow on top of a particular spot of DNA will get transfected and will start to overexpress specific chimeric bait and prey proteins. If these two proteins can interact with each other, they will transactivate the autofluorescent reporter, making that cluster of cells fluorescent (red dots) while the surrounding cells remain negative. A constitutive CMV-driven EGFP construct was printed at the border of the array to monitor transfection efficiency and to establish the coordinates of the printed array after transfection (green dots).

168

Fiebitz and Vanhecke

CAPPIA further increases the speed and cost-effectiveness of the assay. Indeed, protein–protein interaction on these slides is detected without the need for immune-fluorescent staining or enzyme-based reporter detection. Furthermore, signal detection is performed using common DNA array scanners or highthroughput microscopy. Slides can be printed in large batches and stored frozen without losing their efficiency, increasing the flexibility and cost-effectiveness of CAPPIA. Once the slides have been printed, no extensive liquid handling infrastructure is required to perform the assay, in contrast to microwell-based high-throughput assays. Suzuki et al. described a PCR-mediated rapid sample preparation and high-throughput assay system based on the mammalian two-hybrid method (10). Although their method allows for the rapid preparation of high numbers of bait and prey samples, the actual two-hybrid assays are performed in microwells and require semiautomatic multiple dispensers as well as multiple reagents for downstream enzymatic detection of interacting proteins. We showed that the resolution of detection of protein– protein interactions using CAPPIA is comparable to that of enzyme-based mammalian two-hybrid assays performed in microwells. Furthermore, CAPPIA can be used for the quantitative detection of specific protein interactions in different types of mammalian cells and under the influence of different compounds, while offering the advantage of being economical and accessible to research groups equipped with basic cell culture infrastructures. In this chapter, we use CAPPIA to demonstrate the detection of the specific and androgen-dependent interaction between the ligand-binding domain of the androgen receptor (AR-LBD) and the N-terminal domain of the same androgen receptor (AR-NTD) among 160 different prey–bait combinations of proteins or protein domains potentially associated with nuclear receptor function (Fig. 2). In addition, the high sensitivity of detection of protein–protein interactions through CAPPIA is demonstrated by the dose-dependent induction of reporter expression following AR-NTD and AR-LBD interaction in the presence of the synthetic agonist R1881 and the dose-dependent inhibitory effects of two antagonists on this R1881-induced interaction (Fig. 4). Furthermore, the use of different slides that only carry a prey library (Prey-Reporter or PR-slides) with cell lines that carry a stable or transient bait construct increases the flexibility of this assay (11). The high capacity of the cell arrays and the small amounts of reagents that are required make this assay currently one of the most economical high-throughput detection assays for protein–protein interactions in mammalian cells.

High-Throughput Mammalian Two-Hybrid Screening

169

Fig. 2. Application of CAPPIA for the detection of hormone-dependent interactions. A bait, AR-LBD coding for the androgen ligand-binding domain, was tested for its interaction with 17 different preys (samples A–Q, see Table 1). Triplicate spots of each prey-reporterbait (PRB) combination (samples A–Q), positive control (1: p53+SV40T) and negative controls (2: p53+TRAF), 3: SV40T, and 4: TRAF were printed and used to reverse transfect HEK 293T cells in the presence of 10−8 M of the synthetic androgen R1881 (Perkin Elmer) for 3 days. Only in the presence of the androgenic compound, AR-LBD was found to specifically interact with AR-NTD, the N-terminal domain of the androgen receptor. (a) A BIOCCD scanner image of a representative slide. (b) The corresponding relative fluorescence values obtained for the different bait–prey combinations, expressed relative to the fluorescence obtained with constitutively expressed autofluorescent proteins.

2. Materials 2.1. Expression Plasmids

1. Mammalian Two-Hybrid Assay Kit from Stratagene. This kit includes following control plasmids: pBD-NF-kB (a single protein that can bind and activate the reporter), pBD-p53 (control bait plasmid), pAD-SV40T (interacting control prey plasmid), and pAD-TRAF (noninteracting control prey plasmid). We cloned our genes of interest (Table 1) in the bait and prey vectors as described by the vendor (see Note 1). 2. Endo-Free Plasmid Maxi Kit (QIAGEN).

170

Fiebitz and Vanhecke

Table 1 List of preys (samples A–Q) used to screen for protein interaction with the androgen receptor bait (AR-LBD) in Fig. 2 Sample Prey proteins

Accession number

Gene ID

A

Pea3 (full-length)

NM_001986

2118

B

Pea3 N-terminal domain

C

Pea3 middle domain

D

Pea3 C-terminal domain

E

OTEX Menin

NM_139282 NM_130802

158800 4221

F

Menin (aa 1-455)

G

Menin (aa 224-455)

H

Menin (aa 456-615)

I

Menin (aa 224-615)

J

Menin (aa 1-223)

K

NCoR domain

NM_006311

9611

L

SMRT domain AR

NM_0076312 NM_00044

9612 367

M

Hinge region of AR

N

DNA-binding domain (DBD) of AR

O

N-terminal domain (NTD) of AR

P

Ligand-binding domain (LBD) of AR

Q

ALIEN domain

NM_004236

9318

The preys were constructed by cloning genes or indicated domains of genes potentially associated with nuclear receptor function in the pCMV-AD vector. All samples were a kind gift of Dr. Haendler, Bayer Schering Pharma, Germany

2.2. Sample Preparation

1. Gelatin powder (Type B: 225 Bloom; SIGMA©). 2. 0.2% gelatin solution: Dissolve 0.2 g gelatin in 100 ml MilliQ water and heat the solution at 60°C (in a water bath) for 15 min. The solution is immediately sterile-filtered through a 0.45 mm cellulose acetate membrane (Falcon®, Becton Dickinson) (see Note 2). 3. Effectene transfection reagent (Qiagen).

High-Throughput Mammalian Two-Hybrid Screening

171

4. EC-sucrose buffer: Sucrose (Stratagene) is dissolved in EC-buffer (part of the Effectene transfection reagent (Qiagen)) to a final concentration of 0.2 M. The solution is sterile-filtered with a 0.22-mm filter (Filtropur S 0.2, Sarstedt) and stored in aliquots at 4°C. 2.3. Slides

1. Standard 25 × 75 × 1.0 mm slides with low autofluorescence (Menzel-Gläser AG, Germany). 2. Cover glass, size: 16 × 16 mm, thickness: 0.17 mm (Galvoptics Ltd) (see Note 3). 3. VECTABOND™ Reagent (SP-1800, Vector Labs). 4. Poly-l-lysine solution (0.1% (w/v) in water, #P8920 SIGMA). 5. Acetone.

2.4. Printing of the Arrays

1. SciFlexArrayer noncontact piezo-dispensing system (Scienion AG, Germany) with 70 mm nozzle or InstrumentONE highperformance noncontact microdispensing system (M2-Automation, Germany) with 60 mm nozzle. 2. Costar 384-well plates (VWR).

2.5. Reverse Transfection and Fixation

1. Sterile Quadriperm boxes (Vivascience). 2. Complete cell culture media: Dulbecco’s Modified Eagle’s Medium (D-MEM) with 10% fetal bovine serum, penicillin/ streptomycin, and 1% l-glutamine (all from Invitrogen/ GIBCO) (see Note 4). 3. Adherent cell lines: HEK 293, HEK 293 T, PC-3, HeLa, COS7, HepG2 from ATCC. 4. Accutase (PAA) (see Note 5). 5. Sterile phosphate-buffered saline (PBS) (GIBCO). 6. 37% formaldehyde solution (Sigma). 7. Fixing solution: PBS, 4 M sucrose, 3.7% formaldehyde (diluted from 37% solution). 8. Glass cover slips of 22 × 64 mm, thickness No.1 (BDH). 9. Fluoromount-G (Southern Biotech). 10. Nuclear stain: 300 nM 4′,6-Diamidino-2-phenylindole dihydrochloride (DAPI) (SIGMA) (see Note 6). 11. Nail polish.

2.6. Fluorescence Detection and Analysis

1. DNA microarray reader (BIOCCD Image Reader, PE Applied Biosystems) with excitation 470/30 and emission 510/20 filters for detection of EGFP and ZsGreen protein expression.

172

Fiebitz and Vanhecke

2. Picture handling: AxioVision LE (Zeiss) software. 3. Image densiometry analysis: AlphaEase FC Stand Alone Software, version 4.0.0 (Alpha Inotech).

3. Methods 3.1. Plasmid Constructs 3.1.1. Bait and Prey Constructs

Any mammalian two-hybrid system that is based on the transfection and expression of chimeric proteins can be used in CAPPIA (see Notes 7, 8, and 9). Here, we used the expression bait (pCMVBD) and prey (pCMV-AD) vectors from the Mammalian TwoHybrid Assay Kit from Stratagene to generate proteins of interest as GAL4 DBD fusions (bait) and fusions with the transactivation domain of NF-kB (prey). Cloning of the open-reading frames was performed by standard cloning procedures (11) (see Note 10). For the screening of hormone-dependent interactions, we generated ten individual baits and 16 different preys coding for proteins or protein domains potentially associated with nuclear receptor function (Table 1).Positive and negative control plasmids from the kit were pBD-NF-kB, pBD-p53, pAD-SV40T, and pAD-TRAF (Fig. 2).

3.1.2. Reporter Constructs

To improve the ease and speed of detection of protein–protein interactions (PPI) on cell arrays, we created an autofluorescencebased and GAL4-driven reporter plasmid, Gal4-pZsGreen (see Note 11).

3.1.3. Constructs Used to Monitor Transfection Efficiency and Array Outline

A constitutive CMV-driven EGFP construct (pcDNA4-EGFP) was generated by PCR amplification of EGFP from pIRES2EGFP (Clontech) and TA cloning into pcDNA4/HisMax TOPO (Invitrogen) (see Note 12). This construct is printed at the border of the array to monitor transfection efficiency and to establish the coordinates of the printed array after reverse transfection.

3.1.4. Preparation of Plasmid Solutions

All plasmids were purified by Endo-Free Plasmid Maxi Kit (QIAGEN) and dissolved in endotoxin-free TE buffer, as indicated by the manufacturer (see Notes 13 and 14).

3.2. Sample Preparation

As originally described (9), reverse transfection of DNA can be performed in two ways. In the so-called DNA method, the plasmid/gelatin mixtures are first printed on the slides and air-dried. Subsequently, the DNA on the array is treated with transfection reagent before adding the cells. For the second, Lipid-DNA method, arrays are printed with plasmid/gelatin mixtures that already contain the transfection reagent. For CAPPIA, we prefer to use the Lipid-DNA method as the fluorescence signals were consistently brighter as compared to the DNA method (see Note 15).

High-Throughput Mammalian Two-Hybrid Screening

173

The Lipid-DNA method: 1. Plasmids or plasmid mixtures are prepared as follows with total DNA concentrations of 50 ng/ml (see Notes 16, 17, and 18). (a) For mixtures containing one plasmid (for example, pcDNA4-EGFP), 2.5 mg of plasmid is diluted in 15 ml EC-sucrose buffer. (b) For mixtures containing three plasmids, 800 ng of bait, 800 ng of prey, and 800 ng of reporter construct (total amount of around 2.5 mg DNA) are diluted in 15 ml EC-sucrose buffer (see Notes 19 and 20). 2. To each DNA/EC-sucrose buffer, 1.5 ml enhancer (part of Effectene kit) is added, mixed immediately by gentle vortexing, and incubated for 5 min at RT (see Note 21). 3. 7 ml Effectene (Qiagen) is added, mixed by gentle vortexing, and incubated for 10 min at RT. 4. Finally, 25 ml of 0.2% gelatin solution is mixed with the DNA/ Enhancer/Effectene solution so that the final concentration of gelatin in the samples is approximately 0.1% (see Note 22). 5. Prepared samples are incubated for a minimum of 1 h before spotting. They can be stored for a week at 4°C. 3.3. Slide Surface

We compared a large set of different self-made or commercial slides for use in CAPPIA and found that self-made VPL slides (coated with VECTABOND™ Reagent and Poly-l-lysine) offered the best choice regarding the cost per slide and efficiency of transfection (see Notes 23 and 24). VPL slides were made by treating standard 25 × 75 × 1.0 mm slides or 16 × 16 × 0.17 mm cover slips stacked/held in standard slide staining racks (see Note 25): 1. Glass slides are shaken for 2 h in a glass jar containing cleaning solution (70 ml NaOH 1.75 M + 160 ml distilled water + 240 ml ethanol 100%) (see Note 26). 2. Wash 3 × 5 min in distilled water. 3. Wash 5 min in acetone, let slides dry shortly (see Note 27). 4. Treat 5 min in VECTABOND™ solution (7 ml VECTABOND™ Reagent + 350 ml acetone) in a glass jar. 5. Wash 3 × 30 s in distilled water. 6. Dry at 37°C. 7. Transfer the slides to plastic slide staining racks and soak in poly-l-lysine solution (20 ml 0.1% (w/v) poly-l-lysine + 20 ml PBS + 160 ml distilled water) in plastic container (see Note 28).

174

Fiebitz and Vanhecke

8. Shake gently for 45 min at 4°C. 9. Wash briefly in distilled water. 10. Dry at 55°C (see Note 29), store in the dark and under vacuum (see Notes 30 and 31). 3.4. Printing of the Arrays

We used both the “sciFlexArrayer piezo-dispensing system S5” and “InstrumentONE high-performance noncontact micro dispensing system” for automated spotting. These systems are based on noncontact dispensing of nanoliter volumes with piezo capillaries (see Note 32). 1. Samples were plated in 384-well plates (30 ml/well) and plates were centrifuged briefly to remove trapped air bubbles. 2. Arrays were printed with a distance between each feature of 1.0 mm and using nozzles that generate droplets of approximately 400 pl. Every feature of the array was made with a total of 8 nl sample obtained by repetitive dispensing of 20 drops per spot. Triplet features were spotted for each set of bait/prey combination. 3. In order to monitor transfection efficiency and to localize the array after transfection, the control plasmid, pcDNA4-EGFP, is printed at defined places in the array (for example, at the corners or at the borders of the array). The array is printed with a minimum distance from the edges of the glass slide of 2 mm. 4. The arrays were dried for minimum 1 h before transfection, but can also be stored, at the dark and dry, for at least 6 months at 4°C and even longer at –80°C (see Note 33). Although cell arrays can theoretically carry up to 8,000 features per slide (3, 9, 12), the actual number of features per slide is dependent on the type of application. Since the efficiency of the simultaneous transfection of three plasmids is lower than that of a single plasmid and since the reporter will only be expressed after sufficient amounts of both prey and bait proteins are expressed in the same cell, there is a need to increase the number of cells per spot in CAPPIA in order to guarantee the robust detection of interacting proteins. Consequently, the features on CAPPIA slides need to be bigger and hence the slide capacity will be lower. Still, CAPPIA slides can be printed with each containing up to 900 different preys as single spots when using a spot-to-spot distance of 1 mm. This is equivalent to nine standard 96-microwell plates and data can then be collected from a number of identical slides to obtain replicate data points per sample. For the data shown in Fig. 2, slides were printed with 17 different prey–bait combinations as triplicate features in addition to a series of positive and negative controls to localize the array and monitor transfection efficiency as well as background fluorescence.

High-Throughput Mammalian Two-Hybrid Screening

175

Fig. 3. Cell array-based PPI screens in different cell lines. Transfection efficiency and specific protein–protein interaction in different cell lines are demonstrated using microarrays printed with solutions containing Gal4-pZsGreen reporter and plasmids coding for the known interacting p53 (pBD-p53) and SV40-T (pAD-SV40T) hybrid proteins (boxed, line B ). As negative control (line A), Gal4-pZsGreen reporter was cotransfected together with the plasmids encoding the noninteractingproteins p53 (pBD-p53) and TRAF (pAD-TRAF). The pBD-NF-kB control plasmid (pBD-NFkB) expressing the GAL4 DNA-binding domain fused to the transcription activation domain of NF-kB is used as a positive control to monitor transfection efficiency and reporter performance (line C ). A construct expressing EGFP under control of a CMV promoter (line D ) is typically printed as a frame at the periphery of the arrays in order to locate the arrayed features. (a) Examples of images of different adherent cell lines transfected using identical microarray slides. (b) Transfection efficiency as reflected by the level of NF-kB-induced reporter expression is different depending on the cell line tested and is typically lowest for WI-38, COS7, and HepG2. Highest levels of transfection were repeatedly obtained with HEK293T. Comparable results were obtained for PC-3, HEK293, and HeLa. Data shown are from representative experiments and represent mean fluorescence obtained from six features per sample.

3.5. Reverse Transfection

We have performed most CAPPIA experiments with HEK 293T cells (see Note 34), but various other adherent cell lines such as HEK 293, PC-3, HeLa, COS7, HepG2, and WI-38 are also suitable for reverse transfection experiments, as shown in Fig. 3. Adherent mammalian cells were cultured as specified for normal transfection protocols. 1. All cells were cultured at 37°C, 5% CO2 in D-MEM (GIBCO) with 10% Fetal Bovine Serum, penicillin/streptomycin, and 1% l-glutamine (see Note 35). Briefly, the cells are detached with Accutase when approaching confluence on day-1 (d-1) and seeded so that the cells are subconfluent on day 0 (d0) (see Note 36).

176

Fiebitz and Vanhecke

2. For HEK 293T, HEK 293, PC-3, WI-38, and HepG2 1 × 107 cells and for COS7, 5 × 106 cells were seeded on d-1 in 60 cm2 cell culture plates. For HeLa, 5 × 106 cells were precultured in a 145 cm² plate. 3. On the day of transfection, the printed slides are placed in sterile Quadriperm boxes (Vivascience) with the DNA array samples on top. Each sterile Quadriperm box can hold four standard 25 × 75 × 1.0 mm glass slides separated from each other (see Note 37). 4. The cultured cells are detached, counted, and added carefully on top of the printed slides avoiding the addition of the cell solution directly to the printed area of the slide. The boxes are then carefully transferred in an incubator at 37°C and with 5% CO2 (see Note 38). Cells are seeded at 3.5 × 106 (HEK 293T, HEK 293, HEK 293-LBD, and HepG2), 3 × 106 (PC-3, COS 7, and WI-38), and 1 × 106 (HeLa) cells per slide in 8 ml complete media (see Note 39). As described above, sample and slide preparation were identical for all cell lines tested and only the number of cells and time of transfection had to be optimized for each cell type. Using this protocol, we have observed that both the transfection efficiency (levels of the NFkB control) and efficiency of protein–protein interaction (level of reporter signals for known interacting proteins) vary depending on the cell type used. These results are shown in Fig. 3. Moreover, the effects of activators or inhibitors on known protein interactions can be tested by adding compounds to the individual slides at the start of transfection. For example, in Fig. 4, a dose-dependent induction of reporter expression was obtained reflecting the dose-dependent interaction of the ligand-binding domain and N-terminal domain of the androgen receptor by adding different amounts of the synthetic agonist R1881 (Fig. 4a), In addition, we could reiterate the dosedependent inhibitory effects of two antagonists, medroxyprogesterone acetate (MPA) and hydroxyflutamide (OH-Flu), on R1881-induced AR-LBD and AR-NTD interaction (Fig. 4b). Importantly, the resolution of CAPPIA allowed the detection of quantitative differences in antagonist activity. Indeed, the R1881induced interaction was inhibited with 1 nM MPA, whereas almost 100 nM OH-Flu was required to achieve the same level of inhibition. 3.6. Fixation

After 48 or 72 h (depending on the cell line used), slides are fixed and mounted as follows (see Notes 40 and 41): 1. Wash transfected slides 1× with PBS. 2. Fix with PBS/4M sucrose/3.7% formaldehyde solution for 30 min.

High-Throughput Mammalian Two-Hybrid Screening

177

Fig. 4. Dose–response of AR-LBD and AR-NTD interactions to androgenic and antiandrogenic compounds. AR-LBD (bait) and AR-NTD (prey) interaction is analyzed on cell arrays in the presence of different concentrations of agonist and antagonists. Synthetic androgen R1881 (Perkin Elmer) is used as agonistic compound in concentrations between 10−7 and 10−11. Antagonistic compounds in the experiments were hydroxyflutamide (OH-Flu) and medroxyprogesterone acetate (MPA), both tested in concentrations between 10−7 and 10−11 (both from Shering AG, Germany). (a) Dosedependent induction of AR-LBD and AR-NTD interaction by R1881, showing a maximal response from 10−8M onwards. (b) Dose-dependent inhibition of the R1881-induced AR-LBD and AR-NTD interaction by the two antagonists MPA and OH-Flu, respectively. Quantitative analysis of this inhibition reflects the known stronger antagonistic potency of MPA as compared to OH-Flu.

3. Wash 1× with PBS. 4. Stain cell nuclei with DAPI for 5–10 min at RT (see Note 42). 5. Wash 1× with PBS. 6. Remove slide from PBS, drain off most of the PBS. Carefully wipe the underside of the slide dry and place on flat filter paper or tissue.

178

Fiebitz and Vanhecke

7. Add a drop of fluorescence-enhancing mounting fluid (for example, Fluoromount-G) on the fixed monolayer and carefully apply a thin glass cover slip (see Notes 43, 44, and 45). The mounted slides can be stored at 4°C until analysis. Slides that are stored dry, cold, and in darkness are stable for at least 6 months at 4°C (see Note 46). 3.7. Detection and Analysis

Fluorescent signals were detected using a DNA microarray reader (BIOCCD Image Reader) with excitation 470/30 and emission 510/20 filters for EGFP and ZsGreen. Pictures were handled by AxioVision LE, and image densitometry analysis was done with AlphaEase®FC. Fluorescent signals/features can be analyzed by fluorescence microscopes, standard DNA array scanners, and CCD cameras. Analysis of fluorescent features can be performed using any standard DNA array program that is available from the manufacturer of the fluorescence detector. Fluorescence signals were obtained as the sum of pixel values for each feature after background correction. Signals were normalized for transfection efficiency relative to the level of EGFP expression of the control pcDNA4-EGFP vector on the same slide, and the mean value from six spots was calculated for each sample (Examples provided in Figs. 2–4).

4. Notes 1. Typically, one or a few baits are tested against a large library of prey constructs. 2. The type of gelatin is of importance to obtain good reverse transfection (9). Filtration has to be done immediately as the gelatin rapidly cools down and then becomes too viscous to filter. After cooling, the solutions can be stored in aliquots at 4°C for a couple of months. Solutions should be clear. Turbid solutions should be discarded. 3. Cover slips were chosen that are thicker than standard glass slides to reduce the risk of breaking during manipulation. 4. Any standard cell culture media can be used. However, preference is given to the same media formulations that are used to culture the cell line that is to be transfected. 5. Accutase is a mixture of proteases that is used for the dissociation of adherent cells. Also, Trypsin/EDTA solutions can be used. 6. DAPI is dissolved at 300 mM in DMSO and stored at −20°C in aliquots. Before use, the stock solution is diluted 1,000× in PBS.

High-Throughput Mammalian Two-Hybrid Screening

179

7. Typically, most commercial available mammalian two-hybrid systems are based on a protein of interest, the bait, fused to the DBD and a prey protein fused to the activation domain (AD) of a transcriptional activator. The functional transcription factor is reconstituted upon the physical interaction between bait and prey proteins. 8. Cell arrays can also be used for the screening of DNA-binding proteins using mammalian one-hybrid assays. In such assays, cells are transfected with a reporter construct and a protein fused to a transactivator. Upon binding of the fusion protein with a gene-specific promoter upstream of the reporter construct, the transactivator activates the promoter and cells become fluorescent. 9. An alternative system that could be used for CAPPIA is MAPPIT, designed to analyze the interaction of membranebound molecules (13). 10. High-throughput cloning of bait and prey constructs can be achieved by modifying the bait and prey plasmids to incorporate recombination sequences for the easy and fast generation of large number of recombinant proteins. For example, the vectors can be modified to be compatible with the Gateway system of Invitrogen for which large libraries of cloned genes are already available. 11. As described in Fiebitz et al. 2008, the Gal4-pZsGreen was created by cloning the GAL4 upstream activating sequences from pGAL/lacZ (Invitrogen) into the multicloning sites of pZsGreen1-1, a promoter-less vector encoding the autofluorescent protein ZsGreen (Clontech), a fluorescent protein derived from Anthozoa reef corals that is more photostable than EGFP. This construct was used instead of the lacZ reporter of Stratagene. Alternatively, the original lacZ reporter can be used in combination with anti-lacZ antibodies for fluorescence detection. We successfully used the mouse anti-beta galactosidase monoclonal antibody (# ab1047.100 from Abcam) in combination with secondary fluorescent-conjugated antimouse antibodies. 12. Any mammalian expression construct where EGFP or ZsGreen is constitutively expressed can be used. 13. It is not necessary to filter sterilize the plasmid solutions before use in reverse transfection. 14. Other systems of plasmid purification can be used, but should yield plasmid solutions with a minimum concentration of 700 ng/ml. The use of endotoxin-free plasmid preparations is reported to improve transfection efficiencies in mammalian cells.

180

Fiebitz and Vanhecke

15. All solutions should be sterile and manipulations should be made as sterile as possible. 16. For small number of bait/prey combinations, the samples can be prepared manually; for large number of samples, liquid dispensing tools are recommended. 17. The amount of sample to be prepared has to be calculated based on the amount of sample that will be printed per feature in the array, the number of replicate features per sample, and the number of slides to be printed. Automated dispensing systems require a minimum working volume in addition to the volume that will be dispensed to avoid uptake of air in the system. For CAPPIA, typically samples of 50 ml were prepared. 18. TE buffer is used to have equal volumes in every sample. 19. Preferentially, master mixes of bait and reporter are made and aliquots of these are mixed with the different prey constructs to reduce the variability between the samples. 20. Any “triple-transfection” such as a bait, prey, and reporter plasmid described here requires extensive optimization of the transfection protocol. The amount of total DNA has to be considered as well as the concentration of every plasmid. The total amount of DNA should not exceed 60 ng/ml. With DNA concentrations over this limit, cell viability decreased and the signal intensity of the fluorescence was clearly lower. The ratio of plasmids is a compromise between the need of sufficient amount of each plasmid and avoiding excess DNA that would affect cell viability. 21. Any mammalian transfection reagent can be used, but amounts of DNA and transfection reagent and time of incubation will have to be optimized. 22. We found 0.1% gelatin to give the best results. Higher or lower gelatin concentrations result in features/cells with reduced fluorescence following transfection. 23. Vectabond reagent has been developed to treat glass surfaces for efficient binding of tissue sections. We found it improves the binding of poly-l-lysine, required for the attachment of the mammalian cells during transfection, as well the binding of the DNA/gelatin solutions that are spotted on the slides. VPL slides were qualitative equal to GAPS II-coated slides (Corning Incorporated Life Science), used in the original protocol of Sabatini et al, in reverse transfection experiments. 24. Self-made glass slides for use in CAPPIA furthermore allows the preparation of any size of slide/cell array. Smaller glass slides such as cover slips can be used in microwells to test the effect of different compounds on replicate arrays. In addition, smaller arrays require smaller number of mammalian cells,

High-Throughput Mammalian Two-Hybrid Screening

181

permitting the analysis of protein interactions in, for example, primary cells. 25. Slides do not have to be treated under sterile conditions, but all solutions should be kept sterile and slides should be handled with gloves. 26. Glass jars have to be used as NaOH might affect plastic. 27. Glass jars have to be used as aceton might affect plastic. 28. Plastic containers have to be used here as poly-l-lysine will bind the glass containers resulting in lower poly-l-lysine concentrations. 29. The higher temperature helps to fix the Vectabond/poly-llysine coating to the glass slides. 30. Or alternatively, in a closed container with DRIERITE Anhydrous Calcium Sulfate (#22890-229, VWR). 31. Some of the other surfaces that also work fine for reverse transfection are: Lab-Tek™ Chamber Slide™ System (NuncTM), GAPS II-coated slides (Corning Incorporated Life Science), and a number of cell culture-treated plastic surfaces. 32. Contact printing with split-pin or solid-pin arrayers is also possible, but requires additional optimization (9). 33. See introduction. Store in a closed container with DRIERITE Anhydrous Calcium Sulfate (#22890-229, VWR). 34. HEK293-T cells are derived from HEK293 cells and constitutively express the large T antigen of SV40, important for the extrachromosomal amplification of circular plasmid DNA resulting in higher expression levels of transient transfected constructs. 35. Cell lines were only used for reverse transfection experiments up to passage P20, carefully preventing overgrowing during culture. 36. The amount of cells seeded on d-1 is different for each cell type and depends on the growth rate and size of the cells. 37. Since the cells have to stick to the slide and not to the cell culture dish, the reverse transfection is performed in nontreated cell culture dishes. In the original protocol of Sabatini, nontreated square dishes (Becton Dickinson) were used to reverse transfect three slides simultaneously (9). However, we prefer the Quadriperm box from Vivascience as the slides are physical separated. The Quadriperm box is also more convenient when slides are analyzed after different time points of transfection as each slide can be removed from the box without major disturbance of the other slides. 38. Amount of cells depends on the size of the cells and should be sufficient to create a near confluent monolayer. Too many

182

Fiebitz and Vanhecke

cells will decrease the cell viability at the end of the assay. Too few cells will reduce the number of cells that will get transfected, reducing the total fluorescence signal of the feature. 39. The viability of the cells can be improved by carefully replacing 1/3 of medium every day, care should be taken so as not to disrupt the monolayer. 40. The slides can be washed and fixed in the Quadriperm box used for culture by carefully replacing the solutions. Alternatively, the slides can be lifted carefully using a scalpel blade and sequentially transferred by gloved hands to a new container containing the washing and fixing solutions. Washing and fixing are performed with enough solution to cover the whole slide. 41. Care should be taken so as not to disrupt the monolayer of cells during washing and fixing. Especially, care should be taken when using weakly adherent cell lines like HEK293-T cells. 42. Nuclear staining with DAPI is useful when the slides are analyzed with fluorescent microscopes only. 43. The cover slip should be large enough to cover the area where the array was printed. 44. Avoid trapping air bubbles when applying the cover slip. 45. Avoid sliding of the cover slip once it is placed on the monolayer. The cover slip is immediately immobilized on the slide by applying a small drop of nail polish at each corner of the cover slip. Slides are carefully stored at 4°C for at least 2 h before analysis. 46. In order to increase the possible combinatorial screens for protein interactions using cell arrays and hence further improve the high-throughput application of CAPPIA, alternative slides can be printed on which the bait is omitted. Each spot on these so-called prey-reporter- (PR-)slides only contains the reporter and one of the prey constructs. To screen for interacting partners, the bait is then introduced into the cells before adding them to the prey-reporter-arrays. This is done by generation of stably transfected cell lines (PR-stable-bait assay), or alternatively, the cells can be transfected transiently with the bait shortly before being added to the arrays (PR-trans-bait assay) by mixing them with DNA complexed with transfection reagent (11). PR-slides are more suitable for the large-scale screening of novel bait–prey interactions. Up to 900 features can be spotted per slide so that each slide represents comprehensive collections of preys. Since there is no bait on PR-slides, these libraries can be screened with any bait of interest, further increasing the high-throughput application of CAPPIA. In addition, PR-slides can be printed in

High-Throughput Mammalian Two-Hybrid Screening

183

large batches and stored for more than 6 months at –80°C. The slides can be shipped on ice, allowing any research group with access to standard culture facilities to interrogate the prey library with their bait of interest. References 1. Stelzl U, Worm U, Lalowski M, Haenig C, Brembeck FH, Goehler H, Stroedicke M, Zenkner M, Schoenherr A, Koeppen S et al (2005) A human protein-protein interaction network: a resource for annotating the proteome. Cell 122:957–968 2. Luo Y, Batalao A, Zhou H, Zhu L (1997) Mammalian two-hybrid system: a complementary approach to the yeast two-hybrid system. Biotechniques 22:350–352 3. Leonhardt SA, Altmann M, Edwards DP (1998) Agonist and antagonists induce homodimerization and mixed ligand heterodimerization of human progesterone receptors in vivo by a mammalian two-hybrid assay. Mol Endocrinol 12:1914–1930 4. Fields S, Song O (1989) A novel genetic system to detect protein-protein interactions. Nature 340:245–246 5. Uetz P (2002) Two-hybrid arrays. Curr Opin Chem Biol 6:57–62 6. Drees BL (1999) Progress and variations in two-hybrid and three-hybrid technologies. Curr Opin Chem Biol 3:64–70 7. Dang CV, Barrett J, Villa-Garcia M, Resar LM, Kato GJ, Fearon ER (1991) Intracellular leucine zipper interactions suggest c-Myc hetero-oligomerization. Mol Cell Biol 11:954–962

8. Shioda T, Andriole S, Yahata T, Isselbacher KJ (2000) A green fluorescent protein-reporter mammalian two-hybrid system with extrachromosomal maintenance of a prey expression plasmid: application to interaction screening. Proc Natl Acad Sci U S A 97:5220–5224 9. Ziauddin J, Sabatini DM (2001) Microarrays of cells expressing defined cDNAs. Nature 411:107–110 10. Suzuki H, Fukunishi Y, Kagawa I, Saito R, Oda H, Endo T, Kondo S, Bono H, Okazaki Y, Hayashizaki Y (2001) Protein-protein interaction panel using mouse full-length cDNAs. Genome Res 11:1758–1765 11. Fiebitz A, Nyarsik L, Haendler B, Hu YH, Wagner F, Thamm S, Lehrach H, Janitz M, Vanhecke D (2008) High-throughput mammalian two-hybrid screening for proteinprotein interactions using transfected cell arrays. BMC Genomics 9:68 12. Mannherz O, Mertens D, Hahn M, Lichter P (2006) Functional screening for proapoptotic genes by reverse transfection cell array technology. Genomics 87:665–672 13. Eyckerman S, Verhee A, der Heyden JV, Lemmens I, Ostade XV, Vandekerckhove J, Tavernier J (2001) Design and application of a cytokine-receptor-based interaction trap. Nat Cell Biol 3:1114–1119

wwwwwwwwwwwwwwww

Chapter 12 Protein–Protein Interactions: An Application of Tus-Ter Mediated Protein Microarray System Kalavathy Sitaraman and Deb K. Chatterjee Abstract In this chapter, we present a novel, cost-effective microarray strategy that utilizes expression-ready plasmid DNAs to generate protein arrays on-demand and its use to validate protein–protein interactions. These expression plasmids were constructed in such a way so as to serve a dual purpose of synthesizing the protein of interest as well as capturing the synthesized protein. The microarray system is based on the high affinity binding of Escherichia coli “Tus” protein to “Ter,” a 20 bp DNA sequence involved in the regulation of DNA replication. The protein expression is carried out in a cell-free protein synthesis system, with rabbit reticulocyte lysates, and the target proteins are detected either by labeled incorporated tag specific or by gene-specific antibodies. This microarray system has been successfully used for the detection of protein–protein interaction because both the target protein and the query protein can be transcribed and translated simultaneously in the microarray slides. The utility of this system for detecting protein–protein interaction is demonstrated by a few well-known examples: Jun/Fos, FRB/FKBP12, p53/MDM2, and CDK4/p16. In all these cases, the presence of protein complexes resulted in the localization of fluorophores at the specific sites of the immobilized target plasmids. Interestingly, during our interactions studies we also detected a previously unknown interaction between CDK2 and p16. Thus, this Tus-Ter based system of protein microarray can be used for the validation of known protein interactions as well as for identifying new protein–protein interactions. In addition, it can be used to examine and identify targets of nucleic acid–protein, ligand–receptor, enzyme–substrate, and drug–protein interactions. Key words: Tus/Ter, Gateway cloning, Protein microarray, Expression plasmids, Query plasmid, Protein–protein interaction, Cell-free protein synthesis

1. Introduction Over the years, protein microarrays have become more and more popular, for the throughput screening of protein interaction, protein function, and protein profiling (1–3). The development of

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_12, © Springer Science+Business Media, LLC 2011

185

186

Sitaraman and Chatterjee

protein microarray is a significant step toward a genome wide, systematic characterization of proteins. However, protein micro array technology is not that straight forward compared to DNA microarray technology due to the inherent complexity of the protein molecules. Its applicability is hampered by the high cost and technical limitations imposed by the vast requirement of stable, purified, and properly folded proteins. Nevertheless, several different approaches have been developed for the generation of protein microarrays (4–8). Nord et al. in their protein microbead display (9) describe a technology platform to capture proteins via antigen–antibody binding. Using streptavidin coated polystyrene microbeads as solid support, the biotinylated PCR products are immobilized. The PCR product carried a T7 promoter and a FLAG epitope in-frame with two IgG binding domains. Anti-FLAG antibody was then anchored on to the same microbeads, and the protein synthesis was carried out with a coupled transcription and translation system (10). The nascent proteins were captured via FLAG peptide–FLAG antibody interaction. Ramachandran et al. in a microarray format applied a similar antibody-mediated protein detection technology (11, 12). In their method, purified expression construct DNAs were microarrayed via biotin–avidin interaction, and the proteins were synthesized as GST fusions. The array was simultaneously printed with polyclonal GST antibody to capture the newly synthesized GST-fusion proteins following a coupled cell-free transcription and translation reaction. In both these formats, newly synthesized proteins were captured through protein–antibody interaction. Our Tus-Ter based system (13) eliminates the need for antibody or any other capture reagents to immobilize the newly made protein on to the microarray surface. Here, the expression vector DNA, besides directing the synthesis, captures the protein also at the designated location on the microarray surface. This is accomplished through a high affinity DNA–protein interaction between Escherichia coli “Tus” protein and a 20 bp “Ter” sequence. The affinity binding of Tus and Ter was reported to be ~3–7 × 10−13 M (14). All the expression plasmids used for generating the protein microarray carry this Ter sequence, and the protein of interest is synthesized as Tus fusion protein. Thus, the embedded Ter sequence serves as a capture agent for the newly synthesized fusion protein. Since only the plasmid DNA is printed, not only array fabrication is simple, but the array stability is also no longer an issue. We have used this Tus-Ter system to express and capture a number of proteins as well as to evaluate protein–protein interactions (see Subheading 3.7). A detailed description of the method in this chapter will focus on the construction of expression plasmids used in Tus-Ter based protein–protein interaction studies, their expression and detection of the interaction.

Protein–Protein Interactions

187

2. Materials 2.1. Cloning

1. Gateway Cloning System (Invitrogen, Carlsbad, CA). 2. Oligonucleotides can be ordered from numerous suppliers, such as Operon, Huntsville, AL. They generally do not require HPLC or gel purification, and for Gateway reactions the oligo needed is so small that a 50-nmol synthesis scale is more than sufficient. 3. Primers for PCR amplification should be resuspended to a concentration of 10 mM in TE (10 mM Tris–HCl, pH 8.0, 0.1 mM EDTA). 4. 2× Phusion Master Mix HF (NEB, Beverley, MA). 5. Restriction enzymes NheI, MunI, BsrGI, ClaI, NgoMIV, XhoI, XbaI (New England Biolabs, Beverley, MA). 6. QiaQuick PCR purification Kit (Qiagen, Valencia, CA). 7. Agarose gel. 8. PureLink Quick Gel Extraction Kit (Invitrogen, Carlsbad, CA). 9. E. coli ccdB Survival competent cells (Invitrogen, Carlsbad, CA). 10. pTnT vector (Promega, Madison, WI). 11. DH5a chemically competent cells (Invitrogen, Carlsbad, CA). 12. LB agar plates with ampicillin, chloramphenicol (Teknova, Hollister, CA). 13. PureLink Quick Plasmid Miniprep Kits (Invitrogen, Carlsbad, CA). 14. PureLink HiPure Maxi Plasmid Kit (Invitrogen, Carlsbad, CA). 15. 20× SSC (Invitrogen, Carlsbad, CA). 16. 3× SSC: Add 170 ml of water to 30 ml of 20× SSC, mix and store at room temperature (RT).

2.2. Expression Check

1. TnT Quick Coupled Transcription/Translation Systems (Promega, Madison, WI). 2. 30°C incubator or ThermoMixer (Eppendorf). 3. 4–20% Tris-Glycine gel (Invitrogen, Carlsbad, CA). 4. BenchMark Prestained Carlsbad, CA).

Protein

Ladder

(Invitrogen,

5. MagicMark XP Western Protein Standard (Invitrogen, Carlsbad, CA). 6. iBlot Gel Transfer Device (Invitrogen, Carlsbad, CA). 7. iBlot Gel Transfer Stack, Nitrocellulose Filter (Invitrogen, Carlsbad, CA).

188

Sitaraman and Chatterjee

8. SuperSignal West Femto Maximum Sensitivity Substrate (Pierce, Rockford, IL). 9. Penta-His HRP conjugate Kit (Qiagen, Valencia, CA). 10. Monoclonal Anti-FLAG antibody produced in mouse (Sigma, St. Louis, MO). 2.3. Microarray Fabrication

1. MicroGrid II Arrayer and Print Head (BioRobotics, Woburn, MA). 2. 384-Well plate (Genetix, Boston, MA). 3. Fast Slides – nitrocellulose coated (Schleicher & Schuell Bioscience, Keene, NH). 4. Ultrasonicator. 5. MilliQ water. 6. Liquid nitrogen. 7. Incubator set at 80°C for baking the printed slides. 8. Silica gel (Sigma).

2.4. Quality Control of Microarray Printing

1. Safe-Lock 1.5 ml amber tubes Eppendorf (Sigma, St. Louis, MO). 2. Cy3-Spot QC detector oligo (IDT, Coralville, IA). The oligo is provided with 1× hybridization buffer. Add 500 ml of hybridization buffer to the tube containing the oligo, vortex. The final oligo concentration will be 2 mM. Make aliquots in amber tubes and store them at −20°C. Avoid unnecessary freeze-thaw and exposure to light. Stable for 6 months. 3. Polyvinylpyrrolidone (Sigma). 4. Blocking solution: 0.1% (w/v) polyvinylpyrrolidone/0.05% Tween-20. Take one packet of phosphate buffered saline with Tween 20 and weight out 1.0 g of polyvinylpyrrolidone. Dissolve reagents in 1 L of de-ionized water. Store at 4°C. 5. 5% sarkosyl: Dissolve 5 g of sarkosyl (Sigma) in 90 ml of water. Heating at 65°C helps. Make up the volume to 100 ml. Store at RT. 6. Wash buffer 1: 10× SSC (pH 7.0)/0.2% sarkosyl. Mix 5 ml of 20× SSC and 40 ml of 5% sarkosyl with 460 ml of water. Store at RT. 7. Wash buffer 2: 10× SSC, pH 7.0. Mix 250 ml of 20× SSC with equal amount of water; store at RT. 8. Wash buffer 3: 2× SSC (pH 7.0). Mix 100 ml stock 20× SSC with 900 ml of water. 9. Quadri-PERM 4-compartment slide container (Bellco Glass, Vineland, NJ).

Protein–Protein Interactions

10. Axon GenePix Sunnyvale, CA). 2.5. In Situ Expression of Proteins

4000

Scanner

(Molecular

189

devices,

1. Incubators set at 30 and 15°C. 2. TnT Quick Coupled Transcription/translation System (Promega). 3. Arrayed slide. 4. Query plasmid containing the gene coding for the interacting protein. 5. Fast Frame slide holder for four slides (Schleicher & Schuell Bioscience, Keene, NH). 6. Incubation chamber dual well (Schleicher & Schuell Bioscience, Keene, NH). 7. A plastic container to hold the Fast Frame slide holder.

2.6. Prehybridization and Hybridization

1. Phosphate buffered saline, pH 7.4 with Tween 20 (Sigma, St. Louis, MO). 2. Polyvinylpyrrolidone (Sigma). 3. 10× PBS (Invitrogen, Carlsbad, CA). 4. 1× PBS: Dilute 100 ml of 10× PBS with 900 ml of deionized water. 5. Blocking solution: 0.1% (w/v) polyvinylpyrrolidone/0.05% Tween-20 in 1× PBS. Take one packet of phosphate buffered saline and weigh out 1.0 g of polyvinylpyrrolidone. Dissolve reagents in 1 L of deionized water. Store at 4°C. 6. Hybridization solution: 0.1% (w/v) polyvinylpyrrolidone/0.1% Tween-20 in PBS. Take two packets of phosphate buffered saline with Tween 20 and weigh out 1.0 g of polyvinylpyrrolidone. Dissolve reagents in 1 L of deionized water. Store at 4°C. 7. Washing solution: 0.05% (w/v) Tween-20 in 1× PBS. Take one packet of phosphate buffered saline with Tween 20 and dissolve in 1 L of deionized water. Store at RT. 8. Gene-specific antibodies (c-Fos, MDM2, P16 from Santa Cruz Biotechnology, Santa Cruz, CA; and FKBP12 from Abnova, Taipei, Taiwan).

2.6.1. Labeling the Secondary Antibody

1. Antibody to mouse IgG produced in goat (KPL, Gaithersburg, MD). 2. Anti-rabbit IgG, antibody produced in goat (Sigma). 3. Cy3 Mono-Reactive Dye Pack (GE Healthcare). 4. Cy5 Mono-Reactive Dye Pack (GE Healthcare).

190

Sitaraman and Chatterjee

5. Protein Extraction & Labeling Kit (Clontech, Mountain View, CA). 6. Illustra MicroSpin G-25 Chromatography column (GE Healthcare). 2.7. Detection of Expression/ Interaction

1. Axon GenePix 4000 Scanner.

3. Methods The methods described below focus on (a) construction of base plasmid, (b) construction of plasmid that is coding for target protein of interest, (c) construction of query plasmid that codes for the protein that interacts with target protein, (d) microarray fabrication, (e) cell-free expression of proteins in rabbit reticulocyte lysate system, (f) antibody labeling, and (g) confirmation of expression and interaction. 3.1. Cloning

3.1.1. Construction of Base Microarray Plasmid

The Gateway Cloning system (15) was used to make the base microarray plasmid. Tus with a His-tag at the carboxy-end served as the fusion partner for the target templates. 1. A modified “Tus” (E47Q), which has a higher affinity for the Ter DNA sequence, was amplified from plasmid DNA by using Phusion mix. The oligos used were: (a) Forward – 5¢-ATT TTA GCT AGC GGA GGT GCG CGT TAC GAT CTC GTA GAC CGA CTC-3′. (b) Reverse 5¢-TATATT CAA TTG TTA atg atg gtg atg atg gtg ATC TGC AAC ATA CAG GTG CAG CCG TGG-3¢. (c) NheI and MunI sites (bold and underlined) and six-histidine tag (small letters) were incorporated into the oligo as indicated. 2. The PCR product was purified using QiaQuick PCR purification Kit, digested with NheI and MunI, run on an agarose gel, and the fragment was excised. 3. The fragment was then cloned into a derivative pDest47 (Invitrogen, Carlsbad, CA) termed pDest 472 that had been digested with the same enzymes. The transformation was done with ccdB survival cells. 4. The pDest 472-Tus clones obtained were screened by cutting with NheI and MunI and sequence verified. 5. A Ter site (bold) was synthesized by annealing two complementary oligos:

Protein–Protein Interactions

191

(a) 5¢-CCGGC CACTTTAGTTACAACATACTTATT AT-3¢. (b) 5¢-CGATAATAAGTATGTTGTAACTAAAGTG G-3¢. 6. These oligos form a double stranded Ter site, upon annealing with ClaI and NgoMIV overhangs. The annealed oligo was cloned into pDest 472-Tus digested with NgoMIV and ClaI to create pDest Microarray. 7. The clone was verified by sequencing. This is the base pDest microarray plasmid to clone any protein of interest by recombinational cloning. 3.1.2. Construction of Target Plasmids

Most of the genes of interest (Jun, FRB, p53, Cdk2, and Cdk4) we used in our protein–protein interactions studies are entry clones consisting of human open reading frames (ORF) cloned into a Gateway entry vector, p223 (Invitrogen). These genes contain a KOZAK sequence/ATG and no stop codon at the C-terminal end. All of these clones were sequence verified. LR reactions were performed according to manufacturer’s instruction (Invitrogen) using the base pDest microarray vector and entry clones in the donor p223 vector containing the proteins of interest. The LR clones were screened with restriction digestion with BsrGI. Maxipreps of these target plasmids were made using PureLink Maxiprep kit (Invitrogen), and the final pellet was resuspended in 3× SSC (see Note 1). These were the DNAs (see Note 2) we used for printing the array at a concentration of 400 ng/ml.

3.1.3. Construction of Query Plasmids

These are the plasmids added in the cell-free reaction so as to coexpress along with the target DNA on the slide. The genes of interacting proteins (c-Fos, FKBP12, MDM2, and p16) were amplified as XhoI-XbaI or XhoI-XmaI fragments to sub-clone into Promega’s pTnT vector. This vector has both SP6 as well as T7 promoter sites located immediately upstream of the multiple cloning site. These promoters allow the convenient use of a coupled transcription/ translation system. A FLAG tag (see Note 3) was incorporated at the amino- or carboxy-terminus of the gene so that interaction could be monitored either by anti-gene-specific or anti-FLAG antibody. 1. Amplify gene of interest using Phusion Mix with the following primers: (a) Forward – 5¢-TAATAT CTC GAG GCC ACC ATG gat tac aag gat gat gac gat aag GGG TCT-gene specific sequence-3¢. (b) Reverse – 5¢-TTAATAA TCT AGA-gene specific sequence-3¢. (c) XhoI and XbaI sites are indicated as bold and underlined and FLAG tag in small letters.

192

Sitaraman and Chatterjee

2. The PCR product was purified using QiaQuick PCR purification Kit. 3. Purified PCR product was digested with XhoI and XbaI, run on an agarose gel, and the fragment was excised using PureLink Gel purification kit. 4. The fragment was then cloned into pTnT vector that had been digested with the same enzymes to create pTnT-gene of interest vector. 5. The clones were screened and sequence verified. 6. Maxipreps of right clones were made using PureLink maxiprep kit, and the final pellet was resuspended in 3× SSC. 3.2. Expression Check

The expression plasmids of target proteins and query proteins made were first checked for protein expression before spotting them on to microarray slides. Using Promega’s TnT system, the protein synthesis was carried out in a 50 ml volume with rabbit reticulocyte lysate at 30°C (see Note 4). 10 ml of each sample was run on 4–20% Tris-Glycine gels. The gels were blotted using Invitrogen’s iBlot Transfer device following manufacturer’s protocol onto nitrocellulose filters. The blots were probed with either anti-his (for target plasmids) or anti-Flag antibody (for query plasmids) and detected with SuperSignal West Femto Maximum Sensitivity Substrate.

3.3. Microarray Fabrication

The maxi-prepped target plasmids were used to print the nitrocellulose-coated slides at a concentration of 400 ng/ml in 3× SSC. 20 ml of each sample including negative (base plasmid without ORF) and positive (pDest microarray plasmid-GFP) controls was placed in a 384-well source plate. Before printing the array, the solid pins were cleaned using an ultrasonicator. The sonication was done in milliQ water for 10 min. The pin holders (with pins) were dried under liquid nitrogen and used immediately. The instructions to operate the MicroGrid Arrayer can be found in many of the university/institutional Web sites besides the product vendor’s (for example, http://www.grc.nia.nih.gov/ branches/rrb/dna/protocolFAQs.htm; http://www.flychip. org.uk/printsop.php; Graz University of Technology, ErzherzogJohann-University, Austria). 1. Follow the instructions on operating the machine. 2. Keep the humidity at 50% (see Note 5). 3. Maintain the spacing of 0.55 mm between the spots (center to center) and the distance between the sub-array as 1.2 mm. 4. Print three rows of five spots (3 × 5) for each sample with three strikes at each spot.

Protein–Protein Interactions

193

5. After printing, the slides are baked at 80°C for 30 min (see Note 6). The printed slides are stored at RT in slide boxes with a pack of silica gel until use. 3.4. Microarray Printing: Quality Control

The printed arrays need to be quality-controlled before any of the printed slides can be used for protein expression and detection. A random sample of microarray slide was stained and then scanned so as to assess the print uniformity and quality. These checks include substrate defects and spots morphology (Fig. 1). Never touch the array with bare hands. Always use gloves and tweezers to manipulate the microarray slides. Hold the slides by their edges. 1. Block slide in 0.1% polyvinylpyrrolidone/0.05% Tween 20 buffer, 1 h, RT, at 400 rpm in a four chambered staining culture plate on an orbital shaker. 2. Transfer the slide to Hybridization solution containing 400 ml blocking buffer + 10 ml Cy3-Spot QC detector oligo. Incubate for 1 h at RT. 3. Wash 3 min, 10× SSC, 0.2% Sarkosyl. 4. Wash 2 min, 10× SSC. 5. Wash 2 min in 2× SSC. 6. Spin the slides in a centrifuge at 200 × g for a minute. 7. Scan the slide in Axon GenePix 400 Scanner.

3.5. In Situ Expression of Protein

Before the in situ expressions can be performed on the slide, the slides have to be blocked in the blocking solution for 1 h at RT with a gentle shaking at 40 rpm on an orbital shaker. The blocking can be carried out in a plastic four compartment slide container with 5 ml of blocking solution. Now, the slides are ready for expression. 1. Prepare on ice, the rabbit reticulocyte lysate mix. Remove the reagents from storage at −80°C. Rapidly thaw the TnT Quick

Fig. 1. Quality of microarray printing. Plasmid DNAs containing various proteins fused to Tus were printed using a MicroGrid arrayer. The microarray slide was hybridized for 1 h with Cy3-Spot QC detector oligo (IDT, Coralville). Finally, the slide was scanned with a GenePix (Axon) scanner to obtain the image. The image demonstrates the presence of DNA in almost equal intensity in each spot.

194

Sitaraman and Chatterjee

Master Mix by hand warming and place on ice. The other components can be thawed at RT. 2. To a sterile tube, for each slide to be tested, add 160 ml of lysate, 5 ml of 1 mM methionine and 40 mg/ml of query plasmid and make up the volume with nuclease free water to 200 ml. 3. After the blocking, remove as much of the blocking solution as possible, but don’t let the slides dry. 4. Place a dual well incubation chamber on top of the blocked and drained microarray slide so as to expose the printed area on one side of the incubation chamber and the other incubation chamber covers the print-free area. 5. Insert the whole set up, microarray slide and the incubation chamber, carefully into the Fast Frame slide holder. 6. Place the slide holder now containing the printed slide in a suitable plastic container with a lid. 7. Gently pipet the lysate mix containing the query plasmid dropwise onto the printed area of the slide contained inside one of the dual well chambers without introducing any air bubble. 8. Add 200 ml of water to the other print-free chamber (see Note 7). 9. Cover the plastic container with the lid and incubate the set up for 1 h 30 min at 30°C incubator (see Note 7). 10. After the expression, lower the temperature to 15°C and let it sit for another 1 h 30 min for the protein–protein interaction to take place (see Note 8). 11. Proceed with prehybridization.

Fig. 2. Expression of target protein and the detection of protein interactions on a Tus-Ter based microarray system. Four plasmid DNAs (negative controls without any ORF or with GFP and b-globin and a test plasmid Jun) were immobilized onto a nitrocellulose-coated FAST slide. The target proteins were expressed as His-tagged Tus fusion proteins on the slide using 200 ml rabbit reticulocyte lysate containing query plasmid encoding for c-Fos. We detected interaction of the gene-specific anti c-Fos antibody with the region where plasmid encoding Jun was printed.

Protein–Protein Interactions

195

Fig. 3. Rapamycin-mediated interaction of FRB and FKBP: The samples, negative control (GFP) and FRB plasmid, were printed in duplicates (a, b). Using two identical slides, the cell-free protein synthesis reaction was performed in the presence of query plasmid FKBP12 and 10 mM of Rapamycin. The interaction of FRB-FKBP complexes was detected only on the rapamycin-treated slide.

Fig. 4. Protein expression of target proteins and protein interaction: (a) Expression of seven different Tus fusion proteins and their detection by an unlabeled anti-His monoclonal antibody as primary and Cy3, Cy5 labeled goat anti-mouse IgG as secondary antibody. Identical slides printed with the same templates were used for detecting protein–protein interaction in (b, c). p53 was coexpressed along with MDM2 (query plasmid) in a cell-free reaction. (b) The p53 and MDM2 protein complex was detected with unlabeled anti-MDM2 mAB produced in mouse and Cy-labeled anti-mouse IgG. (c) p16 served as the query plasmid for the detection of CDK4 using an unlabeled anti-p16 rabbit mAB as primary, and Cy3 and Cy5-labeled anti-rabbit IgG as the secondary antibody. The images were scanned using a GenePix scanner. Note, we also unexpectedly detected that p16 binds to CDK2 as well.

3.6. Prehybridization and Hybridization

The hybridization is carried out in two steps by indirect labeling. Gene-specific query protein antibody was used as the primary antibody in the first step, and a Cy3 and Cy5-labeled anti-mouse or anti-rabbit IgG was used as the secondary antibody (Figs. 2–4).

196

Sitaraman and Chatterjee

1. When the incubation is about to end, prepare the primary antibody: monoclonal anti-gene specific in mouse or rabbit (see Note 9) at a concentration of 1:1,000 (i.e., 5 ml in 5 ml of hybridization buffer). 2. After TnT reaction is completed, dip and wash the slides briefly in 5 ml of PBS in a plastic slide container. 3. Incubate the slides in the primary antibody solution at 4°C, overnight. 4. In the meantime, prepare the labeled secondary antibody. 3.6.1. Secondary Antibody Labeling with Cy3 and Cy5 Dyes

In this step, proteins are labeled by covalent attachment of multiple fluorophores to each protein molecule. 1. Resuspend the lyophilized antibody to mouse or rabbit IgG in 990 ml of Extraction/Labeling buffer from ClonTech (see Note 10). 2. Label two tubes one for Cy3 and another for Cy5 (see Note 11). 3. Aliquot 90 ml of resuspended antibody in these two tubes. 4. Resuspend a fresh tube of Cy3 and Cy5 (see Note 12) in 110 ml of the same buffer. Mix thoroughly by vortexing. 5. Add 20 ml of Cy3 solution to 90 ml of protein sample. 6. Add 20 ml of Cy5 to the other 90 ml of protein sample. 7. Incubate the tubes on ice for 1 h 30 min. 8. Add 4 ml of Blocking buffer from Protein Extraction and Labeling kit (see Note 13). 9. Incubate for another 30 min on ice. Mix every 10 min. 10. Unused antibody can be stored at −20°C in light protected amber tubes, in aliquots, for future use (see Note 14).

3.6.2. Removal of Unbound Dye

Labeled antibody can be separated from the excess unconjugated dye by gel permeation chromatography. 1. Use an Illustra MicroSpin G-25 Chromatography column to remove unbound label as described below. Two columns/ samples are required. 2. Snap off the bottom tip and place the column in a collection tube. 3. Remove the screw cap and centrifuge for 2 min in a microcentrifuge at 100 × g to remove the packing buffer. Discard the buffer. 4. Wash once with 1× PBS, centrifuge, and discard the flow through. 5. Place the column in an amber 1.5 ml tube.

Protein–Protein Interactions

197

6. Carefully apply the sample (110 ml) directly to the center of the column and centrifuge for 4 min at 100 × g. 7. Load the flow through on a second column placed in a fresh amber centrifuge tube. 8. Centrifuge for 4 min at 100 × g. The labeled antibody is ready for use. 3.6.3. Secondary Antibody Hybridization

1. After the incubation in primary antibody, wash the slides 3× in PBS for 10 min each. Do not allow the arrays to dry out between washes. 2. Prepare the probe by mixing 50 ml of labeled antibody with Cy3 and 50 ml Cy5 labeled antibody in 4 ml of hybridization solution (see Note 15). 3. Hybridize the microarray in labeled secondary antibody for 2 h at RT. 4. Wash 3 times in 5 ml of washing solution for 5 min. 5. Wash once briefly in 1× PBS. 6. Spin the slides at 200 × g for a minute in a centrifuge to remove the excess liquid and dry with cold air. 7. The slides are ready for scanning. Store them at RT in a slide box with lid until ready to use.

3.7. Scanning

1. Slides were scanned on an Axon GenePix 4000 scanner (see Note 16). 2. Fluorescence data were collected and evaluated with the GenePix Pro 5.0 software. 3. Set the scanner at 100% laser power and PMT gain at ~300. 4. The results of protein–protein interactions using Jun/Fos, FRB/FKBP, CDK4/p16, and p53/MDM2 are shown in Figs. 2, 3, and 4, respectively. 5. As can be seen in Fig. 4, we have detected a previously unknown interaction between CDK2 with p16. Initially, CDK2 was chosen for a negative control. However, in our repeated experiments, we found the interaction between CDK2 and p16. Thus, we concluded that this new interaction is valid at least in our in vitro analysis. We are currently trying to validate this finding in an in vivo system (see Note 17).

4. Notes 1. If the DNA is in TE, it has to be desalted and resuspended in 3× SSC. SSC is the printing buffer that has been routinely used for DNA microarrays.

198

Sitaraman and Chatterjee

2. To get consistently pure and high concentration of DNA, we regularly use maxipreps. Column purified miniprep DNA without any RNAse contamination has also been used successfully. 3. The microarrayed template has a HIS 6-tag. So, to distinguish the protein–protein interaction, if any, a different tag (i.e., FLAG tag) is incorporated in query plasmid. Having different tags helps in assessing the protein expression by Western analysis as well. 4. It is critical to maintain the temperature at 30°C. Even with the slightest fluctuation (29 or 31°C) we have noticed considerable decrease in the amount of protein synthesis. 5. The optimal humidity for printing is 50–60%. If the relative humidity is below 40%, use a humidifier; and if above 70%, use a dehumidifier. Bring up the humidity inside the arrayer to 50–70% and maintain this humidity throughout the time of printing. 6. Instead of baking, double stranded target DNA on nitrocellulose microarray can be UV cross-linked at 150–300 mJ. If baking, place the arrays in a lidded glass container, and make sure that the oven is clean. 7. These steps help to maintain the humidity during the incubation. 8. The cell-free reaction is completed within 1 h 30 min at 30°C. But the incubation is continued at a lower temperature for another 1–2 h to allow “Tus-Ter” (DNA-protein) interaction as well as the interaction of target and query protein (protein–protein) to take place. 9. The detection on microarray can be performed in a single step by direct labeling of the anti-FLAG antibody. But we prefer two steps of indirect labeling for the detection of protein–protein interaction. Probing with the gene-specific antibody as the primary and labeled IgG as the secondary antibody gives better results. 10. Optimal labeling of the Cy-dyes occurs at pH 9.3. Sodium bicarbonate/sodium carbonate at a concentration of 0.1 M, pH 9.3 is generally used. The antibody to be labeled is also dissolved in the same buffer. 11. These dyes are considered to be potentially hazardous. Care should be exercised when handling the dyes. 12. Dissolve the Cy3 and Cy5 dyes in the same tube in which the dye is supplied. ~1 mg of protein can be labeled with the contents of each tube. 13. This step is needed to quench the excess unreacted dye. This is necessary to obtain consistent low background. 20 mM Tris, pH 7.5, 500 mM NaCl can also be used as the termination

Protein–Protein Interactions

199

buffer. Buffers containing primary amino groups such as Tris inhibit the conjugation reaction. 14. Rest of the labeled antibody can be stored at −20°C for later use. It is recommended to use freshly labeled antibody for microarray analysis, but we have used it successfully after 2 weeks of storage at −20°C. 15. Once the probe is made, the concentration can be measured, if needed. Try to use at least ~30–50 mg/ml of protein for each slide. 16. One can use any scanner that is compatible with 75 × 25 × 1 mm slides and capable of dual color analysis. The scanner must be capable of measuring Cy3 and Cy5 fluorescent labels. In general, fluorescent dyes are sensitive to light exposure as well as other environmental factors. In order to minimize these risk factors, it is advisable to scan the slides instantly upon finishing the last wash. 17. As stated above, we detected CDK2/p16 interaction in our repeated experiments. CDK2, a Ser/Thr protein kinase, is believed to be involved in the transition G1-S phase of the cell cycle. This new interaction of CDK2 with p16, a tumor suppressor, may suggest a new role of p16 to regulate the function of CDK2 during G1-S phase transition. Of course, the validation of this interaction in an in vivo system must be done.

Acknowledgments This project has been funded in whole or in part with federal funds from the National Cancer Institute, National Institute of Health, under N01-CO-12400. The content of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does the mention of trade names, commercial products, or organizations imply endorsement by U.S. Government. References 1. Panday A, Mann M (2000) Proteomics to study genes and genomes. Nature 405:837–846 2. MacBeath G (2002) Protein microarrays and proteomics. Nat Genet 32:526–532 3. Dietrich HR, Knoll J, van den Doel LR, van Dedem GW, Daran-Lapujade PA, van Vliet JL, Moerman R, Pronk JT, Young IT (2004) Nanoarrays: a method for performing enzymatic assays. Anal Chem 76:4112–4117

4. Ge H (2000) UPA, a universal protein array system for quantitative detection of proteinprotein, protein-DNA, protein-RNA and protein-ligand interactions. Nucleic Acids Res 28:e3 5. Arenkov P, Kukhtin A, Gemmell A, Voloshchuk S, Chupeeva V, Mirzabekov A (2000) Protein microchips: use for immunoassay and enzymatic reactions. Anal Biochem 278:123–131

200

Sitaraman and Chatterjee

6. MacBeath G, Schreiber SL (2000) Printing proteins as microarrays for high-throughput function determination. Science 289:1760–1763 7. Zhu H, Synder M (2003) Protein chip technology. Curr Opin Chem Biol 7:55–63 8. Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P, Lan N, Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P, Dean RA, Gerstein M, Snyder M (2001) Global analysis of protein activities using proteome chips. Science 293:2101–2105 9. Nord O, Uhlen M, Nygren PA (2003) Microbead display of proteins by cell-free expression of anchored DNA. J Biotechnol 106:1–13 10. He M, Wang M-W (2007) Arraying proteins by cell-free synthesis. Biol Mol Eng 24:375–380 11. Ramachandran N, Hainsworth E, Bhullar B, Eisenstein S, Rosen B, Lau AY, Walters JC,

12.

13.

14.

15.

LaBaer J (2004) Self-assembling protein microarrays. Science 305:86–90 Ramachandran N, Raphael JV, Haisworth E, Demirkan G, Fuentes MG, Rolfs A, Hu Y, LaBaer J (2008) Next-generation high-density self-assembling functional protein arrays. Nat Methods 5(6):535–538 Chatterjee DK, Sitaraman K, Baptista C, Hartley JL, Hill TM (2008) Protein microarray on-demand: a novel protein microarray system. PLoS One 3(9):1–5 Coskun-Ari FF, Hill TM (1997) Sequence specific interactions in the Tus-Ter complex and the effect of base pair substitution on arrest of DNA replication in Escherichia coli. J Biol Chem 272:26448–26456 Hartley JL, Temple GF, Brasch MA (2000) DNA cloning using in vitro site-specific recombination. Genome Res 10:1788–1795

Chapter 13 Kinase Substrate Interactions Michael G. Smith, Jason Ptacek, and Michael Snyder Abstract Kinases have become popular therapeutic targets primarily due to their integral role in cell cycle and tumor progression. The efficacy of high-throughput screening efforts is dependent on the development of high quality multiplex tools capable of replacing lower-throughput technologies such as mass spectroscopy or solution-based assays for the study of kinase–substrate interactions. Functional protein microarrays are comprised of thousands of immobilized proteins on glass slides that have been used successfully to identify protein–protein interactions. Here, we describe the application of functional protein microarrays for the identification of the phosphorylation targets of individual protein kinases using highly sensitive radioactive detection and robust informatics algorithms. Key words: Kinase, Protein microarray, Kinase–substrate interaction, Phosphorylation, Multiplex

1. Introduction Traditionally, investigations to understand protein function and protein–protein interactions focused on individual proteins. Recently, however, technologies to analyze proteins in a high-throughput and unbiased fashion have become feasible (1). Protein microarrays, which contain a high density of proteins, enable systematic query of biochemical activities (2–4). The high degree of multiplexing, low sample requirements, and speed of the assays make microarray technology advantageous for human disease research. There are two types of protein microarrays (3). A “functional protein microarray” contains a set of proteins individually produced and positioned in an addressable format on a microarray surface. Functional protein microarrays are useful for identifying binding activities or targets of posttranslational modification enzymes. Production of functional protein microarrays involves Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_13, © Springer Science+Business Media, LLC 2011

201

202

Smith, Ptacek, and Snyder

overexpressing, purifying, and spotting proteins onto the surface of glass slides (5–7). While large amounts of data have been generated using protein microarrays, there are significant challenges in developing robust methods to process the raw data and building reasonable biological hypotheses from the datasets. The second type of protein microarray, the “antibody microarray,” shares similarities with immunoassays and uses antibodies to detect specific probes. These arrays consist of antibodies with specific reactivity spotted in an addressable format. There are two subtypes of antibody arrays: those that make use of direct labels and those that require a second antibody for detection (8). Antibody microarrays require high quality, highly specific antibodies. The availability of such antibodies is a key limiting factor in the preparation of antibody microarrays. Here, we will discuss the use of functional protein microarrays to uncover protein kinase substrates. Protein kinases play a central regulatory role in many cellular and biochemical processes. The human genome encodes approximately 500 kinases, nearly half of which have been mapped to disease loci (9). It is not surprising, therefore, that many biotech and pharmaceutical companies are seeking to discover, develop, and commercialize compounds that can modulate the activity of specific kinases known to be involved in disease processes. Kinase effectors already in clinical use include Gleevec® (Novartis), Iressa® (Astra-Zeneca), Herceptin® (Genentech), Tarceva® (OSI Pharmaceuticals), Erbitux® (ImClone), Nexavar® (Onyx Pharmaceuticals), Sutent (Pfizer), and Tykerb® (GlaxoSmithKline). The biological roles of many kinases have yet to be elucidated, and even well-studied kinases still require further efforts in the identification of their complete range of substrates. Identifying proteins and other biomolecules that interact with a given kinase provides valuable insight into its activity, which can lead to the development of new and improved therapeutics for the treatment of human diseases. In order to identify kinase substrates using high-throughput methodologies, purify a kinase of interest, incubate the kinase on a microarray of spotted human proteins, and then detect the phosphorylated substrates using radiolabeled ATP. A detailed flowchart of how in vitro kinase substrates are identified using protein microarrays is shown in Fig. 1.

2. Materials 1. Purified protein kinase of interest (see section 3.1 Preparing the kinase for requirements), store on ice until use. 2. Spotted protein microarray (e.g., Protoarray® or Sigma’s Panorama®).

Life

Technologies’

Kinase Substrate Interactions

203

Fig. 1. A flowchart for identification of kinase substrates using protein microarrays.

3. [g33P]ATP (3,000 Ci/mmol, 10 mCi/ml). 4. 0.45-mm Filters (Millipore). 5. Clean, covered 4-chamber incubation tray (Greiner or ISC Bioexpress), chilled on ice. 6. Forceps and deionized water. 7. 50-ml Conical tubes. 8. Shaker set to 4°C (capable of circular shaking at 50 rpm). 9. Incubator set to 30°C. 10. 60 × 24-mm Glass coverslips (VWR; or with lifter, VWR). 11. X-ray film cassette and clear plastic wrap. 12. Microarray slide holder and centrifuge equipped with a plate holder (optional).

204

Smith, Ptacek, and Snyder

13. Phosphorimager (e.g., Packard Cyclone and Multisensitive Phosphor Screen (Perkin-Elmer). 14. X-ray film and developer. 15. Microarray data acquisition software (e.g., GenePix® Pro from Molecular Devices). 16. Data analysis software. 17. Blocking buffer (5 ml of buffer is needed for each microarray): 1× PBS, pH 7.4, 1% BSA. 18. Alternative blocking buffer: SuperBlock buffer in TBS (Pierce). 19. Kinase buffer (120 ml of buffer is needed for each microarray): 1% NP-40, 100 mM MOPS, pH 7.2, 100 mM NaCl, 1% BSA, 5 mM MgCl2, 5 mM MnCl2, 1 mM DTT. 20. Wash buffer (80 ml of buffer is needed for each microarray): 0.5% SDS in 10 mM Tris–HCl pH 7.4.

3. Methods Variability is the enemy of high-throughput discovery tools. Efforts to reduce variability will increase the quality of the data and ultimately reduce the workload required to validate candidate substrates. Common pitfalls that may increase variability between arrays include using older, decaying radiolabeled ATP, repeated freezing and thawing the kinase of interest or the functional protein microarrays, dust, bubbles under the coverslip, and temperature fluctuations throughout the experiment. Moreover, each operator can be considered a variable and the number of investigators who perform the experiments should be kept to a minimum. Lastly, analysis tools can also introduce variability. GenePix, for example, can introduce variability by inaccurately defining feature size. Since signal and background values are determined directly as a consequence of the delineation of the feature’s boundaries, care should be taken to ensure that each feature is accurately defined. The methods presented here offer a starting point from which to design substrate identification experiments. Tailoring the protocol, for example with kinase buffers with additional cofactors may be used with success. Lastly, there is an ever-expanding knowledge base in which to rank the hitlists for further validation studies. Most promising substrates, include those that interact with a protein that also interacts with the kinase of interest (e.g., a scaffold or adaptor protein), are in the same localization or functional category, or have a known phospho-motif for the kinase. However, the power of protein microarrays that contain thousands of proteins is that it provides an unbiased approach to sample the proteome, generating novel results that more limited or biased experiments would fail to deliver.

Kinase Substrate Interactions

3.1. Preparing the Kinase

205

You will need 120 ml Kinase Buffer containing your kinase of interest to probe one protein microarray. 1. Prepare a 120 ml dilution of your kinase of interest at a final concentration of 1–50 nM in Kinase Buffer. The optimal concentration for a given kinase must be determined empirically. Initial experiments using 10 and 50 nM of a kinase is recommended to evaluate signal and background intensities. 2. Mix well (do not vortex) and store on ice until use. Immediately return the remaining kinase to −80°C. Kinases lose activity with multiple freeze–thaws; store kinase in multiple aliquots.

3.2. Blocking

1. Allow the protein microarray to equilibrate at 4°C for at least 15 min before blocking. Not doing so may result in condensation on the array, which can reduce protein activity or alter spot morphology. 2. Place one microarray with the spotted protein side facing up into each well of a chilled 4-chamber incubation tray. Often, slides with barcodes are used for sample tracking and easier array handling. In these cases, ensure that the barcoded side of the microarray is facing up. The Quadri-PERM 4-chamber culture dish has an indented numeral at one end of the dish, which is useful not only for sample tracking but also for preventing the microarray from adhering to the bottom of the dish. 3. Using a sterile pipette, slowly add 5 ml Blocking Buffer into each chamber containing an array. Avoid pipetting buffer directly onto the array surface. 4. Incubate the tray for 1 h at 4°C on a shaker set at 50 rpm (circular shaking). 5. After incubation, remove the array from the 4-chamber incubation tray using forceps. Insert the tip of the forceps into the indentation at the numbered end of the tray and gently pry the array upward. Using a gloved hand, pick up the microarray by holding the array by its edges only. Gently remove the excess liquid from array surface by blotting the edge of the array on a paper towel. 6. Proceed immediately to the Probing procedure.

3.3. Probing

1. Place the protein microarray in a 50-ml conical tube with one third of the slide extending outside of the tube. The barcode end of the slide should be outside the tube, facing up. 2. Add 1 ml of [g33P]ATP (3,000 Ci/mmol, 10 mCi/ml) to 120 ml of Kinase Buffer containing diluted kinase (see Subheading 3.1) to obtain a final [g33P]ATP concentration of 33 nM for one microarray. For arrays being probed without

206

Smith, Ptacek, and Snyder

kinase as a negative control, add 1 ml of [g33P]ATP (3,000 Ci/ mmol, 10 mCi/ml) to 120 ml of Kinase Buffer without kinase. Once the ATP is added to the kinase, use the kinase-ATP mixture immediately for probing the array. Do not store the prepared kinase-ATP mixture on ice for more than 2 min prior to use on the array. 3. Pipet mixture gently onto the surface of the protein microarray within the conical tube taking care not to touch or disrupt the array surface. 4. Using forceps, carefully remove a coverslip from package and lay the coverslip on the array to cover the array without trapping any air bubbles. Align the coverslip flush with the top edge of the array to ensure that the printed area of the array is completely covered. Gently adjust the coverslip to remove any air bubbles. 5. Gently slide each array with a coverslip into the conical tube with the printed side (barcode) of the array facing up. Securely fasten a cap onto the conical tube. 6. Place the conical tube horizontally on a flat surface in an incubator set to 30°C with the printed side of the array facing up and the tube as level as possible. If needed, tape the tube to the flat surface to avoid any accidental disturbances. Incubators with wire rack inserts can be useful for properly positioning the conical tube such that the array within is as flat as possible. 7. Incubate the conical tube containing the array for 1 h at 30°C without shaking. 8. Prepare 80 ml of 0.5% SDS Wash Buffer for each array. Remove the conical tube containing the array from incubator and add 40 ml of 0.5% SDS to the tube by dispensing the buffer down the sides of the tube. Avoid pipetting buffer directly onto the array surface. Incubate the array in 0.5% SDS for 1 min at room temperature. Gently move the array in the tube to dislodge the coverslip. Do not remove the coverslip with forceps if the coverslip is not dislodged from the array. Using forceps, carefully remove the dislodged coverslip without touching the array surface. Discard the coverslip appropriately as radioactive waste into the 33P waste bottle. 9. Cap tube and incubate at room temperature for 15 min without shaking. Discard the 0.5% SDS wash as radioactive waste. 10. Repeat the washing step by adding an additional 40 ml 0.5% SDS to the tube, cap the tube, and incubate for 15 min without shaking. Discard the wash as radioactive waste. 11. Add 40 ml of water to the tube and incubate for 15 min at room temperature without shaking. Discard the water wash as radioactive waste, and repeat the wash a second time.

Kinase Substrate Interactions

207

12. Remove the array from the tube using forceps and place in a slide holder. 13. Proceed immediately to drying and scanning the microarray. 3.4. Drying and Scanning the Microarray

1. Dry the array using a table top centrifuge. Spin the array at 800 × g for 3–5 min at room temperature in the slide holder (if using a centrifuge equipped with a plate rotor) or 50 ml conical tube (if using a swinging bucket rotor). Verify that the array is completely dry. 2. Place array in X-ray film cassette, cover with plastic wrap, and overlay with phosphor screen or X-ray film. You can check the radioactivity on the array using a Geiger counter. 3. Expose array to phosphor screen or X-ray film for 3 h being careful not to bump the cassette during this time. Exposure times may need to be adjusted depending on the activity of the kinase (3–24 h or more). 4. Remove phosphor screen from cassette and scan with phosphorimager or develop film using film developer. 5. Obtain 16-bit TIFF image file by scanning X-ray film with scanner or retrieving file from phosphorimaging of phosphor screen. 6. Process image using imaging software (i.e., Prospector Imager or Adobe® Photoshop®). For Prospector Imager, refer to the ProtoArray® Prospector User Guide. For Adobe® Photoshop® process the image as follows: (a) Crop 1″ × 3″ fixed rectangular areas from each TIFF file that correspond to each slide. (b) (optional) Invert the image. (c) Change image file to 2,550 × 7,650 pixels (constrained proportions). (d) Save cropped TIFF image with new name. Note: Do not adjust pixel levels of the file in Adobe® Photoshop® as this will negatively impact the dynamic range of the image. 7. Proceed to Data Acquisition and Analysis.

3.5. Data Acquisition and Analysis

For data acquisition, specific protein array information is required. Often, a .gal (GenePix Array List) file is used to describe the location and identity of all the features on a protein microarray. The microarray data acquisition software can then generate files containing pixel intensity information for all of the features on the array. Other software packages and array lists can be used, but for simplicity, only GenePix will be discussed here. Figure 2 shows an example of a phosphorylated protein array alone and with a grid applied for feature analysis.

208

Smith, Ptacek, and Snyder

Fig. 2. The phosphorylated, radiolabeled protein microarray is exposed to film or a phosphor screen and an image scan of the slide is generated. In GenePix® Pro, an array list is overlayed onto the slide image identifying the features on the array. Each subarray block is then checked for proper placement using landmark control spots of autophosphorylating kinases to guide placement. GenePix® Pro uses the grid to determine signal and background intensity for all the features of the array.

Kinase Substrate Interactions

209

1. Start the GenePix® Pro microarray data acquisition software. Open the saved image (.tif) and load the appropriate array list (.gal file) for the protein microarray. 2. Adjust the subarray grid to ensure that the grid is properly aligned with the spotted protein features in each subarray. It is necessary to have landmark control proteins to align the array; in Fig. 2 there are three pairs of landmark kinases that autophosphorylate and allow each subarray grid’s position to be triangulated precisely. 3. After the grid is properly adjusted and all features are aligned, acquire the signal and background intensity data for each feature by clicking the Analyze button in GenePix® Pro, and save/export the results as a .gpr (GenePix® Results) file.

4. Notes 1. In order to identify kinase substrates, a single active kinase must be purified to homogeneity, incubated with radiolabeled ATP, and applied to a functional protein microarray. 2. The recommended protein concentration for probing a microarray is wide, ranging from 1 to 100 nM. The activity of each kinase will directly determine the working concentration necessary for probing an array. 3. Many kinases are capable of autophosphorylation, and the level of kinase autophosphorylation also influences the required concentration. Too much or too active a kinase may result in high background and a dark array. Too little kinase will result in little useable data. 4. If purifying your own kinase, purify the kinase under native conditions to >90% pure. 5. Ensure that the kinase is active in any one of the number of alternative assays. 6. Make sure that the protein kinase is soluble and active in the buffers used for probing the microarray (see recipes in section 2. Materials); if they are not then you may need to revise your buffer selection. 7. Since many of the proteins that are spotted on the microarray will readily bind ATP, including kinases which possess autophosphorylation ability, a negative control in which radiolabeled ATP is applied to an array in the absence of kinase is essential for proper interpretation of the results. 8. Since each purified kinase will have different activity, a titration of kinase concentration is strongly recommended to determine the optimal working concentration for your kinase.

210

Smith, Ptacek, and Snyder

9. At this point, the data is ready to be analyzed. For each feature spotted onto the microarray, a .gpr file will return a range of data points, including the signal and background values that will be used to determine your final list of substrates, or hitlist. 10. In GenePix, signal and background values are returned as the median of the pixels within a feature or the mean of the pixels within a feature. The median is the preferred metric as it is less susceptible to outliers in the data that can be seen if dust or other contamination ends up on the array during scanning. 11. Background is determined for each feature as a local value, typically 2 or 3 pixels from the features limits. This method of determining signal and background intensities has its limitations, particularly if the slide has uneven signals across the slide. This can occur due to uneven probing or inadequate mixing during the incubation. More problematic can be individual spotted kinases whose signals are so strong that they bleed into neighboring features causing an increased background in the neighboring regions such that genuine signals can be mistakenly subtracted out. Likewise, this is the reason for using lower energy [g33P]ATP and not [g32P]ATP. 12. One method, called ProCAT, to account for these irregularities was developed by Zhu et al. (10). ProCAT employs a multi-step approach to affect background correction, signal normalization, positive spot identification, feature crossreactivity, signal quality inspection, and protein amount normalization. For background correction, rather than assigning the background values as the signal surrounding a particular feature, ProCAT determines the background for the eight features surrounding the feature of interest (using a 3 × 3 window) and assigns the background as the median value from all nine. In this way, local variations can be diminished, and signals can be accentuated. 13. Signal normalization is performed by using a sliding window across the slide within which median signals and median absolute deviations are calculated. These values are then used to correct each feature’s signal such that uneven signal distributions across the slide are reduced. 14. Similarly, positive spot identification also uses a sliding window approach. The default setting analyzes a 9 × 9 window, and those features with signals greater than 2 standard deviations above the local mean are assigned as a hit. Two different filters are then employed. 15. The feature cross-reactivity filter determines which features are deemed positive hits in negative control experiments and removes them from the final hitlist. This removes chip features that bind ATP or are autophosphorylated.

Kinase Substrate Interactions

211

16. The signal quality filter examines the reactivity of duplicate spots, and those features with wide variations between duplicate spots are removed from analysis. This filter clearly assumes that the protein microarray that is used for these experiments contains each feature printed in duplicate. 17. Lastly, signals are adjusted by the amount of protein that is spotted in each feature. This can be a very powerful filter in that often the brightest signals are returned from those features spotted at the highest concentration. However, accurate assessment of the actual spotted amount is often difficult to ascertain. Nonetheless, relative protein abundance can be used to adjust the signals. 18. Alternatives to ProCAT exist that identify kinase substrates by their signal-to-noise ratios both within arrays and by comparison with negative control arrays. Like ProCAT, a series of metrics are generated and features are filtered based on their rank within these filters. 19. The first metric is the signal-to-noise ratio, or Z-factor (11). Signals and standard deviations for each feature are compared to the signals and standard deviations for negative controls. If the signals are high and deviations are low relative to the negative controls, then the Z-factor value will approach a value of 1. Low signals and high deviations will drive Z-factor values toward zero. As a rule of thumb, features with Z-factor values above 0.5 can be considered as hits.

20. Z − factor = 1 − [3 × (s (feature of interest) + s (negative controls))]/ | m(feature of interest) − m(negative controls) | . 21. Another metric is termed the Z-score, which is the number of standard deviations, a features’ signal is from the mean of the entire array. A feature with a Z-score above 3 can be considered a positive. 22. Z − score = [Signal(feature of interest) − m(all features)] / s (all features).

23. Lastly, coefficients of variation (CV) are determined for each pair of duplicate spots. Only those features with low interspot CV will be considered for a final hitlist. 24. Filtering the data against values obtained with a negative control can help reduce false positive calls. Any feature that has a Z-factor score greater than 0.5 or a Z-score value greater than 3 on a negative control slide can be removed from analysis as they represent those proteins that either bind ATP directly or are kinases with robust autophosphorylation capacity.

212

Smith, Ptacek, and Snyder

25. The absolute values of these cut-off values can be adjusted based on the length of the hitlists returned; higher stringency will reduce hitlist length and lower stringency will elongate the candidate substrate list. References 1. Gershon D (2003) Proteomics technologies: probing the proteome. Nature 424:581–587 2. MacBeath G, Schreiber SL (2000) Printing proteins as microarrays for high-throughput function determination. Science 289:1760–1763 3. Phizicky E, Bastiaens PI, Zhu H, Snyder M, Fields S (2003) Protein analysis on a proteomic scale. Nature 422:208–215 4. Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P, Lan N, Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P, Dean RA, Gerstein M, Snyder M (2001) Global analysis of protein activities using proteome chips. Science 293:2101–2105 5. Ptacek J, Devgan G, Michaud G, Zhu H, Zhu X, Fasolo J, Guo H, Jona G, Breitkreutz A, Sopko R, McCartney RR, Schmidt MC, Rachidi N, Lee SJ, Mah AS, Meng L, Stark MJ, Stern DF, De Virgilio C, Tyers M, Andrews B, Gerstein M, Schweitzer B, Predki PF, Snyder M (2005) Global analysis of protein phosphorylation in yeast. Nature 438:679–684 6. Gelperin DM, White MA, Wilkinson ML, Kon Y, Kung LA, Wise KJ, Lopez-Hoyo N, Jiang L, Piccirillo S, Yu H, Gerstein M,

7.

8. 9.

10. 11.

Dumont ME, Phizicky EM, Snyder M, Grayhack EJ (2005) Biochemical and genetic analysis of the yeast proteome with a movable ORF collection. Genes Dev 19:2816–2826 Schweitzer B, Predki P, Snyder M (2003) Microarrays to characterize protein interactions on a whole-proteome scale. Proteomics 11:2190–2199 Haab B (2003) Methods and applications of antibody microarrays in cancer research. Proteomics 11:2116–2122 Manning G, Whyte DB, Martiniez R, Hunter T, Sudarsanam S (2002) The protein kinase complement of the human genome. Science 298:1912–1934 Zhu X, Gerstein M, Snyder M (2006) ProCAT: a data analysis approach for protein microarrays. Genome Biol 7:R110 Zhang JH, Chung TD, Oldenburg KR (2000) Confirmation of primary active substances from high throughput screening of chemical and biological populations: a statistical approach and practical considerations. J Comb Chem 3:258–265

Chapter 14 A Functional Protein Microarray Approach to Characterizing Posttranslational Modifications on Lysine Residues Jun Seop Jeong, Hee-Sool Rho, and Heng Zhu Abstract Functional protein microarrays offer a versatile platform to address diverse biological questions. Printing individually purified proteins in a spatially addressable format makes it straightforward to investigating binary interactions. To connect substrates to their upstream modifying enzymes, such as kinases, ubiqutin (Ub) ligases, SUMOylation E3 ligases, and acetyltransferases, is an especially daunting task using traditional methodologies. In recent years, regulation via various types of posttranslational modifications (PTMs) on lysine residues is emerging as an important mechanism(s) underlining diverse biological processes. Our group has been developing and applying functional protein microarrays constructed for different model organisms to globally identify enzyme–substrate interactions with a focus on lysine PTMs. In particular, we have characterized the pleiotropic functions of a ubiquitin E3 ligase, Rsp5, via identification of its downstream substrates using a yeast proteome chip. Also, we have identified nonhistone substrates of the acetyltransferase NuA4 complex in yeast, and revealed that reversible acetylation on a metabolic enzyme affects a glucose metabolism and contributes to life span. In this chapter, we will provide detailed protocols for the investigation of ubiquitylation and acetylation. These protocols are generally applicable for different organisms. Key words: Protein microarray, Posttranslational modification, Ubiquitylation, Ubiquitin E3 ligase, Rsp5, Acetylation, The NuA4 complex, Lysine residues

1. Introduction Functional protein microarrays by definition are constructed by spotting down hundreds of thousands of individually purified proteins in high density on a solid surface (1–4). Unlike the mass spectrometry technologies, in which en masse affinity-purified proteins are subjected to analysis, functional protein microarrays provide information about direct biochemical and physical interactions among biomolecules. For example, protein microarrays

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_14, © Springer Science+Business Media, LLC 2011

213

214

Jeong, Rho, and Zhu

have been used to profile protein–protein, –lipid, –DNA, –RNA, and –small molecule interactions (4–8). In addition to these binding assays, protein microarrays offer a unique platform to profile covalent modifications. Posttranslational modification (PTM) is an important regulatory mechanism in eukaryotes. While phosphorylation on serine, threonine, and tyrosine residues is well characterized (9, 10), modifications on lysine residues such as ubiquitylation, acetylation, mono-, di- and tri-methylation, SUMOylation, and neddylation are now emerging as important PTMs that are involved in many aspects of cellular functions. Our group is focusing on these mutually exclusive lysine PTMs to better understand the physiological roles and interplay of different signaling pathways. For instance, we and our colleagues set out to identify potential substrates of the yeast NuA4 acetyltransferase complex via covalent acetylation reactions on a yeast proteome chip that contains >5,800 yeast proteins in full-length (11). We identified and validated many nonhistone substrates and by performing in-depth studies on one substrate Pck1, a PEP carboxylase kinase in the cytosol, we found that the acetylation status of Pck1 regulates its enzymatic activity and contributes to longer life span for yeast under nutrient-deprived conditions. This surprising discovery illustrates the power of unbiased, global screening using functional protein microarrays in addressing important biological questions. Ubiquitylation mediates a diverse array of biological functions (Fig. 1). The attachment of ubiquitin (Ub) and poly-Ub chains to a substrate is mediated by a sequential action of E1-activating enzyme, E2-conjugating enzyme, and E3 ligase (12). E3 ligases are broadly classified into RING (really interesting new gene) and HECT (homologous to E6-associated protein C-terminus) domain-containing ligases. There is a variation in RING domains as well (13). There are seven lysine residues in Ub

Fig. 1. Physiological roles of mono- and poly-ubiquitylation in cells. Ubiquitylation of a protein results in the addition of single ubiquitin or multiple ubiquitins as a chain. Ubiquitin chains formed on individual lysine residues (K48, K63, or K29) on a ubiquitin mediate different cellular functions such as proteasome-mediated degradation, DNA repair, and intracellular trafficking.

A Functional Protein Microarray Approach

215

and all of them are shown to form poly-Ub chains in vivo (14). Recent findings indicate that the terminal amine is also a site for ubiquitylation, totaling eight modification sites (12). Among these modifications, the most well-characterized example is the poly-Ub chain formed via K48 linkage, which mediates proteasome-dependent degradation. The K63 chains are involved in DNA repair and intracellular trafficking. Recently, we and another group separately characterized a HECT domain E3 ligase, Rsp5, in yeast using the protein microarray approach (15, 16). Rsp5 belongs to the Nedd4 E3 family. Unlike many RING domain E3 ligases, HECT domain E3 ligases form an intermediate covalent bond with Ub on Cys residues. Rsp5 is highly conserved from yeast to humans, and defects in the human ortholog, Nedd4, can cause congenital disorder. In our study, we predicted previously unknown function of Rsp5 based on the known function of in vivo validated substrates and applied various in vivo assays to confirm the prediction. For example, we predicted that Rsp5 may function in the DNA-damage response via the RNR complex because one of the essential components of the RNR complex, Rnr2, was validated as a bona fide substrate of Rsp5. To demonstrate Rsp5’s new role in DNA damage response, we showed that only upon low dose of hydroxyurea treatment, which is known to target the RNR complex in yeast, did the Rsp5 temperature sensitive strain show growth defect at semipermissive temperature. Further characterization showed that in fact Rnr2 subcellular localization is dependent on the Rsp5 ubiquitylation (15). Construction of functional protein microarrays at high density poses a challenge in every step. Protein samples are usually prepared in a buffer containing ~30% glycerol to prevent evaporation and to ensure protein stability. To fabricate high-density arrays, printing is performed in under a low relative humidity of ~30%. We found that ambient humidity affects printing quality. We are currently using NanoPrinter LM210 (ArrayIt, Inc), which utilizes an enclosed chamber to precisely control the relative humidity inside, a near ideal situation to avoid spotmerging problems on glass. We also found that pilot studies are always helpful to determine printing buffers to achieve high density. Some previous works described general procedures for generation of protein samples and printing (4, 17). Here, we describe detailed protocols for establishing ubiquitylation and acetylation reactions on protein chips. The surface chemistry for printing has to be decided by pilot studies. In our hands, we found noticeable differences in performance for particular applications. For general protein–protein interaction and antibody-based assays, the FullMoon glass surface (FullMoon BioSystems) appears to be the most widely applicable. For HAT (histone acetyltransferase) assay

216

Jeong, Rho, and Zhu

Fig. 2. Acetylation assays on protein chips. X-ray film image of radioisotope detection and comparison of signal intensity and specificity affected by different surface chemistries. An overview of detection of HAT substrate by X-ray film. Landmarks such as histone proteins are visible in individual blocks (top panel). FAST (nitrocellulose-coated), Ni-NTA (tethered with nickel), and FullMoon slides (unknown surface chemistry) were compared for different HAT enzymes and substrate specificity (lower panel). Histone H3 and H4 were used as positive controls, whereas BSA as a negative control H4 is a preferred substrate of NuA4 while H3 is preferred by a different HAT enzyme, SAGA. The substrate specificity is illustrated by the intensity of the signals.

with radio-isotopes, we found that the FAST surface (Whatman) resulted in the highest signal compared to FullMoon, Ni/Cu surface, hydrogel, PVDF, and others (Fig. 2). And again, pilot studies with known biochemical conditions should determine the appropriate surfaces for real experiments.

2. Materials 2.1. Equipment

1. Bench-top centrifuge (Thermo Multifuge 3SR+ centrifuge).

Scientific:

HERAEUS

2. Microarray printer (BioRad: VersArrayer ChipWriter™ Pro System or ArrayIt Corporation: NanoPrinter LM210). 3. Microarray scanner (Molecular Devices: GenePixPro 4000B). 2.2. Protein Microarray Printing

1. 384-well plate (Whatman: 7701-5101). 2. Full Moon protein slides (Bar-coded) (FULL MOON BioSystems: PRT 50B).

A Functional Protein Microarray Approach

217

3. FAST slides (Whatman: 10486111A, Nitrocellulose Coated Slides). 4. Printing pins (ArrayIt Corporation: SMP3). 2.3. Ubiquitylation Reaction

The Ub E1 and E2 enzymes, ubiquitin and its modified forms are available commercially (Boston Biochem, Inc.). In most cases, the E3 ligases have to be purified in a recombinant form in labs following standard protocols. Where applicable, catalog numbers of some ubiquitin products from Boston Biochem are listed in the brackets below. 1. Lifterslip (Thermo Scientific: 25× 60I-2-4789). 2. E1 enzyme: UBE1 (U-100). 3. E2 enzyme: Ubc4 or UbcH5. 4. Ubiquitins (wild-type ubiquitin: U-100H, FLAG-ubiquitin: U-120, myc-ubiquitin: U-115). 5. 4-Well dish (Nalgen Nunc International: 267061). 6. Reaction Buffer: 25 mM Tris–HCl at pH 7.6 containing 50 mM NaCl, 10 mM MgCl2, 4 mM ATP, and 0.5 mM DTT (ATP and DTT are added just before use). 7. Antibodies: rabbit anti-GST (Millipore: AB3282), mouse monoclonal anti-FLAG (Sigma: F7425), goat anti-mouse antibody (Alexafluor 647) (Invitrogen: A21235), goat antirabbit antibody (Alexafluor 555) (Invitrogen: A21429) (see Note 1).

2.4. Protein Acetylation Reaction

1. SuperBlock blocking buffer in TBS (Thermo Scientific: 37535). 2. Reaction buffer (5×): 250 mM Tris–HCl, pH 7.5; 25% glycerol, 0.5 mM EDTA, 250 mM KCl, 5 mM DTT (200×), 0.1 nM Trichostatin A (TSA) (83×), 5 mM PMSF (40×), 5 mM nicotinamide (200×). DTT, TSA, nicotinamide, and PMSF are added just before use. For purified enzymes, HDAC inhibitors (TSA and nicotinamide) may be omitted. 3. 14C-Acetyl CoA (Amersham; CFA729; 50 mCi/mL): 0.5 mL. 4. Kodak BioMax MR Film (8701302).

3. Methods 3.1. Printing Protein Microarrays

Printing protein microarrays is quite challenging even though similar equipments to the DNA microarray printing is used in general. Protein storage buffer contains glycerol and some amount of detergents, but both may work against good spot morphology.

218

Jeong, Rho, and Zhu

For example, excess amount of detergents may lower the surface tension of droplets deposited on the glass surface, resulting in spread-out and merged features. Glycerol may absorb ambient water, which may create merged features as well in high-density printing (2). Thus, formulation of solutions for printing requires that several aspects be considered. To reduce the merged features, printing is done at lower humidity (~30%). We have used both the VersArrayer (BioRad) and NanoPrinter™ (ArrayIt) microarrayers to print protein microarrays. The NanoPrinter has a builtin humidity monitor and dehumidifier. Printing pins are an integral part of the printing process. We use SMP3 or 946MP2 pins from ArrayIt for routine printing. 1. Adjust the relative humidity to 30% by turning on dehumidifier or other dehumidifying equipment. 2. Meanwhile, set a printing program. The most important factor in high-density printing is spacing between features. For yeast printing, we use SMP3 pins from ArrayIt with column and row spacing 256 and 240 mm, respectively. After setting the parameters for printing, try a dry run without loading the pins to make sure that the program runs as expected. 3. Load pins into the print head, and make sure that the pins move freely by gravity. As a final check for the printing program and pins, load a 30% glycerol plate, and print a couple of slides. Make sure that each pin leaves glycerol spots on the blotting or preprinted slides. 4. Load slides into each slot in the microarray printer, and wait until humidity drops to 30% before starting printing. 5. After the printing is done, let the slides sit or cure in the printing station for at least 8 h to overnight so that protein binding is maximal. After the curing, the protein microarray is ready to use. Cured protein microarrays are stored in a −80°C deep freezer. 3.2. Ubiquitylation Reaction on a Protein Microarray

In general, the same reaction mixture for an in vitro reaction is prepared for the reaction on a chip. To reduce the amount of reagents, the reaction is routinely performed using lifterslip. To cover the entire printed area, we use 26 × 60 mm lifterslip. The space between the lifterslip and glass surface can hold 100 mL of volume. We prepare 100–120 mL of reaction mix and distribute the reaction mixture evenly on the printed area. The reaction setup is schematically shown in Fig. 3. The amount of E1, E2, and E3 enzymes used may be determined in vitro before embarking on the chip experiment using known substrates. Identifying physiological E2 and E3 pairs is not trivial. If the physiological E2–E3 pair is known, you may follow the protocols in published literature. However, if the pairing is not known, you may set up

A Functional Protein Microarray Approach

219

Fig. 3. Design of ubiquitylation reactions to identify substrates of the E3 ligase, Rsp5. Right : A mixture of ubiquitin E1/E2/E3 and ubiquitin with ATP in a ubiquitylation reaction buffer is incubated on a yeast proteome chip blocked with 1% BSA. After reactions, the chip is washed under denaturing condition to remove any nonspecific binding signals. To detect ubiquitylated substrates on the chip, anti-Ub antibodies are used. As a negative control, a reaction mixture without adding the E3 ligase is incubated on the chip, which goes through the same procedure. To identify E3 ligase-dependent hits, signals obtained from the E3 ligase reactions are compared with those from the negative control. GST signals, which serve as a surrogate of protein concentration on the chips, are used to further normalize the signals. This step, however, might not be necessary for data analysis because we found that it did not improve the results significantly.

reactions with a panel of E2 enzymes with known substrates. E2 enzymes are important players in determining chain length and types. In general, E1 and E2 enzymes are robust and work in different ranges of concentration. As a starting point, 100 nM of E1 and E2 would be appropriate. For multiple reactions, we routinely use 4-well dishes. 1. Take out chips from deep freezer and plunge directly into TBST at room temperature (RT). 2. Replace the TBST with TBST containing 5% skim milk or 3% BSA for blocking (see Note 2). 3. Incubate the slide on an orbital shaker for 1 h at RT. 4. Wash the slide with TBST 3 times. 5. Equilibrate the slide with 1× reaction buffer without ATP by changing the buffer 3 times. 6. Meanwhile, prepare the reaction mixture without ATP by adding a predetermined amount of E1, E2, E3, and Ub. Keep the reaction mixture on ice. Right before the reaction, add ATP and mix well.

220

Jeong, Rho, and Zhu

7. Take out the slide from the buffer using a pair of forceps, and remove excessive buffer by gently tapping the corner of the slide on lint free paper. It is important not to dry the slide completely or this will result in a smeared final scanned image. Apply the reaction mixture containing all the components on the slide without touching the surface. Gently place the lifterslip on the slide. Make sure no bubble is trapped. If you find bubbles, gentle tapping will help move the bubbles toward the outside of the printed area. 8. Incubate the reaction at RT for 1 h in a humidified chamber (homemade). 9. After the reaction, the lifterslip is removed by flooding with wash buffer. To remove noncovalently bound Ub, TBST with SDS (1%) may be used to wash the slides. 10. Subsequent washing is done by using TBST for 3 times, 5-min each (see Note 3). 11. Meanwhile, Antibody is prepared in TBST 5% milk (rabbit anti-GST and mouse anti-FLAG). 12. Incubate the slides with a mix with primary antibodies for 1 h with gentle shaking. 13. Wash with TBST for 3 times, 5 min each. 14. Incubate with a mix with secondary antibodies labeled with fluorophores (goat anti-rabbit Alexafluor 555 and goat antimouse Alexafluor 647). 15. Incubate for 1 h with gentle shaking. 16. Wash the slide with TBST for 10 min, 3 times. 17. Rinse the slides with distilled water (see Note 4). 18. The slides are spun dry by centrifugation at 712 × g for 3 min. 19. Scan the image. 3.3. Acetylation Assay on a Chip

Acetylated proteins are detected by radiolabeled 14C-Acetyl CoA as donor reagents. Although fluorescence-based detection is superior to radioactive-based detection, the quality of pananti-Ac-lysine antibodies is not good enough for the detection. One main advantage of using radioisotopes as the labeling reagent is that it allows detection of de novo, modifications by the enzyme in the reaction mixture. In this section, we describe radioactivebased method. 1. Take out slides from deep freezer, directly plunge into blocking solution as described above, and incubate for 1 h at RT with gentle rotation. 2. Briefly wash the slides with PBST.

A Functional Protein Microarray Approach

221

3. Equilibrate the slides by washing 2 times with 1× HAT buffer (without DTT, HDAC inhibitors or PMSF) at RT for 5 min each. 4. Meanwhile, assemble the reaction mixture. 5× HAT buffer: 20 mL C-Acetyl CoA (Amersham; CFA729; 50 mCi/mL): 0.5 mL

14

1 M Sodium butyrate: 5 mL 100 mM PMSF: 5 mL HAT: 20 mg (see Note 5) ddH2O to: 100 mL In the control reaction, HAT enzyme is left out. 5. Excess buffer is removed by gently tapping one side of the slide on a paper towel, and apply the reaction mixture evenly across the surface. 6. Using forceps, place lifterslip on the slide. Be careful not to introduce any bubbles. Incubate at 30°C for 1 h. 7. Wash the slide by flooding the slides 3 times with 50 mM NaHCO3–Na2CO3 (pH 9.3). 8. Rinse the slide with PBS. 9. Spin at 712 × g for 3 min to dry the slide. Make sure that slides are completely dried. 10. Place the slide in an X-ray cassette, and put X-ray film on the slides with a direct contact. Expose the film as needed. It usually takes 2 or more weeks. 11. Develop the film. The film is scanned using an office scanner, and the image is processed using Photoshop as follows. 12. Set a HP Scanjet 8300 scanner at the high sensitivity setting with a resolution of 4,800 dpi, and obtain chip images from the autoradiograph film. 13. Process the captured image of the autoradiograph film, and crop the image size as 6.4 × 4.4 cm and format as jpg, after magnification of image. 14. Open the image in adobe Photoshop and click the image in order to edit the image. 15. Edit the image with grayscale in 16-bit/channel mode. 16. Invert image (under Adjustments tab). 17. Compress the image to 53.3% with percentage in image size. 18. Save the image in TIFF format for further analysis. 19. Open the image in GenePixPro 6.0 and load the gal file containing each protein’s ID.

222

Jeong, Rho, and Zhu

20. Align the setting with gal file based on each spot from anti-GST profiling and positive control sets. 21. Adjust the size of the grid based on sizes or form of signals on chip. 22. Quantify the signal intensities using GenePixPro 6.0 by calculating the background and foreground signals on each spot.

4. Notes 1. The choice of detection antibody is up to the individual researchers. If a tagged antibody (FLAG- or HA-tagged Ub) is used, you will detect the Ub that are added during the reaction. However, if anti-Ub is used, signal will be from de novo as well as endogenous Ub. When choosing fluorescent dye, Cy5 or its equivalent Alexafluor 647 is preferred to Cy3 in general. Nitrocellulose-coated slides have an intrinsic autofluoresence when scanned in the Cy3 channel. Also, some protein samples emit autofluorescence at Cy3 channel, which makes it hard to discern specific and nonspecific interaction of probes. 2. Depending on the application, blocking can be done using 5% milk in TBST or PBST instead of BSA. In our hands, blocking with milk generates a better background to signal ratio. 3. During the washing, lift the sides of the slides alternatively with gentle shaking. Due to the flatness of the slides and chamber, small amount of previous samples remain in the space between the slides and bottom surface. By lifting the slide briefly, you can get rid of the reaction mix efficiently. 4. A brief rinse with distilled water is essential for a clean slide image. An incomplete rinse will result in smearing and watermarks on the scanned image due to residual salts. 5. For purified protein complexes from mammalian cells, it is advised to include HDAC inhibitors and protease inhibitors. If a purified enzyme is used, HDAC inhibitor and protease inhibitors can be omitted by adjusting the volume of each component as necessary. The amount of enzyme to be used is determined by an in vitro assay or pilot experiment with known substrate.

Acknowledgments We thank Dr. Wendy Yap for critical comments and editing. This work is in part supported by the NIH.

A Functional Protein Microarray Approach

223

References 1. Chen CS, Korobkova E, Chen H, Zhu J, Jian X, Tao SC, He C, Zhu H (2008) A proteome chip approach reveals new DNA damage recognition activities in Escherichia coli. Nat Methods 5:69–74 2. Paul B, Michael S (2005) Advances in functional protein microarray technology. FEBS J 272:5400–5411 3. Tao SC, Chen CS, Zhu H (2007) Application of protein microarray technology. Comb Chem High Throughput Screen 10:706–718 4. Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P, Lan N, Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P, Dean RA, Gerstein M, Snyder M (2001) Global analysis of protein activities using proteome chips. Science 293:2101–2105 5. Hu S, Xie Z, Onishi A, Yu X, Jiang L, Lin J, Rho H-S, Woodard C, Wang H, Jeong J-S, Long S, He X, Wade H, Blackshaw S, Qian J, Zhu H (2009) Profiling the human proteinDNA interactome reveals ERK2 as a transcriptional repressor of interferon signaling. Cell 139:610–622 6. MacBeath G, Schreiber SL (2000) Printing proteins as microarrays for high-throughput function determination. Science 289:1760–1763 7. Huang J, Zhu H, Haggarty SJ, Spring DR, Hwang H, Jin F, Snyder M, Schreiber SL (2004) Finding new components of the target of rapamycin (TOR) signaling network through chemical genetics and proteome chips. Proc Natl Acad Sci USA 101: 16594–16599 8. Zhu J, Gopinath K, Murali A, Yi G, Hayward SD, Zhu H, Kao C (2007) RNA-binding proteins that inhibit RNA virus infection. Proc Natl Acad Sci USA 104:3129–3134 9. Zhu H, Klemic JF, Chang S, Bertone P, Casamayor A, Klemic KG, Smith D, Gerstein M, Reed MA, Snyder M (2000) Analysis of

10.

11.

12. 13. 14.

15.

16.

17.

yeast protein kinases using protein chips. Nat Genet 26:283–289 Ptacek J, Devgan G, Michaud G, Zhu H, Zhu X, Fasolo J, Guo H, Jona G, Breitkreutz A, Sopko R, McCartney RR, Schmidt MC, Rachidi N, Lee SJ, Mah AS, Meng L, Stark MJ, Stern DF, De Virgilio C, Tyers M, Andrews B, Gerstein M, Schweitzer B, Predki PF, Snyder M (2005) Global analysis of protein phosphorylation in yeast. Nature 438:679–684 Lin Y-Y, Lu J-Y, Zhang J, Walter W, Dang W, Wan J, Tao S-C, Qian J, Zhao Y, Boeke JD, Berger SL, Zhu H (2009) Protein acetylation microarray reveals that NuA4 controls key metabolic target regulating gluconeogenesis. Cell 136:1073–1084 Pickart CM (2001) Mechanisms underlying ubiquitination. Annu Rev Biochem 70: 503–533 Fang S, Weissman AM (2004) A field guide to uniquitylation. Cell Mol Life Sci 61: 1546–1561 Xu P, Duong DM, Seyfried NT, Cheng D, Xie Y, Robert J, Rush J, Hochstrasser M, Finley D, Peng J (2009) Quantitative proteomics reveals the function of unconventional ubiquitin chains in proteasomal degradation. Cell 137:133–145 Lu J-Y, Lin Y-Y, Qian J, Tao S-C, Zhu J, Pickart C, Zhu H (2008) Functional dissection of a HECT ubiquitin E3 ligase. Mol Cell Proteomics 7:35–45 Gupta R, Kus B, Fladd C, Wasmuth J, Tonikian R, Sidhu S, Krogan NJ, Parkinson J, Rotin D (2007) Ubiquitination screen using protein microarrays for comprehensive identification of Rsp5 substrates in yeast. Mol Syst Biol 3:116 Fasolo J, Snyder M (2009) Protein micro arrays. Methods Mol Biol 548:209–222

wwwwwwwwwwwwwwww

Part IV Strategies for Validation of Candidate Targets

wwwwwwwwwwwwwwww

Chapter 15 Multiplexed Detection of Antibodies Using Programmable Bead Arrays Karen S. Anderson Abstract The detection of antibodies in sera has broad applications for detection and monitoring of infectious diseases, autoimmunity, and cancer. Proteomic methods of antigen detection, such as protein microarrays, are excellent clinical discovery tools, but due to both cost and specialization of manufacture, these are limited to screening small numbers of sera. Downstream assays for biomarker validation studies require rapid, reproducible, multiplexed assays for the simultaneous screening of fewer (<100) antigens with hundreds or thousands of sera. Traditional clinical ELISA assays use recombinant proteins, but these are limited by the ability to purify proteins free of cross-reacting contaminants and are limited to one antigen at a time. Here, we describe the application of coupled in vitro protein production with anti-tag capture onto bead arrays, for the rapid multiplexed detection of antibodies in sera. These assays can be readily adapted for detection of any protein-specific infectious, autoimmune, or cancer-specific antibodies. Key words: Bead array, Luminex, Autoantibodies, Biomarkers, ELISA

1. Introduction The induction of antibodies in sera is a highly specific response to changes in protein structure and content. Antibodies to target antigens are well defined for infectious agents, but can be generated to autoantigens in the setting of clinical autoimmunity and cancer-derived proteins (reviewed in ref. (1)). However, for most infectious and autoantigens, detection of the full breadth and specificity of the immune response has been limited by the ability to generate and purify hundreds or thousands of protein assays for serologic detection. The recent development of proteomic technologies, including protein microarrays, phage display, and reversephase protein arrays, has the potential for proteome-wide immune monitoring. These assays have all been used to identify serum antibodies and infectious, autoimmune, and cancer antigens (2–5). Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_15, © Springer Science+Business Media, LLC 2011

227

228

Anderson

However, these assays are generally limited to antigen discovery, as they are both labor-intensive and expensive for high-throughput clinical monitoring of immune responses. When immunogenic antigens are identified using proteomic technologies, it is necessary to validate those select target antigens using larger serum (or plasma) sets. Currently, the standard ELISA assay is the primary assay for the confirmation of antibody biomarkers. ELISA requires recombinant protein expression and purification, and detects antibodies to one antigen at a time. For multiplexed analysis, recombinant protein antigens can be directly or indirectly coupled onto spectrally distinguishable fluorescent beads (Luminex) (6, 7), which can theoretically detect up to 100 antibodies per serum sample, although cross-talk and interference with multiplexed assays remain a concern. As with ELISA assays, these assays require antigen production and purification (8). Here, we describe two related assays for the high-throughput detection of antigen-specific antibodies in patient sera, RAPID ELISA (5) and bead-array ELISA (9). Both assays use rabbit reticulocyte lysate for the in vitro transcription/translation (IVTT) of target antigens from plasmids and can be readily coupled to ORFeome-derived vector systems. Antigens are expressed in vitro and captured via anti-tag antibodies onto plates (RAPID ELISA) or fluorescent beads (Luminex platform). While RAPID ELISA can be used for the detection of one antigen-specific antibody at a time, the bead array allows for multiplexed detection of up to 100 individual antigens from one serum sample (see Note 1).

2. Materials 2.1. In Vitro Protein Expression

1. DNA preparation systems (Qiagen, Valencia, CA or Nucleo bond, Clontech, Mountain View, CA). 2. TNT T7 Coupled Reticulocyte Lysate System (Promega, Madison, WI). 3. RNase Out (Invitrogen, Carlsbad, CA).

2.2. RAPID ELISA (96-Well Format)

1. PBS-T/milk blocking buffer: PBS, 5% fat-free dried milk, 0.2% Tween20, pH 7.4. 2. Serum samples. 3. SuperBlock blocking buffer (Pierce, Rockford, IL). 4. GST-coated 96-well detection module plates (GE Healthcare, Piscataway, NJ). 5. Mouse anti-GST (26H1) MAb (Cell Signaling Technology, Beverly, MA). 6. Goat anti-human IgG-HRP (Jackson Immunoresearch, Westgrove, PA).

Multiplexed Detection of Antibodies Using Programmable Bead Arrays

229

7. Sheep anti-mouse IgG-HRP (GE Healthcare, Piscataway, NJ). 8. SuperSignal ELISA Femto (Pierce, Rockford, IL). 9. Chemiluminescence Plate Reader (PerkinElmer Victor3, Waltham, MA). 10. Plate shaker (Eppendorf Thermomixer R, Westbury, NY). 2.3. Bead-Based Multiplexed ELISA

1. 50 mM MES, pH 5.0. 2. PBS-BN: PBS, 1% BSA, 0.05% sodium azide, pH 7.4. 3. PBS-T: PBS, 0.05% Tween20, pH 7.4. 4. PBS-1% BSA, pH 7.4. 5. Sulfo-NHS. 6. EDC. 7. Slide-a-Lyzer 10K Dialysis cassette. 8. Goat anti-GST antisera (GE Healthcare, Piscataway, NJ). 9. Rabbit anti-FLAG antisera (Sigma-Aldrich, St. Louis, MO). 10. Goat anti-mouse IgG-PE antisera (Jackson Immunoresearch, Westgrove, PA). 11. Mouse anti-GST MAb clone 26H1 (Cell-Signaling Technol ogy, Beverly, MA). 12. Goat anti-human IgG-biotin (Jackson Immunoresearch, Westgrove, PA). 13. Streptavidin-R-Phycoerythrin (Invitrogen, Carlsbad, CA). 14. Carboxylated SeroMap microspheres (see Note 2, Luminex, Austin, TX). 15. Copolymer microfuge tubes (USA Scientific, Ocala, FL). 16. 1.2 mm MultiScreen HTS BV filter plate (Fisher Scientific, Pittsburgh, PA). 17. Sonicator (Cole Parmer, Vernon Hills, IL). 18. Benchtop tube rotator (Fisher Scientific, Pittsburgh, PA). 19. Vacuum manifold (Millipore, Billerica, MA). 20. Luminex 200 machine (Luminex, Austin, TX).

3. Methods Both assays described here, the plate-based RAPID ELISA and the bead-array ELISA, involve the in vitro production of recombinant protein generated immediately prior to use (see Note 3). These proteins are produced from T7-based plasmid vectors using an optimized IVTT system, and captured onto anti-GST plates (RAPID ELISA, Fig. 1a) or anti-GST-coupled fluorescent beads (bead-array ELISA, Fig. 1b). Diluted sera are added, bound IgG

230

Anderson

a

Sera

IVTT

b

Coupling of anti-GST antibody to beads

Capture of GST-

Mix

tagged protein onto anti-GST beads

re-aliquot

Sera

Anti-IgG

Detection

Cell-free expression of GST-tagged DNA

Fig. 1. Schematic of RAPID ELISA and bead-array ELISA for the detection of antibodies in serum. (a) RAPID ELISA. cDNA encoding full-length antigens are expressed as tagged fusion proteins using rabbit reticulocyte lysate. The proteins are captured by anti-tag antibodies in individual wells. Human serum containing antigen-specific antibodies is added and detected with HRP-conjugated secondary anti-IgG antibodies (5). (b) Bead-array ELISA. After expression with rabbit reticulocyte lysate, the recombinant proteins are captured onto anti-GST-coupled microspheres. The beads displaying different antigens are pooled and realiquoted prior to addition of human serum so that there are 5,000 of each color/ antigen bead per serum. Human IgG is detected with PE-conjugated secondary anti-IgG antibodies. Reproduced with permission from Elsevier (9).

is washed, and secondary PE-labeled anti-IgG detection antibodies are added. Bound IgG is detected with a luminometer for RAPID ELISA or with a Luminex 200 instrument for the bead-array ELISA. Classical bead-based (Luminex) arrays rely on covalent crosslinking of antigen to reactive moieties on the bead surface. Each antigen is then “addressable” based on bead color. In this beadarray ELISA, anti-tag antibodies (i.e., GST) are first directly coupled to the beads, and the same antibody is coupled to all the bead colors required for the assay. Both GST and FLAG-tagged proteins can be captured and detected on the beads with this approach. Confirmation of coupling efficiency is done with antiIg secondary antibodies. Anti-GST-coupled beads are prepared in advance, and are stable for at least 1 year. Proteins are captured onto individual beads, pooled for multiplexed detection, and aliquoted for serologic assays.

Multiplexed Detection of Antibodies Using Programmable Bead Arrays

3.1. In Vitro Transcription/ Translation of Recombinant Proteins 3.1.1. DNA Preparation

3.1.2. DNA Precipitation (If Not Already Done)

231

Highly-purified plasmid DNA is required for effective IVTT. For this protocol, a modified T7-based pCITE vector optimized for protein expression was used. This vector contains att sites for subcloning using the Invitrogen Gateway system and encodes a c-terminal GST fusion gene (plasmid available from http://www.hip. harvard.edu) (10). Creator-based vectors (Clontech) and FLAGtagged vectors have also been tested (see Note 3). Standard DNA maxipreps (either Qiagen or Nucleobond) or high-throughput DNA preparations (11) were used. 1. For one well, transfer 300 ml DNA (at least 2 mg) to microfuge tube. 2. Add 30 ml 3 M sodium acetate (10% final). 3. Add 240 ml isopropanol (80% final). Vortex. 4. Spin 5 min at 18,300 × g. Decant. 5. Add 500 ml 80% ethanol. Vortex, spin 5 min at 18,300 × g, and decant. 6. Leave uncapped to dry. 7. Resuspend pellet with 100 ml DEPC-treated water. 8. Spin briefly and quantify at OD280 nm.

3.1.3. In Vitro Transcription/Translation

1. The recombinant proteins are prepared fresh on the day of the ELISA assay. 2. Remove components of reticulocyte lysate kit and RNase Out from freezers. Thaw lysate, RNA Polymerase, and RNase Out on ice. Thaw other components at room temperature (RT) and place on ice (see Note 4). 3. Set water bath to 30°C. 4. For each well of the final RAPID ELISA assay, transfer 500 ng of DNA to microfuge tube. Bring all DNA samples to equal volumes in DEPC water. 5. Keep reaction on ice. For a master mix, add to 1 tube reticulocyte lysate (200 ml), 16 ml reaction buffer, 8 ml polymerase, 4 ml AA mix –Met, 4 ml amino acid mix –Leu, 8 ml RNase Out, and DEPC water for a total volume of 400 ml when mixed with DNA. 6. Add reticulocyte lysate mix to each tube of DNA (25 ml total/ well, including DNA). 7. Incubate DNA/lysate at 30°C for 1.5 h.

3.2. RAPID ELISA (96-Well Format)

1. Block wells of GST-coated 96-well plates with PBS-T/milk overnight at 4°C. Blocking at RT for 4 h is also effective (see Note 5).

232

Anderson

2. Prepare recombinant proteins from IVTT preparation (see Subheading 3.1.3). Add 1× PBS to each tube for a final volume of 50 ml per well. Transfer 50 ml to each well and seal plate. Each serum is tested in duplicate. 3. Incubate at RT for 2 h while shaking at 800 rpm on an orbital shaker. The shaking is important and speeds the reaction. 4. Wash 5× with 200 ml PBS-T/milk per well. 5. Add 100 ml human serum (diluted 1:300 in PBS-T/milk). Optimal dilutions are established for each assay, usually 1:100–1:600. 6. To test protein expression of GST fusions, add mouse antiGST MAb, 1:1,000 in SuperBlock to separate wells. 7. Incubate for 1 h at RT while shaking at 800 rpm on an orbital shaker. 8. Wash 5× with PBS-T/milk. 9. Add 100 ml secondary antibody (anti-human-IgG-HRP for wells with human sera, 1:1,000 in PBS-T/milk or anti-mouseIgG-HRP for wells with mouse anti-GST, 1:1,000 in SuperBlock). Incubate for 1 h at RT while shaking. 10. Wash 5× with 1× PBS. 11. Mix equal volumes each of solutions 1 and 2 from SuperSignal ELISA Femto kit. Add 100 ml mixed solution to each well. Mix 1 min with shaker. 12. Detect by chemiluminescence on Victor3 plate reader or equivalent. 3.3. B ead-Array ELISA

To screen for antibodies using a bead-array (Luminex), anti-GST antisera is first coupled to the individual beads and stored. Antigenic proteins are produced with IVTT as with the RAPID ELISA assay (see Subheading 3.1.3) and captured onto the individual beads prior to mixing for multiplexed analysis (see Note 6).

3.3.1. Preparation of Antibodies for Bead Conjugation

The commercial anti-GST antibodies contain sodium azide, which needs to be removed first with dialysis. 1. Pre-wet dialysis cassette membrane with cold PBS. 2. Dilute antibody in 1 ml PBS. 3. Inject diluted antibody into cassette. Remove air. 4. Dialyze against 3× with 1 L cold PBS over 12–24 h total. 5. Remove antibody from dialysis cassette. 6. Determine concentration by absorbance (mg/ml = UV280/1.6). 7. Aliquot and store in −20°C.

Multiplexed Detection of Antibodies Using Programmable Bead Arrays 3.3.2. Antibody Coupling to Luminex Microspheres

233

This assay has been optimized for the anti-GST antibody/bead ratios shown below. For other anti-tag antibodies, the antibody/ bead ratios should be titered. 1. Volumes listed below are for small scale (2.5 × 106 beads) and large scale [12.5 × 106 beads] (1 ml). 2. Dilute Sulfo-NHS to 50 mg/ml in sterile water, immediately prior to use. 3. Dilute EDC to 50 mg/ml in sterile water, immediately prior to use. 4. Equilibrate microspheres to RT. Rotate 3 min on benchtop rotator. 5. Transfer 2.5 × 106 [12.5 × 106] microspheres to copolymer microfuge tubes. 6. Wash microspheres by centrifuging at 18,300 × g for 2 min. Remove supernatant and add 100 ml [200 ml] sterile water. Resuspend microspheres by vortexing 10 s, sonicating 20 s, and vortexing 10 s. Centrifuge and remove supernatant. 7. Activate microspheres by resuspending in 80 ml 100 mM monobasic sodium phosphate, pH 6.2. 8. Add 10 ml [50 ml] 50 mg/ml sulfo-NHS to microspheres. Vortex. Add 10 ml [50 ml] 50 mg/ml EDC to microspheres. Vortex. 9. Incubate 20 min at RT. Vortex 10 s every 10 min. 10. Wash microspheres 2× with 250 ml [500 ml] 50 mM MES, pH 5.0. 11. To couple the microspheres, resuspend in 100 ml [200 ml] 50 mM MES, pH 5.0. 12. Add 12.5 mg [62.5 mg] goat anti-GST antisera or rabbit antiFLAG antibody. 13. Add 50 mM MES, pH 5.0 to final volume 500 ml [1 ml]. 14. Vortex 10 s. 15. Incubate 2 h, rotating at RT. 16. Wash microspheres with 500 ml [1 ml] PBS-BN. 17. Incubate 30 min, rotating at RT. 18. Wash microspheres 2× with 1 ml PBS-T. 19. Resuspend in 250 ml [1 ml] PBS-BN. 20. Count microsphere suspension with hemocytometer. 21. Store at 4°C wrapped in foil. Beads are stable for at least a year.

3.3.3. Confirmation of Antibody Coupling to Microspheres

Prior to use in ELISA assays, the amount of anti-GST antibody that is covalently attached to the microspheres is confirmed using an independent anti-GST MAb. See Note 7 for technical notes on

234

Anderson

using the vacuum manifold to wash the microspheres and handling the filter plates. Beads of any fluorescent color can be used; colors in adjacent fluorescent spectra can be readily distinguished, and cross-reactivity was not observed. 1. Make twofold serial dilutions of donkey anti-goat IgG-PE detection antibody to total volume per reaction of 50 ml. 2. Resuspend and aliquot coupled microspheres into a copolymer microfuge tube for 5,000 microspheres/reaction. Dilute to 100 microspheres/ml in PBS-1% BSA (50 ml/ reaction). 3. Pre-wet filter plate with 100 ml/well PBS-1% BSA. Vacuum. 4. Resuspend microspheres. Add 50 ml coupled microspheres to appropriate wells in the filter plate. 5. Add 50 ml diluted detection antibody to each well. 6. Incubate 30 min at RT shaking at 750 rpm on an orbital shaker. 7. Wash wells 2× with 100 ml PBS-1% BSA by vacuum. 8. Add 100 ml PBS-1% BSA. Shake for 5 min at 750 rpm on an orbital shaker to resuspend beads. 9. Analyze with a Luminex 200 machine per manufacturer’s instructions. 3.3.4. Bead-Array ELISA, Day 1

The multiplexed bead-array ELISA is a 2-day assay. Antigens are expressed with IVTT on the day of the reaction. For assay planning, sera are tested in duplicate, and protein capture onto the beads is tested in duplicate at the same time using both separate wells and anti-GST MAbs. Controls include the vector control or any control GST fusion protein, including GST alone. Prior to use in ELISA assays, the amount of anti-GST antibody that is covalently attached to the microspheres is confirmed using an independent anti-GST MAb. 1. Express protein antigens by IVTT on the same day as the assay (see Subheading 3.1.3). Use 500 ng DNA per 25 ml reaction. 2. Resuspend coupled microspheres and aliquot enough of the coupled microspheres into a copolymer microfuge tube for 5,000 microspheres/reaction. Dilute to 100 microspheres/ml in PBS-1% BSA (should be 50 ml/reaction). 3. Add 25 ml antigen-expressed IVTT reaction to beads. 4. Mix IVTT protein and beads at RT for 2 h while rotating in dark. 5. Pre-wet filter plate with 100 ml/well PBS-1% BSA. Vacuum and blot.

Multiplexed Detection of Antibodies Using Programmable Bead Arrays

235

6. Spin tubes 2 min at 18,300 × g to pellet the beads. Remove supernatant. Resuspend in 10 ml PBS-1% BSA per reaction. 7. Mix beads in one copolymer microfuge tube to multiplex. Vortex 10 s, sonicate 20 s, vortex 10 s. Add equal volumes of pooled beads to each well. There will be 5,000 beads per antigen per well. 8. Wash 2× with 100 ml PBS-1% BSA per well. 9. Resuspend in 50 ml PBS-1% BSA per well. Shake 5 min at 750 rpm on an orbital shaker. 10. Dilute serum 1:80 in PBS-1% BSA and add 50 ml serum to appropriate wells. Sera are tested in duplicate, and dilutions should be titered. More concentrated sera inhibit the reaction. 11. To test binding of GST protein to the beads, in separate wells, add anti-GST MAb at 1:400 (5 mg/ml). 12. Seal filter plate with tape, cover with lid, and wrap in foil. Incubate overnight at 4°C while shaking at 750 rpm on an orbital shaker. 3.3.5. Bead-Array ELISA, Day 2

1. Shake plate at 750 rpm on an orbital shaker for 30 min at RT. Wash wells 2× with 100 ml PBS-1% BSA. 2. Resuspend microspheres in 50 ml PBS-1% BSA. Shake 5 min at 750 rpm on an orbital shaker. 3. For human serum samples, dilute biotin-labeled goat antihuman IgG antibody to 4 mg/ml in PBS-1% BSA (1:375). Add 50 ml to each well. Add 50 ml PBS-1% BSA to wells for GST detection. 4. Incubate 30 min at RT with shaking. Wash 2× with 100 ml PBS-1% BSA. 5. Resuspend microspheres in 50 ml PBS-1% BSA. Shake 5 min at 750 rpm on an orbital shaker. 6. For human serum samples, dilute streptavidin-R-PE to 4 mg/ ml in PBS-1% BSA (1:250). For GST detection wells, add goat anti-mouse IgG-PE to 1:200. Add 50 ml to each well (see Note 8). 7. Incubate 30 min at RT with shaking. Wash 2× with 100 ml PBS-1% BSA. 8. Resuspend microspheres in 100 ml PBS-1% BSA. Shake until ready to analyze with Luminex, minimum of 5 min at 750 rpm on an orbital shaker. 9. Analyze with Luminex 200 instrument per manufacturer’s instructions (Fig. 2).

0

5

10

15

20

25

30

0

5

10

15

20

25

30

vector control

d

a

p21

CNEBNA- EBNAEBNA1 EBNA1 3A 3B

LMP-2 vector control

e

b

p21

Antigen

CNEBNA- EBNAEBNA1 EBNA1 3A 3B

LMP2

vector control

f

c

p21

CNEBNA- EBNA- LMP-2 EBNA1 EBNA1 3A 3B

Fig. 2. Multiplexed detection of viral antibodies in human sera. Six individual GST-tagged protein antigens, p21, N-EBNA1, C-EBNA1, EBNA-3A, EBNA-3B, and LMP-2, were expressed, loaded onto individual anti-GST-coupled microspheres, and pooled. IgG-specific responses to the viral antigens were detected from six healthy donor sera (a–f). Differential patterns of IgG were readily detected with the bead array. Reproduced with permission from Elsevier (9).

MFI x 103

236 Anderson

Multiplexed Detection of Antibodies Using Programmable Bead Arrays

237

4. Notes 1. The bead-array ELISA (9) and the RAPID ELISA (5) have comparable sensitivities, specificities, and limits of detection as standard protein ELISA for the detection of antibodies in sera to the viral antigen EBNA-1 and the tumor antigen p53. The bead-array ELISA is comparable in sensitivity as direct covalent attachment of recombinant p53 protein to the microspheres. 2. There are different chemical compositions of Luminex microspheres. This assay uses the SeroMap microspheres, but also works with magnetic microspheres, with comparable sensitivities and specificities as the SeroMap microspheres. 3. Protein expression with IVTT can vary from batch to batch. These studies have been optimized for the pCITE-c-terminal GST vector. FLAG-based vectors have also been tested, but not as extensively. Over 100 tumor and viral antigens have been expressed and captured using this assay on Luminex beads, with over 90% of the antigens expressed as determined by GST detection. Both high-quality DNA, and careful use of RNase-free pipets and filter tips are needed for efficient protein expression. Batches of rabbit reticulocyte lysate are tested and ordered in bulk. 4. The cost of RAPID ELISA and the multiplexed bead-array ELISA is approximately $10/antigen, due to the high cost and volume required of IVTT reagent. Dilution of the IVTT greater than 1:2 lowers the sensitivity of detection, but could be used for high-titer antibody detection. Increasing the DNA concentration beyond 500–1,000 ng/reaction had little effect. 5. Some sera (<10%) have persistently high background in these assays. Multiple blocking reagents have been tested on the RAPID ELISA, including goat sera and rabbit sera, and commercial ELISA blocking reagents. Of these, addition of PBS-T/milk after protein expression and capture (prior to sera addition) overnight can decrease the background. 6. Up to eight antigens have been multiplexed at one time using the bead-array ELISA. Theoretically, this assay could be used to multiplex up to 100 antigens (limited only by the number of spectrally distinguishable microspheres). 7. Using the vacuum to wash microspheres can be tricky. Turn on vacuum first before putting any plate on top. Set vacuum manifold to −6 in. Hg by placing solid bottom plate on top. Do not vacuum to dryness, otherwise beads will get stuck in filter (count to about 3 then turn off vacuum). For the filter plate, tape unused wells with 96-well plate sealing tape. Tape the filter plate on top of an upside down lid. Gently blot under drain of filter plate after last vacuum of each wash

238

Anderson

before adding or resuspending beads. Cover wells with a plate lid and wrap with foil to protect the beads from light when not adding reagents. Seal wells with tape if the incubation is overnight. Occasionally beads will get trapped in the filter plates (no bead counts during the assay). These can be rescued by resuspending the beads in the plates with vigorous shaking or vortexing. 8. The assay presented here uses indirect labeling of bound human IgG with a biotinylated anti-human IgG secondary antibody and streptavidin-PE detection. Direct labeling of the secondary antibody with PE also works, but with decreased sensitivities. This can be overcome by overnight incubation of sera.

Acknowledgments The author would like to thank Jessica Wong for excellent technical assistance in the development of these assays, and Dr. Joshua LaBaer for protein microarray expertise. This work was supported by a research grant from the NCI Early Detection Research Network 5U01CA117374. References 1. Anderson KS, LaBaer J (2005) The sentinel within: exploiting the immune system for cancer biomarkers. J Proteome Res 4: 1123–1133 2. Robinson WH, DiGennaro C, Hueber W et al (2002) Autoantigen microarrays for multiplex characterization of autoantibody responses. Nat Med 8:295–301 3. Hudson ME, Pozdnyakova I, Haines K, Mor G, Snyder M (2007) Identification of differentially expressed proteins in ovarian cancer using high-density protein microarrays. Proc Natl Acad Sci U S A 104:17494–17499 4. Anderson KS, Ramachandran N, Wong J et al (2008) Application of protein microarrays for multiplexed detection of antibodies to tumor antigens in breast cancer. J Proteome Res 7:1490–1499 5. Ramachandran NAK, Raphael J, Hainsworth E, Sibani S, Montor W, Pacek M, Wong J, Eljanne M, Sanda MJ, Hu Y, Logvinenko T, LaBaer J (2008) Tracking humoral responses using self assembling protein microarrays. Proteomics Clin Appl 2:1518–1527

6. Kellar KL, Iannone MA (2002) Multiplexed microsphere-based flow cytometric assays. Exp Hematol 30:1227–1237 7. Waterboer T, Sehr P, Michael KM et al (2005) Multiplex human papillomavirus serology based on in situ-purified glutathione s-transferase fusion proteins. Clin Chem 51:1845–1853 8. Opalka D, Pessi A, Bianchi E et al (2004) Analysis of the HIV-1 gp41 specific immune response using a multiplexed antibody detection assay. J Immunol Methods 287:49–65 9. Wong J, Sibani S, Lokko NN, Labaer J, Anderson KS (2009) Rapid detection of antibodies in sera using multiplexed selfassembling bead arrays. J Immunol Methods 350:171–182 10. Ramachandran N, Hainsworth E, Bhullar B et al (2004) Self-assembling protein microarrays. Science 305:86–90 11. Ramachandran N, Raphael JV, Hainsworth E et al (2008) Next-generation high-density self-assembling functional protein arrays. Nat Methods 5:535–538

Chapter 16 A Coprecipitation-Based Validation Methodology for Interactions Identified Using Protein Microarrays Ovidiu Marina, Jonathan S. Duke-Cohan, and Catherine J. Wu Abstract Candidate interactions identified by high-throughput protein microarray screening require rigorous confirmation. Such validation is time-consuming and labor-intensive using conventional techniques. We describe a medium-throughput validation protocol based on coprecipitation of biotin-labeled proteins synthesized in vitro using a rabbit reticulocyte lysate-coupled transcription and translation system. As our experimental system is based on screening for serum antibodies, we also present methods on purifying immunoglobulin from serum and quantifying the amount of coprecipitated (immunoprecipitated) target protein on Western blot. This technique provides a sensitive confirmatory test allowing for the rapid elimination of false positives prior to more extensive validation and analysis of target interactions in their native environment. Key words: High-throughput screening, Validation, Coupled transcription/translation in vitro, Coprecipitation, Image processing

1. Introduction Protein microarrays have evolved in direct response to questions arising from our growing understanding of mammalian cell biological pathways and pathogenesis. In individuals making apparently specific antitumor antibodies, what are the precise targets of this response? Are these targets useful for possible protective vaccine development for a broader population? In processes where an autoimmune component might be suspected, is it possible to screen patient sera for reactivity to endogenously-expressed targets? The complete sequencing of several mammalian genomes has uncovered many potential open reading frames that are associated with disease by quantitative-trait loci analysis, but there is

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_16, © Springer Science+Business Media, LLC 2011

239

240

Marina, Duke-Cohan, and Wu

little understanding of the biochemical interactions of the identified targets (1). The definition of pathways in health or disease is critically dependent upon identifying the likely protein–protein interactions of candidate disease-associated gene products. Protein microarrays allow these questions to be addressed through the high-throughput screening of protein–protein interactions using sera, purified antibodies, cerebrospinal fluid, identified “bait” proteins, or even tissue culture supernatants. The current techniques for array preparation are designed for the high-throughput screening of a large number of proteins but prevent individual refinements that may be required for specific interactions to occur. For example, proteins often exist in complexes, and the multimeric interaction results in a conformation not observed for the protein in isolation (2, 3). Even in isolation, the conformation of a protein ectodomain or cytoplasmic domain may be influenced by association with the nitrocellulose-coated glass slide of the microarray, leading to a different conformation of the target protein and loss or gain of reactivity (4). Protein structure may critically depend on disulfide bonding or preproprotein processing by proteases, glycosylation, phosphorylation, palmitoylation, myristoylation, or a multitude of other potential posttranslational modifications which are not commonly available on microarrayed proteins (5). As a consequence, high-throughput processing is a powerful technique for narrowing the range of potential interactions, but it may result in a significant number of false positives and may miss important interactions occurring under natural conditions. Other considerations concern screening under conditions mimicking cellular locale, such as the oxidizing conditions in the extracellular environment, the reducing conditions in the cytoplasm, or the low pH within the lysosome. These elements are not usually taken into account in high-throughput screening, but are factors that may affect protein domain conformation (6). In addition to processing limitations, actual technical errors may occur during manufacture resulting in the spotting of misidentified proteins (7, 8). Having identified a group of potential target interactions, the ideal validation would be immunoprecipitation of each target from primary cells or representative untransfected cell lines using antibodies against the partner (or bait) protein, and identification of the target on Western blot (coprecipitation). When the screening is for antibody specificity in serum or other physiological fluids, only a method for identifying the target protein in the coprecipitate is needed, either antibody-based, or, more recently, by mass spectroscopic analysis of tryptic peptide digests of the coprecipitate (9). The validation protocol is often undertaken in transfected and overexpressing cells, but is less desirable as the identified interaction cannot be reliably quantified as a fraction of the total material, raising the possibility of a weak or nonspecific interaction driven by the high expression levels of the partner proteins.

Coprecipitation for Protein Microarray Validation

241

Even using a transient transfection system, individual coprecipitations for a large group of proteins are labor-intensive and time-consuming while probing of the reaction in untransfected cells using native proteins requires sensitive antibody reagents for both partners (see Note 1). We describe here a technique that allows rapid expression of test amounts of target protein that are then coprecipitated by specific antibody from a serum sample (immunoprecipitation) (see Note 2). We have focused our research on the targets of immunotherapy using donor lymphocyte infusion (DLI) for relapsed chronic leukemia as a model (10). The procedural scheme is presented in Fig. 1. Following the identification of

Fig. 1. Schematic representation of the experimental protocol. (I) Patient sera or plasma are used to probe protein microarrays for antibody reactivity. (II) Results from different samples, e.g., before and after donor lymphocyte infusion (DLI), are then compared individually (shown) or as a group (not shown) across patients. (III) DNA sequences of the proteins identified as having significantly higher reactivity post- but not pretreatment are then acquired and inserted into an expression vector plasmid. The structure of the plasmid we used contains a T7 polymerase promoter region (pT7), the protein sequence with endogenous or introduced start codon, an additional sequence for a protein tag (tag*), the protein synthesis stop codon, and a region indicating the site for polyadenylation (poly(A)). The protein tag may be omitted but was included in our experiments to fully duplicate the protein sequence found on the protein microarrays. (IV) Protein is then synthesized using a cell-free mammalian expression system which incorporates lysine transfer RNA (tRNA) precharged with biotin-labeled lysine (tRNAlys) in the amino acid mixture. (V) Either the protein tag or the translationincorporated biotin label can be used to identify the synthesized protein antigen. (VI) The same sera or plasma initially screened on the protein array is then used to verify the findings of step II, hit identification. In the experiment presented here, validated hits had reactivity post- but not pretreatment.

242

Marina, Duke-Cohan, and Wu

hits, the appropriate coding DNA is inserted into a vector suitable for protein expression. The proteins are then produced by coupled transcription and translation in a mammalian cell lysate incorporating transfer RNA (tRNA) preloaded with lysine e-amino-labeled with biotin (see Note 3). The incorporation of biotin during translation provides a universal tag minimally destructive to secondary structure yet can be detected with high sensitivity on Western blot using streptavidin (Fig. 2).

Fig. 2. The antibody response against chronic lymphocytic leukemia in one patient treated by donor lymphocyte infusion. (a) The serum reactivity to protein products of genes DAPK3 and ZFYVE19 as well as EBNA1 as control. Sera samples from multiple timepoints before and after therapy were used to immunoprecipitate protein produced in a cell-free mammalian expression system. Increasing serum reactivity following therapy can be evaluated visually. (b) The serum reactivity shown in (a) was quantified as described in Note 15. Briefly, band intensity for each immunoprecipitate product was measured using ImageJ, and mapped onto a standard curve determined using a protein band dilution series. (c) The patient’s disease burden measured by bone marrow biopsy indicates disease remission following treatment that correlates well with the serum antibody response.

Coprecipitation for Protein Microarray Validation

243

Immunoprecipitation results can then be compared visually or quantitatively. As depicted in Fig. 2, not only can the initial findings of the protein microarray be validated, but additional data on the kinetics of an antigen-specific response can be derived and related to clinical parameters.

2. Materials 2.1. P rotein Synthesis

1. Plasmid encoding the target antigen DNA sequences. No specific plasmid is recommended but there must be a strong start codon within a Kozak consensus sequence, an appropriate mammalian stop codon and a consensus T7 promoter upstream of the ATG start codon to allow T7 polymerasedirected transcription (see Note 4). 2. TnT T7 coupled transcription/translation (rabbit) reticulocyte lysate system and Transcend tRNAlys preloaded with lysine-e-biotin (Promega, Madison, WI). 3. Temperature-programmable benchtop microcentrifuge shaker, e.g., Eppendorf Thermomixer.

2.2. Immuno precipitation

1. 5 mL Serum or plasma for each reaction, stored at −80°C until time of use, then at 4°C. 2. 5 mL of Protein synthesized using rabbit reticulocyte lysate per reaction, plus 20% extra to account for pipetting errors. Protein should be stored at −80°C until time of use, avoiding freeze-thaw cycles. 3. Rotator with Eppendorf tube holder. 4. Eppendorf tubes. 5. Phosphate buffered saline (PBS). 6. Tween 20 as a 20% solution in water. 7. Microcentrifuge, with cooling to 4°C if available. 8. Binding buffer: 1× PBS at 4°C. 9. Washing buffer: 1× PBS supplemented with 0.05% Tween 20 at 4°C. 10. Protein A sepharose CL-4B beads (GE Healthcare, United Kingdom) as a 50% suspension in 1× PBS and 0.05% sodium azide stored at 4°C (see Note 5). 11. (Optional, for IgG purification): Spin-X Centrifuge tube filters (0.22 mm cellulose acetate; Costar, Corning, NY); Protein A IgG binding buffer and elution buffer (Thermo Scientific, Rockford, IL).

244

Marina, Duke-Cohan, and Wu

2.3. SDSPolyacrylamide Gel Electrophoresis

1. 4× SDS Laemmli sample buffer: 0.25 M Tris, pH 6.8, 6% SDS, 40% sucrose or glycerol, 0.04% bromophenol blue. 2. Ready Gel Tris–HCl (Bio-Rad, Hercules, CA), with % crosslinking or gradient range dependent upon target size. 3. Nitrocellulose membrane or PVDF membrane (0.45 mm). 4. Whatman paper (Whatman, Florham Park, NJ). 5. SDS-polyacrylamide gel electrophoresis (SDS-PAGE) running buffer: 25 mM Tris, 0.192 M glycine, 0.1% SDS, pH 8.3, stored at room temperature. 6. Gel-blot transfer buffer: 25 mM Tris, 0.192 M glycine, 10% methanol, pH 8.3, stored at 4°C. 7. Prestained molecular weight markers (Kaleidoscope; BioRad, Hercules, CA). Biotin-labeled markers are a useful additional marker (Cell Signaling Technology, Beverly, MA); they are not visible on the transferred blot but are readily apparent after chemiluminescent development of bound streptavidin– horseradish peroxidase (HRP).

2.4. Western Blot

1. Blocking buffer: 0.5% Tween 20, 1× TBS; TTBS-high, kept at room temperature. 2. Washing buffer: 0.05% Tween 20, 1× TBS; TTBS-low, kept at 4°C. 3. Streptavidin–HRP conjugate (1:5,000–20,000 dilution, MP Biomedicals, Solon, OH). 4. Chemiluminescent substrate (Supersignal West Femto, Thermo Scientific; or ECL, GE Healthcare, Little Chalfont, UK). 5. BioMax Light film (Kodak, Rochester, NY) for development; although X-Omat Blue film (Kodak) suffices for initial assay development, the BioMax Light film provides sharper, clearer publication-quality images.

3. Methods The techniques described are designed to function as a mediumthroughput coprecipitation protocol for confirming interactions initially identified by high-throughput screening prior to investing large amounts of time and labor individually studying each interaction. We describe here the confirmation of reactivity to antigens identified by protein microarray screening using serum pretreatment and following tumor immunity secondary to DLI. Nevertheless, the procedure can easily be adapted to analyze any protein–protein interaction (see Note 2). If posttranslational glycosylation of a binding domain is important for interaction,

Coprecipitation for Protein Microarray Validation

245

the incorporation of canine pancreatic microsomal membranes can be included in the coupled transcription/translation reaction to provide the appropriate glycosyltransferases in an endoplasmic reticulum-like environment (see Note 6). The remarkable complexity and high protein concentration of sera, where many proteins actually have a transporting function, may lead to spurious background interactions when screening for new antibody reactivities. One technique to overcome this, if the expected specificity is in the IgG compartment (rather than e.g., IgM or IgA) is to purify the IgG from serum prior to screening. We present here an optional component that allows rapid purification of IgG from small samples of sera (25 mL) with close to 100% efficiency, and yields ideally suited for testing of reticulocyte lysate-expressed protein. This provides the added benefit of being able to standardize the IgG amount used for each immunoprecipitation. Finally, we describe a method for computing a numeric estimate of the relative amount of coprecipitated protein where visual comparison of results is ambivalent. 3.1. Microisolation of IgG from Human Sera (Optional)

1. For each sample, add 500 mL Protein binding buffer, 25 mL serum, and 100 mL of Protein A sepharose slurry (50%) into the upper chamber of Spin-X assembly. 2. Close cap securely, mix gently, then rotate for 1 h at 4°C. 3. Centrifuge complete assembly at 4,000 rpm/1,500 × g for 1 min, discard flowthrough, resuspend beads in 0.7 mL Protein A binding buffer, and incubate with rotation for 5 min at room temperature. 4. Repeat step 3 a further three times. 5. Centrifuge at 4,000 rpm/1,500 × g for 5 min and transfer insert containing beads to clean microcentrifuge tube containing 50 mL Tris–HCl, pH 7.6. 6. Add 100 mL of IgG elution buffer to washed beads in the Spin-X insert, gently mix by swirling with pipette tip (avoiding touching the membrane) and then leave at room temperature for 5 min. 7. Centrifuge at 10,000 rpm/9,300 × g for 2 min. Do not discard flowthrough! 8. Repeat steps 6 and 7 to yield 200 mL of purified IgG. 9. Use any standard protein assay, e.g., microBCA (Thermo Scientific), to calculate yield (usually 200–400 mg).

3.2. Protein Synthesis for Immunoprecipitation

1. Set the temperature on the benchtop mixer/shaker to be used at 30°C. 2. Label nuclease-free PCR tubes. All handling should be done on ice.

246

Marina, Duke-Cohan, and Wu

3. Aliquot plasmid DNA for each reaction into the PCR tubes, bringing volume up to 17.5 mL by addition of nuclease-free molecular biology grade water. The final 50 mL reaction should contain 1.5–2.0 mg of plasmid DNA requiring the stock DNA solution to be at least 100 ng/mL (see Note 7). 4. Half-volume negative control (nuclease-free water only) and positive control (e.g., luciferase) reactions should be performed with each synthesis. 5. Take out the reaction reagents, except for the reticulocyte lysate. Create a master mix for the number of reactions to be performed +1 where each 50 mL reaction contains 2 mL TnT buffer, 0.5 mL methionine, 0.5 mL leucine, 1 mL RNasin ribonuclease (40 U), 2.5 mL Transcend tRNA, 1 mL T7 polymerase (see Note 8). 6. Take out the reticulocyte lysate, warm up in your hand, and place on ice once melted. Add the appropriate amount to the master mix (25 mL for each reaction) and gently mix. 7. Add 32.5 mL of the master mix to the 17.5 mL of DNA for each reaction. Pipette gently to mix, changing pipette tips between tubes. If reactions need to be spun down prior to incubation, a 1.5 mL Eppendorf tube with the cap removed may be used as an adaptor for the PCR tubes containing the reaction mixture. 8. Incubate the PCR tube reactions in the benchtop mixer/ shaker at 950 rpm/30°C for 90 min. 9. To verify synthesis, 0.2–1.0 mL of product from each reaction is diluted in 3–6 mL of 4× SDS Laemmli buffer and visualized by SDS-PAGE and Western blot as outlined below. 3.3. Immuno precipitation

1. Aliquot 5 mL of serum or up to 50 mL purified immunoglobulin (see Note 9) into 1.5 mL Eppendorf tubes. These may be stored tightly sealed at 4°C for several days prior to use. 2. Spin down tube contents at 1,000 rpm/100 × g for 5 s. 3. Place tubes on ice. Add 5 mL reticulocyte lysate product to each tube and pipette up and down gently to ensure mixing. Where multiple reticulocyte lysate reactions giving the same product will be used in an experiment, these should be combined and mixed prior to use. 4. Incubate for 1 h in a cold room or refrigerator without shaking. 5. For each reticulocyte lysate product, create a 1:400 dilution by diluting 0.5 mL of product into 200 mL 4× Laemmli buffer and heating for 8 min at 90°C. This will be used as a positive control and marker on the Western blot for the location of the target protein band.

Coprecipitation for Protein Microarray Validation

247

6. Add 40 mL of 50% Protein A sepharose to each reaction tube (see Note 10). Before the first reaction and every 4–8 reactions, mix the Protein A sepharose stock solution by gently swirling (do not vortex) as the beads tend to settle rapidly. 7. Add 450 mL of binding buffer to each reaction. Immediately close the caps and place tubes onto a rotator at 4°C. Do not vortex or otherwise mix. Once the tubes are rotating, observe each tube to confirm that its contents are moving during the rotation. If not, gently tap the rotator or individual tubes until all tubes have contents moving during the rotation cycle. Leave reactions on the rotator in the cold room for 1 h. 8. Pellet the Protein A beads by centrifuging each reaction for 3 min at 1,000 rpm/100 × g and 4°C. 9. Carefully aspirate the supernatant; it is better to leave some than aspirate beads. 10. Add 1 mL washing buffer to each Eppendorf tube, and invert several times. Do not vortex. 11. Repeat steps 8–10 a further four times. 12. Pellet the Protein A sepharose beads by centrifuging the reactions for 4 min at 14,000 rpm/18,300 × g and 4°C. Aspirate off the supernatant as previously. 13. Pellet the Protein A sepharose beads again by centrifuging for 2 min at 14,000 rpm/18,300 × g and 4°C. Manually remove the remaining supernatant using a fine (gel-loading) pipette tip. Final volumes should be 20–25 mL. 14. Add 25 mL 4× SDS Laemmli buffer to the pelleted beads for each reaction. Immediately place on heating block at 90°C for 8 min. 15. Vortex for 10 s, then centrifuge for 2 min at 14,000 rpm/ 18,300 × g. Place products on ice for immediate use or store at −20°C for use within a week or −80°C for longer periods. 3.4. SDS-PAGE Gel Separation

1. Load 15 mL of supernatant into each lane of the gel. We suggest loading in a set order where the molecular weight markers are followed by the positive control, the immunoprecipitation products and, finally, the 1:400 dilution of the reticulocyte lysate product. 2. Run the gel at 200 V for 30–60 min or as recommended by the gel manufacturer or as appropriate for the protein size and gel concentration.

3.5. Western Blot

1. Remove the gel from the cassette and let equilibrate in transfer buffer on a rotary shaker for 30 min at room temperature to reduce the SDS concentration in the gel.

248

Marina, Duke-Cohan, and Wu

2. For wet transfer, prewet Scotchbrite pads and Whatman filter sheets (two of each for each gel) in transfer buffer. Prewet appropriately-sized nitrocellulose in water or PVDF first in methanol followed by water, then let transfer membrane equilibrate in transfer buffer for 10 min. For semidry transfer, follow manufacturer’s instructions. 3. Place the Whatman paper with gel facing up on top of a wet Scotchbrite pad on the open transfer cassette. Place the nitrocellulose or PVDF membrane on the gel, then add the second Whatman paper, and finally the second wet pad. As each item is added to the transfer cassette, gently and repeatedly roll a 10 mL pipette over the sandwich to remove air bubbles between the layers. Be sure to roll in two perpendicular directions to remove all air bubbles, especially after placing the transfer membrane. 4. Close the transfer cassette, and place in the transfer apparatus filled with transfer buffer, ensuring that the buffer covers the top of the cassette. In order to ensure buffer circulation for cooling, add a small magnetic stir bar and, if appropriate, a cooling element. Place the transfer container on ice or in a cold room on a magnetic stirrer. Transfer at 80 V constant voltage for 2 h, or at 200 mA constant current for 1–1½ h. 5. Stop transfer and remove the transfer membrane. Stain with dilute Ponceau Red to visualize protein transfer. 6. Remove the stain (it can be reused) and destain by repeated rinsing with blocking buffer. 7. Block membranes overnight in blocking buffer at room temperature on a slow shaker. Use a generous amount of blocking buffer (see Note 11). 8. Dilute streptavidin–HRP antibody 1:5,000–20,000 in blocking buffer. Filter this solution through a 0.22 mm filter prior to adding to the membrane. Add enough to ensure that the blot is well-covered and moves freely when placed on the shaker. 9. Incubate 2 h at room temperature on a slow shaker. 10. Carefully decant out the streptavidin–HRP reagent, removing as much as possible, and rinse the blot in wash buffer (TTBS-low). Decant out the wash. 11. Wash the blot 5× with TTBS-low for 5 min each wash on a shaker at room temperature. 12. Drain the washing buffer and add a similar amount of milliQ water. Gently swirl the container four to five times, drain the fluid and replace again with milliQ water. Place on shaker at room temperature for 5 min or until ready to develop, but no longer than 30 min.

Coprecipitation for Protein Microarray Validation

249

13. Prepare the chemiluminescent reagent immediately before use. 14. Partially, but not completely, dry the membrane after removing it from the milliQ water by dabbing the sides onto a dry paper towel. 15. Place the membrane on a clean working surface using a forceps. 16. Pipette the chemiluminescence developing reagent slowly onto the surface of the membrane. 17. Wait the appropriate amount of time (e.g., 5 min) for probing. 18. Pick up the membrane with forceps, and dab the sides onto a dry paper towel making sure the membrane is not fully dry. 19. Wrap the membrane in clear plastic wrap, taking care to remove excess fluid or significant air bubbles. 20. Place wrapped membranes onto an intensifying screen in a film development cassette. At this point, the assembled cassette should be taken to a dark room, film placed on top of the membrane, exposed, and developed. 21. Prepare several exposures to obtain a variety of intensities for each of the observed bands. Thirty seconds is usually a suitable initial exposure (see Note 12). 3.6. Estimating Immunoprecipitated Protein Using ImageJ (Optional)

To estimate relative protein quantities on the immunoblots, the same positive control must be run on every immunoblot to be compared, and multiple exposures must be taken, as described above. Densitometry then proceeds as follows using ImageJ software (see Note 13): 1. Define a rectangle around each protein band of interest, including the positive control(s), and measure the average signal Si and area Ai. Repeat this for all protein bands on one short, intermediate, and long exposure image. Exposures should be selected so that strong, medium, and weak protein bands are not saturated and can be easily visualized, measured, and compared on at least one exposure. 2. Take a measurement Bi of an adjacent area immediately above or below the protein band of interest within each gel lane for each exposure. 3. Calculate measured signal M i = (Bi − Si ) * Ai for each protein band and exposure (see Note 14). 4. Estimate the protein amount Pi for each protein and exposure using:

Pi = 10

−

   max  1 −1 −offset   logbase  steeping   Mi  

,

(1)

250

Marina, Duke-Cohan, and Wu

where the variables max, base, steep, and offset are experimentally determined (see Note 15). Within our experimental system, max was 850, base 2.7, steep 4, and offset 0. 5. Calculate the relative protein quantity Ri = Pi / Pcontrol , which corrects for differences in development between immunoblots for each protein and exposure. 6. True relative protein amount Tk is then calculated for each immunoprecipitated protein as the average of the two Ri measured on different exposures for that protein that are closest to each other in value. This addresses weak protein bands which are not imaged on short exposures and strong bands which are overexposed on long exposures of the Western blot.

4. Notes 1. The identification of interactions using transfected cells overcomes the need for sensitive antibodies to the target protein by allowing incorporation of universal tags such as myc, V5, FLAG, or Xpress. The incorporation of such tags, usually heavily charged at neutral pH, however, may also have subtle effects upon interactions. 2. The procedure is easily adapted to validation of a “bait” protein if a specific antibody is available that does not interfere with the potential binding site, or the “bait” protein carries a tag against which there is a specific affinity reagent. 3. Depending on the protein, 5–10% of the available lysines will be labeled with biotin. For small protein products (20 kDa or lower) with few lysines, it should be taken into account that labeling may not be strong and the concentration of tRNAlys-lysine-e-biotin may need to be increased (11). 4. In this report, we utilize the Gateway cloning system that allows easy exchange of sequence to a variety of vectors for differing downstream purposes (Invitrogen, Carlsbad, CA). The open-reading frames expressed in this report were inserted into pCITE, a vector with a T7 promoter that is optimized for protein expression in vitro (Novagen, Gibbstown, NJ). Transfer to a plasmid is not essential, however, and target sequence may be directly amplified by RT-PCR or PCR using primers that incorporate regulatory sequence and which we have found to function perfectly well in a coupled transcription and translation system optimized for PCR products (Quick TnT PCR; Promega) e.g., as a forward primer: 5¢-gtcggatcctaatacgactcactatagggaacagccaccATG-(genespecific sequence)-3¢

Coprecipitation for Protein Microarray Validation

251

where the italics represent a restriction endonuclease site for subsequent plasmid insertion (if desired), the bold sequence represents the T7 promoter sequence and, in bold italic, the start codon within a consensus Kozak sequence. Likewise, for the reverse primer: 5¢-gtcgcggccgctttttttttTCAATGGTGATGGTGATGATG(gene-specific sequence)-3¢ where the italics represent a restriction endonuclease site, the poly(T) sequence generates a short poly(A) for the desired transcript, the stop codon is in bold characters and the underlined sequence generates a C-terminal His(6) tag. Note that if the His-tag is incorporated and used for purification, then nickel- or cobalt-based chelating resins cannot be used due to interaction with hemoglobin in the reticulocyte lysate. In this instance, zinc-based media should be used (MagZ beads; Promega). The PCR-based technique allows for production in a usual lab setting of up to 100 biotin-labeled proteins from RNA or DNA template in less than a day. 5. It is critical to remove all traces of sodium azide by washing the beads and equilibrating in binding buffer prior to use. Sodium azide inhibits HRP (12). 6. Under these circumstances, sequence encoding a signal peptide is required upstream and in-frame of the desired target sequence. For this purpose, we have found the pSecTag2 vector from Invitrogen to be ideal, but due to the extra processing, the yield of biotinylated protein is significantly reduced (13). If disulfide bond formation is important, then the transcription and translation reactions should be uncoupled (since the system used here incorporates dithiothreitol [DTT]) and, following RNA production, the translation reaction should be performed using the FLEX reticulocyte lysate system (Promega) in the presence of canine microsomal membranes with no addition of DTT. 7. In order to generate labeled protein with minimal interference with structure, 35S-methionine or 3H-leucine is incorporated during the translation stage for some experimental protocols. As a consequence, unlabeled methionine and leucine representation is usually low in cell-free translation mixtures to accommodate this experimental option and it is essential to supplement the reaction mixture with these amino acids if lysine-e-biotin is the label of choice. 8. Standard centrifugal DNA miniprep techniques (e.g., Promega Wizard Plus SV system) usually yield 100 mL at around 200 ng/mL. 9. The amount of IgG to add very much depends upon the range of titers expected. In general, 10–25 mg in 20 mL is a good starting point.

252

Marina, Duke-Cohan, and Wu

10. For all procedures involving pipetting of Protein A sepharose beads, it is recommended that wide-bore 200 mL tips are used or prepared by cutting ~2 mm off the tip of general purpose 200 mL tips. These should be prepared in advance. 11. Although protein-based blocking buffers may be used, we find that TBS containing a high Tween 20 concentration (0.5%; TTBS-high) is optimal for detection of streptavidin– biotin interactions developed using chemiluminescent techniques with minimal background issues. Blocking should always be at room temperature, even for extended periods. 12. Detection by streptavidin–HRP of biotin-labeled proteins usually gives a crystal clear background. If a background signal is significant, this indicates insufficient blocking. We have found that efficient blocking with the TTBS-high requires at least 2 h at room temperature. If the room temperature falls below 20°C (68°F) for several hours, a significant fraction of the Tween 20 comes out of solution resulting in reduced blocking efficacy. Consequently, the TTBS-high storage container should be checked for contents clarity prior to use and may be warmed up to 37°C without any adverse consequences. 13. ImageJ image processing package is available as a free download for several computing platforms from the NIH at: http://rsbweb.nih.gov/ij/. 14. This calculation will give a positive value when a regular Western blot is analyzed, where the lighter background region has a higher numeric representation in ImageJ than the darker band region. Where the protein band is so insignificant that the equation results in a very small or negative value, the equation result should be replaced with a minimal value which is at the low end of the measured range for the experiment. 15. To determine the sigmoid function parameters max (maximum value), steep (steepness of curve), and base (sharpness of curve ends), an experiment using a dilution series of a control protein must be performed. We leave the offset at 0, as exact protein quantities cannot be calculated using our methodology and the offset does not affect relative quantitation. The positive control protein to be used throughout the experiment should be synthesized using reticulocyte lysate as described above, and then run as a 1.5× dilution series on a SDS-PAGE gel and visualized by Western blot as follows. From the 1:10 stock dilution of the protein product in 3× SDS Laemmli buffer, 45 mL should be aliquoted into an Eppendorf tube, after mixing to ensure homogeneity. From this, 15 mL are loaded onto a SDS-PAGE gel, and the 15 mL are replaced with 3× SDS Laemmli buffer. After mixing by

Coprecipitation for Protein Microarray Validation

253

inverting the tube repeatedly and briefly spinning down of the tube, this step is repeated, until 10 wells have been filled on the gel. Additional wells can be loaded with a ladder or with additional dilutions or controls, as desired. The gel is then run, developed, and measurements for multiple exposures are taken, as described previously. The Mi series for each exposure is then plotted as individual curves against relative protein concentration; protein concentration should be displayed using a log scale for easy visualization. A protein concentration of 1 may be assigned to the middle dilution of the series, with the remaining dilutions being assigned relative values, i.e., 0.67 for the higher and 1.5 for the lower dilution in the series. We used Excel (Microsoft, Seattle, WA) to plot this data and then manually estimate the best-fit sigmoid function, using (1) solved for Mi, as:

Mi =

max

1 + base

−[steeping *log10 (Pi ) −offset ]

.

(2)

We defined offset as 0 and the 3 variables to be determined as cells on our spreadsheet, calculating the resulting sigmoid function based on the defined protein concentration series for the gel, and plotting the resulting series on the same plot as the measurement data. In our data, max = –850, base = –2.7, and steepness = –4, which can be used as a starting point for this calculation. Variables are estimated by changing the variable in the spreadsheet and looking for the best match between the estimated and measured curves on the plot, which changes as the variables are changed. The maximum value (max) can be estimated first using the maximal values from the plot. The sharpness of the curve ends (base) can next be determined by visual comparison. Finally, the steepness of the curve (steepness) is determined. The final curve should closely estimate the majority of measured series, with possibly the exception of severe over- and underexposures. To better visualize the match of the estimated and measured curves, the offset may be set to a nonzero value during estimation to better overlap individual measured curves.

Acknowledgments O.M. acknowledges support from a Medical Student Fellowship of the Howard Hughes Medical Institute. C.J.W. acknowledges support from the Department of Defense (W81XWH-07-10080), the Miles and Eleanor Shore Award, NCI (5R21 CA115043-2), the Early Career Physician-Scientist Award of

254

Marina, Duke-Cohan, and Wu

the Howard Hughes Medical Institute, and is a Damon-Runyon Clinical Investigator supported in part by the Damon-Runyon Cancer Research Foundation (CI-38-07). References 1. Cookson W, Liang L, Abecasis G, Moffatt M, Lathrop M (2009) Mapping complex disease traits with global gene expression. Nat Rev Genet 10:184–194 2. Goodsell DS, Olson AJ (2000) Structural symmetry and protein function. Annu Rev Biophys Biomol Struct 29:105–153 3. Yang W, Steen H, Freeman MR (2008) Proteomic approaches to the analysis of multiprotein signaling complexes. Proteomics 8:832–851 4. Ganazzoli F, Raffaini G (2005) Computer simulation of polypeptide adsorption on model biomaterials. Phys Chem Chem Phys 7:3651–3663 5. Witze ES, Old WM, Resing KA, Ahn NG (2007) Mapping protein post-translational modifications with mass spectrometry. Nat Methods 4:798–806 6. Barranco-Medina S, Lázaro JJ, Dietz KJ (2009) The oligomeric conformation of peroxiredoxins links redox state to function. FEBS Lett 583:1809–1816 7. Zhu H, Bilgin M, Bangham R, Hall D, Casamayor A, Bertone P, Lan N, Jansen R, Bidlingmaier S, Houfek T, Mitchell T, Miller P, Dean RA, Gerstein M, Snyder M (2001) Global analysis of protein activities using proteome chips. Science 293: 2101–2105

8. Knight J (2001) When the chips are down. Nature 410:860–861 9. Gstaiger M, Aebersold R (2009) Applying mass spectrometry-based proteomics to genetics, genomics and network biology. Nat Rev Genet 10:617–627 10. Biernacki MA, Marina O, Zhang W, Liu F, Bruns I, Cai A, Neuberg D, Canning CM, Alyea EP, Soiffer RJ, Brusic V, Ritz J, Wu CJ. Efficacious immune therapy in chronic myelogenous leukemia (CML) recognizes antigens that are expressed on CML progenitor cells. Cancer Res 70:906–15 11. Duke-Cohan JS, Wollenick K, Witten EA, Seaman MS, Baden LR, Dolin R, Reinherz EL (2009) The heterogeneity of human antibody responses to vaccinia virus revealed through use of focused protein arrays. Vaccine 27:1154–1165 12. Ortiz de Montellano PR, David SK, Ator MA, Tew D (1988) Mechanism-based inactivation of horseradish peroxidase by sodium azide. Formation of meso-azidoprotoporphyrin IX. Biochemistry 27:5470–5476 13. Duke-Cohan JS, Gu J, McLaughlin DF, Xu Y, Freeman GJ, Schlossman SF (1998) Attractin (DPPT-L), a member of the CUB family of cell adhesion and guidance proteins, is secreted by activated human T lymphocytes and modulates immune cell interactions. Proc Natl Acad Sci U S A 95:11336–11341

Part V Generation of Proteomic Libraries

wwwwwwwwwwwwwwww

Chapter 17 Development of Expression-Ready Constructs for Generation of Proteomic Libraries Charles Yu, Kenneth H. Wan, Ann S. Hammonds, Mark Stapleton, Joseph W. Carlson, and Susan E. Celniker Abstract We describe a method for high-throughput production of protein expression-ready clones. Open-reading frames (ORFs) are amplified by PCR from sequence-verified cDNA clones and subcloned into an appropriate loxP-containing donor vector. Each ORF is represented by two types of clones, one containing the native stop codon for expression of the native protein or amino-terminal fusion constructs and the other made without the stop codon to allow for carboxy-terminal fusion constructs. The expression-ready clone is sequenced to verify that no PCR errors have been introduced. We have made over 11,000 clones ranging in size from 78–6,699 bp with a median of 1,056 bp. This is the largest set of fully sequence-verified “movable ORFs” of any model organism genome project. The donor clone facilitates rapid and simple transfer of the ORF into any expression vector of choice. Vectors are available for expressing these ORFs in bacteria, cell lines, or transgenic animals. The flexibility of this ORF clone collection makes possible a variety of proteomic applications, including protein interaction mapping, high-throughput cell-based expression screens, and functional studies. We have transferred 5,800 ORFs to a vector that allows production of a FLAG-HA tagged protein in Drosophila tissue culture cells with a metallothionein-inducible promoter. These clones are being used to produce a protein complex map of Drosophila from Schneider cells. Key words: cDNAs, Open reading frames, Expression-ready clone collection, Expression clone collection, Tissue culture, Proteomics

1. Introduction Proteomics resources are valuable for the study of protein function on a genomic scale. The generation of easily transferable open reading frame (ORF) clones can be accomplished using recombination-based, commercially available, high-throughput cloning systems (1). Two of the more widely used systems are the Gateway cloning system from Invitrogen and the In-Fusion cloning system from Clontech. Once the donor collection is made

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_17, © Springer Science+Business Media, LLC 2011

257

258

Yu et al.

and sequence verified, ORFs are transferred with high fidelity to any of a number of acceptor vectors for protein expression. As a source of ORFs, we are using the Drosophila Gene Collection (DGC). Clones for inclusion in the DGC were selected from over 263,000 sequenced ESTs (2–4) or isolated using a targeted cDNA screening strategy (5–7). The DGC currently has 18,182 full-insert sequenced clones and contains 10,510 cDNAs whose translation exactly matches a distinct FlyBase annotated peptide (The Gold Collection). We do not have full-length representatives for very long clones, those over 10 kb. We have used the In-Fusion (Clontech) universal cloning system, a highly efficient, directional PCR cloning system that generates a proteomics-ready, easily transferable donor clone. Since this donor clone contains flanking loxP sites, the subsequent transfer of ORFs into a wide variety of acceptor expression vectors is easily accomplished using Cre recombinase (Fig. 1). We are generating two versions of each ORF from the Gold Collection that will allow for the expression of three forms of a given protein. The first version, an ORF that contains its native stop codon, can be used to generate amino-terminal fusion proteins or to express native, untagged proteins (see Note 1). The second version, an ORF without its native stop codon, can be used to generate carboxy-terminal fusion proteins (see Note 2). Each type of donor clone is fully resequenced to a quality that confirms every nucleotide at the vector junctions, as well as correct reading frame and amino acid sequence to match the intended protein. The process flow and verification steps are outlined in Fig. 2. We have designated these two expression-ready resources as XS and XO clone sets. The XS set contains donor clones that retain their stop codon, which can be used for N-terminal protein fusions or proteins without any appended tags, while the XO set consists of donor clones that do not contain their native stop codon, which can be used for C-terminal protein fusions. We have made 11,189 fully sequence-verified donor clones, of which 5,777 can be used for expression of C-terminal fusion

loxP

PCR Amplified DGC ORF Vector/ORF

In-Fusion enzyme

ORF/Vector

SD/HIS Cmr

+

ORF

loxP

Cmr

loxP

DGC Expression-ready Clone

pDNR-Dual Vector

sacB Ampr

SD/HIS

loxP

sacB

Ampr

Fig. 1. Directional cloning into the pDNR-Dual vector. PCR products are mixed with linearized vector and the In-Fusion enzyme to subclone ORFS into the pDNR-Dual vector. Amp r ampicillin resistant selectable marker. sacB Bacillus subtilis gene encoding the enzyme levansucrase allowing growth on high sucrose. SD splice donor site allows splicing in eukaryotic cells.

Development of Expression-Ready Constructs

259

N Inoculate and Prep Gold Clones

Review Agarose Gel Results

Perform PCR with O or NS Primers

Was PCR Fragment Acceptable?

Any Clones with Correct End Sequence? Y

Agarose Gel Electrophoresis

N Rearray Templates

N Clone into pDNR-Dual

Transform into E. coli host

Select Clone w/ Correct End Sequence

Y More isolates available for sequencing ?

N Predicted ORF size >1.5kb? Y

Y DNA Prep and End-Sequence

Y

More Sequence Required?

N

Clone Sequence Verified?

Pick 2 Colonies and Grow Overnight Cultures N Frozen Stocks

Rearray and Sequence Custom Primers

Y Archive ExpressR Clones

Fig. 2. Automated clone production and sequence verification process flow. The entire pipeline with decision points and rework paths is diagrammed with processes indicated by boxes and decision points indicated by diamonds. PCR primers labeled “O Primers” are used to produce clones open at the 3¢ termini to generate carboxy terminal fusion proteins. PCR primers labeled “NS Primers” are used to produce clones with a native stop to generate amino terminal fusion proteins or proteins without any tags.

proteins and 5,412 can be used for expression of untagged or N-terminal fusion proteins (Fig. 3). We have also engineered a number of vectors with affinity tags and promoters for protein expression in tissue culture, in flies and in bacteria. Using these Drosophila expression-ready clones and vectors, we have made a number of expression clone sets (available at http://www.fruitfly. org/ and https://dgrc.cgb.indiana.edu/) suitable for proteomics projects including high-throughput protein complex studies directed toward understanding global protein function. Similar expression clone collections, for example the Human ORFeome (8), although not full-insert sequence verified, have been made primarily using the Mammalian Gene Collection (9) and the Invitrogen Gateway recombinational cloning system. An online human ORFeome database, hORFDB (http://horfdb. dfci.harvard.edu), contains information regarding clones for over 12,000 human genes many that have been implicated in human development and diseases. Our complementary collections allow the study of both Drosophila and human proteins to elucidate the functions of many of the yet uncharacterized orthologs.

260

Yu et al.

Fig. 3. Size distribution of the Drosophila expression-ready clone set. Clones with a native stop (NS) are indicated with diagonal stripes and clones open at the 3¢ end (O) are indicated with solid boxes. For each size range, the bar graph plots the number of ORFs, the number of successful PCRs with subsequent transformants (PCR/Tf), and the number of sequence verified clones.

2. Materials 2.1. Cell Culture and Dilution

1. 2× YT (BD Biosciences). 2. Chloramphenicol (50 mg/mL) for selection of chloramphenicol-resistant clones. 3. Carbenicillin (50 mg/mL) for selection of ampicillin-resistant clones. 4. Deep well blocks for bacterial growth in 96-well format (E&K Scientific).

2.2. Primer Design

1. Primer3 software (http://primer3.sourceforge.net/).

2.3. PCR

1. 10 mM dNTP (NEB). 2. Finnzymes Phusion DNA Polymerase (2 U/mL) (New England Biolabs). 3. Custom Primers ordered in 96-well plates (Illumina). 4. Sterile deionized H2O. 5. 96-well PCR plate (E&K Scientific). 6. PolarSeal foil sealing tape (E&K Scientific).

2.4. Agarose Gel Electrophoresis

1. 1% agarose (Gibco BRL). 2. Running buffer 1× TAE made from 50× stock solution: 242 g Tris base, 57.1 mL glacial acetic acid, 100 mL 0.5 M EDTA ph 8.0/L. Store at room temperature. 3. DNA marker 1 kb ladder (New England Biolabs N3232L). 4. 2× Gel Loading Buffer, diluted from 6× stock solution: 0.25% bromophenol blue, 40%(w/v) sucrose in water.

Development of Expression-Ready Constructs

261

5. Ethidium bromide solution (Fluka BioChemika). 6. Model D3-14 Centipede Gel Electrophoresis System (Owl Separation Systems, 23 cm W × 14 cm L). 2.5. Enzyme Digestion to Remove Vector

1. Dpn I (20 U/mL) and NEB Buffer 4 (New England Biolabs). 2. PCR products (see Subheading 2.3). 3. PolarSeal foil sealing tape (E&K Scientific).

2.6. Purification of PCR Products using 96-Well Gel Filtration Plate Chromatography

1. G-50 Sephadex (Amersham Pharmacia). 2. MultiScreen HV 96-well plate (Millipore). 3. MultiScreen column loader, 45 mL (Millipore). 4. 1.5 oz. soft face hammer (Stanley). 5. MultiScreen centrifuge alignment frame, blue, aqueous applications (Millipore). 6. 96-well round bottom Nunc plate (VWR Scientific). 7. 96-well PCR plate (E&K Scientific).

2.7. In-Fusion Cloning Reaction

1. Purified PCR product (see Subheading 2.6). 2. In-Fusion PCR Cloning Kit (Clontech). 3. 96-well PCR plate (E&K Scientific). 4. PolarSeal foil sealing tape (E&K Scientific).

2.8. T ransformation

1. TAM1 Competent Cells supplied with SOC medium (Active Motif). 2. LB/Amp/X-Gal 100 mm plates [LB (Invitrogen); Ampicillin 100 mg/mL (Sigma); X-Gal 64 mg/mL (bromo-chloro-indolylgalactopyranoside Roche)]. 3. Reaction products (see Subheading 2.7).

2.9. Clone Selection and Growth

1. 2× YT (BD Biosciences). 2. Carbenicillin (50 mg/mL) for selection of ampicillin-resistant clones. 3. 96-Well, 2 mL, square-well, round bottom plate (E&K Scientific). 4. Breathe Seal, breathable sealing tape (E&K Scientific).

2.10. DNA Preparation by Direct Heat Lysis of Bacterial Cells

1. 96-Well PCR plate (E&K Scientific). 2. PolarSeal foil sealing tape (E&K Scientific).

262

Yu et al.

2.11. Clone Verification by Full Insert Sequencing

1. DNA template. 2. End-sequencing primers, -21 (5¢-TGTAAAACGACGGCC AGT-3¢) and PDD3.1 (5¢-GTTTTTTATTGGTGAGAATCC AAGC-3¢). 3. Custom primers. 4. Sequencing reaction buffer 5× concentrate (Applied Biosystems). 5. Big Dye v3.1 ready reaction mix (Applied Biosystems). 6. Ethanol. 7. 125 mM EDTA. 8. Sterile deionized H2O. 9. PolarSeal foil sealing tape (E&K Scientific).

2.12. Informatics Processing

1. Sequence assembly and analysis software Phrap and Consed (http://www.phrap.org). 2. Custom analysis software.

3. Methods We have used a modified Clontech In-Fusion (see Note 3) system to generate the Drosophila ORF collection. ORFs are amplified from cDNA clone templates using PCR primers tailed with sequence homologous to the donor vector. Sized verified PCR products are purified, and then the In-Fusion enzyme is used to insert each PCR amplified ORF into the pDNR-Dual donor vector (In-Fusion cloning reaction) (Fig. 1). Products from the In-Fusion cloning reaction are transformed into TAM1 cells, and two colonies are picked for sequence verification. Each clone is end sequenced and if necessary to verify the full-length sequence, clones containing longer ORFs are selected for custom primer sequencing. Only clones that perfectly match the translation of the target are accepted as successful expression-ready clones. Using this protocol with no rework, our initial success rate, defined by strict criteria of perfect sequence match to the ORF and no relevant errors in the linkers, is 78% of all targets, as shown in Table 1. This is a comparable success rate, to that reported for other high-throughput clone sets (10, 11). We expect our success rate to increase by 5–10% once the analysis of the clones in process (clones passing the verification scripts but not yet full-length sequenced) is complete. 3.1. Cell Culture and Dilution (see Note 4)

1. Inoculate 5 mL bacteria culture cells from cDNA of interest into 600 mL 2× YT/Antibiotic and grow 16–18 h at 300 rpm in a 37°C shaking platform incubator. 2. Make 100 mL of a 1:10 cell dilution in sterile deionized H2O for use in the PCR.

Development of Expression-Ready Constructs

263

Table 1 Cloning statistics Open clones Experiments attemptsa Clones screenedb Successful clonesc (completely sequenced) Successful clonesd (in progress)

Stop clones

Total

7,146

6,606

13,746

16,547

15,305

31,852

5,326

5,388

10,714

405

551

956

a First-pass, nonredundant attempts for transferring an ORF from a DGC clone. Numbers reflect progress at the time of the submission of this article. b During a pilot phase, up to eight colonies per target were screened. Based on our success rate, we determined two colonies per target, as described in this protocol, which were sufficient to achieve an 85% success rate c “Success” is defined as those clones passing the verification scripts (see Subheading 3.12) and have been fully verified d Number of clones passing preliminary tests but not full-length sequenced

3.2. P rimer Design

1. Primers should be designed with 15 bp of overlap to the ends of the linearized donor vector and 8 bp of overlap with the insert.

3.3. P CR (see Note 5)

1. Set up PCRs on ice as described in the table below. Component

Amount Amount (concentration) per well (mL) in final reaction

1:10 Cell dilution (from Subheading 3.1, step 2)

3.0

–

5× Phusion buffer

4.0

1×

10 mM dNTP

0.4

0.2 mM

5 mM 5¢ Oligo

1.2

0.3 mM

5 mM 3¢ Oligo

1.2

0.3 mM

Phusion polymerase (2 U/mL)

0.15

0.015 U/mL

Sterile deionized H2O

10.05

–

Total

20

264

Yu et al.

2. Place reactions into a thermal cycler programmed with the following parameters: Segment

Cycle

Temperature

Time

1

1

98°C

1 min

2

2–4

98°C

10 s

59→50°C, decrease temperature 3°C per cycle

30 s

72°C

2 min

98°C

10 s

50°C

30 s

72°C

2 min

3

3.4. Agarose Gel Electrophoresis

5–17

4

18

72°C

5 min

5

19

4°C

Indefinite

1. In a 1 L Erlenmeyer flask, mix 2.5 g of agarose powder and 250 mL of 1× TAE buffer. 2. Heat the mixture in a microwave for approximately 2 min or until the agarose completely dissolves. 3. Allow the hot agarose mixture to cool to 55°C. Then add 2.5 mL of ethidium bromide solution. Swirl to completely mix. 4. Pour the agarose mixture into a gel cast on a flat surface, avoiding bubbles. 5. Place two evenly spaced 50-tooth combs into the casting tray and allow the agarose to completely solidify. 6. Once the gel solidifies, remove the combs and place the gel and its tray into a Centipede D3-14 gel apparatus filled with enough 1× TAE to completely submerge the gel. 7. In a new PCR plate, combine 5 mL of 2× gel loading buffer with 5 mL of each PCR (from Subheading 3.3). Mix well. 8. Load the entire 10 mL sample. Marker DNA (1KB DNA ladder) is loaded into the first and last wells of each row for sizing. 9. Run the gel at 120 V for approximately 1 h or until the gel loading dye front from the lower sample row is approximately 1 cm from the bottom of the gel. 10. Image the gel in the Bio-Rad Gel Doc XR imager. Poststaining with ethidium bromide is not necessary, but recommended if the PCR products appear faint. 11. Analysis of the gel should reveal single banded PCR products of the expected sizes.

Development of Expression-Ready Constructs

3.5. Enzyme Digestion to Remove Parental Plasmid Template (for Ampicillin-Resistant Clones Only)

265

1. Prepare DpnI reaction mix as follows (see Note 6). Component PCR products (from Subheading 3.3)

Amount per well (mL)

Amount (concentration) in final reaction

15.0

–

NEB buffer 4

2.0

1×

DpnI (20 U/mL)

0.5

0.5 U/mL

Sterile deionized H2O

2.5

–

Total

20

2. Incubate at 37°C for 2 h. 3. Heat inactivate at 80°C for 20 min. 3.6. Purification of PCR Products Using 96-Well Gel Filtration Plate Chromatography

1. Spread Sephadex G-50 onto the MultiScreen column loader using the MultiScreen column loader scraper (http://www. millipore.com/userguides.nsf/docs/p35962). Perform this step in a secondary container to collect excess G-50 powder. Excess G-50 can be returned to the original bottle and reused. Filtration through G-50 is used to exchange the buffer and to reduce concentrations of unincorporated dNTPs and small DNA fragments (<20 bp). Consult Sephadex manufacturer’s literature for size-exclusion limits (http://www1. amershambiosciences.com/aptrix/upp01077.nsf/Content/ Products?OpenDocument&parentid=39901&modulei d=39969). 2. Apply G-50 powder to the wells of the column loader. Visually inspect wells for uniformity. Reapply G-50 if necessary. Scrape away any excess powder from the edges of the loader. 3. Invert a 96-well, MultiScreen HV plate and place it on the loader. Align the plate edge with the metal stopper bar. 4. Grasp loader-plate assembly firmly and invert so the plate is now upright, and below the inverted loader. 5. Tap the loader gently with a soft face hammer several times along the length of the loader. Keep the plate and loader aligned while doing this. Plates containing dry G-50 powder can be stored at room temperature (20–25°C). 6. Add 300 mL sterile deionized H2O to each well. 7. To hydrate G-50 let plate sit at room temperature for 3 h. 8. Place hydrated G-50 plate on top of a Nunc plate fitted with the MultiScreen centrifuge alignment frame to collect eluate. 9. Spin 950 × g for 5 min in an Eppendorf 5810 centrifuge equipped with 96-well plate holders.

266

Yu et al.

10. Discard eluate. 11. Add an additional 150 mL sterile deionized H2O to G-50. 12. Spin 950 × g for 5 min. 13. Discard eluate. 14. Place G-50 plate on top of a new PCR plate (E&K). No adapter between the two plates is necessary. 15. Place assembly on PCR plate base. 16. Transfer all of the PCR products into the G-50 columns, being careful to dispense the sample into the center of each column. 17. Spin at 950 × g for 5 min to collect sample in the PCR plate. 18. Should recover ~15–20 mL of sample. 3.7. In-Fusion Cloning Reaction

1. Prepare cloning reaction as tabulated below (see Note 7). Component

Amount per well (mL)

Amount (concentration) in final reaction

Purified PCR product (see Subheading 3.6)

2.0

~100 ng

10× In-Fusion reaction buffer

1.0

1×

10× Bovine serum albumin (BSA)

1.0

1×

pDNR-dual linearized vector (100 ng/mL)

0.5

50 ng

Diluted In-Fusion enzyme (20 U/mL)

0.5

1U/mL

Sterile deionized H2O

5.0

–

Total

10

2. Incubate at 25°C for 30 min (for clones >3 kb: incubate at 37°C for 15 min, then 50°C for 15 min). 3. Add 20 mL TE to In-Fusion cloning reaction. 4. Seal and store reactions at −20°C. 3.8. T ransformation

1. Thaw TAM1 cells (96-well plate) on wet ice for 10 min. 2. Add 1.5 mL reaction (from Subheading 3.7) to TAM1 cells using an 8-channel pipetter. 3. Tap cells gently to mix. 4. Incubate on ice for 30 min. 5. Heat shock cells at 42°C for 30 s. 6. Place cells back on ice for 2 min.

Development of Expression-Ready Constructs

267

7. Add 150 mL SOC. 8. Incubate 1 h at 225 rpm in a 37°C shaking platform incubator. 9. Plate entire volume onto LB/Carb/X-Gal plates. 3.9. Clone Selection and Growth

1. Pick two white colonies into two, separate, 96-well plates filled with 1.2 mL 2XYT/Carb blocks. 2. Grow for 16–18 h at 300 rpm in a 37°C shaking platform incubator. 3. Make two 15% glycerol frozen stock copies. 4. Use 5 mL of overnight growth for DNA preparation.

3.10. DNA Preparation by Direct Heat Lysis of Bacterial Cells

1. Transfer 5 mL of overnight culture (from Subheading 3.9) into a half-skirt PCR plate. 2. Spin at 3,200 × g for 10 min in the Eppendorf 5810 centrifuge. 3. Invert onto a paper towel and spin at 50 × g for 1 min. 4. Discard paper towel in biohazardous waste. 5. Disinfect rotor bucket, if necessary. 6. Add 5 mL sterile, deionized H2O to each well of the PCR plate. 7. Seal with PolarSeal foil tape. 8. Resuspend by vortexing. 9. Quick spin to 1,000 rpm. 10. Lyse cells at 95°C for 3 min, then cool to 4°C. 11. Quick spin to 1,000 rpm.

3.11. Clone Verification by Full Insert Sequencing

1. Prepare sequencing reactions on ice according to the table below (see Notes 8–9). Use -21 primer to sequence 5¢ end, and PDD3.1 primer to sequence 3¢ end. Component

Amount per Amount (concentration) well (mL) in final reaction

5× Sequencing buffer

1.75

0.875×

1.6 mM sequencing primer

1.0

1.5 pmol

DNA template (from Subheading 3.10)

5.0

–

BigDye Terminator v3.1 ready reaction mix

0.5

–

Sterile deionized H2O

1.75

–

Total

10

268

Yu et al.

2. Place reactions into thermal cycler, set with the following parameters: Segment

Cycle

Temperature (°C)

Time

1

1

85

1 min

2

2–26

96 50 60

10 s 5 s 4 min

3

27

4

Indefinite

3. To remove excess dye-labeled nucleotides from the completed sequencing reactions, add to each 10 mL reaction 2.5 mL of 125 mM EDTA, centrifuge to 180 × g, then add 35.0 mL of nondenatured 95% ethanol (see Note 10). 4. Seal the wells securely with PolarSeal foil sealer. 5. Invert the reaction plate four times to mix. 6. Precipitate the reaction at room temperature for 15 min. 7. Place the reaction plate in a centrifuge with a plate adaptor and spin at 3,000 × g for 30 min. 8. Discard the supernatant by carefully removing the PolarSeal sealer, inverting the reaction plate onto a paper towel folded to the size of the plate, and centrifuging the inverted plate, paper towel unit up to 185 × g. Remove the plate from the centrifuge and discard the paper towel. The supernatants must be removed completely to minimize unincorporated dye terminators remaining in the samples, blot if necessary. 9. Wash pellet with 30 mL of 70% ethanol. 10. Seal the wells securely, then invert the reaction plate four times to mix. 11. Place the reaction plate in the centrifuge and spin for 15 min at 1,650 × g and proceed to the next step immediately (as in step 7). 12. Discard supernatant as in step 8 with an increased spin time at 185 × g for 1 min (start timing when the rotor begins to move). 13. Remove the reaction plate from the centrifuge and discard the paper towel. Dry samples 15 min in a Speed-Vac; make sure that the wells are dry and keep samples protected from light. 14. To load samples immediately onto to a sequence analyzer, resuspend pellets in 15 mL of sterile deionized water to each well. To store, seal the wells as in step 4 and keep in the dark at −15 to −25°C. 15. Load sequencing samples on a conventional capillary electrophoresis instrument (e.g., ABI 3730/3730xl DNA Analyzer).

Development of Expression-Ready Constructs

3.12. Informatics Processing

269

1. Raw sequence data is assembled with phrap (12) and the consensus sequence is analyzed for sequence quality (see Note 11). 2. Clone sequences that pass initial quality tests are further analyzed to determine if the translation of the assembled sequence matches the translation of the target. 3. Clones matching the targets are checked to determine whether the linkers at either end are correct or, if not, if the error in the linker is acceptable. Errors during synthesis of PCR primers can introduce errors in the linkers of the clone sequence, but certain errors are acceptable: for XO clones, a frame shift or substitution before the start codon of the ORF is acceptable if the loxP site is preserved, and a substitution in the 3¢ linker is acceptable if it does not introduce a stop codon. Similarly, a frame shift or substitution in the 3¢ linker of a XS clone after the stop codon is acceptable, as is a substitution in the 5¢ linker for XS clones provided the loxP site is preserved and there is not a frame shift or stop codon introduced. 4. If the translation and linkers are correct, the clone is considered validated and the sequence is imported into the database (see Note 12). 5. If the translation is not correct, further sequence analysis can sometimes identify alternative, potentially successful clones that initially failed verification because of low sequence quality. Custom sequencing primers are designed with Primer3 (13) to select an equally spaced set of oligonucleotides approximately 400 bp apart on the plus strand. The clones are then resequenced using custom primers and the new data assembled into a consensus sequence with Phrap, and the data reanalyzed (see Note 13).

4. Notes 1. N-terminal or untagged proteins can be expressed in E. coli with the combination of XS ORFs and an appropriately engineered bacterial expression vector. An added advantage of the XO clones in the pDNR-Dual is they possess a 5¢ T7 promoter, where protein production using IVT/T can be achieved without moving the ORF and will form C-terminal fusion proteins containing the 6xHIS tag. 2. A slight disadvantage of the system is the lack of flexibility for expression in E. coli; because of the lack of splicing machinery in bacteria, the only choice for C-terminal tags is the 6xHIS epitope, which is “hard-wired” in pDNR-Dual and is carried

270

Yu et al.

over into expression clones derived from XO ORFs. Protein expression in a bacterial host with epitopes other than the 6xHIS tag is currently not feasible in this system but is feasible in the Invitrogen Gateway cloning system. 3. Details of the In-Fusion system can be found at http://www. clontech.com/ 4. For all laboratory procedures in this protocol, wear personal protective equipment including appropriate gloves, eye protection, and lab coat. 5. Although any thermostable polymerase can be used in the PCR, we have had high success with the Phusion DNA polymerase (Finnzymes). 6. The Dpn I enzymatic step is essential to remove cDNA library templates carrying the antibiotic resistant gene for ampicillin but can be omitted in reactions using PCR templates with any other selectable markers. 7. For economy, the In-Fusion cloning reaction as described here is half scale and half volume relative to the manufacturer’s recommended protocol. 8. The BigDye Terminator ready reaction mix has buffering capacity, such that the final buffer concentration is 1×. 9. BigDye 3.1 sequencing reactions are done at 1/16 of the manufacturer’s recommended scale. 10. For sequence reaction purification, we have had the most success with the ethanol/EDTA precipitation protocol described here; however, alternative methods are described in the Applied Biosystems 3730/3730xl DNA Analyzers Sequencing Chemistry Guide (Part Number 04331467, Rev. B). 11. Common failure modes indicating that the clone should not be used for further processing are failures in the sequencing reaction revealed by low quality data and contaminated colonies, characterized by high quality vector sequence but low quality insert sequence. Other observations from the sequence that exclude the clone from further processing are very short consensus sequence, which indicates an empty vector, a large deletion in the insert, or a significant sequence similarity to a cloning intermediate. 12. The main reasons for rejection of clones at step 4 are ORF sequence error, wrong ORF, and errors in the PCR primer within the linker, all likely to originate in the PCR amplification or in primer synthesis errors. Additional causes to reject clones include no insert or parent clone contamination. 13. For the set experiments described here, clones for 1,842/7149 XO targets and 1,961/6606 XS targets required custom primer sequencing to confirm the full-length sequence.

Development of Expression-Ready Constructs

271

Parameters for custom primers designed using Primer3 include 18–25 bp length (optimal size 22), 55–65°C melting temperature (optimal melting temperature 60°), and 30–70% GC content (optimal GC content 50%).

Acknowledgments We thank Soo Park, Xiao Chen, Bhaveen Kapadia, and Bayan Parsa for clone production. This work was supported by NHGRI grant P41HG3487 (SEC) through the Department of Energy under contract no DE-AC02-05CH11231. References 1. Marsischky G, LaBaer J (2004) Many paths to many clones: a comparative look at highthroughput cloning methods. Genome Res 14:2020–2028 2. Rubin GM, Hong L, Brokstein P, EvansHolm M, Frise E, Stapleton M, Harvey DA (2000) A Drosophila complementary DNA resource. Science 287:2222–2224 3. Stapleton M, Carlson J, Brokstein P, Yu C, Champe M, George R, Guarin H, Kronmiller B, Pacleb J, Park S, Wan K, Rubin GM, Celniker SE (2002) A Drosophila full-length cDNA resource. Genome Biol 3, research0080 4. Stapleton M, Liao G, Brokstein P, Hong L, Carninci P, Shiraki T, Hayashizaki Y, Champe M, Pacleb J, Wan K, Yu C, Carlson J, George R, Celniker S, Rubin GM (2002) The drosophila gene collection: identification of putative full-length cDNAs for 70% of D. melanogaster genes. Genome Res 12: 1294–1300 5. Hoskins RA, Stapleton M, George RA, Yu C, Wan KH, Carlson JW, Celniker SE (2005) Rapid and efficient cDNA library screening by self-ligation of inverse PCR products (SLIP). Nucleic Acids Res 33:e185 6. Wan KH, Yu C, George RA, Carlson JW, Hoskins RA, Svirskas R, Stapleton M, Celniker SE (2006) High-throughput plasmid cDNA library screening. Nat Protoc 1:624–632 7. Stark A, Lin MF, Kheradpour P, Pedersen JS, Parts L, Carlson JW, Crosby MA, Rasmussen MD, Roy S, Deoras AN, Ruby JG, Brennecke J, Hodges E, Hinrichs AS, Caspi A, Paten B, Park SW, Han MV, Maeder ML, Polansky BJ, Robson BE, Aerts S, van Helden J, Hassan B, Gilbert DG, Eastman DA, Rice M, Weir M, Hahn MW, Park Y, Dewey CN, Pachter L,

Kent WJ, Haussler D, Lai EC, Bartel DP, Hannon GJ, Kaufman TC, Eisen MB, Clark AG, Smith D, Celniker SE, Gelbart WM, Kellis M, Crosby MA, Matthews BB, Schroeder AJ, Gramates LS, St Pierre SE, Roark M, Wiley KL Jr, Kulathinal RJ, Zhang P, Myrick KV, Antone JV, Gelbart WM, Carlson JW, Yu C, Park S, Wan KH, Celniker SE (2007) Discovery of functional elements in 12 Drosophila genomes using evolutionary signatures. Nature 450:219–232 8. Lamesch P, Li N, Milstein S, Fan C, Hao T, Szabo G, Hu Z, Venkatesan K, Bethel G, Martin P, Rogers J, Lawlor S, McLaren S, Dricot A, Borick H, Cusick ME, Vandenhaute J, Dunham I, Hill DE, Vidal M (2007) hORFeome v3.1: a resource of human open reading frames representing over 10, 000 human genes. Genomics 89:307–315 9. Gerhard DS, Wagner L, Feingold EA, Shenmen CM, Grouse LH, Schuler G, Klein SL, Old S, Rasooly R, Good P, Guyer M, Peck AM, Derge JG, Lipman D, Collins FS, Jang W, Sherry S, Feolo M, Misquitta L, Lee E, Rotmistrovsky K, Greenhut SF, Schaefer CF, Buetow K, Bonner TI, Haussler D, Kent J, Kiekhaus M, Furey T, Brent M, Prange C, Schreiber K, Shapiro N, Bhat NK, Hopkins RF, Hsie F, Driscoll T, Soares MB, Casavant TL, Scheetz TE, Brownstein MJ, Usdin TB, Toshiyuki S, Carninci P, Piao Y, Dudekula DB, Ko MS, Kawakami K, Suzuki Y, Sugano S, Gruber CE, Smith MR, Simmons B, Moore T, Waterman R, Johnson SL, Ruan Y, Wei CL, Mathavan S, Gunaratne PH, Wu J, Garcia AM, Hulyk SW, Fuh E, Yuan Y, Sneed A, Kowis C, Hodgson A, Muzny DM, McPherson J, Gibbs RA, Fahey J, Helton E, Ketteman M, Madan A, Rodrigues S, Sanchez A, Whiting M, Madari A, Young AC, Wetherby KD,

272

Yu et al.

Granite SJ, Kwong PN, Brinkley CP, Pearson RL, Bouffard GG, Blakesly RW, Green ED, Dickson MC, Rodriguez AC, Grimwood J, Schmutz J, Myers RM, Butterfield YS, Griffith M, Griffith OL, Krzywinski MI, Liao N, Morin R, Palmquist D, Petrescu AS, Skalska U, Smailus DE, Stott JM, Schnerch A, Schein JE, Jones SJ, Holt RA, Baross A, Marra MA, Clifton S, Makowski KA, Bosak S, Malek J (2004) The status, quality, and expansion of the NIH fulllength cDNA project: the mammalian gene collection (MGC). Genome Res 14:2121–2127 10. Hu Y, Rolfs A, Bhullar B, Murthy TV, Zhu C, Berger MF, Camargo AA, Kelley F, McCarron S, Jepson D, Richardson A, Raphael J, Moreira D, Taycher E, Zuo D, Mohr S, Kane MF, Williamson J, Simpson A, Bulyk ML, Harlow E, Marsischky G, Kolodner RD, LaBaer J

(2007) Approaching a complete repository of sequence-verified protein-encoding clones for Saccharomyces cerevisiae. Genome Res 17: 536–543 11. Rolfs A, Montor WR, Yoon SS, Hu Y, Bhullar B, Kelley F, McCarron S, Jepson DA, Shen B, Taycher E, Mohr SE, Zuo D, Williamson J, Mekalanos J, Labaer J (2008) Production and sequence validation of a complete full length ORF collection for the pathogenic bacterium Vibrio cholerae. Proc Natl Acad Sci USA 105: 4364–4369 12. The Phred/Phrap/Consed System Home Page: http://www.phrap.org/ 13. Untergasser A, Nijveen H, Rao X, Bisseling T, Geurts R, Leunissen JA (2007) Primer3Plus, an enhanced web interface to Primer3. Nucleic Acids Res 35:W71–W74

Part VI Detection Methods

wwwwwwwwwwwwwwww

Chapter 18 Reverse Phase Protein Microarrays: Fluorometric and Colorimetric Detection Rosa I. Gallagher, Alessandra Silvestri, Emanuel F. Petricoin III, Lance A. Liotta, and Virginia Espina Abstract The Reverse Phase Protein Microarray (RPMA) is an array platform used to quantitate proteins and their posttranslationally modified forms. RPMAs are applicable for profiling key cellular signaling pathways and protein networks, allowing direct comparison of the activation state of proteins from multiple samples within the same array. The RPMA format consists of proteins immobilized directly on a nitrocellulose substratum. The analyte is subsequently probed with a primary antibody and a series of reagents for signal amplification and detection. Due to the diversity, low concentration, and large dynamic range of protein analytes, RPMAs require stringent signal amplification methods, high quality image acquisition, and software capable of precisely analyzing spot intensities on an array. Microarray detection strategies can be either fluorescent or colorimetric. The choice of a detection system depends on (a) the expected analyte concentration, (b) type of microarray imaging system, and (c) type of sample. The focus of this chapter is to describe RPMA detection and imaging using fluorescent and colorimetric (diaminobenzidine (DAB)) methods. Key words: Image acquisition, Colorimetric, Detection, Fluorometric, Immunostaining, Microarray analysis, Protein, Reverse phase protein microarray

1. Introduction Microarrays are miniaturized immunoassays incorporating a bait and capture molecule (1, 2). Protein microarrays exist in two general formats: antibody microarrays (3) or reverse phase protein microarrays (RPMAs) (2, 4). The formats differ in their immobilization and detection strategies, depending on which molecules are used as the bait molecule and which are used as the capture molecule. Protein immobilization on a nitrocellulose substratum occurs via multiple, poorly understood interactions

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_18, © Springer Science+Business Media, LLC 2011

275

276

Gallagher et al.

consisting of van der Waals, electrostatic, and hydrophobic interactions (5). The common element for each type of microarray is the immobilization of bait molecules on a substratum, either as a homogeneous or heterogeneous spot (2, 6–8). Protein microarray technology is not as straightforward as DNA-based microarrays due to the complex structure of proteins and the variety of posttranslational modifications. As yet, an analogous process to PCR amplification for nucleic acid does not exist for proteins, necessitating high-throughput technologies such as microarrays for extracting information from the proteome (9). In addition to challenges for protein microarray immobilization, this molecular variability, coupled with the wide dynamic range of protein concentrations found in any sample, presents particular challenges for detection strategies (6, 9, 10). Protein detection strategies must address the challenges posed by the complex nature of a proteome. The ability to detect low abundant proteins in a complex biological mixture is the first hurdle. In general, amplification techniques (11–14) with stringent amplification chemistries have been developed for chromogenic detection (4, 15) and fluorometric detection (16–20) of proteins. The second hurdle faced by protein microarray detection strategies is the requirement for specific high affinity antibodies and ligands as probe molecules. Antibodies cannot be manufactured with known affinity and specificity. This requires individual validation of antibody specificity and sensitivity prior to use as a probe for protein microarrays (2, 21, 22). The antibody affinity also determines the assay linearity range. Linearity can only be achieved when the concentrations of the analyte and antibody are matched to the affinity constant. Multiplexed formats containing multiple antibodies with varying affinities will not be able to achieve linearity for all analytes in each spot (2). The third hurdle affecting successful protein detection on microarrays is the denatured versus native state of proteins. Selection of lysis conditions, buffers, antibodies, and reagents are dependent on the state of the protein to be analyzed. Peptides, rather than native proteins, are often used as immunogens in antibody production. Peptide-derived antibodies may not bind to proteins in the native state, limiting the ability to detect protein–protein interactions or native, full-length proteins. The detection methods developed for protein microarrays generally depend on the microarray format and substratum. Indirect detection methods generally do not employ amplification techniques due to the direct labeling of the probe, which is either the protein of interest or a second, labeled antibody. Direct detection methods, as described in this chapter, capitalize on signal amplification techniques coupled with a chromogenic or fluorescent probe that is compatible with the substratum. Chromogenic detection of protein microarrays produces permanent signals that are easily visualized for analysis. Commonly used enzymes for the

Reverse Phase Protein Microarrays

277

chromogenic reactions are horseradish peroxidase (HRP) and alkaline phosphatase (AP). These enzymes act on a variety of colorless chemical substrates, each generating a different colored product (9) (see Note 1). Diaminobenzidine (DAB) is commonly used with HRP and has femtomolar sensitivity (4, 22–26). This level of sensitivity is necessary for applications related to biological response monitoring small volume samples such as laser-capture microdissected samples. Protein microarray detection with DAB incorporates a signal amplification procedure based on catalyzed reporter deposition of substrate using commercially available reagents Catalyzed Signal Amplification (CSA kit) (Dako, Inc. Carpinteria, CA) (11–13, 15). The procedure entails blocking endogenous peroxidase, avidin, biotin, and protein activity on the array prior to the addition of primary antibody. A biotinylated secondary antibody, directed against the primary antibody, is the starting point for signal amplification. A streptavidin–biotin complex decorates the secondary antibody. Biotinyl-tyramide deposition in the area of the antibody–streptavidin–biotin complex acts as the amplification reagent. HRP bound to the tyramide oxidizes the DAB, resulting in a brown precipitate. The DAB precipitate is stable and produces an intense signal with relatively low background (4). The signal may be further intensified with nickel, copper, silver, gold, or cobalt to enhance the staining (27). The disadvantage of DAB is its classification as a potential carcinogen and requirement for disposal following state/federal guidelines. Not all substrata are compatible with fluorescence detection strategies due to inherent autofluorescence of the material (5). Fluorophore selection depends on sample type, substratum, and emission characteristics. Cy3 and Cy5 dyes are commonly used for fluorescent detection due to their decreased dye interactions, increased brightness, and the ability to add charged groups to the dyes (e.g., streptavidin) (28). Large dynamic ranges are the hallmark of fluorescent detection systems. Therefore, fluorescence is well-suited for microarrays in which the spots have a total protein content of 1.0 mg/mL or greater, or the arrays are comprised of samples with varying amounts of total protein. Photobleaching and quenching is a disadvantage of fluorescent detection strategies and may decrease the total signal observed on a microarray. In addition to using antibodies to detect specific proteins of interest, it is also useful to measure the total protein content per microarray spot. The total protein content per spot may vary across spots due to differences in protein concentration between samples, sample evaporation during the printing process, or differences in protein content across a dilution series. Total protein content per spot is often used to normalize signal intensities between spots. Total protein staining can be either fluorometric (Sypro Ruby (Invitrogen) (29, 30) or Deep Purple (GE Healthcare)) (31) or colorimetric (colloidal gold (AuroDye™, GE Healthcare)) (32).

278

Gallagher et al.

The dye choice is dependent on (a) expected protein content per spot/sensitivity of the stain, (b) substratum, and (c) detection instrumentation. Colloidal gold produces a permanent photo stable pink/red spot. AuroDye™ Forte’s sensitivity is comparable to silver staining (0.3–10 ng/mL on a gel (33)). AuroDye™ Forte is compatible with nitrocellulose and PVDF (polyvinylidine difluoride) membranes, but not with nylon. Deep purple is a naturally occurring fluorescent compound, epicocconone, derived from the Fungus Epicoccum nigrum (31). Deep Purple is a reversible stain that binds histidine, arginine, and lysine residues. It has an excitation peak of 520 nm and emission peak of 600 nm. Deep Purple images can be acquired with a UV transilluminator or fluorescent scanner with a sensitivity of 0.25–1.0 ng/mm2 protein. Sypro Ruby staining is a permanent fluorescent protein stain, with an excitation wavelength of 280 and emission wavelengths of 450 nm/618 nm. The stain is composed of a heavy metal ruthenium complex. The stain is photostable, allowing long emission lifetime and the ability to measure fluorescence over a longer time frame, minimizing background fluorescence (29, 30). Sypro Ruby stain can detect 0.25–1.0 ng/mm2 of protein on a onedimensional (1D) gel. In this chapter, we describe a method for Sypro Ruby total protein staining of RPMAs. As individualized therapy is rapidly becoming the main focus of cancer research, there is an increasing need for precise and sensitive technology to profile the molecular circuitry of small numbers of tumor cells (2, 10, 21, 34). Our own studies have focused on RPMAs as a proteomic tool for studying cellular signaling pathways and networks that constitute drug targets for personalized therapy (2, 4, 22–26,35–38). Before the development of RPMAs, the activation of signaling pathways relied mainly on either gene expression-based analysis to coordinate upstream signaling or two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) coupled with phosphorylation detection by immunoblotting (39–42). A drawback of coordinate gene transcription profiling is the inability to accurately reflect posttranslational modifications such as protein phosphorylation known to direct cellular signaling processes. RPMAs provide exclusive protein information that gene arrays are not able to generate. RPMAs can be applied to human or animal biopsies, tissue cell aspirates, or body fluids, including bacterial lysates. The immobilization of a panel of samples, controls, and calibrators on a nitrocellulose substrate permits comparison of the activation state of multiple experimental conditions or disease states within the same array for a single endpoint. Each spot printed on an array is representative of the cellular proteome in a sample. The introduction of robotic arraying devices and slide stainers

Reverse Phase Protein Microarrays

279

permits reliable technical precision of the RPMA (22, 38, 43). Identification of low abundance cell signaling proteins and their activation state require signal amplification with optimized detection methods. The choice of a detection system, be it fluorescent or colorimetric, depends on the analyte concentration, the type of microarray imaging system available in the laboratory, and the type of sample being investigated. Precision and accuracy of RPMA data hinges on image acquisition and spot analysis (44). This chapter presents optimized protocols for signal detection, image acquisition, and data analysis of RPMAs. While some of the detection methods are specific to RPMAs, the challenges in their analysis serve to highlight the general issues of signal detection in proteomic analysis.

2. Materials 2.1. Immunostaining RPMA

1. RPMAs printed with whole cell lysates, serum, vitreous, laser-capture microdissected cell lysates, peripheral blood mononuclear cells, or other protein containing body fluids. 2. Reblot™ Mild Antigen Stripping Solution 10× (Chemicon/ Millipore). Prepare 1× solution in deionized water. 3. Phosphate-Buffered Saline (PBS) without calcium or magnesium. 4. I-Block™ Protein Blocking Solution (Applied Biosystems/ Invitrogen). Dissolve 1 g of I-Block Protein Blocking powder in 500 mL of PBS w/o calcium or magnesium on a hot plate with constant stirring (see Note 2). Cool the solution to room temperature and add 500 mL of Tween 20. I-Block solution can be stored at 4°C for 2 weeks. 5. Primary antibodies, and biotinylated, species-specific, secondary antibodies (see Note 3). 6. Dako Autostainer (Dako). 7. CSA kit (Dako). 8. Biotin blocking system (Dako). 9. Antibody diluent with background reducing components (Dako). 10. Tris-Buffered Saline with Tween (TBST) (Dako).

2.1.1. Fluorometric Signal Detection

1. IRDye®680 Streptavidin. (Light sensitive, protect from light) (LI-COR® Biosciences). 2. Bovine Serum Albumin. Prepare 1% BSA in PBS without calcium or magnesium (0.5 g BSA in 50 mL PBS). Store solution at 4°C for up to 2 months.

280

Gallagher et al.

2.1.2. Colorimetric Signal Detection

1. DAB+, Liquid Substrate Chromogen System (Carcinogenic, contact hazard: wear gloves while handling) (Dako).

2.2. Sypro Ruby Total Protein Assay for RPMA

1. Sypro Ruby Protein Fixative Solution (7% v/v Acetic Acid and 10% v/v Methanol in deionized water). Close tightly and store at room temperature. Solution is stable for 2 months. 2. Sypro Ruby Protein Blot Stain (Invitrogen).

2.3. Image Acquisition and Microarray Analysis

There are a variety of scanning devices available for fluorometric and colorimetric detection of slide arrays. Laser scanners supporting both green and red fluorophores can be used for fluorescent detection methods. The RPMA colorimetric detection system is supported by the use of any high-resolution scanner for the acquisition of images. This chapter focuses on the use of the Revolution® 4550 laser scanner and the UMAX Powerlook 2100XL flat bed scanner.

2.3.1. Fluorometric System

1. Revolution® 4550 Scanner and ArraySifter Express® software (VIDAR Systems Corporation). 2. RPMA stained using fluorescent IRDye®680 Streptavidin or Sypro Ruby Protein Blot Stain. 3. Adobe® Photoshop software. 4. Microvigene™ spot analysis software (Vigene Tech).

2.3.2. Colorimetric System

1. UMAX Powerlook 2100XL flat bed scanner (UMAX). 2. RPMA stained with DAB. 3. RPMA stained using fluorescent Sypro Ruby Protein Blot Stain. 4. ImageQuant® spot analysis software (GE Healthcare). 5. Adobe® Photoshop software.

3. Methods The RPMA technique is used to evaluate cell signaling proteins that are activated or modified in a sample population. The multiplex analysis is achieved by probing each array with a validated primary antibody allowing a direct comparison between all samples. The number of slides required for immunostaining directly depends on the number of endpoints of interest. In addition to the test slides, a negative control slide must be included in each staining run to account for any nonspecific background staining produced by the interaction of each sample with the secondary antibody. If the set of selected primary antibodies belong to

Reverse Phase Protein Microarrays

281

Fig. 1. Reverse phase protein microarray (RPMA) detection methods. (a) Colorimetric catalyzed signal amplification (CSA). Biotinyl tyramide binds to a streptavidin biotin complex on the secondary antibody providing signal amplification. Streptavidin-conjugated horseradish peroxidase (HRP) catalyzes the deposition of 3,3′-diaminobenzidine (DAB) tetrahydrochloride, which binds to the biotinyl tyramide complex and also precipitates around the antigen–antibody complex. (b) Fluorescent CSA. Streptavidin-conjugated HRP deposition of the fluorescent molecule is not used in this detection method. The LI-COR IRDye 680 is streptavidin-conjugated and binds directly to the biotinyl tyramide.

ifferent species (i.e., mouse, rabbit), it is necessary to stain one d negative control slide per staining run that corresponds to each of the primary antibody species. During data analysis, negative control values are subtracted from each endpoint to eliminate any nonspecific signal generated by the secondary antibody. The following RPMA staining method includes sections for both fluorometric and colorimetric techniques (Fig. 1). Many of the immunostaining steps are the same for both methods, therefore differences will be highlighted throughout the procedure (see Note 4). 3.1. Immunostaining Reverse Phase Protein Microarray

1. Determine the number of slides to be stained, including negative controls. If the slides were stored at −20°C, allow them to acclimate to room temperature before proceeding. 2. Treat slides with 1× ReBlot Mild Solution for 15 min with constant, low-speed rocking. This step is required for full denaturation of immobilized proteins and facilitates antibody binding (see Note 5). 3. Discard ReBlot solution and wash slides with PBS (without calcium, magnesium or phenol red) twice for 5 min with constant, low-speed rocking. 4. Incubate slides in I-Block solution for at least 60 min at room temperature with constant, low-speed rocking. This step reduces nonspecific antibody binding. Slides can also be left in I-Block solution over night at 4°C if needed.

282

Gallagher et al.

5. Select primary and biotinylated secondary antibodies. 6. Program Dako autostainer as described in Table 1 (see Note 6). 7. Prepare staining solutions in the CSA kit according to manufacturer’s instructions. The Autostainer program automatically calculates the exact amount of each solution needed for

Table 1 Dako Autostainer program for fluorometric and colorimetric staining of RPMAs Reagent category

Reagent

Rinse

Tris-Buffered Saline with Tween (TBST) buffer

Endogenous enzyme block

Hydrogen peroxide

Rinse

TBST buffer

Auxiliary

Avidin

Rinse

TBST buffer

Auxiliary

Biotin

Rinse

TBST buffer

Protein block

Protein block

Rinse

Blow air

Primary antibody

Primary antibody

Rinse

TBST buffer

Rinse

TBST buffer

Auxiliary

TBST buffer

Rinse

TBST buffer

Rinse

TBST buffer

Secondary reagent

Biotinylated secondary antibody

Rinse

TBST buffer

Auxiliary

TBST buffer

Rinse

TBST buffer

Auxiliary

Streptavidin–biotin complex

Rinse

TBST buffer

Auxiliary

TBST buffer

Rinse

TBST buffer

Auxiliary

Amplification reagent

Time (min)

5

10

10

5

30

3

15

3

15

3

15 (continued)

Reverse Phase Protein Microarrays

283

Table 1 (continued) Reagent category

Reagent

Rinse

TBST buffer

Tertiary reagenta

Streptavidin-HRPa

Rinse

TBST buffer

a

Time (min)

Auxiliary

TBST buffera

Rinsea

TBST buffera

a

15

a

3

Switch to toxic waste

–

Substrate

Diaminobenzedine

4

Substrateb

IRDye®680 Streptavidinb

15

Rinseb

dH2O

Rinse

dH2O

Auxiliary

dH2O

a

– a

840

Reagents required for colorimetric staining, omit for fluorometric staining Reagents required for fluorometric staining only, omit for colorimetric staining

a

b

the staining run. Dilute primary and secondary antibodies in Antibody Diluent (Dako) to minimize background signal. (a) Fluorometric signal amplification: prepare a 1:80 dilution of IRDye®680 Streptavidin in 1% BSA solution. Use aluminum foil to protect the diluted IRDye 680 solution from light until it is ready to be loaded in the autostainer. (b) Colorimetric signal amplification: prepare DAB+ Substrate Chromogen System by adding 1 drop of chromogen to every 1.0 mL of DAB substrate. Prepare an adequate volume as calculated by the Dako Autostainer in step 6. 8. Fill Autostainer containers with deionized water and TBST for washes (see Note 7). Ensure that the waste container is empty. 9. Load reagents and blocked RPMA slides on the Autostainer. Do not allow slides to dry during loading; 1× TBST can be used to keep them hyrdated. (a) If staining with IRDye®680 Streptavidin, cover the Autostainer with a dark cloth or plastic bag to protect slides from light during the staining run. 10. Prime water first, followed by TBST buffer, before starting the staining run.

284

Gallagher et al.

11. After the staining run is complete, allow slides to air dry and then remove them from autostainer (see Note 8). To prevent quenching/photobleaching, protect the slides from light if the arrays were stained using a fluorescent dye. 3.2. Total Protein Staining: Sypro Ruby Protein Blot Stain

The concentrations of total, phosphorylated, and cleaved proteins present in different samples can be determined by RPMAs. The signal intensity normalization process during analysis is based on total protein values, allowing the comparison of samples with different protein concentrations (Fig. 2). Sypro Ruby Protein Blot Stain is a reversible fluorescent dye that binds to primary amino groups on proteins in an acidic environment. Images of Sypro Ruby-stained slides can be acquired with a laser scanner or a CCD camera. The dye has two excitation maxima at ~280 nm and at ~450 nm and an emission maximum near 618 nm (29, 30). Typically, RPMAs are printed in batches of 50–100 arrays. The total printing time depends on the arrayer, number of pins, number of samples to be printed, etc. (8). Due to the small sample volumes used to construct arrays, there may be sample evaporation during array printing. As such, the total protein content of a given spot can vary over the series of arrays printed. There may

Fig. 2. Total protein and antibody-stained RPMAs. (a) Sypro Ruby staining. An RPMA stained with Sypro Ruby Protein blot stain detects all samples in which protein is present at a concentration greater than 0.25 ng/mL. The samples were printed in serial twofold dilution curves, in duplicate. Samples 1, 2, and 4 are bovine serum albumin (BSA) standards. Samples 3 and 5 are whole cell lysates prepared from cultured cell lines. (b) Antibody-specific colorimetric staining. An RPMA printed with the same BSA standards and cell line lysates as in (a) was probed with anti-Akt Thr308 (Cell Signaling Technology) followed by signal amplification with a CSA system utilizing DAB. Note that the BSA standards are not visible with an antibody-specific probe.

Reverse Phase Protein Microarrays

285

be an increase in total protein content of the spots from array 1 to array 100. Therefore, multiple array slides can be used to determine an average or median total protein value (see Note 9). 1. If array slides were stored frozen (−20°C), bring selected slide(s) to room temperature. 2. Wash slide(s) in deionized water for 5 min with constant, low-speed rocking. 3. Incubate slides in Sypro Ruby Protein Blot fixative solution at room temperature for 15 min with constant, low-speed rocking. 4. Discard Sypro Ruby Protein Blot fixative solution in an appropriate container and wash slides with deionized water four times for 5 min each. 5. Incubate slides with Sypro Ruby blot stain for a minimum of 30 min. Sypro Ruby is a photo-sensitive dye; therefore, protect the slides from light by covering them with aluminum foil. 6. Discard Sypro Ruby Blot stain in an appropriate container. Rinse slides with deionized water 4× for 1 min each. 7. Allow slides to air dry. Protect stained slides from the light. 8. Acquire slide images with a laser scanner, such as Revolution® 4550 Scanner (VIDAR) or a CCD camera such as the NovaRay (Alpha Innotech). 3.3. Image Acquisition Procedures

Image acquisition is a critical step of RPMA analysis. A colorimetric or fluorescent detection system can be used in conjunction with the CSA method which utilizes a HRP-mediated deposition of biotinyl tyramide at the site of the primary/secondary antibody (Figs. 1 and 2) (12–14). While the colorimetric assay is more sensitive for low abundance proteins, the fluorescence method permits a greater dynamic range in detection. Fluorescent molecules absorb photons of light from an external light source producing excited electrons in a molecule, which then emit light at a different wavelength (45). IRDye®680 streptavidin conjugate (LI-COR Biosciences) emits in the near infra-red with an absorption maximum of 683 nm and an emission maximum of 710 nm.

3.3.1. Fluorometric Image Acquisition

The following are methods to scan microarrays stained with IRDye®680 streptavidin and Sypro Ruby protein blot stain using the Revolution® 4550 Scanner (VIDAR). 1. Turn on the computer and scanner. There is a 5 min warm-up period before the scanner is ready to use (see Note 10). 2. Open the ArraySifter Express® software.

286

Gallagher et al.

To load slides: (a) Open the door by pressing in on the oval area in front of the scanner door. The door will lower automatically. (b) The scanner can only hold four slides at a time. Insert slides by holding the slide by the label, ensure the arrayed spots are facing up, and carefully insert the slide in the chosen slot. Slide #1 will be the first slot on the left. The software automatically recognizes how many slides are currently loaded on the stage by highlighting, in yellow, the position of the occupied slot. (c) Close the door by lifting and pressing on it until it is securely locked. 3. In the main window, under the “Process” section, select the “Channel” to be used for scanning the slides. The Revolution® 4550 scanner is capable of producing dual or single scans. Channel 635 (red) should be selected for slides stained with IRDye®680 and 532 (green) for Sypro Ruby-stained slides. 4. Only one slide can be previewed at any given time. To preview a slide, first deselect additional slide slots by clicking on them. The orange color highlighting the slide slots should change to dark gray and the check mark next to the slide should no longer be visible. After a single slide is selected, the “Preview” button should become visible. DO NOT click on “Preview” until the following steps have been completed: (a) Select “Options” just below the “Preview” button in the “Process” section to open the scanner options window. (b) Select the tab with the slide number you wish to preview. (c) Under “PMT setting,” change the PMT gain for the selected channel (see Note 11). (d) Laser power setting should be set at 75 for both channels. (e) Focus setting adjustment should be set at 0. (f) Select OK and then click on “Preview” (see Note 12). (g) Select “Stop” under the “Process” window to stop previewing a slide. Preview images are scanned at a recommended set resolution of 50 mm and final images at a resolution of 10 mm. If a different resolution is desired, changes can be made under “Resolution” in the “Process” window. (h) Repeat above steps for all loaded slides. 5. After PMT values are chosen for all slides, a final scan can be acquired. (a) Select “Scan” under “Process” window, a “Names for Saved Images” window will appear.

Reverse Phase Protein Microarrays

287

(b) Select “Browse” and the “Select Directory” window will appear. The file path where images can be stored can be entered manually under the “Directory Name.” (c) Under “Path for Saved Images,” “Save Single Channel” should be selected if scanning with either channel 635 or 532. If a Dual scan is desired, there are a few different options available to save an image. Dual-channel setting saves both channels in one single file; Dual-Channel Multiplex saves all three images individually. The option of saving 635 and 532 channels individually is also available under Dual scan. (d) Under “Save Images as (file names),” file names can be assigned for each slide to be scanned (see Note 13). 6. After assigning file names to all slides, select “START.” A “Scanning Information” window will appear on the bottom left only while the image is being acquired and will show both, the “Image Properties,” which shows selected scan properties, and the “Progress Status,” which tracks the scanning progress of a slide and the overall progress of a batch of selected slides when scanning final images. 7. Shut down the scanner. Remove all the slides from the stage, close scanning program, turn off scanner, and then turn off computer. 3.3.2. Colorimetric Image Acquisition

Microarrays stained with DAB can be acquired with any high- resolution (16 bit) scanner provided with a grayscale option. The following method is to be used with a UMAX, Powerlook 2100XL scanner. 1. Place slide with arrayed spots facing down on the scanner (see Note 14). 2. Open Adobe Photoshop and import the image from the scanner using the file menu: File/Import/Magic Scan. 3. Select the “Preview” button under the scanner window. Two boxes will appear: the redbox frames preview area while the dotted blue box defines the “Scan” area. Adjust these two boxes based on the desired scanning area. 4. Under the “Settings” window, select a 14-bit image. 5. Under the “Window” menu, choose select “Scanner Control.” In this window image settings are determined as follows: (a) Ensure “Manual Control” is checked. (b) Select Reflective. (c) Choose Gray 256 scales. (d) Set the resolution to 600 dpi. (e) Select No descreen.

288

Gallagher et al.

(f) Select No filter. (g) Image settings (eyedropper icons): 255 (white), 0 (black), 1.0 (gamma) (see Note 15). 6. Before starting the image acquisition, adjust the blue scan area box to cover the entire slide including the label. All the images will be sent to Adobe Photoshop from which they can be adjusted or saved. 7. After image acquisition is completed, close the Magic Scan program. The images will be opened in Photoshop. Save images as tiff files. 8. Images need to be inverted (white spots on a black background) in Photoshop for compatibility with ImageQuant spot analysis software. 3.4. Microarray Data Analysis

3.4.1. Data Analysis of RPMAs with Microvigene Software

Microvigene software is a spot-finding program, in which an individual object (circle) is drawn automatically around each array feature. Microvigene spot analysis consists of a series of userdefined, editable software parameters that must be set and saved for each array spot layout. ImageQuant is also a spot-finding program, but it can also be used to draw consistently sized objects (grids or circles) around each array feature. The initial ImageQuant setup for a consistent grid area requires a few simple parameter settings. The final output from both software programs is a value representing the pixel intensity of each spot. Software selection is entirely user-dependent and there are many examples of commercially available spot-finding software. 1. Use Adobe Photoshop software to open 16 bit tiff RPMAs images acquired with the Revolution® 4550 or UMAX Powerlook 2100XL scanners. 2. Click Image; rotate canvas to rotate the array slide horizontally so the array slide label is on the right side of the image. If not previously adjusted, invert the slide to visualize black spots with a white background on the slide (click Image/ Adjustments/Invert). Resample the image size as follows: (a) Click on image, image size, and unclick resample image. (b) Change resolution to 600 pixels/in. (c) Click resample the image. (d) Change image size to 3 in. wide × 1 in. high. (e) Click OK. (f) Save image with a new file name. Do not modify the original tiff image. 3. Open Microvigene software. 4. Open the image of the Sypro Ruby protein blot-stained RPMA.

Reverse Phase Protein Microarrays

289

5. Click on the measurement tool. Use the cross hair to measure the diameter of five random spots and calculate the average diameter. At the bottom left of the screen, the diameter (D) value for the spot being measured will be displayed. 6. Select “Options” icon in the main screen (second icon on the tool bar) to open a new window (Fig. 3). Complete the following steps: (a) Select the “Basic” tab and name the template with the name of the experiment. The template must have the . xml extension, or the file will not be saved. Under “Platemap,” select “none” from the drop-down menu. (b) In the “Number” section, select row = 1 and column = 1 under slide if analyzing one slide at a time. If working with two slides, change to Row = 2, Column = 1.

Fig. 3. Example output images from spot analysis software. (a) Microvigene™ microarray analysis software’s main window showing an RPMA with spots in 16 rows and 40 columns. Note each individual feature, or spot, on the array is indicated by a circle. This screen image shows the program parameters used in spot analysis. (b) ImageQuant (ver5.2) can be used to draw grids of a consistent area around each set of features (samples) on an RPMA. This RPMA contains cell lysate samples, standards, and controls printed in varying numbers of dilutions. Note that grid 1 in the upper left of the image consists of a 2 × 4 (rows × columns) grid, whereas grid 50 has 2 × 8 (rows × columns) grid.

290

Gallagher et al.

(c) Under Roi, select row = 1, column = 1 if only one Roi or grid will be used. If more than one grid is necessary, indicate the number of rows and columns in a slide. (d) The “Grid” section will let you select the number of rows and columns of spots that are in the grid. If the arrayed spots cover the entire slide, select Row = 16, Column = 40 and in the Subgrid select Row = 1 and Column = 1. These settings correspond to 16 rows of spots along the y-axis of the array image, and 40 columns of spots along the x-axis of the array image. (e) Under the “Spot Detection” section, set “Sensitivity” to 950. This value can be changed in increments of 25 or 50 at a later time. If set too high, it will incorrectly detect spots that are not present. (f) Indicate the average diameter of the measured spots under “Diameter.” Set “Max Replicate CV” to −1, which will deactivate the replicate CV function. (g) Ensure “Show advanced settings” and “One color only” boxes are selected. (h) Save template. Any time a change is made in the above parameters, the template should be saved. 7. Select “General” tab and ensure the following settings are selected: (a) Setting: Microvigene.ini. (b) Platform: Glass Slide. (c) Comment: Notes about the template can be inserted in this area. (d) View: Circles for empty spots. (e) Roi mode: No relocate. (f) Palette: Gray. (g) Barcode: Code 128. (h) Direction: Left to Right. (i) Low Range (%): 1 – a low range of 1% and high range of 99% allow better spot appearance without the need to change the image. (j) High Range (%): 99. (k) 3D line width: 2.0 – if this value is set to a negative value, the function will not be used by the program. (l) Select the box for “Black on white” barcode. (m) Select the box for “Static Highlight.” (n) Save template.

Reverse Phase Protein Microarrays

291

8. Select the “Grid” tab. The values shown under the “Space” and “Margin” sections will be set automatically once the regions of interest are defined by right clicking inside the roi and selecting “apply all.” Set the following settings: (a) Min signal (std): 2 (see Note 16). (b) Noise Factor: −0.5. (c) Max Spot Shift: the value should be slightly less than half the diameter of the averaged spots for better detection. (d) Flag Mode: Bkg_std. (e) Signal Mode: Variable size circle. (f) Max Circle Shift (pixel): 5. (g) Save template. 9. Select the “Spot” tab and set the following parameters: (a) Background: Variable. (b) Edge Factor: 1 – this function improves the detection of the spot edge. (c) Buffer Zone: 2. (d) Bkg Percentile: 25 or 50 (see Note 17). (e) Smooth: 2 – smoothing minimizes jagged edges of the detected spots. (f) Inflate Size: −2. (g) Under the “Dust” section set: Threshold to −6 Set Max Diameter to a value slightly less than half the spot diameter Min Mean to 0.05 (h) Under the “Spots” section, all available boxes must be unchecked, except Hollow, which should be selected. Positive: Uncheck box (Verify that the background value is a higher magnitude than spot intensities). In the “Poor and Good Spots” section, if average spot diameter is about 16, change Min and Max diameters to the following: Poor Min Diameter: 10 (16 − 6), Good Min Diameter: 12 (16 − 4) Poor Max Diameter: 22 (16 + 6), Good Max Diameter: 20 (16 + 4) Poor Min Solidity: 35, Good Min Solidity: 65 Poor Max Circularity: 3.0, Good Max Circularity: 2.0 Poor Aspect Ratio: 1.5, Good Aspect Ratio: 1.3 (i) Save template.

292

Gallagher et al.

10. Information under the “File” tab does not require any changes. 11. Select “Output” tab and add mean_net, mean_total, cv_total, mean_bkg, and bkg used to the output column. These are the intensity values and calculations that will be produced/ calculated. Set the following parameters: (a) Normalize: All Mean (b) Warp in every: 4 row and col. (c) Exclude top %: 0.0. (d) Fold Change: No Images. (e) Primary Output: Mean. (f) Exclude bottom %: 0.0. (g) Save template and select OK to close the “Options” window. 12. On the main screen, select the project’s template from the drop-down menu at the far right end of the toolbar. 13. If there are any grids on the slide, select “Edit” and then click on “delete all objects.” 14. Click on the “draw roi” icon and place the cross hair on the top left corner besides the first spot. Drag the cross hair across the slide to include all arrayed spots. 15. Click on the “Select” icon, light blue boxes will appear around the slide. Red lines will appear if the mouse is close to the blue boxes. Hold down the shift key and use the arrow to move the red line closer to the spots. Red lines should be approximately one spot apart from the printed spots. 16. Right click on the image and select “apply all.” 17. Save template (third icon from left side of the toolbar). 18. Click on the “Go” icon to find all spots in an array (see Note 18). 19. Save template and then click on the “save text and spots” icon (see Note 19). 20. A platemap must be created to label all samples and their locations on the array. (a) Click on “Edit,” select “Platemap,” then “save as” and name the platemap with the name of the experiment. (b) A pop up #30 window will appear which says no platemap has been assigned, do you want to assign one? Click NO. It is important to familiarize yourself with how the row and column numbers change as you move through the platemap.

Reverse Phase Protein Microarrays

293

(c) The “Editor Screen” will appear, in the drop-down box under dilution select “All Column” and add the desired dilutions for the first row of samples. As the dilutions are filled in, Microvigene will change the dilutions in all subsequent rows to match the first row (see Note 20). (d) Click on “save as” to save the changes. 21. Close Microvigene program and retrieve the platemap file as follows: (a) To open the plate map file, open “My computer,” select “Local Disk (C drive),” open the “program files folder,” and select the folder named “VigeneTech,” open the “Microvigene folder,” select the “Platemap folder,” right click on the platemap, and open with Excel. 22. Make required changes to the sample names in the “Gene ID column.” It is imperative that all dilutions and replicates for the same sample have the identical name or they will appear as different samples in the data output. (a) Designate any blank or empty spaces on the grid as “5” in the “Control column.” (b) Save the file, open Microvigene program, and then open the .vgs file of the total protein slide (Sypro Ruby-stained slide). (c) Select the “Options” icon; from the “platemap” dropdown box, select the platemap assigned to the project. (d) Save template and select OK to close the “Options” window. 23. Select the “Analysis” icon, click on “Plugins,” and then select “Option on Dilution Data.” Under the General Tab select the following settings: (a) Intensity: Mean Net (Mean Net is the average pixel intensity with local background subtraction). (b) Group By: Individual Mean (This indicates that the average intensity of replicate spots will be calculated). (c) Normalize: None. Select “None” for total protein slide and “Total Protein” for antibody slides. (d) Outliers: no flag (Indicates that no flags inserted during spot detection will be altered). (e) Output: Row-based (indicates that every sample value will be listed). (f) The y0 = 100, cv_pct = 25, and linear_pct = 25 are the only boxes required to be selected (see Note 21). (g) Click on the “Curve Fit” tab and select the following settings:

294

Gallagher et al.

Fit Model: Logistic-log Linear: Tangent line (hide) Curve Fit: Only the Log Y box should be checked Replicate: 2 (indicates number of sample replicates) Linear Range: 0.1 Curve Fit: 50 Output: Linear X and Linear Y boxes should be checked for linear scale. If the data are desired in natural log scale, they can be left unchecked Roi Control: No boxes should be checked in this section Scale: Auto x-axis and Auto y-axis should be checked (h) Click OK and close window. 24. Select the “View dilution data” icon, and a new window will appear. Each sample dilution will be listed as a separate line with all corresponding parameters for that sample dilution. 25. Click on “Save total protein.” This file will be saved in the folder where the image of the Sypro Ruby-stained slide is located. The file extension is ref_dxt. 26. Select “Save data to file” to save a regular .dxt file with all output parameters. 27. Select “Save text and spots” icon. 28. In the main screen, select the “View dilution curve” icon. Click on “Save data to file.” 29. Click on “Analysis,” select “Plugins,” and then click on “Option on Dilution Data.” 30. Select the “Browse” button to the right of the “Append total protein” drop-down box to select the total protein slide_ref file, click open, then click on “Append total protein.” Select the total protein_ref file and then “save.” Click OK. 31. Click on the “Save text and spots” icon. 32. To analyze the Negative control slide, open negative control slide image and repeat steps 11 through 19 (see Note 22). (a) If less that 90% of the spots are flagged, click on “Analysis,” select “Plugin,” click on “Option in Dilution Data.” (b) Under “Append total protein,” select “none” from the drop-down menu. Click on the “save negative slide” button. Check that the (nxt) file name is correct and click OK. (c) Click on the “Save settings” icon and then click on “Save text and spots.” (d) Close the image.

Reverse Phase Protein Microarrays

295

33. To analyze an Antibody slide, open an antibody array and repeat steps 11 through 19. (a) Click on “Analysis,” select “Plugins,” and then click on “Option on dilution data.” (b) Under “Normalize” select “Total Protein.” (c) Use the “browse” button, next to the “negative control” drop-down box to select the required negative control slide to be subtracted. (d) Use the “browse” button next to the “Append total protein” drop-down box to select the total protein slide required for normalization. (e) Click OK. A message will appear, “Negative slide file between Protein and Antibody is not the same, Continue?” Click Yes. (f) In the main window, select “View dilution data” and then click on “Save data to file.” (g) Click on append horizontally. A box named “Enter File and Protein” will appear. Name the file to create a spreadsheet containing a collection of all negative control subtracted and total protein normalized values for each of the endpoints. (h) After naming the files click OK. (i) Click on the “Save settings” icon and then click on “Save text and spots.” (j) Close the program and repeat above steps for additional antibody-stained slides. 3.5. Data Analysis of RPMAs Stained with DAB Using ImageQuant Software

1. Open ImageQuant program. 2. Open the inverted tiff image of the scanned slide. A warning will appear “Unable colormap settings for this image?” Click OK. A second warning will appear “This image does not have colormapper field! Would you like to make a copy which is compatible with ImageQuant 5.0?” Click YES and save the image as a “Gel” file. 3. Enlarge image size using F9 key function or adjusting the magnification to a maximum of 500%. 4. Activate the “grid tool” in the left toolbar to make a grid that encompasses all spots for one sample, including all dilutions and replicates. Grid dimension depends on the number of spots per sample. If samples are printed in duplicate in a four point dilution curve, use a 2 × 4 grid. Draw the grid on the first sample (Fig. 3b) (see Note 23). (a) Select the grid with the four-headed black “arrow” tool and copy the grid. Paste the grid on the next set of spots corresponding to sample #2. Each grid is assigned a

296

Gallagher et al.

number based on the order in which the grids are either drawn or copied. (b) Align the grids by selecting them and moving them only when a four-headed black arrow is visible (see Note 24). The grids have to be drawn/copied in numerical order to ensure grid/data correspond to the correct sample order in the Volume Report and Excel data file (see Note 25). (c) Click “Analysis” in the toolbar and select the following settings in “Volume Report Setup.” These settings have to be specified only at the beginning of the first ImageQuant session; they are automatically retained unless changed and can also be set as Preferences. Volume Report Setup: (a) Select “Object Name” box; (b) Select “Volume”; (c) Select “Background Value.” 5. For each slide to be analyzed, copy and paste the grids from the initial array. Adjust the position of the grids using the arrow keys. First set the Background Correction: (a) Select all grids in the inspector window; (b) Choose “Local Average”; (c) Click on “Set.” 6. Select “Volume Report”: (a) Select “Display” box; (b) Click on “Report”; (c) Close the “Inspector” window. 7. Double-click in any cell on the “Volume Report” to open the data in Excel. 8. The pixel intensity is recorded in the “Volume” column. The local area background for each spot is recorded in the “Bkgd” column. 9. Use the “Save As” function to save the Excel file. 10. Close the Excel file. Close the Volume Report file. A warning message appears stating “Data must be saved from Excel activate Excel?” Click “no.” Close the Inspector window. 11. Copy all the grid objects. Close the current gel file image. Open the next tiff array image. Paste the grids on the image. Align the grids and analyze in steps 8–13. 12. Analyze each antibody-stained array, the negative control array and the Sypro Ruby Protein Blot-stained arrays. 13. Compile all the data for every antibody slide with its corresponding negative control. 14. Calculate an average total protein value from the Sypro Rubystained arrays. 15. To normalize data to the total protein concentration per spot, subtract the negative control antibody intensity value from the corresponding antibody intensity. Divide this value by the corresponding spot’s average total protein value. This calculation corrects for any signal generated due to nonspecific binding of the secondary antibody (2, 22, 26, 38, 46).

Reverse Phase Protein Microarrays

297

4. Notes 1. Some chromogens produce alcohol-soluble precipitates, requiring the absence of organic solvents in the test system after the reaction is complete. 2. Heat the I-Block solution on a hot plate at a low to medium temperature for 10–15 min. Do not allow the solution to boil. I-Block is a casein-based protein solution and boiling will alter blocking efficiency by causing protein degradation. 3. Primary antibodies to be used on a RPMA must be validated by Western blotting. Samples similar to those on the array must be used to verify specificity to the protein of interest. A single band on the blot, at the specified molecular weight, provides an indication of affinity for the proposed antibody. 4. The ideal substratum for RPMAs should have minimal autofluorescence, limited nonspecific binding, high surface area to volume ratio, inert to biological molecules, and compatibility with available detection methods. Nitrocellulose, nylon, and polyvinylidine difluoride (PVDF) are microporous surfaces and membranes that are compatible with RPMAs and most fluorescent or colorimetric detection systems. 5. Do not exceed the ReBlot incubation time of 15 min. ReBlot is an alkaline solution that can potentially cause the nitrocellulose to detach from the glass slide. Do not use ReBlot when serum samples are printed on the arrays. The solution will cause diffusion of the sample on the nitrocellulose. 6. The Autostainer is designed to periodically add TBST to the slides at the completion of a staining run if the operator does not terminate the run. The buffer will cause salt crystal formation on the nitrocellulose. The salt crystals interfere with spot detection/analysis. The Autostainer can be programmed to run overnight by including an additional water rinse, for 14 h (840 min), after the slides are rinsed with water following the detection reagent step. This optional water rinse allows the operator to start a staining run in the afternoon and to leave the slides on the autostainer until the following morning without TBST being added to the slides at the completion of the run. 7. Use deionized water to prepare all the solutions including TBST buffer. Clean the water and buffer containers periodically with 70% ethanol to avoid mold/bacterial growth. 8. DAB-stained arrays should be rinsed in running deionized water to remove TBST and DAB residue from the slide. This residue can cause poor spot detection and increased background in the scanned image. Alternatively, the glass slide (not the nitrocellulose) can be wiped clean using a KimWipe® dampened with 70% ethanol or by using an alcohol prep pad.

298

Gallagher et al.

9. To obtain an average total protein value for all array slides constructed during a print run, stain every 25 slides with Sypro Ruby Protein Blot stain. For example, in a study set composed of 100 microarrays, it is suggested to stain slides 12, 37, 62, and 87 with Sypro Ruby Protein blot stain to normalize the total protein values for the microarray set. 10. The 5 min warm-up period is to ensure laser stability. If the scanner is not ready to use, a warning message will appear on the screen. The warning can be canceled and a scan can be performed but is not recommended. 11. Slides stained with IRDye®680 can initially be scanned at a PMT gain of 25 and a gain of 23 for Sypro Ruby-stained slides. PMT values can be changed accordingly after previewing a slide. A spot intensity value of about 65,000 is an indication of saturation. Spot intensity values can be checked during or after previewing a slide by placing the cursor directly on a spot; an image of the spot of interest can be viewed under the “Feature Viewer” window. A good working intensity average is between 40,000 and 50,000. 12. After previewing a slide, PMT values may need to be adjusted. Repeat mentioned steps starting at Subheading 3.3.1, step 4c, to change PMT values. Preview scans can be stopped at any time, especially if spots are clearly saturated (white spots). 13. The default file name is ASE 2009-06-11 14-18-27 Dual 10 mm Slide 1.tif. ASE stands for ArraySifter Express®; 200906-11 is the date the scan was made; 14-18-27 is the time; Dual is the channel selected; 10 mm is the resolution of the final scan; and Slide 1 refers to the position of the slide. 14. Cover slides with a few pieces of white copy paper to avoid possible light artifacts during image acquisition. 15. Each scanner should be calibrated for optimal image acquisition. The gamma setting can be adjusted using the calibrated grayscale strip provided with the scanner. 16. Min Signal is used to determine the cut-off value for spots to be included in the data analysis. It represents the number of standard deviations above local background. 17. The value of Bkg Percentile depends on the level of background on the slide. Use a lower value when the stained array has a high background and use higher value when there is less background. 18. A blue color circle represents a good spot. If there are any purple or green circles, select them and click on the F3 function key to redetect the spots until they change to blue. In some instances, the blue circle will not properly enclose the whole diameter of a spot, and the F3 function can help correct

Reverse Phase Protein Microarrays

299

this problem. Red Xs over spots (flagged spots) represent intensity values that are not 2 standard deviations above background. If there are spots that are flagged and the operator decides to include them in analysis, click on the “flag spots” button and then right click on the spot to correct it. 19. Selecting the “Save template” icon creates a .vgs file that consists of the slide image overlaid with its corresponding grid. In contrast, by clicking on the “save text and spots” icon, a . txt file is created which contains all the raw data. 20. Changing the dilutions for each spot within Microvigene is relatively easy as long as the array has been printed in such a way that every column of spots corresponds to the same dilution. If only the controls have dilutions that are different from the rest of the platemap, they can easily be changed using Excel. 21. y0 Refers to the final spot value; cv_pct is the Coefficient of Variation between replicate spot values; linear_pct refers to the number of spots used to determine the value of y0. 22. If >90% of the spots are flagged, do not subtract the negative control values for the spots. This is because even though the spots are flagged, each spot value is subtracted. Therefore, if the value is negative, it will be added to the antibody value rather than subtracted. 23. Draw the grids by positioning the mouse cursor near the leftmost spot in the first row. Click and drag the grid so the grid encompasses all the spots. The red grid lines should not touch the spots. The red grid line is used to calculate the local area background intensity around each spot. 24. Note that there are two different four-headed arrows (big/ black and small/white). Each arrow has a distinct function. The larger/black arrow is used to move the grid. The smaller white arrow is used to resize the grid. Precise spot analysis is best achieved using the same size grid for each antibody slide. Do not resize the grid once it is drawn unless every grid for each array is resized. If the grid is inadvertently resized, it is best to delete the grid and either redraw it or copy and paste the grid again. The Sypro Ruby Protein Blot-stained arrays will have a unique set of grids compared to the antibodystained arrays due to the differences in image size generated from the laser scanner or CCD camera. 25. Once the dimension of a grid is changed, it cannot be canceled or changed back. The program will only allow the image to be closed and reopened again. For this reason, it is important to save the analysis very often to avoid loosing grids that have been allocated correctly.

300

Gallagher et al.

References 1. Ekins R, Chu F, Biggart E (1990) Multispot, multianalyte, immunoassay. Ann Biol Clin (Paris) 48:655–666 2. Liotta LA, Espina V, Mehta AI, Calvert V, Rosenblatt K et al (2003) Protein microarrays: meeting analytical challenges for clinical applications. Cancer Cell 3:317–325 3. Haab BB (2001) Advances in protein microarray technology for protein expression and interaction profiling. Curr Opin Drug Discov Devel 4:116–123 4. Paweletz CP, Charboneau L, Bichsel VE, Simone NL, Chen T et al (2001) Reverse phase protein microarrays which capture disease progression show activation of prosurvival pathways at the cancer invasion front. Oncogene 20:1981–1989 5. Tonkinson JL, Stillman BA (2002) Nitrocellulose: a tried and true polymer finds utility as a post-genomic substrate. Front Biosci 7:c1–c12 6. Zhu H, Snyder M (2003) Protein chip technology. Curr Opin Chem Biol 7:55–63 7. MacBeath G, Schreiber SL (2000) Printing proteins as microarrays for high-throughput function determination. Science 289: 1760–1763 8. Espina V, Mehta AI, Winters ME, Calvert V, Wulfkuhle J et al (2003) Protein microarrays: molecular profiling technologies for clinical specimens. Proteomics 3:2091–2100 9. Espina V, Woodhouse EC, Wulfkuhle J, Asmussen HD, Petricoin EF III et al (2004) Protein microarray detection strategies: focus on direct detection technologies. J Immunol Methods 290:121–133 10. Celis JE, Gromov P (2003) Proteomics in translational cancer research: toward an integrated approach. Cancer Cell 3:9–15 11. Hunyady B, Krempels K, Harta G, Mezey E (1996) Immunohistochemical signal amplification by catalyzed reporter deposition and its application in double immunostaining. J Histochem Cytochem 44:1353–1362 12. Bobrow MN, Harris TD, Shaughnessy KJ, Litt GJ (1989) Catalyzed reporter deposition, a novel method of signal amplification. Application to immunoassays. J Immunol Methods 125:279–285 13. Bobrow MN, Shaughnessy KJ, Litt GJ (1991) Catalyzed reporter deposition, a novel method of signal amplification. II. Application to membrane immunoassays. J Immunol Methods 137:103–112

14. Bobrow MN, Litt GJ, Shaughnessy KJ, Mayer PC, Conlon J (1992) The use of catalyzed reporter deposition as a means of signal amplification in a variety of formats. J Immunol Methods 150:145–149 15. King G, Payne S, Walker F, Murray GI (1997) A highly sensitive detection method for immunohistochemistry using biotinylated tyramine. J Pathol 183:237–241 16. Wiese R (2003) Analysis of several fluorescent detector molecules for protein microarray use. Luminescence 18:25–30 17. Panchuk-Voloshina N, Haugland RP, BishopStewart J, Bhalgat MK, Millard PJ et al (1999) Alexa dyes, a series of new fluorescent dyes that yield exceptionally bright, photostable conjugates. J Histochem Cytochem 47: 1179–1188 18. Lesaicherre ML, Uttamchandani M, Chen GY, Yao SQ (2002) Antibody-based fluorescence detection of kinase activity on a peptide array. Bioorg Med Chem Lett 12:2085–2088 19. Ekins R, Chu F, Biggart E (1990) Fluorescence spectroscopy and its application to a new generation of high sensitivity, multi-microspot, multianalyte, immunoassay. Clin Chim Acta 194:91–114 20. Bacarese-Hamilton T, Mezzasoma L, Ingham C, Ardizzoni A, Rossi R et al (2002) Detection of allergen-specific IgE on microarrays by use of signal amplification techniques. Clin Chem 48:1367–1370 21. Templin MF, Stoll D, Schrenk M, Traub PC, Vohringer CF et al (2002) Protein microarray technology. Trends Biotechnol 20:160–166 22. VanMeter AJ, Rodriguez AS, Bowman ED, Jen J, Harris CC et al (2008) Laser capture microdissection and protein microarray analysis of human non-small cell lung cancer: differential epidermal growth factor receptor (EGPR) phosphorylation events associated with mutated EGFR compared with wild type. Mol Cell Proteomics 7:1902–1924 23. Gulmann C, Sheehan KM, Conroy RM, Wulfkuhle JD, Espina V et al (2009) Quantitative cell signalling analysis reveals down-regulation of MAPK pathway activation in colorectal cancer. J Pathol 218: 514–519 24. Gulmann C, Espina V, Petricoin E III, Longo DL, Santi M et al (2005) Proteomic analysis of apoptotic pathways reveals prognostic factors in follicular lymphoma. Clin Cancer Res 11:5847–5855

Reverse Phase Protein Microarrays 25. Petricoin EF III, Espina V, Araujo RP, Midura B, Yeung C et al (2007) Phosphoprotein pathway mapping: Akt/mammalian target of rapamycin activation is negatively associated with childhood rhabdomyosarcoma survival. Cancer Res 67:3431–3440 26. Wulfkuhle JD, Speer R, Pierobon M, Laird J, Espina V et al (2008) Multiplexed cell signaling analysis of human breast cancer applications for personalized therapy. J Proteome Res 7:1508–1517 27. Hsu SM, Soban E (1982) Color modification of diaminobenzidine (DAB) precipitation by metallic ions and its application for double immunohistochemistry. J Histochem Cytochem 30:1079–1082 28. Pawley JB (1995) Handbook of biological confocal microscopy. Plenum, New York 29. Berggren K, Steinberg TH, Lauber WM, Carroll JA, Lopez MF et al (1999) A luminescent ruthenium complex for ultrasensitive detection of proteins immobilized on membrane supports. Anal Biochem 276:129–143 30. Berggren KN, Schulenberg B, Lopez MF, Steinberg TH, Bogdanova A et al (2002) An improved formulation of SYPRO Ruby protein gel stain: comparison with the original formulation and with a ruthenium II tris (bathophenanthroline disulfonate) formulation. Proteomics 2:486–498 31. Mackintosh JA, Choi HY, Bae SH, Veal DA, Bell PJ et al (2003) A fluorescent natural product for ultra sensitive detection of proteins in one-dimensional and two-dimensional gel electrophoresis. Proteomics 3: 2273–2288 32. Fowler S (1996) Protein staining and immunodetection using colloidal gold. In: Walker J (ed) Protein protocols handbook. Humana, Totowa, NJ, pp 275–287 33. Switzer RC III, Merril CR, Shifrin S (1979) A highly sensitive silver stain for detecting proteins and peptides in polyacrylamide gels. Anal Biochem 98:231–237 34. Huels C, Muellner S, Meyer HE, Cahill DJ (2002) The impact of protein biochips and microarrays on the drug development process. Drug Discov Today 7:S119–S124 35. Grote T, Siwak DR, Fritsche HA, Joy C, Mills GB et al (2008) Validation of reverse phase protein array for practical screening of potential biomarkers in serum and plasma: accurate

36.

37. 38.

39.

40.

41.

42.

43.

44.

45.

46.

301

detection of CA19-9 levels in pancreatic cancer. Proteomics 8:3051–3060 Korf U, Derdak S, Tresch A, Henjes F, Schumacher S et al (2008) Quantitative protein microarrays for time-resolved measurements of protein phosphorylation. Proteomics 8:4603–4612 Korf U, Wiemann S (2005) Protein microarrays as a discovery tool for studying protein-protein interactions. Expert Rev Proteomics 2:13–26 Espina V, Liotta LA, Petricoin EF III (2009) Reverse-phase protein microarrays for theranostics and patient tailored therapy. Methods Mol Biol 520:89–105 Zhou G, Li H, DeCamp D, Chen S, Shu H et al (2002) 2D differential in-gel electrophoresis for the identification of esophageal scans cell cancer-specific protein markers. Mol Cell Proteomics 1:117–124 Washburn MP, Ulaszek R, Deciu C, Schieltz DM, Yates JR III (2002) Analysis of quantitative proteomic data generated via multidimensional protein identification technology. Anal Chem 74:1650–1657 Gygi SP, Rist B, Gerber SA, Turecek F, Gelb MH et al (1999) Quantitative analysis of complex protein mixtures using isotope-coded affinity tags. Nat Biotechnol 17:994–999 Gorg A, Obermaier C, Boguth G, Harder A, Scheibe B et al (2000) The current state of twodimensional electrophoresis with immobilized pH gradients. Electrophoresis 21:1037–1053 Tibes R, Qiu Y, Lu Y, Hennessy B, Andreeff M et al (2006) Reverse phase protein array: validation of a novel proteomic technology and utility for analysis of primary leukemia specimens and hematopoietic stem cells. Mol Cancer Ther 5:2512–2521 Stanislaus R, Carey M, Deus HF, Coombes K, Hennessy BT et al (2008) RPPAML/RIMS: a metadata format and an information management system for reverse phase protein arrays. BMC Bioinformatics 9:555 Newland J, Jones J (1980) Fluorometry. In: Hicks R, Schenken JR, Steinrauf MA (eds) Laboratory instrumentation. Harper & Row, Hagerstown, MD, pp 61–66 VanMeter A, Signore M, Pierobon M, Espina V, Liotta LA et al (2007) Reverse-phase protein microarrays: application to biomarker discovery and translational medicine. Expert Rev Mol Diagn 7:625–633

wwwwwwwwwwwwwwww

Chapter 19 Förster Resonance Energy Transfer Methods for Quantification of Protein–Protein Interactions on Microarrays Michael Schäferling and Stefan Nagl Abstract Methods based on Förster (or fluorescence) resonance energy transfer (FRET) are widely used in various areas of bioanalysis and molecular biology, such as fluorescence microscopy, quantitative real-time polymerase chain reaction (PCR), immunoassays, or enzyme activity assays, just to name a few. In the last years, these techniques were successfully implemented to multiplex biomolecular screening on microarrays. In this review, some fundamental considerations and practical approaches are outlined and it is demonstrated how this very sensitive (and distance-dependent) method can be utilized for microarraybased high-throughput screening (HTS) with a focus on protein microarrays. The advantages and also the demands of this dual-label technique in miniaturized multiplexed formats are discussed with respect to its potential readout modes, such as intensity, dual wavelength, and time-resolved FRET detection. Key words: FRET, Protein microarray, Biomolecular interaction, Screening, Time-resolved fluorescence

1. The Basis of FRET Detection Energy transfer is a widespread phenomenon in nature and can be observed on many length scales from stellar to subatomic dimensions. Short-range energy transfer processes (typically <2 nm) are often based on electron transfer or exchange mechanisms such as exciton delocalization or Dexter energy transfer and play an important role in materials chemistry and technology (e.g., formation and movement of electron/hole pairs in semiconductor materials) and nature (e.g., electron transport chains to photosynthetic reaction centers in chloroplasts). In the field of

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_19, © Springer Science+Business Media, LLC 2011

303

304

Schäferling and Nagl

Fig. 1. Resonance energy transfer between a donor (D) and acceptor (A) indicates specific interactions of biomolecules. Reproduced with permission from Springer (35).

molecular biology, an energy transfer based on coulombic interactions with longer operational ranges has a huge spectrum of applications. It displays a distance dependence on the scale of a few nanometers (2–10 nm) and therefore matches well with the size of many biological macromolecules such as proteins, lipids, short oligonucleotide strands, sugars, and the like (Fig. 1). This type of energy transfer is now commonly referred to as Förstertype energy transfer, Förster resonance energy transfer, or fluorescence resonance energy transfer, usually abbreviated as FRET or RET. Because of the distance dependence in this length regime, it is most useful and has been extensively used for confirmation of binding events as well as for conformational studies and distance determinations as a “molecular ruler” on the nanometer scale (1–4). The energy that is transferred by means of this mechanism originates from electronically excited states of donor molecules whereby the donor is usually excited via electronic absorption of light. Alternatively, certain molecules can be transferred to an emissive excited state via chemical reactions. This chemically excited state may equally transfer its energy to another compound and the resulting mechanism is then sometimes called chemiluminescence resonance energy transfer (CRET) or bioluminescence resonance energy transfer (BRET) if the molecule is of biological origin (5). In all cases, the excess energy is being transferred into an electronically excited state of a suitable acceptor molecule.

Förster Resonance Energy Transfer Methods

305

This transfer can be realized if the acceptor is in resonance with the transition of the donor molecule back to the ground state or, in other words, the acceptor has to have an appropriate energy gap between its ground and electronically excited state that allows electronic absorption using the energy that would otherwise dissipate via fluorescence or nonradiative decay of the excited donor molecule. In more practical terms, this means that the fluorescence (or generally luminescence) spectrum of the FRET donor has to overlap with the absorbance spectrum of the acceptor and the extent of energy transfer is proportional to this overlap. This is depicted for the FRET pair Alexa 555 (donor) and Alexa 647 (acceptor) in Fig. 2. Note that the donor and the acceptor molecule may also be the same species if it has overlapping absorbance and emission spectra, which is true for most fluorescent dyes. In fact, this is probably the most common form of energy transfer and can be observed in many applications and protocols, although mostly as an undesirable side effect. A quantitative description of the FRET process is rather extensive and discussed in detail elsewhere (see e.g., ref. (6, 7)). However, qualitatively, the phenomenon can be explained using a few principles of classical electrodynamics. The FRET mechanism is based on coulomb interactions between two transient dipoles of the donor and the acceptor

Fig. 2. Spectral representation of energy transfer, exemplary for the Förster (or fluorescence) resonance energy transfer (FRET) pair Alexa 555 (donor) and Alexa 647 (acceptor). Absorbance spectra are straight lines, emission spectra are dashed lines. Spectral overlap between the emission spectrum of the donor and the absorbance spectrum of the acceptor (filled area) is required for FRET to occur.

306

Schäferling and Nagl

molecule, respectively. The transient dipoles result from the involved electronic transitions. Disregarding nonradiative excited state deactivation processes, the transition of an electron from the excited state to the ground state proceeds via a series of very fast oscillations. The electromagnetic field created by these oscillations eventually propagates into infinity and is observed as fluorescence emission. If a second molecule resides very close to the donor molecule undergoing a dipole transition and therefore reaches very deep within its transient electromagnetic field (in the so-called “near field”), its absorption dipole can resonantly couple to the field, provided that a suitable electronic transition exists. An analogy to this process would be a swinging pendulum that causes another one to swing if they are coupled by a connection. The FRET mechanism is highly dependent on the distance between the chromophore molecules as the electromagnetic fields are falling off very sharply in strength with distance. Overall, the FRET efficiency depends on the inverse sixth power of the distance between the molecules. It may be expressed in terms of the FRET quantum yield, which, in analogy to the fluorescence quantum yield, is the ratio of acceptor molecules excited by energy transfer to the number of excited donor molecules. Quantitatively, the following expression for the FRET quantum yield is obtained: −1

f ET

6  R  = 1 +    .  R0   

(1)

Fig. 3. Distance dependence of FRET expressed in terms of the Förster distance R0. Substantial FRET occurs at values of 2 R0 and below; at less than 0.5 R0, there is almost complete energy transfer efficiency.

Förster Resonance Energy Transfer Methods

307

R0 is called the Förster distance. At this donor–acceptor distance (R0 = 1.0 in Fig. 3), a FRET quantum yield of 0.5 or an energy transfer efficiency of 50% is obtained (y-axis in Fig. 3). The constant R0 may be calculated from experimentally accessible parameters using

∞

9, 000 ln10k 2φ F R = I D (λ)ε A (λ)λ 4d λ , 5 4 ∫ 128π N A n 0 6 0

(2)

where f F is the fluorescence quantum yield of the donor in absence of a FRET acceptor, n is the refractive index of the system, NA is Avogadro’s number, k is an orientation factor which is discussed below, and the integral term reflects the spectral overlap between donor emission and acceptor absorbance as already mentioned (Fig. 2). Therefore, R0 is a function of the dyes involved in the process (if there is no spectral overlap or significant luminescence emission from the donor R0, the FRET quantum yield reduces to 0), of the local refractive index, and of the relative orientation of the involved chromophores. R0 values for commonly used donor– acceptor combinations in aqueous solutions are from 2 to 7 nm. It can easily be seen from Fig. 3 that there is a sharp distance dependence centered around R0 and the analytically useful range is roughly from 0.5 to 2 R0. Below and above 0.5–2 R0, the sensitivity towards distance changes is small. If energy transfer is observed in an experimental system, it confirms that a substantial fraction of the involved donor and acceptor chromophores must be within this distance. Lastly, there is also a dependence on spatial orientation of the involved dyes in FRET which is usually expressed as the orientation factor k. Optimal coupling of the transient electromagnetic fields can be achieved when they are in a collinear arrangement (Fig. 4, k 2 = 4). Parallel arrangement of the transition dipoles reduces the FRET efficiency (k 2 = 1); perpendicular orientation prevents FRET. k for all other orientations may be calculated by angular combinations of these ideal states. It is important to note that this orientation dependency is only of relevance when the involved dyes are at well-defined positions and hindered with respect to rotational diffusion. In many experiments such as those involving freely diffusing dyes or also of labeled proteins, antibodies, or nucleotides, there are typically enough rotational degrees of freedom for both donor and acceptor that an orientation dependence is not observed and k 2 = 2/3. This value represents a dynamic average of all possible orientations. However, for fluorophores with restricted conformational flexibility such as in the solid state, e.g., immobilized in membranes, the orientation factor is important for the FRET efficiency and may deviate significantly from k 2 = 2/3.

308

Schäferling and Nagl

Fig. 4. Influence of spatial orientation of the donor and acceptor transition dipoles on FRET efficiency. The highest efficiencies result from a collinear end-to-end (k 2 = 4), or parallel (k 2 = 1) arrangement; acceptor dipoles oriented perpendicular to the donor emission dipole do not display FRET (k 2 = 0).

2. FRET in Biomolecular Assays

Many luminescent donor–acceptor pairs are well established as biomolecular labels (8) and commercially available, and there is an increasing interest in new materials such as luminescent nanocrystals or polymer particles. Novel donor–acceptor pairs for FRET assays and methods for their conjugation to biomolecules have been reviewed by Sapsford et al. (3). In case of biomolecular recognition, the donor–acceptor couple is forced into close spatial proximity, and the resulting energy transfer can be determined. This principle can be used for detection of deoxyribonucleic acid (DNA) strand hybridization (Fig. 1) or protein interactions with receptors or antibodies. Application of intermolecular FRETbased immunoassays to clinical diagnostics or pharmaceutical research presents a large field of research which has been reviewed thoroughly (9, 10). In particular, (pharmaceutical) high-throughput screening (HTS) became a major playing field for FRET assays owing to the fact that interactions of small molecules such as haptens with their corresponding antibodies can be monitored. The multitude of commercially available homogeneous screening assays for protein kinase activity or cyclic adenosine monophosphate

Förster Resonance Energy Transfer Methods

309

(cAMP) testifies the importance of FRET methods in this field. The activity of proteolytic enzymes can be analyzed by means of tailored peptide substrates conjugated to donor and acceptor dyes. In this case, intramolecular FRET is interrupted in response to enzymatic cleavage of the substrate (11, 12). Specific fluorogenic FRET substrates were also designed for other hydrolytic enzymes such as lipases or esterases (13). Furthermore, FRET probes are the basis for widely used quantitative real-time PCR techniques. These include the LightCycler probes (14), the TaqMan probes (15, 16), or the molecular beacons (16, 17). Conformational studies or intracellular assays can be carried out with green fluorescent protein (GFP)-fusion proteins (18–20). Mutants of GFP conjugated to (membrane) proteins or receptors can also act as FRET donors and acceptors. In principle, there are several ways to measure FRET in biomolecular assays. Changes of the donor emission can be monitored either by fluorescence intensity or lifetime measurements. Nonfluorescent energy absorbers, the so-called “dark quenchers,” may also be employed. If the acceptor molecule is also fluorescent, it is possible to monitor changes of the acceptor emission (sensitized emission) as a result of donor excitation. Steady-state or time-resolved fluorescence measurements can be applied to monitor FRET. Finally, the recording of both donor and acceptor emission leads to intrinsically referenced signals. The dual-wavelength detection enables a ratiometric correction as the emission from both wavelength regimes can be divided against each other. This operation makes the method less prone to interferences, which is particularly important for fluorescence imaging microscopy of living cells. Several approaches for the determination of FRET in fluorescence microscopy have been reviewed by Jares-Erijman and Jovin (21). Besides the standard organic dyes used as donor–acceptor pairs, metal–ligand complexes are also suitable donors in FRET systems. Particularly, lanthanide chelates (mainly europium und terbium) are extensively used because of their long luminescence lifetimes in the range of microseconds or even longer and their large R0 Förster distances (22, 23). Typical interferences in fluorescence detection such as scattered excitation light can also be suppressed this way. Nanoparticles of different types may also be used as FRET donors or acceptors, e.g., tetramethylrhodamine (TMR)-doped silica nanoparticles (24), ruthenium-(dpp)3 doped polyacrylonitrile (25), europium(III)-chelate doped polystyrene (26), or upconverting phosphors (27). A selection of FRET donor–acceptor systems is outlined in Table 1.

310

Schäferling and Nagl

Table 1 Förster distances R0 for some commonly employed energy transfer pairs in aqueous solution at room temperature assuming random orientation (k = 2/3) Donor

Acceptor

R0 (nm)

References

Naphthalene lAbs: 290 nm, lEm: 330 nm

Dansyl lAbs: 335 nm, lEm: 525 nm

2.2

(64)

Dansyl lAbs: 335 nm, lEm: 525 nm

Fluorescein lAbs: 490 nm, lEm: 515 nm

3.3

(1)

Pyrene lAbs: 335 nm, lEm: 390 nm

Coumarin lAbs: 375 nm, lEm: 445 nm

3.9

(65)

Fluorescein lAbs: 490 nm, lEm: 515 nm

Tetramethylrhodamine lAbs: 550 nm, lEm: 580 nm

5.4

(66)

Cy3 lAbs: 550 nm, lEm: 565 nm

Cy5 lAbs: 645 nm, lEm: 665 nm

5.4

(67)

Alexa 488 lAbs: 495 nm, lEm: 520 nm

Alexa 594 lAbs: 590 nm, lEm: 615 nm

5.4

(68)

Eu3 + a lAbs: 340 nm, lEm: 615 nm

Cy5 lAbs: 645 nm, lEm: 665 nm

7.0

(1)

Eu3 + a lAbs: 340 nm, lEm: 615 nm

Allophycocyanin lAbs: 650 nm, lEm: 660 nm

9.0

(23)

Tb3 + a lAbs: 330 nm, lEm: 545 nm

Alexa 488 lAbs: 495 nm, lEm: 520 nm

4.6

(69)

Tb3 + a lAbs: 330 nm, lEm: 545 nm

Cy3 lAbs: 550 nm, lEm: 565 nm

6.1

(70)

EGFP lAbs: 490 nm, lEm: 510 nm

DsRed lAbs: 560 nm, lEm: 585 nm

4.7

(71)

CdSe/ZnS QD lAbs: <510 nm, lEm: 530 nm

Cy3 lAbs: 550 nm, lEm: 565 nm

5.0

(72)

CdSe/ZnS QD lAbs: <470 nm, lEm: 490 nm

Dabcyl lAbs: 455 nm, lEm: –

3.4

(73)

a Lanthanide absorbance spectra are strongly dependent on the ligand, while emission spectra are mostly independent from the ligand For indication, the respective spectral maxima are displayed. Spectra and spectral maxima, however, are often strongly dependent on various experimental parameters QD = Quantum Dot

3. FRET Assays on Microarrays Most of the experiments performed with “biochips” or “microarrays” focus on the investigation of DNA interactions to probe gene expression (e.g., by differential gene expression profiling) (28), single nucleotide polymorphisms (SNPs) (29), the determination of pathogens (30), or the identification of genetically modified food and seed (31). While the impact of

Förster Resonance Energy Transfer Methods

311

DNA arrays continuously grows in functional genomics and pharmacogenomics, there is also a trend to focus on the mapping of protein interactions (32). The determination of specific proteins among other regulatory elements is a key in understanding the regulation of various cellular mechanisms, and they can act as biomarkers for certain diseases in diagnostic applications (33). This makes protein or antibody arrays promising tools for protein expression analysis, proteome research, medical diagnosis, detection of toxins in food and feed, and functional arrays for the discovery of binding domains of target molecules and receptors. The trends in protein microarrays have been extensively reviewed in the past years, as well as various aspects of their fabrication, surface chemistry, and fluorescence readout methods (34–42). Up to now, the detection of biomolecular interactions by FRET only found limited dissemination in the DNA or protein array field, although its use is attractive because it can confirm interactions at the molecular level with very high sensitivity. Some commercially available fluorescence scanning systems for microarray readout enable dual-wavelength measurements, thus being equipped for FRET experiments. As the usable wavelengths are typically fixed due to the installed lasers and emission filters, only a certain selection of donor and acceptor dyes is applicable. In addition, these instruments are not capable of time-resolved fluorescence detection. For this purpose, fluorescence imaging systems with charge-coupled device (CCD) detectors or scanning (typically confocal) microscopes based on time-correlated singlephoton counting (TCSPC) may be used. Active complementary metal oxide semiconductor (CMOS) arrays have also been developed for time-resolved fluorescence detection. In this case, the probe can be immobilized directly on or above an array of detectors. This configuration could replace the traditional scanner, resulting in an integrated and portable detection platform. The utility of this approach was demonstrated by means of an oligonucleotide microarray with TR-FRET determination (43). A FRET-based technique for SNP detection on DNA microarrays was demonstrated by Frutos et al. (44). They arrayed oligonucleotides labeled with Cy3 as donor and subsequently incubated complementary probes labeled with the FRET acceptor carboxyx-rhodamine. The probes contained a mismatch, and in a competitive assay, the Cy3-fluorescence increased after addition of a perfectly matching target. Another approach for SNP detection is related to the molecular beacon concept. The immobilized DNA probe had complementary ends and forms a stem-and-loop structure. Both fluorescent donor (TMR) and quencher (4-(dimethylaminoazo)benzene-4-carboxylic acid, Dabcyl) are linked to the two ends of the stem. This nonfluorescent structure becomes fluorescent if the loop hybridizes with its complementary target and the conformational reorganization into a rigid double helix separates the quencher from the fluorophore (45). A similar

312

Schäferling and Nagl

c oncept was used for detection of foodborne pathogens such as Escherichia coli O157:H7. The array contains four specific E. coli genes. In presence of the target gene, the complementary strand hybridizes to the probe, thereby disrupting the stem base-pairing. Thus, the distance between donor and quencher is increased and the FRET efficiency diminished (46). The molecular beacon FRET assembly avoids false negatives on this microarray platform. In another study, peptide nucleic acid (PNA)-encoded libraries consisting of 10,000 protease substrates were hybridized on DNA microarrays. Due to their PNA tags, every peptide of the library could be arranged in a defined position on the DNA microarray containing complementary sequences. The target PNA-peptide conjugates were derivatized with fluorescein as donor and carboxytetramethyl-rhodamine (TAMRA) as acceptor and screened for their protease cleavage specificity using chymopapain and subtilisin as proteolytic enzymes (47). The multiplexed detection of protein–protein interactions by FRET was first shown by means of a tissue microarray for cancer research. The aim of this method is the determination of the autophosphorylation status of epidermal growth factor receptors (EGFR) in head and neck squamous cell carcinoma as a prognostic tool. Phosphorylation of EGFR was detected by immunostaining. The arrayed cell samples were incubated with F4 (a monoclonal antibody to the cytoplasmic domain of the receptor) conjugated to Cy3B dye and FB2 (a monoclonal antibody to the EGFR autophosphorylation site) conjugated to Cy5 dye. Fluorescence lifetime imaging microscopy (FLIM) was used to measure the FRET efficiency between donor (Cy3) and acceptor (Cy5). In this assay the average fluorescence intensity of F4-Cy3b in absence of acceptor is related to the amount of overall EGFR expression, whereas the FRET efficiency is related to EGFR phosphorylation (48). A synthetic peptide microarray for HTS of protein interactions was demonstrated by Usui et al. (49). A library of 112 different derivatives of an a-helical peptide was immobilized on the surface of microwell plates and screened for their interactions with various proteins, e.g., calmodulin. The termini of the immobilized peptide probes were labeled with coumarin and fluorescein which form the FRET system. Interactions between membranes and proteins using FRET can be studied by means of supported lipid bilayer arrays. These were imaged with a total internal reflection fluorescence (TIRF) setup. The supported phospholipid bilayers consisted of two different esterified glycerol-3-phosphoethanolamines which were conjugated to a Fret donor (7-nitro-2-1,3-benzoxadiazol-4-yl) and acceptor (Texas Red), respectively (50). Energy transfer from fluorescein to TMR-doped silica nanoparticles was used to monitor interactions between estrogen receptor alpha (ERa) and steroid receptor coactivator (SRC) on

Förster Resonance Energy Transfer Methods

313

Fig. 5. Heterogeneous estrogen receptor (ER) – steroid receptor coactivator (SRC) interaction FRET assay on a microarray (left, FRET from fluorescein (Fl) to TMR) and homogeneous assay in a microwell plate (right, FRET from TMR to Cy 5). In the former, aldehyde slides were spotted with fluorescein-labeled SRC, blocked with bovine serum albumin (BSA) and incubated with ERa-coated silica nanoparticles. Ligand (L) represents different agonists or antagonists for ERa. Reproduced with permission from ACS (24).

a microarray. The silica nanoparticles were surface-modified with Ni(II)-nitrilotriacetic acid and used to immobilize the His-tagged binding domain of ERa. The nickel chelate-functionalized beads enabled a one-step protein isolation and site-specific fluorescent labeling. The FRET-binding assay with fluorescein-labeled SRC is depicted in Fig. 5. The arrays were imaged with commercially available laser scanner instruments (24). Arrays for pharmaceutical screening applications were constructed with conjugates of GFP and peptides that act as kinase substrate or can be used in ubiquitination assays. Tracer antibodies for the detection of phosphorylation or ubiquitination of these peptide substrates were labeled with a terbium complex as FRET donor. Due to the long-lived luminescence of terbium, TR-FRET measurements can be carried out (51).

4. Use of TimeResolved FRET to Detect Protein Interactions

The implementation of labels with long lifetimes (>1 ms) facilitates the detection of time-resolved fluorescence and of timeresolved FRET (TR-FRET). The luminophores are excited with a short light pulse and their emission is collected only in a defined time gate after a certain time lag. Therefore, fast decaying background fluorescence can be eliminated. The intrinsic fluorescence of biological samples usually decays within a few nanoseconds. The applicability of TR-FRET as detection method for protein interactions on microarrays was first demonstrated in 2005 by means of a model assay with immobilized biotinylated bovine serum albumin (BSA), which was additionally labeled with a longlived reactive platinum-coproporphyrin-derived compound serving as FRET donor. Subsequently, the spots were incubated with different concentrations of streptavidin labeled with the short-lived FRET acceptor Cy5 (38).

314

Schäferling and Nagl

A related concept measuring TR-FRET of donor-labeled immobilized oligonucleotides to acceptor-labeled DNA targets was presented by Schwartz et al. (43). In this approach the fluorescence lifetime of the donor (in this case a quantum dot) indicates hybridization of the target molecules. The advantages of active layer CMOS sensor arrays have been already outlined above. In this study a 64 × 64 array of single-photon avalanche diodes (SPADs) was used. The assay was carried out with immobilized oligonucleotide probes bearing terminal biotin units. The DNA targets were modified with a dark quencher (QSY 21, Invitrogen). Finally, the array was incubated with streptavidinfunctionalized quantum dots. In presence of hybridized target molecules, resonance energy transfer occurs from the quantum dot to the quencher dye. Lifetime detection measurements such as Rapid Lifetime Determination (RLD) provide little sensitivity to background signals, spatial or temporal variations in excitation light source intensity, or detector efficiency. The ratiometric measurement obtains absolute values (lifetimes) rather than relative ones as in case of fluorescence intensity, and particularly, along with FRET-based determinations, large signal changes in response to minor amounts of analyte. RLD represents a typical time-domain lifetime determination method. It should be noted that frequency-domain lifetime detection with sinusoidal-modulated light sources can also be applied for TR-FRET detection (48, 52). In our own studies, we previously generated a competitive model assay and applied biotinylated BSA and streptavidin, labeled with the commonly used fluorescent dyes Alexa 555 and Alexa 647, respectively (Fig. 2) (53). In this assay, the concentration of native (Strept-)avidin could be determined via its competition for biotin-binding sites with Alexa 647-labeled streptavidin using fluorescence lifetime imaging in the donor channel. The fluorescence was excited by 200 ps laser pulses from a mode-locked argon-ion laser, adapted to a fluorescence microscope equipped with an intensified CCD camera as detection unit allowing subnanosecond time resolution. Lifetime images were recorded according to the RLD scheme (54). Typically, fluorescence lifetimes of organic fluorophores are between 0.5 and 5 ns. Certain aromatic molecules and nitrogen heterocycles can decay much slower with lifetimes up to 20 ns. The luminescence lifetimes of compounds containing heavy atoms such as quantum dots, transition metal complexes, or lanthanide chelates can be as long as several milliseconds. Figure 6 displays the time course of a typical RLD measurement, showing an excitation pulse of approximately 200 ps full-width half maximum (solid line). A simulated single exponential decay of the fluorophore with 1 ns lifetime excited by this laser pulse is shown for illustration (dashed line). After the excitation pulse, two sets of images were acquired by means of an intensified CCD camera at set delay times of 2 and 3 ns. Both

Förster Resonance Energy Transfer Methods

315

Fig. 6. The Rapid Lifetime Determination (RLD) method: time course of a laser pulse (straight line), a simulated single exponential fluorescence decay (t = 1 ns, dashed line) and the image acquisition gates. The fluorescence lifetime is a function of the integrated intensity ratios of gate 1 to gate 2. Reproduced with permission from Elsevier (53).

Fig. 7. (a) Schematic of a competitive model assay with immobilized Alexa 555-BSA-biotin and streptavidin with and without Alexa 647 label. (b) Lifetime images of the competitive FRET assay with various amounts of unlabeled streptavidin on epoxysilane-modified glass slides. The screen shows four replicates (top to bottom) with increasing concentrations of streptavidin from the left to the right (0, 20, 40, 60, 80, 100 nmol/L). Reproduced with permission from Elsevier (53).

single images (gates 1 and 2) had an exposure time of 1 ns. In practice, the images of the two different gates are taken separately in subsequent acquisition cycles. In this way, thousands of single images can be integrated and averaged within fractions of seconds. The ratio of the integrated sets of gate 1 to gate 2 is a function of the fluorescence lifetime t. We showed that the lifetime t of the donor drops by up to 80 % (from 1.7 to 0.35 ns) on an array containing only acceptor-labeled streptavidin (Fig. 7b, left) compared to an array containing only native streptavidin (Fig. 7b,

316

Schäferling and Nagl

right), depending on the microarray substrate (Fig. 7). Four replicate spots were used for evaluation (Fig. 7b, top to bottom). In summary, FRET methods appear to be particularly useful for competitive assays, because in absence of analyte the donor fluorescence is quenched. This leads to a low and mostly constant background, even in case of unspecific binding, and a better linearity of the response. Competitive assays are established methods used in analytical protein microarrays (55, 56).

5. Concluding Remarks and Outlook

The use of FRET to monitor multiplexed biomolecular interactions on microchip formats is only at its initial stage. This is remarkable as resonance energy transfer is an established analytical tool that is applied in many bioassays. Furthermore, FRET is a very sensitive method, suitable for single molecule spectroscopy (57). One limitation is that not all commercially available microarray scanners are suited for dual-wavelength FRET experiments. On the other hand, fluorescence imaging microscopes can easily be modified for such measurements. TR-FRET is an attractive alternative for detection at a single wavelength regime. The change of the fluorescence lifetime of the donor is an intrinsically referenced parameter. But its recording requires rather complex and expensive instrumentation such as femtosecond- or picosecond-pulsed lasers and intensified CCD cameras, photon-counting photomultipliers, SPADs, or streak cameras. The progress and increasing availability of pulsed diode lasers which can generate excitation pulses of 100 ps width (and lower) with repetition rates in the MHz range and improved CMOS detectors suitable for time-resolved measurements could help to reduce the instrumental effort. Another obstacle to FRET assays is that they demand dual labeling. Two different strategies can be pursued depending on the type of biomolecular interactions that shall be studied: 1. Intramolecular FRET: Attachment of donor and acceptor dyes at the termini of the same biomacromolecule (e.g., oligonucleotides, oligo- or polypeptides). This approach is appropriate when distinct conformational changes result from binding to a target molecule and the distance between donor and acceptor changes significantly, or the substrate is cleaved by a hydrolytic enzyme. 2. Intermolecular FRET: Probe and target are labeled with donor and acceptor, respectively. The interaction of two binding partners can be monitored directly as donor and acceptor are forced into close proximity. This is a convenient design for

Förster Resonance Energy Transfer Methods

317

analytical and diagnostic assays such as immunoassays. It is applicable to direct sandwich assays as well as competitive assays. Also, functional arrays for studying protein–protein interactions could achieve higher sensitivities by implementation of FRET systems. The development of improved labels with new features is another ongoing process. The distinct sharp emission peaks of terbium complexes can be used for simultaneous multiple wavelength energy transfers to different acceptors. This capacity can be employed for dual-parameter FRET assays (58). Luminescent materials beyond the classic organic or organometallic dyes such as quantum dots or lanthanide-doped upconversion nanocrystals will also have an increasing impact on FRET techniques (59) as well as genetically encoded fluorescent proteins with programmable FRET properties (60). The optoelectronic properties of gold and silver nanoparticles should be pointed out. Under certain conditions, they appear to be accessible to energy transfer at distances (up to several tens of nanometers) that are much larger than those available to conventional FRET systems (1, 61–63, 74 and references therein). The mechanism of this phenomenon appears to be correlated with surface plasmons and is not a Förster-type dipolar interaction. While those particles have not been applied to microarray-type energy transfer studies yet, they promise to extend the available size regime for biochip interaction experiments towards the interior of larger compounds and systems such as certain antibodies, vesicles, or viruses. References 1. Lakowicz JR (2006) Principles of fluorescence spectroscopy, 3rd edn. Springer, Berlin 2. Valeur B (2002) Molecular fluorescence – principles and applications. Wiley-VCH, Weinheim 3. Sapsford KE, Berti L, Medintz IL (2006) Materials for fluorescence resonance energy transfer analysis: beyond traditional donoracceptor combinations. Angew Chem Int Ed 45:4562–4588 4. Roda A, Guardigli M, Michelini E, Mirasoli M (2009) Nanobioanalytical luminescence: Förster-type energy transfer methods. Anal Bioanal Chem 393:109–123 5. Roda A, Guardigli M, Michelini E, Mirasoli M, Pasani P (2003) Analytical bioluminescence and chemiluminescence. Anal Chem 75:462A–470A 6. Clegg RM (1996) Fluorescence resonance energy transfer. In: Wang XF, Herman B (eds)

7.

8. 9.

10. 11.

Fluorescence imaging spectroscopy and microscopy. Wiley, New York, pp 179–252 Enderlein J (2003) Electrodynamics of fluorescence. http://www.joerg-enderlein.de/ fileadmin/downloads/Tutorial.pdf. Accessed 22 July 2009 Selvin PR (2000) The renaissance of fluorescence resonance energy transfer. Nat Struct Biol 7:730–734 Szollosi J, Damjanovich S, Matyus L (1998) Application of fluorescence resonance energy transfer in the clinical laboratory: routine and research. Cytometry B 34:159–179 Williams C (2004) cAMP detection methods in HTS: selecting the best from the rest. Nat Rev Drug Discov 3:125–135 Bremer C, Tung CH, Weissleder R (2001) In vivo molecular target assessment of matrix metalloproteinase inhibition. Nat Med 7: 743–748

318

Schäferling and Nagl

12. Kohl T, Heinze KG, Kuhlemann R, Koltermann A, Schwille P (2002) A protease assay for twophoton crosscorrelation and FRET analysis based solely on fluorescent proteins. Proc Natl Acad Sci U S A 99:12161–12166 13. Yang Y, Babiak P, Reymond JL (2006) Low background FRET-substrates for lipases and esterases suitable for high-throughput screening under basic (pH 11) conditions. Org Biomol Chem 4:1746–1754 14. Lyon E, Wittwer CT (2009) LightCycler technology in molecular diagnostics. J Mol Diagn 11:93–101 15. Smith CJ (2005) Quantitative real-time PCR. In: Osborn AM, Smith CJ (eds) Molecular microbial ecology. Taylor & Francis, New York, pp 151–166 16. Tapp I, Malmberg L, Rennel E, Wik M, Syvanen AC (2000) Homogeneous scoring of singlenucleotide polymorphisms: comparison of the 5¢-nuclease TaqMan assay and Molecular Beacon probes. Biotechniques 28:732–738 17. Goel G, Kumar A, Puniya AK, Chen W, Singh K (2005) Molecular beacon: a multitask probe. J Appl Microbiol 99:435–442 18. Makhina EN, Nichols CG (2001) Mutant GFP-based FRET analysis of K+ channel organization. In: Lopatin A, Nichols CG (eds) Ion channel localization. Humana, Totowa, pp 261–274 19. Arun KHS, Kaul CL, Ramarao P (2005) Green fluorescent proteins in receptor research: an emerging tool for drug discovery. J Pharmacol Toxicol Methods 51:1–23 20. Takanishi CL, Bykova EA, Cheng W, Zheng J (2006) GFP-based FRET analysis in live cells. Brain Res 1091:132–139 21. Jares-Erijman EA, Jovin TM (2003) FRET imaging. Nat Biotechnol 21:1387–1395 22. Kokko T, Kokko L, Lövgren T, Soukka T (2007) Homogeneous noncompetitive immunoassay for 17beta-estradiol based on fluorescence resonance energy transfer. Anal Chem 79:5935–5940 23. Mathis G (1993) Rare earth cryptates and homogeneous fluoroimmunoassays with human sera. Clin Chem 39:1953–1959 24. Kim SH, Jeyakumar M, Katzenellenbogen JA (2007) Dual-mode fluorophore-doped nickel nitrilotriacetic acid-modified silica nanoparticles combine histidine-tagged protein purification with site-specific fluorophore labeling. J Am Chem Soc 129:13254–13264 25. Kürner JM, Wolfbeis OS, Klimant I (2002) Homogeneous luminescence decay timebased assay using energy transfer from nanospheres. Anal Chem 74:2151–2156

26. Valanne A, Lindroos H, Lövgren T, Soukka T (2005) A novel homogeneous assay format utilising proximity dependent fluorescence energy transfer between particulate labels. Anal Chim Acta 539:251–256 27. Kuningas K, Ukonaho T, Päkkilä H, Rantanen T, Rosenberg J, Lövgren T, Soukka T (2006) Upconversion fluorescence resonance energy transfer in a homogeneous immunoassay for estradiol. Anal Chem 78:4690–4696 28. Shen Y, Wu BL (2009) Microarray-based genomic DNA profiling technologies in clinical molecular diagnostics. Clin Chem 55: 659–669 29. Hinds DA, Stuve LL, Nilsen GB, Halperin E, Eskin E, Ballinger DG, Frazer KA, Cox DR (2005) Whole-genome patterns of common DNA variation in three human populations. Science 307:1072–1079 30. Uttamchandani M, Neo JL, Ong BNZ, Moochhala S (2008) Applications of microarrays in pathogen detection and biodefence. Trends Biotechnol 27:53–61 31. Ahmed FE (2004) DNA-based methods for GMO detection: historical developments and future prospects. In: Ahmed FE (ed) Testing of genetically modified organisms in foods. Haworth, Binghamton, pp 221–253 32. Hakes L, Pinney JW, Robertson DL, Lovell SC (2008) Protein-protein interaction networks and biology – what’s the connection? Nat Biotechnol 26:69–72 33. Caiazzo RJ Jr, Maher AJ, Drummond MP, Lander CI, Tassinari OW, Nelson BP, Liu BCS (2009) Protein microarrays as an application for disease biomarkers. Proteomics Clin Appl 3:138–147 34. Hartmann M, Roeraade J, Stoll D, Templin MF, Joos TO (2009) Protein microarrays for diagnostic assays. Anal Bioanal Chem 393: 1407–1416 35. Jonkheijm P, Weinrich D, Schroeder H, Niemeyer CM, Waldmann H (2008) Chemical strategies for generating protein biochips. Angew Chem Int Ed 47: 9618–9647 36. Seidel M, Niessner R (2008) Automated analytical microarrays: a critical review. Anal Bioanal Chem 391:1521–1544 37. Kersten B, Wanker EE, Hoheisel JD, Angenendt P (2005) Multiplex approaches in protein microarray technology. Expert Rev Proteomics 2:499–510 38. Schäferling M, Nagl S (2006) Optical technologies for the read out and quality control of DNA and protein microarrays. Anal Bioanal Chem 385:500–517

Förster Resonance Energy Transfer Methods 39. Nagl S, Schaeferling M, Wolfbeis OS (2005) Fluorescence analysis in microarray technology. Microchim Acta 151:1–21 40. Templin MF, Stoll D, Schwenk JM, Pötz O, Kramer S, Joos TO (2003) Protein microarrays: promising tools for proteomic research. Proteomics 3:2155–2166 41. Schäferling M, Schiller S, Paul H, Kruschina M, Pavlickova P, Giammasi C, Kambhampati D (2002) Application of self-assembly techniques in the design of biocompatible protein microarray surfaces. Electrophoresis 23: 3097–3105 42. MacBeath G (2002) Protein microarrays and proteomics. Nat Genet 32:526–532 43. Schwartz DE, Gong P, Shepard KL (2008) Time-resolved Förster-resonance-energytransfer DNA assay on an active CMOS microarray. Biosens Bioelectron 24:383–390 44. Frutos AG, Pal S, Quesada M, Lahiri J (2002) Method for detection of single-base mismatches using bimolecular beacons. J Am Chem Soc 124:2396–2397 45. Fang X, Liu X, Schuster S, Tan WJ (1999) Designing a novel molecular beacon for surface-immobilized DNA hybridization studies. J Am Chem Soc 121:2921–2922 46. Kim H, Kane MD, Kim S, Dominguez W, Applegate BM, Savikhin S (2007) A molecular beacon DNA microarray system for rapid detection of E. coli O157:H7 that eliminates the risk of a false negative signal. Biosens Bioelectron 22:1041–1047 47. Diaz-Mochon JJ, Bialy L, Bradley M (2006) Dual colour, microarray-based, analysis of 10,000 protease substrates. Chem Commun 38:3984–3986 48. Kong A, Leboucher P, Leek R, Calleja V, Winter S, Harris A, Parker PJ, Larijani B (2006) Prognostic value of an activation state marker for epidermal growth factor receptor in tissue microarrays of head and neck cancer. Cancer Res 66:2834–2843 49. Usui K, Takahashi M, Nokihara K, Mihara H (2004) Peptide arrays with designed a-helical structures for characterization of proteins from FRET fingerprint patterns. Mol Diver 8:209–218 50. Castellana ET, Cremer PS (2007) Imaging large arrays of supported lipid bilayers with a macroscope. Biointerphases 2:57–63 51. Vogel KW, Riddle SM, Horton RA, Robers MB, Michaud GA (2008) FRET-based kinase and ubiquitination assay using a fluorescent fusion protein substrate and a luminescent metal complex, and screening applications. PCT Int. Appl. WO 2008011601

319

52. Herman P, Lin HJ, Lakowicz JR (2003) Lifetime-based imaging. In: Vo-Dinh T (ed) Biomedical photonics handbook. CRC, Boca Raton, pp 9.1–9.30 53. Nagl S, Bauer R, Sauer U, Preininger C, Bogner U, Schäferling M (2008) Microarray analysis of protein-protein interactions based on FRET using subnanosecond-resolved fluorescence lifetime imaging. Biosens Bioelectron 24:397–402 54. Woods RJ, Scypinski S, Cline Love LJ, Ashworth HA (1984) Transient digitizer for the determination of microsecond luminescence lifetimes. Anal Chem 56:1395–1400 55. Hartmann M, Schrenk M, Doettinger A, Nagel S, Roeraade J, Joos TO, Templin MF (2008) Expanding assay dynamics: a combined competitive and direct assay system for the quantification of proteins in multiplexed immunoassays. Clin Chem 54:956–963 56. Cervino C, Sauceda JC, Niessner R, Knopp D (2008) Mycotoxin analysis by automated flow-through immunoassay with chemoluminescence readout. Luminescence 23:206–207 57. Roy R, Hohng S, Ha T (2008) A practical guide to single-molecule FRET. Nat Methods 5:507–516 58. Kokko T, Liljenbäck T, Peltola MT, Kokko L, Soukka T (2008) Homogeneous dual-parameter assay for prostate-specific antigen based on fluorescence resonance energy transfer. Anal Chem 80:9763–9768 59. Kuningas K, Päkkilä H, Ukonaho T, Rantanen T, Lövgren T, Soukka T (2007) Upconversion fluorescence enables homogeneous immunoassay in whole blood. Clin Chem 53: 145–146 60. Campbell RE (2009) Fluorescent-proteinbased biosensors: modulation of energy transfer as a design principle. Anal Chem 81:5972–5979 61. Malicka J, Gryczynski I, Fang J, Kusba J, Lakowicz JR (2003) Increased resonance energy transfer between fluorophores bound to DNA in proximity to metallic silver particles. Anal Biochem 315:160–169 62. Rosi NL, Mirkin CA (2005) Nanostructures in biodiagnostics. Chem Rev 105:1547–1562 63. Ray PC, Fortner A, Darbha GK (2006) Gold nanoparticle based FRET assay for the detection of DNA cleavage. J Phys Chem B 110:20745–20748 64. Haas E, Wilchek M, Katchalski-Katzir E, Steinberg IZ (1975) Distribution of end-toend distances of oligopeptides in solution as estimated by energy transfer. Proc Natl Acad Sci U S A 72:1807–1811

320

Schäferling and Nagl

65. Borochov-Neori H, Montal M (1989) Rhodopsin-G-protein interactions monitored by resonance energy transfer. Biochemistry 28:1711–1718 66. Kosk-Kosicka D, Bzdega T, Wawrzynow A (1989) Fluorescence energy transfer studies of purified erythrocyte calcium ATPase. Calcium regulated activation by oligomerization. J Biol Chem 264:19495–19499 67. Ha T, Rasnik I, Cheng W, Babcock HP, Gauss GH, Lohman TM, Chu S (2002) Initiation and re-initiation of DNA unwinding by the Escherichia coli Rep helicase. Nature 419: 638–641 68. Schuler B, Lipman EA, Eaton WA (2002) Probing the free-energy surface for protein folding with single-molecule fluorescence spectroscopy. Nature 419:743–747 69. Kokko T, Kokko L, Soukka T (2009) Terbium(III) chelate as an efficient donor for

70. 71. 72.

73.

74.

multiple-wavelength fluorescent acceptors. J Fluoresc 19:159–164 Selvin PR (2002) Principles and biophysical applications of lanthanide-based probes. Annu Rev Biophys Biomol Struct 31:275–302 Patterson GH, Piston DW, Barisas BG (2000) Förster distances between green fluorescent protein pairs. Anal Biochem 284:438–440 Medintz IL, Clapp AR, Melinger JS, Deschamps JR, Mattoussi H (2005) A reagentless biosensing assembly based on quantum dot-donor Förster resonance energy transfer. Adv Mater 17:2450–2455 Kim JH, Morikis D, Ozkan M (2004) Adaptation of inorganic quantum dots for stable molecular beacons. Sens Actuators B 102:315–319 Selvin PR, Rana TM, Hearst JE (1994) Luminescence resonance energy transfer. J Am Chem Soc 116:6029–6030

Chapter 20 Label-Free Detection with Surface Plasmon Resonance Imaging Christopher Lausted, Zhiyuan Hu, and Leroy Hood Abstract Sensors based on surface plasmon resonance have demonstrated, over the last 2 decades, to be an effective method of studying biomolecular interactions without the need for labeling. Recently, it has been adapted to high-throughput use for imaging microarray binding in real time. This provides a promising platform – a label-free protein microarray system – for the study of disease. In this example, antibody microarrays are used to efficiently profile the secretion of proteins from a cell line exposed to varying concentrations of a toxic compound. Key words: Surface plasmon resonance, Label-free microarray, Molecular interactions, Protein binding, Protein profiling, Kinetics analysis, Affinity, Refractometry

1. Introduction Surface plasmon resonance (SPR) spectroscopy is a versatile tool for the study of the binding and kinetics of unlabeled biomolecules. Small changes in the refractive index of a coating or solution near the SPR-active surface are detected with high sensitivity and with fast temporal resolution. In SPR Imaging (SPRI), the active surface is a gold-coated microarray spotted with a large number of capture molecules such as antibodies, proteins, peptides, DNA, RNA, or small molecules. Protein-binding microarrays – typically containing spotted antibodies – show great promise as a highthroughput tool for the differential analysis of healthy and diseased states. Like the fluorescence and chemiluminescence-labeled variety, SPRI microarrays allow users to study equilibrium protein binding levels. Additionally, binding can be observed in real time, enabling kinetic analysis.

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_20, © Springer Science+Business Media, LLC 2011

321

322

Lausted, Hu, and Hood

Early SPRI examples of protein-binding binding microarrays include double-stranded DNA arrays used to bind transcription factors (1, 2). Subsequently, antibody arrays have been used to profile protein expression levels in complex samples such as crude cell lysates (3, 4) and serum (5). In these cases, the limits of detection reached the low ng/mL level, given sufficiently specific antibodies. This is less sensitive than comparable sandwich immunoassay-based assays, but sufficient to detect at moderately expressed proteins. Similarly, serum has been applied to arrays of glycans (6) and citrullinated peptides (7) to measure antibody titers. Other applications include the use of lectin arrays to analyze the kinetics and affinity of protein–glycan interactions (8) and Fab fragment arrays for antibody development based on affinity ranking (9). In these cases, the arrays were found to be highly reusable. Depending on the ligand and its immobilization, surfaces have been regenerated to analyze scores of different samples. Antibody arrays have also been used to profile surface proteins by capturing whole cells such as baby hamster kidney cells via basic fibroblast growth factor (10) and lymphocytes via CD markers (11). However, arrays used for whole-cell capture have not been demonstrated to be reusable. In contrast to fluorescence imaging of dried microarrays, SPRI measures binding events on a microarray inside a fluid-filled flow cell. It measures changes in the effective refractive index – the thickness of the adsorbed layer – near the metal surface of the microarray. When collimated, p-polarized light is reflected at greater than the critical angle, the incident photons are coupled into surface plasmons, leading to a decrease in the reflected light intensity. Binding events at the microarray features increase the local refractive index, changing the reflectance pattern of the surface. The reflectance across the entire array surface is recorded in real-time using a camera. This provides several advantages. First, quality control is simplified since the microarray features are immediately visible. No arrays need to be sacrificed for staining to detect printing errors. The initial image of each array shows clearly the area and mass of each antibody (or other capture reagent) immobilized at each feature. Second, sample preparation is simplified. No labeling effort is required and no labeling artifacts are introduced. Third, real-time monitoring provides for instant feedback, simplified assay optimization, kinetic analysis, and true affinity calculation. Fourth, the arrays are often reusable. Background signal does not necessarily accumulate, as fluorophores are not allowed to dry on the chip surface. Commercial systems for array use have become available recently from GE/Biacore, GWC Technologies, GenOptics, K-MAC, IBIS, and Plexera. These systems can generally be used with most conventional pin or inkjet arraying robots. SPR-active substrates are gold-coated glass or plastic chips; in some cases the

Label-Free Detection with Surface Plasmon Resonance Imaging

323

format is a standard 25 × 76 mm slide. While a number of surface coatings and coupling chemistries are available, most vendors also provide a bare gold surface. A wide selection of capture reagents, including antibodies and thiol-terminated nucleic acids, can be reliably and irreversibly immobilized on this surface by physical adsorption. The surface is desirable as it requires no activation step, and coupling efficiency does not decline over the course of a long printing process. In the protocol below, an antibody array is printed, blocked, and used to profile the proteins secreted into culture medium by the HepG2 human liver carcinoma cell line. To attenuate the normal secretion of proteins, the cells are treated with three concentrations of troglitazone (TGZ), a hepatotoxic thiazolidinedione (12).

2. Materials 2.1. Reagents

1. Phosphate buffered saline pH 7.4 (PBS) (Gibco/Invitrogen, Carlsbad, CA). 2. Spotting buffer: 10% (v/v) dimethyl sulfoxide (DMSO) in 1× PBS. 3. Blocking solution: 2% (w/v) nonfat dry milk in 1× PBS. 4. Running buffer (mPBST): 0.05% (w/v) nonfat dry milk and 0.005% (w/v) Tween 20 (Sigma, St. Louis, MO) in 1× PBS. 5. Dulbecco’s Modified Eagle’s Medium (DMEM) (Gibco/ Invitrogen). 6. Growth medium: DMEM supplemented with 10% heatinactivated fetal bovine serum, 1% penicillin:streptomycin solution (100×) and 1% l-glutamine (all from Gibco/ Invitrogen). 7. Running Buffer (1×): 0.05% (w/v) nonfat dry milk, 0.05% (v/v) Protease Inhibitor Cocktail (Sigma, St. Louis, MO), and 0.005% (w/v) Tween 20 in 1× PBS. 8. Running Buffer (100×): 5% (w/v) nonfat dry milk, 5% (v/v) Protease Inhibitor Cocktail (Sigma, St. Louis, MO), and 0.5% (w/v) Tween 20 in 1× PBS. 9. Regeneration Buffer: 0.2% orthophosphoric acid (85% solution, Sigma) in deionized water. 10. Calibration Solution: 1.3% glycerol (v/v) (99% pure, Sigma) in running buffer. This provides an increase of 1.8 × 10−3 refractive index units (1,800 mRIU) over plain running buffer. 11. Cleaning Solution: 0.05% Tween 20 in deionized water. 12. Hepatotoxin: TGZ powder (98% pure, Sigma).

324

Lausted, Hu, and Hood

2.2. Equipment

1. Robotic microarrayer: VersArray® (Bio-Rad, Hercules, CA) or equivalent pin-spotting microarray printer. 2. Microarray pins: Stealth® SMP7 (Arrayit, Sunnyvale, CA) deliver 2.3 nL producing a spot diameter of approximately 230 mm on glass slides. On gold SPR slides, a square feature is produced with a width of approximately 300 mm. A minimum of one and a maximum of four pins are used for printing. 3. Arrayer source plates: Uniplate® (Whatman, Clifton, NJ) 384-well, V-bottom, polypropylene plates contain the capture molecules to be arrayed. 4. Microarray blocking station: The ZeptoFOG® (Zeptosens, Witterswil, Switzerland) device is used to apply blocking solution as an aerosol to freshly printed SPR arrays. This is necessary to maintain good spot morphology when using bare gold slides. 5. SPR Imager: PlexArray® running Instrument Control Module v1.0 and Data Analysis Module software v1.0 (Plexera LLC, Seattle, WA). 6. Gold-coated array substrates: NanoCapture Gold® (Plexera) slides are 25 × 76 × 1 mm in size and coated on one side with a 47 nm gold layer. Slides are provided with adhesive covers, which will be attached to form a 30 mL flow cell. 7. Nitrogen jet: An oil-free blow gun is attached to a nitrogen gas cylinder regulated to 200 kPa (30 psig). 8. Vacuum chamber: A vacuum desiccator, preferably constructed of clear polycarbonate with an internal diameter greater than 240 mm, and a vacuum source capable of drawing −50 kPa gauge pressure. 9. T-150 Tissue Culture Flask (Becton Dickinson, Franklin Lakes, NJ).

2.3. Biological Materials

1. Capture antibodies: Antibodies from stock concentrations of greater than 500 mg/mL. 2. Control antibody: Purified NA/LE Mouse IgG1 k isotype control (BD Biosciences, San Jose, CA). 3. Liver cell line: HepG2 (ATCC No. HB-8065).

3. Methods In this protocol, an array is printed containing 128 unique antibodies targeting proteins expressed in hepatocytes. Four replicates of each antibody are printed. In addition, every third

Label-Free Detection with Surface Plasmon Resonance Imaging

325

antibody printed is a nonbinding control, distributing numerous negative controls evenly over the surface of the array. The total number of array features in 768, which fit easily within the viewable area of the SPR imager. While these procedures are specific to the VersArray microarray spotter and the PlexArray SPR imager, the materials and the order of the steps may be very similar for other brands of instruments. Actions required by the PlexArray Instrument Control Module version 1 software (ICM) are given in parentheses. 3.1. Preparation of Secretome Samples

1. HepG2 human hepatocellular liver carcinoma cells are grown to confluency in DMEM-based growth media at 37°C in the presence of 5% CO2. 2. Cells are split at a 1:6 ratio into T-150 flasks. Flasks are allowed to grow to 85% confluency before drug treatment. 3. Drug treatment media is prepared with 0, 10, 50, or 100 mM TGZ and 0.1% DMSO in DMEM. 4. Immediately prior to treatment, the cells are washed twice in serum-free DMEM and then treated with 0, 10, 50, or 100 mM TGZ for 16 h. The medium is collected cleared of any cellular debris by centrifugation at 15,000 rpm for 5 min. The medium is stored at −80°C until use.

3.2. Printing Antibody Arrays

1. Prepare the antibody spotting source plate. Dilute capture antibodies to 100–500 mg/mL in 10% DMSO in PBS. Fill the wells in a Whatcom 384-well plate with 6–12 mL of antibody solution. Start filling from well A1 and proceed to A24 before moving to B1. Place control antibody in every third well. With 128 experimental antibodies and 64 controls, wells A1 through H24 are filled. 2. Cover the plate and store at 4°C until printing. Use a plastic cover for short-term storage or cover the wells with aluminum tape and seal it down with a rubber roller for long-term storage. 3. Configure the arrayer for single-pin printing. The pitch or center-to-center spacing can be set to 450 mm in the horizontal and 350 mm in the vertical. Single-pin printing maximizes the number of spots that can be fit in the active area of the SPR. A rectangular pitch improves the imaging as SPRI camera observes the microarray from approximately 60º. The PlexArray rotates the microarray 90º counterclockwise and foreshortens the image in what will appear to be the vertical. The position and size of the viewable region of the SPR slide, along with a typical four-replicate printing geometry, is shown in Fig. 1. 4. Set the humidity control to 60% relative. Set the room thermostat to between 18 and 22°C.

326

Lausted, Hu, and Hood

a 25.4 14.0

39.0

14.0 76.2

5.7

b

Fig. 1. Printing of a 768-feature microarray containing 192 antibodies. (a) The geometry of the microarray slide, as viewed by the ESI spotter. The antibodies are printed on the 14 × 14 mm SPR-active area, which is right-of-center. Printing starts with the antibody in source plate well A1 at the top left corner of the active area and proceeds from left to right, then top to bottom. The array is divided into four regional replicates. (b) The geo metry of the microarray slide, as viewed on the SPR imager, is rotated 90º counterclockwise. The antibody from well A1 now appears located at the lower left corner of the array, antibody A2 is above it, and antibody A24 is at the top left corner of the array. The last antibody printed, H24, appears at the top right corner of the array. The pattern is repeated in four rectangular regions as each antibody is printed in quadruplicate. (c) SPR image of the unused 192-antibody microarray immersed in running buffer. Unlike a fluorescence microarray, the quality of the microarray printing can be immediately evaluated. Pixel intensity from the SPR image is proportional to the mass of material (antibody) immobilized on the surface. Based on intensity, antibody densities on this microarray range from 0 to 23 ng/mm2. Intensities vary greatly by antibody but remain consistent between replicates.

Label-Free Detection with Surface Plasmon Resonance Imaging

327

5. Print with a single SMP7 pin. It is possible to print with up to four pins, but with some space wasted. Configure the arrayer to wash the pin twice using 15 s of sonication and drying. Blot the pin four times on a dummy glass slide before printing on the gold slides. With each dip of the pen, print not more than eight features. If more features are needed (e.g., more than four slides with two replicate features), the pin needs to be redipped in the source well. Wash the pin between every antibody/well (see Note 1). 6. After printing, place the arrays in a humid chamber at 4°C overnight. 7. Block the arrays using the Zeptofog device. Fill the cup with 0.1% nonfat dry milk in deionized water. Place up to four arrays in the system and run for 1 h. By this time, the arrays will be uniformly wet (see Notes 2 and 3). 8. Wash in arrays in PBS for 1 min. Wash the arrays in deionized water for 1 min. 9. Dry the arrays with a jet of nitrogen. Pack the arrays in slide holders. Place the holders in a nitrogen-filled jar and store the jar at 4°C. 10. Create a Gene Array List (GAL) file. A GAL file contains the location and identity of each microarray feature. GAL files allow microarrays to be divided into subregions called “blocks.” Each feature is specified with a block number, row number, column number, and description. This array contains four blocks (replicates) each with 24 rows and eight columns. Use the arrayer manufacturer’s software to create the GAL file. 3.3. Binding Experiment

1. Start preparing the array the day before use. Block the array with Blocking Solution (2% milk) at 4°C, overnight. 2. Remove the array from the blocking solution and wash. Transfer the array to a 50 mL centrifuge tube filled with PBS and agitate for 30 min. Rinse the array in deionized water for 1 min and dry with a jet of nitrogen. 3. Power up the SPR instrument and launch the control software. Toggle on the flow cell temperature control and adjust the set point to 20°C. Toggle on the well plate temperature control and adjust the set point to 4°C. Toggle on the in-line degasser for the running buffer. Power up the SPR at least 30 min before using to allow the light source to warm and the temperature control to equilibrate. (ICM: Launch the “Instrument Control” module. This software uses a tabbed interface, the first of which is the “Instrument Setup” tab. The switches and set points are found here.)

328

Lausted, Hu, and Hood

4. Degas the Running Buffer, Calibration Solution, and Regeneration Solution. Partially open the lids of their containers, place them in a desiccator, and draw a vacuum of at least 50 kPa for 30 min. 5. Load reagents and samples. Place Running Buffer in location 0, Regeneration Solution in location 1, deionized water in location 2, and Cleaning Solution in location 3. (ICM: Select the “Method Setup” tab and use the “Reagent Table” button.) 6. Load the autosampler. Load the 96-well × 1 mL sample plate with 800 mL of each culture supernatant as follows: (A1) Calibration Solution, (A2–A4) Running Buffer, (A5) 0 mM TGZ control, (A6) 100 mM TGZ treatment, (A7) 40 mM TGZ treatment, (and A8) 10 mM TGZ treatment. Place replicate samples in sample plate rows B, C, etc. Configure the software with the location, name, and fluid flow profile of each sample. The flow profile requires a flowrate of 1 mL/s, an association duration of 600 s, and a dissociation duration of 300 s. (ICM: Select the “Method Setup” tab and use the “Analyte Table” button.) 7. Define the sample injection protocol. Define the regeneration protocol as a brief injection of 360 mL of Regeneration Buffer followed by 900 mL of Running Buffer, both at 4 mL/s. Configure the experiment to consist of three cycles of regeneration followed by the injections of the analytes in autosampler as described above. (ICM: Select the “Method Setup” tab and use the “Method Builder” table.) 8. Attach the flow cell. Insert the slide into the manufacturer’s jig. Check that the gold surface is face-up and the printed area is at the back of the jig. Remove the protective paper from the adhesive and insert the flow cell into the jig. Clamp the two pieces together and wait 10 min. After 10 min, remove the assembled flow cell from the jig. Check for any air gaps between the adhesive and the gold that might allow leakage. Reclamp if necessary. 9. Fill the flow cell with deionized water. Use a 200 mL pipetter to fill the flow cell, carefully avoiding bubbles (see Note 4). 10. Mount the flow cell on the prism. Apply two drops of index matching fluid to the SPRI prism. Place the flow cell in the flow cell holder and close the holder on to the prism. Use the live video display to check for the presence of bubbles in the index fluid. These appear as pairs of matching white and black bubbles. If a bubble appears, open the holder and clean off the index fluid using the supplied lint-free tissues and optical cleaning solution. Repeat if necessary. (ICM: Select the “Load Instrument” tab to view the live video display.)

Label-Free Detection with Surface Plasmon Resonance Imaging

329

11. Prime all channels of the fluid handling system. Flush all channels with running buffer, then prime the channels that handle reagents and analytes. Last, load the flow cell. Displace the deionized water with running buffer. (ICM: The “Load Instrument” tab also contains the priming functions.) 12. Choose nine “regions of interest” (ROIs). Select nine square ROIs of any size, evenly spaced over the entire viewable area. (ICM: Use the “Assign ROI” tab.) 13. Perform the SPR angle scan. Plot reflectance from roughly 50–65º to determine the best angle for monitoring real-time binding. At the lower angles, the gold surface of the micro array will reflect 100% of the incident light. Between 55 and 60º, reflectance will drop dramatically as the scan approaches the resonance angle. Choose the angle where the average SPR curve is at the 30% point (less than midway) between the minimum and maximum signal values. (ICM: Select the “Set SPR Angle” tab. Specify a scan from position 0–30 mm and allow the system to choose the 30% reflectance point. An example of SPR reflectance curve is shown in Fig. 2.) 14. Redefine the ROIs. With the SPR angle chosen, the new antibody array will now appear as grid of light features on a dark background. Brighter features indicate a higher density of immobilized material. Move and resize the nine ROIs to fit representative antibody features. The large, square-tipped pins used for printing produce square features, so square

Fig. 2. The reflectance at nine regions-of-interest (ROIs) on the SPR-active surface drops from nearly 100 to 10% as the positioning arm moves from 9 to 25 mm (where 25 mm corresponds to an observation angle of approximately 60º). To monitor binding, the angle is fixed at the low end of the linear region, usually between 30 and 35% reflectance. In this case, a position of 21.0 mm is selected to produce 33% reflectance from the leftmost SPR curve.

330

Lausted, Hu, and Hood

ROIs can be selected. Choose ROIs from various regions of the SPR image. Choose antibodies of varying intensity and choose antibodies where binding is expected. These ROIs will be plotted in real time as the experiment proceeds. (ICM: Use the “Assign ROI” tab.) 15. Start the binding experiment. Disable background subtraction before starting. Each ROI has a set of “satellite” or reference ROIs that can be used to compensate for bulk refractive index changes. By disabling this, the real time plot more clearly shows the injection of low-index regeneration solutions, high-index calibration solution, and the elevated-index of the culture supernatants. Note the time at which the calibration solution is injected. The background subtraction switch does not affect subsequent data analysis. (ICM: Use the “Run” tab.) 16. Clean the fluidic system. After the experiment is completed, flush all fluid channels thoroughly with Cleaning Solution followed by deionized water. 3.4. Data Analysis

While the Instrument Control Module as used here displays SPR sensorgrams in real-time for nine microarray features, postanalysis by Data Analysis Module is required to generate the full 768 sensorgrams. The Data Analysis Module contains a number of curve fitting capabilities, but here it is used only to produce calibrated sensorgrams. Selected sensorgrams can then be imported into software packages such as CLAMP for kinetic analysis (13) or Bioconductor for expression profile analysis (14). In the following steps, the sensorgrams are extracted from a movie file and calibrated to units of refractive index. While the PlexArray sensorgrams are provided in units of percent reflectance, other systems report angular units, spectral units, or other “Resonance Units” (usually equivalent to 10−6 refractive index units). By calibrating binding levels to refractive index units, comparisons can be drawn between the measurements performed by different systems. 1. Extract sensorgrams from the raw movie file. PlexArray: Launch the “Data Analysis” module and choose the “Video Setup” tab. Load the new AVI movie file and the corresponding GAL file. Align the ROIs generated from the GAL file with the microarray image. Use the “Measure Intensities” feature to extract SPR sensorgrams from the movie ROIs. This process may take a long time. Export the data to the spreadsheet (XLS) format. While the Data Analysis module is capable of calculating the kinetic constants of a simple 1:1 interaction model, more sophisticated analysis and outlier detection must be performed manually or by third-party software (see Note 5).

Label-Free Detection with Surface Plasmon Resonance Imaging

331

2. Calibrate the data to refractive index units. This can be accomplished by loading the data into spreadsheet software. The data from each ROI occupies four columns: identification, time, foreground intensity, and background intensity. Using the calibration injection time recorded earlier, locate the data row corresponding to calibration solution injection. Also, locate a data row immediately preceding the calibration. Use the difference in signal between these two data points, corresponding to a change of 1,800 mRIU in bulk refractive index, to calculate a calibration factor for each ROI foreground and background. Apply these calibration factors to the data. 3. Perform background subtraction. Subtract each ROI background level from the foreground level. This step removes changes in the sensorgram due to bulk refractive index changes (see Note 6). An example of the background-subtracted sensorgrams from four serial injections of culture media onto an SPR microarray is shown in Fig. 3.

Fig. 3. SPR array analysis of culture media from hepatocytes exposed to three concentrations of potentially toxic Troglitazone (TGZ). Through the use of surface regeneration, four samples are applied to the same antibody microarray over the course of an hour. Each cycle consists of 6 min of binding followed by 3 min of washing and then regeneration. Decreased rates of binding indicate diminished levels of plasminogen and transferrin secretion by cells treated with 40 or 100 mM TGZ.

332

Lausted, Hu, and Hood

4. Notes 1. A single pin is used here to print 192 in four replicates, but a wide range of geometries can be arrayed. In this example, an array of 24 rows by 32 columns was printed in an area of 10.8 × 11.2 mm or 121 mm2. As the sensing region is 14 × 14 mm or 196 mm2, the array was only 62% full. With single-pin printing, nearly 1,240 features can be arrayed at this density. Printing speeds can also be accelerated fourfold by the use of four pins. However, a configuration of 2 × 2 pins on a standard 4.5 mm spacing limits the geometry to four blocks of 10 × 12 antibody features, or a mere 480 features in 76 mm2. 2. Milk blocking is a highly effective blocking protein for bare gold surfaces and complex samples such as cellular lysates and serum. For more inert surfaces and simpler samples, casein, BSA, and fish gelatin are often effective. 3. The ZeptoFog blocker will not “fog” with a milk protein solution concentration above 0.1%. Higher concentrations of casein or BSA can be used. 4. Persistent bubbles can be removed from the flow cell by drying and then refilling the flow cell. The nitrogen jet can be used to purge the deionized water from the flow cell. Avoid drying a flow cell that contains running buffer. Salt and protein will precipitate and stick irreversibly to the gold surface. 5. Replicates are very important in SPR experiments. It is common to discard the data from the first two samples injected into the flow cell as these will often be outliers. 6. Sensorgram background subtraction may be performed using the local background (the gold surface immediately surrounding the antibody) or by using a nearby control antibody. Often, less nonspecific binding is observed when using the control antibody. However, calibration errors increase as the distance between the experiment and the control antibodies increase. References 1. Smith EA, Wanat MJ, Cheng Y, Barreira SVP, Frutos AG, Corn RM (2001) Formation, spectroscopic characterization, and application of sulfhydryl-terminated alkanethiol monolayers for the chemical attachment of DNA onto gold surfaces. Langmuir 17(8): 2502–2507 2. Shumaker-Parry JS, Aebersold R, Campbell CT (2004) Parallel, quantitative measurement

of protein binding to a 120-element doublestranded DNA array in real time using surface plasmon resonance microscopy. Anal Chem 76(7):2071–2082 3. Kyo M, Usui-Aoki K, Koga H (2005) Labelfree detection of proteins in crude cell lysate with antibody arrays by a surface plasmon resonance imaging technique. Anal Chem 77(22):7115–7121

Label-Free Detection with Surface Plasmon Resonance Imaging 4. Usui-Aoki K, Shimada K, Nagano M, Kawai M, Koga H (2005) A novel approach to protein expression profiling using antibody microarrays combined with surface plasmon resonance technology. Proteomics 5(9):2396–2401 5. Lausted C, Hu Z, Hood L (2008) Quantitative serum proteomics from surface plasmon resonance imaging. Mol Cell Proteomics 7(12):2464–2474 6. de Boer A, Hokke C, Deelder A, Wuhrer M (2008) Serum antibody screening by surface plasmon resonance using a natural glycan microarray. Glycoconj J 25(1):75–84 7. Lokate AMC, Beusink JB, Besselink GAJ, Pruijn GJM, Schasfoort RBM (2007) Biomolecular Interaction Monitoring of Autoantibodies by Scanning Surface Plasmon Resonance Microarray Imaging. J Am Chem Soc 129(45):14013–14018 8. Karamanska R, Clarke J, Blixt O, MacRae J, Zhang J, Crocker P, Laurent N, Wright A, Flitsch S, Russell D, Field R (2008) Surface plasmon resonance imaging for real-time, label-free analysis of protein interactions with carbohydrate microarrays. Glycoconj J 25(1):69–74 9. Wassaf D, Kuang G, Kopacz K, Wu Q-L, Nguyen Q, Toews M, Cosic J, Jacques J,

333

Wiltshire S, Lambert J, Pazmany CC, Hogan S, Ladner RC, Nixon AE, Sexton DJ (2006) High-throughput affinity ranking of antibodies using surface plasmon resonance microarrays. Anal Biochem 351(2): 241–253 10. Peelen D, Kodoyianni V, Lee J, Zheng T, Shortreed MR, Smith LM (2006) Specific Capture of Mammalian Cells by Cell Surface Receptor Binding to Ligand Immobilized on Gold Thin Films. J Proteome Res 5(7):1580–1585 11. Suraniti E, Sollier E, Calemczuk R, Livache T, Marche PN, Villiers MB, Roupioz Y (2007) Real-time detection of lymphocytes binding on an antibody chip using SPR imaging. Lab Chip 7(9):1206–1208 12. Masubuchi Y (2006) Metabolic and nonmetabolic factors determining troglitazone hepatotoxicity: a review. Drug Metab Pharmacokinet 21(5):347–356 13. Myszka DG, Morton TA (1998) CLAMP: a biosensor kinetic data analysis program. Trends Biochem Sci 23(4):149–150 14. Dudoit S, Gentleman RC, Quackenbush J (2003) Open source software for the analysis of microarray data. Biotechniques Suppl 45–51

wwwwwwwwwwwwwwww

Part VII Data Analysis Techniques for Protein Function Microarrays

wwwwwwwwwwwwwwww

Chapter 21 Data Processing and Analysis for Protein Microarrays David S. DeLuca, Ovidiu Marina, Surajit Ray, Guang Lan Zhang, Catherine J. Wu, and Vladimir Brusic Abstract Protein microarrays are a high-throughput technology capable of generating large quantities of proteomics data. They can be used for general research or for clinical diagnostics. Bioinformatics and statistical analysis techniques are required for interpretation and reaching biologically relevant conclusions from raw data. We describe essential algorithms for processing protein microarray data, including spot-finding on slide images, Z score, and significance analysis of microarrays (SAM) calculations, as well as the concentration dependent analysis (CDA). We also describe available tools for protein microarray analysis, and provide a template for a step-by-step approach to performing an analysis centered on the CDA method. We conclude with a discussion of fundamental and practical issues and considerations. Key words: Protein microarray, Concentration dependent analysis, Z score, Differential expression analysis, Bioinformatics

1. Introduction Protein microarray technology offers direct detection and quantification of protein expression, the endpoint of both molecular and cellular function in health and disease. The potential of protein microarrays is great in research applications and future clinical diagnostics due to their high-throughput with minimal sample requirements (1). Protein microarrays have been used for elucidation of protein function and signaling (2), protein–protein interactions (3), detection of bacteria and toxins (4), drug discovery (5), and identification of protein biomarkers (6). This technology can be used to measure a range of protein properties including protein–protein interactions, protein–phospholipid interactions, protein kinase substrates, small molecule targeting, and antibody–antigen interactions (7–9). Thus, the potential for Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_21, © Springer Science+Business Media, LLC 2011

337

338

DeLuca et al.

protein microarrays covers a wide range of applications in both research and diagnostics. The large quantity of data collected through protein microarray analysis necessitates a series of computational processing steps to properly arrive at biological conclusions (10). Protein microarrays, however, are still in their infancy in comparison to nucleotide microarrays. Thus, a comprehensive analysis software solution has yet to be produced. Software packages for protein microarray data analysis are available (11), but have limitations compared to their more mature DNA microarray analysis counterparts; while some tools address protein microarray-specific issues such as protein concentration correction, (12, 13) the comprehensive analysis and statistics packages available for DNA chips (14, 15) are still lacking. Some authors have used DNA microarray analysis methods for protein microarrays (16). However, the lessons of DNA microarray analysis are only a starting point for developing robust protein microarray approaches. Future protein microarray analysis pipelines are likely to combine components specific to protein microarrays with generic microarray analytical methods. The purpose of this chapter is to provide the reader with the background and framework necessary to develop a successful data analysis strategy for protein microarrays. In line with a computational approach, Subheading 2 contains neither reagents nor devices, but rather a series of key algorithms followed by a list of software tools which implement them. This catalog represents the computational toolbox from which the investigator may draw the necessary components for data analysis. It should therefore be noted that it is not necessary to combine all of these tools into one analysis and that it is up to the reader to pick and choose depending on the specific application. Some of these techniques have an origin in DNA microarray analysis, such as spot-finding on slide images, Z-score calculations, and significance analysis of microarrays (SAM). Other techniques have been developed solely for protein microarray analysis, such as the concentration dependent analysis (CDA). To aid the reader in assembling these components together into a complete strategy, Subheading 3 contains a simple yet inclusive step-by-step analysis of protein microarray data. In this example, we walk through a CDA-based methodology for measuring differential antibody expression in leukemia patients before and after immunotherapy. Here, we utilize ProtoArray from Invitrogen, a commercially available high-density protein microarray platform. However, the method can be generalized to other platforms and applications. To address technical issues that arise in following this method, including crossplatform applicability, the chapter concludes with Subheading 4 in which the pitfalls and practical considerations are discussed.

Data Processing and Analysis for Protein Microarrays

339

2. Materials 2.1. A nalytical Tools 2.1.1. Determination of Spot Intensity

2.1.2. Z-Score Analysis

Computational processing of microarray data begins with the acquisition of a digital image representation of the signal intensity of the protein spots on the microarray. A software is necessary to determine the boundaries of spots in the scanned image, and to convert the pixel values into a file correlating protein identifiers with their numeric signal intensities. The GenePix Pro software (Molecular Devices, Union City, CA) is typically used for this task. The user must first align a grid of circles, representing spot boundaries, to the scanned image. The position and size of the circles must then be adjusted to ensure proper alignment. GenePix offers the capability to calculate signal intensity by determining the mean or median background in the vicinity of the circle as well as the mean or median signal within the spot. The calculation using the median is very robust and is not typically affected by small boundary misalignments or small artifacts. The program then creates an output file, with the file extension: GAL. In a typical microarray analysis, the investigator wants to determine which signals are significantly different from the expected values. Calculating the Z score, also called normal score, is a convenient method used for this task. The Z-score equation is: ZS =

SS − m , s

where Zs is the Z score for the s th spot, Ss is the signal for that spot, m is the mean signal across all spots, and s is the standard deviation across all spots. Thus, the Z score represents the distance of a given spot’s signal from the mean signal in units of standard deviations. When the population of signals has a normal distribution, samples with a Z score of 3 or greater in magnitude are in the 99.7th percentile. 2.1.3. ConcentrationDependent Analysis

A problem that is unique to protein microarrays is the variety in quantity of spotted material on the chip (13). Higher concentrations of spotted proteins produce higher absolute signals. Simply dividing the signal by concentration does not correct for this properly, because this correction favors the signal of spots with low protein concentration. Instead, a Z score using the CDA technique can be calculated (13). First, signals are sorted by their spotted protein concentration. Subsequent calculations are then performed within sliding windows in which the concentrations of spotted proteins are similar. The Z score for the given signal is

340

DeLuca et al.

then calculated using the mean and standard deviation of the signal values within that window: ZS =

SS − mw , sw

where Zs is the Z score for the sth spot, Ss is the signal for that spot, mw is the mean signal for the spots within the window, and sw is the standard deviation for spots within the window. If the concentrations within a given window vary too greatly (as defined by a reasonable threshold), the window is contracted by exclusion of the values at the edges. Additionally, outliers are defined as spots having signals greater than a defined number of standard deviations away from the mean (the usual default value is 3). Outliers are removed iteratively, with the mean and standard deviation recalculated after outliers are removed, and then any new outliers identified. The iteration stops when the recalculation identifies no further outliers within the window. 2.1.4. Differential Expression Analysis

A typical protein microarray analysis scenario involves the comparison of protein expression in samples under two different conditions A and B, for example before versus after treatment, healthy versus infected, etc. Working with Z scores, the difference (Zdiff) can simply be calculated as the difference between the Z scores of each protein spot under the two conditions, ZA−ZB, for all given spots. Alternately, differences can be expressed as a percentage, or fold increase. That is, Zmult is calculated as ZA/ZB. In practice, Zdiff tends to call false positives when the values are very large (unpublished exploratory work). Conversely, Zmult tends to over-represent spots where signals are very small. Therefore, it has proven useful to combine these two tests and select values where both Zdiff and Zmult are above a threshold (13). Figure 1 provides an example of a differential antibody expression analysis comparing the immunological profiles of leukemia patients before and after immunotherapy.

2.1.5. Differential Expression Analysis with Repeat Experiments

SAM, originally developed for nucleic acid microarrays (18), has also been applied to protein microarrays (16). This method addresses the problem of a large number of chance hits that are generated when large datasets are analyzed using standard T test P values. SAM addresses this problem by applying the T test in a gene-specific manner (18). A difference score, dp for a protein, p, is calculated by: dp =

x A ( p) − x B ( p) , sp

where x A (p) is the average signal for protein, p under conditions A, and x B (p) is the average signal for that protein under conditions B. In the denominator, sp is the standard deviation specific

Data Processing and Analysis for Protein Microarrays

341

Fig. 1. Differential expression analysis comparing antibody expression in CML patients before and after immunotherapy (17). Differential expression analysis was applied to determine antibody reactivity that was significantly increased after therapy (triangles). Such hits are expected to be found high above the diagonal.

to the protein, p, across all repeated measurements. In essence, this score reflects the differences in expression of each protein relative to the standard deviations of the repeated measurements. This score is then compared to a threshold to determine significance. In the author’s original study, a threshold of dp >1.2 resulted in a false discovery rate of 18% (18). Increasing the threshold would lower the false discovery rate at the expense of missing some true hits. 2.2. Software Tools

The following software tools are implemented versions of the analytical methods described in Subheading 2.1. Proprietary as well as free-for-use software are included. Only software which was available at the time of publication has been included.

2.2.1. GenePix Pro

GenePix Pro from Molecular Devices, Union City, CA. This software package is able to process a scanned microarray slide and interpret the spots to produce a file of signal intensities – the starting data for microarray analysis. URL: http://www.moleculardevices.com/ pages/software/gn_genepix_pro.html.

2.2.2. Prospector

Prospector from Invitrogen, Carlsbad, CA. This software calculated Z scores for data from their ProtoArray platform. The output format from GenePix Pros (GAL) can be read directly into this program. URL: http://tools.invitrogen.com/ content.cfm?pageid=10400.

2.2.3. ProtMAT

ProtMAT from Dana-Farber Cancer Institute, Boston, MA. This is an online webtool, which implements the CDA algorithm.

342

DeLuca et al.

It is capable of reading data files in the Prospector output format. CDA Z-score calculations as well as differential expression analysis can be preformed. URL: http://cvc.dfci.harvard.edu/protmat/. 2.2.4. Significance Analysis of Microarrays

SAM from Stanford University, Stanford, CA. The authors of the SAM algorithm provide an implementation for download which requires R, Microsoft Excel, and Windows 2000 or XP. The implementation is capable of producing graphs and analyzing time series as well as gene set enrichment. URL: http://www-stat. stanford.edu/~tibs/SAM/.

3. Methods The following steps represent a simple yet inclusive pipeline for analyzing protein microarray data. This example focuses on differential antibody expression analysis using the Invitrogen ProtoArray platform but can be modified to accommodate other platforms. This analysis centers on the CDA method, but could be supplemented with additional tools or algorithms mentioned in Subheading 2. 3.1. S ignal Acquisition

1. Scan the microarray slide to produce an image file (see Notes 1–3). 2. Open image file in GenePix software. 3. To localize spots within the scanned image, an array file (*.GAL) specific to the microarray version must be loaded. This file contains a grid of circles in blocks, which define the borders of the spots (see Notes 1–3). 4. Loading the GAL file provides a grid of circles, which define the spot boundaries. Align these spots by shifting the blocks and rows of circles into position. 5. Once the blocks are aligned, the resulting signals can be exported to a results file (*.GPR).

3.2. Format Preparation

1. Open the Prospector application from Invitrogen (see also Notes 4 and 5). 2. Select “Immune Response Profiling” as the type of application for this analysis. 3. Load the GPR file(s) produced in the last step of the previous section (Subheading 3.1, step 5). 4. Clicking the “Show” button will produce the correct data format for the next step and will automatically open the data in Excel. 5. In excel, save the data as a text file (tab-delimited).

Data Processing and Analysis for Protein Microarrays

3.3. H it Calling

343

1. To perform hit calling by the CDA method, open the ProtMAT website in a web browser (see Subheading 2.2.3). 2. Load the text Subheading 3.2.

file

generated

in

the

last

step

of

3. On the Settings page, set the threshold to 3. 4. Turn off the log data option. 5. Leave the additional settings to their default values. 6. Review the list of hits. If the list is unmanageably long, or much too short then click the Calculate tab and adjust the threshold accordingly (see Notes 1, 2, 6–8). 7. Use the Export tab to save the results. You may choose what kind of data to export. Select settings, hits, and statistics (see Notes 3, 6–8). 3.4. Comparing Multiple Microarrays

1. Open the ProtMAT website in a web browser (see Subheading 2.2.3). 2. Load two of the data files in the format produced by the instructions in Subheading 3.2. 3. Continue the analysis by performing steps 3–7 of Subheading 3.3.

4. Notes When following the steps outlined in Subheading 3, the investigator is likely to encounter a series of technical issues. Many of these issues are addressed in the following section. However, in addition to the technical issues, there are important fundamental considerations that have to be taken into account in order to ensure the successful application of protein microarray technology. Therefore, (see Note 1), which relates directly to the steps in (see Notes 6-8), is succeeded by (see Notes 9–14), which addresses the broader challenges facing analysts. 1. High-resolution scanning is required to provide enough pixels per spot for accurate measurement. A pixel size between 10 and 25 mm will produce quality results. The higher resolution 10 mm scan may be desirable for publication figures. 2. Saturation should be assessed by viewing a histogram of pixel intensities. If a large number of pixels have the maximum value, the PMT gain may have to be adjusted. 3. When aligning circle boundaries to the image file in GenePix Pro 5.0, there is an automated block alignment feature which can expand or contract circle size (Fig. 2). Experience has

344

DeLuca et al.

Fig. 2. Aligning spot boundary grid to scanned slide using GenePix Pro software using the automated alignment mode without proper manual verification or parameter settings (left ) compared to a manual alignment (right ) that preserves spot boundary size and spacing. (a) Some spots are off-center; the spots on the microarrays are actually quite nicely arrayed so they should be evenly spaced. (b) Some spots are exceedingly small and (c) some spots are exceedingly large; although spot size can vary slightly, the boundaries represented here do not reflect reasonable size variation. (d) Spots with vertical lines were identified by the alignment program as unreliable and are flagged as such when the data are exported.

shown that it is best not to apply this feature and to adjust spots on a block-by-block or row-by-row basis. If automated mode is preferred, it is recommended to adjust the parameters, such as minimum and maximum spot size. Whether a manual or automated method is applied, all alignments should be visually verified to reduce erroneous signal identification during downstream processing. Spotted protein concentration can be acquired from the protein microarray manufacturer. Alternatively, it can be determined by the investigator specifically for the lot of arrays used in the investigation. 4. To acquire spot concentrations from the manufacturer, follow their instructions. For the ProtoArray protein microarrays, load the GPR file into the Prospector software provided by the manufacturer. Prospector is capable of automatically downloading the “Protein Information File” containing spotted protein concentrations. Alternatively, it can be acquired from the Invitrogen website by providing the product bar code. 5. To determine protein concentrations when these are not available, several microarrays from a given printing lot must be selected. Ideally, 1–2 microarrays are selected from the beginning, middle, and end of the printing lot. These are probed

Data Processing and Analysis for Protein Microarrays

345

with an antibody against the tag that is constructed into the synthesized proteins, for example glutathione-S-transferase (GST). Medians of these measured fluorescence values are then determined using steps in Subheading 3.1. These values can then be formatted as needed for further analysis. 6. When performing the CDA analysis, the most important parameter is the threshold. It may be reasonable and necessary to increase the threshold value to limit the number of hits if the quantity is unmanageable for further downstream experimentation. If few or no hits are found, however, the threshold should not be reduced below 3 because of the loss in statistical significance. 7. On the ProtMAT website, there is a summary statistics page. Inspect the skewness and kurtosis to assess whether the data points are distributed normally. The skewness and kurtosis (with base distribution as normal) must be of small magnitude for the data to follow an approximate normal distribution. One could use D’Agostino’s K-squared test for skewness (19) and Anderson–Darling test for kurtosis for testing departures from zero (20). One could also directly perform a test of normality, using a separate Anderson–Darling test or Shapiro–Wilks test (21, 22). If the assumption of normality is rejected by any of these tests, downstream analyses based on the normality assumptions are not appropriate. In such situations, consultation with a statistician to propose appropriate parametric or nonparametric methodology is recommended. 8. When exporting the results, it is also possible to output all the datapoints, as well as the controls values. This will produce large files, and should be done separately from the hits and statistics data. 9. The mathematical and computational tools, algorithms, and methods presented above represent the essential components of a protein microarray analysis process. However, successful application of protein microarray technology requires proper experimental design. Factors such as cell purity, time scale, tissue type can have a larger impact on the experiment than which tools are used in the analysis. 10. Artifacts can be caused by inefficient lysis buffer, inconsistent sample processing, varying optimal conditions for different protein interactions, nonspecific associations, and requirements for appropriate conformation (23). 11. Beyond the scope of this chapter, more complex modeling is certainly possible. Normalizing has been shown to improve hit-calling in certain cases, although such transformations can

346

DeLuca et al.

also lead to information loss (10). Inclusion of SAM analysis would address signal fluctuations at the protein level (18). Alternatively, a Wilcoxon Rank-Sum could provide a useful nonparametric approach (24). Furthermore, more complex designs require formal statistical approach for analysis development. This is a recommended way of ensuring that such analysis tools are applied appropriately. 12. There are currently no standardized file formats for representing protein microarray data. Consequently, file conversion and data processing are necessary in many cases. For instance, running the CDA analysis tool referenced above requires data in a format that is arrived at by using the same software as we have described in Subheading 3. When this is not the case, such projects require programming ability or inclusion of a bioinformatician. 13. The reproducibility of protein microarray experiments has been determined to be quite high by performing repeat experiments (unpublished observations). While multiple measurements are desirable to overcome variability in sampling handing or biological factors, little correction for variability in the array as a platform is necessary.

Acknowledgments C.J.W. acknowledges support from the Department of Defense (W81XWH-07-1-0080), the Miles and Eleanor Shore Award, NCI (5R21CA115043-2), the Early Career Physician-Scientist Award of the Howard Hughes Medical Institute, and is a Damon-Runyon Clinical Investigator supported (in part) by the Damon-Runyon Cancer Research Foundation (CI-38-07). O.M. acknowledges support from a Medical Student Fellowship of the Howard Hughes Medical Institute. References 1. Hartmann M, Roeraade J, Stoll D, Templin MF, Joos TO (2009) Protein microarrays for diagnostic assays. Anal Bioanal Chem 393:1407–1416 2. Wolf-Yadlin A, Sevecka M, MacBeath G (2009) Dissecting protein function and signaling using protein microarrays. Curr Opin Chem Biol 13:398–405 3. Coleman MA, Beernink PT, Camarero JA, Albala JS (2007) Applications of functional protein microarrays: identifying protein-protein interactions in an array format. Methods Mol Biol 385:121–130

4. Ehricht R, Adelhelm K, Monecke S, Huelseweh B (2009) Application of protein arraytubes to bacteria, toxin, and biological warfare agent detection. Methods Mol Biol 509:85–105 5. Michaud GA, Salcius M, Zhou F, Papov VV, Merkel J, Murtha M, Predki P, Schweitzer B (2006) Applications of protein arrays for small molecule drug discovery and characterization. Biotechnol Genet Eng Rev 22:197–211 6. Kerschgens J, Egener-Kuhn T, Mermod N (2009) Protein-binding microarrays: probing disease markers at the interface of proteomics and genomics. Trends Mol Med 15:352–358

Data Processing and Analysis for Protein Microarrays 7. Hall DA, Ptacek J, Snyder M (2007) Protein microarray technology. Mech Ageing Dev 128:161–167 8. Lubomirski M, D’Andrea MR, Belkowski SM, Cabrera J, Dixon JM, Amaratunga D (2007) A consolidated approach to analyzing data from high-throughput protein microarrays with an application to immune response profiling in humans. J Comput Biol 14:350–359 9. MacBeath G, Schreiber SL (2000) Printing proteins as microarrays for high-throughput function determination. Science 289:1760–1763 10. Brusic V, Marina O, Wu CJ, Reinherz EL (2007) Proteome informatics for cancer research: from molecules to clinic. Proteomics 7:976–991 11. Zhu X, Gerstein M, Snyder M (2006) ProCAT: a data analysis approach for protein microarrays. Genome Biol 7:R110 12. White AM, Daly DS, Varnum SM, Anderson KK, Bollinger N, Zangar RC (2006) ProMAT: protein microarray analysis tool. Bioinformatics 22:1278–1279 13. Marina O, Biernacki MA, Brusic V, Wu CJ (2008) A concentration-dependent analysis method for high density protein microarrays. J Proteome Res 7:2059–2068 14. Dennis G Jr, Sherman BT, Hosack DA, Yang J, Gao W, Lane HC, Lempicki RA (2003) DAVID: database for annotation, visualization, and integrated discovery. Genome Biol 4:P3 15. Reich M, Liefeld T, Gould J, Lerner J, Tamayo P, Mesirov JP (2006) GenePattern 2.0. Nat Genet 38:500–501 16. Hueber W, Kidd BA, Tomooka BH, Lee BJ, Bruce B, Fries JF, Sonderstrup G, Monach P,

347

Drijfhout JW, van Venrooij WJ, Utz PJ, Genovese MC, Robinson WH (2005) Antigen microarray profiling of autoantibodies in rheumatoid arthritis. Arthritis Rheum 52:2645–2655 17. Biernacki MA, Marina O, Zhang W, Liu F, Bruns I, Cai A, Neuberg D, Canning CM, Alyea EP, Soiffer RJ, Brusic V, Ritz J, Wu CJ (2010) Antigen targets of remission-inducing immune therapy are expressed on CML progenitor cells. Cancer Res 70(3): 906–915 18. Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci USA 98:5116–5121 19. D’Agostino DB, Belanger A, D’Agostino RB (1990) A suggestion for using powerful and informative tests of normality. Am Stat 44:316–321 20. Anderson TW, Linfeng Y (1996) Adequacy of asymptotic theory for goodness-of-fit criteria for spectral distributions. J Time Ser Anal 17:533–552 21. Shapiro SS (1990) How to test normality and other distributional assumptions, Revth edn. ASQC, Milwaukee, WI 22. Shapiro SS, Wilk MB (1965) An analysis of variance test for normality (complete samples). Biometrika 52:591–611 23. Schena M (2005) Protein microarrays. Jones and Bartlett, Sudbury, MA 24. Guyon I, Weston J, Barnhill S, Vapnik V (2002) Gene selection for cancer classification using support vector machines. Machine Learn 46:389–422

wwwwwwwwwwwwwwww

Chapter 22 Database Resources for Proteomics-Based Analysis of Cancer Guang Lan Zhang, David S. DeLuca, and Vladimir Brusic Abstract Biological/bioinformatics databases are essential for medical and biological studies. They integrate and organize biologically related information in a structured format and provide researchers with easy access to a variety of relevant data. This review presents an overview of publicly available databases relevant to proteomics studies in cancer research. They include gene/protein expression databases, gene mutation and single nucleotide polymorphisms databases, tumor antigen databases, protein–protein interaction, and biological pathway databases. Automated information retrieval from these databases enables efficient large-scale proteomics data analysis. Key words: Database, Proteomics, Protein microarray, Protein expression, Single nucleotide polymorphisms, Mutation, Tumor antigen, Protein–protein interaction, Biological pathway

1. Introduction Proteomics is the large-scale study of the protein complement to the complete genome (1). Proteomics is more complicated than genomics because an organism’s proteome varies from cell to cell. Considering alternative gene splicing, posttranslational modifications, and individual coding variants, the number of proteins in a human is estimated to be 106 – a much bigger number than that of the protein-coding genes, which ranges from 20,000 to 25,000. Proteomics focuses mainly on the study of expression, function, and structure of proteins. It deploys quantitative detection of proteins in cells and tissues, often based on comparison of different conditions, such as healthy vs. diseased, differentiated vs. undifferentiated, treated vs. nontreated (e.g., by drugs), and others. It promises to advance our knowledge and improve the prevention

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0_22, © Springer Science+Business Media, LLC 2011

349

350

Zhang, DeLuca, and Brusic

and treatment of cancer. Proteomics provides unique tools for discovery of biomarkers and therapeutic targets (2). Proteomics data are generated by state-of-the-art highthroughput technologies as well as traditional biological experiments. Protein microarrays (also known as biochips or proteinchips) are used to determine the presence and amount of proteins in biological samples. While concepts of protein microarrays and nucleotide microarrays are somewhat similar, there is not always a direct correlation between mRNA levels and corresponding proteins expression. Correlation coefficient between protein and mRNA abundance in yeast is only r = 0.4, indicating that mRNA data are not predictive of protein expression levels (3). Understanding molecular mechanisms therefore requires the direct study of proteins. Furthermore, protein arrays are critical for detection of posttranslational modifications which are important for study of oncogenesis (4). However, both nucleotide and protein data are necessary for study of protein functions. For example, most protein variations are inferred from genetic variations (5). There are two main types of protein microarrays: analytical arrays and functional arrays. Analytical microarrays use wellcharacterized molecules, such as antibodies and peptide-MHC complexes, as probes. They can be used for protein expression profiling, biomarker identification, cell surface marker/glycosylation profiling, and clinical diagnosis. For example, a popular type of analytical array is the antibody microarray, on which a collection of capture antibodies are spotted and fixed to enable detecting of antigens (6). It is often used for detecting protein expression from cell lysates. Analytical microarrays are also used for detection of special biomarkers from serum or urine for diagnostic applications. In functional microarrays, many different proteins, even the total proteome of an organism, are used as probes. They are used to detect protein activities, such as protein–protein, protein–DNA, protein–drug, and protein–peptide interactions (7). The interpretation of the data generated by microarrays depends heavily on the access and use of existing knowledge scattered across multiple databases. Thus, database applications are essential proteomics tools, given the complexity of biomarker identification and functional characterization of proteins and their interactions. The vast amount of data produced by proteomics experiments coupled with their heterogeneity poses challenges in translating the raw results into relevant and understandable information. Because of the significant data management issue, databases have become a standard component of bioinformatics pipelines for the high-throughput analysis of data from proteomics experiments. Databases provide researchers with a wide variety of biologically relevant data, including gene and protein sequences and related

Database Resources for Proteomics-Based Analysis of Cancer

351

feature annotation. When these data are stored in a structured repository, they provide a foundation for bioinformatics analysis tools. For example, DAVID (The Database for Annotation, Visualization, and Integrated Discovery) knowledgebase (8) is a backend database used by DAVID bioinformatics tools (9). Cancer is a large group of diseases in which a group of cells grows uncontrollably. Malignant cells can destroy adjacent tissues and migrate from the original site to other parts of the body (metastasis). Proteomics has increasingly been applied to cancer studies including cataloguing of the proteome, discovery of cancer-specific biomarkers, uncovering molecular mechanisms of disease, and identification of novel protein targets for therapeutic intervention (10, 11). Several proteomic approaches, such as serological identification of antigens by recombinant expression cloning (SEREX, SErological identification of antigens by Recombinant EXpression cloning) (12), serological proteome analysis (SERPA), and protein microarrays (7), two-dimensional gel electrophoresis, liquid chromatography, mass spectrometry, isotope-coded affinity tag have been developed for identification of tumor-associated antigens (TAAs) and the autoantibodies that recognize them. This chapter introduces a number of databases related to proteomics studies in cancer with brief description of their content and functionality. We have included databases containing gene/ protein expression data produced by microarray studies, nextgeneration sequencing, and other high-throughput experiments, gene mutation and SNP databases, tumor antigen databases and T-cell tumor antigen databases, and protein interaction and pathway databases.

2. Databases Comprehensive, general-use protein databases are valuable starting points for data analysis in proteomic studies. One is the UniProt database (www.ebi.ac.uk/uniprot/), a central resource for protein sequences and functional information (13). UniProt is created by combining Swiss-Prot (high-quality manually curated and validated proteins), TrEMBL (computationally translated and annotated proteins), and PIR (a pan-database resource integrator). UniProt provides information on protein functions, domains structures, posttranslational modifications, and cross-links to about 100 external databases. Another useful general database is GeneCards (http://www.genecards.org/), which offers genomic, proteomic, transcriptomic, genetic, and functional information on all known and predicted human genes (14). The information provided by GeneCards includes orthologies, disease relationships,

352

Zhang, DeLuca, and Brusic

mutations and single nucleotide polymorphisms (SNP), gene expression, gene function, pathways, protein–protein interactions (PPIs), and so on. While general-use protein databases provide essential information on specific proteins, proteomics-based cancer research relies on a spectrum of specialized databases. These can be classified into groups according to their content and functionality. They include gene/protein expression databases, cancer mutation and SNP databases, tumor antigen databases, databases of cancer-associated genes, protein interaction and pathway databases, and others. In the following sections, we have chosen to include only those databases which fulfill the following criteria: (1) the databases are currently active and freely available online; and (2) the databases have been published in peer-reviewed journals and are widely used by research communities. 2.1. Gene/Protein Expression Databases

The constant advancement of high-throughput technologies for measuring gene and protein expression has resulted in an explosive growth of expression datasets. Storing these results in public databases ensures their availability to researchers to enhance their studies. The goal is to correlate these datasets with observable phenotypes to make a meaningful interpretation. The first step toward this goal is storing the data in a common format. Standardized formatting subsequently facilitates downstream large-scale integrative analyses. Several commonly used microarray databases are listed in Table 1. Gene Expression Omnibus (GEO) (15) at NCBI (National Center for Biotechnology Information), CIBEX (Center for Information Biology gene EXpression database) at NIG (National Institute of Genetics) in Japan (16), and ArrayExpress at EBI (European Bioinformatics Institute) (17) are public repositories that archive microarray, next-generation sequencing, and other

Table 1 Online databases on gene/protein expression Database

Description

URL

GEO

The largest public repository archiving microarray, www.ncbi.nlm.nih.gov/geo next-generation sequencing, and other highthroughput experimental data developed at NCBI

CIBEX

A public database for microarray data at NIG in Japan

http://cibex.nig.ac.jp/

ArrayExpress

A repository for microarray-based gene expression data developed at EBI

www.ebi.ac.uk/arrayexpress

CGED

A database with gene expression profiles and accompanying clinical information

http://lifesciencedb.jp/cged/

Database Resources for Proteomics-Based Analysis of Cancer

353

high-throughput experimental data such as serial analysis of gene expression (SAGE) and proteomics data submitted by the scientific community. These databases are in compliance with major community-derived scientific reporting standards such as “Minimum Information About a Microarray Experiment” (MIAME). As of 2009, GEO contains over 10,000 experiments comprising 300,000 samples and 16 billion individual abundance measurements, for over 500 organisms. CIBEX hosts data from 62 experiments, 105 arrays, and 1970 hybridizations. There are three components in ArrayExpress: the ArrayExpress Repository containing functional genomics experiments and supporting data, the ArrayExpress Warehouse hosting gene expression profiles and other biomeasurements, and the ArrayExpress Atlas which is a summary database and meta-analytical tool of ranked gene expression across multiple experiments and various biological conditions (18). As of 2009, the Repository contains information on more than 6,000 experiments comprising approximately 200,000 assays. On 5 Nov 2009, GEO and ArrayExpress agreed to exchange ultrahigh-throughput sequencing data to create a merged and comprehensive dataset. This complements the exchange of underlying raw data between SRA (Sequence Read Archive) and ERA (European Nucleotide Archive). Raw sequencing data submitted to ArrayExpress or GEO will be sent to ERA or SRA, respectively. Cancer Gene Expression Database (CGED) hosts data on breast, colorectal, hepatocellular, esophageal, thyroid, and gastric cancers. (19). Refer to Penkett and Bahler (20) for a more detailed review on public microarray databases. 2.2. Gene Mutation and SNP Databases

Cancer is a complex disease which arises as a consequence of accumulated mutations in somatic cells of an individual. Identification of the mutated genes that have implication in oncogenesis is the central aim of cancer research. Quite a few cancer-related gene mutation databases have been developed and are shown in Table 2. SNP databases provide an important complement to databases cataloguing cancer-associated genetic change. They can rapidly provide information concerning whether genetic change in cancer tissue represents a tumor-specific mutation or is a result of benign genetic variation. Certain genetic variants are also increasingly appreciated as a source of predisposition for developing cancer, other disease conditions, or therapeutic response (21). Although all mutation and SNP databases are nucleotidebased, they are important for characterization of protein functions. It is essential that the inferred protein mutations are checked against protein databases to identify mutations that are functionally important.

2.2.1. Databases that Catalog Tumor-Associated Mutations

Catalog of Somatic Mutations In human Cancer (COSMIC) is currently the most comprehensive resource for information on somatic mutations in human cancer (22). As of 2009, the latest

354

Zhang, DeLuca, and Brusic

Table 2 Online databases on human gene mutations and SNPs Database

Description

URL

COSMIC

Catalog of somatic mutations in human http://www.sanger.ac.uk/perl/ cancer containing >50,000 mutations genetics/CGP/cosmic

Human gene mutation A collection of germline mutations in www.hgmd.org/ database nuclear genes that are associated with human-inherited disease MutationView KMcancerDB

Human gene mutation database; The number of genes in the databases is less than what was claimed in the paper

http://mutview.dmb.med.keio. ac.jp/

Cancer gene census

A catalog of mutations in more than 400 cancer implicated genes

www.sanger.ac.uk/genetics/ CGP/Census/

IARC TP53 mutation database

TP53 gene variations identified in human populations and tumor samples

http://www-p53.iarc.fr/

CDKN2a database

The variants of CDKN2A recorded in human disease

https://biodesktop.uvm.edu/ perl/p16

Androgen receptor gene mutations database

374 Published mutations, most being http://androgendb.mcgill.ca/ point mutations identified in patients with androgen insensitivity syndrome

Breast cancer information core

A repository for all mutations and polymorphisms in genes related to breast cancer

http://research.nhgri.nih.gov/ bic/

GAC

A collection of gene mutations, loss of heterozygosity, and/or chromosome changes in tumors from humans, mice, or rats gathered from peerreviewed journals

www.niehs.nih.gov/research/ resources/databases/gac/

SNP500 cancer database

Information on over 3,400 SNPs in genes important in cancer

http://snp500cancer.nci.nih.gov/

dbSNP

A general catalog of genome variations

www.ncbi.nlm.nih.gov/projects/ SNP/

MedRefSNP

Information about 36,199 unique SNPs collected from the PubMed and OMIM databases

www.medclue.com/medrefsnp

Chromosomal abnormalities in cancer

A compilation of human cancers associated with chromosome aberrations

www.slh.wisc.edu/cytogenetics/ cancer/

Database Resources for Proteomics-Based Analysis of Cancer

355

release version 44 contains 101,138 mutations resulting from analysis of 13,501 genes and 419,018 tumors. Human Gene Mutation Database (HGMD) has a collection of germline mutations in nuclear genes in association with human-inherited disease (23). The database contains 93,347 mutation entries. The public version of the database is free for academic/nonprofit usage and a registration is required. MutationView, a human gene mutation database, contains mutation data for approximately 300 genes, most of which are involved in monogenic diseases. Among them, 42 genes are cancer-related and have been placed in a separate database named KMcancerDB (24). The Cancer Gene Census is a catalog of the genes for which mutations have been causally implicated in cancer (25). To date, more than 1% of all human genes are implicated via mutation in cancer. Of these cancerassociated genes, approximately 90% have somatic mutations, 20% bear germline mutations that predispose to cancer, and 10% show both somatic and germline mutations. In addition to these broad-scoped cancer mutation databases, there are more specialized databases which focus on specific cancer proteins and protein families. For example, P53 functions as a tumor suppressor through the regulation of the cell cycle. In humans, P53 is encoded by the Tp53 gene, which is the most frequently mutated gene in human cancer. The International Agency for Research on Cancer (IARC) Tp53 Mutation Database contains more than 10,000 entries of Tp53 somatic mutations, 144 entries of Tp53 germline mutations, and 13 entries of p53 polymorphisms (26). Another important cancer gene represented in a specialist database is cyclin-dependent kinase inhibitor 2A (CDKN2A or P16). Cdkn2a is a major melanoma predisposition gene. The CDKN2a Database provides information on germline and somatic variants of the Cdkn2a tumor suppressor gene recorded in human disease (27). Another specialist database, the Androgen Receptor Gene Mutations Database, shows associations between mutations in the androgen receptor (Ar) gene and male prostate cancer (28). Androgen regulated genes are critical for the development and maintenance of the male sexual phenotype. Some germline mutations in Ar are associated with the occurrence of breast cancer in males suffering from partial androgen insensitivity, while somatic mutations in the Ar are associated with metastatic prostate cancer. The database contains 605 entries of reported mutations and 70 AR-interacting proteins (28, 29). Breast Cancer Information Core at National Human Genome Research Institute facilitates the detection and characterization of breast cancer susceptibility genes. To use the database, membership is needed, which is open to all and can be obtained through online registration. The Genetic Alterations in Cancer (GAC) database is a webbased system for collecting and summarizing data reported in the

356

Zhang, DeLuca, and Brusic

publications for genetic alterations in various types of human and rodent tumors (30). The system also provides a tool for assessing groups of related genes for mutations, allelic loss, and homozygous deletions based on species and tumor site. It gives users a broader view of tumor development and facilitates comparative analysis of multiple genes in genetic pathways to cancer rather than focusing on single genes. 2.2.2. SNP Databases

SNPs are a class of genetic markers that aid in identifying individuals at high risk of developing certain cancers and developing tailored medication. For example, the Tp53 Arg/Pro polymorphism at codon 72 may be linked to an increased risk of lung cancer (31). Several SNP databases are listed in Table 2. SNP500 Cancer Database is a component of the National Cancer Institute Cancer Genome Anatomy Project. It provides sequence and genotype assay information for over 13,400 SNPs which are useful in mapping complex diseases, such as cancer (32). The database provides gene locations and >200 bp of surrounding annotated sequence (including nearby SNPs). Furthermore, frequency information and per subpopulation as well as calculation of Hardy-Weinberg equilibrium for each subpopulation are also provided. The dbSNP was set up at NCBI to serve as a central repository for genetic variation (33). MedRefSNP provides integrated information about SNPs collected from the PubMed and OMIM databases (34).

2.2.3. Databases Cataloguing other Genetic Abnormalities

Chromosomal abnormalities can be caused by mutations which change the number of chromosomes (numerical abnormalities) or change the structure of the chromosome (structural abnormalities). One of the important causes of cancer is gene translocations and gross gene deletions. The Chromosomal Abnormalities in Cancer website, hosted by the Wisconsin State Laboratory of Hygiene, provides information on several human cancers associated with chromosome aberrations.

2.3. Tumor Antigen Databases

Since the identification of MAGEA1 as a tumor antigen recognized by cytolytic T lymphocyte on human melanoma, the number of characterized tumor antigens has exponentially increased (35). The identification of tumor antigens remains a high priority in cancer research and is an essential component in developing immune-based strategies to combat cancer. The databases listed in Table 3 are useful resources for the study of immune responses against tumors. The term cancer-testis (CT) antigen was proposed by Scanlan and colleagues to encompass a heterogeneous groups of antigens, which show restricted expression in cancer and testis and restricted immunogenicity in cancer patients (36). CT antigens are ideal targets for cancer immunotherapy due to their restricted expression pattern.

Database Resources for Proteomics-Based Analysis of Cancer

357

Table 3 Online databases on human tumor antigens Database

Description

URL

CTdatabase

A repository of cancer-testis antigen data

http://www.cta.lncc.br/

CAPD

An analysis system for cancerrelated data

www.bioinf.uni-sb.de/CAP/

Cancer immunome database

A repertoire of antigens eliciting antibody responses in cancer patients

http://ludwig-sun5.unil.ch/ CancerImmunomeDB/

Cancer immunity peptide Four data tables containing 129 database tumor antigens with defined T-cell epitopes

http://www.cancerimmunity.org/ peptidedatabase/Tcellepitopes.htm

TANTIGEN

http://cvc.dfci.harvard.edu/tadb/

A human tumor T-cell antigen database

CT database provides information on CT antigens, including gene names and aliases, RefSeq accession numbers, genomic location, known splicing variants, gene duplications, and additional family members. It also provides gene expression at the mRNA level in normal and tumor tissues, manually curated data related to mRNA and protein expression, antigen-specific immune responses in cancer patients, and links to PubMed for relevant CT antigen articles (37). The Cancer-Associated Protein Database (CAPD) was built upon SEREX database, in which the sequences were obtained by screening cDNA expression libraries using serum from cancer patients as probes and sequencing individual reactive clones (38). The database also contains microarray, epigenetic, and immunostaining data. It aims to provide information covering all the gene products against which an immune response has been documented in cancer patients. The Cancer Immunome database is a continuation of the SEREX database in a more organized form. It is an access point of information about all of the gene products against which an immune response has been documented in cancer patients (39). The development of T-cell immunity against cancer has the potential to effective rejection and elimination of tumor cells, and hence, T cell-defined tumor antigens are a particular focus of several databases. The Cancer Immunity Peptide database provides four static data tables, containing 129 human tumor antigens with defined T-cell epitopes (40). Among them, 45 entries are tumor antigens resulting from mutations, 29 are shared tumorspecific antigens, 12 are differentiation antigens, and 43 are antigens overexpressed in tumors. For each tumor antigen, a link to

358

Zhang, DeLuca, and Brusic

GeneCards (14) and literature reference are provided. The database is relatively simple without any query function or analysis tool. A list of mouse and human tumor T-cell antigens were reported in (41). The Tumor T-Cell Antigen Database (TANTIGEN) is a data source and analysis platform for cancer vaccine target discovery focusing on human tumor-derived HLA ligands and T-cell epitopes. It contains 4,006 curated antigen entries representing 251 unique proteins. TANTIGEN also provides information on experimentally validated T-cell epitopes and HLA ligands, antigen isoforms, antigen sequence mutations, and tumor antigen classification. Analysis tools integrated in the database include search tool for querying the dataset, multiple sequence alignment of antigen isoforms, sequence similarity search using BLAST, visual display of T-cell epitopes/HLA ligands, and prediction of binding peptides of 15 HLA Class I and Class II alleles. TANTIGEN is the most comprehensive database on Tumor T-cell antigens so far. 2.4. Databases of Cancer-Associated Genes

This section provides descriptions of two databases that integrate multiple heterogeneous datasets, including molecular data, clinical data, and experimental data, together with computational analysis tools to advance the cancer research. The Cancer Genome Anatomy Project (CGAP) aims to improve detection, diagnosis, and treatment for cancer patients through the analysis of the gene expression profiles of normal, precancer, and cancer cells (42). Its website provides genomic data for humans and mice, including transcript sequence, gene expression patterns, SNPs, clone resources, and cytogenetic information. The Mitelman Database of Chromosome Aberrations in Cancer (http://cgap.nci.nih.gov/Chromosomes/Mitelman) is part of CGAP. It is one of the largest online catalogs of cytogenetic aberrations in cancer, containing 56,694 cases as of 2009. The database relates chromosomal abnormalities to tumor characteristics (43). The Mouse Tumor Biology Database (MTBD) supports the use of the mouse as a model system of hereditary and induced cancers (44). The database provides access to tumor names and classifications, tumor incidence and latency data in different strains of mice, tumor pathology reports and images, information on genetic factors that are associated with tumor biology, and the references associated with these data (Table 4).

2.5. Protein Interaction and Pathway Databases

Biological pathways are the blueprints of cellular actions and they describe the roles of genomic entities in various cellular mechanisms. Human PPI data are important for understanding molecular signaling networks and the functional roles of biomolecules. Approaches involving pathway and PPI data are useful for analyzing microarray data and for generating testable hypotheses. Much effort has been put into pathway studies and many pathway

Database Resources for Proteomics-Based Analysis of Cancer

359

Table 4 Databases of cancer-associated genes Database

Description

URL

Cancer genome anatomy project

Containing genomic data for humans and mice, including transcript sequence, gene expression patterns, SNPs, clone resources, and cytogenetic information

http://cgap.nci.nih.gov.ezp-prod1.hul. harvard.edu/

MTBD

Supports the use of the mouse as a model system of hereditary cancer

http://tumor.informatics.jax.org/ mtbwi/

databases have been developed and made available online. Table 5 contains a list of databases providing information on PPI and biological pathways. Pathguide is a meta-database which contains information about 302 biological pathway resources (45). They include databases on metabolic pathways, signaling pathways, transcription factor targets, gene regulatory networks, genetic interactions, protein–compound interactions, and PPIs. Pathguide serves as a starting point for biological pathway analysis. A recent paper reviewed the major databases of human pathways and discussed how to use the information for the reconstruction of signaling pathways (46). The KEGG pathway database is a collection of manually drawn pathway maps representing our knowledge on the molecular interaction and reaction networks involved in metabolism, genetic information processing, environmental information processing, cellular processes, and pathogenesis (47). The BioCarta website catalogs and summarizes classical pathways as well as newly suggested pathways information on more than 120,000 genes from multiple species, including human and mouse. Reactome is an expert-curated knowledgebase of human reactions and pathways. As of 2009, it hosts 2,975 human proteins, 2,907 reactions, and 4,455 literature citations (48). The Pathway Interaction Database (PID) hosts 100 human Pathways containing 6,298 interactions curated by domain experts from US National Cancer Institute and Nature Publishing Group. It also encompasses 329 human pathways containing 7,418 interactions imported from BioCarta and Reactome (49). The Human Pathway Database (HPD) combines heterogeneous human pathway data from PID, Reactome, BioCarta, KEGG, or indexed from the Protein Lounge Web sites (50). So far, HPD contains

360

Zhang, DeLuca, and Brusic

Table 5 Protein interaction and pathway database Database

Description

URL

Pathguide

A metadatabase providing an overview of web-accessible biological pathway and network databases

http://www.pathguide.org/

KEGG pathway database

A collection of manually drawn pathway maps

www.genome.jp/kegg/ pathway.html

BioCarta pathway

A collection of pathways for multiple species

http://www.biocarta.com/ genes/index.asp

Reactome

A curated resource for human pathway data

http://www.reactome.org/

PID

A collection of curated pathways related to human molecular signaling, regulatory events, and key cellular processes

http://pid.nci.nih.gov/

HPD

Providing combined view connecting human proteins, genes, RNAs, enzymes, signaling, metabolic reactions, and gene regulatory events

http://bio.informatics.iupui. edu/HPD

NetPath

A catalog of annotations for cancer and immune signaling pathways

www.netpath.org/

HAPPI

One of the most comprehensive public compilation of human protein interaction information

http://bio.informatics.iupui. edu/HAPPI/

HomoMINT

An inferred human network based on orthology mapping of protein interactions discovered in model organisms

http://mint.bio.uniroma2. it/HomoMINT

IntAct

An open-source, open data molecular interaction database and toolkit

www.ebi.ac.uk/intact

999 human pathways and more than 59,341 human molecular entities. A set of analysis tools is also provided in HPD to allow searching, managing, and studying human biological pathways. NetPath is a component of Human Protein Reference Database (HPRD), which is a centralized platform to visually depict and integrate information pertaining to domain architecture, posttranslational modifications, interaction networks, and disease association for each protein in the human proteome (51). NetPath has 20 annotated immune and cancer signaling pathways involving 1,682 molecules and 1,800 interactions. The HAPPI database integrates protein interaction data from multiple public databases, including HPRD, BIND, MINT, STRING, and OPHID. A measure of reliability (rank levels from 1 to 5) has been given to each entry in the database. As of 2008,

Database Resources for Proteomics-Based Analysis of Cancer

361

the database contains 142,956 nonredundant, medium to high-confidence level human protein interaction pairs among 10,592 human proteins (52). HomoMINT extends PPIs experimentally verified in models organisms to the orthologous proteins in human (53). As of 2009, the database has 2,4439 interactions of 8,041 proteins. The curated data can be analyzed in the context of the high-throughput data and displayed in graphics by a tool named MINT Viewer. Data in IntAct are obtained from the literature or from direct data depositions by expert curators. As of September 2009, it contains over 200,000 curated binary interaction entries. IntAct provides a two-tiered view of the interaction data: a simplified and tabular view and a specialized view providing the full annotation of interactions, interactors and their properties (54). Detailed review and evaluation of public human PPI databases can be found in (55, 56).

3. Discussion Here we have reviewed and summarized five groups of databases related to proteomics studies in cancer research. We have included databases containing gene/protein expression data produced by microarray studies, next-generation sequencing, and other high-throughput experiments, gene mutation and SNP databases, tumor antigen databases, databases of cancer-associated genes, and protein interaction and pathway databases. For more cancerrelated databases, refer to the 2009 database special issue of Nucleic Acids Research (www.oxfordjournals.org/nar/database/ subcat/8/33) (57). The relevant databases can be found in subheading “Cancer gene databases” under heading “Human Genes and Diseases.” An increasing use of databases is data mining (58). These applications involve systems that combine data from multiple specialized databases and analytic tools enabling detailed analysis. For example, BiomarkerDigger (http://biomarkerdigger.org) performs data analysis, searching, and metadata-gathering (59). When gathering metadata, it searches proteome DBs for PPI, Gene Ontology annotations, protein domains, human genetic disorders, and tissue expression profile information. These diverse sources are integrated into protein data sets that are accessed through a search function in BiomarkerDigger. The identification of a serological biomarker for hepatocellular carcinoma by comparison of plasma and tissue proteomic data sets from healthy volunteers and cancer patients was demonstrated by using this resource (59). The vast amount of information generated from cancer research has presented both a challenge and an opportunity for

362

Zhang, DeLuca, and Brusic

researchers worldwide. Researchers have invested lots of effort and time in cleaning, annotating, and organizing the data produced from proteomics studies and put them into specialized biological databases. Information retrieval from proteomics databases is the starting point in performing downstream bioinformatics analyses. Making use of these data enhances understanding of the disease and advance anticancer treatments.

Acknowledgments We thank Dr. Catherine J. Wu for thoughtful reviews of this manuscript. References 1. James P (1997) Protein identification in the post-genome era: the rapid rise of proteomics. Q Rev Biophys 30:279–331 2. Koomen JM, Haura EB, Bepler G, Sutphen R, Remily-Wood ER, Benson K, Hussein M, Hazlehurst LA, Yeatman TJ, Hildreth LT, Sellers TA, Jacobsen PB, Fenstermacher DA, Dalton WS (2008) Proteomic contributions to personalized cancer care. Mol Cell Proteomics 7:1780–1794 3. Gygi SP, Rochon Y, Franza BR, Aebersold R (1999) Correlation between protein and mRNA abundance in yeast. Mol Cell Biol 19:1720–1730 4. Ivanov SS, Chung AS, Yuan ZL, Guan YJ, Sachs KV, Reichner JS, Chin YE (2004) Antibodies immobilized as arrays to profile protein post-translational modifications in mammalian cells. Mol Cell Proteomics 3:788–795 5. Bamford S, Dawson E, Forbes S, Clements J, Pettett R, Dogan A, Flanagan A, Teague J, Futreal PA, Stratton MR, Wooster R (2004) The COSMIC (catalogue of somatic mutations in cancer) database and website. Br J Cancer 91:355–358 6. Kopf E, Zharhary D (2007) Antibody arrays – an emerging tool in cancer proteomics. Int J Biochem Cell Biol 39:1305–1317 7. Tao SC, Chen CS, Zhu H (2007) Applications of protein microarray technology. Comb Chem High Throughput Screen 10:706–718 8. Sherman BT, da Huang W, Tan Q, Guo Y, Bour S, Liu D, Stephens R, Baseler MW, Lane HC, Lempicki RA (2007) DAVID knowledgebase: a gene-centered database integrating

9.

10.

11. 12.

13. 14.

15.

heterogeneous gene annotation resources to facilitate high-throughput gene functional analysis. BMC Bioinformatics 8:426 da Huang W, Sherman BT, Tan Q, Kir J, Liu D, Bryant D, Guo Y, Stephens R, Baseler MW, Lane HC, Lempicki RA (2007) DAVID Bioinformatics Resources: expanded annotation database and novel algorithms to better extract biology from large gene lists. Nucleic Acids Res 35:W169–W175 Rajcevic U, Niclou SP, Jimenez CR (2009) Proteomics strategies for target identification and biomarker discovery in cancer. Front Biosci 14:3292–3303 Grantzdorffer I, Carl-McGrath S, Ebert MP, Rocken C (2008) Proteomics of pancreatic cancer. Pancreas 36:329–336 Sahin U, Tureci O, Schmitt H, Cochlovius B, Johannes T, Schmits R, Stenner F, Luo G, Schobert I, Pfreundschuh M (1995) Human neoplasms elicit multiple specific immune responses in the autologous host. Proc Natl Acad Sci USA 92:11810–11813 Boutet E, Lieberherr D, Tognolli M, Schneider M, Bairoch A (2007) UniProtKB/Swiss-Prot. Methods Mol Biol 406:89–112 Rebhan M, Chalifa-Caspi V, Prilusky J, Lancet D (1998) GeneCards: a novel functional genomics compendium with automated data mining and query reformulation support. Bioinformatics 14:656–664 Barrett T, Troup DB, Wilhite SE, Ledoux P, Rudnev D, Evangelista C, Kim IF, Soboleva A, Tomashevsky M, Marshall KA, Phillippy KH, Sherman PM, Muertter RN, Edgar R (2009) NCBI GEO: archive for high-throughput

Database Resources for Proteomics-Based Analysis of Cancer

16.

17.

18.

19.

20. 21.

22.

23.

24. 25.

functional genomic data. Nucleic Acids Res 37:D885–D890 Ikeo K, Ishi-i J, Tamura T, Gojobori T, Tateno Y (2003) CIBEX: center for information biology gene expression database. C R Biol 326:1079–1082 Rocca-Serra P, Brazma A, Parkinson H, Sarkans U, Shojatalab M, Contrino S, Vilo J, Abeygunawardena N, Mukherjee G, Holloway E, Kapushesky M, Kemmeren P, Lara GG, Oezcimen A, Sansone SA (2003) ArrayExpress: a public database of gene expression data at EBI. C R Biol 326:1075–1078 Parkinson H, Kapushesky M, Kolesnikov N, Rustici G, Shojatalab M, Abeygunawardena N, Berube H, Dylag M, Emam I, Farne A, Holloway E, Lukk M, Malone J, Mani R, Pilicheva E, Rayner TF, Rezwan F, Sharma A, Williams E, Bradley XZ, Adamusiak T, Brandizi M, Burdett T, Coulson R, Krestyaninova M, Kurnosov P, Maguire E, Neogi SG, Rocca-Serra P, Sansone SA, Sklyar N, Zhao M, Sarkans U, Brazma A (2009) ArrayExpress update – from an archive of functional genomics experiments to the atlas of gene expression. Nucleic Acids Res 37:D868–D872 Kato K, Yamashita R, Matoba R, Monden M, Noguchi S, Takagi T, Nakai K (2005) Cancer gene expression database (CGED): a database for gene expression profiling with accompanying clinical information of human cancer tissues. Nucleic Acids Res 33:D533–D536 Penkett CJ, Bahler J (2004) Navigating public microarray databases. Comp Funct Genomics 5:471–479 Zeng ZY, Xiong W, Zhou YH, Li XL, Li GY (2006) [Advances in high-density whole genome-wide single nucleotide polymorphism array in cancer research]. Ai Zheng 25:1454–1458 Forbes SA, Bhamra G, Bamford S, Dawson E, Kok C, Clements J, Menzies A, Teague JW, Futreal PA, Stratton MR (2008) The catalogue of somatic mutations in cancer (COSMIC). Curr Protoc Hum Genet. Chapter 10, Unit 10 11 Stenson PD, Mort M, Ball EV, Howells K, Phillips AD, Thomas NS, Cooper DN (2009) The human gene mutation database: 2008 update. Genome Med 1:13 Shimizu N, Ohtsubo M, Minoshima S (2007) MutationView/KMcancerDB: a database for cancer gene mutations. Cancer Sci 98:259–267 Futreal PA, Coin L, Marshall M, Down T, Hubbard T, Wooster R, Rahman N, Stratton MR (2004) A census of human cancer genes. Nat Rev Cancer 4:177–183

363

26. Hernandez-Boussard T, Rodriguez-Tome P, Montesano R, Hainaut P (1999) IARC p53 mutation database: a relational database to compile and analyze p53 mutations in human tumors and cell lines. International Agency for Research on Cancer. Hum Mutat 14:1–8 27. Murphy JA, Barrantes-Reynolds R, Kocherlakota R, Bond JP, Greenblatt MS (2004) The CDKN2A database: integrating allelic variants with evolution, structure, function, and disease association. Hum Mutat 24:296–304 28. Gottlieb B, Trifiro M, Lumbroso R, Pinsky L (1997) The androgen receptor gene mutations database. Nucleic Acids Res 25:158–162 29. Gottlieb B, Beitel LK, Wu JH, Trifiro M (2004) The androgen receptor gene mutations database (ARDB): 2004 update. Hum Mutat 23:527–533 30. Jackson MA, Lea I, Rashid A, Peddada SD, Dunnick JK (2006) Genetic alterations in cancer knowledge system: analysis of gene mutations in mouse and human liver and lung tumors. Toxicol Sci 90:400–418 31. Hu Y, McDermott MP, Ahrendt SA (2005) The p53 codon 72 proline allele is associated with p53 gene mutations in non-small cell lung cancer. Clin Cancer Res 11:2502–2509 32. Packer BR, Yeager M, Burdett L, Welch R, Beerman M, Qi L, Sicotte H, Staats B, Acharya M, Crenshaw A, Eckert A, Puri V, Gerhard DS, Chanock SJ (2006) SNP500Cancer: a public resource for sequence validation, assay development, and frequency analysis for genetic variation in candidate genes. Nucleic Acids Res 34:D617–D621 33. Smigielski EM, Sirotkin K, Ward M, Sherry ST (2000) dbSNP: a database of single nucleotide polymorphisms. Nucleic Acids Res 28:352–355 34. Rhee H, Lee JS (2009) MedRefSNP: a database of medically investigated SNPs. Hum Mutat 30:E460–E466 35. van der Bruggen P, Traversari C, Chomez P, Lurquin C, De Plaen E, Van den Eynde B, Knuth A, Boon T (1991) A gene encoding an antigen recognized by cytolytic T lymphocytes on a human melanoma. Science 254:1643–1647 36. Scanlan MJ, Gure AO, Jungbluth AA, Old LJ, Chen YT (2002) Cancer/testis antigens: an expanding family of targets for cancer immunotherapy. Immunol Rev 188:22–32 37. Almeida LG, Sakabe NJ, deOliveira AR, Silva MC, Mundstein AS, Cohen T, Chen YT, Chua R, Gurung S, Gnjatic S, Jungbluth AA, Caballero OL, Bairoch A, Kiesler E, White SL, Simpson AJ, Old LJ, Camargo AA,

364

38.

39. 40.

41. 42.

43.

44.

45. 46.

47.

48.

49.

50.

Zhang, DeLuca, and Brusic Vasconcelos AT (2009) CTdatabase: a knowledge-base of high-throughput and curated data on cancer-testis antigens. Nucleic Acids Res 37:D816–D819 Donnes P, Hoglund A, Sturm M, Comtesse N, Backes C, Meese E, Kohlbacher O, Lenhof HP (2004) Integrative analysis of cancer-related data using CAP. FASEB J 18:1465–1467 Jongeneel V (2001) Towards a cancer immunome database. Cancer Immun 1:3 Novellino L, Castelli C, Parmiani G (2005) A listing of human tumor antigens recognized by T cells. Cancer Immunol Immunother 54:187–207 Van den Eynde BJ, van der Bruggen P (1997) T cell defined tumor antigens. Curr Opin Immunol 9:684–693 Strausberg RL, Buetow KH, Greenhut SF, Grouse LH, Schaefer CF (2002) The cancer genome anatomy project: online resources to reveal the molecular signatures of cancer. Cancer Investig 20:1038–1050 Schaefer C, Grouse L, Buetow K, Strausberg RL (2001) A new cancer genome anatomy project web resource for the community. Cancer J 7:52–60 Krupke D, Naf D, Vincent M, Allio T, Mikaelian I, Sundberg J, Bult C, Eppig J (2005) The Mouse Tumor Biology Database: integrated access to mouse cancer biology data. Exp Lung Res 31:259–270 Bader GD, Cary MP, Sander C (2006) Pathguide: a pathway resource list. Nucleic Acids Res 34:D504–D506 Bauer-Mehren A, Furlong LI, Sanz F (2009) Pathway databases and tools for their exploitation: benefits, current limitations and challenges. Mol Syst Biol 5:290 Kanehisa M, Goto S, Furumichi M, Tanabe M Hirakawa, M (2010) KEGG for representation and analysis of molecular networks involving diseases and drugs, Nucleic Acids Res 38:D355–D360 Matthews L, Gopinath G, Gillespie M, Caudy M, Croft D, de Bono B, Garapati P, Hemish J, Hermjakob H, Jassal B, Kanapin A, Lewis S, Mahajan S, May B, Schmidt E, Vastrik I, Wu G, Birney E, Stein L, D’Eustachio P (2009) Reactome knowledgebase of human biological pathways and processes. Nucleic Acids Res 37:D619–D622 Schaefer CF, Anthony K, Krupa S, Buchoff J, Day M, Hannay T, Buetow KH (2009) PID: the Pathway Interaction Database. Nucleic Acids Res 37:D674–D679 Chowbina SR, Wu X, Zhang F, Li PM, Pandey R, Kasamsetty HN, Chen JY (2009) HPD: an

51.

52.

53.

54.

55. 56.

57.

58. 59.

online integrated human pathway database enabling systems biology studies. BMC Bioinformatics 10(Suppl 11):S5 Keshava Prasad TS, Goel R, Kandasamy K, Keerthikumar S, Kumar S, Mathivanan S, Telikicherla D, Raju R, Shafreen B, Venugopal A, Balakrishnan L, Marimuthu A, Banerjee S, Somanathan DS, Sebastian A, Rani S, Ray S, Harrys Kishore CJ, Kanth S, Ahmed M, Kashyap MK, Mohmood R, Ramachandra YL, Krishna V, Rahiman BA, Mohan S, Ranganathan P, Ramabadran S, Chaerkady R, Pandey A (2009) Human Protein Reference Database – 2009 update. Nucleic Acids Res 37:D767–D772 Chen JY, Mamidipalli S, Huan T (2009) HAPPI: an online database of comprehensive human annotated and predicted protein interactions. BMC Genomics 10(Suppl 1):S16 Persico M, Ceol A, Gavrila C, Hoffmann R, Florio A, Cesareni G (2005) HomoMINT: an inferred human network based on orthology mapping of protein interactions discovered in model organisms. BMC Bioinformatics 6(Suppl 4):S21 Aranda B, Achuthan P, Alam-Faruque Y, Armean I, Bridge A, Derow C, Feuermann M, Ghanbarian AT, Kerrien S, Khadake J, Kerssemakers J, Leroy C, Menden M, Michaut M, Montecchi-Palazzi L, Neuhauser SN, Orchard S, Perreau V, Roechert B, van Eijk K, Hermjakob H (2009) The IntAct molecular interaction database in 2010. Nucleic Acids Res 38(Database issue):D525–31 Lehne B, Schlitt T (2009) Protein-protein interaction databases: keeping up with growing interactomes. Hum Genomics 3:291–297 Mathivanan S, Periaswamy B, Gandhi TK, Kandasamy K, Suresh S, Mohmood R, Ramachandra YL, Pandey A (2006) An evaluation of human protein–protein interaction data in the public domain. BMC Bioinformatics 7(Suppl 5):S19 Galperin MY, Cochrane GR (2009) Nucleic acids research annual database issue and the NAR online molecular biology database collection in 2009. Nucleic Acids Res 37:D1–D4 Brusic V, Zeleznikow J (1999) Knowledge discovery and data mining in biological databases. Knowledge Eng Rev 14:257–277 Jeong SK, Kwon MS, Lee EY, Lee HJ, Cho SY, Kim H, Yoo JS, Omenn GS, Aebersold R, Hanash S, Paik YK (2009) BiomarkerDigger: a versatile disease proteome database and analysis platform for the identification of plasma cancer biomarkers. Proteomics 9:3729–3740

Index A Absorption/covalent printing........................................... 86 Accutase.............................................................171, 175, 178 Acetone......................................................................171, 173 Acetylation..............................................214–217, 220–222 Acuity 4.0..........................................................144, 146, 147 Adherent cells............................ 53, 167, 171, 175, 178, 182 Affinity......................17, 29, 57, 58, 71, 83, 86, 98, 108, 109, 116, 120, 122, 146, 149, 155–157, 159, 186, 190, 213, 250, 259, 276, 297, 322, 351 printing................................................................. 83, 86 reagent................................. 57, 149, 186, 250, 276, 322 Allo-antibodies..................................................... 82, 87–88 Amino acid activation...................................................112, 123, 124 coupling.............................................110, 113, 123, 124 pentafluorophenyl esters................................... 107, 112 Ampicillin............. 91, 95, 102, 187, 258, 260, 261, 265, 270 Analytical methods............................ 38, 316–317, 338, 341 Androgen receptor (AR)................................168–170, 176, 177, 354, 355 Angiogenin (Ang).......................................................... 6, 7 Anti-beta galactosidase................................................... 179 Antibodies...............3–12, 15–27, 29–39, 42, 43, 47–52, 69, 81–102, 108–110, 118, 119, 121, 122, 130, 138, 140–141, 143, 145, 146, 149, 150, 153, 155–157, 159, 179, 186, 189–190, 194–199, 202, 217, 219, 220, 222, 227–242, 245, 250, 275–277, 279–285, 293, 295–297, 299, 307, 308, 311, 312, 317, 321–327, 329–332, 338, 340–342, 344–345, 350, 357 autoantibody.......130, 131, 142–143, 150, 151, 155–156 profiling.............................................130, 155, 157–158 screening....................................................155, 240, 245 validation....................................... 51–52, 240, 250, 276 Antibody-based detection...........4–6, 10, 11, 16, 30, 32, 33, 81–102, 118, 119, 186, 202, 222, 227–238 Antigens......................4–5, 9, 11, 12, 16, 20, 23, 24, 81–102, 129–147, 149–153, 155, 156, 158–160, 181, 186, 191, 196, 227, 228, 230, 234–237, 241, 243, 244, 281, 337, 350–352, 356–358, 361 isoform............................................................. 101, 359 Anti human IgG Alexa647 conjugate........................ 91, 93

Anti-phosphoprotein antibodies................................ 43, 51 Aptamer characterization.................................................... 57–59 high-throughput optimization.............................. 57–65 length minimization................................................... 63 microarray............................................................. 57–65 AR. See Androgen receptor Array format........................................................17, 75, 166 Autoantibodies..................130, 131, 142–143, 149–151, 153, 155–156, 159, 351 immunoglobulin G labeling of........................... 133, 138–140, 143, 146, 189, 195, 198, 229–230, 235, 238 purification of..................... 132–133, 137–138, 144, 145, 243, 245, 324 profiling.............................................130, 142–143, 155 Auto-fluorescence.................... 123, 171, 172, 222, 277, 297 Autoimmune disease.............................................. 130, 227 Autoimmunity.........................................130, 131, 227, 239 Automation...............................98, 106, 114, 142, 147, 171, 282–283, 286, 288, 291, 296, 342, 344 AxioVision LE....................................................... 172, 178

B Bacterial adhesins............................................................. 72 Bacterial invasion.............................................................. 44 Bait...................130, 166–170, 172–174, 177–180, 182–183, 240, 250, 275, 276 BCL2.............................................................................. 150 Bead array...................................................... 29–36, 227–238 BIOCCD image reader...........................167, 169, 171, 178 Bioinformatics.................................... 85, 87, 144, 146, 346, 350–352, 362 Biological pathway..........................................239, 358–360 Bioluminescence............................................................. 304 Biomarker................................ 8, 15–27, 129–147, 152, 158, 228, 311, 337, 349–351, 361 Biomolecular interaction.........................202, 304, 311, 316 Biotin.............................................. 16, 31, 34, 36, 116, 146, 186, 229, 235, 241–244, 250–252, 277, 281, 282, 314, 315 Biotinylation................................6, 10, 11, 18, 22, 186, 236, 277, 279, 282, 313, 314

Catherine J. Wu (ed.), Protein Microarray for Disease Analysis: Methods and Protocols, Methods in Molecular Biology, vol. 723, DOI 10.1007/978-1-61779-043-0, © Springer Science+Business Media, LLC 2011

365

Protein Microarray for Disease Analysis 366 Index

Bovine serum albumin (BSA)............... 8, 12, 17, 41, 43–46, 49, 59, 61, 69, 70, 75, 88, 90, 93, 122, 133, 140, 145, 204, 216, 219, 222, 229, 234, 235, 266, 279, 283, 284, 313–315, 332 Buffer blocking................................12, 17, 19, 41, 47, 48, 52, 69, 70, 74, 75, 93, 99, 108, 109, 118, 119, 122, 132, 133, 139, 193, 196, 204, 205, 217, 228, 244, 248, 252 crossdown......................................................... 109, 119 incubation.....................75, 109, 120, 121, 133, 140, 145 regeneration.......................................109, 120, 323, 328 spotting.......................................... 41, 47, 108, 118, 323 transfer..............................10, 11, 18, 118, 140, 244, 248 tris-buffered saline (TBS)...........................91, 122, 204, 217, 244, 252 tween-tris buffered saline (T-TBS)...................108, 122, 220, 244, 248, 279, 282, 283, 297

C Cancer breast................................................... 150, 158, 353–355 melanoma..................................................150, 355, 356 prostate......................................................130, 150, 355 vaccine target............................................................ 358 Candida........................................................................... 40 CAPPIA. See Cell array protein–protein interaction assay Capping........................................... 30, 107, 113, 123, 124 Carbohydrates.......................................................15, 67, 71 Catalyzed signal amplification (CSA).................... 277, 281 CDA. See Concentration dependent analysis cDNAs............... 94, 150, 154–156, 230, 258, 262, 270, 357 Cell array protein–protein interaction assay (CAPPIA)................................... 165–183 Cell arrays............................................................... 165–183 Cell culture................................9, 38, 42–43, 52, 87, 91–93, 95–98, 108, 118, 120, 131, 134–135, 145, 168, 171, 176, 178, 181, 260, 262, 284, 323 293 Cell line.......................94, 169, 171, 175, 176, 181, 182 Cell lysate arrays..............................................130, 143, 322 Cell monolayer........................................166, 167, 181, 182 Cell signaling...........................37, 38, 40, 44, 155, 228, 229, 244, 279, 280, 284 Cellulose membrane b-alanine................................................................... 107 amino-alkyl linked membranes.................107, 110–112, 117, 123–124 amino functionalization.............................110, 123–124 CAPE membrane............................................. 110, 111 cleavage......................................................108, 115–118 esterified membranes.........................110–111, 123–124 preparation.........................................107–108, 110–118 TFA-soluble membrane........................................... 109 TOTD membrane............................................ 110, 111

Charge-coupled device (CCD) camera..............50, 53, 178, 284, 285, 299, 311, 314, 316 Chemiluminescence....................... 109, 119, 122, 125, 229, 232, 244, 249, 252, 304, 321 Chemokine................................................................... 5, 11 Chip blocking................................................41, 47–48, 219 Chlamydia................................................................................40 Chronic lymphocytic leukemia............................... 241, 242 CLAMP kinetic analysis software.................................. 330 Cloning....................................68, 71, 72, 76, 170, 172, 179, 187, 190–192, 231, 250, 251, 257, 258, 259, 261–263, 266, 270 Colorimetric assay.......................................................... 285 Competitive assay............................................311, 314–317 Concentration dependent analysis (CDA)................ 85–86, 338–343, 345, 346 Conjugation................................................ 8, 10, 18, 24, 41, 43, 46, 59, 69, 82, 91, 93, 110, 124, 155, 157, 179, 188, 199, 214, 230, 232, 244, 281, 285, 308, 309, 312, 313 Contact printing............................................................. 181 Contact protein printer..................................................... 93 Co-precipitation..................................................... 239–254 COS7.................................................................171, 175, 176 Coupling cycle................................. 105, 113–115, 117, 123 Cre. See Cre/loxp Cre/loxp..................................................................258, 269 CSA. See Catalyzed signal amplification Cy3......................................8, 35, 59, 70, 134, 141, 157, 158, 188, 189, 193, 195–199, 222, 277, 310–312 Cy5....................................41, 42, 70, 72, 110, 121, 134, 141, 146, 190, 195–199, 222, 277, 310, 312, 313 Cytokine................................................................... 4–7, 10, 11

D DAPI. See 4’,6-Diamidino–2-phenylindole dihydrochloride Dark quencher........................................................ 309, 314 Data analysis........................... 10–11, 22–23, 39, 42, 49–50, 62–63, 77, 93, 100–101, 134, 137, 144, 147, 153, 204, 207–209, 219, 279, 281, 288–296, 298, 324, 330–331, 337–346, 350, 351 Database...............................................................62, 349–362 DBD. See DNA binding domain Denatured nickel affinity chromatography............83, 92, 97 Deprotection................................... 108, 111, 115–117, 124 Detection alkaline phosphatase (AP)................................ 109, 118 antibody.......................3–13, 16–23, 30, 33, 43, 81–102, 118, 119, 195, 202, 222, 227–238, 277, 308, 322, 350 chemiluminescence...........................................109, 119, 122, 125, 232 fluorescence.......................... 17–23, 109–110, 120–121, 171–172, 179, 220, 277, 285, 309, 311, 313

Protein Microarray for Disease Analysis 367 Index

horse-radish peroxidase (HRP, POD)................... 5, 69, 75, 77, 91, 96, 98, 109, 118, 119, 155, 157, 188, 228–230, 232, 244, 248, 251, 252, 276–277, 281, 283, 285 methods.........................5, 108–110, 118–121, 146, 276, 279–281, 285, 297, 313 staining...................................... 109, 119, 168, 284, 322 X-ray film......................................................... 119, 216 DGC. See Drosophila gene collection Diabetes...............................................................6, 150, 158 Diagnosis.....................................................8, 311, 350, 358 4’,6-Diamidino–2-phenylindole dihydrochloride (DAPI)..................................171, 177, 178, 182 Diastolic blood pressure...................................................... 7 Differential expression analysis............................... 340–342 DLI. See Donor lymphocyte infusion D-MEM. See Dulbecco’s Modified Eagle’s Medium DNA binding domain (DBD)......... 166, 170, 172, 175, 179 DNA method..................................................172, 173, 186 DNA microarray reader.......................................... 171, 178 DNA microarrays......................... 57–65, 81, 154–156, 166, 167, 186, 198, 213–214, 217, 276, 311, 312, 321, 322, 338 DNA polymerase................................... 68, 73, 76, 260, 270 Donor-acceptor pairs.............................................. 308–310 Donor lymphocyte infusion (DLI)..................241, 242, 244 Dose-dependence............................................168, 176, 177 Drosophila gene collection (DGC)........................ 258, 263 Dulbecco’s Modified Eagle’s Medium (D-MEM).............................171, 175, 323, 325

E EBNA. See Epstein–Barr virus nuclear antigen EC-buffer............................................................... 171, 173 ECL. See Enhanced chemiluminiscence E.coli clones.................................................................... 259 Effectene transfection reagent................166, 167, 170–173, 180, 182 EGFP. See Enhanced green fluorescent protein ELISA. See Enzyme-linked immunosorbent assay Endo-Free Plasmid Maxi Kit................................. 169, 172 End-stage renal disease (ESRD).................................... 5–6 Energy transfer....................................................... 303–317 Enhanced chemiluminiscence (ECL)......................96, 109, 119, 122, 125, 229, 232, 244, 249, 321 Enhanced green fluorescent protein (EGFP)................ 167, 171–175, 178, 179, 310 Enteropathogenic Escherichia coli (EPEC)....................... 40 Enzyme-linked immunosorbent assay (ELISA)............ 4, 5, 69–71, 74–75, 77, 82, 84, 85, 88, 151–152, 155, 228–237 EPEC. See Enteropathogenic Escherichia coli Epithelial cells........................................................... 38–40, 52, 130

Epitope tags GST...................................................................... 83, 84 V5........................................................ 83, 88, 89, 94, 95 6xHis................................... 83, 86, 88, 94–95, 269–270 Epoxide coating...................................................... 108, 122 Epstein–Barr virus nuclear antigen (EBNA)..........153, 158, 236, 237, 242 ESRD. See End-stage renal disease Excited state........................................................... 304–306 Expression clone collection............................................ 259 Expression profiling...........................................85, 330, 350 Expression-ready clone collection.......................... 257–271 Expression system coupled transcription-translation in vitro................. 156

F FAST slides...............................................93, 188, 194, 217 FBS. See Fetal bovine serum Fetal bovine serum (FBS).............. 9, 12, 131, 171, 175, 323 Filter paper.......107, 108, 110–112, 114, 118, 120–122, 177 FLAG.............. 186, 188, 191–192, 198, 217, 220, 229, 230, 231, 233, 237, 250, 291 FLAG-HA..................................................................... 222 Flow cell .................................. 136, 322, 324, 327–329, 332 Fluidic cells................................................................. 42, 48 Fluorescence...................... 4, 6, 7, 10, 17–23, 30, 35, 42, 46, 49–50, 53, 61–63, 65, 77, 87, 89, 101, 109–110, 120–121, 123, 141, 147, 169, 171–172, 174, 175, 178–182, 197, 220, 277, 278, 285, 297, 304–307, 309, 311, 321, 322, 326, 344–345 lifetime......................................................312, 314–316 Fluorescent label.................................. 48, 59, 134, 199, 313 Fluorochrome............................................................. 82, 88 Fluorometric assay.................................................. 275–299 Fluoromount-G...................................................... 171, 178 Fmoc building block............................................112, 116, 117 chemistry.................................................................. 216 removal of protecting-group............................. 113, 114 Formaldehyde......................................................... 171, 176 Förster distance..................................................... 306–307, 309, 310 Förster resonance energy transfer........................... 303–317 FPLC chromatographic system............................ 92, 94–95 Free peptides..................................................106, 110, 115, 117–118, 124 Fusion proteins (amino and carboxy terminal).............. 258, 259, 269

G GAD65......................................................................150, 158 GAL4..............................................................172, 175, 179 GAL (file format). See Gene array list Gal4-pZsGreen...............................................172, 175, 179

Protein Microarray for Disease Analysis 368 Index

GAPS II coated slides............................................ 180, 181 Gateway system/technology................... 179, 187, 190, 231, 250, 257, 259, 270 Gelatin..................................... 170, 172, 173, 178, 180, 332 Gel electrophoresis.................. 244, 259–261, 264, 278, 351 Gene array list (GAL).................... 142, 207, 209, 221, 222, 261, 327, 330, 339, 341, 342 Gene expression..............3, 81, 278, 310, 351–353, 357–359 profile....................................................3, 352, 353, 358 GenePix.................................70, 75, 93, 101, 102, 141, 142, 189, 190, 193, 195, 197, 204, 207–210, 216, 339, 341–344 GenePix Pro 6.0........................................22, 146, 221, 222 GenePix results (GPR)........................... 101, 142, 144, 147, 209, 210, 342, 344 Gentamycin protection assay............................................ 53 Glass slides........................16, 71, 86, 88, 93, 108, 115, 118, 121, 150, 159, 166, 167, 173, 174, 176, 178, 180, 181, 201–202, 240, 290, 297, 315, 324, 327 aldehyde surface.........................................108, 122, 124 epoxide coating......................................................... 108 Glutathione-S-transferase (GST)................ 71, 83, 84, 155, 157, 186, 217, 219, 228, 230–237, 344–345 Glycan......................... 15–16, 19, 22–24, 27, 67, 68, 71, 322 Glycerol...........8, 26, 69, 91, 92, 98, 101, 108, 121, 131, 132, 137, 215, 217, 218, 244, 267, 312, 323 Glycomics................................................................... 67–77 Glycoprofiling................................................................ 350 Glycoprotein........................................ 15–17, 19, 20, 71, 74 Glycosylation..................5, 15–27, 68, 72, 82, 240, 244–245 GPR (file format). See GenePix results Graft versus host disease (GVHD)................................... 87 Graft versus leukemia (GVL)........................................... 87 GST. See Glutathione-S-transferase GTPases.........................................................................38, 40 GVHD. See Graft versus host disease GVL. See Graft versus leukemia

H HEK 293.................................................171, 175, 176, 181 HEK 293 T............................. 169, 171, 175, 176, 181, 182 HeLa cells 38, 40, 42–44, 51, 52, 171, 175, 176 HepG2.......................................................171, 175, 176, 325 HepG2 cell line...................................................... 323, 324 High-throughput.............................. 3–12, 37, 67, 129, 201, 202, 204, 228, 231, 276, 321, 337, 350–353, 361 High-throughput screening (HTS)........................... 15–27, 165–183, 229, 240, 244, 308, 312 HIV-p24................................................................88, 94, 97 HLA ligand.................................................................... 358 Hormone-dependence............................................ 169, 172 Host cell signalling................................................38, 40, 44 Host-pathogen interaction................................... 37–53, 68

HTS. See High-throughput screening Human..............................16, 24, 29, 40, 59, 81–83, 88, 130, 131, 134, 145, 191, 201, 202, 215, 259, 278, 349, 351, 354, 355– 361 Human blood........................16, 24, 99–100, 150, 230, 232, 235, 245 Human IgG.................................. 84, 88, 91, 145, 153, 230, 235, 238 H-Y antigens/H-Y proteins DDX3X.......................................................... 87–88, 94 DDX3Y.......................................................... 87–90, 94 EIF1AX.....................................................87–88, 94, 97 EIF1AY.....................................................87–88, 94, 97 RPS4X............................................................ 87–88, 94 RPS4Y............................................................ 87–88, 94 UTX............................................................... 87–88, 94 UTY...................................................................87–90, 94 ZFX................................................................ 87–88, 94 ZFY....................................................................87–90, 94 Hybridization..........................17, 20, 24, 25, 38, 39, 59, 65, 70, 146, 155, 157, 159, 166, 168, 175, 188, 189, 193–197, 308, 311, 312, 314, 353 Hydroxyflutamide (OH-Flu)................................. 176, 177

I IA2, 150 IC/PBS. See Interstitial cystitis/painful bladder syndrome IgG purification......................................132–133, 137–138, 145, 243, 245 IL–12. See Interleukin–12 Image acquisition..................................... 50, 209, 279, 280, 285–288, 298, 315 ImageJ......................................................242, 249–250, 252 Imidazole..................................... 86, 88, 91–93, 95–98, 107 Immunoassay.................................... 5, 30, 50, 51, 202, 275, 308, 316–317, 322 Immunoblotting........................................38, 249, 250, 278 Immunoglobulin E........................................................... 59 Immunoprecipitation................................82, 240, 241–243, 245–247, 249–250 Immunoprofile....................................................... 149–159 Immunostaining......................................279–284, 312, 357 Individualized therapy.................................................... 278 In situ activation..............................................112, 123, 156 InstrumentONE high-performance....................... 171, 174 Interleukin–12 (IL–12).................................................. 6–7 Interstitial cystitis/painful bladder syndrome (IC/PBS)...............................130, 131, 143, 144 Intracellular parasites.................................................. 37–38 Invasive bacteria................................................... 37–38, 40 In vitro infection..............................................38–44, 52–53 IPTG. See Isopropyl-b-D-thiogalactoside IRB approval.......................................................... 132, 151 Isopropyl-b-D-thiogalactoside (IPTG)................91, 95, 96

K Kinase...................................38, 40, 199, 214, 308, 313, 355 Kinase-substrate interaction............................201–212, 337 Kinetics....................................................115, 243, 321, 322

L Label-free.............................................................24, 321–332 Lab-Tek™ Chamber Slide™, 181 LacZ. See b-galactosidase Lanthanide..............................................309, 310, 314, 317 Large T antigen of SV40................................................ 181 LB-Ampicillin (LB-Amp) plates..................................... 95 LB-Amp plates. See LB-Ampicillin (LB-Amp) plates LBD. See Ligand binding domain Lectin........................................................ 15–27, 67–77, 322 microarray..................................................15–27, 67–77 l-glutamine................................................40, 171, 175, 323 Library............................................... 168, 178, 270, 283, 312 Ligand binding domain (LBD)....... 168, 169, 170, 176, 177 Linker biotin............................................................................116 coupling.....................................................116–117, 122 hydrazinobenzoic acid (HBA).......................... 116, 117 Lipid bilayer array.......................................................... 312 Lipid-DNA method............................................... 172–173 Listeria................................................................................40 Loxp. See Cre/loxp Luminex............................ 5, 30, 35, 228–230, 232–235, 237 Lysates............. 29, 37–41, 43–45, 47, 49, 51, 74, 76, 82, 95, 98, 130–137, 140, 142–145, 149, 156, 190, 192, 194, 228, 230, 231, 237, 242, 243, 245–247, 251, 278, 279, 284, 289, 322, 332, 350 Lysis buffer..............................41, 44, 52, 53, 69, 74, 76, 92, 96–98, 102, 131, 134, 345 Lysozyme...............................................................69, 74, 97

M Macroarrays.....................................106, 107–108, 110–115 MALDI. See Matrix-assisted laser desorption/ionization Mammalian cells........................ 52, 83, 166–168, 175, 179, 180, 222, 239, 241, 242 Mammalian protein-protein interaction trap (MAPPIT)................................................... 179 Mammalian two hybrid.......................................... 165–183 MAPK. See Mitogen-activated protein kinase MAPPIT. See Mammalian protein-protein interaction trap Mass spectrometry (MS).............................. 16, 17, 24, 145, 147, 213, 351 Mastoparan.................................................................... 116 Matrix-assisted laser desorption/ionization (MALDI)................................................. 24–26 Mean fluorescence intensity (MFI)......................10, 32, 84, 87, 101, 102, 236

Protein Microarray for Disease Analysis 369 Index Medroxyprogesterone acetate (MPA)..................... 176, 177 Melanoma inhibitor of apoptosis (ML-IAP)................. 150 Membrane b-alanine.................................... 107, 110, 113, 121, 123 amino-alkyl linked............. 107, 110–112, 117, 123–124 amino functionalization.................................... 110, 121 CAPE............................................................... 110, 111 cleavage......................................108, 115–118, 121–122 esterification..................................................... 110, 121 PEG.....................................................................115, 124 TFA-soluble..............................................108, 109, 124 TOTD.............................................................. 110, 111 Methallothionein inducible promoter............................ 257 MFI. See Mean fluorescence intensity Mfold.....................................................................62, 63, 65 mHA. See Minor histocompatibility antigens Microarrays.............................4, 16, 29, 41, 58, 67, 81, 106, 129, 149, 166, 185, 201, 213, 227, 239, 275, 310, 321, 337, 350 analysis..........................10, 15–27, 30, 67–77, 106, 115, 134, 144, 147, 155, 199, 276, 280, 288–295, 337–346 forward phase microarrays.....................29–30, 129–130 printer 4.................................................. 70, 93, 98, 118, 132, 216, 218, 324 printing..............................9, 12, 17, 86–87, 93, 98, 132, 137, 150, 152, 153, 188–189, 192, 193, 197, 216–218, 277, 326, 344 printing and blocking of............................137, 188, 193 quality control....................................188–189, 193, 322 reverse phase microarrays.............. 29–30, 129–130, 227, 275–299 scanner...................................8, 10, 12, 62, 93, 100, 121, 134, 141, 216, 316 Micro flow system............................................................ 48 Microtiter plates (MTP)............................6–12, 30, 32–34, 41, 45, 48, 69, 118 Minimal binding domain................................................. 63 Minor histocompatibility antigens (mHA)...................... 87 Mitogen-activated protein kinase (MAPK)..................... 40 pathways..................................................................... 40 ML-IAP. See Melanoma inhibitor of apoptosis MOI. See Multiplicity of infection Molecular interaction analysis................................ 359, 360 Monolayer........................................ 166, 167, 178, 181, 182 MPA. See Medroxyprogesterone acetate MS. See Mass spectrometry MTP. See Microtiter plates Multiplex........................... 4–6, 18, 22, 29–30, 58, 81, 82, 86, 88, 129, 201, 227–238, 276, 280, 287, 312, 316 Multiplexed assay..................................................... 35, 228 Multiplicity of infection (MOI)........................... 44, 52–53 Multivariate approach........................................................ 3 Mycobacterium........................................................40, 42–43

Protein Microarray for Disease Analysis 370 Index

N Nanoparticle............................................309, 312–313, 317 NAPPA. See Nucleic acid programmable protein array Native nickel affinity chromatography............83, 92, 97, 98 Nebulization............................................................... 41, 48 NF-kB.................................................................... 172, 175 Non-contact microarray spotter........................................ 41 Non-contact piezo-dispensing system............................ 171 Nonradiative decay......................................................... 305 N-terminal domain (NTD).............................168–170, 176 Nucleic acid programmable protein array (NAPPA).............................................. 149–159

O Open reading frames (ORFs).............................83, 94, 101, 172, 191, 192, 194, 239–240, 250, 257–259, 262, 263, 269, 270 Organic compounds............................................... 105–106

P p53.......................................................... 150, 158, 175, 191, 195, 197, 237, 355 pAD-SV40T...................................................169, 172, 175 pAD-TRAF....................................................169, 172, 175 Pathogen..........................................37–53, 67–68, 310–312 Pathogenesis........................................................... 239, 359 pBD-NF-kB...................................................169, 172, 175 pBD-p53............................................................169, 172, 175 PBXL–3...................................................................5, 6, 8, 10 PC–3................................................................. 171, 175–176 pcDNA4-EGFP.............................................172–174, 178 pcDNA4/HisMax TOPO.............................................. 172 pCMV-AD............................................................ 170, 172 pCMV-BD..................................................................... 172 PCR. See Polymerase chain reaction PCR primer.............................................259, 262, 269, 270 Pellet................ 32, 34, 53, 64, 74, 76, 95–97, 102, 134, 145, 191, 192, 231, 235, 247, 268 Penicillin/streptomycin...................... 40, 131, 171, 175, 323 Peptide array macroarray.................................................. 106, 114 microarray...................................106, 109–110, 115, 118, 120–121 free peptide........................ 106, 110, 115, 117–118, 124 immobilization................................................. 116, 124 reconstitution............................................................ 118 solution......................................................115, 120, 124 synthesis........................................................... 105–125 transfer...................................................................... 118 unprotected peptides................................................. 116 pGAL/lacZ.................................................................... 179

Phosphoprotein/protein ratio........................................... 51 Phosphorylation................................. 38–40, 43, 51, 52, 82, 214, 240, 278, 312, 313 status......................................................................... 312 Photo multiplier tubes (PMT)................... 12, 22, 100, 141, 142, 146–147, 197, 286–287, 298, 316, 343 pIRES2-EGFP.............................................................. 172 Planar waveguide excitation....................................... 50, 53 Plasma.................................30, 32, 34, 35, 59, 81–102, 108, 157, 159, 228, 241, 243, 361 Plasmid..............................40, 68, 73, 76, 83, 101, 150, 155, 166, 167, 169–170, 172–175, 179–181, 186, 187, 189–195, 198, 228, 229, 231, 241, 243, 246, 250–251, 265 PMT. See Photo multiplier tubes Poly-l-lysine............................................171, 173, 180, 181 Polymerase chain reaction (PCR)...................34, 68, 71–73, 76, 85, 94, 168, 172, 186, 187, 190, 192, 245, 246, 250, 251, 258–267, 269, 270, 276, 309 Post-translational modifications glycosylation....................................15–27, 82, 244–245 phosphorylations..........38–40, 43, 51, 82, 214, 240, 278 Pre-activated derivatives......................................... 107, 112 Prey...........................166–170, 172–174, 177–180, 182, 183 Prey-reporter- (PR-) slides......................168, 169, 182–183 Primer design......................................................... 260, 263 Printing buffer....................8, 9, 11, 17, 18, 88, 90, 197, 215 Printing substrates glycosylation......................................................... 16, 72 nitrocellulose....................................................... 86, 278 Probing solution............................................................. 108 ProCAT.................................................................. 210, 211 Prognosis............................................................................ 8 Prospector................................................207, 341–342, 344 Protein.....................................3, 15, 29, 38, 57, 67, 81, 106, 129, 149, 165, 185, 201, 213, 227, 239, 257, 275, 304, 321, 337, 349 arrays................................37–53, 94, 140, 150, 166, 207, 227, 241, 311, 344, 350 binding............................64, 67, 86, 115, 119, 120, 132, 156, 179, 201, 214, 218, 235, 245, 321, 322 detection........................................................... 186, 276 expression................................20, 53, 69, 74, 76, 82, 83, 91–93, 95–98, 171, 192, 193, 195, 198, 228, 231, 232, 237, 241–242, 250, 257–259, 270, 311, 322, 337, 340, 350–352, 357, 361 gel quantification........................................ 44, 303–317 kinase................................... 40, 199, 202, 209, 308, 337 kinase B...................................................................... 40 labeled protein..............61, 109, 119, 122, 251, 252, 307 microarray..............29, 30, 82–86, 94, 98–100, 152–155, 185–199, 201–203, 208, 213–222, 215, 239, 241, 275–299, 311, 337–346, 350

Protein Microarray for Disease Analysis 371 Index

antigen........................................................ 129–147 blocking................................... 12, 99, 108, 205, 324 DNA..........................57–65, 81, 166, 167, 171, 178, 186, 190, 191, 193, 194, 197, 213–214, 217, 276, 311, 312, 338 nucleic acid................................................. 150, 155 printing......................................86–87, 93, 216–218 replicates....................................................6, 10, 152 reproducibility........................................47, 156, 346 scanning....................... 134, 141–144, 146, 207, 311 zone variation......................................152–153, 155 phosphorylation................... 38, 202, 207, 208, 278, 284 profiling............ 30, 35, 53, 130, 185, 214, 322, 323, 350 protein interaction (PPI)..................105–125, 165–183, 185–199, 215, 240, 244, 276, 303–317, 312, 317, 337, 352, 358, 359, 361 slide............................................................216, 293–295 solution........................................ 95, 108, 109, 297, 332 synthesis................................... 156, 186, 192, 195, 198, 241, 243, 245–246 translation................................................................. 231 Protein kinase B (Akt)...........................................40, 51, 52 Proteomics...........................29, 37, 129, 227, 228, 257–271, 278, 279, 349–362 ProtMAT.........................................................341–343, 345 Protoarray™........... 83, 84, 85, 202, 207, 338, 341, 342, 344 PR-stable-bait assay....................................................... 182 PR-trans-bait assay......................................................... 182

Q QuadriPERM®..........................93, 108, 118, 120, 145, 171, 176, 181, 182, 189, 205 Quantum yield....................................................... 306, 307

R R1881...................................................... 168, 169, 176, 177 Rabbit reticulocyte lysate........................156, 190, 192–194, 228, 230, 237, 243 Radiolabel........................................ 202, 204, 208, 209, 220 Recombinant antigen arrays............................. 79–102, 130 Recombinant lectins................................................... 67–77 Recombinant protein expression...........................82, 91–93, 95–98, 228, 230 Recombination................................................179, 257, 259 Referenced fluorescence intensity (RFI)..................... 49–51 Referencing...........................21, 41, 43, 46, 47, 49, 77, 153, 309, 316, 330, 346 Regeneration............109, 120, 125, 322, 323, 328, 330, 331 Regions of interest (ROI)................................291, 329–331 Reporter........ 30, 35, 166–169, 172–176, 179, 180, 182, 277 Reverse-phase protein arrays.......................................... 227 Reverse phase protein microarray........................... 275–299 Reverse protein arrays (RPA)......................37–53, 129–130

Reverse transfection........................................169, 171, 172, 175–176, 178–181 RFI. See Referenced fluorescence intensity ROI. See Regions of interest RPA. See Reverse protein arrays

S Salmonella............................................. 38, 40, 42–43, 51–53 Salmonella-containing vacuole (SCV).............................. 40 SAM. See Significance analysis of microarrays Sandwich style immunoassay.............................................. 5 ScanArray Express.......................................................... 146 SciFlexArrayer........................................................ 171, 174 sciFlexArrayer piezo–dispensing system S5................... 174 Screening.......................................................................15–27 SCV. See Salmonella-containing vacuole SDS-PAGE. See Sodium dodecyl sulfate polyacrylamide gel electrophoresis Selection marker ampicillin.......................................................... 258, 270 carbenicillin...................................................... 260, 261 chloramphenicol..........................................40, 187, 260 SELEX. See Systematic evolution of ligands by exponential enrichment Sensorgram......................................................330, 331, 332 Sequencing..........................18, 58, 60, 62–65, 82, 179, 186, 190–192, 239, 241, 250–252, 257–260, 262, 263, 267–270, 350–353, 356–359, 361 Serum............. 3–9, 15–27, 29–30, 34, 40, 81–102, 121, 130, 132, 134, 138, 144, 150, 154–159, 227, 228, 230, 232, 235, 240–243, 246, 279, 297, 322, 325, 332, 350, 357 screening............................ 153, 155, 159, 240, 244, 245 screening study..........................................151, 158–159 Shigella.............................................................................. 40 Signal corrections................................................... 178, 339 Signaling.................. 15, 37–38, 40, 42–44, 49, 51, 155, 214, 228, 229, 244, 278–280, 284, 337, 358–360 Significance analysis of microarrays (SAM)...................................338, 340, 342, 346 Single nucleotide polymorphisms (SNPs)......101, 310, 311, 351–356, 354, 356, 358, 359, 361 Single-stranded DNA (ssDNA)............................58, 59, 62 SNPs. See Single nucleotide polymorphisms Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE)......................... 74, 76, 77, 95, 96, 98, 244, 246, 247, 252 Software..................... 10, 26, 42, 60–62, 70, 75, 87, 93, 101, 102, 121, 134, 137, 142, 144, 146, 147, 172, 197, 204, 207, 209, 249, 260, 262, 280, 285, 286, 288–296, 324, 325, 327, 328, 330, 331, 338, 339, 341–342, 344, 346 Solid phase peptide synthesis..........................105, 112, 116

Protein Microarray for Disease Analysis 372 Index

Solvents dichloromethane (DCM)..................107, 108, 114, 115 dimethylformamide (DMF).............106, 107, 109–114, 123, 125 dimethylsulphoxide (DMSO).........................31, 69, 74, 107, 113, 178, 323, 325 ethanol (EtOH)............................ 76–77, 106, 107, 109, 111–113, 115, 119, 123, 173, 231, 262, 268, 270, 297 methanol (MeOH)................... 106, 107, 109, 111–113, 115, 119, 120, 123, 244, 248, 280 N-methylpyrrolidone (NMP)...................106, 112, 113, 115–117 SopB........................................................................... 40, 51 SopE........................................................................... 40, 51 SopE2......................................................................... 40, 51 SPOT synthesis cellulose membranes..........................105–108, 110–115 coupling solutions............................................. 105, 112 SPOT macroarray..............................106–108, 110–115 SPOT robot.....................................................41, 44, 46 SPOT technology..............................105–106, 114, 115 Spotting................. 17, 19, 43–47, 53, 71, 93, 108, 110, 113, 116–118, 124, 125, 153, 173, 174, 192, 201–202, 213, 240, 323–325 microplates................................................41, 44, 46–47 SPR. See Surface plasmon resonance ssDNA. See Single-stranded DNA Stable transfection.......................................................... 182 Staining bromophenol blue (BPB).......................................... 107 Statistical analysis false discovery rate............................................ 158, 341 overfitting data..................................................... 6, 158 Steady-state fluorescence................................................ 309 Stealth microarray printhead............................................ 93 Stealth micro spotting prints............................................ 93 Stock solution..............................33, 46, 109, 112, 178, 246, 247, 260 Storage amino-acid solutions................................................. 112 membranes............................................................... 123 microarray slides....................................................... 118 Streptavidin....................... 6, 8, 10, 16, 18, 32, 35, 116, 143, 146, 186, 229, 235, 238, 242, 252, 277, 279–283, 285, 313–316 Streptavidin-HRP........................... 244, 248, 252, 281, 283 Stripping.......................................... 109, 120, 260, 279, 298 Sucrose.............................108, 171, 173, 176, 244, 258, 260 Supernatant................. 9, 44–46, 53, 74, 76, 83, 95–97, 102, 123, 134, 233, 235, 240, 247, 268, 328, 330 Surface plasmon resonance (SPR) angle scan................................................................. 329 applications............................................................... 322 binding......................................................321, 327–330 blocking.....................................................323, 324, 327

data analysis...............................................324, 330–331 equipment................................................................. 324 imaging instruments......................................... 325, 327 materials....................................................325, 326, 329 printing arrays................................................... 324–327 region of interest (ROI).................................... 329–330 sample preparation............................................ 322, 325 Suspension bead array................................................ 29–36 SV40............................................................................... 181 Systematic evolution of ligands by exponential enrichment (SELEX)......................................................... 58

T T-cell epitope.....................................................87, 357, 358 T7 expression..........................................154, 156, 231, 250 T24 human bladder cancer cell line.........130, 131, 134, 145 Time-resolved fluorescence.....................309, 311, 313, 316 Tissue culture...........................9, 40, 43, 134, 240, 259, 324 Transcription.................................. 154, 156, 166, 175, 179, 186, 187, 189, 191, 228, 231, 242–245, 250, 251, 278, 322, 359 Transcriptional activating domain (AD)................ 166, 179 Transfection..................... 166, 167, 170–176, 178–182, 241 Transformation........................73, 74, 76, 83, 144, 147, 190, 259, 261, 262, 266–267, 345–346 Translocated bacterial effectors................................... 38, 40 Transplantation...............................................84–85, 87–88 Triple-transfection.......................................................... 180 Troglitazone (TGZ)................................323, 325, 328, 331 TSA. See Tyramide signal amplification Tumor antigen....................85, 237, 351, 352, 356–358, 361 Tumor-specific mutation........................................ 353–357 Two hybrid............................................................. 165–183 Type III secretion............................................................. 38 Type III secretion effector................................................ 38 Type IV secretion............................................................. 38 Tyramide signal amplification (TSA)..............155, 157, 217

U Ubiquitin (Ub)................................................214, 217–219 E3 ligase..............................................38, 214, 217, 219 Universal cloning system................................................ 258 Urea..................................................91, 92, 98, 109, 123, 131

V Vascular cell adhesion molecule–1 (VCAM–1).............. 6, 7 VCAM–1. See Vascular cell adhesion molecule–1 VECTABOND™ reagent......................171, 173, 180, 181 Vectors....................... 68, 71, 73, 76, 95, 169, 170, 172, 178, 179, 186, 187, 191, 192, 228, 229, 231, 234, 236, 237, 241–242, 250, 251, 257–259, 261–263, 266, 269, 270 Viral antigens................................................... 90, 235–237 Virulence factor.....................................................38, 39, 42 VPL slides.............................................................. 173, 180

Protein Microarray for Disease Analysis 373 Index

W

Z

Western blotting....................................... 11, 52, 82, 85, 95, 96, 98, 106, 240, 242, 244, 246–250, 252, 297

Z-factor.......................................................................... 211 Z-score............................................................211, 338–342 ZsGreen...........................................................171, 178, 179