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
Biological Microarrays Methods and Protocols
Edited by
Ali Khademhosseini Department of Medicine, MIT Division of Health Sciences and Tech, Harvard Medical School, Cambridge, MA, USA
Kahp-Yang Suh Department of Mechanical and Aerospace E, Seoul National University, Kwanak-gu, Shinlim-dong 56-1-1, 151-742, Seoul, Korea, Republic of (South Korea) Gwanak-ro 599, Gwanak-gu, Seoul 151-742, Republic of Korea
Mohammed Zourob Biophage Pharma, Montreal, QC, Canada
Editors Prof. Dr. Ali Khademhosseini Department of Medicine MIT Division of Health Sciences and Tech Harvard Medical School Cambridge, MA USA
[email protected]
Dr. Mohammed Zourob Biophage Pharma Montreal, QC Canada
[email protected]
Kahp-Yang Suh Department of Mechanical and Aerospace E Seoul National University Kwanak-gu, Shinlim-dong 56-1-1, 151-742 Seoul, Korea, Republic of (South Korea) Gwanak-ro 599, Gwanak-gu, Seoul 151-742, Republic of Korea
[email protected]
ISSN 1064-3745 e-ISSN 1940-6029 ISBN 978-1-934115-95-4 e-ISBN 978-1-59745-551-0 DOI 10.1007/978-1-59745-551-0 Springer New York Dordrecht Heidelberg London Library of Congress Control Number: 2010938723 © 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 Microarrays are spatially ordered arrays with ligands chemically immobilized in discrete spots on a solid matrix, usually a microscope slide. Microarrays are a high-throughput large-scale screening system enabling simultaneous identification of a large number of target molecules (up to several hundred thousand) that bind specifically to the immobilized ligands of the array. Microarrays represent a promising tool for clinical, environmental, and industrial microbiology since the technology allows relatively rapid screening and identification of large number of specific analytes or genetic determinants simultaneously. The successful use of microarrays requires attention to unique issues of experimental design and execution. This book provides an overview of the methodology and applications of biological microarrays in various areas of biological and biomedical research. This book presents a significant and up-to-date review of the various biological microarrays, recognition elements, their immobilization, characterization techniques by a panel of distinguished scientists. This work is a comprehensive approach to the biological microarrays area presenting a thorough knowledge of the subject and an effective integration of these biological entities on microarray surfaces in order to appropriately convey the state-of-the-art fundamentals and applications of the most innovative approaches. This book comprises of 18 chapters written by 50 researchers actively working in USA, Canada, Germany, Spain, Korea, China, and the UK. The authors were requested to adopt a pedagogical tone in order to accommodate the needs of novice researchers such as graduate students and post-doctoral scholars as well as of established researchers seeking new avenues. This has resulted in duplication of some material, which we have chosen to retain, because we know that many readers will pick only a specific chapter to read at a certain time. We have divided this book into two major sections. The first part (Chaps. 1–9) comprises nine chapters, which are devoted to the application of biological microarrays including DNA/RNA, apatmer, proteins, tissues, oligonucleotides, carbohydrates, biomaterials, cells, bacteria, and virus microarrays. The second part (Chaps. 10–18) describes in detail the different techniques used for generating the microarray platforms. The second part divided into four subsections including photolithography (microfluidic-based approaches and cells and proteins patterns using photolithography), bioprinting (microspotters, microprinting), soft lithography (microcontact, micromolding, microstructure surface based on chemical vapor deposition, permeability of microvascular tubes), and microarray bioinformatics. It covers the theory behind each technique and delivers a detailed state-ofthe-art review for all the new technologies used. This book is intended to be a primary source both on fundamental and practical information of where the biological microarray area is now and where it is headed in the future. We anticipate that the book will be helpful to academics, practitioners and professionals working in various fields to name a few biologist, biotechnologists, biochemists, analytical chemists, biomedical, physical, microsystems engineering, nanotechnology, medicine, food, bioterrorism and security as well as allied health, health care, and surveillance. Since
v
vi
Preface
the fundamentals were also reviewed, we believe that the book will appeal to advanced undergraduate and graduate students who study in areas related to biological microarrays and biosensors. We gratefully acknowledge all authors for their participation and contributions, which made this book a reality. We give many thanks to Prof. John M. Walker for his guidance and patience. Last, but not least, we thank our families for their patience and enthusiastic support of this project.
Cambridge, MA Seoul, Korea Montreal, QC
Ali Khademhosseini Kahp-Yang Suh Mohammed Zourob
Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
v ix
Part I Application of Biological Microarray 1 RNA and DNA Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Stuart C. Sealfon and Tearina T. Chu 2 Aptamer Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eva Baldrich 3 Oligonucleotide Microarrays for Identification of Microbial Pathogens and Detection of Their Virulence-Associated or Drug-Resistance Determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dmitriy V. Volokhov, Hyesuk Kong, Keith Herold, Vladimir E. Chizhikov, and Avraham Rasooly 4 Protein Microarrays Printed from DNA Microarrays . . . . . . . . . . . . . . . . . . . . . . Oda Stoevesandt, Mingyue He, and Michael J. Taussig 5 Lithographically Defined Two- and Three-Dimensional Tissue Microarrays . . . . . Esther W. Gomez and Celeste M. Nelson 6 Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ku-Lung Hsu, Kanoelani Pilobello, Lakshmipriya Krishnamoorthy, and Lara K. Mahal 7 Cell Microarrays Based on Hydrogel Microstructures for the Application to Cell-Based Biosensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Won-Gun Koh 8 Fabrication of Bacteria and Virus Microarrays Based on Polymeric Capillary Force Lithography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Pil J. Yoo 9 3D Polymer Scaffold Arrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Carl G. Simon Jr., Yanyin Yang, Shauna M. Dorsey, Murugan Ramalingam, and Kaushik Chatterjee
3 35
55
95 107
117
133
147 161
Part II Methods for Microarray Generation 10 PDMS Microfluidic Capillary Systems for Patterning Proteins on Surfaces and Performing Miniaturized Immunoassays . . . . . . . . . . . . . . . . . . . 177 Mateu Pla-Roca and David Juncker 11 Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 195 Ji Youn Lee and Alexander Revzin
vii
viii
Contents
12 Generation of Protein and Cell Microarrays on Functionalized Surfaces . . . . . . . . Yoo Seong Choi and Chang-Soo Lee 13 Microprinting of Liver Micro-organ for Drug Metabolism Study . . . . . . . . . . . . . Robert C. Chang, Kamal Emami, Antony Jeevarajan, Honglu Wu, and Wei Sun 14 Microcontact Printing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yunyan Xie and Xingyu Jiang 15 Micromolding for the Fabrication of Biological Microarrays . . . . . . . . . . . . . . . . . Ashley L. Galloway, Andrew Murphy, Jason P. Rolland, Kevin P. Herlihy, Robby A. Petros, Mary E. Napier, and Joseph M. DeSimone 16 Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Yaseen Elkasabi and Joerg Lahann 17 Methods for Forming Human Microvascular Tubes In Vitro and Measuring Their Macromolecular Permeability . . . . . . . . . . . . . . . . . . . . . . . Gavrielle M. Price and Joe Tien 18 Microarray Bioinformatics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Robert P. Loewe and Peter J. Nelson
207 219
239 249
261
281 295
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 321
Contributors Eva Baldrich • Instituto de Microelectrónica de Barcelona (IMB-CNM), Barcelona, Spain Robert C. Chang • Department of Mechanical Engineering and Mechanics, Drexel University, Philadelphia, PA, USA Kaushik Chatterjee • Polymers Division, National Institute of Standards and Technology, Gaithersburg, MD, USA Vladimir E. Chizhikov • Center for Biologics Evaluation and Research, Food and Drug Administration, Kensington, MD, USA Yoo Seong Choi • Department of Chemical Engineering, Pohang University of Science and Technology, Pohang, Korea Tearina T. Chu • Departments of Pharmacology and Systems Therapeutics, Mount Sinai School of Medicine, New York, NY, USA Joseph M. DeSimone • North Carolina State University, Raleigh, NC, USA Shauna M. Dorsey • Polymers Division, National Institute of Standards and Technology, Gaithersburg, MD, USA Yaseen Elkasabi • Material Science and Engineering, University of Michigan, Ann Arbor, MI, USA Kamal Emami • Radiation Physics Laboratory, NASA Johnson Space Center, Houston, TX, USA Ashley L. Galloway • Liquidia Technologies Research, Triangle Park, NC, USA Esther W. Gomez • Departments of Chemical Engineering and Molecular Biology, Princeton University, Princeton, NJ, USA Mingyue He • The Babraham Institute, Cambridge, UK Kevin P. Herlihy • Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Keith Herold • Department of Bioengineering, University of Maryland, College Park, MD, USA Ku-Lung Hsu • Department of Chemistry and Biochemistry, University of Texas at Austin, Austin, TX, USA Antony Jeevarajan • Radiation Physics Laboratory, NASA Johnson Space Center, Houston, TX, USA Xingyu Jiang • National Center for NanoScience & Technology, Beijing, China David Junker • Bio-Medical Engineering Department, McGill University, Montreal, QC, Canada Won-Gun Koh • Department of Chemical and Biological Engineering, Yonsei University, Seoul, Korea Hyesuk Kong • Center for Biologics Evaluation and Research, Food and Drug Administration, Kensington, MD, USA
ix
x
Contributors
Lakshmipriya Krishnamoorthy • Institute for Cellular and Molecular Biology, University of Texas at Austin, Austin, TX, USA Joerg Lahann • University of Michigan, Ann Arbor, MI, USA Chang-Soo Lee • Department of Chemical and Biological Engineering, Chungnam National University, Daejeon, Korea Ji Youn Lee • Department of Biomedical Engineering, University of California, Davis, CA, USA Robert P. Loewe • Medical Policlinic, Ludwig Maximillians, University of Munich, Munich, Germany Lara K. Mahal • Chemistry and Biochemistry Department, The University of Texas at Austin, Austin, TX, USA Andrew Murphy • Liquidia Technologies Research, Triangle Park, NC, USA Mary E. Napier • Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill NC, USA Celeste M. Nelson • Department of Chemical Engineering, Princeton University, Princeton, NJ, USA Peter J. Nelson • Medical Policlinic, Ludwig Maximillians, University of Munich, Munich, Germany Robby A. Petros • Department of Chemistry, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA Kanoelani Pilobello • Institute for Cellular and Molecular Biology, University of Texas at Austin, Austin, TX, USA Gavrielle M. Price • Department of Biomedical Engineering, Boston University, Boston, MA, USA Murugan Ramalingam • Polymers Division, National Institute of Standards and Technology, Gaithersburg, MD, USA Avraham Rasooly • National Institutes of Health, National Cancer Institute, FDA, Bethesda, MD, USA Alexander Revzın • Department of Biomedical Engineering, University of California, Davis, CA, USA Jason P. Rolland • Liquidia Technologies Research, Triangle Park, NC, USA Stuart C. Sealfon • Neurology, Mount Sinai School of Medicine, New York, NY, USA Oda Stoevesandt • Babraham Bioscience Technologies, Cambridge, UK Wei Sun • Department of Mechanical Engineering and Mechanics, Drexel University, Philadelphia, PA, USA Joe Tien • Department of Biomedical Engineering, Boston University, Boston, MA, USA Dmitriy V. Volokhov • Center for Biologics Evaluation and Research, Food and Drug Administration, Kensington, MD, USA Honglu Wu • Radiation Physics Laboratory, NASA Johnson Space Center, Houston, TX, USA Yunyan Xie • National Center for NanoScience & Technology, Beijing, China Yanyin Yang • Polymers Division, National Institute of Standards and Technology, Gaithersburg, MD, USA
Contributors
xi
Pil J. Yoo • Chemical Engineering, Sungkyunkwan University, Seoul, Korea Michael J. Taussig • Babraham Bioscience Technologies, Cambridge,UK Carl G.Simon Jr • Polymers Division,National Institute of Standards and Technology, Gaithersburg, MD, USA Mateu Pla-Roca • Bio-Medical Engineering Department, McGill University, Montreal, QC, Canada
Part I Application of Biological Microarray
Chapter 1 RNA and DNA Microarrays Stuart C. Sealfon and Tearina T. Chu Abstract The development of microarray technology has revolutionized RNA and deoxyribonucleic acid (DNA) research. In contrast with traditional biological assays, microarrays allow the simultaneous measurement of tens of thousands of messenger RNA (mRNA) transcripts for gene expression or of genomic DNA fragments for copy number variation analysis. Over the past decade, genome-wide RNA or DNA microarray analysis has become an essential component of biology and biomedical research. The successful use of microarrays requires attention to unique issues of experimental design and execution. This chapter provides an overview of the methodology and applications of RNA and DNA microarrays in various areas of biological research. Key words: RNA, DNA, Expression, Comparative genomic hybridization, cDNA, BAC, Microarray, Copy number variation, Transcripts
1. Introduction Deoxyribonucleic acid (DNA) carries the hereditary information content of the genome, which is organized into discrete functional genes that regulate and encode individual RNAs. Genome size varies in organisms ranging from bacteria containing 1–5 million bases and 1,000–3,000 genes (1) to humans containing three billion bases and 20,000–25,000 protein-coding genes. Genes encoding protein are dynamically regulated and produce messenger RNA (mRNA) that are translated into protein. The human genome also contains thousands of nonprotein encoding RNA genes and large areas of regulatory and noncoding sequence (2). Measuring mRNAs indicate the level of gene activity and provide a snapshot of the biosynthetic state of the cell or tissue. The expression level of each gene can be influenced by a combination of genetic or environmental factors. The genetic factors include Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_1, © Springer Science+Business Media, LLC 2011
3
4
Sealfon and Chu
DNA polymorphisms in the regulatory regions of genes (such as the promoter/enhancer regions), and variations in the number of copies of the gene [copy number variation (CNV)]. Environment factors such as temperature, stress, nutrition, or exercise can lead to changes in extracellular hormones or intracellular signaling molecules that influence the expression level of genes. The expression levels of the genes of a cell determine the cell type, developmental stage, cell functions, and/or pathological state. However, it must be noted that the measurement of mRNA levels provides an imperfect reflection of protein levels and activity. The concentration of a particular protein is controlled not only by the level of its mRNA, but also by the rate of mRNA translation into protein and of protein degradation. Other modifications of protein, such as phosphorylation, are also important determinants of activity. With these limitations in mind, measurement of global mRNA expression provides insight into the overall level of gene activity and protein expression. Many human diseases involve altered gene expression (3). Small genomic deletions and duplications (1 kb to 10 Mb) constitute up to 15% of all mutations underlying human monogenic diseases (4). Thus, the study of small regions of chromosomal variation provides insight into the pathogenesis of many diseases. Changes in gene expression can arise from polymorphisms, deletions, or insertions in protein coding or regulatory sequences of DNA. Changes in gene expression can also arise from altered regulation of mRNA production in response to various signaling mechanisms or stimuli. In contrast with classical Mendelian genetics involving hereditable defects of a single gene locus, many diseases are polygenetic and have clusters of genes that may contribute to the pathological state (5–11). Microarray techniques that allow detection of small regions of DNA deletions or duplications play an important role in mapping diseases with a complex hereditary etiology. Microarray technology was first introduced in 1995 by Patrick Brown and colleagues (12). The first microarray was generated using complementary DNAs (cDNA) derived from polymerase chain reaction (PCR) products. The array was printed using a home-made robot and was used to measure the gene expression patterns in parallel of 48 Arabidopsis thaliana genes. Advances in microarray technology and the decoding of the human genome (13–15) as well as the genome of many other species (16–20), now make it feasible to assay simultaneously the expression level of tens of thousands of mRNA transcripts. We use the term RNA microarrays to refer to arrays used to measure RNA levels, whereas DNA microarrays measure DNA sequence or levels. RNA microarrays have been widely used to identify regulated genes, pathways, or gene networks in a variety of cells and tissues when two or more related biological conditions are compared.
RNA and DNA Microarrays
5
These approaches provide insight into biological mechanisms or cellular programs such as cell cycle progression (21–23), embryonic development (24–26), cell fate determination (27, 28), hormone responsive gene regulation programs (29–31), and drug or disease model -mediated gene expression changes (32–35). Microarrays have also been widely to define disease-associated gene regulation, gene expression patterns in disease subtypes, and gene biomarkers of various disease states such as cancers (36–40), infectious diseases (41–43), inflammatory disease (44, 45), neurological diseases (46–48), and psychiatric disorders (32, 49–51). In addition, microarray approach has been used in pharmacogenomic/ toxicogenomic studies for drug discovery, and for determining the mechanisms of therapeutic or side effects of specific drugs (35, 52–55). In the development of many human diseases, for example tumors, chromosomal damage leads to gain or loss of genomic material (4, 56–58). Comparative genomic hybridization (CGH) allows the study of the entire genome for variations in DNA copy number. Originally, metaphase chromosomes were used to represent the genome (59). This approach has limited resolution (~5–10 Mb for single copy gains and losses) and is technically difficult in requiring optimal chromosome metaphase spreads. Array-CGH (aCGH) circumvents some of these technical difficulties, and offers higher resolution (~1–100 kb intermarker spacing). This technique is useful for the detection of deletions or duplications of chromosomal regions or gene CNV in a comparison between individuals with altered disease states (60–62). Microarrays are important in cancer biology in identifying new tumor subtypes and prognostic groups. For example, breast cancer is a heterogeneous disease comprising many biological subtypes. After diagnosis, 30–40% of patients will develop metastases and die of the disease within 15 years. The selection of adjuvant chemotherapy is currently based on prognostic and predictive factors including age, tumor size, histological grade, hormone receptor status, Her2/neu status, menopausal status, and lymph node status (63, 64). Although these “classical” factors are effective based on general population statistics, they poorly predict the outcome for the individual patient because of the heterogeneity of this disease. Using RNA microarrays, Perou et al. was the first to divide breast carcinomas into distinct subtypes, based on the unique gene expression patterns of a subgroup of genes, called the “gene set” or “gene signature” (37). In a follow-up study, this gene set was used to predict the prognosis of 78 breast carcinoma patients (40). Using similar approaches, other workers have developed gene sets to predict the development of metastases and the prognosis of different groups (65–68). Microarrays have also been used to predict response to therapy (69–71). Using a DNA CGH microarray, Gonzalez-Neira et al. identified a genetic
6
Sealfon and Chu
classifier based on specific somatic genetic aberration of regions on chromosome 3p, 3q, and 5q to differentiate BRCA1 (breast cancer risk gene) mutation carriers from nonBRCA1 carriers of breast cancer patients (72). In a similar manner, array-CGH technology has been used in the characterization of different features of breast cancer including tumor stage classification and diagnostic/ prognostic subgroups of patients (for review see ref. 73). These studies raise the possibility that prediction of prognosis may be improved and that treatment for many diseases may be individualized based on gene analysis or gene signature patterns. The first diagnostic microarray to influence drug selection and dosage, the Roche AmpliChip Cytochrome P450 test, was approved by the FDA in 2004. Large-scale DNA or RNA analysis can be performed using microarrays generated by cloned libraries or synthetic oligonucleotides. The latter, nucleotide microarrays, will be discussed in a different section of this volume. In this chapter, we provide an introduction to the use, sample quality control, and sample preparation for cDNA-based RNA microarrays and BAC-based CGH DNA microarrays. The selection of reagents that are used in the materials and methods is based on authors’ preferences and experience. 1.1. RNA Microarray
Printing cDNA microarrays on glass slides was the first microarray technique developed and is still commonly used. cDNAs are amplified from individual clones in a library. Each cDNA fragment representing an individual gene of interest is immobilized on a glass slide that has been coated with DNA-binding chemicals such as amino silane or poly-l-lysine. These slide arrays can be printed as whole genome microarrays or with a focused selection of genes of interest. The two-color cDNA microarray assay is illustrated in Fig. 1. In a typical slide microarray experiment, the mRNAs from experimental samples to be compared (such as test vs. control) are reverse transcribed and are then labeled with two different detectable fluorescent markers (typically Cy3 vs. Cy5 or compatible Alexa dyes). When the amount of RNA in each sample is limited, such as assays from few or single cells, the RNAs may be subjected to an amplification procedure prior to fluorescent labeling. The two labeled samples are mixed and then hybridized to a microarray. After the excess of labeled probes is removed by washing, the intensity of each fluorophore at each array location is read using a laser scanner. Hybridization intensity is represented by the amount of fluorescent emission, which provides an estimate of the relative amount of each transcript present in the different samples. These arrays are printed using a library containing the sequences of interest. In a library, each bacteria clone carries a plasmid containing a unique sequence derived from the mRNA of a gene.
RNA and DNA Microarrays Test
RNA QC
Labeling QC
7
Control
or Total RNA RT & label or Amplify &
or cRNA
Fig. 1. Two-color RNA microarray assay. Test and control samples are reverse transcribed and labeled with different fluorophores. The labeled samples are competitively hybridized to the microarray that has been printed at each location with a different DNA sequence for a gene of interest. Determination of the relative fluorescence obtained at each array location, after normalization, reflects the relative levels of expression of each specific mRNA in the original samples. RT reverse transcription, QC quality control.
The library cDNAs, ranging from 500 bp to 2 kb, are amplified by PCR using the specific flanking primers of the gene, according to each clone library. The amplified cDNA fragments are purified, and the concentration determined. The end product of each clone is quality controlled by gel electrophoresis and printed using a specialized printing robot. 1.1.1. RNA QC
A slide RNA microarray offers a high level of flexibility in terms of labeling choices. The starting amount of RNA plays a significant role in influencing the choice of method used for target production, since it determines the level of amplification required. Three examples of labeling protocols are included in this section. Regardless of which method is chosen, the test and control samples must be properly matched to minimize the background variation.
1.2. DNA Microarray
CGH allows partial or entire genome analysis for variations in DNA copy number. In a CGH DNA microarray, artificial chromosomes from bacteria (BAC) containing known genomic
8
Sealfon and Chu
fragments are immobilized on a glass surface and hybridized with a mixture of different fluorescence-labeled test and control DNAs. The ratio of the fluorescent intensities of the fluorophores is measured for each feature on the array. This ratio provides a relative measure of the difference in gene copy number between the samples. This technique is useful for a comparison between individuals with altered disease states in order to detect tissue-specific deletions or duplications of chromosomal regions. BAC genomic libraries can be purchased or custom made by an industrial provider or academic core facility, e.g., National Human Genome Research Institute, Wellcome Trust Sanger institute, Children’s Hospital Oakland Research Institute, Roswell Park Cancer Center, Clemson University Genomics Institute, Arizona Genomics Institute, Arabidopsis Biological Resource Center (ABRC) at Ohio State University, etc. The detailed protocol for preparation of the BAC DNAs is usually provided with the source library obtained, and the fabrication of a microarray slide is performed in an array printing service center. Briefly, BAC clones are streaked on LB-agar plates containing the appropriate antibiotic and grown overnight at 37°C. A single colony is inoculated in TB media containing the appropriate antibiotic and placed in a shaking incubator at 37°C for 16 h. From this culture, DNA is isolated and amplified through ligation-mediated PCR where a genomic DNA clone is digested with a restriction enzyme and a universal primer adaptor is ligated to serve as a priming site for PCR amplification. The amplified genomic DNA fragments are purified, the concentrations are determined and normalized, and then they are used for the fabrication of array slides for CGH. 1.2.1. DNA QC
Good quality of genomic DNA generally increases the sensitivity and accuracy of the array CGH assay. The DNA must be pure and free of contaminants, especially other sources of genomic DNA. Genomic DNA can be extracted from various sources, e.g., blood, buccal cells, cultured cells, tissue, and paraffin-embedded tissue. Many commercial kits targeting a particular sample source produce high-quality genomic DNA. Methods that include boiling or strong denaturants that may generate single-stranded DNA are not suitable to use. The purity of the DNA can be determined by the 260/280 spectrophotometer absorbance ratio. Ratio of 1.8 in a 10-mM Tris-HCl buffer typically represents pure DNA, whereas lower value for protein contamination and higher value for RNA contamination or degraded samples. The approximate average size of genomic DNA can be viewed on a 1% agarose gel. High-quality genomic DNA will run as a major peak at approximately 10–20 kb on the gel. Whole genome amplification (WGA) was developed in 1992 (74, 75) as a way to increase the amount of DNA from limited samples such as forensics and genetic disease research.
RNA and DNA Microarrays
9
Various WGA techniques have been developed. One approach, multiple displacement amplification (MDA) provides unbiased and accurate amplification of whole genomes (76, 77). This method utilizes Phi29 DNA polymerase, producing micrograms of high-molecular weight DNA fragments from as little as 10 ng of starting DNA. The end products are suitable for array CGH assay.
2. Materials 2.1. RNA QC 2.1.1. Direct Labeling
See Note 1 for preparing the mixture. 1. Oligo-dT18. 2. Superscript II. 3. 100 mM dNTP set. 4. Cy3- or Cy5- dUTP. 5. RNAsin. 6. RNAse H. 7. RNAseOne. 8. Minelute Cleanup Kit.
2.1.1.1. Stock Solutions and Master Mixtures
10× low dTTP dNTPs stock solution: Reagents
Amount
dGTP (100 mM)
25 ml
dATP (100 mM)
25 ml
dCTP (100 mM)
25 ml
dTTP (100 mM)
10 ml
DEPC-water
415 ml
Total
500 ml
Master Mix for making fluorescent cDNA target: Reagents
Amount
5× First strand buffer
6 ml
0.1 M DTT
3 ml
10× Low dTTP dNTP mix
0.8 ml
Cy-3 or Cy-5 dUTP (1 mM)
2 ml
RNAsin
1 ml
Total
12.8 ml
10
Sealfon and Chu
2.1.2. Indirect Labeling
1. Oligo-dT18. 2. Superscript II. 3. 100 mM dNTP set. 4. aa-dUTP. 5. Sodium bicarbonate. 6. CyDye PostLabeling Reactive Dye Pack. 7. K2HPO4. 8. KH2PO4. 9. 0.5 M EDTA. 10. NaOH. 11. HCl. 12. MinElute cleanup kit. 13. NaAcetate. 14. DMSO.
2.1.2.1. Stock Solutions and Master Mixtures
Phosphate buffers (1 M Potassium phosphate, pH 8.5): Check pH with pH paper. Reagents
Amount
1 M K2HPO4
9.5 ml
1 M KH2PO4
0.5 ml
Total
10 ml
Phosphate wash buffer (5 mM KPO4, pH 8.5, 80% EtOH): Reagents
Amount
1 M Phosphate buffer, pH 8.5
0.5 ml
Water
15.25 ml
95% EtOH (alcohol)
84.25 ml
Total
100 ml
Phosphate elution buffer: Reagents 1 M Phosphate buffer, pH 8.5
Amount 4 ml
Water
996 ml
Total
1,000 ml
RNA and DNA Microarrays
11
100 mM Amino-allyl dUTP: Store in −20°C Reagents
Amount
aa-dUTP
1 mg
Water
19.1 ml
50× Labeling Mix (2:3 aa-dUTP:dTTP): Store in −20°C Reagents
Amount
dGTP (100 mM)
5 ml
dATP (100 mM)
5 ml
dCTP (100 mM)
5 ml
dTTP (100 mM)
3 ml
aa-dUTP (100 mM)
2 ml
Total
20 ml
0.3 M Sodium bicarbonate, pH 9.0: Check pH with pH paper. Use for 1 day only. Reagents
Amount
Sodium bicarbonate
1 g
dH2O
40 ml
NaOH (10 N)
180 ml
aa-dUTP-labeled cDNA target:
2.1.3. Small Sample Labeling
Reagents
Amount
5× First strand buffer
6 ml
0.1 M DTT
3 ml
50× Aminoallyl-dNTP mix
0.6 ml
Total
9.6 ml
1. Low input RNA amplification kit, see Note 2. 2. RNeasy MinElute Cleanup kit. 3. 50 mM aa-UTP. 4. 100 mM NTP set. 5. Sodium bicarbonate. 6. CyDye PostLabeling Reactive Dye Pack. 7. Hydroxylamine. 8. Fragmentation buffer.
12
Sealfon and Chu
2.1.3.1. Stock Solutions and Master Mixtures
Low UTP NTP mixture: Aliquot 100 ml per tube and store at −80°C. Reagents
Amount
100 mM ATP
100 ml
100 mM GTP
100 ml
100 mM CTP
100 ml
100 mM UTP
75 ml
DEPC-water
25 ml
Total
400 ml
25 mM aa-UTP mixture: Aliquot 50 ml per tube and store at −80°C. Reagents
Amount
50 mM aa-UTP
100 ml
DEPC-water
100 ml
Total
200 ml
0.3 M Sodium bicarbonate at pH 9.0: Check pH with pH paper. Use for 1 day only. Reagents
Amount
Sodium bicarbonate
1 g
dH2O
40 ml
NaOH (10 N)
180 ml
cDNA Master Mix: Reagents
Amount
5× First strand buffer
4 ml
0.1 M DTT
2 ml
10 mM dNTP
1 ml
MMLV RT
1 ml
RNase OUT
0.5 ml
Total
8.5 ml
RNA and DNA Microarrays
In vitro transcription Master Mix:
2.1.4. Hybridization and Scanning (see Note 4)
Reagents
Amount
Water
5.7 ml
4× Transcription buffer
20 ml
0.1 M DTT
6 ml
Low UTP NTP Mix
16 ml
50% PEG
6.4 ml
RNAse OUT
0.5 ml
Inorganic pyrophosphatase
0.6 ml
aa-UTP (25 mM)
4 ml
T7 RNA polymerase
0.8 ml
Total
60 ml
1. Human or mouse Cot-1 DNA. 2. Poly(dA). 3. Transfer RNA (tRNA). 4. Formamide. 5. Succinic anhydride. 6. n-Methyl-pyrrilidinone. 7. NaBorate. 8. 50× Denhardt’s solution. 9. 20× SSPE. 10. 20× SSC. 11. SDS. 12. SS salmon sperm DNA. 13. Raised-edge coverslip. 14. Microarray hybridization cassette.
2.1.4.1. Stock Solutions
20× Blocking mixture: Reagents
Amount
Poly(dA)
40 mg
tRNA
80 mg
Human or mouse Cot-1 DNA 2,000 mg
13
14
Sealfon and Chu
Add 1/10 volume of 3 M NaAcetate, pH 5.2, then precipitate the content with 2.5 × volume of 100% ethanol. Wash the pellet 1× with 70% ethanol and air-dry. Resuspend in 20-ml filtered Mili-Q water. Hybridization solution: Reagents
Amount
Formamide
35 ml
20× SSPE
20 ml
20% SDS
2.5 ml
50× Denhardt’s solution
5 ml
Water
37.5 ml
Total
100 ml
Prehybridization solution: Reagents
Amount
Formamide
35 ml
20× SSPE
20 ml
10% SDS
5 ml
50× Denhardt’s solution
5 ml
SS salmon sperm DNA (10 mg/ml)
2 ml
Water
33 ml
Total
100 ml
2.2. DNA QC
1. REPLIg Mini Kit.
2.2.1. Materials for Genomic DNA Amplification
2. Mini Quick Spin Column.
2.2.1.1. Stock Solutions and Master Mixtures (see Note 5)
Buffer D1 (sufficient for 15 reactions): Reagents
Amount
Reconstituted DLB buffer
5 ml
Nuclease-free water
35 ml
Total
40 ml
Buffer N1 (sufficient for 15 reactions): Reagents
Amount
Stop solution
8 ml
Nuclease-free water
72 ml
Total
80 ml
RNA and DNA Microarrays
15
Master Mix for amplification: Add the master mix components in the order listed in the table below. After addition of water and reaction buffer, briefly vortex and centrifuge the mixture before the addition of the DNA polymerase. The master mix should be kept on ice and used immediately upon the addition of the DNA polymerase.
2.2.2. Direct Labeling
Reagents
Amount
Nuclease-free water
10 ml
Reaction buffer
29 ml
DNA polymerase
1 ml
Total
40 ml
1. BioPrime Array CGH Genomic DNA Labeling module. 2. 100 mM dNTP set. 3. 1 M Tris–HCl, pH 8.0. 4. 0.5 M EDTA, pH 8.0. 5. 1 mM Cy3 or Cy5-labeled dCTP. 6. Microcon YM 30.
2.2.2.1. Stock Solutions and Master Mixtures
10× dNTP Mix (0.5 mM dCTP, 2 mM dATP, 2 mM dGTP, 2 mM dTTP in TE buffer): Reagents
Amount
dGTP (100 mM)
4 ml
dATP (100 mM)
4 ml
dTTP (100 mM)
4 ml
dCTP (100 mM)
1 ml
1 M Tris–HCl, pH 8.0
2 ml
0.5 M EDTA, pH 8.0
0.4 ml
DEPC-water
184.6 ml
Total
200 ml
Master Mix for DNA labeling: Add the component in order listed. Reagents
Amount
10× dNTP
10 ml
Cy3 or Cy5-labeled dCTP
4 ml
Klenow
2 ml
Total
16 ml
16
Sealfon and Chu
2.2.3. Hybridization and Scanning
1. 3 M NaAcetate, pH 5.2. 2. Ethanol, 100%, 80%. 3. Yeast RNA. 4. Herring sperm DNA. 5. Cot-1 DNA. 6. TE buffer, pH 8.0. 7. Formamide. 8. Dextran sulphate. 9. 10% Tween-20. 10. SSC. 11. 1 M Tris buffer pH 7.4. 12. Raised-edge coverslip. 13. Microarray hybridization cassette.
2.2.3.1. Stock Solutions and Master Mixtures
Reagents
Amount
Formamide
500 ml
Dextran sulphate
100 mg
Tween-20
1 ml
20× SSC
100 ml
1 M Tris buffer, pH 7.4
10 ml
Nuclease-free water
~389 ml
Total
1 ml
Pre-/hybridization buffer (50% formamide, 10% dextran sulphate, 0.1% Tween-20, 2× SSC, 10 mM Tris–HCl, pH 7.4).
3. Methods 3.1. RNA QC
Accurate measurement of transcripts requires RNA samples that are free of degradation, which can differentially affect individual sequences. The quality of the total RNA should be verified by two methods – spectrophotometer-based assay and visualization of the ribosomal RNA (rRNA). The 260/280 spectrophotometer absorbance ratio is the simplest test to assess RNA quality. Ratios between 1.9 and 2.1 in a 10-mM Tris-based buffer typically represent high quality, pure RNA. Values considerably below this range suggest DNA, protein, or chemical contamination. Values greater than this range suggest the presence of degraded RNAs. A more sensitive method to assess the integrity of RNA is to
RNA and DNA Microarrays
a
17
c RIN=9.6 18S rRNA
RIN=5.5
28S rRNA
Leading marker
b
d RIN=8.3
RIN=2.8
Fig. 2. Representative Bioanalyzer electropherogram showing RNA samples of varying quality. RIN: RNA integrity number, ranging from 10 to 0 for the best to worst RNA integrity. RNAs in panels (c) or (d) are not suitable for RNA microarray assay.
visualize the rRNA component of total RNA. This can be achieved by performing an RNA gel electrophoresis; however, the process is tedious and requires micrograms levels of RNA. The Bioanalyzer (Agilent Technologies, Inc., Palo Alto, CA), a microfluidics-based platform, is a satisfactory way to assess RNA quality using small quantities of RNA. Representative Bioanalyzer readouts are depicted in Fig. 2. In addition to providing a “gel-like” image of the sample, this system derives an RNA integrity number or RIN that is useful in estimating the overall quality of total RNA samples. 3.1.1. Direct Labeling Methods
Reagents’ list and amounts are given in Subheading 2.1.1 to prepare Stock Solutions or Master Mixtures. 1. Resuspend 20–50 mg of total RNA or 1–2 mg of mRNA in DEPC-H2O to make the final volume of 13.2 ml. 2. Add 2 ml of oligo-dT18 (2 mg/ml). 3. Final volume: 15.2 ml. 4. Incubate at 65°C for 10 min. 5. Place on ice for at least 2 min.
18
Sealfon and Chu
3.1.1.1. To Make Fluorescent cDNA Target
1. Add the mixed content, given in Subheading 2.1.1, to the annealed RNA mix. 2. Add 2 ml of superscript II into the mixture. 3. Final volume: 30 ml. 4. Incubate at 42°C for 1 h. 5. Add another 1 ml of superscript II. 6. Incubate for another 1 h. 7. Heat at 94°C for 2 min.
3.1.1.2. To Degrade RNA
1. Add 48 ml dH2O to each tube. 2. Add 9 ml 10× RNAse One buffer. 3. Add 2 ml RNAse One. 4. Incubate at 37°C for 10 min. 5. Heat 94°C for 1 min.
3.1.1.3. To Cleanup cDNA Targets by MinElute Column
1. Keep Cy5 and Cy3 separate for MinElute cleanup to measure CyDye incorporation, if desired. 2. Add 9 ml 3 M NaAcetate, pH 5.2 to the sample tube. 3. Add 495 ml binding buffer to each sample. 4. Assemble the MinElute column on the provided 2-ml collection tubes. 5. Load the entire mixture to a MinElute column. 6. Centrifuge for 1 min at 10,000 RCF. Discard the flowthrough and reuse the 2-ml tube. 7. Add 750 ml PE wash buffer to the column. 8. Centrifuge at 10,000 RCF for 1 min. Discard the flowthrough and reuse the 2-ml tube. 9. Repeat steps 7 and 8. 10. Centrifuge again at maximum speed for 1 min to remove residual EtOH. 11. Place column in a fresh 1.5-ml tube. Add 10 ml of water (pH 7.5) to elute. 12. Allow elution water to stand for at least 2 min before spinning. 13. Centrifuge at maximum speed for 1 min. Add 10 ml of water (pH 7.5) to elute. 14. Allow elution water to stand for at least 2 min before spinning. 15. Centrifuge at maximum speed for 1 min. 16. Proceed to “Analysis of Target Labeling Reaction by NanoDrop Spectrophotometer” in Subheading 3.1.3.7.
RNA and DNA Microarrays
19
17. Combine equal amount of the exp- and control-labeled cDNA targets. Bring the final volume to 18.5 ml. Reduce volume by speedvac if necessary. 3.1.2. Indirect Labeling Methods
Reagents’ list and amounts are given in Subheading 2.1.2 to prepare Stock Solutions or Master Mixtures.
3.1.2.1. To Anneal RNA
1. Resuspend 20–50 mg of total RNA or 1–2 mg of mRNA in DEPC-H2O to make the final volume to 16.4 ml. 2. Add 2 ml of oligo-dT18 (2 mg/ml). 3. Final volume: 18.4 ml. 4. Incubate at 65°C for 10 min. 5. Quick spin and place on ice for at least 2 min.
3.1.2.2. To Make aa-dUTP-Labeled cDNA Target
1. Add the mixed content given in Subheading 2.1.2 to the annealed RNA mix. 2. Add 2 ml superscript II into the mixture. 3. Final volume: 30 ml. 4. Incubate at 42°C for 2 h. 5. Add another 1 ml superscript II. 6. Incubate for another 1 h. 7. Add 1 ml 0.5 M EDTA, see Note 6.
3.1.2.3. RNA Hydrolysis
1. Heat mixture at 95°C for 3 min. 2. Quick spin and immediately place on ice for at least 2 min. 3. Add 15 ml 1 M NaOH. 4. Mix and incubate at 65°C for 15 min. 5. Quick spin and put the tube on ice. 6. Add 15 ml 1 M HCl. 7. Total volume: 62 ml.
3.1.2.4. Targets Purification (see Note 7)
1. Add 6 ml 3 M NaAcetate, pH 5.2 to each sample tube. 2. Add 340 ml binding buffer to each sample. 3. Assemble the MinElute column on the 2-ml collection tubes provided. 4. Load the entire mixture to a MinElute column. 5. Centrifuge for 1 min @ 10,000 RCF. Discard the flowthrough and reuse the 2-ml tube. 6. Add 750 ml phosphate wash buffer to the column. 7. Centrifuge at 10,000 RCF for 1 min. Discard the flow-through and reuse the 2-ml tube. 8. Repeat steps 6 and 7.
20
Sealfon and Chu
9. Centrifuge again at maximum speed for 1 min to remove residual EtOH. 10. Place column in a fresh 1.5-ml tube. Add 10 ml of phosphate elution buffer to elute. 11. Allow elution buffer to stand for at least 2 min before spinning. 12. Centrifuge at maximum speed for 1 min. Add 10 ml phosphate elution buffer to elute. 13. Allow elution buffer to stand for at least 2 min before spinning. 14. Centrifuge at maximum speed for 1 min. 15. Dry sample completely in a speedvac. 3.1.2.5. Coupling aa-cDNA to Cy Dye Ester
1. Resuspend sample in 6 ml water. 2. Add 3 ml 0.3 M Na2CO3 buffer, pH 9.0. 3. Total volume: 9 ml. 4. Add 11 ml high-quality DMSO to one tube of Cy3 or Cy5 dye. 5. Vortex to mix thoroughly. Keep dye in the dark until ready to use (do not prepare dye >1 h before using). Make sure that no water gets into the dye/DMSO mix at any point. 6. Transfer the dye mix to sample tube. 7. Total volume: 20 ml. 8. Incubate at RT in the dark for 1 h.
3.1.2.6. Target Purification
1. To the tube, add 70 ml H2O and 10 ml 3 M NaOAc, pH 5.2. 2. Add 500 ml binding buffer. 3. Assemble the MinElute column on the provided 2-ml collection tubes. 4. Load the entire mixture to a MinElute column. 5. Centrifuge for 1 min at 10,000 RCF. Discard the flowthrough and reuse the 2-ml tube. 6. Add 750 ml PE buffer to the column. 7. Centrifuge at 10,000 RCF for 1 min. Discard the flowthrough and reuse the 2-ml tube. 8. Repeat steps 6 and 7. 9. Centrifuge again at maximum speed for 1 min to remove residual EtOH. 10. Place column in a fresh 1.5-ml tube. Add 10 ml of water (pH 7.5) to elute. 11. Allow elution buffer to stand for at least 2 min before spinning. 12. Centrifuge at maximum speed for 1 min. Add 10 ml of water (pH 7.5) to elute.
RNA and DNA Microarrays
21
13. Allow elution water to stand for at least 2 min before spinning. 14. Centrifuge at maximum speed for 1 min. 15. Put sample on ice and in the dark. 16. Proceed to “Analysis of Target Labeling Reaction by NanoDrop Spectrophotometer” in Subheading 3.1.3.7, if desired. 17. Combine the test- and control-labeled cDNA targets. Bring the final volume to 18.5 ml. Reduce volume by speedvac if necessary. 3.1.3. Small Sample Labeling Methods
3.1.3.1. RNA Annealing
Reagents’ list and amounts are given in Subheading 2.1.3 to prepare Stock Solutions or Master Mixtures. Preset a PCR program with lid heat off as follows: ●
65°C 10 min
●
4°C 5 min
●
4°C pause
●
40°C 2 h
●
65°C 15 min
●
4°C 5 min
●
4°C hold
1. Use 50 ng to 5 mg of total RNA per reaction. If possible, start with 2 mg of total RNA and make the final volume to 6.5 ml. 2. Add 5 ml T7 promoter primer. 3. Place the tube in a preprogrammed PCR machine. 4. Incubate at 65°C for 10 min, 4°C for 5 min and pause.
3.1.3.2. cDNA Synthesis
1. Prewarm 5× first strand buffer at 80°C for 3–4 min. Quick spin the tube and keep at RT until use. 2. To each sample tube, add 8.5 ml of cDNA Master Mix. 3. Incubate samples at 40°C for 2 h, 65°C for 15 min, 4°C for 5 min, then 4°C hold (see Note 8).
3.1.3.3. In Vitro Transcription
1. Prewarm the 50% PEG solution at 40°C for 1 min. 2. In a separate tube, prepare a master mix, given in Subheading 2.1.3 at RT immediately prior to use. 3. Add 60 ml of transcription master mix to each sample tube. 4. Total volume: 80 ml. 5. Place tubes on PCR machine. Run program 6. 40°C 2 h 7. 4°C hold 8. Add 20 ml RNase-free water to each sample tube to make a total volume of 100 ml.
22
Sealfon and Chu
3.1.3.4. cRNA Purification with RNeasy MinElute Column
See Note 9 for cRNA purification. 1. Prewarm RNase-free water at 50°C for at least 10 min. 2. Add 350 ml Buffer RLT, and mix thoroughly. 3. Add 250 ml ethanol (96–100%) to the mixture, and mix thoroughly by pipetting. Do not centrifuge. Continue immediately with step 4. 4. Apply the sample (700 ml) to an RNeasy MinElute column placed in a 2-ml collection tube (supplied). Close the tube gently, and centrifuge for 30 s at ³10,000 × g. 5. Discard the flow-through and collection tube. 6. Transfer the RNeasy column into a new 2-ml collection tube (supplied). Pipet 500 ml Buffer RPE onto the RNeasy column. Close the tube gently, and centrifuge for 1 min at ³10,000 × g to wash the column. Discard the flow-through. Reuse the collection tube in step 6. 7. Add 500 ml 80% ethanol to the RNeasy column. Close the tube gently, and centrifuge for 1 min at ³10,000 × g. Discard the flow-through. Reuse the collection tube and centrifuge for additional 2 min. 8. To elute, transfer the RNeasy column to a new 1.5-ml collection tube. Pipet preheated 10 ml RNase-free water directly onto the RNeasy silica-gel membrane. Close the tube gently, let sit at room temperature for 1 min, and centrifuge for 1 min at ³10,000 × g to elute. 9. Repeat step 8 once. 10. Quantitate cRNA yield by spectrophotometer.
3.1.3.5. Cy Dye Coupling Reaction
1. Add 11 ml of high-quality DMSO to each dye tube. Mix thoroughly and keep in dark. 2. Use 5 mg of aa-modified cRNA and vacuum dry (not to complete dryness). 3. Adjust volume to 6 ml. 4. Add 3 ml 0.3 M sodium bicarbonate buffer, pH 9.0 to sample tube. 5. Transfer the Cy-DMSO dye solution to sample tube and mix well. 6. Total volume: 20 ml. 7. Incubate in the dark at RT for 1 h. 8. Add 4.5 ml 4 M hydroxylamine solution to the mixture and incubate for 15 min in the dark at RT. 9. Add 5.5 ml DEPC-water to the labeled cRNA. 10. Total volume: 30 ml.
RNA and DNA Microarrays 3.1.3.6. Labeled cRNA Purification with RNeasy MinElute Column (see Note 8)
23
1. Prewarm RNase-free water at 50°C for at least 10 min. 2. Add 105 ml Buffer RLT, and mix thoroughly. 3. Add 75 ml ethanol (96–100%) to the mixture, and mix thoroughly by pipetting. Do not centrifuge. Continue immediately with step 4. 4. Apply the sample mixture (210 ml) to an RNeasy MinElute column placed in a 2-ml collection tube (supplied). Close the tube gently, and centrifuge for 30 s at ³10,000 × g. 5. Discard the flow through and collection tube. 6. Transfer the RNeasy column into a new 2-ml collection tube (supplied). Pipet 500 ml Buffer RPE onto the RNeasy column. Close the tube gently, and centrifuge for 1 min at ³10,000 × g to wash the column. Discard the flow-through. Reuse the collection tube in step 6. 7. Add 500 ml 80% ethanol to the RNeasy column. Close the tube gently, and centrifuge for 1 min at ³10,000 × g. Discard the flow-through. Reuse the collection tube and centrifuge for additional 1 min. 8. To elute, transfer the RNeasy column to a new 1.5-ml collection tube. Pipet preheated 10 ml RNase-free water directly onto the RNeasy silica-gel membrane. Close the tube gently; let it sit at room temperature for 1 min, and centrifuge for 1 min at ³10,000 × g to elute. 9. Repeat step 8 once. 10. Proceed to “Analysis of Target Labeling Reaction by NanoDrop Spectrophotometer” in Subheading 3.1.3.7 if desired. 11. Combine equal amount of the labeled cRNAs, approximately 100 pmol of Cy3 and 50 pmol Cy5 for each hybridization reaction. 12. Bring the volume to 20 ml with Nuclease-free water (or reduce volume in a speedvac if necessary). Do not dry completely. 13. Proceed to hybridization.
3.1.3.7. Analysis of Target Labeling Reaction by NanoDrop Spectrophotometer
1. Start the NanoDrop software. 2. Click the MicroArray tab. 3. Before initializing the instrument as requested by the software, clean the sample loading area with nuclease-free water. 4. Load 1.0 ml of nuclease-free water to initialize. Then, click OK. 5. Once the instrument has initialized, select RNA-40 (for cRNA), ssDNA-33 (for cDNA), or DNA-50 (for genomic DNA) as the Sample type (use the drop down menu) or according to your sample.
24
Sealfon and Chu
6. Make sure the Recording button is selected. If not, click Recording so that the readings can be recorded, saved, and printed. 7. Blank the instrument by pipetting 1.0 ml of nuclease-free water or elution buffer (whatever the samples are in) and click Blank. 8. Clean the sample loading area with a laboratory wipe. Pipette 1.0 ml of the sample onto the instrument sample loading area. Type the sample name in the space provided and click Measure (see Note 10). 9. Similarly, measure the RNA, ssDNA, or DNA absorbance by clicking the NucleicAcid tab in the main menu. 10. Print the results. If printing the results is not possible, record the following values: ●
Cyanine 3 or cyanine 5 dye concentration (pmol/ml)
●
RNA, ssDNA, or DNA absorbance ratio (260/280 nm)
●
cRNA, ssDNA, or DNA concentration (ng/ml)
11. Determine the yield and specific activity of each reaction as follows: ●
●
●
Use the concentration of RNA or DNA (ng/ml) to determine the mg RNA or DNA yield as follows: (Concentration of RNA or DNA) × (elution volume)/1,000 = mg of RNA or DNA. Use the concentrations of RNA or DNA (ng/ml) and cyanine 3 or cyanine 5 (pmol/ml) to determine the specific activity as follows: (Concentration of Cy3 or Cy5)/ (Concentration of RNA or DNA) × 1,000 = pmol Cy3/mg RNA or DNA. Use the A260 and A550 (for Cy3) or A650 (for Cy5) to determine the base-to-dye ratio as follows: Base/dye for Cy3™ = [A260 × 150,000 (cm–1 M–1)]/(A550 × 6,600). Base/dye for Cy5™ = [A260 × 250,000 (cm–1 M–1)]/ (A650 × 6,600). The base-to-dye ratio should be 40–80 for both Cy3 and Cy5.
12. Examine the yield and specific activity results. 13. If the yield is <0.8 mg and the specific activity is <30 pmol Cy3 or Cy5, do not proceed to the hybridization step. Repeat sample preparation. 3.1.4. Hybridization and Scanning 3.1.4.1. Prehybridization of the Combined Targets
Reagents’ list and amounts are given in Subheading 2.1.4 to prepare Stock Solutions or Master Mixtures. 1. Preheat a 50°C water bath. 2. In a tube, prepare the hybridization solution, given in Subheading 2.1.4.
RNA and DNA Microarrays
25
3. Add 40.5 ml hybridization solution to 18.5 ml labeled-target solution. 4. Add 1 ml 20× blocking solution. 5. Total volume: 60 ml. 6. Heat at 95°C for 1 min. 7. Centrifuge at max speed for 5 min and transfer the supernatant to a clean tube. 8. Prehybridize target at 50°C for 1 h. Template 18 x 18 mm
3.1.4.2. Preparation of Denatured Microarray Slide for Hybridization
1. Boil 200 ml of water in a beaker. Make sure the water level is high enough to immerse the entire array on a microarray slide. 2. Place the microarray slide on a piece of paper (a 3″ × 5″ Post-It works fine) and draw an outline around it. Then, mark the area that the array is printed on. See the example below: The array is visible when printed with SSC or other salt-containing printing buffer. If not, obtain the printed area from the microarray slide provider. 1. Prepare slide blocking solution right before start as follows – Dissolve 1 g of succinic anhydride (sigma) in 63 ml of n-methyl-pyrrilidinone. To this, add 7 ml of 0.2 M NaBorate, pH 8.0, and stir until dissolved. 2. Vapor moisturize arrayed DNA (array down ~2 s) over boiling water. 3. Quickly place in Stratalinker (array up). 4. Cross-link for 250 mJ (optimized for silane slide). 5. Vapor moisturize again. 6. Heat snap slide on a hot plate (~2 s or it easily breaks). 7. Soak array slides in blocking solution for 15–30 min with gentle shaking (the solution should always cover the entire array area). 8. Rinse slide in 0.1% SDS for 30 s. 9. Rinse slide in filtered water for 1 min with one quick change of water. 10. Boil slide in 95°C water for 5 min. 11. Duck into ice-cold ethanol for 1 min.
26
Sealfon and Chu
12. Remove excess ethanol from slides by spinning the slide in a 50-ml conical tube at 405 RCF, RT for 5 min (array to face down when spinning). 13. Slides are now ready for hybridization. On the other hand, salt deposit of the array on the slide is no longer visible (see Note 11). 3.1.4.3. Prehybridization of Microarray
1. In a tube, prepare the prehybridization solution given in Subheading 2.1.4. 2. Carefully place a precleaned coverslip on top of array (use the predrawn microarray slide template). Make sure that the edges of the coverslip never touch the array area. 3. Load 60 ml prehybridization solution over the array. 4. Add 10 ml water in each corner of the microarray cassette to maintain humidity. 5. Place slide in a hybridization cassette, and incubate at 50°C for 1 h. Coverslip should be cleaned with 0.1% SDS and rinsed thoroughly with filtered water and quick dried.
3.1.4.4. Hybridization
1. Remove the coverslip by rinsing the prehybridized slide in water for 2 min. 2. Spin the slide at 405 RCF for 5 min. 3. Carefully place a new precleaned coverslip on top of array (use the predrawn microarray slide template). Make sure that the edges of the coverslip never touch the array area. 4. Load 60 ml prehybridized target-hybridization solution over array. 5. Place the microarray slide in a hybridization cassette with 10 ml water in each corner. 6. Hybridize in a 50°C water bath for 17 h in the dark.
3.1.4.5. Array Wash
1. Place the microarray slide in a 50-ml tube with 2× SSC/0.1% SDS. 2. Shake gently so that coverslip falls off from the slide. 3. Place slide in a slide holder and put in a glass dish filled with a 400-ml solution containing 0.2× SSC/0.1% SDS. Make sure the array is completely immersed. 4. Shake slide for 5 min. 5. Repeat steps 3 and 4. 6. Dip slide in 0.2× SSC to remove SDS. 7. Place slide in a slide holder and put in a glass dish filled with a 400-ml solution containing 0.2× SSC. Make sure the array is completely immersed. 8. Shake slide for 5 min.
RNA and DNA Microarrays
27
9. Repeat steps 7 and 8. 10. Place slide in a 50-ml tube and spin at 405 RCF for 5 min to dry slide (array facing down when spinning). 11. Keep the slide in the dark until scanning. 3.1.4.6. Microarray Slide Scan
3.2. DNA QC 3.2.1. Genomic DNA Amplification 3.2.1.1. Presetting
1. Scan arrays as soon as possible on a microarray scanner. 2. Adjust the PMTs, so that the spots are just slightly below saturation for both channels. Reagents’ list and amounts are given in Subheading 2.2.1 to prepare Stock Solutions or Master Mixtures.
1. Preheat a water bath or heating block to 30°C. 2. Prepare sufficient Buffer D1 (denaturation buffer) and Buffer N1 (neutralization buffer) for the total number of WGA reactions.
3.2.1.2. Denature DNA
1. Dilute genomic DNA sample to 10–20 ng in TE buffer to a final volume of 2.5 ml. 2. Add 2.5 ml Buffer D1 to the DNA. Mix by vortexing and centrifuge briefly. 3. Incubate the samples at room temperature for 3 min. 4. Add 5 ml Buffer N1 to the samples. Mix by vortexing and centrifuge briefly. 5. Total volume: 10 ml.
3.2.1.3. Amplification
1. Thaw DNA polymerase on ice. Thaw all other components at room temperature, vortex, and then centrifuge briefly. 2. Prepare a master mix on ice, given in Subheading 2.2.1. Mix and centrifuge briefly. 3. Add 40 ml of the master mix to 10 ml of denatured DNA. 4. Incubate at 30°C for 16 h. 5. Heat the sample for 3 min at 65°C to inactivate the DNA polymerase. 6. Cool on ice.
3.2.1.4. Purify the End Product by Roche Quick Spin Column
1. Vigorously invert the column several times. 2. Remove the top cap, and then snap off the bottom tip. 3. Remove excess buffer by centrifuging at 1,100 × g for 2 min at RT into a 1.5-ml tube. 4. Discard the elute tube and place the column onto a new 1.5-ml tube.
28
Sealfon and Chu
5. Add 50 ml sample to the center of the column bed, and centrifuge at 1,100 × g for 4 min at RT. 6. Check concentration and yield by a spectrophotometer. 7. Store amplified DNA at –20°C. 3.2.2. Direct Labeling
Reagents’ list and amounts are given in Subheading 2.2.2 to prepare Stock Solutions or Master Mixtures.
3.2.2.1. DNA Labeling
1. In an amber tube, combine 40 ml 2.5× Random Primer with 1 mg genomic DNA, and bring volume to 84 ml with nucleasefree water. 2. Denature the DNA mixture at 95°C for 10 min, and immediately cool on ice for 5 min. 3. Quick spin. 4. Prepare a master mix. 5. Mix well and quick spin. 6. Incubate at 37°C for 16 h.
3.2.2.2. Cleanup of Labeled DNA by Microcon YM30
1. Quick spin the sample tube before opening the lid. 2. Add the reaction mixture (~100 ml) to a microcon YM30. 3. Add 300 ml dH2O. 4. Mark tube and column for orientation. 5. Centrifuge at RT for 8 min at 10,000–12,000 × g (or until almost all the liquid is filtered through the membrane). 6. Wash 3× with 400 ml dH2O; discard flow through. 7. Invert column and place it into a new amber tube. 8. Spin at 1,000–4,000 × g for 1 min to collect the target. 9. Proceed to “Analysis of Target Labeling Reaction by NanoDrop Spectrophotometer” in Subheading 3.1.3, if desired.
3.2.3. Hybridization and Scanning 3.2.3.1. Precipitation of Pellet for Slide Prehybridization (Prehyb Pellet)
Reagents’ list and amounts are given in Subheading 2.2.2 to prepare Stock Solutions or Master Mixtures. 1. Combine 80 ml 10 mg/ml Herring sperm DNA (preheat at 70°C for 5 min prior to use) and 100 ml Cot-1 DNA (1 mg/ml). 2. Add 18 ml 3 M NaAcetate, pH 5.2. 3. Add 500 ml 100% ethanol. Mix gently. 4. Precipitate at −80°C for at least 2 h or −20°C overnight. 5. Centrifuge at max speed (14,000 ´ g) for 20 min at 4°C. 6. Remove supernatant; add 500 ml 70% ethanol. 7. Re-centrifuge at max speed for 5 min. 8. Remove supernatant and air-dry the pellet.
RNA and DNA Microarrays 3.2.3.2. Labeled-DNA Precipitation
29
1. Combine equal cDNA amount (~7–10 mg) of cy3- and cy5label targets. 2. Add 100 ml 1 mg/ml Cot-1 DNA. 3. Vortex, then quick spin. 4. Add 1/10 volume of NaAcetate and 2.5× 100% ethanol. Mix thoroughly. 5. Precipitate at −80°C for at least 2 h. 6. Centrifuge at max speed (³14,000 ´ g) for 20 min at 4°C. 7. Remove supernatant; add 500 ml 80% ethanol. Recentrifuge at max speed for 5 min. 8. Remove supernatant with pipetman; Recentrifuge at max speed for 1 min. 9. Remove supernatant with a fine tip, and air-dry pellet for 3–5 min.
3.2.3.3. Prehybridization of Labeled Targets
1. Preheat hybridization buffer in a 75°C heat block. 2. To the target pellet, add 54 ml preheated hybridization buffer. 3. Incubate at 75°C for 2 min. 4. Make sure everything is resuspended. 5. Add 6 ml yeast tRNA (100 mg/ml in H2O). 6. Denature the tube at 75°C for 10 min. 7. Incubate at 37°C for 60 min. Keep in the dark.
3.2.3.4. Prehybridization of Microarray Slide
1. Pretreatment of microarray slide, follow the protocol of “Preparation of Denatured Microarray Slide for hybridization” in Subheading 3.1.4. 2. Resuspend the prehyb pellet in 65 ml hybridization buffer as described previously to make prehybridization buffer. 3. Denature the tube at 75°C for 10 min. 4. In the meantime, carefully place a precleaned coverslip on top of array (use the predrawn microarray slide template). Make sure that the edges of the coverslip never touch the array area. 5. Load 60 ml prehybridization solution over the array. 6. Add 10 ml water in each corner of the microarray cassette to maintain humidity. 7. Place slide in a hybridization cassette and incubate at 37°C for 1 h.
3.2.3.5. Hybridization
1. Remove the coverslip by dipping the prehybridized slide in water for 2 min. 2. Spin at 405 RCF for 5 min. 3. Carefully place a new precleaned coverslip on top of array.
30
Sealfon and Chu
4. Load 60 ml prehybridized target-hybridization solution over array. 5. Place the microarray slide in a hybridization cassette. 6. Hybridize in a 37°C water bath for 24–48 h in the dark. 3.2.3.6. Array Wash
1. Place the microarray slide in a 50-ml tube with 2× SSC/0.05% Tween-20 at 42°C for 5 min. 2. Shake gently so that coverslip falls off from the slide. 3. Transfer slide into a preheated solution of 50% formamide/2× SSC/0.1% Tween-20 at 42°C for 15 min in a Petri dish. 4. Transfer slide into a prewarmed solution of 2× SSC/0.05% Tween-20 at 42°C for 10 min in a rocking Petri dish. 5. Transfer the slide to a solution of 1× PBS/0.05% Tween-20 at RT for 10 min. 6. Dip slide in water. 7. Place slide in a 50-ml tube, and spin at 405 RCF for 5 min to dry the slide (array facing down when spinning). 8. Keep the slide in the dark until scanning.
3.2.3.7. Microarray Slide Scan
3.3. Conclusion
1. Scan arrays as soon as possible on a microarray scanner. 2. Adjust the PMTs so that the brightest spots are just slightly below saturation for both channels. Microarray technology is the most widely used tool for genomewise analyses of gene expression and CNV in basic science and clinical research. RNA and DNA microarrays are platforms available to perform this function. These arrays are mostly produced by core facilities in an academic setting. Several options for sample amplification or labeling methods developed by academia or industry are available. Special attention to sample preparation and handling are critical for obtaining high-quality assay results.
4. Notes 1. (a) Prepare experiment and control samples at the same time. One will be labeled with Cy-3 and the other one with Cy-5. (b) Standard molecular biology procedures to avoid RNA degradation should be followed carefully. 2. For small sample amplification and direct labeling, you can follow the instruction supplied with the Agilent low input RNA amplification kit. The method described below is a modified version of the indirect labeling method.
RNA and DNA Microarrays
31
3. (a) Water used should be filtered through 0.2 mm membrane. Mili-Q water (Millipore system) or equivalent is satisfactory. (b) Cy5 destruction has been reported after 10–30 s of exposure to ozone levels above 5–10 ppb, especially during the array washing, drying, and scanning steps. Problems have occurred at different geographical locations, and several potential solutions have been proposed (78–81). 4. (a) REPLI-g Mini DNA polymerase should be thawed on ice (see step 6). All other components can be thawed at room temperature (15–25°C). (b) Buffer D1 and Buffer N1 should not be stored longer than 3 months. (c) Make sure 500 ml of nuclease-free water has been added to DLB buffer. Reconstituted Buffer DLB can be stored for 6 months at –20°C. Buffer DLB is pH labile. Avoid neutralization with CO2. 5. If necessary, stop here and store the reaction at −20°C overnight. 6. Do not use Qiagen PE wash or elution buffer. Use home-made phosphate wash buffer and phosphate elution buffer instead. 7. This kit does not require cDNA purification after this step. 8. Ensure that b-ME is added to Buffer RLT before use. Read manufacturer’s instructions for detailed protocol. 9. Be sure to clean the sample loading area between measurements, and ensure that the baseline is always flat at 0, which is indicated by a thick black horizontal line. If the baseline deviates from 0 and is no longer a flat horizontal line, reblank the instrument as before, and then remeasure the sample. 10. Microarray slides should be kept in a slide box in a dessicator. (a) Do not introduce any ink on the slide. It will generate fluorescent background. (b) All microarray slides should have been baked at 80°C for 2 h after printing.
Acknowledgments Our microarray research is supported by grants from NIH NIDDK and by a research contract from NIAID. References 1. Fleischmann, R.D., et al., (1995) Whole-genome random sequencing and assembly of Haemophilus influenzae Rd. Science, 269(5223): p. 496–512. 2. International Human Genome Sequencing Consortium, (2004). Finishing the euchromatic sequence of the human genome. Nature, 431(7011): p. 931–945.
3. Schena, M., (1996).Genome analysis with gene expression microarrays. Bioessays, 18(5): p. 427–431. 4. Vissers, L.E.L.M., et al., (2005). Identification of disease genes by whole genome CGH arrays. Hum. Mol. Genet, 14(suppl_2): p. R215–223. 5. Fu, S., et al., (2008). Peripheral arterial occlusive disease: global gene expression analyses sug-
32
Sealfon and Chu
gest a major role for immune and inflammatory responses. BMC Genomics, 9(1): p. 369. 6. Lenk, G.M., et al., (2007). Whole genome expression profiling reveals a significant role for immune function in human abdominal aortic aneurysms. BMC Genomics, 8: p. 237–249. 7. Seo, D., et al., (2004). Gene expression phenotypes of atherosclerosis. Arterioscler Thromb Vasc Biol, 24: p. 1922–1927. 8. Singhal, S., et al., (2003). Gene expression profiling of malignant mesothelioma. Clin Cancer Res, 9: p. 3080–3097. 9. Schena, M., et al., (1996). Parallel human genome analysis: microarray-based expression monitoring of 1000 genes. Proc Natl Acad Sci USA, 93(20): p. 10614–10619. 10. DeRisi, J., et al., (1996). Use of a cDNA microarray to analyse gene expression patterns in human cancer. Nat Genet, 14(4): p. 457–460. 11. Milano, A., et al., (2008).Molecular subsets in the gene expression signatures of scleroderma skin. PLoS One, 3(7): p. e2696. 12. Schena, M., et al., (1995). Quantitative monitoring of gene expression patterns with a complementary DNA microarray. Science, 270(5235): p. 467–470. 13. Lander, E.S., et al., (2001). Initial sequencing and analysis of the human genome. Nature, 409(6822): p. 860–921. 14. Bentley, D.R., et al., (2008). Accurate whole human genome sequencing using reversible termi nator chemistry. Nature, 456(7218): p. 53–59. 15. Venter, J.C., et al., (2001). The sequence of the human genome. Science, 291(5507): p. 1304–1351. 16. Cerdeno-Tarraga, A.M., et al., (2003). The complete genome sequence and analysis of Corynebacterium diphtheriae NCTC13129. Nucleic Acids Res, 31(22): p. 6516–6523. 17. Rew, D.A., (2004). The sequencing of the rat genome. Eur J Surg Oncol, 30(8): p. 905–906. 18. Tettelin, H., et al., (2001). Complete genome sequence of a virulent isolate of Streptococcus pneumoniae. Science, 293(5529): p. 498–506. 19. Nierman, W.C., et al., (2001). Complete genome sequence of Caulobacter crescentus. Proc Natl Acad Sci USA, 98(7): p. 4136–4141. 20. Check, E., (2002). Draft mouse genome makes public debut. Nature, 417(6885): p. 106. 21. Galitski, T., et al., (1999). Ploidy regulation of gene expression. Science, 285(5425): p. 251–254. 22. Tu, B.P., et al., (2005). Logic of the yeast metabolic cycle: temporal compartmentalization of cellular processes. Science, 310(5751): p. 1152–1158.
23. Iyer, V.R., et al., (1999). The transcriptional program in the response of human fibroblasts to serum. Science, 283(5398): p. 83–87. 24. Plescia, C., C. Rogler, and L. Rogler, (2001). Genomic expression analysis implicates Wnt signaling pathway and extracellular matrix alterations in hepatic specification and differentiation of murine hepatic stem cells. Differentiation, 68(4–5): p. 254–269. 25. Lee, K.-H., D.-H. Yu, and Y.-S. Lee, (2009). Gene expression profiling of rat cerebral cortex development using cDNA microarrays. Neurochemical Res, 34(6): p. 1030–1038. 26. Nikolova-Krstevski, V., et al., (2008). Gene expression analysis of embryonic stem cells expressing VE-cadherin (CD144) during endothelial differentiation. BMC Genomics, 9(1): p. 240. 27. Zhu, H., H. Yang, M. R. Owen, (2007). Combined microarray analysis uncovers selfrenewal related signaling in mouse embryonic stem cells. Syst Synth Biol, 4(1): p. 171–181. 28. Chen, V.C., et al., (2008). Notch signaling respecifies the hemangioblast to a cardiac fate. Nat Biotech, 26(10): p. 1169–1178. 29. Kakar, S.S., et al., (2003). Identification of distinct gene expression profiles associated with treatment of LbetaT2 cells with gonadotropinreleasing hormone agonist using microarray analysis. Gene, 308: p. 67–77. 30. Malyala, A., et al., (2004). Suppression subtractive hybridization and microarray identification of estrogen-regulated hypothalamic genes. Neurochem Res, 29(6): p. 1189–1200. 31. Auger, C.J., H.M. Jessen, and A.P. Auger, (2006). Microarray profiling of gene expression patterns in adult male rat brain following acute progesterone treatment. Brain Res, 1067(1): p. 58–66. 32. Elin Lehrmann, W.J.F., (2008). Transcriptional correlates of human substance use. Ann N Y Acad Sci, 1139(Drug Addiction: Research Frontiers and Treatment Advances): p. 34–42. 33. Runne, H., et al., (2008). Dysregulation of gene expression in primary neuron models of Huntington’s disease shows that polyglutaminerelated effects on the striatal transcriptome may not be dependent on brain circuitry. J Neurosci, 28(39): p. 9723–9731. 34. Salas, R., et al., (2008). Nicotine relieves anxiogenic-like behavior in acetylcholinesteraseR overexpressing mice but not in wild-type mice. Mol Pharmacol, 74(6): p. 1641–1648, mol.108.048454. 35. Bergström, A., et al., (2007). Molecular pathways associated with stress resilience and drug resistance in the chronic mild stress rat model of depression – a gene expression study. J Mol Neurosci, 33: p. 201–215.
RNA and DNA Microarrays 36. Golub, T.R., et al., (1999). Molecular classification of cancer: class discovery and class prediction by gene expression monitoring. Science, 286(5439): p. 531–537. 37. Perou, C.M., et al., (2000). Molecular portraits of human breast tumours. Nature, 406(6797): p. 747–752. 38. Garber, M.E., et al., (2001). Diversity of gene expression in adenocarcinoma of the lung. Proc Natl Acad Sci USA, 98(24): p. 13784–13789. 39. Alizadeh, A.A., et al., (2001). Towards a novel classification of human malignancies based on gene expression patterns. J Pathol, 195(1): p. 41–52. 40. Sørlie, T., et al., (2001). Gene expression patterns of breast carcinomas distinguish tumor subclasses with clinical implications. Proc Natl Acad Sci USA, 98(19): p. 10869–10874. 41. Murphy, D.J. and J.R. Brown, (2007). Identification of gene targets against dormant phase Mycobacterium tuberculosis infections. BMC Infectious Disease, 7: p. 84. 42. Helbig, K.J., et al., (2008). Differential expression of the CXCR3 ligands in chronic hepatitis C, and their modulation by HCV in vitro. J Virol, 83(2): p. 836–846, JVI.01388-08. 43. Mikhailovich, V., et al., (2008). DNA microarrays in the clinic: infectious diseases. Bioessays, 30(7): p. 673–682. 44. Wright, H.J., et al., (2008). Periodontitis associates with a type 1 IFN signature in peripheral blood neutrophils. J Immunol, 181(8): p. 5775–5784. 45. Kaminski, N., et al., (2000). Global analysis of gene expression in pulmonary fibrosis reveals distinct programs regulating lung inflammation and fibrosis. Proc Natl Acad Sci USA, 97(4): p. 1778–1783. 46. Kinter, J., T. Zeis, and N. Schaeren-Wiemers, (2008). RNA profiling of MS brain tissues. Int MS J, 15(2): p. 51–58. 47. Satoh, J-i., et al., (2008). Neuromyelitis optica/ Devic’s disease: gene expression profiling of brain lesions. Neuropathology, 28(6): p. 561–576. 48. Grünblatt, E., et al., (2007). Comparison analysis of gene expression patterns between sporadic Alzheimer’s and Parkinson’s disease. J Alzheimers Dis, 12(4): p. 291–311. 49. Mash, D.C., et al., (2007). Gene expression in human hippocampus from cocaine abusers identifies genes which regulate extracellular matrix remodeling. PLoS One, 2(11): p. e1187. 50. Colantuoni, C., et al., (2008). Age-related changes in the expression of schizophrenia susceptibility genes in the human prefrontal cortex. Brain Struct Funct, 213(1): p. 255–271. 51. Iwamoto, K., M. Bundo, and T. Kato, (2005). Altered expression of mitochondria-related genes in postmortem brains of patients with
52.
53.
54. 55. 56.
57.
58.
59.
60.
61.
62.
63.
64.
33
bipolar disorder or schizophrenia, as revealed by large-scale DNA microarray analysis. Hum Mol Genet, 14(2): p. 241–253. Chin, K.-V. and A.N.T. Kong, (2002). Application of DNA microarrays in pharmacogenomics and toxicogenomics. Pharmaceutical Res, 19(12): p. 1773–1778. Duncan, C.E., A.F. Chetcuti, and P.R. Schofield, (2008). Coregulation of genes in the mouse brain following treatment with clozapine, haloperidol, or olanzapine implicates altered potassium channel subunit expression in the mechanism of antipsychotic drug action. Psychiatr Genet, 18(5): p. 226–239. Debouck, C. and P.N. Goodfellow, (1999). DNA microarrays in drug discovery and development. Nat Genet, 21(1 Suppl): p. 48–50. Gerhold, D.L., R.V. Jensen, and S.R. Gullans, (2002). Better therapeutics through microarrays. Nat Genet, 32(Suppl): p. 547–551. Kresse, S., et al., (2008). DNA copy number changes in high-grade malignant peripheral nerve sheath tumors by array CGH. Mol Cancer, 7(1): p. 48. Ankita, P., et al., (2008). Validation of a targeted DNA microarray for the clinical evaluation of recurrent abnormalities in chronic lymphocytic leukemia. Am J Hematol, 83(7): p. 540–546. Shaffer, LG., et al., (1993). Familial interstitial deletion 11(p11.12p12) associated with parietal foramina, brachymicrocephaly, and mental retardation. Am J Med Genet, 45(5): p. 581–583. Kallioniemi, A., et al., (1992). Comparative genomic hybridization for molecular cytogenetic analysis of solid tumors. Science, 258(5083): p. 818–821. Mizuguchi, T., et al., (2008). Microarray comparative genomic hybridization analysis of 59 patients with schizophrenia. J Hum Genet, 53(10): p. 914–919. Jiang, M., et al., (2008). Simultaneously detection of genomic and expression alterations in prostate cancer using cDNA microarray. Prostate, 68(14): p. 1496–1509. Staaf, J., et al., (2008). Detection and precise mapping of germline rearrangements in
BRCA1, BRCA2, MSH2, and
MLH1 using zoom-in array comparative genomic hybridization (aCGH). Hum Mutat, 29(4): p. 555–564. National Institutes of Health, (2001). Consensus conference 2000: adjuvant therapy for breast cancer. National Institutes of Health Consensus Development Conference Statement November 1–3, 2000. Cancer Control, 8(1): p. 55. Goldhirsch, A., et al., (2003). Meeting highlights: Updated international expert consensus
34
65. 66.
67.
68.
69.
70.
71.
72.
Sealfon and Chu on the primary therapy of early breast cancer. J Clin Oncol, 21(17): p. 3357–3365. van ’t Veer, L.J., et al., (2002). Gene expression profiling predicts clinical outcome of breast cancer. Nature, 415(6871): p. 530–536. van de Vijver, M.J., et al., (2002). A geneexpression signature as a predictor of survival in breast cancer. N Engl J Med, 347(25): p. 1999–2009. Wang, Y., et al., (2005). Gene-expression profiles to predict distant metastasis of lymph-nodenegative primary breast cancer. Lancet, 365(9460): p. 671–679. Foekens, J.A., et al., (2006). Multicenter validation of a gene expression-based prognostic signature in lymph node-negative primary breast cancer. J Clin Oncol, 24(11): p. 1665–1671. Chang, J.C., et al., (2003). Gene expression profiling for the prediction of therapeutic response to docetaxel in patients with breast cancer. Lancet, 362(9381): p. 362–369. Ma, X.-J., et al., (2004). A two-gene expression ratio predicts clinical outcome in breast cancer patients treated with tamoxifen. Cancer Cell, 5(6): p. 607–616. Ayers, M., et al., (2004). Gene expression profiles predict complete pathologic response to neoadjuvant paclitaxel and fluorouracil, doxorubicin, and cyclophosphamide chemotherapy in breast cancer. J Clin Oncol, 22(12): p. 2284–2293. Gonzalez-Neira, A., et al., (2007). Genomewide high-density SNP linkage analysis of nonBRCA1/2 breast cancer families identifies various candidate regions and has greater power
than microsatellite studies. BMC Genomics, 8(1): p. 299–. 73. Climent, J., et al., (2007). Characterization of breast cancer by array comparative genomic hybridization (1).(MINIREVIEW/ MINISYNTHESE)(Clinical report). Biochem Cell Biol, 85(4): p. 497–508. 74. Telenius, H., et al., (1992). Degenerate oligonucleotide-primed PCR: general amplification of target DNA by a single degenerate primer. Genomics, 13(3): p. 718–725. 75. Zhang, L., et al., (1992). Whole genome amplification from a single cell: implications for genetic analysis. Proc Natl Acad Sci USA, 89(13): p. 5847–5851. 76. Hosono, S., et al., (2003). Unbiased wholegenome amplification directly from clinical samples. Genome Res, 13(5): p. 954–964. 77. Paez, J.G., et al., (2004). Genome coverage and sequence fidelity of phi29 polymerase-based multiple strand displacement whole genome amplification. Nucleic Acids Res, 32(9): p. e71. 78. Brown, P.O., (2007). Ozone News [cited 1; Available from: http://cmgm.stanford.edu/ pbrown/protocols/OzoneNewsVol1.pdf. 79. Brown, P.O. Ozone News 2. 2007 [cited; Available from: http://cmgm.stanford. edu/pbrown/protocols/OzoneNewsVol2. pdf. 80. Fare, T.L., et al., (2003). Effects of atmospheric ozone on microarray data quality. Anal Chem, 75(17): p. 4672–4675. 81. Klass, D. and D. Hogan, (2007). The Ozone Annihilator [cited; Available from: http:// cmgm.stanford.edu/pbrown/protocols/ Ozone_Prevention.pdf.
Chapter 2 Aptamer Arrays Eva Baldrich Abstract In less than 40 years, aptamers have consolidated their role in biosensor development. Chemically related to nucleic acid probes, production of aptamers against targets of various sizes and compositions places them as ideal capture elements, alternative to more consolidated molecules such as antibodies. Thanks to their chemical simplicity and production, as well as their unique characteristics, aptamers have been successfully integrated in several innovative approaches. The incorporation of aptamers into the existing microarray technologies has lead to the reporting of various detection strategies, including direct fluorescence detection of fluorescent reporters, fluorescence anisotropy, FRET, SPR imaging, and electrochemical detection. Key words: Aptamer, Aptamer array, Aptasensor, Reagentless detection
1. Introduction 1.1. Aptamers: Description, Advantages, and Drawbacks
Aptamers are artificial nucleic acid ligands, selected in vitro from DNA/RNA random pools against specific nonnucleic acid targets. The reported aptamers have shown equal or higher affinity and specificity for their targets than their equivalent antibodies. Furthermore, aptamers have been selected against a vast variety of targets, including small molecules, drugs, peptides, and hormones, and also complex objectives such as proteins, spores, and whole cells, showing surprising versatility compared to other biorecognition components (1–6). The fact that aptamers are selected and produced in vitro eludes the use of animals and related ethical concerns, ensures no batch-to-batch variation, and allows selection under nonphysiological conditions and toward small molecules and toxins. The whole procedure is potentially automated and easier, quicker, and cheaper than antibody production. From an integration point of view, aptamers are
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_2, © Springer Science+Business Media, LLC 2011
35
36
Baldrich
smaller and less complex than antibodies (5–25 kDa vs. 150 kDa), and are easier to modify during or after synthesis, favoring immobilization and labeling. In addition, the unique chemical and structural characteristics of nucleic acids permit aptamer reversible denaturation and thus the design of truly reusable devices. The main concerns regarding the real applicability of aptamers, related to inherent nucleic acids properties such as sensitivity to nuclease attack and chemical simplicity, are being circumvented in the shape of spiegelmers and chemically modified aptamers. The fact that each single DNA/RNA sequence can adopt multiple conformation, reducing assay efficiency and increasing cost, can be also minimized by careful folding characterization and assay optimization. 1.2. Aptamer Arrays
The first reports on aptamer arrays exploited either optical fiber arrays (7, 8), or agarose beads deposited in the wells of micromachined flow chips (9). Later works, however, used glass slides to produce aptamer arrays in a variety of assay formats that, taking advantage of detection strategies already proved for DNA, permitted detection of an assortment of targets (Fig. 1). Gold and
a
c
b
d
Fig. 1. Scheme of different fluorescence aptameric array assay formats. (a) Target capture, followed by labeling with a reactive fluorophore. (b) Sandwich assay using aptamer and labeled antibody as capture and detector biocomponents. (c) Competition between native and labeled target variants. (d) Molecular beacon format exploiting the use of a labeled aptamer immobilized on surface; target binding induces changes in fluorophore emission.
Aptamer Arrays
37
coworkers introduced the use of photoaptamers that, remaining covalently bound to their targets after photoactivated cross-linking, allowed of highly stringent washes and efficient removal of nonspecific binders (10–12). In view of the enhanced signal-to-noise ratios and low detection limits registered, the SomaLogic team developed a 17-plex photoaptamer array on activated glass slides (10). Detection was based in a sandwich format, using either NH-reactive fluorophores or fluorophore-labeled antibodies, with detection limits below 10 fM for several analytes measured in 10% serum. Ellington et al. reported on a number of arrays manufactured using the lysozyme, ricin, IgE, and thrombin RNA/DNA aptamers on streptavidin slides (13–15). As a novelty, they defined a “universal buffer” in which the four aptamers retained acceptable affinity for the fluorophore-labeled analytes at concentrations over seven orders of magnitude (10–107 pg/mL). In a different approach, the team at the company Archemix immobilized fluorescein-labeled RNA/DNA aptamers on streptavidin slides (16). In this format, target binding is directly measured as changes in fluorescence polarization anisotropy in a completely reagentless format. The system detected and quantified four different proteins in the presence of serum and bacterial cell lysates. Alternatively, Lin self-assembled the aptamer into high-density nanoarrays, following modification with a fluorescent nucleotide analog near the target-binding site (17). Target binding generated increase in fluorescence, measurable by confocal fluorescence microscope imaging down to the low nanomolar range. Finally, Bera Aberem spotted onto silanized slides an aptamer, Cy3-labeled and hybridized with a chromic, cationic, water-soluble polythiophene (18). In the absence of target, the polymer quenched Cy3 emission; target binding induced polymer displacement and fluorescence increase. Alternative approaches include the development of surface plasmon resonance imaging (SPRi) and electrochemical aptamer arrays (19–21).
2. Materials 2.1. Physical Support
1. Microscope slides, modified so as to incorporate on surface reactive groups (amino, aldehyde, maleimide, thiol, and epoxy) or components (streptavidin and biotin) are provided, among others, by Xenopore Corp. (Hawthorne, NJ), Pierce– Thermo Fisher Scientific Inc. (Rockford, IL), Corning (Corning, NY), Nanocs Inc. (New York, NY), Nalge Nunc International (Rochester, NY), Genetix (New Milton, UK),
38
Baldrich
and Erie Scientific (Portsmouth, NH). In some cases, customized coated glass slides can be produced. 2. Streptavidin-coated agarose and silica beads can be obtained from Sigma-Aldrich and Pierce-Thermo Fisher Scientific Inc. (Rockford, IL). 3. Gold-coated slides and substrates can be purchased from a number or companies, including Aldrich, Platypus Technologies (Madison, WI), Phasis (Geneva; Switzerland), Nanocs Inc. (New York, NY), and Asylum Research (Santa Barbara, CA); and Gentel Biosciences, Inc. (Madison, WI) produces gold-coated substrates especially optimized for SPRi. 4. Electrodes of different geometries, composition, and complexity can be obtained from providers such as BASi (West Lafayette, IN), DropSens (Oviedo, Spain), BVT Technologies, a.s. (Brno, Czech Republic), Palm Instruments BV (Houten; The Netherlands), Applied BioPhysics (Troy, NY), and Princeton Applied Research (Oak Ridge, TN). 2.2. Aptamer Production and Manipulation
1. Once its sequence is known, an aptamer is produced by classical DNA/RNA synthesis. In the lack for synthesis facilities, the aptamer can be ordered to any company commercially providing oligonucleotides. In this case, double check the sequence ordered and ensure that appropriate spacers/linkers are being added to the extreme chosen for aptamer immobilization/modification (see Note 1). 2. For novel aptamers, at least three companies produce customized aptamers: RNA-tec (Leuven, Belgium), AptaRes (Luckenwalde, Germany), and Nascacell (Munchen, Germany). 3. Reconstitute lyophilized aptamers to 100–500 mM using sterile water or binding buffer. Store frozen in small aliquots in order to avoid repeated thaw-freeze cycles (see Note 2). 4. The most widely used binding buffers, which often correspond to the solutions employed in aptamer SELEX, are the following ones: – PBS (10 mM phosphate buffer, 138 mM NaCl, 2.7 mM KCl, pH 7.4). – PBS, 1–5 mM MgCl2. – 5 mM NaH2PO4, 5 mM KH2PO4, 2 mM MgCl2. – 10–100 mM Tris–HCl, pH 7.5, 50–150 mM NaCl, 0–5 mM MgCl2. – 20 mM Tris–acetate, pH 7.4, 140 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2. – 10–50 mM Hepes, pH 7.4, 0–150 mM NaCl, 0–5 mM KCl, 0–5 mM MgCl2.
Aptamer Arrays
39
5. When possible, sterile plasticware, pipette tips with filter, and separate pipettes (specific for work with DNA/RNA) should be used. If working with RNA aptamers, additional measures should be considered such as using RNAse-free material, treating all solutions and surfaces with an RNAse inhibitor (such as DEPC), and working in a space physically separated and labeled for work with RNA. 2.3. Aptamer Immobilization and Labeling
1. Use an aptamer produced or modified with an amino/thiol/ biotin group at one of the extremes, incorporating the appropriate spacer/linker (see Note 1). 2. Unless otherwise stated, use the following solutions. Binding buffer: sterile binding buffer of choice (see Subheading 2.2). Blocking buffer: binding buffer supplemented with 0.1% Tween (see Note 3). Washing buffer: binding buffer containing 0.05–0.1% Tween. 3. Amine-silane-based aptamer immobilization. Activation solution (1): 0.05 M dioxane solution of carbonyldiimidazole. Activation solution (2): 5% (v/v) glutaraldehyde in PBS, pH 7. Washing solution (1): dioxane and diethyl ether. Washing solution (2): PBS, pH 7. Chemical blocking: 1 M ethanolamine, pH 8.5, prepared in sterile water or binding buffer. 4. Thiol-silane-based aptamer immobilization. Chemical blocking: 0.1 M mercaptoethanol prepared in sterile water or binding buffer. 5. Self-assembly of thiolated aptamer on gold surfaces. Use 0.1 M in KH2PO4, pH 3.8. 6. Aptamer labeling with amine-reactive reagents. Reaction buffer: freshly prepared 0.1 M tetraborate buffer, pH 8.5 (see Note 4). Alternatively, 0.1 M sodium bicarbonate buffer pH 8–9 can be used (see Note 5). For labels/reagents insoluble in water, dissolve in dimethyl sulfoxide (DMSO) or in dimethyl formamide (DMF). Caution: DMF is a possible carcinogen and should be manipulated wearing protections. 7. Aptamer conjugation to COOH-bearing electroactive labels. Reaction buffer: 10 mM PBS or HEPES (4-(2-hydroxyethyl)1-piperazineethanesulfonic acid), pH 7.4. For labels/ reagents insoluble in water, dissolve in DMSO or in DMF. Caution: DMSO and DMF are possible carcinogens and should be manipulated wearing protections. Cross-linkers: 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide (EDC) and N-hydroxy succinimide (NHS). 8. Aptamer precipitation. 3 M NaCl, cold ethanol, and cold 70% (v/v) ethanol.
40
Baldrich
9. Aptamer modification by transcription (see Note 6). Dithiothreitol (DTT), each of the four unmodified ribonucleotides, modified ribonucleotide of choice, DNA-dependent RNA polymerase (e.g., SP6, T3, or T7), and RNAse-free DNAse. Transcription buffer (often provided 5×): 60 mM Tris–HCl, pH 8.0, 10 mM NaCl, 40 mM MgCl2, and 8.0 mM spermidine. 2.4. Target Labeling
1. For labels/reagents insoluble in water, dissolve in DMSO or in DMF. Caution: DMSO and DMF are possible carcinogens and should be manipulated wearing protections. 2. Reaction buffer for succinimidyl and STP esters: 0.1–0.2 M sodium bicarbonate buffer, pH 8.3 (see Note 5). 3. Reaction buffer for TFP esters, isothiocyanates, and sulfonyl chlorides: 0.1–0.2 M sodium bicarbonate buffer, pH 9.0 (see Note 5). 4. Stop reaction: freshly prepared 1.5 M hydroxylamine hydrochloride, pH 8.5. 5. Gel filtration/desalting columns: G25, P-10, D-Salt, or similar (provided by Pharmacia; Supelco-Sigma Aldrich; GE Healthcare; and Thermo Scientific Pierce among others). 6. Centrifuge-filter devices and spin columns: Microcon, VectaSpin, Centri-Spin, Centrex, Quick-Spin, or similar (provided by Millipore; Whatman; Princeton Separations; Aldrich; Roche Applied Science; and GE Healthcare).
2.5. Regeneration Solutions
Table 1 summarizes the most widely used regeneration solutions.
3. Methods Aptamer immobilization for microarray production can take advantage to certain extent of the fabrication and spotting techniques already developed for the more consolidated DNA microarrays. However, as target binding absolutely depends on aptamer folding into the appropriate conformation, aptamers performance can be impaired following immobilization. For this reason, using long spacers is essential for optimal aptamer performance, with length and chemical composition of the spacer strongly influencing detectability (see Note 1). Even if aptamers, modified with the appropriate reactive groups, can be successfully conjugated on silanized glass/silica or self-assembled on gold-modified slides, most authors chose using streptavidin-coated surfaces that provide better blocking against nonspecific adsorption and a much faster and easier aptamer immobilization procedure. For this purpose, the aptamer can be
Aptamer Arrays
41
Table 1 Aptamer regeneration solutions reported by several authors Reagent type
Regeneration solution
Target type
Buffer
Binding buffer Water
Small molecules Small molecules
Temperature
Hot water, 80°C
Small molecules
Concentrated salt
2 M NaCl
Protein
pH
10–100 mM NaOH 12 mM NaOH, 1.2% EtOH 0.2 M Glycine–HCl, pH 2.2 0.1 M HCl
Protein Protein Protein Protein
Chaotropic
7–8 M Urea 5–6 M Guanidinium hydrochloride
Protein Protein
Chelating
50 mM EDTA 10 mM EGTA
Protein Small molecules
Detergent
0.03–10% SDS
Protein
Mixture
0.1 M Sodium citrate, 10 mM EDTA, 7 M Urea, pH 5.0 1 M NaCl in 0.02% Triton X-100 7 M Urea in 25 mM Tris–HCl for 3 min at 70°C
Protein and small molecules Protein Protein
obtained commercially biotinylated, or chemically or enzymatically biotinylated in-home. Although polylysine slides work well for the immobilization of oligonucleotide probes and nucleic acid detection, they have been reported to be unsuitable for aptamer immobilization. This is likely to be due to aptamer unfolding following electrostatic interaction between its negatively charged phosphodiester backbone and positively charged polylysines. Assay performance, on the other hand, strongly depends on the binding buffer of choice. In this respect, the unique characteristics of aptamers make it difficult to define universal assay conditions. For this reason, most authors just use a buffer similar to that used for aptamer selection. Nevertheless, it has been demonstrated that buffer optimization can generate improved results or facilitate assay performance. Once an aptamer optimal binding buffer is defined, all the assay steps, from aptamer immobilization to target detection, should be carried in that buffer. Fluorescence aptamer arrays mostly exploit competition, often using a suboptimal fluorophore-labeled target, or sandwich assay formats (Fig. 1). Alternatively, the molecular beacon format takes advantage of an immobilized aptamer whose free extreme has been modified with a fluorophore. In this case, aptamer rearrangement induced by target binding translates into a change
42
Baldrich
of either fluorescent emission or fluorescence polarization anisotropy. SPRi and impedance electrochemical arrays, on the other hand, are based in truly reagentless formats, in which target binding directly translates into changes in signal transduction (Fig. 2a, b). a Light source
Angle shift
F1
F1
1
Intensity
Prism
F2
F2
2
12
FeCN
e–
e–
10
Zim/Ω (x102)
FeCN
8
2
14 12 10 8 6
6 4
16
0
10 20 30 Target conc.
40
[Target]
b
Rct /Ω (X102)
Angle (F)
0 0
4
8
12
16
20
Zre/Ω (x102)
c
e–
300
[Target]
Current/nA
400
200
100
0 –0.3 –0.2 –0.1 0.0
0.1
0.2
0.3
0.4
0.5
0.6
E/V vs. Ag-AgCl
Fig. 2. Scheme of different SPRi and electrochemical aptameric array assay formats. In SPRi (a) and impedance sensing (b), target capture translates into direct label-less signal transduction. (c) Among the various electrochemical detection strategies, the molecular beacon format exploits the use of an aptamer immobilized on surface, which bears on its free extreme an electroactive label. Target binding correlates with a change in distance between surface and label, and thus change in electron transfer rate, in a reagentless assay format.
Aptamer Arrays
43
Other electrochemical sensing strategies, such as the amperometric or voltametric approaches, remind their fluorescent counterparts, with the exception that electroactive labels are used in place of classical fluorophores (Fig. 2c). In these specific cases, aptamer self-assembly onto gold electrodes is preferred to streptavidin capture. In the following sections, we summarize the most used protocols for aptamer immobilization and modification, and for the production of labeled targets. We will finish with a discussion on optimization of binding conditions and sensor regeneration. 3.1. Glass Slide Pretreatment and Silanization
1. Sequentially sonicate the glass/silica slide for 5 min in chloroform, 5 min in acetone, and five more minutes in isopropyl alcohol. 2. Extensively rinse with sterilized water and sonicate for 1 h in 3 M NaOH. 3. Thoroughly rinse with sterile water and dry under a nitrogen flow. 4. Dissolve the silane of choice in ethanol or methanol to final concentration 0.5% (v/v). 5. Silanization can be carried out by immersion of the glass surface into the silane solution for 15–60 min at room temperature. Alternatively, drop the solution on a compartment, place the slide on top, without touching the liquid with the aid of a physical support, and incubate overnight at room temperature for vapor deposition. 6. Rinse with ethanol or methanol, dry, and bake for 15 min at 110°C.
3.2. Pretreatment of Gold-Coated Surfaces and Gold Electrodes 3.2.1. Gold-Coated Slides
1. Gold-coated slides have to be handled carefully as to avoid scratching/damaging the gold cover. Wear gloves as not to leave fingerprints. 2. Although the best option is to seek for the provider’s advice, high-quality slides are usually sufficiently cleaned by wiping the surface with an optical wipe, cotton-tipped swab, or a smooth brush moistened in high-purity alcohol. 3. If the facilities are available, the slides can be treated with UV/ozone for 10–25 min, or by oxygen or argon plasma etching (time to be experimentally determined), followed by ultrasonic washing for 1 min in isopropanol and/or extensive washing with water. 4. Reports exist on treatment of gold-coated surfaces with piranha (3/1 v/v H2SO4/H2O2. Caution!) or chromic acid for up to 30 min, but some substrates do not survive in good shape in these harsh conditions and can present gold detaching.
44
Baldrich
3.2.2. Electrodes
1. In addition to cleaning (as described in the previous section) to remove organic, inorganic, and ionic species not native to the surface, electrodes are often electrochemically activated as to promote more uniform metallic surfaces (e.g., to reduce the concentration/thickness of oxide on gold and/or remove adsorbed components). 2. Macroelectrodes (such as disk electrodes) are regularly cleaned by mechanical polishing with diamond, alumina, or silicon carbide of different particle sizes (BASi supplies 0.05 mm alumina polish and 1, 3, 6, and 15 mm diamond polishes), followed by ultrasonic washing with alcohol or water according to the provider’s instructions. 3. Most electrodes can be electrochemically activated by cyclic voltammetry. For gold screen-printed electrodes, cycle between 0 and +1.4 V at scan rate 100 mV/s in 0.1 M H2SO4 until a sharp stable peak of surface gold oxide reduction is obtained (10–20 cycles in most cases). For microfabricated electrodes, cycle the potential from 0.8 to −1.5 V (vs. Ag/ AgCl) in 0.1 M KCl until reproducible voltammograms are obtained. Hydrogen evolution at the cathode and oxygen generation at the anode accompany cleaning. 4. The electrode state can be traced by cyclic voltammetry in 1 mM ferrocyanide, 0.1 M KCl at the beginning, and at the end of each experiment.
3.3. Aptamer Immobilization (See Note 7) 3.3.1. Immobilization of Biotinylated Aptamer on Streptavidin-Coated Glass Slides
1. Wash the streptavidin-coated slide 2–3 times with binding buffer. 2. Dissolve the biotinylated aptamer in binding buffer and spot onto the surface. Aptamer concentration should be optimized in each case, but concentrations in the range of 100 nM to 1 mM will work in most cases. 3. Incubate for 30 min at room temperature (see Note 8). 4. Wash three times with washing buffer and store at 4°C in binding buffer (see Note 9).
3.3.2. Covalent Conjugation of NH-Aptamer to Amino-Silanized Surfaces
1. Activate the amine-modified slides by immersion for 1 h in either (1) a 0.05 M dioxane solution of carbonyldiimidazole or (2) 5% glutaraldehyde. 2. Wash with (1) dioxane and diethyl ether or (2) PBS, and dry under a nitrogen stream. 3. Spot the amino-modified aptamer, dissolved to 1–10 mM in binding buffer, and incubate for 30–60 min at 37°C or overnight at room temperature (see Note 8). 4. Rinse with binding buffer and block any remaining active groups by incubating in 1 M ethanolamine for 15 min (see Note 9).
Aptamer Arrays
45
5. Rinse with washing buffer and store at 4°C until used, either immersed in binding buffer, or dried under an argon or nitrogen stream. 3.3.3. Covalent Conjugation of SH-Aptamer to Thiol- or Maleimide-Silanized Surfaces
1. Spot the thiol-modified aptamer, dissolved in binding buffer to a concentration of 0.1–1 mM. 2. Incubate for 30–60 min at 37°C or overnight at room temperature (see Note 8). 3. Rinse with binding buffer and block any remaining active groups by incubating in 0.1 M mercaptoethanol for 15 min (see Note 9). 4. Rinse extensively with water and washing buffer, and incubate with binding buffer for at least 15 min before use.
3.3.4. Self-Assembly of SH-Aptamers onto Gold Surfaces
1. Spot the thiol-modified aptamer, containing a spacer of the appropriate length (to be optimized), dissolved in KH2PO4 buffer to a concentration of 0.1–1 mM. 2. Incubate for 60 min at room temperature (see Note 8). 3. Rinse with sterile water. 4. Spot a thiolated molecule of choice whose chain length has to be shorter than the aptamer spacer (e.g., mercaptohexanol has been widely used for this purpose), dissolved to 0.1 M in KH2PO4, or in ethanol depending on its solubility. This molecule will fill any pin holes existing on the aptamer cover, and will displace loosely bound molecules. Incubate for 60 min at room temperature. 5. Rinse extensively with water and washing buffer, and incubate with binding buffer for at least 15 min before use.
3.4. Aptamer Modification by Label/ Biotin Incorporation 3.4.1. Incorporation of Commercial Amine-Reactive Compounds to Amine-Modified Aptamer
1. This protocol is appropriate for the most commonly used amine-reactive reagents (e.g., fluorophore, biotin): sulfosuccinimidyl esters, isothiocyanates, sulfonyl chlorides, and tetrafluorophenyl esters (see Note 10). Label quantity has been calculated for an aptamer of 15–25 nucleotides. For optimal results, the most favorable aptamer:label molar ratio has to be experimentally determined by assaying 3–4 different ratios. 2. Dissolve 100 mg of the aptamer in 0.1 M tetraborate buffer (see Note 5). The final volume should be below 100 ml. If necessary, precipitate and redissolve the aptamer. 3. Immediately before use, dissolve the amine-reactive label in DMSO to a final concentration of 20 mg/mL, and vortex until completely dissolved. 4. While gently stirring, add drop-to-drop 10 ml of the fluorophore to each 100 mg of aptamer (the concentration of DMSO must not exceed 10% in the final reaction).
46
Baldrich
5. Incubate for 4–6 h at room temperature, in agitation, and protected from light. 6. Purify the modified aptamer by ethanol precipitation as follows. Add one-tenth volume of 3 M NaCl and mix. Add two and a half volumes of cold absolute ethanol. Mix again and incubate at −20°C for 30 min. Vortex for 1–2 min as to wash away any molecules attached to the tube walls. Centrifuge at ~12,000 × g for 30 min, using a refrigerated centrifuge, if possible. Carefully remove the supernatant, rinse the pellet twice with two volumes of cold 70% ethanol, and dry briefly avoiding that the labeled aptamer becomes completely dry. Dissolve in sterile water or binding buffer, aliquot, and freeze. 7. Alternatively, the labeled aptamer can be purified by reversephase HPLC or by gel electrophoresis. 3.4.2. Incorporation of COOH-Bearing Electroactive Labels to Amine-Modified Aptamer
1. This protocol is appropriate for most molecules exhibiting COOH-groups (e.g., ferrocene carboxylic acid). Label quantity has been calculated for an aptamer of 15–25 nucleotides. For optimal results, the most favorable aptamer:label molar ratio has to be experimentally determined by assaying 3–4 different ratios. 2. Dissolve 100 mg of the aptamer in 10 mM PBS or HEPES, pH 7.4, to a final volume of 100 ml. If necessary, precipitate and redissolve the aptamer. 3. Immediately before use, dissolve the label in DMSO (or DMF) to a final concentration of 200 mM and vortex until completely dissolved. 4. While gently stirring, add 1 ml of the label to each 100 mg of aptamer (the label will be at a final concentration of approximately 2 mM, equivalent to a tenfold molar excess vs. the aptamer. DMSO/DMF final concentration must not exceed 10%). 5. Add a 100-fold molar excess of EDC (e.g., 1 ml of a stock prepared by dissolving 3.82 mg in 200 ml PBS or HEPES) and NHS to a final concentration 5 mM (e.g., 2.5 ml of a 200 mM stock prepared by dissolving 2.3 mg in 100 ml PBS or HEPES). 6. Incubate for 2 h at room temperature, in agitation, and protected from light. 7. Purify the modified aptamer by ethanol precipitation as follows. Add one-tenth volume of 3 M NaCl and mix. Add two and a half volumes of cold absolute ethanol. Mix again and incubate at −20°C for 30 min. Vortex for 1–2 min as to wash away any molecules attached to the tube walls. Centrifuge at ~12,000 × g for 30 min, using a refrigerated centrifuge, if possible.
Aptamer Arrays
47
Carefully remove the supernatant, rinse the pellet twice with two volumes of cold 70% ethanol, and dry briefly avoiding that the labeled aptamer becomes completely dry. Dissolve in sterile water or binding buffer, aliquot, and freeze. 8. Alternatively, the labeled aptamer can be purified by reversephase HPLC or by gel electrophoresis. 3.4.3. RNA Aptamer Modification by Incorporation During Transcription
1. Use a commercially modified nucleotide (e.g., biotinylated, fluorophore-labeled) and, if required, deprotect following the provider’s instructions. 2. Unless otherwise advised by the provider, prepare 20 ml reactions containing 800 ng of the template DNA (see Note 6), 30 mM DTT, 5 mM of each of the three unmodified nucleotides, 2 and 3 mM of native and biotinylated species for the modified nucleotide, and 2 U of T7 RNA polymerase, in transcription buffer. 3. Incubate for 4 h at 37°C. 4. Treat with 1 ml of RNAse-free DNAse for 15–30 min at 37°C, or according to the provider’s instructions. 5. Purify by chromatography, gel electrophoresis, or ethanol precipitation as described above. Dissolve in sterile binding buffer containing RNAse inhibitor, aliquot, and freeze.
3.5. Target Labeling 3.5.1. Target Labeling with an Amine-Reactive Fluorophore
1. This protocol is appropriate to label amine-containing proteins using the most common amine-reactive reagents: sulfosuccinimidyl esters, isothiocyanates, sulfonyl chlorides, and tetrafluorophenyl esters (see Note 10). For smaller targets, specific protocols will have to be searched. 2. Dissolve the protein target in 0.1 M sodium bicarbonate buffer, pH 8–9, to a final concentration 2–20 mg/mL (see Notes 5 and 11). 3. Immediately before use, dissolve the amine-reactive fluorophore in DMSO to 10 mg/mL and vortex until complete dissolution. 4. While vortexing the protein solution, add drop-by-drop 5–10 ml (0.05–0.1 mg) of reactive fluorophore per mg of reacted target protein (the concentration of DMSO must not exceed 10% in the final reaction). In general, it is recommended to assay in parallel at least three different protein:fluorophore molar rates. 5. Incubate for 1 h at room temperature with continuous stirring and protected from light. For sulfonyl chlorides, incubate at 4°C instead. 6. The reaction can be optionally stopped by adding 0.1 mL of freshly prepared 1.5 M hydroxylamine per mL of reaction, and incubating for 1 h at room temperature.
48
Baldrich
7. Depending on the target size, the conjugate can be recovered through a gel filtration column using the buffer of choice, and concentrated using a centrifuge filter/spin device. Alternatively, the conjugate can be dialyzed or purified by column chromatography. 8. Determine the degree of fluorescence labeling according to the provider’s protocol. 9. Add bovine serum albumin (BSA) to a final concentration of 1–10 mg/mL, or any other stabilizer of choice, and store at 4°C. For extended storage, either add glycerol up to 50% or distribute in small aliquots and freeze at −20°C. 3.6. Optimal Conditions for Target Capture
1. Every single aptamer molecule can fold into different structures, but only one or few of them exhibit on surface the binding pockets or clefts for the specific recognition of the target (see Note 7). Aptamer folding will thus be affected by a variety of external factors, such as incubation temperature and buffer composition, and will often depend on the presence of certain ions. For example, aptamer folding into hairpin and stem-andloop structures usually requires the presence of magnesium ions, and quadruplex formation is favored and stabilized by the binding of cations such as sodium or potassium. 2. This implies that changes in the assay binding conditions (buffer ionic composition or strength, assay temperature, etc.) can affect aptamer folding, affinity for the target, and assay performance. This is of extreme importance when, for example, an aptamer evaluated in vitro is to be used in real samples of different ion concentrations. 3. For these reasons, aptamers are regularly assayed under similar conditions as those employed in their selection procedure (normally available in the literature and summarized in the methods section). Nevertheless, some works indicate that aptamer optimal working conditions are not necessarily the SELEX ones, especially if a multiplexed assay is to be optimized. In these cases, series of optimization experiments should be performed for each aptamer or structure type. In any case, and contrary to what happens with immuno and nucleic acids arrays, it is difficult to define universal conditions in the case of the aptamer arrays. 4. In the same way, optimal binding time has to be experimentally optimized for each aptamer-target pair and experimental setup, with authors having reported incubations from just few minutes in flow systems to up to 2 h in bulk experiments. 5. Although most aptamers perform well at room temperature in the appropriate binding buffer, in some few cases higher temperatures (i.e., 37°C) are required. Nonetheless,
Aptamer Arrays
49
higher temperatures usually contribute to increase the level of nontarget nonspecific adsorption, and complicate assay setup. 6. When required, nonspecific adsorption can be decreased by adding 0.05–0.1% of Tween to the binding buffer. 7. In all cases, incubations have to be followed by extensive washing (e.g., 10 min flowing), or at least three serial washes with detergent-containing washing buffer. 3.7. Array Regeneration and Reutilization
1. Compared to antibody-based immunosensors, which are difficult to regenerate preserving their properties and performance, aptasensors are potentially regenerated. Being nucleic acids, aptamers can be submitted to repeated cycles of denaturation and renaturation without seriously damaging their structure, and thus performance. 2. Wash the array with washing buffer, followed by a wash with binding buffer. 3. The binding of small targets to an aptamer, for example drugs and small molecules, is highly reversible. So, it is possible to regenerate these aptasensors by just washing extensively with water, washing buffer, or binding buffer, until the target has been completely removed. 4. In the case of more complex targets, the various strategies reported for aptamer regeneration are mainly directed toward disruption of the aptamer folding, disturbance of the aptamer– target interaction, and/or target denaturation (see Table 1). They range from the use of high temperature (to denature aptamer folding), to the use of concentrated salt solutions (to disrupt binding interactions and denature aptamer folding), surfactants (to wash away the analyte), chaotropic components (to disrupt noncovalent molecular interactions), or chelating agents (to remove metals responsible for aptamer secondary or tertiary structure). Even if extreme pHs should be avoided, because they can damage nucleic acids, successful results have been reported for the use of diluted NaOH, glycine, and HCl, which are believed to disrupt noncovalent interactions of analytes bound to aptasensors. Nevertheless, the choice of the regeneration procedure will be conditioned by the aptamer immobilization strategy (i.e., it is meaningless to regenerate the aptamer with a component that damages the streptavidin–BSA cover used for its immobilization) and by target nature. 5. Regeneration should always be followed by a final rinse with washing and binding buffer as to completely eliminate the regeneration agent.
50
Baldrich
3.8. Detection Strategy 3.8.1. Fluorescence Aptamer Microarrays
1. Fluorescence aptamer arrays exploit the use of fluorescent labels such as fluorophores and, more recently, quantum dots. 2. In the simplest approach, captured target is treated with a reactive fluorescent component, and thus labeled (Fig. 1a). For this system to work, the target has to present the reactive groups targeted by the label, and the aptasensing surface has to be 100% specific (as any nontarget components nonspecifically trapped onto the surface will be also labeled). 3. In most cases, competition assay formats are developed, in which a fluorescently labeled target competes in solution with the native target (i.e., present in the samples) for the immobilized aptamer (Fig. 1b). As an alternative, labeled aptamer can be used if competition occurs between native target in solution and modified target immobilized on the chip for the aptamer in solution. For competition optimization, several concentrations of the labeled component are assayed. The concentration generating 70–80% of the maximal signal registered is considered optimal for competition performance. In this respect, changing the concentrations of immobilized and labeled biocomponents allows the assay linear range to be shifted if required. 4. The sandwich assay formats usually provide extremely high sensitivity (Fig. 1c). However, few aptamers present two of more binding sites per target unit, and two or more aptamers have been seldom reported against different epitopes in the same target. Consistently, sandwich microarrays often take advantage of mixed aptamer/antibody formats. Using aptamer and antibody as capture and detection biocomponent, respectively, offers the additional advantage that the latter, being much bigger, can incorporate a higher number of label units, and thus generate “amplified” signals. 5. The unique characteristics of aptamers allow optimization of an alternative assay format using molecular beacons. This format takes advantage of an immobilized aptamer whose free extreme has been modified with a fluorophore (Fig. 1d). In this case, aptamer rearrangement induced by target binding translates into a change of either fluorescent emission or fluorescence polarization anisotropy.
3.8.2. Surface Plasmon Resonance Imaging
1. SPR is an optical label-free detection strategy based on the fact that, under certain conditions, part of a polarized light beam hitting a gold-coated prism is absorbed by the free electrons at the metal surface. The photons absorbed are in this way converted into surface plasmon waves, a phenomenon traceable in real time in the shape of a dip in light reflectivity.
Aptamer Arrays
51
Adjustments that occur at the biochip gold surface, such as those generated by target capture, will directly induce a change of resonance, and thus, a modification in light reflectivity. 2. In the case of SPRi, the whole biochip surface is visualized using a video camera. This enables the researcher to functionalize the biochip in an array format (e.g., with different aptamers) and to obtain data separately and simultaneously from the different spots. SPRi provides data in real time and does not require more biocomponent modification/labeling than aptamer immobilization (Fig. 2a). 3.8.3. Electrochemical Aptamer Microarrays
1. Target capture onto the aptasensor surface induces its partial blocking and thus increase of the interfacial electron-transfer resistance. This effect can be monitored in real time by impedance spectroscopy, without the need to use labels or labeled biocomponents (Fig. 2b). Nevertheless, several authors have described sandwich formats that, making use of nanoparticle or enzyme-labeled biocomponents, contribute to the increase in impedance signals registered in an important way. 2. Other electrochemical sensing strategies, such as the amperometric or voltametric approaches, remind their fluorescent counterparts, with the exception that fluorophores are substituted by electroactive labels such as redox mediators or metal nanoparticles. In this case, target capture correlates with changes in electron transfer rate or efficiency (Fig. 2c). 3. Electrode apta functionalization is mostly based on aptamer self-assembly onto gold instead of using streptavidin capture. In these cases, the length of the spacers used during aptamer immobilization has to be carefully optimized so as to provide surface specificity against nonspecific adsorption and efficient electron transfer.
4. Notes 1. Longer spacers generate better aptamer performance, in terms of signal amplitude, reproducibility, and selectivity, except for electrochemical assay formats in which the use of long spacers interferes with electron transfer. Spacers of DNA, hydrocarbon-based, and oligo(ethylene oxide) composition have been successfully used. With few exceptions, aptamer immobilization through its two extremes generates very similar results. It is thus recommended to order 3¢ modifications, significantly cheaper than 5¢ changes.
52
Baldrich
2. TE buffer (10–50 mM Tris–HCl, pH 8.0, 1 mM EDTA) can also be used, where EDTA contributes to inhibit nuclease activity. However, ion quelation interferes in aptamer folding, and buffer exchange might be necessary previous to any experiment. 3. Although some authors describe the use of BSA as a blocking agent, according to our experience, this molecule is too big for most aptamers. It is appropriate, on the other hand, for blocking the surface previous to aptamer capture (e.g., streptavidin-coated surfaces). 4. Tetraborate buffer can be aliquoted and frozen, but each aliquot should be thawed and used only once. If the frozen buffer precipitates, vortex until completely resuspended. 5. Avoid buffers containing primary amines, such as Tris, that might compete for conjugation with the amine-reactive compound. 6. RNA aptamer production and/or modification by transcription is carried out from an appropriate double stranded DNA template that incorporates a promoter for the RNA polymerase used. The reaction should be carried out under conditions that exclude contamination with RNases. All plasticware and solutions should thus be nuclease free. 7. The requirement for a denaturation step previous to aptamer immobilization or assay performance, in order to improve aptamer proper folding, is unclear. Among the few existing studies, some indicate that denaturing may be unnecessary or deleterious for certain aptamers, as it is the case of the quadruplex-folding 15-nucleotide thrombin-binding aptamer, while others demonstrate that aptamers submitted to denaturation show slightly better limits of detection and reproducibility, as happens for the hairpin-forming aptamer against HIV Tat protein. Denaturation can be carried out by incubating for 3–5 min at 70 and 90°C for single-strain RNA and DNA aptamers, respectively, followed by gradual cooling to 4°C over 1–2 min. Alternatively, aptamers can be denatured by washing in 7–8 M urea, followed by refolding in the appropriate binding buffer. 8. In order to avoid the spots drying during incubation, place the surface into a humid chamber (a closed box or tube with a humid towel inside will serve). In some cases, especially when using highly saline buffers, addition of glycerol up to a final 1–10% concentration can help to prevent desiccation. 9. Optionally, the slides can be additionally blocked to reduce nonspecific adsorption by incubating for 1 h at RT in blocking buffer. In this case, add 0.05–0.1% Tween to all buffers subsequently used.
Aptamer Arrays
53
10. Tetrafluorophenyl and succinimidyl esters are preferred because they form very stable amide bonds (TFPs are less susceptible to hydrolysis and provide better conjugation ratios in aqueous media). Isothiocyanates are also commonly used, but the resulting thiourea product has been reported to deteriorate over time. Sulfonyl chlorides, on the other hand, are more reactive and may conjugate also to aromatic amines, and are more difficult to manipulate. 11. Amine-reactive fluorophores react with nonprotonated aliphatic amine groups, including protein terminal amine and e-amino groups in lysines. While the pKa value of the latter is around 10.5, the terminal amine is lower. Using a slightly basic buffer will thus target the reaction toward the two groups that are kept protonated under these conditions. Using a buffer closer to neutral instead will favor labeling of the amine terminal.
Acknowledgments The author is supported by an I3P fellowship from the Consejo Superior de Investigaciones Científicas (CSIC, Spain). References 1. Baldrich E, Campàs Homs M, O’Sullivan CK. Aptamers: Powerful molecular tools for therapeutics and diagnostics. In: Rapley R, Harbron S, eds. Molecular Analysis and Genome Discovery. London: John Wiley & Sons, Ltd; 2004:193–217. 2. Baldrich E, O’Sullivan CK. Aptamers as analytical tools. In: Grimes CA, Dickey EC, Pishko MV, eds. Encyclopedia of Sensors. Pennsylvania, USA: American Scientific Publishers; 2006: 275–295. 3. Hesselberth J, Robertson MP, Jhaveri S, Ellington AD. In vitro selection of nucleic acids for diagnostic applications. Rev Mol Biotechnol 2000;74:15–25. 4. James WC. Aptamers. In: Meyers RA, ed. Encyclopedia of Analytical Chemistry. Chichester: John Wiley & Sons, Ltd; 2001: 4848–4871. 5. Jayasena SD. Aptamers, an emerging class of molecules that rival antibodies in diagnostics. Clin Chem 1999;45:1628–1650. 6. Osborne SE, Matsumura I, Ellington AD. Aptamers as therapeutic and diagnostic reagents: Problems and prospects. Curr Opin Chem Biol 1997;1:5–9.
7. Kleinjung F, Klussmann S, Erdmann VA, Scheller FW, Fürste JP. High-affinity RNA as a recognition element in a biosensor. Anal Chem 1998;70:328–331. 8. Lee M, Walt DR. A fiber-optic microarray biosensor using aptamers as receptors. Anal Biochem 2000;282:142–146. 9. Kirby R, Cho EJ, Gehrke B, et al. Aptamer-based sensor arrays for the detection and quantitation of proteins. Anal Chem 2004;76:4066–4075. 10. Bock C, Coleman M, Collins B, et al. Photoaptamer arrays applied to multiplexed proteomic analysis. Proteomics 2004;4:609–618. 11. Golden MC, Collins BD, Willis MC, Koch TH. Diagnostic potential of photoSELEXevolved ssDNA aptamers. J Biotechnol 2000;81:167–178. 12. Petach H, Gold L. Dimensionality is the issue: Use of photoaptamers in protein microarrays. Curr Opin Biotechnol 2002;13:309–314. 13. Cho EJ, Collett JR, Szafranska AE, Ellington AD. Optimization of aptamer microarray technology for multiple protein targets. Anal Chim Acta 2006;564:82–90. 14. Collett JR, Cho EJ, Lee JF, et al. Functional RNA microarrays for high-throughput screening
54
Baldrich
of antiprotein aptamers. Anal Biochem 2005; 338:113–123. 15. Collett JR, Eun JC, Ellington AD. Production and processing of aptamer microarrays. Methods 2005;37:4–15. 16. McCauley TG, Hamaguchi N, Stanton M. Aptamer-based biosensor arrays for detection and quantification of biological macromolecules. Anal Biochem 2003;319: 244–250. 17. Lin C, Katilius E, Liu Y, Zhang J, Yan H. Selfassembled signaling aptamer DNA arrays for protein detection. Angew Chem Int Ed Engl 2006;45:5296–5301. 18. Bérn Abérem M, Najari A, Ho A-A, et al. Protein detecting arrays based on cationic
polythiophene-DNA-aptamer complexes. Adv Mater 2006;18:2703–2707. 19. Li Y, Hye JL, Corn RM. Detection of protein biomarkers using RNA aptamer microarrays and enzymatically amplified surface plasmon resonance imaging. Anal Chem 2007;79: 1082–1088. 20. Li Y, Lee HJ, Corn RM. Fabrication and characterization of RNA aptamer microarrays for the study of protein-aptamer interactions with SPR imaging. Nucleic Acids Res 2006;34(22): 6416–6424. 21. Xu D, Xu D, Yu X, Liu Z, He W, Ma Z. Labelfree electrochemical detection for aptamerbased array electrode. Anal Chem 2005;77: 5107–5113.
Chapter 3 Oligonucleotide Microarrays for Identification of Microbial Pathogens and Detection of Their Virulence-Associated or Drug-Resistance Determinants Dmitriy V. Volokhov, Hyesuk Kong, Keith Herold, Vladimir E. Chizhikov, and Avraham Rasooly Abstract Microarrays are spatially ordered arrays with ligands chemically immobilized in discrete spots on a solid matrix, usually a microscope slide. Microarrays are a high-throughput large-scale screening system enabling simultaneous identification of a large number of labeled target molecules (up to several hundred thousand) that bind specifically to the immobilized ligands of the array. DNA microarrays represent a promising tool for clinical, environmental, and industrial microbiology since the technology allows relatively rapid identification of large number of genetic determinants simultaneously, providing detailed genomic level information regarding the pathogen species, including identification of their virulenceassociated factors and the presence of antibiotic resistance genes. In this chapter, we describe key aspects and methodologies important for the development and use of DNA microarrays for microbial diagnostics. Key words: Microarray, Microbial pathogens, Virulence factors, Food safety
1. Introduction Microarrays are promising tool for large-scale genomic (and other biomolecular) analysis allowing relatively rapid and simultaneous identification of a large number of genetic determinants. The origin of microarray technology can be traced to work in the 1960s, which established the basic elements of DNA microarray technology: nucleic acid hybridization and chemical immobilization of nucleic acids to solid matrixes (DNA-bound membranes, glass slides) (1–3), different technologies for immobilization of capturing ligands and their detection on solid matrixes, and the development Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_3, © Springer Science+Business Media, LLC 2011
55
56
Volokhov et al.
of various immunological binding assays (4, 5). These methodologies are, in many ways, low-density “array” technologies. The current microarray technology emerged in the late 1980s, expanding the basic approach of ELISA/dot-blot into high-density arrays by reducing the size of the array elements, increasing the number of elements of the array, and automating ligand binding to the solid matrix (6–8). These developments lead to introduction of the concepts of “multi-analyte” immunoassays and “microspots” distributed on a solid support detected by fluorescence or enzymatic labeling. There are several array technologies, which share several features: multitarget analysis, specific binding or hybridization of the target, and labeling of the target molecules. These technologies all have similar main steps in designing and implementing DNA microarrays, which are (1) probe development, (2) array fabrication, (3) sample preparation, (4) assay, (5) detection, and (6) data analysis. These steps are shown in Fig. 1. 1.1. Applications of Microarray Technology
DNA microarray technologies are used today mainly for analysis of gene expression and for genotyping (genomic analysis) enabling simultaneous analysis of a large number of molecular markers. Microarray analysis of microbial genomics has several practical applications including medical diagnostics, food safety assays, drug safety assays for vaccine analysis, medical device testing, and microbial security applications.
1.1.1. Analysis of Gene Expression
The original and among the most commonly used applications of microarray technology is the study of gene expression, comparing Oligonucleotide Design Expression: cDNA
Genomic DNA
Genomic: Target amplification
Target Labeling
Array Printing
Hybridization Scanning
Data analysis
Fig. 1. Development and application of microarrays for microbial analysis. The key steps for development and application of microarrays for microbial analysis are oligonucleotide design for array spotting and target amplification (if PCR is used). Array spotting and target labeling (cDNA, DNA, or PCR amplicons). Hybridization, scanning, and data analysis.
Microarrays Detection of Bacteria
Tissue A
57
Tissue B RNA Extraction mRNA purification cDNA synthesis
Cy3
Cy5
cDNA Labeling Hybridization Washing Imaging and data analysis
Fig. 2. Microarray differential gene expression experiment steps. Cellular extracts from two tissues to be analyzed are processed to obtain total RNA followed by mRNA purification; the mRNA is converted to cDNA by reverse transcriptase. The cDNA from each sample is labeled with a fluorescent dye unique to that sample (e.g., Cy3 and Cy5). The two labeled cDNAs are then mixed and hybridized to the microarray. After washing, the array is scanned using excitation for both dyes, and data analysis of the fluorescent signals is performed. The relative intensity of each of the dyes for each of the microarray spots is measured, and the relative abundance of each mRNA is calculated.
and quantifying the relative abundance of mRNA between two (or more) samples (Fig. 2). In such experiments, cells from each sample are processed to obtain mRNA, which is converted to cDNA by reverse transcriptase. Each cDNA sample is then labeled with a specific fluorescent dye (e.g., Cy3 and Cy5). The two labeled cDNA samples are then mixed and hybridized to the microarray. After washing to remove the nonspecific bound cDNA, the array is scanned usually by a laser scanner, which excites each of the dyes. The user then measures the fluorescent signal of each dye from each spot. The array scanning data is analyzed to calculate and compare the relative intensity of each of the dye to assess the relative abundance of each mRNA. The data can also be used in cluster analysis to determine patterns of gene expression of the sample. The focus of this chapter is not on analysis of gene expression but on the application of microarrays for microbial genomic analysis and genotyping. 1.1.2. Genomic Analysis and Genotyping
Genotyping analysis uses the microarray to determine whether specific sequences are present in a DNA sample (Fig. 3). Although the technology could be used to characterize an entire genome,
58
Volokhov et al.
Cells DNA Extraction (DNA Amplification)
Q.C. Oligonucleotide
(ssDNA) Cy3
Cy5
Target Labeling Hybridization Washing
Imaging and data analysis Q.C. Data
Array Data
Fig. 3. Microarray microbial genomic analysis. For genotyping, genomic DNA is extracted from a cell, amplified using PCR (optional), converted to single-strand form (optional), and labeled with a fluorescent dye. If needed, a specific quality control oligonucleotide labeled with a different fluorescent dye is mixed with the target DNA. The microarray is printed with each spot containing the quality control oligonucleotide (complement to the oligo mixed with the sample DNA) plus the labeled target DNA. The mixed DNA sample is hybridized to the microarray followed by washing, scanning, and data analysis. The quality control scan is run to verify proper printing and hybridization of the microarray; the resulting image has a signal at every location on the microarray. A data scan is done to determine whether the sample DNA hybridizes to the specific probe at each location.
in practice for microbial analysis it is often limited to a search for smaller number of features within a microbial genome such as marker sequences for a particular gene or species or single nucleotide polymorphisms (SNP). Unlike the applications for gene expression, which are based on mRNA analysis, genomic DNA is used for genotyping. The genomic DNA is extracted from a cell, purified, amplified using polymerase chain reaction (PCR) (if necessary) and, as for gene expression analysis, labeled with a fluorescent dye. After hybridization, a data scan is then done to determine whether the sample DNA contains a specific sequence by hybridization to the specific probes. Although only one dye is sufficient, for some applications a specific oligonucleotide, labeled with a different fluorescent dye, is mixed with the sample for quality control purposes. In such experiments, the microarray is printed so that each spot contains the quality control oligonucleotide (complementary to the
Microarrays Detection of Bacteria
59
oligonucleotide mixed with the sample DNA) plus a probe sequence. The mixed DNA sample is then hybridized to the microarray followed by washing, scanning, and data analysis. The quality control scan can be used to verify proper printing of the microarray, and the resulting image has two signals at every spot of the microarray. 1.1.3. Target Sequence for Microbial Microarray Analysis
For microbial analysis, two types of DNA sequences are used: microbial ribosomal RNA and signature sequences representing characteristic genes. Signature sequences may include various virulence factors, antibiotic resistance genes, and conserved indels (insertions/deletions) (9–12). Such sequences provide additional information about the classification and nature of the target organism, and its pathogenicity while ribosomal sequence analysis is used as a tool for bacterial classification.
1.1.3.1. Identification of Bacteria on the Basis of Ribosomal DNA Polymorphisms
In prokaryotes, the rRNA genetic loci, rrn, contains the genes for all three conserved ribosomal RNA sequences, 16S, 23S, and 5S. These genes are separated by spacer regions, which exhibit a high degree of length and sequence variation at the levels of genus and species (13). Several such ribosomal sequences have been used for microbial identification including the 16S rDNA, 23S rDNA, and the 16S–23S rDNA spacer region. The most commonly used is the 16S rDNA sequence (14). The variability of 23S rDNA is similar to that of the 16S–23S rDNA spacer region (14), but its variable region is larger than that of the spacer region, making it more suitable for bacterial identification. Microbial ribosomal RNA and the corresponding genes (rrn) are widely used for evolutionary classification (15) because (1) ribosomal RNAs are found in all organisms and are structurally and functionally conserved; (2) ribosomal RNAs are abundant, readily isolated, and identified; (3) their sequence contains both highly conserved and variable regions; (4) they are genetically stable, mutating very slowly, and not exhibiting the horizontal gene transfer found with many other prokaryotic genes; and (5) for many microbial species, rrn sequence data are publicly available.
1.1.3.2. Identification of Bacterial Virulence Factors and Antibiotic Resistance Genes
It is important to identify bacterial virulence factors and antibiotic resistance genes for several reasons. First, such genes can be used to identify the pathogen. Second, some bacterial species are pathogens only when they express particular virulence factors. For example, Staphylococcus aureus is a food pathogen primarily because it expresses the heat-stable enterotoxins (SEs). Similarly, the Bacillus anthracis, virulence factors pagA, lef, cya, and capsule synthesis genes capA, capB, and capC all are required for B. anthracis pathogenicity and useful for pathogen characterization.
60
Volokhov et al.
Many virulence factors are shared among various microbial species. For example, the invasion plasmid antigen (coded by the ipaH gene) (16) is associated with the invasive phenotypes of both Shigella and enteroinvasive Escherichia coli (EIEC). Similarly, shigalike toxins (SLT), slt-I and slt-II (17), which characterize the highly pathogenic E. coli O157:H7 strain are similar to the toxin originally identified in Shigella. eaeA (intimin) (18), which mediates the adherence of the organism to host cells, can be found in both E. coli and Shigella. Microarray analysis can not only identify the pathogens but also determine whether they have the usual virulence genes or have acquired others and can differentiate between pathogenic and nonpathogenic strains of the same bacteria. Several published microarray applications are based on genetic markers in bacterial genomes that are associated with various pathogens and virulence factors such as those for E. coli O157:H7, (19, 20), Listeria (21), Campylobacter (22), Clostridium perfringens (23), Staphylococcal enterotoxins (24), and B. anthracis virulence factors (25). In addition to pathogen-specific microarrays, several pathogen-specific microarrays were combined into a single DNA chip (26) for simultaneous analysis of several species and their virulence factors including S. aureus and S. aureus enterotoxin genes, Listeria spp., Campylobacter spp., and Clostridium perfringens, which represent the majority of food microbial pathogens. A larger microarray with over 53,000 oligoprobes was used for multipathogen detection (142 unique diagnostic regions of 11 bacteria, 5 RNA viruses, and 2 eukaryotes) (27, 28). Genomic markers were used to produce a high sensitivity microarray for biowarfare pathogen detection (10 fg of B. anthracis) for which PCR-amplified target sequences were used to target 18 potential biowarfare agents (28). In addition to virulence factors, it is clinically important to determine the antibiotic resistance profile of the contaminating organism enabling planning of appropriate treatment and monitoring the spread of antibiotic-resistant microorganisms. A microarray for analysis for erythromycin resistance (29) was developed and tested with many S. aureus strains. In separate studies, several multiple tetracycline (tet) resistance genes and b-lactamase blaTEM-1 genes (30) and microarrays have been used for mapping of mutations of pyrazinamide-resistant Mycobacterium tuberculosis strains (31). 1.2. Microarray Design and Fabrication 1.2.1. Microarray Bioinformatics
A bioinformatics approach is useful at several stages of microarray design and interpretation. The primary tools for such analysis include sequence analysis (revolving around melting temperature prediction), sequence search, sequence alignment, and oligo design for PCR primer and probe applications. The best method for melting temperature prediction is based on the nearest neighbor
Microarrays Detection of Bacteria
61
thermodynamic model (32–34), while other simplified models are based on GC content. With many microbial sequences now in the public domain, such bioinformatics-based analyses are an integral part of any microarray project. If sequence information is available for the microbe of interest, available computer tools provide the ability to perform systematic analysis of the genome to discover optimum primer and probe choices under specified constraints. Such analyses can include a multiple sequence alignment step when multiple organisms or strains are expected in the samples. These tools allow the user to create a microarray with temperature-matched probes that assay the targets of interest while avoiding significant similarity to other regions of the genome or to other organisms that may be present as contaminants. Computerized analysis of PCR primer sets can easily detect hairpin and self-annealing potential so as to maximize amplification yield. Many commercial and public domain bioinformatics programs can be found in the literature (35), each of which has a different set of strengths. In our work, we wrote a program, called Oligo Design, that is designed to support microbial microarray bioinformatics design tasks (36). For the purposes of microarray design and utilization, a common task is to compare a short sequence, such as a PCR primer or a microarray probe, against a longer sequence such as a genome or a set of genomes. The purpose of such a search is either to find a close match or to confirm that no significant similarity exists. Many local alignment algorithms have been devised and are freely available, starting from dynamic programming methods such as Smith–Waterman (37) and progressing to fast heuristic algorithms. The search algorithms use sophisticated strategies to accelerate the search, which becomes critical when searching a genome length sequence. Among many search algorithms (38), two have gained wide acceptance: LAlign (39–41) and BLAST (42, 43). Oligo Design implements a Smith–Waterman algorithm that provides local alignment for short sequences (36). Multiple alignment is an extension of local alignment to multiple sequences. In general, multiple alignment involves aligning all sequences to each other and then looking for a consensus sequence that provides the best alignment among all sequences in the input set. As with local alignment, multiple alignment utilizes a set of rules to create an alignment score. The programs search for high scoring segments between the input sequences. The rule sets used are somewhat arbitrary, giving weights to gap openings, gap extensions, exact matches, and mismatches. As a result, the alignments that result must be interpreted carefully since the rules imposed by evolution might be somewhat different. However, the rules are simple enough that one can look at them and conclude that they seem reasonable. Although sequence alignment is not an exact science, these techniques are widely used as the
62
Volokhov et al.
best option for similarity searching. Many multiple alignment algorithms and software packages are available (38). In our work, we used Clustal because it is freely available and well regarded (44, 45). Multiple alignment of a set of closely related microbial genomes (or segments of genomes) provides much useful information for microarrary design. If the input sequences represent the target organisms, the alignment can point toward conserved regions with little sequence variability as well as regions, which have diverged. Conserved regions can be candidate sites for universal primers that can anneal to all targets allowing amplification of multiple targets with a single primer set. If these conserved regions straddle a highly variable region, the amplicon can become the target for a multiorganism assay. Depending on the level of variability found, species and strain identification can be possible using a microarray that has probes for the variable region. PCR primers are designed to anneal, one to each strand of the target, in such a way that the 3¢-ends of the primers point toward each other and the primer pair straddles the amplicon of interest. Primer design is based on knowledge of the sequence of the target, and thus is a bioinformatics task. During the PCR, the polymerase extends the double-stranded molecule starting from the 3¢-end of the primer. Thus, the binding strength and specificity of the 3¢-end are important for high yield PCR. Other design requirements are that the melting temperatures of a pair of primers should be close (ideally equal), the primer set should not exhibit significant self-annealing (which results in competition between the primers and the target), and the amplicon should be of an appropriate length (typically 200–1,000 bp in our work). Computer-based primer design allows a long sequence to be efficiently searched to find all primer sets that meet a particular set of design criteria. Several programs are available to assist with primer design including standalone programs (46) and Web-based tools provided by oligo vendors. Oligo Design has a powerful set of primer design tools that automatically evaluate melting temperature, self-annealing, and 3¢-end stability as the program searches a target gene sequence. A separate tool allows the user to input a primer set, and the program evaluates each primer, finds the amplicon, and diagnoses the potential for primer dimers (36). Once the target genes have been identified, the next task in microarray design is to design the probes. Probe length can vary from very long (100–1,000 nt) for cDNA sequences or PCRderived probes to shorter oligonucleotide probes (20–100 nt). The choice of probe length depends on a number of trade-offs between hybridization specificity, selectivity, and the array cost. A key variable in probe design is the melting temperature, and the probe length is the most important variable in determining melting temperature, which also depends on the sequence
Microarrays Detection of Bacteria
63
(the GC content is a coarse measure of sequence), salt concentration, and DNA concentration. One of the main advantages of short oligoprobes is the ability to discriminate between a perfect match and a single mismatch. With short oligoprobes (20–30 nt in length), one mismatched base will reduce the melting temperature significantly and will consequently reduce hybridization (reduce the signal intensity at a spot). The level of hybridization for a mismatched sequence depends mainly on the nature of the mismatch, the position of the mismatch within the probe (hybridization is most sensitive to mismatches in central positions of the probes), the GC content, and the length of the probe. A second advantage of short oligoprobes is that they allow independent testing of several speciesspecific regions within each gene enabling effective coverage of the target sequence with more (but shorter) oligoprobes. This reduces the probability of misidentification. A third advantage is that short oligonucleotides reduce the cost of microarray production. Advantages of longer oligoprobes (50–70 nt) include higher temperature hybridization, reduction of background interference, and the ability to hybridize with double-stranded DNA target sequences (47). Microarray probe design starts with identification of the genes of interest. In a typical microarray assay, the entire microarray is held at the same temperature during the hybridization incubation step. Thus, all of the probes on a particular array should ideally have the same melting temperature. This design goal is problematic for certain in situ array printing technologies, which often used fixed length probes instead. Although length is a primary variable in determining melting temperature, the sequence dependence is also significant such that the melting temperature for a probe of 20 bp in length can range from 55 to 80°C with the higher temperatures associated with sequences with higher GC content (salt concentration is 1 M, DNA concentration is 50 mM). Data analysis can normalize these differences somewhat but frequently, there is insufficient data to normalize with confidence. For custom-made arrays, a much better approach is to design temperature-matched probes from the start. This can be done by allowing variable length probes and by choosing sequences within the gene that meet temperature criteria. A computerized search of a gene, with temperature constraints, will usually result in a set of probe choices that provide redundant coverage of the gene and which are expected to yield uniform intensity hybridization if the gene is present in the target. 1.2.2. Microarray Fabrication Technologies
The main technologies for DNA microarray fabrication are robotic spotting of presynthesized oligoprobes onto a solid matrix and in situ synthesis of oligoprobes on the array solid matrix. The in situ synthesis involves synthesis of oligonucleotides on the
64
Volokhov et al.
surface of the array in multiple cycles where each cycle adds one base (nucleotide). Such technologies enable the printing of large large-scale, high-density microarrays. However, in situ technologies are relatively expensive and complicated and thus are less practical for smaller microarray projects. 1.2.2.1. Robotic Spotting of Presynthesized Oligoprobes into the Solid Matrix (Contact Printing)
This method, developed and publicized through Pat Brown’s laboratory (48), uses presynthesized oligonucleotides, which are spotted on an array surface. The method is called either contact printing or spotting. The robotic spotting process uses pins to transfer the probe solution from a reservoir (e.g., 96-well plate) onto the array solid matrix (e.g., glass microscope slides coated with a binding layer). The printing head can carry several pins, which are moved to the washing and drying stations to minimize cross-contamination, before dipping into a microtiter plate containing the ligand solutions to be printed. The pins are designed to deposit tiny drops (nanoliter volume) of liquid on the array. The oligonucleotide spotting is followed by irreversible binding of the oligonucleotides to the array surface. Contact printing is versatile and enables the printing of many and varied ligands in addition to DNA probes. Contact printing is probably the most practical way for small laboratories to produce microarrays. The robotic spotting system is relatively simple, inexpensive and is probably the most practical way for small laboratories to produce microarrays. Contact printing is versatile and can be used for printing of tissues and other ligands such as cDNA, PCR products, protein, peptide, antibody, or carbohydrate.
1.2.2.2. Surface Chemistry for Microarray Oligoprobe Binding
Chemical modification of DNA probes is required for the covalent attachment of DNA probes to a solid substrate. The most common methodology for such attachment is based on adding a terminal amino group to the 5¢-end of the oligonucleotide during synthesis to facilitate effective covalent binding to the substrate.
1.3. DNA Microarray Experiments
There are several laboratory steps in microarray experiments: target preparation, target labeling, target hybridization, and microarray washing. These manipulations are followed by microarray image acquisition and data analysis.
1.3.1. Target Preparation and Labeling
The first step in a microarray experiment is the preparation of the target being analyzed by the microarray. The target can be PCR amplicons, genomic DNA, cDNA, or RNA. The preferred target for hybridization is single-stranded DNA (ssDNA), so several protocols have been developed for generating ssDNA. The target is labeled by either directly incorporating molecules of a fluorescent dye (e.g., Cy3 or Cy5) during synthesis or incorporating the dye after synthesis.
Microarrays Detection of Bacteria 1.3.1.1. Target Preparation for Microarray Analysis
65
For differential gene expression analysis, mRNA from two samples is converted to cDNA by reverse transcriptase, and each cDNA is labeled unique fluorescent dye. The two labeled cDNA samples are then mixed and hybridized to the microarray. There are numerous methods for both mRNA and cDNA preparation so these topics will not be discussed here. For microbial genomic analysis, a common approach for target preparation employs both PCR and microarray technologies. The assay starts with PCR amplification of target DNA followed by microarry analysis for identification of the amplicon. This method harnesses both the sensitivity of PCR, which can amplify even very small amounts of genetic material, while hybridization on the microarray provides a highly specific assay and allows parallel analysis of multiple sequences simultaneously. Several PCR-based approaches have been developed to amplify several DNA sequences simultaneously including univeral primers and multiplex PCR. Univeral primer amplification utilizes PCR primers matching conserved sequences that flank a variable region. The universal primers can be designed to incorporate degeneracy and nonspecific bases (such as inosine). For example, 16 staphylococcal enterotoxins were amplified in a single PCR with universal primers (24). The amplicons were then identified using an oligonucleotide microarray designed to hybridize to variable regions of each enterotoxin gene. Multiplex PCR may be used when the detection of multiple genes is required. In multiplex PCR, several specific primer pairs are used to target different sequences in the same PCR, instead of running separate PCRs, to amplify several genes simultaneously in one tube (21, 22, 26). For effective multiplex PCR, the primers have to be carefully selected to minimize interference between primer pairs. Amplification products of PCR are usually double-stranded DNA molecules, which may be used in several ways for hybridization. In general, the hybridization efficiency of double-stranded DNA is lower than the hybridization of single-stranded target DNA. Various available approaches include the following: ●
●
●
●
Double-stranded PCR products can be used directly for hybridization on microarrays made with long oligoprobes. Fragmentation of double-stranded target DNA (e.g., using DNase or sonication) is often necessary for use with microarrays made with short oligoprobes (20–25 nt). Single-stranded target can be produced directly by asymmetric PCR. Single-stranded target can be generated from PCR amplicons by a primer extension (PE) reaction.
66
Volokhov et al. ●
●
●
1.3.1.2. Target Labeling for Microarray Analysis
Single-stranded target can be generated by strand separation using biotinylated primers and streptavidin-coated magnetic beads. Single-stranded target can be generated by in vitro transcription by RNA polymerase utilizing the T7 promoter tag attached to the 5¢-end of the primer, which serves as a recognition site for RNA polymerase. This method is recommended for obtaining very long single-stranded targets (up to 10 kb). Single-stranded target can be generated by asymmetric PCR, an amplification strategy where an excess of one primer is used resulting in linear amplification of one strand in the final PCR cycles.
Commonly used fluorescent labels used for microarray analysis include Cy3 (excited by a green laser) and Cy5 (excited by a red laser). In the most common differential gene expression experiments, two samples are used where each sample is labeled with a different dye. For genotyping, multiple dyes may be used for quality control purposes where the sample is labeled with a one dye, and a second dye can be used for quality control (see below). There are several common methods used to label target DNA, including direct methods of label incorporation during target synthesis and indirect incorporation after target synthesis. ●
●
●
●
Direct incorporation of fluorescently labeled nucleotides (e.g., with Cy3 or Cy5) can be done by reverse transcriptase for cDNA or by DNA polymerase during PCR. However, Cy5 is incorporated less efficiently by reverse transcriptase than Cy3, introducing an incorporation bias in the double label expression experiments. In addition, such transcripts are also biased toward the 3¢-end of the mRNA. The labeled nucleotides may also decrease the yield of PCRs. Indirect labeling is based on incorporation of an aminoallyl-modified dCTP during target synthesis (e.g., cDNA or PCR). The PCR is followed by reaction of the resulting amplicon with an active ester of the dye. This indirect labeling method results in higher total yield of the amplification/ labeling reaction. In addition, the rate of incorporation of Cy3 and Cy5 to the cDNA is similar, eliminating the incorporation bias found for direct labeling. For labeling PCR amplicons, a fluorescent label or other moiety, such as biotin, can be incorporated during the PCR amplification or attached to a primer. The target can then be labeled by post-PCR labeling with streptavidin labeled dye. Chemical processes are available to add fluorescent cyanine labels to RNA (or cDNA) directly, thus making the reverse
Microarrays Detection of Bacteria
67
transcription step unnecessary (49). However, one limitation of this method is that it labels mRNA (the sense strand), and many commercial oligonucleotide microarrays use oligoprobes based on the sense strand. Thus, this method is not used widely. ●
●
1.3.2. Hybridization and Washing
Another approach does not require prehybridization labeling or amplification. Instead, after hybridizing, the RNA to an oligonucleotide microarray, the bound RNA molecules are detected in a second hybridization step using gold nanoparticles modified poly-T oligos, which (50) hybridize to the poly-A tail of the captured mRNA molecules. This approach can give ~1,000 times more sensitivity than fluorescent-based methodologies. However, this method is limited to eukaryotic expression analysis since prokaryotic mRNA does not have the poly-A tail. As with other direct labeling methods, the orientation of the probe is a factor in the application of this approach. Quantum-dot methodologies can also be used for labeling (51) in a manner similar to the gold nanoparticles described above. Such labels can be visualized optically (absorbance, resonance light scattering, colorimetric, surface-enhanced Raman spectroscopy) or electrochemically.
In hybridization, the DNA probes on the microarray and the complimentary labeled DNA (or RNA) target anneal to form a double-stranded molecule. Unbound targets are washed off the array, with the result that the only fluorescent signals on the array are emitted from spots where labeled target sequence is hybridized. This is a quantitative measurement in which the level of fluorescence from each of the spots is measured and used to calculate the absolute or relative amount of DNA bound to each spot on the array. With double label experiments, the relative signal measured for each label is interpreted as the ratio of the transcript abundance. Stringent hybridization conditions minimize nonspecific binding and decrease background signal. Hybridization stringency depends on several parameters including probe size, target concentration, salt concentration, temperature, formamide concentration, hybridization chamber configuration, and time. In general, long oligoprobes and high hybridization temperature increases stringency while high salt concentration decreases stringency. Formamide tends to denature DNA and disrupt secondary structure increasing the hybridization stringency by lowering the temperature at which hybridization will occur. A wide range of formamide concentrations (between 0 and 50%) have been used in microarray practice. Na+ is commonly used in the hybridization buffer at about 1 M concentration. The less Na+ present, the
68
Volokhov et al.
greater the stringency of the hybridization. In some cases, carrier DNA may be added to the hybridization solution to reduce nonspecific annealing. Most hybridization reactions take 12–24 h, with shorter times (a few hours) for short probes with high target DNA concentration. For hybridization, the array is placed in a hybridization chamber, and the labeled target DNA in the appropriate hybridization solution (including buffer, salt, and formamide) is introduced into the chamber. The sealed cassette is then immersed in a temperature-controlled environment (water bath or an incubator) for hybridization. The speed of the hybridization reaction can be increased through mixing. For larger-scale operations, robotic hybridization systems have been developed to automatically handle multiple chips enabling computer control of the temperature of the target and slide, and the washing cycles. Computer control may reduce the variability of microarray experiments. The aim of washing is to remove unbound target assuring that only the DNA complementary to each spot will remain bound to the microarray. After hybridization, the array is typically washed with a low salt-detergent buffer (e.g., 0.1× SSC and 0.1% SDS) or with a high-temperature wash. Normally, several washing cycles with increasing stringency (decreasing salt concentration) are used to remove nonspecifically annealed target molecules, maximizing the significance of a hybridization signal. 1.3.3. Array Scanning and Data Analysis
After washing, the bound labeled targets on the microarray are analyzed, normally by fluorescence detection. The scanners used for array analysis are normally optical detectors that scan the surface of the slide and detect the fluorescent signals from the dye. The most common scanners for microarray detection are confocal scanning devices in which a laser is used to excite the dyes and the emission is measured by a photomultiplier tube (PMT), or similar detector, and converted to a digital signal. In most systems, twocolor arrays are scanned with two lasers; a green laser (for Cy3; excitation wavelength is 550 nm and emission wavelength is 581 nm) and a red laser (for Cy5; excitation wavelength is 649 nm and emission wavelength is 670 nm). Confocal laser scanners provide high-resolution imaging with a very narrow depth of focus, thus reducing background artifacts. Commercially available confocal imaging systems provide 5–10 mm resolution, which is the best available resolution and sensitivity. For low-density microarrays with one dye, imaging with a charge-coupled device (CCD) camera or flatbed scanner is an alternative to confocal scanning. Unlike confocal scanning, such imaging often involves illumination and detection of a large portion of the slide (~1 cm2) simultaneously. CCD instruments often use continuous wavelength light sources, such as arc lamps, thereby eliminating the need for multiple lasers. The limitation of
Microarrays Detection of Bacteria
69
such low-cost scanners is generally lower pixel resolution (i.e., approximately 20–50 mm) and higher background signal resulting in lower sensitivity. The output of a confocal laser scanner is usually two monochrome images: one for each of the two wavelengths measured, which are combined for data analysis. For two-color differential expression studies, these images are combined to create red– green–yellow false color images of microarrays, with yellow used when both wavelengths are emitted. Quantitation is usually accomplished by superimposing a grid over the microarray image and computing an average intensity value for each microarray element. Large expression microarrays generate a large number of data points, which are geared toward uncovering expression clustering patterns in the data. As a consequence of the experimental design, there are many sources of variability in expression microarrays (e.g., sample preparation, amount of mRNA, quality of mRNA, labeling biases, and efficiency); this is in addition to the differential gene expression, which is studied. This experimental approach requires very complex data analysis and sophisticated bioinformatics with the aim of “clustering” the data, which provides a useful tool for extracting underlying gene expression information. Several analytical methodologies were developed including “supervised” or “unsupervised” analysis methods. Supervised methods require a preexisting classification (e.g., a subset of data used as “training set,” knowledge of gene function or regulation, phenotype, tissue origin, or cell type) while in unsupervised methods, no preexisting classification and no additional information besides the expression data itself is used. The main purpose of clustering methods is to group genes based on similarity of expression profiles (i.e., genes that are expressed together most frequently) and to adjust the profile for experimental variability. Many aspects of the bioinformatics related to expression microarrays were reviewed recently (52) and are beyond the scope of this manuscript, which is focused on genomic microarrays. Unlike expression arrays where the main objective is finding associations between genes, microarrays for genomic analysis are mainly used to detect the presence of genes or alleles. Thus, the data analysis and bioinformatics for these experiments is much different than the gene expression analysis. Moreover, the fewer spots on typical low-density genomic arrays provide a relatively smaller amount of data. Although in our work, we have used a second dye to improve quality control, the primary data is still from a single dye. Regarding the reproducibility of the method, microarray technologies are relatively complex with a large number of variables that can influence the results. Many discrepancies in microarray assay results have been reported in the literature, especially
70
Volokhov et al.
when using different microarray platforms (53–55). For genomic analysis, reproducibility can be obtained through design of the array and the technique with an emphasis on quality control (21, 22, 25). 1.3.4. Quality Control
To assure uniformity of array printing and hybridization conditions, quality control (QC) elements can be incorporated into the array during fabrication. In one method, each oligoprobe spot is spiked with a QC oligonucleotide of nonbacterial origin (21). This QC oligoprobe is designed to hybridize with a complementary Cy3labeled oligonucleotide mixed with the target DNA. Laser scanning of the microchip at 543 nm generated a QC image that was used to evaluate the efficiency of the spotting and hybridization steps (see QC images, Fig. 4).
Fig. 4. Microarray for Listeria species identification based on the iap gene. Species identification based on differences in iap gene sequences. Genomic DNAs of six reference strains were amplified by using universal iap-specific primers followed by separation of PCR products in a 1.5% agarose gel (a). Lanes : M, 100-bp DNA ladder mix; 1, L. monocytogenes ; 2, L. ivanovii ; 3, L. welshimeri ; 4, L. innocua ; 5, L. seeligeri ; 6, L. grayi. Species-specific DNAs were hybridized with the iap microchip. (b) QC microarray showing the chip features hybridized to the QC oligonucleotide labeled with Cy3, the microarray contains individual oligonucleotides specific to L. monocytogenes (row a), L. ivanovii (row b), L. welshimeri (row c), L. innocua (row d ), L. seeligeri (row e), and L. grayi (row f ). (c) The Cy5-labeled target amplicons from each species shown in (a) are hybridized to the microarray; image panels are numbered in accordance with the species numeration in (a).
71
Microarrays Detection of Bacteria
1.4. Example of DNA Microarray for Microbial Genotyping
To demonstrate the application of microarray technologies for microbial analysis, here we provide an example of such an application: microarray analysis and identification of Listeria species based on the iap gene sequence. In the original work (21), a rapid microarray-based assay for the reliable detection and discrimination of six species of the Listeria genus, L. monocytogenes, L. ivanovi, L. innocua, L. welshimeri, L. seeligeri, and L. grayi, was described. The approach used in that study involves one-tube multiplex PCR amplification of six target bacterial virulence factor genes (iap, hly, inlB, plcA, plcB, and clpE). Here, as an example, we present the analysis of only one gene: iap. The iap gene encodes the 60-kDa secreted invasion-associated protein p60, for which homologues were found in all Listeria species. The iap genes from all six Listeria species were amplified using a single set of “universal” primers binding the regions conserved among all Listeria species producing similar size amplicons (644–722 bp in length depending on species, as shown in Fig. 4a). Following the PCR amplification, fluorescently labeled targets were synthesized using primer extension of the PCR products in the presence of only reverse primers, generating Cy5-labeled ssDNA of iap sequences. Discrimination between Listeria species was done by hybridization of amplified ssDNA samples with immobilized oligonucleotide probes recognizing variable DNA regions specific for each species. Ten unique oligonucleotide probes were immobilized on glass microchips in rows representing each Listeria species. Along with the probes, quality control oligos were also included. The rows of spots in the QC scan are shown in Fig. 4b, which validates the microarray fabrication and hybridization steps by providing a control image of the entire microarray. As shown in Fig. 4c, after hybridization with the amplicons shown in Fig. 4a, all designed oligonucleotide probes could recognize and hybridize predominantly, with minimal crosshybridization, with samples representing each of the six Listeria species. The results of hybridization demonstrate high species specificity and unambiguous discrimination among all six species.
2. Materials 2.1. Bioinformatics
Use of the software Subheading 3.7.
in
this
list
is
described
under
1. Oligo Design (http://www.enme.umd.edu/bioengineering). 2. LAlign (http://fasta.bioch.virginia.edu/fasta_www2/fasta_ www.cgi?rm=lalign). 3. BLAST (http://www.ncbi.nlm.nih.gov/blast/Blast.cgi). 4. Clustal (http://www.clustal.org/).
72
Volokhov et al.
2.2. General Material and Supplies
The general materials in this list are commonly used laboratory material, so no specific supplier specified unless it is a specialty item. 1. Microcentrifuge. 2. Boiling water bath. 3. Heat block. 4. Thermal cycler. 5. 15-ml test tube. 6. 1.5-ml test tubes. 7. Micropipet and tip. 8. Microcentrifuge. 9. Tube Racks with cover to keep microcentrifuge tubes closed while boiling (Electron Microscopy Sciences, Hatfield, PA). 10. UV/ VIS Spectrophotometer. 11. Electrophoretic equipment (gel box, power supply, agarose, ethidium bromide, UV transilluminator, and camera).
2.3. Microbial Samples Preparation
1. Media: general purpose media; Brain Heart Infusion (BHI).
2.4. Isolation of Total DNA from Microbial
1. TE buffer (10 mM Tris–Cl, pH 7.5, 1 mM EDTA).
2.5. Preparation of Single-Stranded DNA or RNA Targets for Microarray Analysis
1. HotStarTaq DNA Polymerase 10× buffer and 2.5 mM MgCl2 (QIAgene, Chatsworth, CA). Store at −20°C.
2.5.1. Preparation of Single-Stranded DNA 2.5.1.1. Preparation of Single-Stranded DNA Targets Using Primer Extension Reaction
2. Microbial strains (E. coli, Salmonella, Shigella, Listeria, Campylobacter, Staphylococci, Bacilli, and Bacillus spp.) isolated or obtained from the American Type Culture Collection (ATCC), Manassas, VA.
2. Solvents: Phenol, Chloroform, anhydrous ethanol.
2. HotStarTaq DNA Polymerase (QIAgen, Chatsworth, CA). Store at −20°C. 3. 10 mM deoxynucleotide triphosphates (dNTPs; Invitrogen, Carlsbad, CA). Store at −20°C. 4. Forward and reverse primers for PCR amplification of selected bacterial genetic marker(s) and preparation of ssDNA target for microarray analysis ((21) #65; (24) #538; (26) #57; (25) #64); adjust the concentration of primer stocks to 200 mM and store the stock at −20°C; prepare the primer working stock solutions with concentration 20 mM and keep them at −20°C. 5. UltraPure™ DEPC-Treated Water (Invitrogen). Store at room temperature. 6. QIAquick PCR Purification Kit (QIAgen).
Microarrays Detection of Bacteria 2.5.1.2. Preparation of ssDNA Targets Using Magnetic Bead-Based Strand Separation
73
1. HotStarTaq DNA Polymerase 10× buffer and 2.5 mM MgCl2 (QIAgene, Chatsworth, CA). Store at −20°C. 2. HotStarTaq DNA Polymerase (QIAgen, Chatsworth, CA). Store at −20°C. 3. 10 mM dNTPs (Invitrogen, Carlsbad, CA). Store at −20°C. 4. Forward and reverse primers for PCR amplification of selected bacterial genetic marker(s). Forward PCR primer should contain a biotin moiety at the 5¢-end of oligonucleotide. The 5¢-end biotinylation of oligonucleotide is currently offered by many service biotech companies. Adjust the concentration of primer stocks to 200 mM and store the stock at −20°C; prepare the primer working stock solutions with concentration 20 mM and keep them at −20°C. 5. UltraPure™ DEPC-Treated Water (Invitrogen). Store at room temperature. 6. 2× binding buffer for magnetic DNA strand separation [10 mM Tris–HCI, pH 7.5, 2 M NaCl, 1 mM ethylenediaminetetraacetic acid (EDTA)]. 7. 1× binding buffer for magnetic DNA strand separation. 8. 0.1 M NaOH. 9. 3 M Sodium acetate, pH 5.8. 10. Dynabeads® M-280 Streptavidin (Invitrogen). Store at +4°C. 11. Magnetic particle concentrator Dynal MPC-S (Invitrogen). 12. QIAquick PCR Purification Kit (QIAgen).
2.5.2. Preparation of Single-Stranded RNA Targets Using T7 Polymerase
1. Forward and reverse primers for PCR amplification of selected bacterial genetic marker(s) ((29) #63). Reverse PCR primer should contain the sequence of T7 promoter at the 3¢-end. 2. MEGAscript T7 High Yield Transcription Kit (Ambion, Austin, TX). Store at −20°C. 3. QIAquick PCR Purification Kit (QIAgen). 4. UltraPure™ DEPC-Treated Water (Invitrogen). Store at room temperature. 5. CentriSep Spin Columns (Princeton Separations, Adelphia, NJ).
2.6. Labeling Microarray Hybridization Targets with Fluorescent Dyes 2.6.1. Labeling of ssDNA During Primer Extension Reaction
1. HotStarTaq DNA Polymerase 10× buffer and 2.5 mM MgCl2 (QIAgen, Chatsworth, CA). Store at −20°C. 2. HotStarTaq DNA Polymerase (QIAgen, Chatsworth, CA). Store at −20°C. 3. 10 mM dNTPs (Invitrogen, Carlsbad, CA). Store at −20°C. 4. 25 nM Cy5-dCTP (GE Healthcare Bio-Sciences Corp., Piscataway, NJ).
74
Volokhov et al.
5. Reverse primers for PE reaction (21, 24–26). 6. UltraPure™ DEPC-Treated Water (Invitrogen). Store at room temperature. 7. CentriSep Spin Columns (Princeton Separations, Adelphia, NJ). 8. UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro (GE Healthcare Bio-Sciences Corp., Piscataway, NJ). 2.6.2. Labeling of ssDNA and ssRNA Using the MICROMAX ASAP RNA 2.6.2.1. Labeling System
1. MICROMAX ASAP RNA Labeling Kit (Perkin-Elmer, Boston, MA). 2. CentriSep Spin Columns (Princeton Separations, Adelphia, NJ). 3. UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro (GE Healthcare Bio-Sciences Corp., Piscataway, NJ).
2.7. Preparation of Microarray ssDNA Hybridization Targets for Nanogold/Silver Enhancement Microarray Detection Method
1. Label It Biotin Labeling Kit (Mirus Bio, Madison, WI); 1× Mirus labeling buffer, Mirus Label It Biotin reagent.
2.8. Microarray Fabrication and Hybridization Assay
1. Oligonucleotide probes: As an example please see (20–22, 24–26, 29, 56, 57) for details of the oligonucleotide probes used for detection and discrimination of B. anthracis by microarray (25). The 5¢-end of each microarray oligoprobe should be modified during the synthesis using the TFA Aminolink CE reagent (PE Applied Biosystems, Foster City, CA) to allow immobilization to the surface of silylated (aldehyde surface chemistry) glass slides (CEL Associates, Inc., Houston, TX) or any other suitable slide surfaces (e.g., CodeLink Activated slides).
2.8.1. Preparation of Oligonucleotides for Spotting
2. CentriSep column (Princeton Separations, Adelphia, NJ).
2. Dimethyl sulfoxide (Sigma-Aldrich): Store aliquots frozen at −20°C. 3. 384-well PCR plate (Marsh Bio Products, Rochester, NY). 4. Silylated (aldehyde surface chemistry) glass slides (CEL Associates, Inc., Houston, TX): Store desiccated at room temperature or refrigerated. 5. Contact microspotting robotic system PIXSYS (Cartesian Technologies, Inc., Ann Arbor, MI).
5500
6. Microspotting pin CMP7 or equivalent (Arrayit, Sunnyvale, CA). 7. Dry-seal desiccator (PGC Scientific, Frederick, MD).
Microarrays Detection of Bacteria
75
8. Silylated glass slides blocking solution: 0.25% NaBH4 solution in water. The water solution of NaBH4 is unstable and should be freshly prepared every time. Store NaBH4 at room temperature in a dry-seal desiccator. 9. 20× SSC: Dissolve 175.3 g of NaCl and 88.2 g of sodium citrate in 800 ml of distilled H2O. Adjust the pH to 7.0 with a few drops of 1 M HCl. Bring the final volume to 1 l with additional ddH2O. Store at room temperature for up to 1 month. 10. 4× SSC, 0.1% (w/v) SDS (sodium dodecyl sulfate). 11. Eppendorf centrifuge 5810R (Eppendorf AG, Hamburg, Germany) or equivalent. 12. Boekel Slide Moat Hybridization Incubator 240000 (MG Scientific, Pleasant Prairie, WI, USA) or equivalent. 2.8.2. Reagents for Quality Control of Microarray Printing and Hybridization Conditions 2.8.3. Microarray Hybridization and Posthybridization Slide Processing 2.8.3.1. Microarray Hybridization and Posthybridization Slide Processing for Fluorescence-Based Detection
1. FluoroLink Cy3 dye kit (Amersham Pharmacia Biotech, Inc.). 2. 5¢Aminolink QC oligonucleotide. 1. Glass coverslips (LifterSlip coverslips; Erie Scientific Company, Portsmouth, NH, USA). 2. ArrayIt® Hybridization Cassette chambers (Arrayit, Sunnyvale, CA). 3. Boekel/Grant Premier Digital Water Bath, Models PB-600 (MG Scientific, Pleasant Prairie, WI, USA) or equivalent. 4. 2× hybridization buffer (10× Denhardt’s solution, 12× SSC buffer, and 0.1% Tween 20). 5. 6× SSC, 0.2% (w/v) Tween 20. 6. 2× SSC solution. 7. 1× SSC solution. 8. Eppendorf centrifuge 5810R (Eppendorf AG, Hamburg, Germany) or equivalent. 9. ScanArray 5000 (Perkin-Elmer). 10. ScanArray Express software (Perkin-Elmer).
2.8.3.2. Microarray Hybridization and Posthybridization Slide Processing for Nanogold/ Silver Enhancement Detection Method
1. Glass coverslips (LifterSlip coverslips; Erie Scientific Company, Portsmouth, NH, USA). 2. ArrayIt® Hybridization Cassette chambers (Arrayit, Sunnyvale, CA). 3. Micromax hybridization buffer III (Perkin-Elmer, Boston, MA). 4. Gold-labeled streptavidin, 5 nm (Kirkegaard and Perry Laboratories, Gaithersburg, MD).
76
Volokhov et al.
5. Silver Enhancement solutions A and B, Silver Enhancer Kit (store at −20°C) (Nanosphere, Northbrook, IL). 6. 6× SSC buffer. 7. 6× SSC buffer containing 0.4% gelatin and 0.1% Tween 20. 8. 0.6 M Sodium nitrate (NaNO3). 9. Eppendorf centrifuge 5810R (Eppendorf AG, Hamburg, Germany) or equivalent.
3. Methods The microarray protocol presented below was established and optimized for detection and discrimination of B. anthracis by microarray hybridization (25). 3.1. Microbial Samples Preparation
3.1.1. Culturing on BHI Plates
There are numerous media and culturing conditions for various microorganisms, so detailed discussion of microbial culturing is beyond the scope of this chapter; for detailed culturing and isolation method, the FDA BAM (58) is a good source. For the work on food microbial pathogens (E. coli, Salmonella, Shigella, Listeria, Campylobacter, Staphylococci, Bacilli, and Bacillus spp.) described here, BHI was used for culturing. BHI is a rich none selective media suitable for culturing many microorganisms. Culturing can be done in broth or plates, most of the work done in our lab for foodborne pathogens was done with platting on BHI, which provide more cells. 1. Incubate the plates overnight at 37°C. 2. Wash the plates with TE buffer (e.g., 10 ml). 3. Centrifuge the cell suspension, (e.g., 8,000 × g for 10 min) and resuspend the cells in 3-ml TE buffer.
3.1.2. Culturing in BHI Broth
1. Inoculate 3 ml of BHI in 10-ml tube with a single colony. 2. Incubate the culture at 37°C with vigorous shaking (250 cycles/min in a rotary shaker) over night or until the bacteria reach the late log phase of growth (i.e., an OD600 of approximately 0.6). 3. Remove aliquots (1.5–2 ml) of the bacterial culture to a fresh microcentrifuge tube and harvest by centrifugation (e.g., 13,000 × g for 10 min). 4. Discard the supernatant and resuspend in 100-ml TE buffer.
3.2. Isolation of Total DNA from Microbial Samples
There are numerous methods for microbial DNA extraction; microbial DNA isolation is beyond the scope of this chapter, several good manual with protocols for DNA isolation are available (59, 60).
Microarrays Detection of Bacteria
77
For our work on food microbial pathogens, we used two relatively simple methods: boiling method and phenol–chloroform extraction; in general, the boiling method is more rapid and is suitable for large number of sample preparations, whereas phenol–chloroform extraction provides cleaner DNA. 3.2.1. Isolation of Total DNA with Boiling
1. Overnight grown cell on BHI plates washed with TE buffer (e.g., 10 ml), followed by centrifugation (e.g., 8,000 × g for 10 min), and suspended in 3 ml TE buffer, Vortex to mix. 2. The cell divides the cell suspension into ~0.7-ml aliquots in 2-ml minifuge tubes. 3. The tubes closed very tightly and boiled for 10 min followed by centrifugation at 14,000 × g for 10 min a minfuge to remove denatured proteins and bacterial membranes. 4. Transfer the supernatant (containing the DNA) to clean 1.5-ml tube, avoid cell debris. 5. The presence of genomic DNA in all prepared samples was confirmed by 1% agarose gel electrophoresis followed by staining with ethidium bromide.
3.2.2. Phenol–Chloroform Extractions of Total DNA
1. Freshly grown bacteria (e.g., 2.5-ml broth) resuspended in 0.5-ml water (approximately 108 cells/ml). 2. The cell treated with lysozyme (50 mg/ml) for 2 h at 37°C (60) to disrupt bacterial walls. 3. Proteins are removed from solution by two sequential phenol–chloroform (1:1) extractions. 4. Bacterial DNA is precipitated by three volumes of anhydrous ethanol. 5. The DNA pellets are dried in a vacuum and resuspended in 300 ml of water. 6. The presence and quality of genomic DNA were confirmed by 0.8% agarose gel electrophoresis followed by gel staining with ethidium bromide.
3.3. PCR Amplification
Several PCR amplification approaches are useful for DNA amplification. The conditions for each approach depend on many factors including the target sequence, primer sequences, and primer length and have to be optimized for each assay. Standard PCR amplification describes the basic protocol for PCR amplification. Multiplex PCR protocol describes amplification of several target sequence in the same tube; generally multiplex PCR requires a great deal of optimization and the PCR primers used have to be compatible. Asymmetric PCR amplification, which performs with an excess of the reverse primers, generates ssDNA suitable for hybridization. Multiplex-asymmetric PCR amplification
78
Volokhov et al.
combines multiplex-asymmetric PCR amplification and Universal primers PCR amplification, which use for simultaneously amplification of several target sequences, which share conserved regions that are used for primer design. 3.3.1. Standard PCR Amplification
1. Set up a 50-ml PCR mixture as following: 1× HotStarTaq DNA Polymerase buffer. 2 U of HotStarTaq DNA Polymerase. 2.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 200 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 300 nM of forward and reverse primers. Depending on the methods selected for preparation of microarray hybridization target, ssDNA, or ssRNA, primers have to contain either 5¢-biotin or T7 promoter sequence at 3¢-end, respectively. 0.1–0.2 mg of total bacterial DNA as a template. UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions (25): initial denaturing at 95°C for 15 min; 40 cycles at 94°C for 40 s, 55°C or 60°C for 40 s, 72°C extension for 60 s; and final extension at 72°C for 10 min. 3. Hold the reaction at 4°C. 4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol. 5. Measure the target amplicon yield using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro.
3.3.2. Multiplex PCR Amplification
The conditions for multiplex PCR amplification depend on many factors including the target sequence, primer sequences, primer length, etc. The example given here is for identification of Listeria species. 1. Set up a 30-ml PCR mixture as following: 1× HotStarTaq DNA Polymerase buffer. 1.5 U of HotStarTaq DNA Polymerase. 2.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 200 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 600 nM of forward and reverse primers.
Microarrays Detection of Bacteria
79
~200 ng total bacterial DNA as a template. UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions: initial activation at 95°C for 15 min; 40 cycles at 94°C for 20 s, 60°C for 20 s, and 72°C for 50 s; and a final extension at 72°C for 10 min. 3. Hold the reaction at 4°C. 4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol. 5. Measure the target amplicon yield using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro. 6. The presence of amplified PCR products can be detected by 2% agarose gel electrophoresis in 1× TBE buffer with ethidium bromide staining and photographed in UV light. 3.3.3. Asymmetric PCR Amplification
The conditions for asymmetric PCR amplification depend on many factors including the target sequence, primer sequences, primer length, etc. The example given here is for amplification of Bacillus cereus group virulence factors (56). 1. Set up a 50-ml PCR mixture as following: 1× HotStarTaq DNA Polymerase buffer. 4 U of HotStarTaq DNA Polymerase. 3.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 400 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 50 nM of forward primer. 500 nM of reverse primers. 30–200 ng total bacterial DNA as a template. UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions: initial denaturing at 95°C for 15 min; 35 cycles at 94°C for 30s, 54°C for 40 s, and 72°C extension for 45 s; and final extension at 72°C for 7 min. 3. Hold the reaction at 4°C. 4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol. 5. Measure the target amplicon yield using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro.
80
Volokhov et al.
6. The presence of amplified PCR products can be detected by 2% agarose gel electrophoresis in 1× TBE buffer with ethidium bromide staining and photographed in UV light. 3.3.4. MultiplexAsymmetric PCR Amplification
The conditions for multiplex-asymmetric PCR amplification depend on many factors including the target sequence, primer sequences, primers length, etc. The example given here is for amplification of Bacillus cereus group virulence factors (56). In this example, the primers were divided into three groups with four sets of primers in each group. 1. Set up a 50-ml PCR mixture as following: 1× HotStarTaq DNA Polymerase buffer. 4 U of HotStarTaq DNA Polymerase. 3.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 200 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 50 nM of forward and 500 nM of reverse primers for each of the genes. 300–500 ng total bacterial DNA as a template. UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions: initial activation at 95°C for 15 min; 35 cycles at 94°C for 45 s, 54°C for 1 min, and 72°C for 1 min; and extension at 72°C for 10 min. 3. Hold the reaction at 4°C. 4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol. 5. Measure the target amplicon yield using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro. 6. The presence of amplified PCR products can be detected by 2% agarose gel electrophoresis in 1× TBE buffer with ethidium bromide staining and photographed in UV light.
3.3.5. Universal Primers PCR Amplification
The conditions for universal primers PCR amplification depend on many factors including the target sequence, primer sequences, primer length, etc. The example given here is for amplification of Staphylococcal enterotoxins (24). 1. Set up a 50-ml PCR mixture as following: 1× HotStarTaq DNA Polymerase buffer. 5 U of HotStarTaq DNA Polymerase.
Microarrays Detection of Bacteria
81
3.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 200 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 700 nM forward universal primer mixture and 1,400 nM reverse universal primer mixture. 600–800 ng total bacterial DNA as a template. UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions: initial denaturing at 95°C for 15 min; 7 cycles at 94°C for 1 min, 40°C for 1 min, and 72°C for 1 min; 35 cycles at 94°C for 1 min, 45°C for 1 min, and 72°C for 1 min; and final extension at 72°C for 7 min. 3. Hold the reaction at 4°C. 4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol. 5. Measure the target amplicon yield using UV/Visible Spectrophotometer e.g. Ultraspec 2100 pro. 6. The presence of amplified PCR products can be detected by 2% agarose gel electrophoresis in 1× TBE buffer with ethidium bromide staining and photographed in UV light. 3.4. Preparation of Single-Stranded DNA or RNA Targets for Microarray Analysis 3.4.1. Preparation of Single-Stranded DNA 3.4.1.1. Preparation of Single-Stranded DNA Targets Using Primer Extension
1. Set up a 50-ml ssDNA reaction mixture as following: 1× HotStarTaq DNA Polymerase buffer. 2 U of HotStarTaq DNA Polymerase. 2.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 200 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 600 nM of reverse primer. 200 ng of the PCR product as a template from the first step (Subheading 3.3.1, step 1). UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions: initial denaturing at 95°C for 15 min; 40 cycles at 94°C for 40 s, 55°C or 60°C for 40 s, 72°C extension for 60 s; and final extension at 72°C for 10 min. 3. Hold the reaction at 4°C. 4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol.
82
Volokhov et al.
5. Measure the target amplicon yield using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro. 6. Save the target PCR amplicon at −20°C until use. 3.4.1.2. Preparation of ssDNA Targets Using Magnetic Bead-Based Strand Separation
1. Set up PCR using biotinylated reverse primer of 16S–23S ITS region (Volokhov et al. 2006): 1× HotStarTaq DNA Polymerase buffer. 2 U of HotStarTaq DNA Polymerase. 2.5 mM MgCl2. 250 mM of each dNTP (dATP, dGTP, dCTP, and dTTP). 200 nM of each forward and reverse primers. UltraPure™ DEPC-Treated Water added to the final volume of 50 ml (add first). 2. Perform the PCR amplification using thermocycler with the following cycle conditions: initial denaturing at 95°C for 15 min; 35 cycles at 94°C for 30 s, 60°C for 30 s, 72°C extension for 2 min; and final extension at 72°C for 10 min. 3. Hold at 4°C. Check the presence of amplified PCR products by electrophoresis in 1% TAE agarose gels with ethidium bromide followed by ultraviolet amplicon visualization. 4. Prepare 50-ml Streptavidin Magnetic Dynabeads (Dynabeads M-280 Streptavidin reagent; Invitrogen). Transfer 50 mL of Dynabeads M-280 suspension into 1.5-ml centrifuge tube and place the tube into magnetic particle concentrator Dynal MPC-S or equivalent for 1 min. Replace the storage buffer for 100 ml of 1× binding buffer, remove the tube from concentrator and thoroughly mix the suspension. Place the tube back into concentrator for 1 min and replace the supernatant to 50 ml of 2× binding buffer. 5. Mix 50-ml stock Dynabeads with 50-ml purified biotinylated PCR product. 6. Incubate for 30 min with intense agitation. 7. Separate magnetic beads with the bound PCR products using a magnetic particle concentrator MPC-S (Dynal MPC-S) (Invitrogen). 8. Discard supernatant. 9. Wash magnetic beads one more time with 100 ml 1× binding buffer. 10. Place the tube into magnetic concentrator for 1 min and discard the supernatant. 11. Elute the ssDNA from the beads using 50 ml of 0.1 M NaOH for 5 min using magnetic concentrator.
Microarrays Detection of Bacteria
83
12. Neutralize NaOH by addition of 5 ml of 3 M CH3COONa (pH 5.8). 13. Purify ssDNA using the CentriSep column (Princeton Separations, Adelphia, NJ) equilibrated with the deionized water. 14. Measure the target amplicon yield using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro. 15. Save the ssDNA at −20°C until the use. 3.4.2. Preparation of Single-Stranded RNA Targets Using T7 Polymerase
1. Set up a 30-ml single-stranded RNA (ssRNA) reaction mixture as following: 1× MEGAscript T7 reaction buffer. 2 ml of MEGAscript T7 Enzyme Mix. 5 mM of dATP, dUTP, dCTP, and dGTP each. 0.1–0.5 mg of the PCR product as a template from the first PCR step using reverse primers containing the sequence of T7 phage promoter (Subheading 2.3). UltraPure™ DEPC-Treated Water to a final volume of 30 ml (add first). 2. Incubate the reaction mixture at 37°C for 2–4 h. Hold the reaction at 4°C until used. 3. Remove the unincorporated dNTPs using a CentriSep Spin Columns according to the manufacturer’s protocol.
3.5. Labeling Microarray Hybridization Targets with Fluorescent Dyes 3.5.1. Labeling of ssDNA During Primer Extension Reaction
1. Set up a 50-ml ssDNA reaction mixture as following: 1× HotStarTaq DNA Polymerase buffer. 2 U of HotStarTaq DNA Polymerase. 2.5 mM MgCl2 (supplied with HotStarTaq DNA Polymerase). 200 mM of each dATP, dGTP, and dTTP. 40 nM dCTP. 40 nM Cy5-dCTP. 600 nM of reverse primer. 200 ng of the PCR product as a template from the first step (Subheading 3.3.1, step 1). UltraPure™ DEPC-Treated Water to a final volume of 50 ml (add first). 2. Run the PCR amplification using a thermocycler with the following cycle conditions: initial denaturing at 95°C for 15 min; 40 cycles at 94°C for 40 s, 55°C or 60°C for 40 s, 72°C extension for 60 s; and final extension at 72°C for 10 min. 3. Hold the reaction at 4°C.
84
Volokhov et al.
4. Clean up the target amplicon from the unincorporated primers and dNTPs using QIAquick PCR Purification Kit according to the manufacturer’s protocol. 5. Assess the amount of the Cy5 dye incorporated into ssDNA by measuring the ratio of the absorbance at 649 nm to the absorbance at 260 nm using UV/ Visible Spectrophotometer e.g. Ultraspec 2100 pro. The typical ratio of 649/260 should be about 0.15–0.25, which corresponds to 1.5–3 dye moieties per 100 nucleotides of ssDNA. 6. Save the labeled target PCR amplicon at −20°C until the hybridization. 3.5.2. Preparation of Fluorescent Labeling of ssDNA and ssRNA Microarray Hybridization Targets Using the MICROMAX ASAP RNA Labeling System
Use the MICROMAX ASAP RNA Labeling Kit for equally efficient incorporation of Cy5 fluorophore into either ssRNA or ssDNA molecules. 1. Set up a 50-ml PCR mixture as follows: 10 mg of ssRNA or ssDNA sample (from Subheadings 3.3.2 and 3.3.3). 1 ml of ASAP Cyanine-5 Labeling Reagent. ASAP Labeling Buffer to the final volume of 50 ml (add first). 2. Run the labeling reaction using a thermocycler at 85°C for 40 min. 3. Hold the samples at 4°C until used. 4. Remove the unincorporated dNTPs by purification through the CentriSep Spin Columns according to the manufacturer’s protocol. 5. Quantitate the DNA/RNA concentration by diluting the nucleic acids as 1/50 dilution in water and measure A260. 6. Adjust the nucleic acid concentration to 0.5–1.0 mM in the 1× Hybridization Buffer. 7. Store the labeled target DNA or PCR at −20°C until the hybridization.
3.6. Preparation of Microarray ssDNA Hybridization Targets for Nanogold/Silver Enhancement Microarray Detection Method
1. Set up a 50-ml reaction mixture using Label It Biotin Labeling Kit (Mirus Bio, Madison, WI): 1× Mirus labeling buffer. 0.1–0.5 mg of purified ssDNA. 3–5 ml of Mirus Label It Biotin reagent. 2. Incubate the reaction at 37°C for 1 h. 3. Clean up biotinylated DNA samples using the water-equilibrated CentriSep columns.
Microarrays Detection of Bacteria
85
4. Dry biotinylated DNA samples under vacuum. 5. Reconstituted in the Micromax hybridization buffer III (Perkin-Elmer, Boston, MA). 6. Adjust concentration of ssDNA to 0.05 mM. 7. Store the hybridization target at −20°C until the use. 3.7. DNA Microarray Design and Bioinformatics in Microarray Design and Use
The primary tools for bioinformatic analysis in microarray design and use include sequence analysis (revolving around melting temperature prediction), sequence search, sequence alignment, and oligo design for PCR primer and probe applications. Sequence alignment includes pairwise local alignment and multiple sequence alignment.
3.7.1. Local Alignment Using Oligo Design
The Oligo Design alignment tool is primarily for alignment of short sequences (maximum length of a few thousand bp). The advantage of using this tool is that it is integrated with a set of sequence manipulation tools that allow many tasks to be completed within a single environment. 1. Load the query and target sequences into either one of the two input windows on the main page. 2. Click Align. 3. Results are presented including a map of the alignment and the expectation value.
3.7.2. Local Alignment Using LAlign
LAlign provides a comparable but alternative search algorithm to BLAST and thus, together, they provide information to better understand the alignment output. 1. Open the web interface for LAlign (http://fasta.bioch.virginia. edu/fasta_www2/fasta_www.cgi?rm=lalign). 2. Copy the query sequence (i.e., the probe or primer) into the query sequence text box, which accepts either FASTA format (including plain text) or accession/GI number. 3. Copy the target sequence into the second sequence input box. 4. Check the search parameters to make sure they match the application. In particular, use the DNA:DNA program for DNA oligo searches. The default parameters are appropriate for most simple searches. 5. Click Align Sequences. 6. Read the output to find the alignment score and other statistics; output includes a map of the alignment showing exact matches, gaps, and mismatches. The statistics include a bit score and an expectation value, indicating the significance of the match.
86
Volokhov et al.
3.7.3. Local Alignment Using Blast
BLAST is maintained by National Center for Biotechnology Information (NCBI) as the primary alignment tool provided. BLAST is widely used and referenced. 1. Open the Web interface for LAlign (http://www.ncbi.nlm. nih.gov/blast/Blast.cgi). 2. Select the “nucleotide blast” program for DNA searches. 3. Enter the accession/GI number or FASTA sequence for the query sequence (i.e., the probe or primer). 4. Specify an NCBI database to search against. There are many options at this step, and the choices have a significant effect on the speed of the search and the turnaround time. Options range from the entire NCBI nucleotide collection (nr/nt – for nonredundant nucleotides) to organism-specific sequences. Organism-specific searches can be run by first selecting the Others database option and then entering organism-specific text in the Organism input box. After typing a few letters, the program provides a suggested list of organisms and groups of organisms. Alternatively, the taxid (taxonomy identifier) of an organism can be entered directly. 5. The results are provided (after a computational delay) in several forms. Particularly if a large database was searched, the results can be lengthy. Two key statistics are provided for each sequence match reported: (1) the bit score and (2) the expectation value. The bit score equals twice the sequence length for an exact match. Mismatches and gaps reduce the bit score. Thus, a high bit score is indicative of a close match. The expectation value is an estimate of the probability that the match that was found could have arisen by chance. Thus, a low expectation value is important to conclude that a search result represents homology and not just a random match. The expectation value depends primarily on the length of the database since the chance of a random match increases when the database is large.
3.7.4. Multiple Alignment Using ClustalX
Many multiple alignment algorithms and software packages are available (38). In our work, we used Clustal because it is freely available and well regarded (44, 45). 1. Start the ClustalX program. 2. Obtain the sequences to be aligned in FASTA format. NCBI is a good source for sequence data. One way to start the process is to get a target gene sequence for one organism. Then, use BLAST to find that gene region in other organisms. Once the gene is located, the relevant sequence can be downloaded in FASTA sequence with a few clicks. Append all of the FASTA sequences to be aligned into a single text file. The FASTA header line marks the start of a new sequence.
Microarrays Detection of Bacteria
87
3. From within ClustalX, choose File|Load Sequences. This step displays a color coded line for each sequence where each color represents a base. 4. Choose Alignment|Do Complete Alignment to initiate the multiple alignment computation. The computation may take some time (easily a few minutes), depending on the size and number of sequences in the input file. After the calculation completes, the sequences are typically reordered into groups representing the closest sequence similarity. In addition, the sequence display shows shifted sequences according to the calculated alignment. Assuming significant similarity, the color coded sequence display should now show unicolor columns where identical bases are aligned. The alignment may also show gaps and mismatches in some columns. 5. The aligned sequences can be saved in several file formats for further analysis. The GCG/MSF format creates a flat file that is easily readable by other programs. 3.7.5. PCR Primer Design
PCR primers are designed to amplify targets for microarray analysis. The primers are designed to anneal, one to each strand of the target (and not to any other sequence) and to amplify only the target sequences so primer design is based on knowledge of the sequence of the target, as well as other similar sequences in the genome and thus is a bioinformatics task. 1. Load target sequence in input window. 2. Specify a series of primer selection parameters including length range, melting temperature range, maximum length of tandem repeats, amplicon length range, and maximum DG for hairpins, 3¢-end annealing, self-annealing, and primer pair annealing. 3. Click Find Primers. 4. The search takes some time, depending on the target length. The results are presented as a series of primer pairs that meet all criteria. Sample results are shown in Fig. 5.
3.7.6. Primer Pair Evaluation
1. Input forward and reverse primers (these boxes are filled automatically if the user navigated to the primer pair evaluation from the primer selection tool). 2. Input the target sequence. 3. Click Check Primer Set. 4. The results are produced as a document that can be scrolled through or printed. An example of the output of this tool is provided as Fig. 6.
88
Volokhov et al.
PCR Primer Analysis Summary Input summary: Primer length range: Melting temperature target (°C): Max dG for hairpin (kcal/mol): Max number of tandem pairs: Max dG for self anneal (kcal/mol): Target start position: Max amplicon length: Eliminate neighbor primers:
18 to 25 55 +/- 1 2 1 2 1000 1
Max length of tandem repeats: 3 Max dG of 5 bp at 3' end (kcal/mol): 3 Max dG for primer pair anneal (kcal/mol): 2 Target end position: Min amplicon length: 200
Each forward primer has a group of reverse primers listed below it # Amp Len Pos. Length Sequence 2 349 20 ATGTATTTACTTTCCAAATT ----- 754 1102 20 ATGAAGAACTATCGTATTAT ----- 434 782 19 ACTTCTTATTGCTCATTAT 759 20 GGATAATAATGAGCAATAAG 55.50 ----- 748 1506 19 TCCATCGTATCTTTAATAG 768 19 TGAGCAATAAGAAGTTAAT 55.56 ----- 610 1377 18 TATCCTCCAGATTCATTA 1181 22 GAAACAATTATTATGTCTTTAA 55.34 ----- 326 1506 19 TCCATCGTATCTTTAATAG 1230 18 CAGAAGGTTCATTTAGTA 54.44 ----- 277 1506 19 TCCATCGTATCTTTAATAG 1232 19 GAAGGTTCATTTAGTAGTA 54.96 ----- 275 1506 19 TCCATCGTATCTTTAATAG
Tm(C) 54.92 55.25 55.30
6
54.91
7
54.78
9
54.91
10
54.91
11
54.91
Fig. 5. Primer search within the piplc gene of Bacillus thuringiensis (gi:2815229).
3.7.7. Microarray Probe Design
Similar to PCR primers, probes are designed to anneal to specific targets for microarray analysis. The probes are designed to anneal only to the target sequences (and not to any other sequence). The melting temperature of all the probes in the arrays has to be very similar, and this requires knowledge of the sequence of the target, as well as other similar sequences in the genome. 1. Enter target sequence. 2. Specify a series of probe selection parameters including length range, melting temperature range, maximum length of tandem repeats, amplicon length range, and maximum DG for hairpins, 3¢-end annealing, and self-annealing. 3. Click on Oligo Probes. 4. Results are presented as a list of all probes that satisfy the input criteria. Sample results are shown in Fig. 7.
3.8. QC and Microarray Fabrication 3.8.1. Preparation of Cy-Labeled QC Oligonucleotides 3.8.2. Microarray Fabrication
1. Follow the recommendations described in the manufacturer’s protocol accompanying the FluoroLink Cy3 dye kit (Amersham Pharmacia Biotech, Inc.). 2. The efficiency of dye incorporation was determined by the UV–visible spectrum in the 220- to 700-nm range. 1. Adjust the concentration of each oligoprobe to 100 mM in 50% dimethyl sulfoxide. Add the quality control oligonucleotide (QCprb) to a final concentration of 10 mM.
Microarrays Detection of Bacteria Primer Length Forward primer length = Reverse primer length = Target length = 1563
20 19
Amplicon Details Forward primer location (5' end) = Reverse primer location (5' end) = Amplicon length = 434 Primer Dimer Evaluation Length of 3' complementarity = dG (kcal/mol) = 0.
349 783
0
Hair Pin Evaluation Forward primer hairpin length = Reverse primer hairpin length = Melting Forward Reverse Forward Reverse
89
4 0
Temperature primer melting temp (C) = 54.92 primer melting temp (C) = 55.3 primer dG (kcal/mol) = 23.96 primer dG (kcal/mol) = 23.51
Potential Primer Pair Annealing 5' ATGTATTTACTTTCCAAATT 3' dG (kcal/mol) = 1. 3' TATTACTCGTTATTCTTCA 5' Forward Primer Potential Hairpin Formation 5' ATGTATTTACTTTCCAAATT 3' dG (kcal/mol) = 1.52 Potential Self Annealing 5' ATGTATTTACTTTCCAAATT 3' dG (kcal/mol) = 1.52 3' TTAAACCTTTCATTTATGTA 5' Reverse Primer Potential Hairpin Formation Potential Self Annealing 5' ACTTCTTATTGCTCATTAT 3' dG (kcal/mol) = -0.47 3' TATTACTCGTTATTCTTCA 5'
Fig. 6. Primer pair analysis for the second primer pair found in Fig. 4.
Oligo Probe Summary Input summary: Primer length range: Melting temperature target (°C): Max dG for hairpin (kcal/mol): Max number of tandem pairs: Max dG for self anneal (kcal/mol): Target start position: Max amplicon length: #
Position
Length
1 2 3 4 5
411 412 413 414 415
20 20 21 20 19
18 to 25 55 +/- 1 2 1 2 1000
Max length of tandem repeats: 2 Max dG of 5 bp at 3' end (kcal/mol): 3 Max dG for primer pair anneal (kcal/mol): 2 Target end position: Min amplicon length: 200
Tm (°C)
Sequence
55.50 54.40 55.73 54.64 54.12
GGATAATAATGAGCAATAAG GATAATAATGAGCAATAAGA ATAATAATGAGCAATAAGAAG TAATAATGAGCAATAAGAAG AATAATGAGCAATAAGAAG
Fig. 7. Probe search results. Restrictive inputs were used to minimize the output. In particular, this gene has many tri-nucleotide repeats, which were eliminated here by restricting the maximum repeat length to 2.
90
Volokhov et al.
2. Transfer 10 ml of each oligoprobe solution into a 384-well PCR plate. 3. Deposit the oligoprobes onto the surface of silylated (AmineAldehyde) glass slides using a contact microspotting robotic system equipped with a microspotting pin (see Notes 1–3). 4. After spotting, place the slides for 10 min into Boekel Slide Moat Hybridization Incubator installed at 80°C. 5. Mark the hybridization (array) area(s) on other (back surface) side of the slide using a Diamond Pencil. 6. Block residual reactive groups on the slide surface by treating the slides with 0.25% NaBH4 fresh water solution for 15 min. 7. Remove traces of the NaBH4 solution by five washes of the slides with distilled water for 1 min each. 8. Remove traces of water from the slide surface by centrifugation of slides at 128 × g for 3 min. 3.9. Microarray Hybridization 3.9.1. Hybridization with Fluorescent-Labeled ssDNA or ssRNA Targets
Hybridization between microarray oligoprobes and fluorescently labeled ssRNA (ssDNA) samples is conducted in the 1× hybridization buffer at 53°C for 30 min. 1. Mix 10 ml of Cy5-labeled ssRNA (or ssDNA) sample (from Subheadings 3.5.1 and 3.5.2) with a Cy3-QC probe at molar ratio 10:1. Denature at 95°C for 1 min then place the tube in ice for few seconds. 2. Place a glass coverslip(s) on microarray slide to cover the hybridization (array) area completely. The glass coverslip will prevent evaporation of the reaction solution during incubation. 3. Carefully place the hybridization mixture on microarray slide [on the hybridization (array) area under the glass coverslip]. 4. Place the microarray slide into ArrayIt® Hybridization Cassette chamber and then place the chamber into Boekel/Grant Premier Digital Water Bath set at 53°C. Wait for 30 min. 5. After hybridization, wash the slides for 1 min in 6× SSC containing 0.2% Tween 20, followed by three consecutive washes (1 min each) with 6× SSC buffer, 2× SSC buffer, and 1× SSC buffer. 6. Remove traces of buffer from the slide surface by centrifugation at 128 × g for 3 min.
3.9.2. Hybridization with Biotynilated ssDNA Followed by Nanogold/ Silver Enhancement of Hybridization Signals
1. Denature 10 ml of biotin-labeled ssDNA hybridization target (from Subheading 3.6) at 95°C for 1 min. Chill on ice. 2. Place a glass coverslip(s) on microarray slide to cover the hybridization (array) area completely. The glass coverslip will prevent evaporation of the reaction solution during incubation.
Microarrays Detection of Bacteria
91
3. Carefully place the hybridization mixture on microarray slide [on the hybridization (array) area under the glass coverslip]. 4. Place the microarray slide into ArrayIt® Hybridization Cassette chamber and place the chamber into the Boekel/Grant Premier Digital Water Bath settled at 45°C. Wait for 30 min. 5. After hybridization, wash away coverslips using 6× SSC containing 0.1% Tween 20. 6. Wash slide once for 1 min with 6× SSC, 0.4% gelatin, 0.1% Tween 20 and three times for 30 s with 6× SSC, 0.1% Tween 20. Remove traces of the buffer by centrifugation at 800 rpm (128 RCF) for 3 min. 7. Place a glass coverslip(s) on microarray slide to cover the hybridization (array) area completely. 8. Add 6–10 ml of gold-labeled streptavidin (5 nm), diluted 1:5 (v/v) by 6× SSC, 0.4% gelatin, 0.1% Tween 20. 9. Incubate at room temperature for 30 min. 10. Wash slides once with 6× SSC solution containing 0.1 % Tween 20, two times with 6× SSC followed two more times with 0.6 M NaNO3. 11. Remove traces of buffer from the slide surface by centrifugation at 128 × g for 3 min. 12. Soak slides for 4–5 min into silver enhancement solution containing 1:1 (v/v) mixture of reagent A and B provided by Nanosphere. 13. Wash slides several times with tap water. Dry under air. 3.10. Fluorescent Signal Detection and Data Processing
1. Generate the fluorescent images of processed microarray slides using the array scanner.
3.11. Detection of Silver Enhanced Signals from Nanogold Particles
1. The resulting pictures of hybridization patterns were taken through a universal 3× magnifier lamp using any available digital camera.
2. Measure and compare the fluorescent signals from each spot using ScanArray Express software (see Note 3).
2. A black background was found to provide the most efficient visualization of silver spots on the array slide surface.
4. Notes 1. Quadruplication of each oligoprobe on microchip increases the reliability of mutation detection, reduces the probability of misinterpretation of microarray data that could occur due
92
Volokhov et al.
to potential failure during microarray fabrication, and allows for application of statistical approaches for analysis of microarray data. 2. During the spotting procedure, the humidity inside the microspotter environmental chamber should be maintained at 75–80%. The settings of the spotting program should be chosen to provide 450-mm intervals between spots (center to center). Training with technical experts in the field of microarray printing, slide scanning, and image analysis may be necessary. Alternatively, microarray slides may be customary printed using commercially available microarray service. 3. Training with the technical experts of Perkin-Elmer may be necessary before performing microarray data analysis using the ScanArray 5000 and ScanArray Express software systems. References 1. Gillespie, D. and S. Spiegelman, A quantitative assay for DNA-RNA hybrids with DNA immobilized on a membrane. J Mol Biol, 1965. 12(3): pp. 829–42. 2. Southern, E.M., Detection of specific sequences among DNA fragments separated by gel electrophoresis. J Mol Biol, 1975. 98(3): pp. 503–17. 3. Alwine, J.C., D.J. Kemp, and G.R. Stark, Method for detection of specific RNAs in agarose gels by transfer to diazobenzyloxymethylpaper and hybridization with DNA probes. Proc Natl Acad Sci USA, 1977. 74(12): pp. 5350–4. 4. Engvall, E. and P. Perlmann, Enzyme-linked immunosorbent assay (Elisa) quantitative assay of immunoglobulin-G. Immunochemistry, 1971. 8(9): pp. 871–4. 5. Kafatos, F.C., C.W. Jones, and A. Efstratiadis, Determination of nucleic acid sequence homologies and relative concentrations by a dot hybridization procedure. Nucleic Acids Res, 1979. 7(6): pp. 1541–52. 6. Ekins, R., F. Chu, and J. Micallef, High specific activity chemiluminescent and fluorescent markers: their potential application to high sensitivity and ‘multi-analyte’ immunoassays. J Biolumin Chemilumin, 1989. 4(1): pp. 59–78. 7. Ekins, R., F. Chu, and E. Biggart, Multispot, multianalyte, immunoassay. Ann Biol Clin (Paris), 1990. 48(9): pp. 655–66. 8. Ekins, R. and F.W. Chu, Microarrays: their origins and applications. Trends Biotechnol, 1999. 17(6): pp. 217–8.
9. Gupta, R.S. and E. Griffiths, Critical issues in bacterial phylogeny. Theor Popul Biol, 2002. 61(4): pp. 423–34. 10. Gupta, R.S., The phylogeny and signature sequences characteristics of Fibrobacteres, Chlorobi, and Bacteroidetes. Crit Rev Microbiol, 2004. 30(2): pp. 123–43. 11. Karlin, S. and L. Brocchieri, Heat shock protein 70 family: multiple sequence comparisons, function, and evolution. J Mol Evol, 1998. 47(5): pp. 565–77. 12. Gupta, R.S. and V. Johari, Signature sequences in diverse proteins provide evidence of a close evolutionary relationship between the Deinococcus-thermus group and cyanobacteria. J Mol Evol, 1998. 46(6): pp. 716–20. 13. Ludwig, W. and K.H. Schleifer, Bacterial phylogeny based on 16S and 23S rRNA sequence analysis. FEMS Microbiol Rev, 1994. 15(2–3): pp. 155–73. 14. Anthony, R.M., T.J. Brown, and G.L. French, Rapid diagnosis of bacteremia by universal amplification of 23S ribosomal DNA followed by hybridization to an oligonucleotide array. J Clin Microbiol, 2000. 38(2): pp. 781–8. 15. Olsen, G.J., et al., Microbial ecology and evolution: a ribosomal RNA approach. Annu Rev Microbiol, 1986. 40: pp. 337–65. 16. Hartman, A.B., et al., Sequence and molecular characterization of a multicopy invasion plasmid antigen gene, ipaH, of Shigella flexneri. J Bacteriol, 1990. 172(4): pp. 1905–15. 17. Nagano, I., et al., Detection of verotoxinproducing Escherichia coli O157:H7 by multiplex
Microarrays Detection of Bacteria
18.
19.
20. 21. 22.
23.
24.
25.
26.
27.
28.
29. 30.
31.
polymerase chain reaction. Microbiol Immunol, 1998. 42(5): pp. 371–6. Yu, J. and J.B. Kaper, Cloning and characterization of the eae gene of enterohaemorrhagic Escherichia coli O157:H7. Mol Microbiol, 1992. 6(3): pp. 411–7. Sazhin, O.V. and S.F. Borisov, Accommodation coefficient of tangential momentum on atomically clean and contaminated surfaces. J Vac Sci Technol A, 2001. 19(5): pp. 2499–503. Chizhikov, V., et al., Microarray analysis of microbial virulence factors. Appl Environ Microbiol, 2001. 67(7): pp. 3258–63. Volokhov, D., et al., Identification of Listeria species by microarray-based assay. J Clin Microbiol, 2002. 40(12): pp. 4720–8. Volokhov, D., et al., Microarray-based identification of thermophilic Campylobacter jejuni, C. coli, C. lari, and C. upsaliensis. J Clin Microbiol, 2003. 41(9): pp. 4071–80. Al-Khaldi, S.F., et al., Genotyping of Clostridium perfringens toxins using multiple oligonucleotide microarray hybridization. Mol Cell Probes, 2004. 18(6): pp. 359–67. Sergeev, N., et al., Simultaneous analysis of multiple staphylococcal enterotoxin genes by an oligonucleotide microarray assay. J Clin Microbiol, 2004. 42(5): pp. 2134–43. Volokhov, D., et al., Identification of Bacillus anthracis by multiprobe microarray hybridization. Diagn Microbiol Infect Dis, 2004. 49(3): pp. 163–71. Sergeev, N., et al., Multipathogen oligonucleotide microarray for environmental and biodefense applications. Biosens Bioelectron, 2004. 20(4): pp. 684–98. Wilson, K.H., et al., High-density microarray of small-subunit ribosomal DNA probes. Appl Environ Microbiol, 2002. 68(5): pp. 2535–41. Wilson, W.J., et al., Sequence-specific identification of 18 pathogenic microorganisms using microarray technology. Mol Cell Probes, 2002. 16(2): pp. 119–27. Volokhov, D., et al., Microarray analysis of erythromycin resistance determinants. J Appl Microbiol, 2003. 95(4): pp. 787–98. Call, D.R., et al., Identifying antimicrobial resistance genes with DNA microarrays. Antimicrob Agents Chemother, 2003. 47(10): pp. 3290–5. Wade, M.M., et al., Accurate mapping of mutations of pyrazinamide-resistant Mycobacterium tuberculosis strains with a scanning-frame oligonucleotide microarray. Diagn Microbiol Infect Dis, 2004. 49(2): pp. 89–97.
93
32. SantaLucia, J., Jr., A unified view of polymer, dumbbell, and oligonucleotide DNA nearestneighbor thermodynamics. Proc Natl Acad Sci USA, 1998. 95(4): pp. 1460–5. 33. Sugimoto, N., et al., Improved thermodynamic parameters and helix initiation factor to predict stability of DNA duplexes. Nucleic Acids Res, 1996. 24(22): pp. 4501–5. 34. Breslauer, K.J., et al., Predicting DNA duplex stability from the base sequence. Proc Natl Acad Sci USA, 1986. 83(11): pp. 3746–50. 35. Google, Google Directory>Science>Biology> Bioinformatics>Software. 2008; Available from: http://www.google.com. 36. Herold, K.E. and A. Rasooly, Oligo Design: a computer program for development of probes for oligonucleotide microarrays. Biotechniques, 2003. 35(6): pp. 1216–21. 37. Smith, T.F. and M.S. Waterman, Identification of common molecular subsequences. J Mol Biol, 1981. 147(1): pp. 195–197. 38. Lyon, Alignment software. 2008; Available from: http://pbil.univ-lyon1.fr/alignment. html. 39. Huang, X.Q. and W. Miller, A time-efficient, linear-space local similarity algorithm. Adv Appl Math, 1991. 12(3): pp. 337–357. 40. Pearson, W., LAlign program, part of the FASTA program set. 2008; Available from: http://fasta. bioch.virginia.edu/fasta_www2/fasta_ www.cgi?rm=lalign, http://www.ch.embnet.org/ software/LALIGN_form.html. 41. Pearson, W.R. and D.J. Lipman, Improved tools for biological sequence comparison. Proc Natl Acad Sci USA, 1988. 85(8): pp. 2444–8. 42. Altschul, S.F., et al., Issues in searching molecular sequence databases. Nat Genet, 1994. 6(2): pp. 119–29. 43. Altschul, S.F., et al., Basic local alignment search tool. J Mol Biol, 1990. 215(3): pp. 403–10. 44. Larkin, M.A., et al., Clustal W and Clustal X version 2.0. Bioinformatics, 2007. 23(21): pp. 2947–8. 45. Clustal, Clustal download site. 2008; Available from: http://www.clustal.org/. 46. Google, Science>Biology>Biochemistry and Molecular Biology>Methods and Techniques> PCR>Software. 2008; Available from: http:// www.google.com. 47. Bodrossy, L. and A. Sessitsch, Oligonucleotide microarrays in microbial diagnostics. Curr Opin Microbiol, 2004. 7(3): pp. 245–54. 48. Schena, M., et al., Quantitative monitoring of gene expression patterns with a complementary
94
49.
50.
51.
52.
53. 54.
Volokhov et al. DNA microarray. Science, 1995. 270(5235): pp. 467–70. Badiee, A., et al., Evaluation of five different cDNA labeling methods for microarrays using spike controls. BMC Biotechnol, 2003. 3(1): p. 23. Lyon, L.A., M.D. Musick, and M.J. Natan, Colloidal Au-enhanced surface plasmon resonance immunosensing. Anal Chem, 1998. 70(24): pp. 5177–83. Lian, W., et al., Ultrasensitive detection of biomolecules with fluorescent dye-doped nanoparticles. Anal Biochem, 2004. 334(1): pp. 135–44. Raychaudhuri, S., et al., Basic microarray analysis: grouping and feature reduction. Trends Biotechnol, 2001. 19(5): pp. 189–93. Jarvinen, A.K., et al., Are data from different gene expression microarray platforms comparable? Genomics, 2004. 83(6): pp. 1164–8. Yauk, C.L., et al., Comprehensive comparison of six microarray technologies. Nucleic Acids Res, 2004. 32(15): p. e124.
55. Aguilar, Z.P., W.R. Vandaveer, and I. Fritsch, Self-contained microelectrochemical immunoassay for small volumes using mouse IgG as a model system. Anal Chem, 2002. 74(14): pp. 3321–9. 56. Sergeev, N., et al., Microarray analysis of Bacillus cereus group virulence factors. J Microbiol Methods, 2005. 57. Sergeev, N., et al., Microarray analysis of Bacillus cereus group virulence factors. J Microbiol Methods, 2006. 65(3): pp. 488–502. 58. AOAC International, Bacteriological analytical manual (BAM). 8th ed. (revision A). 1998, Gaithersburg, MD: AOAC International. 59. Sambrook, J. and D.W. Russell, Molecular cloning: a laboratory manual. 3rd ed. 2001, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. 60. Chachaty, E. and P. Saulnier, Isolation chromosomal DNA from bacteria, in The nucleic acid protocols: handbook, R. Rapley, Editor. 2000, Totowa, NJ: Humana Press Inc. pp. 29–32.
Chapter 4 Protein Microarrays Printed from DNA Microarrays Oda Stoevesandt, Mingyue He, and Michael J. Taussig Abstract Protein arrays are miniaturised and highly parallelised formats of interaction-based functional protein assays. Major bottlenecks in protein microarraying are the limited availability and high cost of purified, functional proteins for immobilisation and the limited stability of immobilised proteins in their functional state. In contrast, protein-coding DNA is readily available by PCR, and DNA arrays can be stored over prolonged times without deterioration. This chapter presents a method for the rapid and economical “printing” of replicate protein microarrays directly from a single DNA array template using cell-free protein synthesis, termed “DNA array to protein array,” DAPA. The procedure is a truly enabling technology, making customised protein microarrays affordable for laboratories with no access to routine microarray spotting. The experimental effort involved for the printing of a protein array from the template DNA array is comparable to the assembly of a Western blot. Key words: Protein array, Protein microarray, Cell-free protein synthesis, Protein immobilisation
1. Introduction Proteomics and systems biology require technologies for highthroughput, multiplexed analysis of protein expression levels and protein function. Protein arrays are miniaturised and highly parallelised formats of interaction-based assays using a minimum amount of samples and reagents. Among the interactions that can be assayed are protein–protein, protein–antibody, protein–nucleic acid, and protein–small molecule (1). Therefore, protein arrays are increasingly applied for profiling protein expression, identification of biomarkers for diagnostics, studying protein interactions, and discovery of protein signalling pathways (2). One of the impediments limiting the more widespread use of protein arrays is the preparative effort involved in expressing and purifying large numbers of functional proteins for immobilisation.
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_4, © Springer Science+Business Media, LLC 2011
95
96
Stoevesandt, He, and Taussig
Commercially available protein arrays are therefore very costly. Another major problem is the limited stability of functional proteins in immobilised state on an array, especially when the array is stored over time periods of days to weeks before usage. Cell-free protein synthesis systems have been used to overcome these problems (3–8). These systems direct the synthesis of proteins directly from added DNA templates, providing a rapid means for conversion of genetic information into functional proteins. As cell-free protein synthesis systems are open for addition of further components, they allow the creation of optimised environments, for example correct protein folding or post-translational modifications (9). By coupling cell-free synthesis and simultaneous in situ protein immobilisation on the array surface, we have developed two cell-free methods, termed PISA (3, 6) and DAPA (7), for making protein arrays on demand directly from PCR DNA molecules. Here, we describe the details of the recent DAPA procedure (Fig. 1). In DAPA, a slide with a DNA microarray encoding a set of tagged proteins is assembled face-to-face with a second slide, functionalised with the tag-capturing reagent. A membrane soaked
Fig. 1. Principle (a) and sample results (b) of DAPA.
Protein Microarrays Printed from DNA Microarrays
97
with an in vitro transcription and translation system is positioned between the two slide surfaces. Tagged proteins synthesised from the immobilised DNA diffuse towards the capture slide surface where they adhere, creating the protein array corresponding to the DNA array template, with well-defined protein spot morphologies described by Gaussian intensity profiles. The DNA template slide can be re-used in further cycles of DAPA. We have demonstrated that at least 20 cycles of protein array printing are feasible using a single DNA template array (7). In this protocol, we describe 1. The PCR-based generation of constructs for cell-free expression in DAPA, encoding proteins with a C-terminal double (His)6 tag (10) for immobilisation. 2. The immobilisation of the PCR constructs on epoxy-activated slides for generation of template DNA microarrays. 3. The DAPA procedure itself for printing a protein microarray from the template DNA array, using an Escherichia coli (E. coli) lysate-based system for cell-free protein synthesis.
2. Materials 2.1. DNA Encoding Proteins of Interest
The protocol below assumes the availability of DNA (cDNA or cloned) encoding the proteins of interest for arraying by DAPA.
2.2. Primers for Construction of DNA Templates for DAPA
The primers described here are suitable for generation of constructs for cell-free expression in E. coli S30 extracts. The numbering follows the primer designations in the overview schematic drawing of the construction process (Fig. 2). 1. T7-for: 5′-GATCTCGATCCCGCG-3′: Forward primer for generating the fragment with T7 sequence (see Subheading 2.3.1) in combination with T7-rev (primer 2). 2. T7-rev: 5′-CATGGTATATCTCCTTCTTAAAG-3′: Reverse primer for generating the fragment with T7 sequence in combination with T7-for (primer 1). 3. GENE-for: 5′-CTTTAAGAAGGAGATATACCATG(N)15–25-3′: Forward primer for PCR amplification of a target gene in combination with GENE-rev (primer 4). It contains a sequence (underlined) overlapping with the T7 fragment and 15–25 nucleotides from the 5′-sequence of the target gene. 4. GENE-rev: 5′-CACCGCCTCTAGAGCG(N)15–25-3′: Reverse primer for PCR amplification of a target gene in combination with GENE-for (primer 3). It contains a sequence (underlined)
98
Stoevesandt, He, and Taussig
Fig. 2. PCR strategy to generate constructs for cell-free expression on DAPA arrays. The primers used are as follows: (1) T7-for, (2) T7-rev, (3) GENE-for, (4) GENE-rev, (5) LTT-for, (6) LTT-rev, (7) Cy5-T7-for (fluorophore-coupled derivative of 1), (8) NH2-LTT-rev (NH2-functionalised derivative of 6).
overlapping with the linker sequence and 15–25 nucleotides complementary to the 3′-sequence of the target gene. 5. LTT-for: 5′-GCTCTAGAGGCGGTGGC-3′: Forward primer for PCR generation of a fragment encoding linker, protein tag, and termination region (see Subheading 2.3.2) in combination with LTT-rev (primer 6). 6. LTT-rev: 5′-TCCGGATATAGTTCCTCC-3′: Reverse primer for PCR generation of a fragment encoding linker, protein tag, and termination region in combination with the LTT-for (primer 5).
Protein Microarrays Printed from DNA Microarrays
99
7. Cy5-T7-for: 5′-Cy5-GATCTCGATCCCGCG-3′: Fluorophorecoupled T7-for (primer 1). Forward primer for PCR amplification of a full length DAPA construct in combination with NH2-LTT-rev (primer 8). The Cy5-label allows detection of the immobilised PCR product on DNA microarrays. 8. NH2-LTT-rev: 5′-NH2-TCCGGATATAGTTCCTCC-3′: NH2functionalised LTT-rev (primer 6). Reverse primer for PCR amplification of a full length DAPA construct in combination with Cy5-T7-for (primer 7). The NH2-group allows immobilisation of the PCR product on epoxy-activated slides. 2.3. Plasmids Encoding Generic Elements for Cell-Free Protein Expression 2.3.1. Plasmid Including the T7 Domain
2.3.2. Plasmid Encoding Linker, Double (His)6-tag and Termination Sequence
2.4. Further Reagents and Kits for Molecular Biology and Cell-Free Protein Expression
The control plasmid included in the RTS100 E. coli HY kit (Roche) is used. It contains the T7 promoter (underlined), the ribosome-binding site (underlined italics), and the start codon ATG (bold) within the following sequence: 5 ′ G AT C T C G AT C C C G C G A A AT TA ATA C G A C T C A C TATA G G G A G A C C A C A A C G G T T T C C C T C TA G A AATAATTTTGTTTAACTTTAAGAAGGAGATA TACCATG-3′. Using standard molecular biology techniques a plasmid is created, containing a DNA insert encoding a flexible 19 amino acid linker (lowercase), a double (His)6-tag (underlined), two consecutive stop codons (bold), a poly(A) tail, and a transcription termination region (italics) (3). The double-(His)6 tag used here has shown improved affinity for arraying proteins on Ni-NTA modified surfaces compared to a conventional single-(His)6 tag (3, 10, 11). The detailed sequence of the insert is as follows: 5′-GCTCTAGAggcggtggctctggtggcggttc tggcggtggcaccggtggcggttctggcggtggcAAAC G G G C T G AT G C T G C A C AT C A C C AT C A C C AT C A CTCTAGAGCTTGGCGTCACCCGCAGTTCGGTGGTCA CCACCACCACCACCACTAATAA(A)28CCGCTGAGCAAT A A C TA G C ATA A C C C C T T G G G G C C T C TA A A C G G G TCTTGAGGGGTTTTTTGCTGAAAGGAGGAAC TATATCCGGA-3′. 1. Desoxy nucleotides (Sigma, UK). 2. Taq DNA polymerase and buffers (Qiagen, UK). 3. GenElute™ Gel Extraction kit (Sigma, UK). 4. GenElute™ PCR Clean-Up kit (Sigma, UK). 5. RTS100 E. coli HY (5 prime, UK) cell-free protein expression system for production of up to 20 mg of protein in a 50 ml reaction.
100
Stoevesandt, He, and Taussig
6. Durapore 0.22 mm membrane filters (Millipore, UK). 7. Ni-NTA-coated microscope slide (Xenopore, USA). 8. Nexterion™ slide E (epoxysilane coated) (Schott Nexterion, UK). 2.5. Buffers for DNA Microarray Spotting
1. PBS-Tween: PBS pH 7.4, 0.05% Tween20. 2. 6× Spotting buffer: 300 mM sodium phosphate, pH 8.5. 3. Quenching buffer: 0.1 M Tris–HCl, pH 9.0. Add ethanolamine (Sigma, UK) to a final concentration of 50 mM immediately before use. 4. 0.1% (v/v) Tween-20. 5. 1 mM HCl. 6. 100 mM KCl. 7. Saturated NaCl solution for humidified chamber: 30% NaCl in H2O, boil to dissolve, cool down.
3. Methods 3.1. Generation of PCR Constructs for Cell-Free Expression in DAPA
3.1.1. Generation of PCR Fragments for Assembly
The PCR constructs contain the essential elements for cell-free transcription and translation, including a T7 promoter, a ribosome-binding site, start and stop codons, a poly(A) tail, and a transcription termination region. For immobilisation of proteins on a Ni-NTA modified surface in DAPA, a sequence encoding a linker and a double (His)6-tag is included downstream of the target gene (see Note 1). To simplify the PCR construction, these common upstream and downstream elements are assembled in plasmids which are used as templates for PCR. Figure 2 shows the PCR construction process in overview. 1. Set up standard 50 ml PCR reactions (see Note 2): (a) 5 ml 10× PCR buffer (b) 10 ml 5× Q solution (c) 4 ml dNTP (2.5 mM each) (d) 1.5 ml Forward primer (16 mM) (e) 1.5 ml Reverse primer (16 mM) (f) 1–10 ng Template (g) 1.25 U Taq (h) H2O to 50 ml
Protein Microarrays Printed from DNA Microarrays
101
With the following combinations of primers and templates: Forward primer
Reverse primer
T7-for
Template
PCR product (Fig. 2)
T7-rev
Plasmid encoding T7
Upstream T7 fragment (101 bp)
GENEfor
GENE-rev
Gene of interest
Target gene fragment (encoding protein of interest with deleted stop codon and short flanking 5′ and 3′ sequences)
LTT-for
LTT-rev
Plasmid encoding linker, tag, termination sequences
Downstream fragment (encoding linker, double (His)6 tag and termination sequences; 249 bp)
PCR programme: 30 Cycles: 94°C, 30 s → 54°C, 30 s → 72°C, 80 s 1 Cycle: 72°C, 480 s End: hold 10°C 2. Analyse the PCR products by electrophoresis on a 1% agarose gel. 3. Isolate the expected fragments by extracting them from the gel (use kit as specified by manufacturer). 4. Determine the concentration and purity of the cleaned PCR product by absorption at 260 and 280 nm or by gel electrophoresis and comparison with DNA marker bands. 3.1.2. Assembly of PCR Fragments to Complete Construct for Cell-Free Expression
1. Set up a 25 ml assembly PCR reaction as follows: (a) 2.5 ml 10× PCR buffer (b) 5 ml 5× Q solution (c) 1 ml dNTP (2.5 mM each) (d) Mix of upstream T7 fragment, target gene fragment, and downstream fragment in equimolar ratio (total DNA 50–100 ng) (e) 0.67 U Taq (f) H2O to 25 ml PCR programme for fragment assembly: 8 Cycles: 94°C, 30 s → 54°C, 60 s → 72°C, 60 s End: hold 10°C
102
Stoevesandt, He, and Taussig
2. Amplify the assembled product by a second 50 ml PCR: (a) 5 ml 10× PCR buffer (b) 10 ml 5× Q solution (c) 4 ml dNTP (2.5 mM each) (d) 1.5 ml Primer T7-for (16 mM) (e) 1.5 ml Primer LTT-rev (16 mM) (f) 2 ml Product of assembly PCR product (step 1 above) (g) 1.25 U Taq (h) H2O to 50 ml PCR programme: 30 Cycles: 94°C, 30 s → 54°C, 60 s → 72°C, 80 s 1 Cycle: 72°C, 480 s End: hold 10°C 3. Analyse the PCR product by electrophoresis on a 1% agarose gel and purify the DNA if required (see Note 3). Construct identity can be further confirmed by PCR mapping using primers annealing at various positions along the desired sequence (see Note 4). The resulting PCR construct may be stored at −20°C for at least 6 months. 3.1.3. Re-amplification of Construct with Labelled Primers for Immobilisation on Arrays
1. Set up standard 50 ml PCR as in step 2 Subheading 3.1.2, but with labelled primers: (a) Cy5-T7-for (instead T7-for) (b) NH2-LTT-rev (instead LTT-rev) 2. Analyse the PCR product for correct size by agarose gel electrophoresis, purify on a spin column (use kit as specified by manufacturer), and elute in H2O. 3. Determine the concentration and purity of the cleaned PCR product by absorption at 260 and 280 nm or by gel electrophoresis and comparison with DNA marker bands. A DNA concentration of 100–200 ng/ml is recommended for DAPA template array spotting (see Subheading 3.2) (see Note 5).
3.2. Generation of DNA Arrays as Templates for DAPA
1. Add one volume of 6× spotting buffer to five volumes of the labelled PCR product (see Subheading 3.1.3). 2. Spot DNA samples on epoxysilane slides (see Note 6) with spot-to-spot distances of 1 mm and volumes per spot of 5–10 nl. 3. Incubate spotted slides in a humidified chamber (see Note 7) at RT for 1 h. 4. Incubate slides at 60°C for 30 min.
Protein Microarrays Printed from DNA Microarrays
103
5. Wash the slides at RT as follows: (a) 1× with 0.1% Tween-20 for 5 min (b) 2× with 1 mM HCl for 2 min (c) 1× with 100 mM KCl for 10 min (d) 1× with ddH2O for 1 min 6. Quench remaining epoxy groups by incubating slides in 0.1 M Tris–HCl pH 9.0, 50 mM ethanolamine at 50°C for 15 min. 7. Wash slides with ddH2O for 1 min and dry by pressurised air. 8. Scan slides in microarray scanner to confirm immobilisation of Cy5-labelled DNA. The slides are ready for use and can be stored in the dark at 4°C. 3.3. DAPA: Printing Protein Arrays from the Template DNA Array
Use a glass slide holder similar to a prototype designed by us. Figure 3 shows a schematic cross-section of the holder and the DAPA assembly process. 1. Cut a Durapore membrane filter large enough to cover the area of the DNA template array. 2. Prepare 10 ml E. coli cell-free lysate (according to the instructions of the manufacturer) for every 1 cm² of the membrane. 3. Assemble the DAPA sandwich in the slide holder in the following order (numbering refers to Fig. 3): (a) Bottom plate (1) (b) Rubber spacer (2) (c) Layer of parafilm (3) (d) Ni-NTA-coated slide (4), with the protein-capturing surface facing up (see Note 6)
Fig. 3. Schematic cross-section of DAPA assembly. (1) Bottom plate, (2) rubber spacer, (3) parafilm, (4) Ni-NTA-coated slide, protein-capturing surface facing up, (5) membrane filter soaked with cell-free lysate, (6) DNA array template slide, DNA surface facing down, (7) parafilm, (8) rubber spacer, (9) top plate.
104
Stoevesandt, He, and Taussig
(e) Cell-free lysate, distributed on the surface of the Ni-NTA slide (f) Membrane filter (5), allowing it to soak up the lysate (see Note 8) (g) DNA array template slide (6), with DNA surface facing down (h) Layer of parafilm (7) (see Note 9) (i) Rubber spacer (8) (j) Top plate (9) Ensure even pressure on the slide sandwich. 4. Incubate the assembled slide holder at 30°C for 2–4 h (see Note 10). 5. Disassemble the slide sandwich and (a) Wash the Ni-NTA slide with the DAPA protein array on it with PBS-Tween. Do not dry the DAPA array before application in order to avoid denaturation of the immobilised proteins. (b) Wash the DNA template slide with ddH2O, dry by pressurised air, and store at 4°C for use in further DAPA cycles. 3.4. Downstream Usage of DAPA Arrays
The DAPA protein array is now ready for immediate use in an assay of choice. Downstream handling protocols and detection protocols will be dependent on the individual application. To control for expression of proteins and their immobilisation on the array, it is advisable to perform immunofluorescence staining with appropriate reagents against the arrayed proteins (Fig. 1b) (7).
4. Notes 1. The location of a tag should be tested at both the N- and C-terminus of the protein to make sure it is accessible and does not affect protein activity. C-terminal immobilisation tags are preferable, as their presence guarantees that the entire protein is synthesised. 2. The T7 fragment and the downstream fragment are usually produced in a larger quantity and stored at −20°C for use as required. 3. In case multiple PCR bands are generated, the PCR fragment with the expected size is isolated by gel extraction and used as the template for PCR re-amplification (see Subheading 3.1.3). In general, PCR fragments without purification can be directly used for protein synthesis in cell-free systems.
Protein Microarrays Printed from DNA Microarrays
105
4. A construct can be confirmed by PCR mapping, which is performed by using a combination of various primers annealing at different positions in the construct. If all PCR products have the expected size, it suggests the correct construction. 5. If the eluted PCR product is below this range, it can be concentrated in a vacuum centrifuge. The Cy5-label of the purified PCR product is usually not detectable by absorption (as there is only one Cy5 fluorophore per DNA double strand). 6. Mark glass slides and their orientation with a diamond-tipped pen. Any possible glass splinters or dust from the slide surfaces can be removed by using pressurised air. 7. This can be prepared using a box containing saturated NaCl solution and a raised platform for incubation of slides. 8. The soaking process takes a few seconds. It is crucial to avoid drying of the cell-free lysate within the membrane filter. 9. The parafilm must form an airtight seal around the slide sandwich (Fig. 3) in order to prevent evaporation of cell-free lysate soaked in the membrane filter between the two slides. 10. The incubation time may be varied depending on the downstream applications.
Acknowledgements Research at the Babraham Institute is supported by Biotechnology and Biological Sciences Research Council (BBSRC), UK. The Protein Technology Group at Babraham Bioscience Technologies is a partner in the EC FP6 CA 026008 ProteomeBinders, and in. References 1. Hall D A, Ptacek J, Snyder M. Protein microarray technology. Mech Ageing Dev 2007;128:161–167. 2. Bertone P, Snyder M. Advances in functional protein microarray technology. FEBS J 2005; 272:5400–5411. 3. He M, Taussig M J. Single step generation of protein arrays from DNA by cell-free expression and in situ immobilization (PISA method). Nucleic Acid Res 2001;29:e73. 4. Ramachandran N, Hainsworth E, Bhullar B, Eisenstein S, Rosen B, Lau A Y, Walter J C, LaBaer J. Self-assembling protein mircoarrays. Science 2004;305:86–90.
5. Angenendt P, Kreutzberger J, Glokler J, Hoheisel J D. Generation of high density protein microarrays by cell-free in situ expression of unpurified PCR products. Mol Cell Proteomics 2006;5: 1658–1666. 6. He M, Taussig M J. DiscernArray™ technology: a cell-free method for the generation of protein arrays from PCR DNA. J Immunol Methods 2003;274:265–270. 7. He M, Stoevesandt O, Palmer E A, Khan F, Ericsson O, Taussig M J. Printing protein arrays from DNA arrays. Nat Methods 2008;5: 175–177.
106
Stoevesandt, He, and Taussig
8. He M, Stoevesandt O, Taussig M J. In situ synthesis of protein arrays. Curr Opin Biotechnol 2008;19:4–9. 9. He M. Cell-free protein synthesis: applications in proteomics and biotech-nology. N Biotechnol 2008;Aug2025:126–132. 10. Khan F, He M, Taussig M J. A double-His tag with high affinity binding for protein immo-
bilisation, purification, and detection on Ni-NTA surfaces. Anal Chem 2006;78:3072–3079. 11. Steinhauer C, Wingren C, Khan F, He M, Taussig M J, Borrebaeck C A. Improved affinity coupling for antibody microarrays: engineering of double-(His)6-tagged single framework recombinant antibody fragments. Proteomics 2006;6:4227–4234.
Chapter 5 Lithographically Defined Two- and Three-Dimensional Tissue Microarrays Esther W. Gomez and Celeste M. Nelson Abstract Traditional methods to study normal and pathological development of tissues have been limited by difficulties in controlling experimental conditions and quantifying biological processes of interest. Here we describe methods to create microarrays of engineered tissues that enable controlled and quantitative investigations. Using soft lithography-based techniques, extracellular matrix proteins can be microcontact printed or micromolded to make two- and three-dimensional micropatterned scaffolds. The ultimate form and resulting properties of the tissue construct are dictated by the geometry of the patterned extracellular matrix components. This chapter describes elastomeric stamp fabrication, microcontact printing and micromolding of extracellular matrix proteins, cell culture in micropatterned substrata, and quantitative immunofluorescence analysis of micropatterned tissues. Key words: Tissue engineering, Microfabrication, Organotypic culture, Epithelial
1. Introduction Understanding the processes involved in both normal and pathological tissue development is crucial to the engineering of tissue constructs for therapeutic and diagnostic purposes (1). Studies in vivo are difficult to control, observe, and quantify. As a result, much effort has been directed toward culturing organs ex vivo. Despite progress made in this field, organ cultures tend to be difficult to maintain as they require fresh tissues and are often uncontrollable. Engineered tissues offer unique benefits over organ culture approaches by allowing for the investigation of developmental processes and cellular behaviors within tissues in a controlled and quantitative manner. By defining the properties of the extracellular matrix (ECM) environment and the specific
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_5, © Springer Science+Business Media, LLC 2011
107
108
Gomez and Nelson
biochemical factors that are presented, one can precisely control the spatial organization and behaviors of the cells that make up the engineered tissue. One can then study how properties of the tissue, such as geometry and form, contribute to the control of biological events including proliferation, apoptosis, gene expression, and differentiation. Geometry can be chosen to best recapitulate the system of interest with, for example, two-dimensional (2D) engineered tissues approximating epithelial sheets and three-dimensional (3D) engineered tissues approximating epithelial tubes. Additionally, engineered tissues are advantageous due to the fact that they can be multiplexed into microarray formats thereby enabling quantitative analysis and high-throughput assays. Several techniques have been developed to pattern microarrays of ECM proteins in 2D onto rigid substrate. Early experiments used photolithographic methods to deposit adhesive islands of defined size onto nonadhesive substrate in order to study the effects of anchorage on cellular behavior (2). In recent years, soft lithographic techniques, which use elastomeric stamps to either print proteins using contact or adsorb proteins using microfluidics, have become popular (3–5). Soft lithography approaches have enabled the formation of complex patterns and gradients of ECM proteins on 2D substrate (6, 7). Similarly, elastomeric membranes containing holes can be used to mask regions of substrate thus enabling stenciling of proteins onto surfaces (8). More recently, microarrays of ECM proteins have been printed using both a standard DNA spotter and the atomic force microscopy technique of dip-pen nanolithography (9, 10). Sacrificial layers, such as aluminum thin films, have been used to pattern combinations of proteins and bioactive molecules onto silica (11). Of the techniques outlined, soft lithography offers the benefits of low cost and ease of use. Recently, much effort has been directed toward tailoring the biochemical and mechanical properties of 3D ECMs to more closely mimic the natural cellular microenvironment. Photopo lymerization has been used to form hydrogels through the activation of light-sensitive photoinitiator molecules to encapsulate cells and to create scaffolding materials (12). Through the use of multilayer photopatterning platforms, increased complexity can be built into engineered materials (13). Applications of 2D soft lithography have been extended to 3D by using elastomeric stamps as molds for macromolecular gels and hydrogel systems (14–19). Additionally, soft lithographic techniques have been used for layer-by-layer deposition of biopolymers, which exploits the use of alternating layers of cell-adhesive and cell-repellant polysaccharides and proteins, to pattern cellular co-cultures (20). Another strategy for patterning 3D ECMs relies on the use of a combination of microfluidics and sacrificial materials, such as
Tissue Microarrays
109
paraffin, matrigel, and gelatin, to create internal cavities, channels, and networks with gels (21, 22). Although progress has been made in patterning 3D ECMs, much work is still required in order to replicate the complex properties and architectures of ECMs found in vivo. Here we describe soft lithography-based techniques to pattern 2D and 3D epithelial tissue microarrays using microcontact printing and micromolding approaches, respectively. In both the 2D and 3D patterning methods the geometry of the ECM is controlled, thus dictating the geometry and form of the tissue and the individual and collective behaviors of the cells that make up the tissue. Micropatterned tissues can be treated with biological molecules of interest, and the behaviors of the cells within the tissues tracked statistically by analyzing the spatial distributions of specific cellular markers within the tissue constructs. Here, we outline the procedures for (1) casting elastomeric stamps from patterned templates; (2) microcontact printing islands of ECM proteins onto slides to create 2D tissue microarrays; (3) micromolding collagen gels to create 3D tissue microarrays; and (4) immunofluorescence analysis of micropatterned tissues.
2. Materials 2.1. Stamp Preparation
1. Patterned silicon wafer.
2.2. Two-Dimensional Tissue Microarrays
1. Patterned PDMS stamp.
2. Poly(dimethyl siloxane) (PDMS; Sylgard 184, Dow Corning).
2. 22-mm glass coverslips (Fisher Scientific). 3. Spin-coater. 4. Extracellular matrix protein (fibronectin, BD Biosciences). 5. Phosphate-buffered saline (PBS). 6. Pluronic F108 Pastille, 1% (w/v) solution in PBS (BASF Corporation).
2.3. ThreeDimensional Tissue Microarrays
1. Patterned PDMS stamp. 2. 35-mm tissue culture dish. 3. Ethanol. 4. Bovine serum albumin (BSA, Calbiochem), 1% (w/v) solution in PBS. 5. 10× Dulbecco’s Modified Eagle’s Medium (DMEM/F12, Sigma).
110
Gomez and Nelson
6. 1:1 Dulbecco’s Modified Eagle’s Medium:Ham’s F12 Nutrient Mixture (DMEM/F12 (1:1), Hyclone), supplemented with 2% fetal bovine serum (Atlanta Biologicals), 50 mg/mL gentamicin (Invitrogen), and 5 mg/mL insulin (Sigma). 7. 0.1 N NaOH. 8. Collagen (bovine dermal or rat tail, BD Biosciences). 9. Glass coverslips, 15-mm diameter. 10. Ice. 2.4. Immunoflu orescence Staining and Image Analysis
1. PBS. 2. Paraformaldehyde, 4% (w/v) solution in PBS (Electron Microscopy Sciences). 3. IgePal CA 630, 0.5% (v/v) solution in PBS (Sigma). 4. Block buffer and antibody dilution buffer: 10% (v/v) goat serum/0.1% (v/v) Triton X-100/PBS. 5. Primary antibody: Phospho-p44/42 MAP Kinase (Thr202/ Tyr204) Antibody (Cell Signaling Technology). 6. Secondary antibody: Alexa 594 goat anti-rabbit (Invitrogen). 7. Nuclear stain: Hoechst 33258 (Invitrogen). 8. Glass coverslips (Fisher). 9. Mounting medium: Fluormount G (Southern Biotech). 10. Photoshop, ImageJ, or another image analysis program.
3. Methods Here, we describe soft lithography-based methods for creating 2D and 3D tissue microarrays (Fig. 1). First, PDMS stamps with defined patterns are prepared from a silicon wafer master. The patterned PDMS stamp is then used for microcontact printing or micromolding of ECM proteins for 2D and 3D ECM microarrays, respectively. Cells are then seeded on the ECM microarrays to form engineered tissues. The tissues can then be fixed, stained, and analyzed for spatial distributions of specific cellular markers. The methods described outline the following: 1. Preparation of the stamp. 2. Fabrication of 2D tissue microarrays. 3. Fabrication of 3D tissue microarrays. 4. Staining and imaging the tissues. 5. Analysis of the spatial distribution of cellular behaviors.
Tissue Microarrays
111
Fig. 1. Schematic of patterning two- and three-dimensional tissues.
3.1. Stamp Preparation
1. Prepare 60 g of 10:1 (w/w) PDMS polymer:curing agent solution. Mix thoroughly and place the mixture in a vacuum dessicator to remove air bubbles. 2. Pour degassed PDMS mixture onto patterned silicon wafer master (see Note 1). 3. Bake at 60°C for 2 h to cure the PDMS. 4. Carefully peel the PDMS from the surface of the master. 5. Cut PDMS patterned by master into stamps of the desired size.
3.2. Two-Dimensional Tissue Microarrays
1. Spin coat a thin layer of PDMS onto the surfaces of glass coverslips. 2. Bake at 60°C for 2 h to cure PDMS. 3. Treat PDMS-coated coverslips for 7 min in UV/ozone cleaner before use in 2D microarray patterning. This oxidizing treatment increases the wettability of the PDMS substration thus allowing for both microcontact printing of protein and adsorption of Pluronics F108 (see steps below) (23). 4. Sterilize PDMS stamps with ethanol. Dry the stamps thoroughly with a vacuum aspirator.
112
Gomez and Nelson
5. Coat the PDMS stamps with a solution of 25 mg/mL of fibronectin in PBS. Incubate for 2 h at room temperature. 6. Rinse once with PBS. 7. Dry the stamps with a steady stream of compressed nitrogen. 8. Stamp fibronectin onto the surface of a UV/ozone-treated PDMS-coated glass coverslip. Press lightly and then lift directly upward to remove stamp. 9. Flood the dish with a solution of 1% (w/v) Pluronics F108 in PBS. Incubate for 15 min. 10. Rinse twice with PBS. Leave patterned coverslips in PBS until ready to plate cells. 11. Plate cells on fibronectin patterned coverslips in cell culture media. Place in incubator and allow cells to adhere to fibronectin islands. Rinse to remove excess cells that have not adhered (see Note 2). 3.3. ThreeDimensional Tissue Microarrays
1. Cut PDMS into stamps that are ~5 mm cubes. Cut two small rectangles per sample from a thin sheet of polymerized PDMS to use as supports. Place stamps into Petri dish feature-side up. 2. Sterilize the PDMS stamps, PDMS supports, and 15-mm diameter coverslips with ethanol. Dry thoroughly with a vacuum aspirator. 3. Coat the feature side of the PDMS stamps with a solution of 1% BSA in PBS. Using a pipette tip, gently scrape the surface of the PDMS to remove air bubbles from the PDMS surface. Incubate for at least 30 min at room temperature (see Note 3). 4. Prepare a neutralized solution of collagen by mixing stock collagen with 0.1 N NaOH and 10× DMEM on ice. Mix thoroughly without introducing air bubbles (see Note 4). Adjust to the desired collagen concentration by adding 1× DMEM/F12. 5. Aspirate the BSA from PDMS stamps with a vacuum pipette. Rinse the BSA-coated surface twice with neutralized collagen (~30 mL). 6. Pipette a drop of neutralized collagen to the top of the PDMS stamp (~30 mL). 7. Flip over the collagen-coated stamp and place on top of two supports. 8. Pipette ~30 mL of collagen to the center of 15-mm round coverslips to make lids. 9. Place the dishes and lids in a 37°C incubator for 30 min. 10. Prepare a concentrated suspension of cells and keep them on ice. 11. Remove PDMS stamp from collagen gels by lifting directly upward with sterilized tweezers.
Tissue Microarrays
113
12. Add ~30 mL of resuspended cells to the collagen gel. Monitor the sample with a phase-contrast microscope. When cells have settled into the wells, wash the sample by holding the dish at a 45° angle and gently pipetting 400 mL of cold media across the surface to remove excess cells. Repeat up to three times. Place the sample in a 37°C incubator for 5 min to allow the cells to adhere to the collagen. 13. Remove samples from the incubator and gently place a collagen lid on each sample. Add 2.5 mL of culture media to each sample and return the sample to the incubator. Observe the sample after ~24 h for tubule formation. 3.4. Immunoflu orescence Staining and Analysis
After seeding cells onto the 2D and 3D micropatterned ECM arrays, cellular and tissue properties, such as projected cell area and tissue form, can be observed by phase-contrast microscopy (see Fig. 2a, d). Likewise, samples can be fixed and stained for markers of interest. Here, as an example we describe the spatial distribution of phosphorylated extracellular signal-regulated kinase (ERK1 and ERK2) in 2D and 3D epithelial tissues. After treatment with epidermal growth factor (EGF), ERK1 and ERK2 are phosphorylated and are then translocated to the nucleus where they promote transcription of target genes of the mitogen-activated protein kinase (MAPK) signaling pathway.
Fig. 2. Two- and three-dimensional mammary epithelial tissue microarrays. (a) Phase-contrast image of 2D mammary epithelial tissue microarray. (b) Gray-scale fluorescence microscopy image of 2D tissue stained for phosphorylated ERK1/2. (c) Color-coded frequency map of nuclear localized phosphorylated ERK1/2 in 2D tissue. (d) Phase-contrast image of 3D mammary epithelial tissue microarray. (e) Gray-scale fluorescence microscopy image of 3D mammary tissue stained for phosphorylated ERK1/2. (f) Frequency map of total phosphorylated ERK1/2 in 3D mammary epithelial tubule. Scalebars: (a, d) 100 mm; (b, c, e, f ) 25 mm.
114
Gomez and Nelson
1. Treat sample with 25 ng/mL EGF and place in an incubator for 15 min. 2. Remove sample from the incubator and aspirate media. Rinse the sample once with PBS. Aspirate the PBS and replace with fixative solution (4% paraformaldehyde in PBS). Incubate at room temperature for 15 min. Wash fixed samples three times with PBS. 3. Incubate in 0.5% IgePal CA 630 in PBS twice for 10 min each time. 4. Incubate in 0.1% TritonX-100 in PBS for 15 min at room temperature. 5. Block sample with blocking buffer for 2 h at room temperature. Rinse sample once with PBS. 6. Apply diluted primary antibody (1:500) and incubate overnight at 4°C. For 2D samples, rinse sample with PBS three times for 5 min. For 3D, rinse with PBS for 5 h at room temperature. 7. Apply diluted secondary antibody (1:1,000). Incubate for 1–2 h at room temperature in the dark or overnight at 4°C for 2D and 3D tissues, respectively. For 2D samples, rinse sample with PBS three times for 5 min. For 3D, rinse with PBS for 5 h at room temperature. 8. Apply diluted nuclear stain (1:10,000) and incubate for 20 min at room temperature. Rinse sample with PBS. Mount samples on cover slides. 9. Observe and image using a fluorescence microscope. Take 50 images of tissues that have been aligned using the eyepiece or a stage micrometer on a fluorescence microscope. 10. Using an image analysis software convert gray-scale images into black-and-white images using binarize function. 11. Add the black-and-white images together. 12. Convert the gray-scale image into a color-coded frequency map using the Indexed Color mode in Photoshop (see Fig. 2c, f).
4. Notes 1. Silicon masters can be silanized to aid in removal of the PDMS. Place the master in a vacuum dessicator with a glass slide containing a drop of (tridecafluoro-1,1,2,2,tetrahydrooctyl)-1-tricholorosilane (Sigma-Aldrich). Evacuate the chamber. After 1–2 min isolate the chamber and allow for vapor silanization reaction to proceed for 2 h. Alternatively, a PDMS master can be made from the silicon master. First create a PDMS master with posts from the silicon wafer.
Tissue Microarrays
115
Silanize the PDMS master as described above. Use PDMS master with posts to make a PDMS master with holes. 2. Observe samples every half hour to determine when cells begin to adhere to ECM islands. The concentration of plated cells and the plating time can be modified to best achieve desired number of cells per island. 3. Stamps can be incubated overnight with BSA at 4°C. 4. The neutralized collagen solution can be quickly spun down in a centrifuge to remove air bubbles, if needed.
Acknowledgments This work was supported by grants from the NIH (CA128660 and GM083997), Susan G. Komen for the Cure (FAS0703855); and the David & Lucile Packard Foundation. C.M.N. holds a Career Award at the Scientific Interface from the Burroughs Wellcome Fund. E.W.G. was supported by post doctoral fellowships from the New Jersey Commission on Cancer Research and Susan G. Komen for the cure. References 1. Langer, R., and Vacanti, J. P. (1993) Tissue engineering. Science 260, 920–926. 2. O’Neill, C., Jordan, P., and Ireland, G. (1986) Evidence for two distinct mechanisms of anchorage stimulation in freshly explanted and 3t3 swiss mouse fibroblasts. Cell 44, 489–496. 3. Chen, C. S., Mrksich, M., Huang, S., Whitesides, G. M., and Ingber, D. E. (1997) Geometric control of cell life and death. Science 276, 1425–1428. 4. Whitesides, G. M., Ostuni, E., Takayama, S., Jiang, X., and Ingber, D. E. (2001) Soft lithography in biology and biochemistry. Annu. Rev. Biomed. Eng. 3, 335–373. 5. Nelson, C. M., Jean, R. P., Tan, J. L., Liu, W. F., Sniadecki, N. J., Spector, A. A., and Chen, C. S. (2005) Emergent patterns of growth controlled by multicellular form and mechanics. Proc. Natl. Acad. Sci. USA 102, 11594–11599. 6. Tien, J., Nelson, C. M., and Chen, C. S. (2002) Fabrication of aligned microstructures with a single elastomeric stamp. Proc. Natl. Acad. Sci. USA 99, 1758–1762. 7. Jeon, N. L., Dertinger, S. K. W., Chiu, D. T., Choi, I. S., Stroock, A. D., and Whitesides, G. M.
8.
9.
10.
11. 12.
13.
(2000) Generation of solution and surface gradients using microfluidic systems. Langmuir 16, 8311–8316. Ostuni, E., Kane, R., Chen, C. S., Ingber, D. E., and Whitesides, G. M. (2000) Patterning mammalian cells using elastomeric membranes. Langmuir 16, 7811–7819. Flaim, C. J., Chien, S., and Bhatia, S. N. (2005) An extracellular matrix microarray for probing cellular differentiation. Nat. Methods 2, 119–125. Wilson, D. L., Martin, R., Hong, S., CroninGolomb, M., Mirkin, C. A., and Kaplan, D. L. (2001) Surface organization and nanopatterning of collagen by dip-pen nanolithography. Proc. Natl. Acad. Sci. USA 98, 13660–13664. Jackson, B. L., and Groves, J. T. (2007) Hybrid protein-lipid patterns from aluminum templates. Langmuir 23, 2052–2057. Nguyen, K. T., and West, J. L. (2002) Photopolymerizable hydrogels for tissue engineering applications. Biomaterials 23, 4307–4314. Tsang, V. L., Chen, A. A., Cho, L. M., Jadin, K. D., Sah, R. L., DeLong, S., West, J. L., and
116
14.
15.
16.
17.
18.
Gomez and Nelson Bhatia, S. N. (2007) Fabrication of 3d hepatic tissues by additive photopatterning of cellular hydrogels. Faseb J. 21, 790–801. Tang, M. D., Golden, A. P., and Tien, J. (2003) Molding of three-dimensional microstructures of gels. J. Am. Chem. Soc. 125, 12988–12989. Tang, M. D., Golden, A. P., and Tien, J. (2004) Fabrication of collagen gels that contain patterned, micrometer-scale cavities. Adv. Mater. 16, 1345–1348. Nelson, C. M., Vanduijn, M. M., Inman, J. L., Fletcher, D. A., and Bissell, M. J. (2006) Tissue geometry determines sites of mammary branching morphogenesis in organotypic cultures. Science 314, 298–300. Fukuda, J., Khademhosseini, A., Yeo, Y., Yang, X. Y., Yeh, J., Eng, G., Blumling, J., Wang, C. F., Kohane, D. S., and Langer, R. (2006) Micromolding of photocrosslinkable chitosan hydrogel for spheroid microarray and cocultures. Biomaterials 27, 5259–5267. Nelson, C. M., Inman, J. L., and Bissell, M. J. (2008) Three-dimensional lithographically defined organotypic tissue arrays for quantitative analysis of morphogenesis and neoplastic progression. Nat. Protoc. 3, 674–678.
19. Jongpaiboonkit, L., King, W. J., Lyons, G. E., Paguirigan, A. L., Warrick, J. W., Beebe, D. J., and Murphy, W. L. (2008) An adaptable hydrogel array format for 3-dimensional cell culture and analysis. Biomaterials 29, 3346–3356. 20. Fukuda, J., Khademhosseini, A., Yeh, J., Eng, G., Cheng, J. J., Farokhzad, O. C., and Langer, R. (2006) Micropatterned cell co-cultures using layer-by-layer deposition of extracellular matrix components. Biomaterials 27, 1479–1486. 21. Bettinger, C. J., Weinberg, E. J., Kulig, K. M., Vacanti, J. P., Wang, Y. D., Borenstein, J. T., and Langer, R. (2006) Three-dimensional microfluidic tissue-engineering scaffolds using a flexible biodegradable polymer. Adv. Mater. 18, 165–169. 22. Golden, A. P., and Tien, J. (2007) Fabrication of microfluidic hydrogels using molded gelatin as a sacrificial element. Lab Chip 7, 720–725. 23. Tan, J. L., Liu, W., Nelson, C. M., Raghavan, S., and Chen, C. S. (2004) Simple approach to micropattern cells on common culture substrates by tuning substrate wettability. Tissue Eng. 10, 865–872.
Chapter 6 Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome Ku-Lung Hsu, Kanoelani Pilobello, Lakshmipriya Krishnamoorthy, and Lara K. Mahal Abstract The mammalian cell surface is rich with carbohydrate polymers involved in a diversity of biological recognition events. Dynamic alterations of surface glycans mediate cell–cell communication in the immune system and host specificity of bacterial and viral pathogens. In addition, altered surface glycosylation has been implicated in disease progression of many cancers and may serve as important new targets for therapeutics. Despite the importance of glycosylation, the systematic analysis of sugars, i.e., glycomics, has lagged behind the well-studied disciplines of genomics and proteomics. This deficiency is due in part to the unique analytical challenges presented by glycans and the overwhelming diversity of sugars in nature. New microarray technologies have provided a high-throughput methods with which to probe the glycome. Our laboratory has pioneered a shown ratiometric two-color lectin microarray method that rapidly evaluates differences in the glycosylation of mammalian cells. Herein, we present a detailed protocol of our lectin microarray methodology for the differential analysis of mammalian glycomes. Key words: Lectin, Microarray, Glycomics, Carbohydrate, Glycan, Glycosylation, Cancer, Pathogen, Differentiation
1. Introduction The cell surface is a densely packed assortment of glycosylated proteins and lipids that function in a myriad of biological processes. The carbohydrate motifs embedded within these complex polymers encode information that mediate many biological events including cell–cell interactions, differentiation, and the host tropism of pathogens (1). Unlike proteins and nucleic acids, carbohydrates exist as both linear and branched polymers that can differ in the linkages between monomers and in the anomeric stereochemistry, in addition to varying in monomer Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_6, © Springer Science+Business Media, LLC 2011
117
Hsu et al.
composition, resulting in enormous structural diversity. These issues pose a challenge to the high-throughput profiling of glycosylation, a critical element in the systems-level study of glycosylation, i.e., glycomics (2, 3). In response, we have developed lectin microarray technology that utilizes the innate ability of natural carbohydrate-binding proteins to recognize complex glycan motifs, as a rapid method for glycan analysis and applied it to the glycomic characterization of glycoproteins, bacteria, and mammalian cells (4–8). Lectin microarrays consist of a series of carbohydrate-binding proteins (lectins) immobilized onto a glass slide as high density spots in a defined layout. Hybridization of fluorescently labeled samples to the microarray gives a visual binding pattern that provides structural information about the glycosylation status of samples. We have recently extended this technology to the characterization of mammalian cell surface glycans using a sensitive ratiometric method (Fig. 1) (8). Hybridization of fluorescently labeled cell membrane-derived micellae against an orthogonally labeled biological reference sample allows the comparison of samples across multiple slides and examination of subtle differences in glycosylation between samples. This ratiometric lectin microarray approach has the advantage of increased resolution and improved quality control for the differential analysis of mammalian glycomes (8).
t
118
Co
A
4
Fig. 1. Schematic of a ratiometric lectin microarray experiment (from (8)). Orthogonally labeled cellular samples are hybridized to a subarray on a slide. The comparative lectin binding pattern of the sample allows us to elucidate the relative level of glycan epitopes present in the cellular membrane.
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome
119
2. Materials 2.1. Manufacture of Lectin Microarray
1. Commercially available lectins (E.Y. Labs, San Mateo, CA and Vector Labs, Burlingame, CA). 2. Nexterion Slide H (Schott, North America, Elmsford, NY). Slides should be stored at −20°C in a dessicator. 3. Spotbot Arrayit Microarrayer with SMP3 pins (Telechem International, Sunnyvale, CA). 4. 384-well plates. 5. Standard centrifuge with adapters to spin 384-well plates (same outer dimensions as a standard 96-well plate).
2.2. Sample Preparation
1. 0.5 M Ethylenediaminetetraacetic acid (EDTA, in ddI H2O, pH 8). 2. PBS (0.1 M NaH2PO4, 0.15 M NaCl, pH 7.2–7.4). 3. Cell sonicator with 1/8″ microtip horn (Branson Inc, Danbury, CT). 4. Ultracentrifuge (either tabletop or floor model). 5. Cy dye buffer (0.1 M NaCO3 in H2O, pH = 9.3). A stock buffer may be made but the pH will change over time and will need adjustment. 6. 25-Gauge needle, 1-mL plastic syringe. 7. DC protein assay (Bio-Rad, Hercules, CA).
2.3. Dye Conjugation
1. NHS-Cy3 or -Cy5 dye (GE Healthcare Life Sciences, Piscataway, NJ). This reagent is light sensitive and should be stored at 4°C. 2. Slide-A-Lyzer cassette (Pierce, Rockford, IL).
2.4. Hybridization and Scanning
1. Blocking reagent (50 mM ethanolamine in 50 mM sodium borate, pH 8.0). 2. PBST (PBS, 0.005% Tween-20). 3. Ovalbumin, porcine mucin (Sigma, St. Louis, MO). These should be stored at 4°C. 4. Coplin jars. 5. Hybridization cassettes (1 × 24 configuration, Telechem International). 6. Slide spinner (Labnet International, Edison, NJ) or a standard floor centrifuge with slide rack adaptors may be used. 7. GenePix 4000B scanner (Molecular Devices, Sunnyvale, CA).
120
Hsu et al.
3. Methods Our current lectin microarray consists of commercially available lectins that are mainly plant derived. We have also now included some recombinant lectins of bacterial origin (4). All lectins are printed at concentrations optimized to give a minimum signal of 1,000 arbitrary fluorescence units (A.U.) under fixed scanning conditions with glycoprotein standards. Each slide provides 24 total subarrays of which at least two need to be used for quality control hybridizations with glycoprotein standards. This adds an important level of quality control that can help identify inactive and misprinted lectins. In addition, all lectins are printed in multiple replicates ensuring more precise measurement of lectin activity. 3.1. Lectin Microarray Print Procedure
1. Remove Nexterion H slides from −20°C freezer, and let thaw to room temperature before use. 2. Dilute lectins to the recommended concentration in the appropriate print buffer (see Table 1 and Note 1). Load 10 ml of each lectin mixture into a 384-well plate in the desired order and centrifuge the plate at 50 × g for 5 min prior to use. 3. To ensure consistent print quality, the SMP3 pins should be cleaned before each use. Place pins in a floatable pin rack, and sonicate for 5 min in an ultrasonic water bath filled with 5% Micro Cleaning Solution (Telechem International, Inc.). Sonicate the pins for an additional 5 min in distilled water. Remove and dry pins (see Note 2). 4. The printhead on the Spotbot Microarrayer should also be cleaned before each use. Prepare 50 mL of hot (65°C) printhead cleaning solution (Telechem International, Inc.). Scrub extensively the outer surface and pin-holes of the printhead using a brush. Rinse the printhead extensively with distilled water and dry with forced air (see Note 3). 5. Insert the pin carefully into the printhead, and test the movement of the pin in the printhead by pushing the pin up and down several (5–10) times (see Note 4). 6. Turn on the Spotbot and all accessory devices. Open the Multiple Microarray Format SpoCLe Generator and select the desired print configuration (see Note 5). 7. Insert the preprint slide(s), the Nexterion H slide(s), and the 384-well plate loaded with lectin mix (see Note 6). 8. Check the humidity of the print chamber. Ideally, the humidity should be kept around 50–60% during the entire print process.
500 1,000 1,000 500 500 500
500 500 500 500 1,000 500 500 500 500
ABA AAL AAA PNA AIA BPA Blackbean BDA Con A CCA CAA CPA CA CSA CVN DSA
Agaricus bisporus
Aleuria aurantia
Anguilla anguilla
Arachis hyogaea
Artocarpus intergrifolia (Jacalin)
Bauhinia purpurea
Black bean crude
Bryonia dioica
Canavalia ensiformis
Cancer antennarius
Caragana arborescens
Cicer arietinum
Colchicum autumnale
Cystisus scoparius
Cyanovirin
Datura stramonium
1,000
500
APA
Abrus precatorius
[Print] mg/mL
Abbreviation
Lectin
Table 1 Panel of lectins in microarray
Lactose
Mannose
Galactose
Galactose
Lactose
Galactose
Lactose
Mannose
Galactose
Lactose
Galactose
Galactose
Galactose
Fucose
Fucose
Galactose
Galactose
Carbohydrates
(continued)
GlcNAcb-1,4GlcNAc oligomers and LacNAc
a-1,2 Mannose
b-GalNAc, terminal
Terminal Gal b-OR
Complex
GalNAc/Gal (monosaccharides best)
9-O-Acetyl Sia and 4-O-Acetyl Sia
Branched and terminal mannose, terminal GlcNAc
GalNAc
GalNAc
Primarily Gal b-1,3 or 1,4 but will also bind b-GalNAc more weakly
a-GalNAc
Terminal Gal b-OR
a-Fuc
Fuc
Gal b-1-3GalNAc
Gal b-1,3GalNAc (TF antigen) > Gal
Rough specificity a
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome 121
500 1,000 1,000 500 2,500 500 1,000 500 500 500 1,000 500 500 500 1,000 500 500 500 1,000 1,000
DBA ECA EEA GNA Gal-1 SBA GRFT GS-I GS-II HPA HHL HMA IAA LAA LcH LFA LPA LTL LEA MAA-I
Dolichos biflorus
Erythrina cristagalli
Euonymus eurpaeus
Galanthus nivalis
Galectin-1
Glycine max
Griffithsin
Griffonia simplicifolia I
Griffonia simplicifolia II
Helix pomatia
Hippeastrum hybrid
Homaris americanus
Iberis amara
Laburnum alpinum
Lens culinaris
Limax flavus
Limulus polphemus
Lotus tetragonolobus
Lypersicon esculentum
Maackia amurensis
[Print] mg/mL
Abbreviation
Lectin
Table 1 (continued)
Lactose
GlcNAc
Fucose
Lactose
Lactose
Mannose
GlcNAc
Lactose
Lactose
Mannose
Galactose
GlcNAc
Galactose
GlcNAc
Galactose
Lactose
Mannose
Lactose
Galactose
Galactose
Carbohydrates
LacNAc
b-1,4GlcNAc oligomers
Terminal a-Fuc, Lex
a-Sia
a-Sia
Complex (man/GlcNAc core with a-1,6 Fuc)
GlcNAc oligomers
GalNAc
Sialic acid
a-1,3 Mannose and a-1,6 mannose
a-GalNAc terminal
Terminal GlcNAc
a-Galactose
Mannose, GlcNAc
Terminal GalNAc
LacNAc
Terminal a-1,3 mannose
Blood Groups B and H
LacNAc and GalNAc
GalNAca-OR
Rough specificity a
122 Hsu et al.
1,000 500 1,000 500 500
500 500 1,000 1,000 1,000 1,000 500 500 500 1,000 500 500
NPA PAA LBA PHA-E PHA-L PEA, PSA PSL PTA galactose PTL-I PTL-II RCA RPA SNA SVN STA SJA TKA RTA
Narcissus pseudonarcissus
Persea americana
Phaseolus lunatus
Phaseolus vulgaris-L
Phaseolus vulgaris-L
Pisum sativum
Polyporus squamosus
Psophocarpus tetragonolobus
Psophocarpus tetragonolobus
Psophocarpus tetragonolobus
Ricin B chain
Robinia pseudoacacia
Sambucus nigra
Scytovirin
Solanus tuberosum
Sophora japonica
Trichosanthes kirilowii
Trifolium repens
1,000
500
MAA
Maackia amurensis
1,000
MAA-II
Maackia amurensis
GlcNAc
Galactose
Galactose
GlcNAc
Mannose
Lactose
Lactose
Lactose
Galactose
Galactose
Galactose
Lactose
Mannose
Galactose
Lactose
Galactose
GlcNAc
Mannose
Lactose
Lactose
2-Deoxy-Glu (continued)
b-Gal, LacNAc but Sia-a-2,3 or 2,6 inhibits best
GalNAc
GlcNAc oligomers, LacNAc
a-1,2 Mannose
a-2,6 Sialic acid on LacNAc
Complex
b-Gal/GalNAc
a-1,2 Fucosylated LacNAc
a-GalNAc
Gal
a-2,6 Sialic acid
Man
b-1,6 Branched trimannosyl core N-linked glycans
Complex
GalNAca-1,3[Fuca-1,2]Gal
Unknown
Terminal and internal Man
a-2,3 Sialic acid
a-2,3 Sialic acid
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome 123
1,000 1,000 500 500 500 500 500 500 500 500
WGA TL UEA UEA-II UDA VFA VGA VVA VVA (man) WFA
Tritiicum vulgare
Tulipa sp.
Ulex europaaeus I
Ulex europaaeus II
Uritica dioica
Vicia fava
Vicia graminea
Vicia villosa
Vicia villosa
Wisteria floribunda
Galactose
Galactose
Galactose
Galactose
Galactose
GlcNAc
GlcNAc
Fucose
GlcNAc
GlcNAc
Carbohydrates
GalNAc
Man
GalNAc
O-linked Gal b-1,3GalNAc clusters
Man > Glc > GlcNAc
GlcNAc b-1,4GlcNAc oligomers and high mannose epitopes
GlcNAc oligomers
a-Fucose
GlcNAc
b-GlcNAc, sialic acid, GalNAc
Rough specificity a
a Specificity shown is very rough and was obtained from a variety of sources including the Consortium for Functional Glycomics Carbohydrate Array Analysis and the Handbook of Plant Lectins (1998, Wiley and Sons)
[Print] mg/mL
Abbreviation
Lectin
Table 1 (continued)
124 Hsu et al.
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome
125
A humidity control device may be used to maintain the appropriate relative humidity (see Note 7). 9. Turn on the cooling plate and set the temperature to ~8°C. This is important to help maintain lectin activity during extended and overnight prints. 10. Once the temperature and humidity have reached the desired parameters, start the printing process. Periodically monitor the printing process to make sure that everything is functioning properly. 11. After the printing process has finished, turn off the cooling plate and allow the slides to remain in the chamber for 1 h at ~50% relative humidity. This helps to ensure maximum coupling efficiency. 12. Once printed, slides can be stored at −20°C in moisture barrier bags. In our hands, these arrays are stable for up to 2 weeks, without any noticeable decreases in lectin activity. 3.2. Preparation of Micellae
The isolation of membrane glycoproteins and glycolipids for ratiometric lectin microarray analysis requires harvesting of cells without the use of proteases or detergents. Trypsin, a common protease used in cell cultures, has been shown to preferentially bind N-linked glycoproteins, potentially biasing the glycan pools. Detergents solubilize glycolipids leading to samples that do not accurately represent the glycome. Thus, our protocol avoids the use of both reagents, forming micellae from the physical disruption of membranes. Previous work has shown no significant differences in the lectin binding patterns from cells labeled prior to or after lysis, validating that our micellae accurately represent cellsurface glycans (8). In addition, the presence of glycolipids in our micellae samples was verified by TLC, followed by resorcinol staining for sialic acids that are a major component of gangliosides. 1. Adherent cells should be harvested without trypsin, which cleaves cell surface glycans in a biased manner. Cells are incubated with 0.5 M EDTA for 15 min at room temperature and harvested with a cell scraper. 2. Suspension or adherent cells are centrifuged at 200 × g for 10 min at 4°C. A pellet should form at the bottom. The samples should be kept at 4°C (i.e., on ice) from this point onward. 3. Cells are resuspended in cold PBS at a concentration of 3–10 million cells per 3 mL of PBS. 4. Using a cell sonicator with a 1/8 in. microtip, sonicate the cells in 5 s pulses with 10 s intervals at 70% power. Sonication homogenizes the cell membranes creating micellae. While this may include organellar membranes, we have validated that the glycomic composition of these samples match that of the surface via lectin histology (Fig. 2, see Note 8).
126
Hsu et al.
Fig. 2. Comparison of Yang Correlation data from cellular micellae to fluorescence microscopy. (a) Comparative Yang Correlation data derived from a dye-swapped pair of lectin microarrays are shown for ConA, HPA, and WGA binding to the Chinese Hamster Ovary cells Pro-CHO 5 (Pro-5) and Lec8 (adapted from (8)). Green indicates stronger binding to Pro-5, while red indicates stronger binding to Lec8. (b) Fluorescence microscopy confirms that data obtained by microarray methods are accurate to the cell surface (8).
5. Micellae are pelleted at 100,000 × g for 60 min in an ultracentrifuge. The supernatant should be removed immediately after centrifugation to maintain the integrity of the pellet. 6. The pellet is resuspended in 250–500 mL of Cy dye buffer and homogenized by passing it through a 25-gauge needle as many times as necessary, typically 10× (see Note 9). 7. After homogenization, the protein concentration of the samples is determined using the DC protein assay.
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome
3.3. Sample Labeling
127
1. It is important to make sure that sample labeling is consistent. For every milligram of protein, we use 60 mg of the appropriate NHS-Cy dye. 2. Incubate the samples with Cy dye for 45 min at room temperature in the dark (i.e., covered in foil). 3. After 45 min, the dyed samples are dialyzed. The most effective dialysis method is to use the Slide-A-Lyzer cassettes. However, for multiple samples, this can be highly time consuming. An alternative is to use the Pierce mini-dialysis units. These are more effective for sample volumes under 100 ml (see Note 10). 4. Float the dialysis units in 2 L of PBS at 4°C and stir on low in the dark to prevent loss of fluorescence. 5. The dialysis buffer should be changed once after 30 min, and the samples should then dialyze overnight at 4°C in the dark. 6. After dialysis, sample concentrations can be redetermined by the DC assay to check for protein loss. Labeled samples should be immediately aliquotted and snap frozen in liquid nitrogen for storage in the dark at −20°C.
3.4. Hybridization of Samples and Scanning
1. The printed slide is allowed to come to room temperature for approximately 10–15 min. 2. The slide is gently immersed in blocking solution for 1 min, with the printed side facing down. 3. The slide is then slowly immersed in a coplin jar filled with blocking solution. The ethanolamine reacts with the free, unreacted sites on the slides, preventing them from binding nonspecifically to proteins. 4. The slide is blocked for 1 h at room temperature on the benchtop. 5. After 1 h, the blocking solution in the coplin jar is discarded and replaced with a solution of PBST. The slide is rinsed gently with PBST by carefully swirling the coplin jar. The slides are gently rinsed 3× with PBST, followed by a final rinse with PBS. 6. The slide is then dried using a slide spinner for 30 s. It is important to ensure that the slide is dry prior to use. 7. The slide is fitted tightly within the hybridization cassette. 8. The samples are diluted in PBST to obtain the desired sample concentrations (see Note 11) and are carefully added to each subarray at one corner in such a way that liquids are not pipetted directly over the printed array (see Note 12). For single-color experiments, a unique Cy3- or Cy5- labeled sample is added to each subarray. For ratiometric dual-color
128
Hsu et al.
experiments, two differentially labeled samples are added to a single subarray in equal amounts. For example, equal amounts of Cy3- labeled sample A and Cy5- labeled sample B are added to the same subarray (Fig. 1). 9. For dye-swap experiments, two subarrays are used for the analysis of orthogonally labeled pair of analytes. For example, equal amounts of Cy3- labeled sample A and Cy5- labeled sample B are added to one subarray, while equal amounts of Cy3- labeled sample B and Cy5- labeled sample A are added to another subarray. 10. For every array, two subarrays are used to ensure quality control of the lectin microarray. Typically, 10 mg of Cy3- or Cy5labeled glycoproteins with known glycosylation patterns, such as ovalbumin or porcine mucin, are used for this purpose. 11. The samples are hybridized to the array for 2 h with gentle rocking on a shaker. The array is covered with aluminum foil to protect the samples from light (see Note 13). 12. After 2 h, the samples are carefully removed from the array using a pipette (a multichannel pipetter works well for this). 13. The arrays are rinsed by addition of 200 mL of PBST to each array and gently rocking on a shaker for 5 × 3 min. 14. Any residual liquid is removed after the last wash, the slide is carefully removed from the cassette and immersed in a coplin jar with PBS. The slide is rinsed in PBS for 5 min on a shaker. 15. The slide is then dried using a slide spinner and scanned using a GenePix 4000B scanner (see Note 14). 3.5. Single Color and Ratiometric Analysis
Data from the microarray are extracted using a standard fluorescent slide scanner and image analysis software. For single color analysis, signals are considered positive if the signal to noise ratio is greater than 5. For ratiometric analysis, single color analysis of the two samples is used to determine positive signals. Outlier points are determined using the Grubbs outlier test at the 95% confidence interval to account for experimental variations (www.graphpad.com). For ratiometric two-color analysis, Yang’s correlations are calculated to compensate for dye bias, and hierarchical clustering allows comparison between sample sets. The carbohydrate composition of samples is inferred using the carbohydrate specificity profiles of lectins obtained from the Consortium for Functional Glycomics (www.functionalglycomics.org) and literature. 1. The spots on the scanned slide image are aligned and segmented in GenePix Pro 5.1 (or above) with circular alignment and local background subtraction.
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome
129
2. The data are exported to a .txt file and imported into Microsoft Excel as a tab delimited file. 3. We currently use the single color signal to noise (S/N) data as a reference for positive signals. Signal to noise is calculated from the median signal (S, not background subtracted) and the median local background (N). We consider signals with S/N > 5 to be positive (see Note 16). 4. Yang’s correlations are calculated for ratiometric data using two dye-swapped arrays. The median background subtracted fluorescence values are used (e.g., “F635-B median” column in GenePix for Cy5). The background subtracted fluorescence data are tested for outliers using the Grubbs outlier test (www.graphpad.com). The remaining values are transformed to Log base 2 values and averaged. We then perform the following operation: [(Log2 (Cy5 Cell B signal) − Log2 (Cy3 Cell A signal)) − (Log2 (Cy3 Cell B signal) − Log2 (Cy5 Cell A signal))]/2, which is the equivalent of taking the average of the log2 ratios for the dye swapped pair (the Yang’s correlation, 9, see Note 17). 5. We compare samples by hierarchically clustering the results in Cluster (10) using centered Pearson correlation as a distance metric and average linkage analysis. The significance of a node value can be determined by comparing the correlation coefficients to the p-values.
4. Notes 1. Lectins are resuspended in PBS or the supplier’s recommended buffer at concentrations of 1 mg/mL. Some lectins may need to be rocked gently overnight to completely resuspend. Lectins are aliquoted (10–20 ml), snap frozen in a liquid nitrogen bath, and stored for up to 1 year at −80°C. 2. When cleaning the SMP3 pins, do not use forced air on the pin tip as it may result in damage. It is okay to dry the base of the pin using forced air. Also, be careful not to drop the pin as it will need to be replaced. 3. When cleaning the printhead, be sure that it is completely dry before inserting the pin. Any leftover moisture can cause the pin to get stuck during the printing process resulting in misprints. 4. When testing the movement of the pin in the printhead, cover the pin with the supplied sheath to prevent damage to the pin tip. Never move the pin by pushing on the pin tip.
130
Hsu et al.
5. When using the Multiple Microarray Format SpoCLe Generator program, be sure to adjust the offsets so that there is space between lectins and the subarray edges. If the lectins are printed too close to the edges, the SuperMask frame (Telechem International, Inc.) may contact the lectins. 6. Preprint slides can be plain glass slides or old slides that have been cleaned. 7. When using a humidity controller, be sure to allow the sensor to adjust to the humidity in the chamber for ~45 min before beginning to print. 8. During sample sonication, it is important that the samples are either on ice during or immediately following sonication because the samples can overheat. 9. After pelleting the cellular micellae, the homogenization of the sample is much easier to do immediately after centrifugation. We have not seen differences in lectin microarray signals with samples stored over night at 4°C. However, the homogenization is more difficult. 10. During sample dialysis, we have found that making sure the mini-dialysis units are flush with the floater and not pressed too deeply or unevenly results in homogenously dialyzed samples. While this method is easier to work with for large sets, it is less reliable. 11. For cellular micellar samples, we typically use 10 mg of each labeled sample per subarray in 100 mL total volume (diluted with PBS), though lesser amounts (~1–3 mg) have been used successfully in our lab. 12. Care should be taken during the addition of samples or during rinses to avoid pipetting liquids directly on the printed part of the array. Usually, an empty corner of the subarray is used for this purpose. 13. The fluorescent samples should be protected from light during the entire procedure using aluminum foil. 14. It is important to ensure that the slide is dry in all the steps that utilize the slide spinner. 15. For analyzing the specificity of our lectin microarray, inhibition experiments are performed with a small panel of carbohydrates including mannose, galactose, N-acetyl glucosamine, fucose, and lactose. The array is preincubated with the inhibitory sugars for 30 min, followed by the addition of the sample. 16. In data analysis using signal to noise ratios, occasionally, a trend in background affects the S/N. For example, the background may increase as with the rows in each subarray. Care should be taken in the washes by aspirating wash buffer from
Ratiometric Lectin Microarray Analysis of the Mammalian Cell Surface Glycome
131
each side of the well, if possible. Background trends and anomalies in general should be considered when comparing data and trying to validate dual color data. 17. The single color S/N values should be used to help exclude ratiometric data on nonpositive lectins. Because the ratiometric data are a ratio, spurious values are obtained when the lectins are not positive.
References 1. Gabius, H. J., Siebert, H. C., Andre, S., Jimenez-Barbero, J., and Rudiger, H. (2004) Chembiochem 5, 740–64. 2. Mahal, L. K. (2008) Anticancer Agents Med Chem 8, 37–51. 3. Raman, R., Raguram, S., Venkataraman, G., Paulson, J. C., and Sasisekharan, R. (2005) Nat Methods 2, 817–24. 4. Hsu, K. L., Gildersleeve, J. C., and Mahal, L. K. (2008) Mol Biosyst 4, 654–62. 5. Hsu, K. L., and Mahal, L. K. (2006) Nat Protoc 1, 543–9.
6. Hsu, K. L., Pilobello, K. T., and Mahal, L. K. (2006) Nat Chem Biol 2, 153–7. 7. Pilobello, K. T., Krishnamoorthy, L., Slawek, D., and Mahal, L. K. (2005) Chembiochem 6, 985–9. 8. Pilobello, K. T., Slawek, D. E., and Mahal, L. K. (2007) Proc Natl Acad Sci USA 104, 11534–9. 9. Yang, Y. H., Dudoit, S., Luu, P., Lin, D. M., Peng, V., Ngai, J., and Speed, T. P. (2002) Nucleic Acids Res 30, e15. 10. Eisen, M. B., Spellman, P. T., Brown, P. O., and Botstein, D. (1998) Proc Natl Acad Sci USA 95, 14863–8.
Chapter 7 Cell Microarrays Based on Hydrogel Microstructures for the Application to Cell-Based Biosensor Won-Gun Koh Abstract Cell-based biosensors constitute a promising field that has numerous applications ranging from pharmaceutical screening to detection of pathogen and toxicant. The trends toward miniaturization of cell-based biosensor continue to spur development of cell microarray integrated into microfluidic devices. For cellbased biosensors to be useful for larger applications, several technical goals must be realized. First, the cell-patterning method used to generate multi-phenotypic array can accommodate multiple cell lines without major losses of cell viability, maintain total isolation of each cell phenotype, provide for the adequate mass transfer of dissolved gases and nutrients, and easy enough to allow for mass production. Second, cells on microarray must be cultured in three-dimensional environment as they do in real tissue to obtain accurate response of cells against target analyte. Third, physiological status of micropatterned cells must be monitored non-invasively. As one solution to satisfy these requirements, we prepare cell microarrays using microfabricated poly(ethylene glycol)(PEG) hydrogel. Arrays of hydrogel microstructures encapsulating one or more different cell phenotypes can be fabricated using photolithography or photoreaction injection molding, and can be incorporated within microfluidic network. Finally, we demonstrate the potential application of cell-containing hydrogel microarrays for toxin detection by monitoring toxin-induced change of cell viability and intercellular enzymatic reaction. Key words: Cell-based biosensor, Cell microarray, Hydrogel microstructures, Microfluidic devices, Cell encapsulation, Microfabrication
1. Introduction Cell-based biosensors are devices that use living cells as the biorecognition elements to detect a broad range of agents and have been receiving attention recently because of potential applications of high content drug screening and detection of biological warfare agents and pathogens (1). Of particular importance is the application of cell-based biosensor to drug screening, Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_7, © Springer Science+Business Media, LLC 2011
133
134
Koh
because cells represent the ultimate target for pharmaceuticals (2). In these cell-based biosensors, analytes (i.e., drug candidates) act on immobilized cells, and physiological changes such as cell metabolism and viability are detected by electrochemical or optical methods. Although cell-based systems often exhibit a long response time and poor selectivity, they have several advantages over other biosensor based on enzymes or antibodies, including extremely good sensitivity, the ability to detect large number of analytes and the easiness of growing and isolating sensing components (i.e., the cells). The most distinct advantage of cell-based systems over conventional biosensor system is that they can offer functional information, i.e., information about the effects of stimulus on a living system, which includes the effects of stimuli on cell death (toxicity) as well as cell function. Currently, most of the cell-based assays are still being performed out in 96-well plates; however, the move toward higher density plate format and system miniaturization is essential to improve the performance or functionality of cell-based biosensor systems (3). To move toward assay miniaturization, significant efforts have focused on the fabrication of cell microarray using a variety of cell-patterning techniques such as photolithography or soft lithography (4–17). Cell microarray are being coupled with fluorescent or electrochemical technologies and incorporated into microfluidic devices to detect the physiological changes of cells by external environment (1). In most of these applications, anchorage-dependent cells are immobilized on a two-dimensional substrate. However, in a two-dimensional system, non-adherent cells are difficult to immobilize and adherent cells such as fibroblasts and hepatocytes are in an unnatural environment; i.e., in tissue they exist in a three-dimensional hydrogel matrix consisting of proteins and polysaccharides (i.e., the extracellular matrix). As a result, the response of these cells to target molecules may be very different than that of the same cells in their native tissue. One strategy to overcome the problems associated with a two-dimensional culture system is to encapsulate cells inside a three-dimensional hydrogel matrix. Originally, cell encapsulation technologies using hydrogels were developed for tissue engineering or therapeutic cell transplantation to prevent the rejection of transplanted cells by the host’s immune system. Hydrogels have been widely used because of their high water content, softness, pliability, biocompatibility, and easily controlled mass transfer properties, essential for allowing transport of nutrients to and waste products from cells (18–22). Recently, micropatterned hydrogels were prepared using microfabrication technique for the development of electrochemical and optical sensors (23, 24). Hydrogel micropattern could be also fabricated inside microfluidic devices for use as “smart” flow controllers or various sensor applications (25).
Cell Microarrays Based on Hydrogel Microstructures
135
In this chapter, we first describe the fabrication of cell microarray using poly(ethylene glycol)(PEG)-based hydrogel. Arrays of hydrogel microstructures encapsulating one or more different cell phenotypes can be fabricated using photolithography or photoreaction injection molding using microfluidic networks made from poly(dimethylsiloxane) (PDMS) (26–28). Next, we describe the incorporation and culture of hydrogel-encapsulated mammalian cells within microfluidic devices (29). Finally, investigation of viability and intracellular enzyme reactions of encapsulated cells is described as examples of cell-based assay systems that can potentially be used in biosensing application (29, 30).
2. Materials 2.1. Substrate Preparation
1. 3-(Trichlorosilyl)propyl methacrylate (TPM) (Fluka Chem.)
2.2. Microchannels for Injection Molding and Microfluidic Device
1. Silicon master, which has negative patterns of desired microchannels defined with SU-8 photoresist (Microlithography Chemical Corp.).
2.3. PEG Hydrogel Microarray
1. Poly(ethyelene glycol) diacrylate (PEG-DA) (MW 4,000) (Polysciences).
2. Hydrogel peroxide (30 wt% in water), sulfuric acid (30 wt% in water), carbon tetrachloride, and n-heptane.
2. PDMS (Sylgard 184, Dow Corning), which is composed of prepolymer and curing agent.
2. Acryloyl-PEG-n-hydroxysuccinimide ester (acryloyl-PEGNHS, MW 3,400) (Nektar). 3. Gly–Arg–Gly–Asp–Ser (GRGDS) peptide (CalBioChem). 4. 2-Hydroxy-1[4-(hydroxyethoxy)phenyl]-2-methyl-1-propanone (Ciba) as a photoinitiator. 5. Photomask (advanced reproductions). 2.4. Cell Culture
1. Murine 3T3 fibroblast, SV-40 transformed murine hepatocyte, and SV-40 transformed murine peritoneal macrophage cell (American Type Culture Collection). 2. Dulbecco’s modified Eagle’s medium (DMEM), RPMI 1640 medium, fetal bovine serum (FBS), antibiotic/antimycotic solution, dexamethasone, trypsin, ethylenediaminetetraacetate (EDTA), sodium chloride, sodium phosphate, and potassium phosphate monobasic (Sigma).
2.5. Fluorescence Detection
1. Calcein-AM (molecular probes). 2. Live/dead viability/cytotoxicity assay kit (molecular probes).
136
Koh
2.6. Equipment
1. Oxygen plasma (Harrick Scientific Co.). 2. Syringe pumps (Harvard Inst.). 3. UV spot lamp (EXFO Corp.). 4. A Zeiss Axiovert 200 microscope equipped with an integrated color CCD camera (Carl Zeiss Inc.). 5. Image analysis software (KS 300, Carl Zeiss Inc.).
3. Methods In this section, we describe experimental protocol to create single or multi-phenotype cell microarrays by combining cell encapsulation and micropatterning process. First, cell-containing PEG hydrogel microarrays are fabricated on the TPM-modified glass substrates using photolithography or photoreaction injection. Mammalian cells encapsulated in PEG hydrogel microarray can be also prepared and grown under static culture conditions in microfluidic devices. Finally, encapsulated cells are examined for their response to the addition of model chemotoxins to demonstrate possible application of cell microarray to biosensor. The methods described outline the following: 1. Surface modification of glass substrate 2. Preparation of PDMS microchannel molds 3. Fabrication of cell-containing hydrogel microarray 4. Incorporation of cell microarray into microfluidic device 5. Toxin detection using cell microarray 3.1. Surface Modification of Glass Substrate
When PEG hydrogel microarrays are generated on glass substrate without surface modification, array elements are easily delaminated upon hydration due to swelling of the cross-linked PEG hydrogel. To prevent delamination, a self-assembled monolayer of TPM on glass is used to create a reactive surface onto which the gel is covalently affixed during photopolymerization (see Note 1). 1. Glass substrates are immersed in piranha solution, consisting of a 3:1 ratio of 30% sulfuric acid and 30% hydrogen peroxide (caution: this mixture reacts violently with organic materials and must be handled with extreme care), for 30 min at 80°C, washed with deionized water thoroughly and dried under nitrogen. 2. Substrates were then treated for 5 min, at room temperature, in a 1 mM solution of TPM in a 4:1 ratio of heptane–carbon tetrachloride in an N2 atmosphere, followed by washing with hexane and water.
Cell Microarrays Based on Hydrogel Microstructures
3.2. Preparation of PDMS Microchannel Molds 3.2.1. Fabrication of Master Using SU-8 Photoresist on Silicon Wafer 3.2.2. Preparation of PDMS Mold
137
1. SU-8 50 photoresist is spin-coated for 20 s at 2,000 rpm and baked for 20 min at 95°C on a horizontal hot plate. 2. Photoresist-coated wafers are then exposed to 365 nm UV light through the photomasks, which have designed microchannels, followed by post-bake at 95°C for 15 min. 3. Finally, dissolving away the unpolymerzied photoresist using developer leaves a positive relief that serves as a master. 1. PDMS precursor is prepared by mixing a PDMS prepolymer with curing agent in a 10:1 ratio by weight. 2. This mixture is poured onto the silicon master and then placed in vacuum desiccators to evacuate the bubbles created during mixing. 3. The PDMS is cured in an oven at 60°C for at least 2 h and the replica is peeled from the master. 4. Several holes are punched through PDMS replica using 16-gauge needle to access the microchannels. 5. PDMS replicas are treated with oxygen plasma for 1 min to change its hydrophobic surface to hydrophilic. 6. Oxidized PDMS microchannels are placed by hand on the TPM-modified glass to form an enclosed channel. Here, reversible, conformal sealing with TPM-modified glass surfaces was used. (These PDMS microchannel systems are used as mold inserts for photoreaction injection molding.)
3.3. Cell Culture
1. Murine fibroblasts are cultured in DMEM with 4.5 g/L glucose and 10% FBS. SV-40 transformed murine hepatocytes are cultured in DMEM containing 1.0 g/L glucose, 200 nM dexamethasone, and 4% FBS. Both phenotypes are grown to confluence in 75 cm2 polystyrene tissue culture flasks, and confluent cells are subcultured every 2–3 days by trypsinization with 0.25% (w/v) trypsin and 0.13% (w/v) EDTA. 2. SV-40-transformed murine macrophages are cultured in RPMI 1640 medium with 2 mM l-glutamine containing 1.5 g/L sodium bicarbonate, 4.5 g/L glucose, 10 mM HEPES, 1.0 mM sodium pyruvate, and 10% FBS. Confluent cells are subcultured every 2–3 days by cell scraping. 3. All cell lines are incubated at 37°C in 5% CO2 and 95% air.
3.4. Fabrication of Cell-Containing Hydrogel Microarray
Cell microarray is prepared by encapsulating mammalian cells inside hydrogel microarray to mimic in vivo environment where cells exist in a three-dimensional hydrogel matrix consisting of proteins and polysaccharides so that more accurate response of cells against analytes can be obtained. PEG hydrogel is modified
138
Koh
with Arg–Gly–Asp (RGD) peptide sequence to promote adhesion and spreading of encapsulated cells. Two different methods are used to create hydrogel microarray. First method is using spincoating and photolithography, while second method is using photoreaction injection molding with microfluidic networks. 3.4.1. Preparation of Hydrogel Precursor Solution
1. Incorporation of the GRGDS peptide: the peptide is conjugated to PEG by reacting the peptide with acryloyl-PEGNHS. The peptide is dissolved to a final concentration of 1 mg/mL in culture media and then 10% w/v of acryloylPEG-NHS is dissolved in peptide solution and reacted at room temperature for at least 2 h (see Note 4). 2. Final precursor solution is prepared by adding 10% w/v of PEG-DA and 0.1% v/v 2-hydroxy-1-[4-(hydroxyethoxy) phenyl]-2-methyl-1-propanone in ethanol as photoinitiator to solution containing acrylated peptide (see Notes 2 and 3). 3. The solution is sterilized by filtration and added to cell suspension (see Note 7).
3.4.2. Cell Microarray Using Photolithography (Fig. 1 Part A)
1. The cell-containing polymer suspension is spin-coated onto functionalized glass substrates to form uniform fluid layer. 2. This layer is covered with a photomask and exposed to 365 nm UV light for 30 s through the photomask. (Proximity
Fig. 1. Part A. (a) Schematic of photoencapsulation of mammalian cells inside hydrogel microarray. (b) SEM and fluorescence images of resultant hydrogel microarray encapsulating fibroblasts (from ref. 26).
Cell Microarrays Based on Hydrogel Microstructures
139
Fig. 1. (continued) Part B. (a) Schematic diagram of the photoreaction injection molding process for creating hydrogel microstructures (from refs. 27). (b) Fabrication of 6 × 6 array of hydrogel microstructure encapsulating three phenotypes of cells using photoreaction injection molding: fibroblasts (upper row), macrophage (middle row), hepatocytes (bottom row) (from refs. 28).
140
Koh
photolithography is used to avoid contact between photomask and precursor solution) (see Note 6). 3. Upon exposure to UV light, only exposed regions undergo free-radical-induced gelation and become insoluble in common PEG solvents such as water. As a result, desired microstructures are obtained by washing away unreacted precursor solution with PBS or cell culture medium so that only the hydrogel microstructures remain on the substrate surface. 4. During the UV-light-induced gelation process, cells suspended in the precursor solution are encapsulated in the resultant hydrogel microstructures. 5. Glass substrates with cell-containing microstructures are immersed in cell culture media and incubated in a 5% CO2. 3.4.3. Cell Microarray Using Photoreaction Injection Molding (Fig. 1 Part B)
PDMS microchannel systems are used as mold inserts for photoreaction injection molding. Photoreaction injection molding offers several advantages over previously described methods of encapsulating mammalian cells in hydrogel microstructures. For example, a small volume of cell-containing precursor solution is sufficient to fill and be photopolymerized inside microchannel, whereas cell-patterning techniques based on spin-coating requires a much larger volume of precursor solution because of solution loss during the spin-coating procedure. Another important advantage of photoreaction injection molding is the possibility to encapsulate different phenotypes on the same array. 1. Each independent microchannel is filled with cell-containing precursor solution. (For the preparation of multi-phenotype cell microarray, precursor solutions containing different cells are introduced to different microchannels) (see Note 5). 2. To make various patterns of hydrogel microstructures, a photomask is aligned over the microchannels. 3. UV exposure for 30 s. 4. PDMS microfluidic networks are quickly removed from the glass substrate. 5. Final hydrogel microstructures are obtained after washing processes.
3.5. Incorporation of Cell Microarray into Microfluidic Device (Fig. 2)
Microfluidic systems offer several advantages, including decreased sample volume, fewer cells, shorter reaction time, and the ability to perform many experiments in parallel. Microfluidic devices are well suited for biological experiments at cellular level since microchannels within these devices can mimic the physical size found in vivo. Because of small size of microchannels, microfluidic devices also allow adequate oxygenation and fast nutrition diffusion. Such an environment helps cells to easily maintain local
Cell Microarrays Based on Hydrogel Microstructures
141
Fig. 2. (a) Fabrication of cell-containing hydrogel microstructure within a microchannel (First image: A cell-containing hydrogel precursor solution in a microchannel. Second image: Gelation of the hydrogel inside the microchannel after exposure to UV light through a photomask. Third image : A cell-containing hydrogel microstructure inside a microchannel after the removal of unreacted precursor solution.) (b) Hydrogel microarray encapsulating macrophages within microfluidic devices (cell-containing hydrogel microarrays are prepared inside 200 mm-wide microchannels or a 2 mm × 2 mm chamber connected to 100 mm-wide microchannels.) (from refs. 27 and 29).
microenvironment than in macro-scale cell culture flasks and exist in less stressful, more in vivo like surroundings, which can lead to more accurate observation on cellular behavior in response to external stimuli. 1. To create microfluidic system, PDMS replica of desired design and cover glass are placed in low-energy plasma cleaner and oxidized at medium power for 1 min. 2. After removal from the plasma cleaner, two substrates are brought into conformal contact and irreversible sealing forms spontaneously. 3. Enclosed microchannels are treated with dilute TPM solution in perfluorooctane for 10 min immediately after sealing to enhance the adhesion of hydrogel microstructure inside the microchannels. 4. Cell-containing hydrogel precursor solution is introduced to the microfluidic network and exposed to UV light through the photomask (see Note 8).
142
Koh
5. After photopolymerization, cell culture media is introduced into channels at 10–100 mL/min with a syringe pump not only to remove unreacted precursor solution from microchannels but also to culture encapsulated cells inside microfluidic devices. 6. Cell culture media inside microchannels was replaced with fresh media every day. 7. These microfluidic devices containing encapsulated mammalian cells are placed onto Petri dish filled with deionized water or culture media to prevent evaporation of cell culture media from microchannels and incubated at 37°C in 5% CO2 and 95% air. 3.6. Toxin Detection Using Cell Microarray
Cell microarray can be used to detect toxic compounds. Here, we describe experimental protocol for the optical detection of sodium azide, a model toxin, using cell-containing hydrogel microarrays. After incubation with sodium azide, response of encapsulated cells against sodium azide is monitored by observing intercellular reaction or cell viability.
3.6.1. Cell Viability Assays
1. A live/dead viability/cytotoxicity fluorescence assay is used to investigate the viability of encapsulated cells. This assay uses SYTO 10 and Dead Red as fluorophores to distinguish living cells and dead cells. SYTO 10 stains live cells green and Dead Red stains dead cells red. 2. For this assay, 2 mL of two fluorophores are added to 1 mL of cell culture media to make the staining solution. 3. The staining solution is introduced to the encapsulated cells and incubated for 20 min in the dark, at room temperature. 4. Viability of cells encapsulated in hydrogel microarrays is imaged and analyzed using A Zeiss Axiovert 200 microscope equipped with fluorescent optical package. 5. A difference in viability is observed between the cells exposed to sodium azide and unexposed cells. Most of the encapsulated cells are alive before the dose of sodium azide; however, when the cells are exposed to sodium azide, the azide anion kills the cells in the hydrogel microstructures as anticipated, and this results in cells that stained red in the assay (Fig. 3a).
3.6.2. Intercellular Reactions
1. Calcein-AM is used to investigate the intercellular reaction of encapsulated cells after exposure to sodium azide. Calcein-AM is colorless and nonfluorescent which can permeate cell membrane. Once inside cells, ester group of calcein-AM are hydrolyzed by esterase enzymes contained within cells to yield calcein, which is fluorescent-emitting green light. If cells are
Cell Microarrays Based on Hydrogel Microstructures
143
Fig. 3. (a) Results of live/dead fluorescence viability assay for encapsulated cells before and after the exposure to sodium azide. (b) Left : Fluorescence images of encapsulated macrophages incubated with calcein-AM, which show an initial (A and C) and 2-h time lapse (B and D) images of a single array element upon exposure to 0 mM (A and B) and 100 mM (C and D) sodium azide. Right: Change of fluorescent intensity by exposure to different concentrations of sodium azide (from refs. 29 and 30).
damaged, they will lose the ability to convert calcein-AM into calcein, which results in decrease of fluorescence intensity. 2. 10 mM Calcein-AM solution is prepared by diluting calceinAM stock solution dissolved in anhydrous dimethylsulfoxide (DMSO) with cell culture media. 3. Encapsulated cells are incubated with this solution for 30 min inside microchannels and then observed with fluorescence microscope. 4. In general, azide damages cellular mitochondria, effectively blocking the flow of electrons to oxygen and halting the production of ATP in cells. Without ATP production, cell viability is gradually diminished, decreasing the fluorescence intensity of cells reacting with calcein-AM (Fig. 3b).
144
Koh
4. Notes 1. Surface modification of glass surface with TPM should be carried out under nitrogen environment because TPM can react with water in the air. When TPM solution contact with air, transparent solution become opaque. 2. Various molecular weights of PEG can be used. However, for the long-term cultures, higher molecular weight of PEG is more desirable due to the enhanced mass transfer for the supply of nutrient and removal of metabolic wastes. 3. PEG-DA with higher molecular weight of PEG (>4,000 Da) can be prepared by reaction PEG with acyloyl chloride. In brief, PEG (20 g) is dissolved in 200 mL of dry benzene under nitrogen and heated at 40°C until fully dissolved. The solution is cooled in an ice bath, followed by the addition of 0.7 mL of triethylamine and 1.13 mL of acryloyl chloride. The mixture is then heated to reflux for 2 h, followed by stirring overnight at room temperature under nitrogen. The solution is filtered to remove the amine salts formed during the reaction, and then the polymer is precipitated in n-heptane. The final product is isolated as a powder by subsequent drying at room temperature in a vacuum oven. Fourier transform infrared (FTIR) spectroscopy can confirm that hydroxyl groups on the PEG polymer are acrylated to near completion. 4. Incorporation of RGD into hydrogels is achieved by functionalizing the amide terminus of the peptide with an acrylate moiety, enabling the adhesion peptide to copolymerize rapidly with PEG-DA during photopolymerization. Without RGD peptide incorporation, encapsulated cells appear rounded after 24 h and slowly spread over the course of several days due to the non-adhesive nature of PEG hydrogels toward proteins and, hence, toward cells. 5. For the photoreaction injection molding of PEG hydrogels, the gel precursor solution must completely fill the microchannels. Since reversible sealing cannot withstand high pressure in the microchannels, the precursor solution should fill the channel by capillary action. For the precursor solutions described here, however, both PDMS and TPM-modified glass surfaces are hydrophobic; therefore, the solution cannot flow through the channel by capillary action. To solve this problem, PDMS microchannels are treated with oxygen plasma to make them hydrophilic. Oxygen plasma treatment lowered the contact angle of channel surfaces with water to almost zero, allowing channels to be easily filled with the gel precursor solution via capillary action.
Cell Microarrays Based on Hydrogel Microstructures
145
6. For the photopolymerization, the spot UV lamp should be placed at a height such that the cell-containing precursor solutions received a light intensity of approximately 10 mW/cm2. 7. The number of cells in one hydrogel microstructure can be controlled by adjusting the cell density in the hydrogel precursor solution or size of hydrogel microstructures. 8. When hydrogel microarrays are fabricated within microfluidic devices, the introduced precursor solution should be allowed to reach a stationary state and then exposed to UV light to avoid deformed structures. References 1. Pancrazio, J. J., Whelan, J. P., Borkholder, D. A., Ma, W., and Stenger, D. A., Annals of Biomedical Engineering, 73: 697–711 (1999). 2. Castel, D., Pitaval, A., Debily, M., and Gidrol, X., Drug Discovery Today, 11: 616–622 (2006). 3. Sundberg, C. J., Current Opinion in Biotechnology, 11: 47–53 (2000). 4. Bhatia, S. N., Yarmush, M. L., and Toner, M., Journal of Biomedical Materials Research, 34: 189–199 (1997). 5. Kleinfeld, D., Journal of Neuroscience, 8: 4098–4120 (1988). 6. Matsuda, T., Inoue, K., and Sugawara, T., ASAIO Journal, 36: M559–M562 (1990). 7. Matsuda, T., and Inoue, K., ASAIO Journal, 36: M161–M164 (1990). 8. Matsuda, T., and Sugawara, T., Journal of Biomedical Materials Research, 29: 749–756 (1995). 9. Chen, C. S., Biotechnology Progress, 14: 356– 363 (1998). 10. Singhvi, R., Kumer, A., Lopez, G. P., Stephanopoulos, G. N., Daniel, I. C., Wang, D., Whitesides, G. M., and Ingber, D. E., Science, 264: 696–698 (1994). 11. McDonald, J. C., Duffy, D. C., Anderson, J. R., Chiu, D. T., Wu, H., Schueller, O. J. A., and Whitesides, G. M., Electrophoresis, 21: 27–40 (2000). 12. Delamarche, E., Bernard, A., Schmid, H., Bietsch, A., Michel, B., and Biebuyck, H., Journal of American Chemical Society, 120: 500–508 (1998). 13. Ito, Y., Biomaterials, 20: 2333–2342 (1999). 14. Kane, R. S., Takayama, S., Ostuni, E., Ingber, D. E., and Whitesides, G. M., Biomaterials, 20: 2363–2376 (1999). 15. Jung, D. R., Kapur, R., Adams, T., Giuliano, K. A., Mrksich, M., Craighead, H. G., and Taylor, D. L., Critical Reviews in Biotechnology, 21: 111–154 (2001).
16. Park, T. H., and Schuler, M. L., Biotechnology Progress, 3: 335–373 (2003). 17. Whitesides, G. M., Ostuni, E., Takayama, S., Jiang, X., and Ingber, D. E., Annual Review of Biomedical Engineering, 3: 335–373 (2001). 18. Quinn, C. P., Pathak, C. P., Heller, A., and Hubbell, J. A., Biomaterials, 16: 389–396 (1995). 19. Csoregi, E., Quinn, C. P., Schmidtke, D. W., Lindquist, S. E., Pishko, M. V., Ye, L., Katakis, I., and Heller, A., Analytical Biochemistry, 66: 3131–3138 (1994). 20. Pathak, C. P., Sawhney, A. S., and Hubbell, J. A., Journal of the American Chemical Society, 114: 8311–8312 (1992). 21. Mann, B. K., Gobin, A. S., Tsai, A. T., Schmedlen, R. H., and West, J. L., Biomaterials, 22: 3045–3051 (2001). 22. Mellott, M. B., Searcy, K., and Pishko, M. V., Biomaterials, 22: 929–941 (2001). 23. Sirkar, K., and Pishko, M. V., Analytical Chemistry, 70: 2888–2894 (1998). 24. Revzin, A. R., Yadavalli, V., Koh, W., Deister, C., Hile, D., Mellott, M., and Pishko, M. V., Langmuir, 17: 5440–5447 (2001). 25. Beebe, D. J., Moore, J. S., Bauer, J. M., Yu, Q., Liu, R. H., Devadoss, C., and Jo, B., Nature 404: 588–590 (2000). 26. Koh, W. G., Revzin, A., and Pishko, M. V., Langmuir, 18: 2459–2462 (2002). 27. Koh, W. G., and Pishko, M. V., Langmuir, 19: 10310–10316 (2003). 28. Koh, W., Itle, L. J., and Pishko, M. V., Analytical Chemistry, 75: 5783–5789 (2003). 29. Koh, W. G., and Pishko, M. V., Analytical and Bioanalytical Chemistry, 385: 1389–1397 (2006). 30. Itle, L. J., and Pishko, M. V., Analytical Chemistry, 77: 7887–7893 (2005).
Chapter 8 Fabrication of Bacteria and Virus Microarrays Based on Polymeric Capillary Force Lithography Pil J. Yoo Abstract There is a growing interest on the fabrication of bacteria and virus microarray owing to their great potential in many biological applications ranging from diagnostic devices to advanced platforms for fundamental studies on molecular biology. Over the past decade, a number of studies with regard to the biomolecular patterning have been presented. Capillary force lithography (CFL) for polymeric thin films can provide well-ordered microarray structures over a large area in a facile and cost-efficient way while maintaining its biocompatibility during a process. Patterned polymeric structures can be utilized either to physical barriers for the confinement of bacteria or to physicochemical template for the subsequent binding of viruses. In this chapter, we have shown that the patterned structures of poly(ethylene glycol) (PEG) containing polymer enables a selective binding of Escherichia coli, leading to a physically guided microarray of bacteria. Additionally, we demonstrate the fabrication of virus microarray of M13 viruses via electrostatic interactions with a prepatterned microstructure of polyelectrolyte multilayers. Key words: Bacteria microarray, Virus microarray, E. coli, M13 bacteriophages, Patterning, Capillary force lithography, Polydimethylsiloxane, Layer-by-layer assembly, Polyelectrolyte multilayers
1. Introduction The ability to spatially locate and anchor bacteria or viruses at the microscale in a precise manner affords useful platforms for biosensors, disease markers, and drug-screening applications (1–4). Particularly, bacteria are strong candidates for sensing applications since analytical specificity can be readily manipulated via genetic engineering and their micro-organisms are relatively robust as compared to mammalian cells (5). Additionally, with a growing threat of vial outbreaks and the reemergence of previously eradicated viruses, more efficient and reliable methods to identify viruses and to discover the relevant drugs must be investigated (6). Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_8, © Springer Science+Business Media, LLC 2011
147
148
Yoo
Biopatterning technology has advanced at a rapid pace in the last decade with a variety of patterning methodologies for immobilizing and functionalizing these biomaterials onto a designed matrix. Among the diverse approaches, capillary force lithography for polymeric matrix layer has received much attention due to its attractiveness in economical feasibility, straightforward process, accessible scalability, and biocompatibility (7). Here, we focus on two representative methods for the presentation of bacteria and viruses into ordered micropatterns. First method for bacteria microarray fabrication utilizes the strategy of physical confinement of biomolecular species within a polymeric pattern engraved surface (8, 9). In order to prevent a nonspecific binding of bacteria over surface, a bio-repellent barrier layer that is including poly(ethylene glycol) polymer is synthesized (10). Then, the coated polymeric layer is patterned with the capillary force lithographic method, which renders a substrateexposed polymeric microarray structure. After a selective decoration with antibodies onto the substrate-exposed regions, bacteria are subsequently deposited through host–parasite interactions. As a result, the micropatterned array where bacteria are confined by the physical walls of polymeric structure can be obtained. Second method for virus microarray is employing electrostatic interactions between viruses and functionalized polymeric patterns. Layer-by-layer assembly of polyelectrolyte multilayers is adopted to prepare the polymeric base layer (11). Then, the solvent-assisted capillary molding, a modification of conventional capillary force lithographic method, is performed under ambient environment to pattern the polyelectrolyte multilayers that are prone to be thermally cross-linked (12, 13). The next deposition of charged viruses atop the prepatterned polymeric structure gives rise to the electrostatic binding between viruses and charged polyelectrolyte multilayers and thus leaves a micropatterned array where the viruses are placed onto the underlying polymeric patterns.
2. Materials 2.1. Escherichia coli Growth and Virus Amplification
1. TBS buffer: 50 mM Tris–HCl and 150 mM NaCl, pH 7.5. Store at room temperature. 2. Luria-Bertani (LB) medium: Dissolve 10 g/L tryptone (BD), 5 g/L yeast extract (BD), and 10 g/L NaCl in deionized water. Autoclave and store at room temperature. 3. LB agar: Dissolve 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, and 15 g/L agar in deionized water. Autoclave and store at room temperature.
Bacteria and Virus Microarrays
149
4. LB plate: LB medium with agar of 15 g/L is sterilized by autoclaving. 5. LB-Tet plate: Autoclave LB medium with agar of 15 g/L. Add 1 mL of tetracycline hydrochloride stock (20 mg/mL in ethanol, Aldrich). Pour the stock onto plate (Petri dish) and store them at 4°C in dark. Final concentration of tetracycline hydrochloride is 20 mg/mL. 6. YT medium containing 5 mM MgCl2 (×2): Add 16 g of tryptone, 10 g of yeast extract, and 5 g of NaCl to 900 mL of deionized water. Dissolve all the components and adjust the solution pH to 7.0 with 5 N of NaOH. Finally, adjust volume to 1 L and autoclave. 7. M13 virus (bacteriophage): 50 mL of ER2738 [Wild-type M13 virus, Escherichia coli (E. coli) strain ER2738, New England Biolabs] is overnight cultured in LB medium from a colony (see Note 1). 8. PEG/NaCl: Prepare solution of 20 wt% PEG (crystalline powder, average Mn ~ 8,000, Aldrich) with 2.5 M of NaCl. Autoclave and store at room temperature. 2.2. Fabrication of Bacteria Microarrays 2.2.1. Synthesis of Nonbiofouling Copolymer of Poly (TMSMA-r-PEGMA)
1. Monomer 1: 2.5 g of 3-(trimethoxysilyl) propyl metharylate (TMSMA) (10 mmol, Aldrich) is used as received. This reagent is moisture-sensitive; therefore, it should be transferred with a double-ended needle under nitrogen purging. 2. Monomer 2: 4.75 g of poly(ethylene glycol) methyl ether methacrylate (PEGMA) (10 mmol, Aldrich, average Mn ~ 475) is used as received. 3. Initiator: 16.5 mg of 2,2′-azobisisobutyronitrile (AIBN) (0.1 mmol, Aldrich, 98%) is used as received. AIBN is a free radical initiator and very reactive and explosive. So, special caution is required. 4. Inhibitor remover: Prepacked column for removing hydroquinone and monomethyl ether hydroquinone (Aldrich), used as received. 5. Reaction solvent: 10 mL of tetrahydrofuran (Aldrich, anhydrous, inhibitor free, 99.9%) is used as received.
2.2.2. Patterning of Nonbiofouling Copolymer Film and Fabrication of Bacteria Array
1. Substrates: Diced silicon substrates (25 × 25 mm) from 8-in. silicon wafer (orientation ~100, test grade, Silicon Quest). 2. Silicon master pattern: Prepare the master using a silicon wafer that had been thermally oxidized to a thickness of 1 mm. Emboss micrometer scaled line-and-space patterns or dot array patterns onto the wafer surface by photolithography. 3. Trichloroethylene: Purchased from Aldrich, used as cleaning solvent for the substrates.
150
Yoo
4. Methyl alcohol: Purchased from Aldrich, used as cleaning solvent for the substrates. 5. Polydimethylsiloxane (PDMS) mold: Prepare the replicated PDMS molds using thermally casting prepolymer (Sylgard 184, Dow Corning) onto the complimentary relief structures of a silicon master. To thermally cure the PDMS molds, mix the curing agent and the prepolymer with a ratio of 1:10 (wt%), and incubate at 70°C for 6 h. Then, peel off from the master and cut into desired sizes (see Note 2). 6. Antibody: Anti-M13 P3 monoclonal antibody (New England Biolabs), store at −20°C. 2.3. Fabrication of Virus Microarrays 2.3.1. Layer-by-Layer Assembly of Polyelectrolyte Multilayers
1. Cationic polyelectrolyte: Dissolve linear polyethylenimine (LPEI, 25,000 MW, Polysciences) in deionized water to 30 mM based on the repeat-unit molecular weight, e.g., 1.50 g LPEI in 1 L of water (see Note 3). Then, filter out the impurities and nonsoluble residues through filtered bottle (pore size ~0.2 mm). Store the filtered solution at room temperature at least for 2 weeks before use. 2. Anionic polyelectrolyte: Dissolve poly(acrylic acid) (PAA, 90,000 MW, 25% aqueous solution, Polysciences) in water at 20 mM based on the repeat-unit molecular weight, e.g., 5.77 g PAA (25% aqueous solution) in 1 L of water. Store the filtered solution at room temperature at least for 2 weeks before use. 3. Substrates: Glass substrate (VWR international, 75 × 25 mm) or diced silicon substrates (60 × 25 mm) from 8-in. silicon wafer (orientation ~100, test grade, Silicon Quest).
2.3.2. Preparation of Polymeric Microarray Using Solvent-Assisted Capillary Molding
1. Silicon master pattern: Prepare the master using a silicon wafer that had been thermally oxidized to a thickness of 1 mm. Emboss micrometer scaled line-and-space patterns or dot array patterns onto the wafer surface by photolithography. 2. PDMS mold: Prepare the replicated PDMS molds using thermally casting prepolymer (Sylgard 184, Dow Corning) onto the complimentary relief structures of a silicon master. To thermally cure the PDMS molds, mix the curing agent and the prepolymer with a ratio of 1:10 (wt%), and incubate at 70°C for 6 h. Then, peel off from the master and cut into desired sizes.
2.3.3. Proteins Binding onto Virus Array
1. PBS buffer: Dilute phosphate buffer solution (pH 7.2, Aldrich) to 1/10 of the original concentration in deionized water and sterilize by filtering.
Bacteria and Virus Microarrays
151
2. Blocking agent: Prepare Tween 20 (Aldrich) stock solution to 0.2% (v/v) solution by serial dilution of original solution in PBS (see Note 4). 3. Virus-binding protein: Biotin-conjugated anti-fd bacteriophage (Sigma), which is developed in rabbits using repeated injections of M13 bacteriophage as the immunogen, is diluted to 3 mg/mL in the prepared phosphate buffer solution. 4. Fluorescence-labeled Streptavidin: Streptavidin conjugated Alexa Fluors 430 and 568 (molecular probes, green and red fluorophore, respectively) is diluted to 5 mg/mL in the prepared phosphate buffer solution.
3. Methods In this method section, the protocols for E. coli growth and M13 bacteriophage amplification are firstly presented as preparatory materials for the fabrication of bacteria and virus microarray. Bacteria microarray can be obtained using a scheme of the physical confinement of bacteria within patterned polymeric microstructures. In the course of bacteria ordering, alleviating nonspecific adsorption of bacteria onto polymeric micropatterns is very important to obtain the pattern selectivity of the microarray. The synthesized PEG containing nonbiofouling polymer is thus introduced as a nonbiofouling material. For the fabrication of virus microarray, instead of using the strategy of the physical confinement, the ability of physicochemical templating of the patterned polymeric structures can be exploited. To provide these properties, layer-by-layer assembled polyelectrolyte multilayers are employed as a patternable polymeric layer. It is worthwhile noting that the viruses are electrostatically bound onto the protruded surface of polymeric patterns of the polyelectrolyte multilayers (templated assembly), which is contrasted to the case of bacteria microarray, where bacteria are placed only in the recessed regions surrounded by topographically patterned polymeric barriers (confinement-induced assembly). 3.1. E. coli Growth in LB Media 3.1.1. Day 1: Media Preparation
This protocol gives the steps to grow E. coli in LB media from a stab culture. The culture volume is 250 mL. 1. Prepare 250 mL of liquid LB media and aliquot 50 mL each into 250 mL flasks. If necessary, adjust pH to 7.0 with 5 M NaOH. 2. Prepare 100 mL of LB agar.
152
Yoo
3. Prepare 80% (v/v) glycerol solution to make glycerol stocks of cell strains. 4. Autoclave the LB media and LB agar prepared in steps 1–3. Both of LB media and LB agar would become clear light yellow solutions after autoclaving. LB agar can be stored in 60°C oven before use. 5. Prepare stock antibiotic solutions (50 mg/mL ampicillin stock) by filter sterilizing. Distribute 1 mL aliquots into sterilized Eppendorf tubes, and store them at −20°C in a dark condition. 3.1.2. Day 2: Inoculation by Streaking
1. Cool down the autoclaved LB agar to below 50°C (could be touched by hands), then add 100 mL ampicillin stock solution into 100 mL LB agar. Pour the LB agar into five petri dishes and let them solidify. 2. Inoculate the agar plates from stab culture by streaking. Flame a wire loop and cool on a spare sterile agar plate. Dip the wire loop into the stab culture, and streak an inoculum of E. coli across one corner of a fresh agar plate. Flame and cool the wire loop again. Pass it through the first streak and then streak again across a fresh corner of the plate. 3. Incubate the plate upside down in 37°C incubator for 12–24 h until colonies develop.
3.1.3. Day 3: Grow E. coli in LB Media
1. Remove plates from incubator. Seal the plates with parafilm and keep them at 4°C if not use immediately. 2. Before transfering colonies from LB agar plates into LB liquid media, add 50 mL ampicillin stock solution into each flask, which contains 50 mL liquid media. 3. Take agar plates from refrigerator. Pick one colony with blunt toothpicks to inoculate one 250 mL shake flask. 4. Incubate all flasks overnight at 37°C with 250 rpm.
3.1.4. Day 4: Make Master Bank and Working Bank of Cell Strains
1. Harvest the cell culture. 2. Get two cryogenic vials (1.8 mL) for each cell strain. Add 200 mL sterilized 80% glycerol into each vial first. Then, add 800 mL overnight cell culture into each vial (the final concentration of glycerol is 16%). Mix them well. 3. Label the vials and keep them at −80°C as glycerol stocks. One vial of each cell strain works as the master bank, and the other works as the working bank.
3.2. Virus Amplification and Extraction
1. Set up an overnight culture of ER2738 in LB-Tet from a colony. Dilute overnight culture 1:100 in 2× YT medium containing 5 mM MgCl2.
Bacteria and Virus Microarrays
153
2. Add the neutralized phage elute (~1 mL) to the diluted 20 mL ER2738 culture and incubate at 37°C with vigorous shaking for 4–5 h. 3. Transfer the culture to a centrifuge tube and spin 10 min at 10,000 rpm at 4°C. 4. Transfer the supernatant to a fresh tube and re-spin. 5. Pipet the upper 80% of the supernatant to a fresh tube. The volume of PEG/NaCl added to the supernatant is 1/5 of the supernatant volume. Allow phage to precipitate at 4°C for 1 h (see Note 5). 6. Spin PEG precipitation 15 min at 10,000 rpm, at 4°C. Decant supernatant, re-spin briefly, and remove residual supernatant with a pipette. 7. Suspend the pellet in 1 mL TBS. 8. Transfer the suspension to a microcentrifuge tube and spin for 5 min at 4°C to pellet residual cells. 9. Transfer the supernatant to a fresh microcentrifuge tube and re-precipitate with 1/6 volume of PEG/NaCl. Incubate on ice 15–60 min. 10. Microcentrifuge for 10 min at 4°C. Discard supernatant, respin briefly, and remove residual supernatant with a micropipette. 11. Suspend the pellet in 200 mL TBS (0.02% NaN3). 12. Microcentrifuge for 1 min to pellet any remaining insoluble matter. 13. Transfer the supernatant to a fresh tube. This is the amplified elute. Take 10 mL to titer. The rest are stored at 4°C. For a long-term storage of amplified phage, dilute 1:1 with sterile glycerol (20%) and store at −20°C. 3.3. Fabrication of Bacteria Microarrays 3.3.1. Synthesis of Nonbiofouling Copolymer of Poly(TMSMA-r-PEGMA) (see Note 6)
1. Prior to polymerization, flow the neat PEGMA through the inhibitor removal column. 2. Place 10 mmol PEGMA, 10 mmol TMSMA, and 0.01 mmol AIBN in a vial and dissolve in 10 mL of anhydrous tetrahydrofuran. 3. Degas the mixture for 20 min using an Ar gas stream and seal the vial with a Teflon-lined screw cap. 4. To carry out the polymerization, place the mixture at 70°C for 24 h with a slight stirring, then evaporate the solvent under vacuum condition. As a result, a viscous polymeric liquid of poly(TMSMA-r-PEGMA) is obtained (see Note 7).
154
Yoo
Fig. 1. (a) Experimental procedure of fabricating bacteria arrays using capillary force lithography and host–parasite/virus–antibody interactions. (b) Scanning electron microscope (SEM) image of a single Escherichia coli (E. coli ) on a silicon substrate. Bar scale is 1 mm. (c) Fluorescent micrograph for FITC-labeled P3 antibody of 50 mm circle pattern. (d) Optical micrograph for subsequently adsorbed E. coli onto the same pattern. (e) SEM micrograph for single arrays of E. coli along the channel direction. Pattern spacing is 1 mm. 3.3.2. Patterning of Nonbiofouling Copolymer Film and Fabrication of Bacteria Array
1. Clean silicon wafer or glass substrate by ultrasonification in trichloroethylene and methanol for 5 min each, then dry in nitrogen. The native oxide is not removed from the surface and would exist. 2. Place a few drops of a 1–10 wt% solution of poly(TMSMA-rPEGMA) on a silicon wafer and spin-coat at 1,000 rpm for 10 s. 3. Place patterned PDMS mold carefully onto the polymer surface with ensuring a conformal contact (capillary force lithography). 4. Store the samples overnight at room temperature to allow for complete evaporation of the solvent. Ellipsometric measure-
Bacteria and Virus Microarrays
155
ment reveals that the patterned film thickness ranges from 32 to 370 nm depending on the solution concentration of poly(TMSMA-r-PEGMA) (see Note 8). 5. To selectively attach the antibodies inside the polymeric patterns, prepare the diluted solution of P3 antibodies at 1:100 in PBS (see Note 9). 6. Spread a few drops of the antibody solution evenly on the surface for an hour. Then, rinse the surface with diluted PBS buffer for 1 min. 7. Place the samples in a solution of E. coli (cultured with M13 viruses) for 4 h to attach bacteria onto the patterned antibodies with an aid of the virus–antibody (host–parasite) interactions. Finally, rinse the surface with diluted PBS buffer for 1 min. 3.3.3. Characterization and Imaging of Virus Array
1. The patterned bacteria array is characterized with atomic force microscope (AFM, Veeco, Dimension 3100 with Nanoscope IIIa controller) in dry condition. In order to minimize any possible misreading during data acquisition on topologically patterned surface, use in a slow scanning tapping mode (0.5– 1.0 Hz of scan speed). 2. Scanning electron microscope can be used to capture the images of bacteria array as well. Prior to imaging, deposit 5 nm thicknesses of the Au layer on the patterned surface to prevent charging.
3.4. Fabrication of Virus Microarrays 3.4.1. Layer-by-Layer Assembly of Polyelectrolyte Multilayers
1. Clean a silicon wafer or a glass substrate by ultrasonification in trichloroethylene and methanol for 5 min each, then dry in nitrogen. 2. Plasma-treat the cleaned substrate with a conventional plasma cleaner for 30 s (PDC-001, Harrick Scientific Corp.) to prepare an initial negatively charged surface. 3. Adjust pH of LPEI and PAA solutions to 4.7 carefully with diluted solutions of hydrochloric acid and sodium hydroxide. 4. Prepare the polyelectrolyte multilayers of LPEI and PAA using a programmable slide stainer (HMS-70, Microm) with a deposition condition of 15 min adsorption of polyelectrolyte and followed by three sequential washing steps in the bath of deionized water (2 min for each washing step). To compensate the negative surface charge of the substrate, deposit positively charged LPEI firstly onto the substrate. For the deposition of 200 nm thick film of LPEI/PAA, prepare 11.5 bilayers of LPEI/PAA, (LPEI/PAA 4.7/4.7)11.5 (see Note 9). 5. Remove the loaded samples from the slide stainer and store at ambient temperature in dry condition.
156
Yoo
Fig. 2. (a) Experimental procedure of fabricating virus arrays using solvent-assisted capillary molding and electrostatic interactions between viruses and polyelectrolyte multilayers. (b) Three-dimensional atomic force microscope (AFM) image for assembled viruses atop patterned template of polyelectrolyte multilayers (1.5 mm circle pattern). (c) Twodimensional AFM image for assembled viruses (phase mode, Z-range = 30°). (d) Phase mode AFM images for assembled viruses from “before biotinylation” (left ) to “after biotinylation” (right ). (e) Green fluorescence image of Alexa Fluor 430 streptavidin bound viral dot array. (Reproduced from [13] with permission from American Chemical Society).
3.4.2. Preparation of Polymeric Microarray Using Solvent-Assisted Capillary Molding (see Note 10)
1. Place the prepared samples of the LPEI/PAA film in a mildly heated humid chamber (normally at 80°C, 100% relative humidity) for 30 min to soften the film prior to contact with PDMS mold. 2. Place the PDMS mold onto a vapor-exposed polymeric film with a slight pressure of a few bars to promote the capillary
Bacteria and Virus Microarrays
157
molding. After 3–4 h of heating in a vapor-saturated environment, detach the PDMS mold from the sample surface and the replicated patterns of the polyelectrolyte multilayers remained on the surface. 3. To remove the very thin residual layers of polymer from the mold contact regions, treat the patterned surface with 1 min of plasma cleaning, during which an air plasma is irradiated onto the sample with an etching rate of about 20–30 nm/ min for the polymeric layer. As a result, one can assure the clean removal of residual polymeric layer from the interstitial regions between patterns. 3.4.3. Electrostatic Self-Assembly of M13 Viruses on Polymeric Micro Array
1. Dilute the stock solution of wild-type M13 virus in deionized water to the final concentration of 109–1010 molecules/mL. 2. Adjust the solution pH of virus to 4.8 by adding 0.01 M HCl or 0.01 M NaOH, near the isoelectric point of M13 viruses (see Note 11). 3. Drop-dispense the pH-adjusted virus solution (~200 mL/ cm2) on prepared polymer surface for 20–30 min at ambient temperature. During an adsorption process, negatively charged M13 viruses were electrostatically bound on the positively charged top surface of LPEI/PAA multilayer, finally leading to an ordered monolayer structure of M13 viruses on the surface. 4. Rinse the virus assembled surfaces with deionized water several times to remove loosely bound viruses and dried by blowing with nitrogen.
3.4.4. Proteins Binding onto Virus Array
1. To attach biotin groups onto assembled M13 viruses, dropdispense a diluted solution of the biotin-conjugated anti-fd bacteriophage (~200 mL/cm2) onto the patterned virus surface for 5 min. Then, rinse the surface with diluted PBS buffer for 1 min. 2. For the biotin–streptavidin coupling reaction, similar procedures can be applied by using fluorophore-tagged streptavidin. Drop-dispense a diluted solution of the streptavidin conjugated Alexa Fluors 430 or 568 (~200 mL/cm2) onto the patterned biotins atop viruses for 20 min. Then, rinse the surface with diluted PBS buffer several times and dry with nitrogen.
3.4.5. Characterization and Imaging of Virus Array
1. The patterned assembly of viruses is mainly characterized with AFM (Veeco, Dimension 3100 with Nanoscope IIIa controller) in dry condition. In order to minimize any possible misreading during data acquisition on topologically patterned surface and to enhance the image resolution, operate
158
Yoo
in a slow scanning tapping mode (0.5–1.0 Hz of scan speed) with super sharp silicon probes (Pacific Nanotechnology, SSSNCH, typical tip radius of curvature ~2 nm). 2. Fluorescence microscopic images of the labeled virus array are obtained with microscope (Zeiss Axioplan 2, Carl Zeiss).
4. Notes 1. The M13 bacteriophage, a virus that only infects bacteria, is composed of ~2,700 major coat proteins helically stacked around its single-stranded DNA, rendering a monodispersed and semi-flexible filamentous structure (880 nm in length and 6.6 nm in width). 2. In spite of being cured and solidified, cured PDMS contains considerable amounts of unthethered siloxane chains that can provide lubrication property to the surface. This lubrication effect lowers the surface energy of the PDMS mold, which is required surface property as a mold material. Therefore, freshly made PDMS molds are to be used for a molding process. 3. Due to its linear and regular structure, LPEI forms a crystalline solid when unprotonated and is thus insoluble in water above its pKa (~4.5). To dissolve LPEI into water, therefore, a small amount of HCl solution was added to a mixture of LPEI and deionized water until all the LPEI was dissolved. 4. Tween 20 detergent is employed to minimize any nonspecific (e.g., electrostatic or van der Waals interactions) binding of the antibodies or proteins onto the prepared patterns of polyelectrolyte multilayers and the substrate. 5. PEG is a long-chain polymeric compound which, in the presence of salt, absorbs water, thereby causing macromolecular assemblies such as phage particles to precipitate. 6. Random copolymer of poly(TMSMA-r-PEGMA) is comprised of “anchor part” of trialkoxysilane and a “function part” of PEG. Incorporation of the surface-reactive trimethoxysilyl group in the monomer can allow the copolymer to form multiple covalent bonds onto the oxide surface to provide multiple PEG immobilizations. Consequently, nonspecific protein adsorption and biomolecules adhesion can be significantly reduced. 7. The molecular weight of poly(TMSMA-r-PEGMA) was Mn = 26,000 with Mw/Mn = 1.88 as measured by gel permeation chromatography (GPC) relative to monodisperse polystyrene standards.
Bacteria and Virus Microarrays
159
8. Capillary force lithography of polymeric molding is a powerful method to construct microscale polymeric patterns on the surface. One potential concern with this technique is whether the substrate surface can be completely exposed, which is a prerequisite to enable the successful attachment of biological species. The PEG moieties in poly(TMSMA-r-PEGMA) allows a full exposure of the substrate surface during a molding process. 9. The nomenclature (A/B m/n)X is used to denote a multilayer film of X layer pairs of cationic A and anionic B deposited at pH m and n, respectively. When X includes 0.5, cationic A is the final adsorbed layer and thus the outermost surface of the multilayer. 10. Generally, polymeric molding process is an efficient way of generating physically patterned polymeric structures. Among several approaches, capillary force lithography is a promising candidate that utilizes the fluidic mobility of polymers at high temperature conditions (above the Tg of the polymer) to induce a fluidic mobility for a molding. Using thermal mobility, however, is not applicable to most layer-by-layer assembled polyelectrolyte multilayers film systems because in the dry state, the ionic complexes might be considered as networks based on multivalent ionic crosslinks that impede flow, and furthermore, the heating of polyacid/polyamine multilayers to high temperatures usually brings about thermal crosslinking between cationic and anionic species. To overcome this challenge, instead of utilizing thermally induced mobility, one should turn to another means of achieving polymeric mobility via softening of the multilayer film with suitable solvents (solvent-assisted capillary molding). For polyelectrolyte multilayers, water vapor can be a relevant solvent. Notably, cured PDMS does not significantly take up water (<1 wt% of maximum water uptake), a problematic deformation of the mold thus can be avoided during a molding process. Moreover, a good gas permeability of the PDMS mold ensures the capillary action in the confined environment of patterns. 11. Notably, no buffer was used for the preparation of the virus solution to prevent affecting the underlying PEM patterns, which are susceptible to reorganization or dissociation in a strong ionic environment. Similarly, very careful pH adjustment of the virus solution was performed, and as a result, the total concentration of added salts from highly diluted HCl and NaOH was maintained at <1 mM.
160
Yoo
References 1. Falconnet, D., Csucs, G., Grandin, H. M., and Textor, M. (2006) Surface engineering approaches to micropattern surfaces for cellbased assays. Biomaterials 27, 3044–3063. 2. Rowan, B., Wheeler, M. A., and Crooks, R. M. (2002) Patterning bacteria within hyperbranched polymer film templates. Langmuir 18, 9914–9917. 3. Lee, K. B., Kim, E. Y., Mirkin, C. A., and Wolinsky, S. M. (2004) The use of nanoarrays for highly sensitive and selective detection of human immunodeficiency virus type 1 in plasma. Nano Lett. 4, 1869–1872. 4. Cheung, C. L., Camarero, J. A., Woods, B. W., Lin, T., Johnson, J. E., and De Yoreo, J. J. (2003) Fabrication of assembled virus nanostructures on templates of chemose lective linkers formed by scanning probe nanolithography. J. Am. Chem. Soc. 125, 6848–6849. 5. Koh, W. G., Revzin, A., Simonian, A., Reeves, T., and Pishko, M. (2003) Control of mammalian cell and bacteria adhesion on substrates micropatterned with poly(ethylene glycol) hydrogels. Biomed. Microdevices 5, 11–19. 6. Wood, M. J. and Moellering, Jr. R. C. (2003) Microbial resistance: bacteria and more. Clin. Infect. Dis. 36, S2–S3.
7. Suh, K. Y., Kim, Y. S., and Lee, H. H. (2001) Capillary force lithography. Adv. Mater. 13, 1386–1389. 8. Suh, K. Y., Khademhosseini, A., Yoo, P. J., and Langer, R. (2004) Patterning and separating infected bacteria using host-parasite and virus-antibody interactions. Biomed. Microdevices 6, 223–229. 9. Suh, K. Y., Khademhosseini, A., Jon, S., and Langer, R. (2006) Direct confinement of individual viruses within polyethylene glycol (PEG) nanowells. Nano Lett. 6, 1196–1201. 10. Jon, S., Seong, J., Khademhosseini, A., Tran, T. N. T., Laibinis, P. E., and Langer, R. (2003) Construction of nonbiofouling surfaces by polymeric self-assembled monolayers. Langmuir 19, 9989–9993. 11. Hammond, P. T. (2004) Form and function in multilayer assembly: new applications at the nanoscale. Adv. Mater. 16, 1271–1293. 12. Yoo, P. J., Nam, K. T., Qi, J., Lee, S. K., Park, J., Belcher, A. M., and Hammond, P. T. (2006) Spontaneous assembly of viruses on multilayered polymer surfaces. Nat. Mater. 5, 234–240. 13. Yoo, P. J., Nam, K. T., Belcher, A. M., and Hammond, P. T. (2008) Solvent-assisted patterning of polyelectrolyte multilayers and selective deposition of virus assemblies. Nano Lett. 8, 1081–1089.
Chapter 9 3D Polymer Scaffold Arrays Carl G. Simon Jr., Yanyin Yang, Shauna M. Dorsey, Murugan Ramalingam, and Kaushik Chatterjee Abstract We have developed a combinatorial platform for fabricating tissue scaffold arrays that can be used for screening cell–material interactions. Traditional research involves preparing samples one at a time for characterization and testing. Combinatorial and high-throughput (CHT) methods lower the cost of research by reducing the amount of time and material required for experiments by combining many samples into miniaturized specimens. In order to help accelerate biomaterials research, many new CHT methods have been developed for screening cell–material interactions where materials are presented to cells as a 2D film or surface. However, biomaterials are frequently used to fabricate 3D scaffolds, cells exist in vivo in a 3D environment and cells cultured in a 3D environment in vitro typically behave more physiologically than those cultured on a 2D surface. Thus, we have developed a platform for fabricating tissue scaffold libraries where biomaterials can be presented to cells in a 3D format. Key words: Array, Biomaterials, Cell adhesion, Cell proliferation, Combinatorial methods, Polymer scaffold
1. Introduction Combinatorial methods, which have been successful at accelerating pharmaceutical research (1, 2), are being developed and applied to accelerate biomaterials research (3–11; reviewed in 11). Meredith et al. (3) fabricated combinatorial polymer libraries containing orthogonal gradients of poly(e-caprolactone):poly (D,L-lactic acid) (PCL:PDLLA) blend composition and annealing temperature. Osteoblasts cultured on the libraries displayed “hot spots” where the blend morphology enhanced differentiation. Anderson et al. (4) used an automated spotter to deposit monomers onto glass slides, which were photopolymerized to create polymer arrays. Polymers that supported human embryonic stem Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_9, © Springer Science+Business Media, LLC 2011
161
162
Simon et al.
adhesion and proliferation were identified. Disney and Seeberger (5) coupled monosaccharides to slides for combinatorial screening of bacterial adhesion and found mannose to be the best surface. Smith et al. (6) cultured fibroblasts on polymer arrays made in 96-well plates and observed a range of cellular metabolic activities that were used in computational modeling that identified polymer descriptors that were predictive of cell response. Anderson et al. (7) used an array spotter for making polymer library arrays of several polymers and their blends and observed that blends containing poly(ethylene glycol) tended to inhibit cell adhesion. Flaim et al. (8) robot-spotted several extracellular matrix proteins and their blends into arrays and observed that hepatocyte differentiation was enhanced when collagen IV was present. Simon et al. (9) made gradients of polymer composition and used an automated image acquisition and analysis system to determine that cell adhesion and proliferation were enhanced in specific regions of the libraries. Finally, Gallant et al. (10) used “click” chemistry to synthesize gradients of surface-coupled RGD cell adhesive peptides to identify optimal coupling densities for cell adhesion and proliferation. These studies demonstrate many approaches for fabricating biomaterial arrays and how they can be used to identify the material formulations that promote different cellular behaviors. Although previous methods for rapid screening of cell– biomaterial interactions have primarily focused on 2D films or surfaces (3–11), biomaterials are frequently used to fabricate 3D scaffolds (12), cells exist in vivo in a 3D environment and cells cultured in a 3D environment in vitro typically behave more physiologically than those cultured on a 2D surface (13–15). For these reasons, a combinatorial approach in which cell–biomaterial interactions are screened using a 3D polymer scaffold configuration will provide more relevant information regarding cell responses to test biomaterials. Toward this end, we have developed methods for fabricating combinatorial arrays of polymer scaffolds where scaffold composition and properties are varied (16, 17). The scaffold arrays can be used for rapid screening of cell response to identify scaffold compositions that best support cell adhesion, proliferation, and differentiation. Herein, we provide protocols for fabricating the polymer scaffold arrays and how to use them in screening cell response.
2. Materials 1. NaCl (Sigma). 2. Sudan IV (Sigma). 3. 1,4-Dioxane (Sigma).
3D Polymer Scaffold Arrays
163
4. Poly(D,L-lactic acid) (PDLLA), Mw 100,000 g/mol (LactelBirmingham Polymers). 5. Poly(D,L-lactic acid), Mw 103,000 g/mol (Absorbable Polymers International). 6. Poly(D,L-lactic acid), Mw 109,000 g/mol (MedisorbAlkermes). 7. Poly(desaminotyrosyl-tyrosine ethyl ester carbonate) (pDTEc), Mw 183,000 g/mol (gift from Joachim Kohn at New Jersey Center for Biomaterials). 8. Poly(desaminotyrosyl-tyrosine octyl ester carbonate) (pDTOc), Mw 122,800 g/mol (gift from Joachim Kohn at New Jersey Center for Biomaterials). 9. Poly(e-caprolactone) (PCL), Mw 80,000 g/mol (Sigma). 10. MC3T3-E1 mouse osteoblast cell line (Riken Cell Bank, Hirosaka, Japan). 11. a-Modification of Eagle’s minimum essential medium (Cambrex Bio Science). 12. Fetal bovine serum (Gibco). 13. Kanamycin sulfate (Sigma). 14. Paraformaldehyde (Sigma). 15. Triton X-100 detergent (Sigma). 16. Bovine serum albumin (Sigma). 17. Sytox green cell nucleus stain (Invitrogen). 18. Ethylene oxide sterilizer (Anderson Products). 19. T-junction (Cole-Parmer). 20. Static mixer (Cole-Parmer). 21. Polypropylene 96-well plates (Sigma). 22. Female Leur connectors (Cole-Parmer). 23. Gradient syringe pumps (New Era Pump Systems).
3. Methods 3.1. Preparing the 96-Well Plates
1. Sieve NaCl using #60 and #40 sieves to yield NaCl of size range 0.250–0.425 mm. 2. Use polypropylene 96-well plates of standard dimensions (6.5 mm well diameter). Polypropylene plates have better solvent resistance than standard polystyrene plates. 3. Put 0.13 g of the sieved NaCl into the top four rows of a 96-well plate (48 wells total). If there is not enough salt in the wells, then the polymer solution will cover the salt and
164
Simon et al.
form a polymer skin on top of the salt during freeze drying. The salt should be higher than the polymer solution so that the scaffolds have open pores on their top surface. 4. Level the surface of the salt in each well, so that uniform scaffolds are fabricated. A tool can be machined from plastic for this purpose and inserted in each well to level the salt (Fig. 1a, b). 3.2. Preparing Polymer Solutions
1. The protocol described herein yields a scaffold array where the composition of Sudan IV red dye is varied. This is for system demonstration only. The red dye can be omitted when fabricating scaffold arrays for cell screening experiments when scaffold formulations are being varied and tested. 2. Make two vials of polymer solution where 1 g of polymer is dissolved in 10 mL of dioxane in each vial (see Note 1). Label one vial RED and the other CLEAR. Depending on the density of the polymer used, this will yield a solution of approximately 9.4% (mass/volume). 3. To the RED vial, add 1 mL of Sudan IV solution [0.3% Sudan IV (mass/volume) in dioxane]. This will yield a RED polymer solution of approximately 8.5% (mass/volume) with 0.03% Sudan IV (mass/volume). 4. To the CLEAR vial, add 1 mL of dioxane as a control. This will yield a CLEAR polymer solution of approximately 8.5% (mass/volume).
3.3. Preparing the Two-Syringe Pump System
1. The parts shown in Fig. 1a must be assembled into the mixing apparatus shown in Fig. 1c. First, a section of Tygon tubing (2 cm long, 3.2 mm internal diameter) should be placed on each end of the static mixer. The plastic T-junction should be connected to one end of the mixer, and the end of a pipette tip (2.5 cm long) should be inserted into the tubing on the opposite end of the mixer. Tubing should be placed on the open ends of the T-junction, and then the female Luer connectors should be inserted into the tubing connected to the T-junction. The Luer connectors will provide connections for the syringes. 2. The static mixer is made of steel, is 135 mm long (Fig. 1a, c), and has an internal diameter g of 4.8 mm with a volume of 1.1 mL (Cole-Parmer). The mixer contains 18 mixing elements (helices). The mixing elements are helices that mix and fold the polymer solutions 218 times (218 = 262,144; the exponent of “18” comes from the 18 mixing elements) (Fig. 1d). 3. The tubing and syringes should be wetted and rinsed with dioxane. Syringes filled with dioxane can be attached to the Luer connectors and used to inject dioxane into the tubing.
3D Polymer Scaffold Arrays
165
Fig. 1. (a) Parts required for making scaffold arrays. (b) NaCl porogen in 96-well plates is leveled/flattened using a home-made plastic tool. (c) The two-syringe pump system for making scaffold arrays is shown. The CLEAR polymer solution is in the left pump in a 1-mL syringe. The RED polymer solution is in the right pump in a 2.5-mL syringe. Flow from the syringe pumps comes together at a T-junction in the middle of the image, goes into the steel static mixer, and comes out of the pipette tip at the bottom of the image. (d) A diagram of the interior of the static mixer. The helices in black are made of metal and are part of the mixer. The helices mix and fold the flowing polymer solution depicted by the red flow lines. (e) Polymer solution being deposited into a 96-well plate as it elutes from the static mixer. The 96-well plate contains NaCl porogen in each well. (f) After the polymer solution is deposited into the 96-well plate, the plate is frozen in liquid nitrogen for 5 min before being freeze-dried. (g) A completed PDLLA scaffold array after freeze-drying and salt leaching where Sudan IV red dye is varied is shown.
166
Simon et al.
Shake out the tubing and syringes so that they are wet but not holding excess dioxane. 4. Two syringe pumps should be set up to face each other, so that the syringes feed into the opposing ends of the T-junction as shown in Fig. 1c. The left syringe pump should hold the 1 mL of CLEAR polymer solution in a 1 mL syringe (4.5 mm internal diameter). The right syringe pump should hold 2.5 mL of RED polymer solution in a 2.5 mL syringe (7.25 mm internal diameter). 5. The syringe pumps should be placed on “lab jacks” so that the syringes, tubing, and T-junction can be leveled. A ring stand with a clamp can be used to support the end of the static mixer with the pipette tip as shown in Fig. 1c. There should be a slight decline (5°) in the static mixer as it goes from the T-junction toward the pipette tip to help prevent bubbles. However, if there is too much of a decline, then the polymer solution will drip out ahead of the back pressure, causing bubbles and inconsistencies in composition. 6. The pipette tip itself should also be angled downward at approximately 30°C. The down angle helps drops to form and fall cleanly into the 96-well plates. If there is not enough down angle, then the droplets can run back along the pipette tip and be difficult to get into the wells of the 96-well plate. If the pipette tip has too much down angle, then the polymer solution will drip out ahead of the back pressure causing bubbles and inconsistencies in composition. 7. The syringe pumps must be programmable and must be able to ramp up or down between two different velocities. The left pump holding the CLEAR polymer should be programmed to start at 0.5 mL/min and ramp down linearly to 0 mL/min over 72 s (0.3 mL). The right pump holding the RED polymer should be programmed to start at 0 mL/min and ramp up linearly to 0.5 mL/min over 72 s (0.3 mL) and then hold at 0.5 mL/min for 5 min. The RED polymer pump continues to pump after the CLEAR polymer pump has stopped in order to push the mixed polymers out of the mixer, through the pipette tip and into the 96-well plate. 3.4. Mixing and Deposition of the Scaffold Arrays
1. The CLEAR and RED polymer solutions should both be primed down the tubing up to the T-junction. This can be achieved by turning the syringe pumps on and off as appropriate. Next, an additional 325 mL of the CLEAR polymer should be pumped from the left pump (the solution will flow into the static mixer a little bit). There is cross-mixing during the fabrication process, and leading with a bolt of 325 mL of the CLEAR makes it possible to get the widest possible range of compositions in the library.
3D Polymer Scaffold Arrays
167
2. Make sure both syringe pumps are reset to the beginning of their respective pumping programs. Start both pumps simultaneously to begin mixing and deposition. 3. After a couple of minutes, polymer solution will begin to elute from the pipette tip. Catch two drops of polymer solution in each well of the 96-well plate for 36 wells as shown in Fig. 1e (see Notes 2 and 3). 4. Control scaffolds of pure CLEAR and pure RED can also be prepared if desired. For the CLEAR controls, CLEAR polymer solution is pumped at constant speed (0.5 mL/min) from a single syringe pump and deposited at two drops per well into the 96-well plate containing 0.13 g of sieved salt per well. RED controls can be prepared in the same manner but using RED polymer solution. Completed control scaffolds of pure CLEAR and pure RED polymers are visible in the completed scaffold array in Fig. 1g (fourth row from top of plate, six CLEAR on left and six RED on right). 5. The wet 96-well plate containing the polymer solutions should be centrifuged for 2 min with a lid at 2,000 rpm (210 rad/s) in a swinging bucket centrifuge, using 96-well plate holders. This gently forces the polymer solutions to the bottom of each well ensuring that the scaffolds will be of uniform shape. 6. Freeze the 96-well plate in liquid nitrogen as shown in Fig. 1f (wear safety glasses!). Place the 96-well plate in a shallow pan and very gently pour in liquid nitrogen until the 96-well plate is submerged. This must be done slowly to avoid disturbing the scaffolds before they are frozen. Use a spatula to hold the plate below the surface of the liquid nitrogen (the plate will float). Leave the plate in the liquid nitrogen for 5 min to insure a thorough freeze. When ready to place on the freezedryer, gently pour off the liquid nitrogen and immediately place in the freeze-dryer (see Note 4). 7. Freeze-dry scaffolds overnight at 13 Pa (100 mTorr) or less using a liquid nitrogen trap to catch dioxane. The trap should be bypassed at the end of the day (before the overnight), so that the dioxane will not return to vapor phase when the liquid nitrogen is gone (see Notes 5 and 6). 3.5. Salt Leaching and Drying Scaffold Arrays
1. After overnight freeze-drying, remove scaffolds from freezedryer and leach salt in water. Plates can be placed in shallow pans and held under water using weights. Rubber-coated lead rings designed for holding down bottles in a water bath work very well for this. Cover pan with foil to keep dust out during leaching.
168
Simon et al.
2. When submerging the plates in water for salt leaching, air bubbles may be present in the wells on top of the scaffolds, which can inhibit the salt leaching. A transfer pipette can be used to very gently force water into the wells to displace the bubbles. If this is done too forcefully, then the scaffolds can become dislodged from the plate. 3. Water should be changed daily for a total of 5 d of salt leaching (see Note 7). After 5 days, remove the 96-well plate from the water and gently place it upside down on clean paper laboratory towels to absorb excess water. The plates should be handled gently to prevent the scaffolds from becoming dislodged. 4. Air-dry scaffold arrays for 3 d while protected from dust (in a drawer, or wrapped lightly with a paper towel). Do not dry the scaffolds in an oven as this will distort them (they collapse even in a low temperature oven 37°C). A dried, completed scaffold array is shown in Fig. 1g (see Notes 8–10). 3.6. Characterizing Scaffold Arrays Using Sudan IV Dye Absorbance
1. Sudan IV absorbance can be used to track scaffold composition in the arrays in order to characterize array fabrication (see Note 11). For this analysis, sieved NaCl should be omitted from the 96-well plate. Two drops of polymer solution should be deposited directly into the wells of the 96-well plate as they elute from the pipette tip and static mixer. Add 0.1 mL of dioxane to each well, mix, and read absorbance of each well at 490 nm (Abs490) using a plate reader. Control CLEAR and RED wells can also be prepared and read for calibration. 2. The Sudan IV analysis is demonstrated in Fig. 2e using PDLLA scaffolds (16). The Abs490 values for six PDLLA 96-well scaffold arrays were averaged and plotted. Control CLEAR and RED wells were also prepared, read, averaged, and plotted. The plot shows that a linear (R = 0.99) and reproducible change in Sudan IV concentration is present in the libraries, which spanned 90% of the composition range from pure CLEAR to pure RED (7% RED–97% RED).
3.7. Scaffold Porosity and Microstructure
1. The 96-well array scaffolds are too small for accurate determination of porosity by gravimetric analysis (weighing scaffolds to determine mass and using calipers to determine volume). Thus, gravimetric measurements to determine scaffold porosity were made using larger scaffolds (PDLLA, Mw = 103,000; Mw = “relative molecular mass” for the purposes of this work) fabricated by the same process of freeze-drying and salt leaching (16). As described in Simon et al. (16), the total porosity of the scaffolds is 97% (n = 8, SD = 0.1%) calculated using the equation, “1–[(m/d)/v] = total porosity”, where m is mass of the scaffold (g), d is polymer density (g/mL), and v is volume
3D Polymer Scaffold Arrays
169
Fig. 2. (a–c) SEM images of the interior pores of a PCL scaffold made by the scaffold array process. Panel (a) is lower magnification and shows several pores, (b) shows a single pore formed by a NaCl crystal at the very bottom of a scaffold in the 96-well plate (scaffold bottom is on right) and (c) shows the small pores in the scaffold walls and struts that form during the freeze-drying process as the dioxane sublimes. (d) Stereomicroscope image of a PDLLA scaffold made by the scaffold array process. (e) The change in composition of the 96-well scaffold libraries was determined by tracking Sudan IV red dye. Six 96-well libraries were fabricated and Sudan IV was determined by measuring Abs490. Abs490 for six 96-well libraries was averaged and plotted against well number. Control clear (black square) and red (black triangle) scaffolds were also analyzed. Error bars are SDs, and the line is linear regression (R = 0.99). Ninety percent of the range from 0 to 100% is covered (7–97% red). (f–h) The nuclei of MC3T3-E1 osteoblasts cultured on pure PCL scaffolds made by the library approach were stained to fluoresce green using Sytox green. Cells were cultured for 1 d (f), 7 d (g), or 14 d (h).
of scaffold (mL). Of this total porosity of 97%, the theoretical porosity from NaCl leaching was 83% calculated using the equation, “(m/d)/v = NaCl Porosity”, where m is mass of NaCl used (g), d is NaCl density (2.165 g/mL), and v is scaffold volume (mL). The remaining 14% porosity (97– 83% = 14%) is due to voids in the scaffold walls that form as the dioxane sublimes during freeze-drying. 2. Large pores (0.2–0.4 mm) formed by salt leaching are shown in Fig. 2a, while small pores (< 10 mm) formed by dioxane sublimation are shown in Fig. 2c. In Fig. 2b, a large, cubic pore formed by salt leaching is visible in the middle of the
170
Simon et al.
panel, while smaller “dioxane-sublimation” pores are visible in the walls of this “salt-leach” pore. 3. Scaffold porosity is essentially independent of the polymers used because the scaffolds are so highly porous and because the densities of different polymers are similar [PDLLA = 1.3 g/mL (Absorbable Polymers International), PCL = 1.1 g/mL (Sigma–Aldrich), pDTEc = 1.2 g/mL, pDTOc = 1.2 g/mL] (17). Although the density of different polymers can vary slightly, the polymer accounts for only 3% of the scaffold volume. Thus, a 10% change in polymer density will result in only a 0.3% change in the total porosity of the scaffolds. 3.8. Cell Culture on Scaffold Arrays
1. Sterilize scaffold arrays using an ethylene oxide sterilizer (Anprolene AN74i, Anderson Products) and degas under vacuum in a desiccator for 2 d. 2. When ready to seed cells, add 0.2 mL of medium per well and place under vacuum for 2 min (bubbles may form). Briefly release the vacuum and reapply for 2 min. This procedure removes air from the interior scaffold pores filling them with medium. 3. Scaffolds can be seeded with up to 100,000 cells per well (in 0.2 mL of medium) and successfully cultured for 2 weeks with two medium changes per week without the medium becoming acidic (turning yellow from cell waste accumulation) (unpublished observations). 4. Standard cell assays can be performed to assess cell adhesion, morphology, and proliferation in scaffold arrays, such as fixing and staining for microscopy or soluble colorimetric assays for cell viability and counting (16, 17). 5. MC3T3-E1 osteoblasts cells cultured for 1 d, 7 d, or 14 d on control PCL scaffolds made by the library approach are shown in Fig. 2f–h (see Note 12). Ten thousand cells were seeded on each scaffold and the nuclei were stained to fluoresce green. Note that the number of cells increases with increasing incubation time, indicating that cells can adhere and proliferate on scaffolds made by the library approach. Also note that the pores of the scaffold are visible in Fig. 2h, as the cells have become confluent.
4. Notes 1. Safety glasses should be worn when fabricating scaffolds arrays. There is potential for splashing of solvent and liquid nitrogen. 2. The average drop size during deposition can be determined by counting the number of drops that fall in a given time (1 min) at a known flow rate (0.5 mL/min). When this was
3D Polymer Scaffold Arrays
171
done on five different days using PDLLA, the average drop size was 15.2 mL (SD = 2.0 mL, n = 5) [16]. Similar drop sizes were observed for pDTEc, pDTOc, and PCL. 3. The dead volume of the tubing system including the T-junction, static mixer, pipette tip, and tubing connectors is ≈ 2 mL. Approximately, 0.85 mL of CLEAR polymer (0.07 g) and 2 mL of RED polymer (0.17 g) are used to make each scaffold array, although only ≈ 1 mL is actually deposited into the 96-well plates to yield the scaffold arrays. More RED than CLEAR is required since additional RED is required to push the gradient through the mixer and into the 96-well plate. The gradient takes ≈ 4 min to pump: 72 s of mixing and ≈ 3 min of pumping for the gradient to get pushed out of the mixer and into the salt trough. 4. When preparing several scaffold arrays at once, they can be temporarily frozen in liquid nitrogen and stored in a −80°C freezer for up to 2 h so that all the arrays can be loaded onto the freeze-dryer at one time. Make sure to reimmerse the arrays for 5 min in liquid nitrogen to refreeze them before putting them on the freeze-dryer. 5. The 96-well plates are bulky and may not fit in the glass bell jars that typically come with freeze-dryers. In this case, the 96-well plates can be placed in the manifold of the freeze-dryer. However, a trap may have to be installed between the freezedryer manifold and the pump to catch the dioxane. Alternatively, a secondary chamber with a large opening for loading in the 96-well plates could be attached to the freeze-dryer. A vacuum oven, Parr Bomb (steel reaction vessel), or vacuum desiccator may work (make sure they can withstand the low pressure or they could implode!). A trap can be placed between the secondary chamber and the freeze-dryer to catch dioxane. 6. Dioxane sublimes quickly enough during the freeze-drying process, so the scaffolds stay frozen. If the polymers being used to make the scaffold arrays are not soluble in dioxane and a different solvent must be used, then it may be necessary to keep the scaffolds cold during freeze-drying. This can be achieved by packing the plates in dry ice. [1,4-Dioxane physical values from manufacturer (Sigma–Aldrich): density = 1.033 g/mL; boiling point = 101°C; melting point = 11°C; Mw = 88.1 g/mol]. 7. Five days of salt leaching is sufficient to remove all the salt from all of the scaffolds. When scaffolds were leached for only 3 d, residual NaCl was occasionally found in some of the scaffolds. 8. If two drops of polymer solution are added to each well of the 96-well plate at an average drop size of 15.2 mL, then 30.4 mL of polymer solution will end up in each well. This
172
Simon et al.
equals 2.6 mg of PDLLA using 8.5% (mass/volume) PDLLA solutions. Thus, the mass of each scaffold in a completed array is ≈ 2.6 mg and each scaffold contains ≈ 10 mg of Sudan IV (0.4% by mass). Finished scaffolds are cylinders 6.5 mm in diameter and 2 mm high. 9. The following polymers have been used successfully to make scaffold arrays as described herein: (a) Poly(d,l-lactic acid): PDLLA ●
●
●
Mw 100,000 g/mol, Lactel-Birmingham Polymers (unpublished results of the authors) Mw 103,000 g/mol Absorbable Polymers International [16] Mw 109,000 g/mol, Medisorb-Alkermes (unpublished results of the authors)
(b) Poly(desaminotyrosyl-tyrosine ethyl ester carbonate): pDTEc, Mw 183,000 g/mol [17] (c) Poly(desaminotyrosyl-tyrosine octyl ester carbonate): pDTOc; Mw 122,800 g/mol [17] (d) Poly(e-caprolactone): PCL, Mw 80,000 g/mol, Sigma– Aldrich (Figs. 1c, e–g and 2a–c) 10. Scaffolds can be gently removed from the 96-well plates with a needle for analysis. Gently slide the needle down the side of the scaffold against the wall of the well and pry the scaffold. Figure 2d shows a scaffold that has been removed from the plate and imaged by stereomicroscopy. 11. When making scaffold libraries where the scaffold composition is actually varied (not just red dye), FTIR can be used to characterize the composition. Control samples of known composition can be made to establish an FTIR calibration curve for assessing the scaffold arrays [9, 17–20]. 12. The MC3T3-E1 murine osteoblast cell line (Riken Cell Bank, Hirosaka, Japan) was used as a model for osteoblasts [21] and cultured as described [9]. Low passage cultures (>6) at 80% confluency were passaged with trypsin for experiments. Medium was a-modification of Eagle’s minimum essential medium (Cambrex Bio Science) supplemented with 10% volume fraction fetal bovine serum (Gibco) and 0.060 mg/mL kanamycin sulfate (Sigma). For staining, cells were fixed (4% volume fraction paraformaldehyde), permeabilized (0.5% volume fraction Triton X-100), and blocked (1% mass fraction bovine serum albumin). Cells were fluorescently stained in phosphate buffered saline (PBS) containing 5 mmol/L Sytox green (Invitrogen). Scaffolds were removed from 96-well plates and observed on slides while moist with PBS. Cells were imaged by epifluorescence microscopy at 100× magnification (10× objective).
3D Polymer Scaffold Arrays
173
Acknowledgments We thank Sheng Lin-Gibson (NIST) and Nancy Lin (NIST) for the insightful discussions and Ed Parry (ADA-NIST) for machining. M.R. and K.C. acknowledge support from the National Academies of Science Research Associateship. S.M.D. acknowledges support from the NIST-NSF Summer Undergraduate Research Fellowship (SURF). The “standard deviation” (SD) is the same as the “combined standard uncertainty of the mean” for the purposes of this work. This work was supported by NIST and NIH/NIBIB R21 EB006497-01. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH, NIBIB, or NIST. This article, a contribution of the National Institute of Standards and Technology, is not subject to US copyright. Certain equipment and instruments or materials are identified in the paper to adequately specify the experimental details. Such identification does not imply recommendation by NIST; neither does it imply that the materials are necessarily the best available for the purpose. References 1. Dooley, C. T., Chung, N. N., Wilkes, B. C., Schiller, P. W., Bidlack, J. M., Pasternak, G. W., Houghten, R. A. (1994) An all d-aminoacid opioid peptide with central analgesic activity from a combinatorial library. Science 266, 2019–2022. 2. Rohrer, S. P., Birzini, E. T., Mosley, R. T., Berk, S. C., Hutchins, S. M., Shen, D.-M., Xiong, Y., Hayes, E. C., Parmar, R. M., Foor, F., Mitra, S. W., Degrado, S. J., Shu, M., Klopp, J. M., Cai, S.-J., Blake, A., Chan, W. W. S., Pasternak, A., Yang, L., Patchett, A. A., Smith, R. G., Chapman, K. T., Schaeffer, J. M. (1998) Rapid Identification of subtype-selective agonists of the somatostatin receptor through combinatorial chemistry. Science 282, 737–740. 3. Meredith, J. C., Sormana, J.-L., Keselowsky, B. G., Garcia, A. J., Tona, A., Karim, A., Amis, E. J. (2003) Combinatorial characterization of cell interactions with polymer surfaces. J. Biomed. Mater. Res. 66A, 483–490. 4. Anderson, D. G., Levenberg, S., Langer, R. (2004) Nanoliter-scale synthesis of arrayed biomaterials and application to human embryonic stem cells. Nat. Biotechnol. 22, 863–866. 5. Disney, M. D., Seeberger, P. H. (2004) The use of carbohydrate microarrays to study carbohydrate–cell interactions and to detect pathogens. Chem. Biol. 11, 1701–1707.
6. Smith, J. R., Seyda, A., Weber, N., Knight, D., Abramson, S., Kohn, J. (2004) Integration of combinatorial synthesis, rapid screening, and computational modeling in biomaterials development. Macromol. Rapid Commun. 25, 127–140. 7. Anderson, D. G., Putnam, D., Lavik, E. B., Mahmood, T. A., Langer, R. (2005) Biomaterial microarrays: rapid microscale screening of polymer–cell interaction. Biomaterials 26, 4892–4897. 8. Flaim, C. J., Chien, S., Bhatia, S. N. (2005) An extracellular matrix microarray for probing cellular differentiation. Nat. Methods 2, 119–125. 9. Simon, Jr., C. G., Eidelman, N., Deng, Y., Kennedy, S. B., Sehgal, A., Khatri, C. A., Washburn, N. R. (2005) Combinatorial screening of cell proliferation on poly(l-lactic acid)/poly(d,l-lactic acid) blends. Biomaterials 26, 6906–6915. 10. Gallant, N. D., Lavery, K. A., Amis, E. J., Becker, M. L. (2007) Universal gradient substrates for “click” biofunctionalization. Adv Mater. 19, 965–969. 11. Simon, Jr., C. G., Yang, Y., Thomas, V., Dorsey, S. M. and Morgan, A. W. (2008) Cell interactions with biomaterials gradients and arrays. Comb. Chem. High Throughput Screen, in press.
174
Simon et al.
12. Shea, L. D., Wang, D., Franceschi, R. T., Mooney, D. J. (2000) Engineered bone development from a pre-osteoblast cell line on three-dimensional scaffolds. Tissue Eng. 6, 605–617. 13. Benya, P. D., Shaffer, J. D. (1982) Dedifferentiated chondrocytes reexpress the differentiated collagen phenotype when cultured in agarose gels. Cell 30, 215–224. 14. Hall, H. G., Farson, D. A., Bissell, M. J. (1982) Lumen formation by epithelial cell lines response to collagen overlay: a morphogenetic model in culture. Proc. Natl Acad. Sci. USA 79, 4672–4676. 15. Abbott, A. (2003) Biology’s new dimension. Nature 424, 870–872. 16. Simon, Jr., C. G., Stephens, J. S., Dorsey, S. M., Becker, M. L. (2007) Fabrication of combinatorial polymer scaffold libraries. Rev. Sci. Instrum. 78, 0722071–0722077. 17. Yang, Y., Becker, M. L., Bolikal, D., Kohn, J., Zeiger, D. N., Simon, Jr., C. G. (2008) Combinatorial polymer scaffold libraries for
18.
19.
20.
21.
screening cell–biomaterial interactions in 3D. Adv. Mater. 20, 2037–2043. Eidelman, N., Simon, Jr., C. G. (2004) Characterization of combinatorial polymer blend composition gradients by FTIR microspectroscopy. J. Res. Natl. Inst. Stand. Technol. 109, 219–231. Simon, Jr., C. G., Deng, Y., Eidelman, N., Washburn, N. R. (2004) High-throughput method for determining modulus of polymer blends. Macromol. Rapid Commun. 25, 2003–2007. Yang, Y., Dorsey, S. M., Becker, M. L., LinGibson, S., Schumacher, G. E., Flaim, G. M., Kohn, J., Simon, Jr. C. G. (2008) X-ray imaging optimization of 3D tissue engineering scaffolds via combinatorial fabrication methods. Biomaterials 29, 1901–1911. Sudo, H., Kodama, H.-A., Amagai, Y., Yamamoto, S., Kasai, S. (1983) In vitro differentiation and calcification in a new clonal osteogenic cell line derived from newborn mouse calvaria. J. Cell Biol. 96, 191–198.
Part II Methods for Microarray Generation
Chapter 10 PDMS Microfluidic Capillary Systems for Patterning Proteins on Surfaces and Performing Miniaturized Immunoassays Mateu Pla-Roca and David Juncker Abstract In this chapter, we describe the fabrication and use of microfluidic capillary systems (CSs) made in soft, transparent polydimethylsiloxane (PDMS). Sixteen microfluidic CSs, each containing a loading pad, a microchannel, and a capillary pump are engraved in a single PDMS chip. The CSs are used for two applications, firstly to pattern fibronectin on glass surfaces to locally control the adhesion of cultured cells to the substrate, and secondly to carry out multiplexed miniaturized immunoassays. Key words: Microfluidics, Miniaturizated immunoassays, Micromosaic immunoassays, Protein patterning
1. Introduction Microfluidic systems can transport minute amounts on solutions and replace macroscopic tubes, pipettes, vessels, and dishes in a myriad of applications. They help save reagents by reducing sample size from milliliters to microliters, decrease the time to result by enhancing mass transport, and allow many reactions to be carried out simultaneously on a small chip. These characteristics have driven the adoption of microfluidic systems in many different fields and notably in chemistry, biology, and medicine. 1.1. Microfluidic Capillary Systems
A microfluidic capillary system (CS) is made of three elements: a loading pad, a microchannel, and a capillary pump, each with a precisely designed geometry. The three elements are structured into a flat substrate, and the recessed (inner) surface is made
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_10, © Springer Science+Business Media, LLC 2011
177
178
Pla-Roca and Juncker
hydrophilic. A liquid delivered to the loading pad of a CS therefore spontaneously fills the microchannel and capillary pump by capillary forces alone. The direction and the flow rate are defined by the capillary forces and the flow resistance, both of which depend on the geometry of the three elements of the CS, and which can therefore be controlled (and preprogrammed) by modifying the geometry. “Capillary systems” and “microfluidic networks” have both been used to describe arrays of microchannels that are filled by capillary force and used for immunoassays and surface patterning. In this chapter, we use the word CS to reflect the fact that the microchannels have been designed (via the use of capillary effects) to perform specific microfluidic operations for a particular set of applications. 1.2. Patterning Proteins and Performing Immunoassays
Owing to the reversible sealing properties of the polydimethylsiloxane (PDMS), capillary systems made of PDMS (PDMS-CSs) are used in the present protocol. A section of the open microchannels of the PDMS-CSs can be covered reversibly by placing a flat substrate onto the engraved PDMS surface, and form a temporally closed microchannel for patterning proteins or carrying out immunoassays on the surface of a substrate (Fig. 1). Solutions are loaded into the loading pads with a conventional pipette and automatically flow through the microchannels by capillary forces (Fig. 1a, b). The reagents in the solution adsorb or react with the surface while faithfully replicating the pattern defined by the microchannels. If the solution is a physiological buffer containing proteins (e.g., capture antibodies or cell adhesion proteins), the proteins will adsorb to the substrate and form a pattern of parallel lines. After rinsing the channels with buffer and double-distilled water, the substrate is peeled off from the PDMS-CSs (Fig. 1c). The nonpatterned areas are “back-filled” by incubating the substrate with a drop of solution containing a “blocking agent,” such as the bovine serum albumin protein, which adsorbs to the surface and prevents further binding of other proteins (Fig. 1d).
Fig. 1. Using PDMS-CSs for patterning surfaces with proteins and for performing miniaturized immunoassays. (a) A substrate is sealed over the open microchannels of the CSs. (b) Antibodies or proteins are patterned on the substrate by flowing solutions through the microchannels. Three different types of miniaturized immunoassays (“drop,” “microchannel,” and “micromosaic”) can be performed following the three different protocols (a–c). (a) After peeling off the substrate from the PDMS-CSs, the substrate patterned with lines of capture antibodies is incubated with a drop of sample solution and will bind target analytes in the sample. (b) The substrate is kept in place, and each channel is sequentially flushed with minute amounts of different sample solutions, thus allowing the detection of one analyte per channel. (c) A micromosaic immunoassay is performed by removing the substrate, rotating it by 90°, and flowing different samples in each “column” across all “lines” of capture antibodies. When a target analyte corresponding to an immobilized antibody is present in the sample, it will be captured at the intersection of the corresponding line and column. Such a micromosaic immunoassay allows detecting multiple analytes in multiple samples simultaneously.
PDMS Microfluidic Capillary Systems for Patterning Proteins
179
180
Pla-Roca and Juncker
The patterned surface can be used for cell cultures (Fig. 1e) if cell adhesion proteins were patterned (1, 2) or for performing immunoassays if antibodies were patterned. A “drop” immunoassay (Fig. 1a) is performed by incubating the patterned substrate surface with a drop of a sample solution (Fig. 1f). Each line of immobilized capture antibody specifically binds its target analytes in the sample in a concentration-dependent manner. Fluorescent detection antibodies are then delivered over the substrate and bind only to the line where the target analytes have been captured and immobilized (Fig. 1g). The result of the assay is visualized with a fluorescence microarray scanner or a fluorescence microscope (Fig. 1h). In one experiment, it is thus possible to detect multiple analytes contained in a small sample (3). Alternatively, it is possible to perform all the steps of an immunoassay without removing the substrate from the engraved PDMS (Fig. 1b) by sequentially delivering and flowing solutions of blocking agent, samples, and labeled detection antibodies through the CSs (Fig. 1i–m). We call this approach protocol a “microchannel” immunoassay. Each microchannel corresponds to a single analyte only, and repeated delivery of the sample to multiple channels is necessary to analyze multiple analytes. This protocol further reduces sample consumption and incubation time because a few hundred nanoliters can suffice for tens of minutes of continuous flow and sample replenishment within the microchannels, and thus enable high-sensitivity assays (4, 5). It is also possible to measure multiple analytes in multiple samples simultaneously by first patterning m) capture antibodies, then removing the substrate, rotating it by 90°, and sealing it against a second microengraved PDMS-CSs (Fig. 1c). The second PDMS-CS is used to flow the blocking agent, (n) different samples and detection antibodies across the capture lines (Fig. 1o–q). In this scenario, each intersection represents a unique combination, and m × n different binding reactions can be performed at once on a single chip. After peeling off the substrate from the CSs (Fig. 1r), the binding is analyzed by fluorescence imaging and each binding event forms a bright square at the intersection (Fig. 1s); because the overall pattern appears as a “mosaic” of squares (Fig. 2), these assays have been dubbed “micromosaic” immunoassays. Such assays have been used to quantify cross-reactivity between antibodies (6), measure DNA hybridization rates (7), detect protein biomarkers (8) (Fig. 2d), viruses, and bacteria analytes (9), and perform competitive immunoassays (10). In the following protocol, we provide some basic design guidelines on how to make and use PDMS-CSs and a step by step description on how to pattern proteins and to carry out the miniaturized immunoassays presented in Fig. 1.
PDMS Microfluidic Capillary Systems for Patterning Proteins
181
Fig. 2. Micromosaic immunoassays performed on a PDMS substrate. (a) Fluorescence image of a capture antibody (against goat immunoglobulin) labeled with a green fluorescent dye that was patterned and immobilized as lines. (b) Fluorescence image showing the bound analyte (goat immunoglobulin labeled with a red fluorescent dye), bound at the intersections and forming squares. (c) View of the 16 × 16 intersections of the micromosaic assay. The lanes 8 and 9 were flushed with buffer only and serve as a negative control. The full assay protocol is described in this chapter. (d) Simultaneous detection of the cardiac biomarkers myoglobin (Mb), cardiac Troponin I (cTnI), S100a and C-reactive protein (CRP), and B-type natriuretic peptide (BMP) in human plasma. Four capture antibodies (aMb, acTnI, aS100a, and a CRP) were patterned as horizontal “lines” on a PDMS substrate and used to capture biomarkers spiked into plasma samples (from healthy subjects) delivered in the vertical “columns.” The captured biomarkers were detected by filling the vertical “columns” with a cocktail of fluorescently labeled detection antibodies (reproduced from 8 with permission of Elsevier Science).
1.3. Designing PDMS-CSs
CSs can be made in silicon (Si), glass, or PDMS. PDMS has spontaneous and reversible sealing properties; therefore,PDMSCSs are advantageous as they can be sealed to any type of smooth surface. This property allows using glass and gold substrates, coated with silanes and thiols, respectively, or polymer substrates made of polystyrene, polycarbonate, or poly (methylmethacrylate). The glass surfaces used in this protocol are coated with a selfassembled monolayer of a silane with epoxy groups prior to use (11). The epoxy groups on the surface ensure a good attachment of the proteins via a covalent reaction between the epoxy group and their amino groups. PDMS-CSs are made by replica molding (12) of a master that is typically made of a Si wafer, structured with an inverse pattern of the microchannels. The PDMS-CSs presented here comprise an array of 16 independent CSs, each containing a loading pad, an open microchannel, and a capillary pump (Fig. 3a–c). Once the loading pad is filled with a solution, the negative capillary pressure at the liquid–air interface at the filling front produces a continuous flow, without the need for peripheral pumps, or controllers, or connections (Fig. 3d). The key to spontaneous filling is to “activate” the inside surface of the PDMS microchannels and render it hydrophilic, which can be done by exposing the chip to an oxygen plasma or to ozone.
182
Pla-Roca and Juncker
Fig. 3. (a) PDMS-CSs chip with loading pads, microchannels, and capillary pumps. Each channel has been loaded with a different colored solution. (b, c) Detailed view of a loading pad and capillary pumps. (d) Schematic showing how reagent solutions are delivered to the loading pads by means of a micropipette. Capillary pumps continuously flush sample through the reaction area, where the adsorption or binding events are occurring. (e) Multiple solutions and large volumes can be flushed and drained sequentially using a flow promoter (i.e., a clean room paper or a tissue) in contact with the capillary pumps.
The flow rate (Q) of a CS can be calculated and is proportional to the capillary pressure (Pc) formed at the liquid–air interface of the filling front, divided by overall flow rate resistance (R) of the CS, Q ~ Pc/R. Both Pc and R are defined by the cross section of the conduit as follows: for conduits with a square cross section, the Pc scales with the width w of the conduit as w−1 and the resistance R scales as w4 (5). Thus, by reducing the cross section w by a factor of 2, for example, Pc doubles, R increases by a factor 16,
PDMS Microfluidic Capillary Systems for Patterning Proteins
183
and Q decreases by a factor of 8. One important parameter that needs to be considered when designing a CS is that as the channel is being filled over an increasing length l, the flow resistance R also continuously increases proportionally to the filled length, R ~ l for a channel with a constant cross section. In addition, the resistance of all upstream sections needs to be added as well. The characteristic dimension of the microstructures in the capillary pumps is made smaller than the one in the loading pad to create a stronger capillary pressure and generate unidirectional flow toward the pump. The arborescent architecture of the capillary pumps provides a high capillary pressure and a large volume while not significantly adding to the overall flow resistance when being filled because all channels run parallel to one another. The capillary pump also serves as a waste or as a sample concentrator. The total volume of each capillary pump is only a few hundred nanoliters, but larger volumes can be loaded in each CS and flowed through each assay area if a tissue (or any kind of wicking material) or evaporation is used to drain liquid in one or several capillary pumps (Fig. 3e). A capillary retention valve (CRV) can be used as a control element for entirely self-regulated and autonomous CSs. CRVs prevent undesired drainage of the assay area and trapping of bubbles, and are particularly useful when multiple solutions are added sequentially to many channels, but are beyond the scope of this chapter. Interested readers can find additional information on the design and use of CRVs in 4, 5. To design a CS, the following parameters need to be taken into account: 1. CS chip. The larger the number of CSs, the more difficult it becomes to load them rapidly without error. The chip shown (Fig. 3a) includes 16 CSs each 30 mm long. 2. Loading pads. The size of the loading pads and the spacing between them have to be sufficiently large in order to allow manual delivery of solutions with a micropipette. In the presented design, the loading pad area is approximately 0.7 mm2. 3. Microchannels. The length (l), width (w), depth (d), and spacing (s) are critical parameters to be considered. The assay area is 8 mm long, and each microchannel is 150 mm wide, 30 mm deep, and the channels are arrayed with a pitch of 300 mm. CSs have a high flow rate. For high-sensitivity immunoassays, we recommend 10–30 mm deep channels and aspect ratios between 1:2 and 2:1 (5). The length of the microchannels on the assay area of the CS array should be sufficiently long to allow an easy rotation of the substrate when a “micromosaic” immunoassay is performed. 4. Capillary pumps. A trade-off has to be made between a small pump with a limited capacity and a large pump with a big footprint, which will limit the number of CSs that can be arrayed on a chip.
184
Pla-Roca and Juncker
2. Materials 2.1. General Material and Reagents
1. Polystyrene Petri dishes (150 mm × 15 mm 60 mm × 15 mm, untreated, Corning, MA).
and
2. Vacuum desiccator and a vacuum line (1 mm Hg). 3. Plasma chamber (Plasma Line 415, Tegal Corporation, CA) connected to air or oxygen. Alternatively, an ozone production system can be used. 4. Leveled oven with thermometer. 5. Nitrogen blowgun or duster (ChemTronics, Ultra Jet, GA). 6. Tweezers (SPI, 2A). Round flat tip. 7. Microscope slides (Fisher, 25 cm × 75 cm, precleaned). 8. Powder-free gloves (Fisher). 9. Tape (Scotch Ruban Magic tape). 10. Clean room paper (Perotex 100, Perotech Sciences inc., Montreal, Canada). 11. Absorbent paper (Kimwipe, SPI supplies, PA). 12. Nitrogen blowgun or duster (ChemTronics, Ultra Jet, GA). 13. Micropipette (0.5–2 and 100–1,000 mL). 14. Scalpel or cutter. 15. Double-distilled water. 16. 96% ethanol and 75% ethanol. 2.2. Fabrication of the Mold
The microfluidic PDMS-CSs are made by PDMS replica molding (12) of either SU-8 2015 (MicroChem, Newton, MA) photoresist patterned on a silicon wafer (Fig. 4) or a structured SOI wafer. Mold fabrication is a standard procedure that can be performed in most microfabrication facilities. Molds can be obtained from large universities through their respective microfabrication services (13).
Fig. 4. (a) 4″ silicon wafer patterned with CSs made of SU-8 that serve as a mold for PDMS replica molding. (b) Detail of an array of CS with the capillary pumps, loading pads, and microchannels.
PDMS Microfluidic Capillary Systems for Patterning Proteins
185
1. Photomask with the design (Fineline Imaging, Colorado Springs, CO) (see Note 1). Patterns have to be clear on the photomask because SU-8 is a negative resist. 2. SU-8 2015 (Microchem, Newton, MA). 3. HMDS, Hexamethyldisilazane (reagent grade 99%, Sigma– Aldrich). SU-8 adhesion promoter. 4. Spin coater. 5. Photolithographic aligner or collimated UV light source. 6. SU-8 developer (Microchem, Newton, MA). 7. Silicon substrates (4″) (Monto Silicon Technologies). 8. TPFS, Trichloro-1H,1H,2H,2H-perfluorooctylsilane (97%, Sigma-Aldrich). 9. Digitally controlled hot plate (VWR, 810-HPS). 10. Glass Petri dish (5″). 2.3. Molding of PDMS-CSs and Flat PDMS Substrate
1. PDMS Sylgard 184 prepolymer (Dow Corning, Midland, MI).
2.4. Coating of Cover Slips with an Epoxy Silane
1. Round or rectangular cover slips (Fisher brand). 2. Toluene (97%, Fisher Scientific, ACS reagent). 3. Epoxy silane, 3-Glycidoxypropyldimethoxymethylsilane (3-GPS, 97%, Sigma-Aldrich). 4. Cover slip staining racks (two units) and staining dishes (two units) (Fisher Scientific).
2.5. Reagents, Proteins, and Antibody Solutions
1. Phosphate buffered saline (PBS) (Fisher Bioreagents, 1X, pH 7.4) and double-distilled water filtered with syringe filters (FisherBrand, 0.22 mm, 13 mm syringe PVDF sterile). 2. Carbonate–bicarbonate buffer, pH 9.6, prepared following manufacturer’s instructions (Carbonate–bicarbonate buffer capsule, Sigma-Aldrich). Filter the buffer with a syringe filter after preparation. 3. Capture antibody solution: Chicken antigoat immunoglobulin labeled with Alexa Fluor 488 dye (AF488, green fluorescence) from Invitrogen diluted at 100 mg/mL with PBS. Store at 2–8°C and protect from light. Bring to room temperature before using it. 4. Blocking solution: 1% (w/v) bovine serum albumin (BSA) (Sigma-Aldrich) in filtered PBS (Fisher Bioreagents, 1X, pH 7.4). 5. Assay buffer: 0.1% (w/v) BSA (Sigma-Aldrich) and 0.1% (w/v) Tween-20 (Sigma-Aldrich) in PBS.
186
Pla-Roca and Juncker
6. Sample solution: Goat immunoglobulin labeled with Alexa Fluor 647 dye (AF647, red fluorescence) from Invitrogen is diluted at 0.1 mg/mL with assay buffer. Store at 2–8°C and protect from light. Bring to room temperature before using it. 7. Fibronectin solution: Sterile cell culture-tested fibronectin (Sigma-Aldrich) is diluted at 100 mg/mL with carbonate– bicarbonate buffer. No BSA should be added. Store at 2–8 C and bring to room temperature before using it. 2.6. Patterning of Fibronectin on Epoxy-Coated Cover Slips for Cell Culture Applications
1. PDMS microfluidic capillary systems (PDMS-CSs). 2. Epoxy silane-coated cover slips. 3. PBS, fibronectin solution, and blocking solution. The solutions should be used at room temperature. 4. Dulbecco’s Modified Eagle Media (DMEM) and fetal bovine serum (sterile filtered and cell culture tested) both from Sigma-Aldrich. 5. 4% (w/v) Paraformaldehyde (95% powder, Sigma-Aldrich) solution in PBS. 6. Toluene (97%, Fisher scientific, ACS reagent). 7. Myoblast cells suspension at 1 × 106 cells/mL (C2C12 cell line, American Type Culture Collection, VA). 8. Microscope with differential interference contrast (DIC) or phase contrast.
2.7. Miniaturized Immunoassay
1. PDMS-CSs. 2. Microscope slide (Fisher, 22 cm × 75 cm, precleaned). 3. Square PDMS substrate cut to size (~1 cm2). 4. Blocking, capture antibody, and sample solutions. The solutions should be used at room temperature. 5. Fluorescence scanner with 488 and 633 nm laser excitations and FITC and Cy5 dyes emission filters or fluorescence microscope with the corresponding filters.
3. Methods 3.1. Fabrication of PDMS-CSs and Substrates 3.1.1. Fabrication of PDMS-CSs
1. Activate the surface of the CSs mold in a plasma chamber for 1 min (see Note 2). Place the activated mold immediately in a desiccator, together with a few drops of TPFS on a cover slip. Connect to vacuum for 1–5 min, close the desiccator valve, and keep the mold for 1 h in the silane atmosphere. The TPFS coating will prevent PDMS from sticking to the mold during replica molding.
PDMS Microfluidic Capillary Systems for Patterning Proteins
187
Fig. 5. Tape the edges of the 4″ silicon CSs mold on a plastic Petri dish before pouring PDMS prepolymer.
2. Transfer the coated mold into a glass Petri dish and place it in an oven at 110°C for 10 min. This will ensure the complete reaction of the silane with the surface. 3. Place the CSs mold on the bottom of a polystyrene Petri dish (150 mm × 15 mm) and use tape to fix it to the Petri dish and cover the edges (Fig. 5). Pour PDMS prepolymer (Sylgard 184) until it forms a 3-mm thick layer. 4. Degas by placing the dish in a vacuum chamber for 5–30 min depending on the vacuum level. Check with a microscope to ensure that the structures are bubble- free. If not, place under vacuum for an additional 5–10 min. 5. Fully cure the PDMS in the oven at 60°C for 12 h (see Note 3). 6. Use a scalpel to cut the cured PDMS along the edge of the silicon wafer. Use flat tweezers to peel off the replica from the mold. If handled carefully, the mold can be reused many times. The TPFS needs to be applied once every few months only. Inspect the mold under a microscope after each replication. If PDMS residues are observed, especially in the channels area, discard the mold. 7. Cut the PDMS-CSs from the PDMS replica and place tape on top of the engraved CSs to prevent dust contamination of the patterned surface. Keep the rest of the replicas face down in a Petri dish or in a dust-free environment for future use. 3.1.2. Fabrication of Flat PDMS Substrate
1. A flat PDMS substrate is made by curing PDMS on a polystyrene Petri dish (150 mm × 15 mm). Pour the PDMS mixture to reach a level of at least 1 mm. Degas for 5 min. Fully cure the
188
Pla-Roca and Juncker
PDMS in the oven at 60°C for 3–12 h and cut out a PDMS block with the desired size. The surface of PDMS cured in contact with the Petri dish is used as assay surface. 3.1.3. Coating of Cover Slips with Epoxy Silanes
1. Fill a staining dish with 99 mL of toluene and cover it with its lid. 2. Place cover slips on the staining rack and plasma treat for 1 min in order to activate the glass surface. 3. Add 1 mL of 3-Glycidoxypropyldimethoxymethylsilane (3-GPS) to the toluene and mix it well. Immediately immerse the activated cover slips for 20 min. 4. Rinse the cover slips with fresh toluene and blow dry with nitrogen (30 s). Place the cover slips on a new staining rack and cure in the oven for 30 min at 110°C. 5. Rinse with ethanol, blow dry, and keep in a dust-free environment for future use.
3.2. Patterning Fibronectin on Cover Slips for Cell Culture Applications
Fibronectin is a well-characterized multifunctional extracellular matrix protein that mediates cell adhesion, migration, and plays an important role in many cell–surface interactions and wound healing. Fibronectin-coated surfaces promote cell attachment. To maintain sterility, the following steps should be performed in a cell culture environment or in a biosafety cabinet. Tweezers should be sterilized under a flame or by immersion in a 75% ethanol solution before using them. 1. Use a scalpel to cut out several 2 cm × 5 cm pieces of clean room paper. 2. Use gloves in order to manipulate the PDMS-CSs. Rinse the gloves with ethanol before using them. Peel off the protecting tape from the PDMS-CSs. Rinse the microstructures with ethanol and dry them with nitrogen (1 min). 3. Place the cleaned PDMS-CSs on a glass slide, with the microstructures facing up. This will allow better handling of the PDMS-CSs during the following steps. Place the PDMS-CSs in the plasma chamber for 1 min in order to render the PDMS-CSs hydrophilic (see Note 4). Stamp the activated PDMS-CSs several times against a flat PDMS surface. This treatment will neutralize the reactive groups on the surface of the CS but will keep the inside of microchannels hydrophilic. This step is important to prevent solution delivered to the loading pads from spreading over the surface, and generally helps minimize leakage and ensure proper function of the CS (e.g., spontaneous filling). 4. Place an epoxy-coated cover slip over the hydrophilized PDMS-CSs. Make sure not to block the loading pads with the cover slip. With a glass scribe, mark a small R (for Rear) near the edge of the cover slip to mark the back side. Slightly
PDMS Microfluidic Capillary Systems for Patterning Proteins
189
press with the tweezers in order to ensure conformal contact with the PDMS-CSs. Sterilize the setup for a few minutes under UV light and place it in a polystyrene Petri dish (150 mm × 15 mm) along with a wet tissue or paper. This will prevent evaporation during incubation. 5. Load 2 mL of fibronectin solution in the channels, cover the Petri dish with its lid, and incubate for 30 min (see Note 5). Shorter incubation times will result in lower levels of protein attached on the surface. 6. Place the short side of the clean room paper piece (2 cm × 5 cm) over the capillary pumps until all channels are drained. Slightly humidify the paper edge prior to use in order to ensure a better contact with the capillary pumps. 7. Rinse with PBS (3 × 2 mL) and peel off the cover slip. Replace the clean room paper whenever it is saturated. 8. Peel off the PDMS-CSs and place the patterned cover slip with the fibronectin lines facing up in a small Petri dish. Use the previously inscribed small “R” as an orientation mark. Incubate for 2 h at 4°C in order to ensure fibronectin adhesion to the surface. Rinse the cover slip with PBS (2 × 2 mL) and add 2 mL of blocking solution. Ensure that the cover slip is perfectly immersed. Incubate for 1 h at room temperature. Remove the blocking solution and rinse several times with filtered PBS (see Note 6). 9. Transfer the cover slip in a new small Petri dish and add 2 mL of a myoblast cells suspension at a density of 1 × 106 cells/mL in DMEM media, with 20% (v/v) fetal bovine serum media. After a 24 h incubation (37°C, and 5% CO2 atmosphere), the attachment of the cells to the areas where the fibronectin was patterned can be visualized using a microscope (Fig. 6).
Fig. 6. Differential interference contrast (DIC) micrographs showing the preferential attachment of myoblast cells on fibronectin patterns after 24 h incubation (37°C and 5% CO2 atmosphere).
190
Pla-Roca and Juncker
For imaging purposes, the cells can be fixed by adding 2 mL of 4% (w/v) paraformaldehyde PBS solution to the Petri dish for 10 min at 37°C. Rinse the cover slips with PBS (3 × 2 mL) and double-distilled water (1 × 2 mL) . 3.3. Miniaturized Immunoassays
1. Use a scalpel to cut several (2 cm × 5 cm) pieces of clean room paper. 2. Use a scalpel to cut a 1 cm2 piece of PDMS substrate. Before peeling it off from the Petri dish, make a notch in one of the corners. This notch will help proper alignment of the substrate in later procedures (see Note 7). Using the tweezers, peel off the substrate, pulling from the edges in order to avoid tearing of the PDMS. Place the PDMS substrate in a Petri dish in order to protect it from dust. 3. Peel off the tape that protects the PDMS-CSs from dust, rinse the microstructures with ethanol, and dry them under a stream of nitrogen for several seconds. Use powder-free gloves to manipulate the PDMS-CSs. 4. Place the cleaned PDMS-CSs chip in a microscope glass slide, with the microstructures facing up. Place the PDMS-CSs in a plasma chamber for 1 min in order to hydrophilize the surface (see Note 4). Bring the hydrophilized PDMS-CSs several times into contact with a flat PDMS surface. This treatment will minimize leakages and facilitate the loading of solutions on the microchannels. 5. Place the PDMS substrate on top of the PDMS-CSs channels. Press slightly in order to ensure conformal contact. Make sure not to block the loading pads with the substrate. (Fig. 7a). Place the microscope slide with the immunoassay setup surrounded by a wet tissue or paper in a polystyrene Petri dish (150 mm × 15 mm). This will prevent evaporation during incubation procedures. 6. Load 2 mL of PBS buffer in microchannels 8 and 9 as negative controls. Load all other microchannels with 2 mL of capture antibody solutions (chicken antigoat immunoglobulin labeled with AF488 has been used in the present protocol for demonstration purposes) (Fig. 7a). Cover the Petri dish with its lid in order to prevent evaporation, and incubate antibodies for 1 min. 7. Place the short side of the clean room paper (2 cm × 5 cm) over the capillary pumps until all channels are drained (Fig. 7b). Do not let dry for an extended time (see Note 8). Any of the three miniaturized immunoassay variants can be performed at this stage (Fig. 1a–c, see Subheading 1). The protocol for the “drop” immunoassay (Fig. 1, protocol A) follows the same steps as for the “microchannel” (Fig. 1, protocol B)
PDMS Microfluidic Capillary Systems for Patterning Proteins
191
Fig. 7. (a) Loaded CSs with capture antibody solution. A PDMS substrate with a notch has been placed on the assay area. (b) Draining capture antibody solution by means of the clean room paper placed over the capillary pumps. (c) Microchannels loaded with sample solutions. The notch on the PDMS substrate is used as orientation in the course of the 90º substrate rotation. For visualization purposes, color dye has been added to the loaded solutions. Scale bar 2 mm.
and the “micromosaic” immunoassay (Fig. 1, protocol C) which are described below, but, with the exception that whenever filling of the microchannels is described, a drop of solution is to be applied to the patterned PDMS surface instead. This drop further needs to be spread over the entire surface of the substrate using a glass cover slip, and incubated for ten times longer than the time indicated for steps carried out in a microfluidic channel. 8. Remove the wet clean room paper and load the channels with 2 mL of BSA blocking solution. Incubate for 1 min. Unload the channels with a new piece of clean room paper and keep it in place. 9. Rinse the channels (3 × 2 mL) with PBS and (1 × 2 mL) with double-distilled water. Replace the clean room paper if saturated. 10. (C only) Peel off the PDMS substrate and blow it dry with the nitrogen blowgun (10 s). Use the tweezers for this process. Rotate the PDMS substrate by 90° and place it over a new PDMS-CS previously activated in the plasma chamber. Use the notch for orientation (Fig. 7c). 11. (C only) Load the channels with 2 mL of BSA blocking solution. Incubate for 1 min. Unload the channels with a new piece of clean room paper. 12. Load 2 mL of sample solution into each CS and incubate for 3 min (see Note 9). Drain the channels with clean room paper. The incubation time will be dependent on the binding capacity of the capture antibodies used and the analyte concentration to be detected. The longer the incubation, the higher the sensitivity. In the present protocol, a solution containing goat immunoglobulin AF 647 was used as the sample, and the corresponding result is shown in Fig. 2a. For a full micromosaic immunoassay as the one shown in Fig. 2b, steps 13 and 14 need to be carried
192
Pla-Roca and Juncker
out to deliver the labeled detection antibodies that will bind to the immobilized analyte. 13. Rinse the microchannels (3 × 2 mL) with PBS. Replace the clean room paper whenever it is saturated. 14. Load the microchannels with the corresponding fluorescently labeled detection antibodies in assay buffer and incubate them for 20 min. The recommended concentration is 1 mg/mL (see Note 10). 15. Rinse the microchannels (3 × 2 mL) with PBS and (1 × 2 mL) double-distilled water. Use a clean room paper in order to drain each solution. 16. Peel off the PDMS substrate and dry it under a stream of nitrogen (30 s). The immunoassay result is revealed by imaging with a fluorescence microscope or scanner (see Note 11). The result of the demonstration mosaic immunoassay is shown in Fig. 2a (see Subheading 1).
4. Notes 1. The design of the photomask can be made using Inkscape (open source software), CorelDraw (Corel Corporation) or Adobe Illustrator (Adobe systems), or more specialized software such as Clewin 4.0 (Vieweb, Netherlands) or Autodesk (AutoCAD, Autodesk). The photomask can be made of mylar foil or glass depending on the requirements of the photolithographic process (i.e., aligner). 2. Plasma will clean the mold surface and activate it for functionalization. Time and power vary depending on the plasma chamber used, and the processing time should be kept short to prevent heating and damaging of the SU-8 features on the mold. Alternatively, an ozone chamber (Ozomax Corporation, http://www.ozomax.com) can be used for activation, but the processing time will need to be increased. 3. Curing of PDMS can be accelerated by using a higher temperature up to 90°C. At 90°C, curing takes only 1 h, but shrinkage of PDMS becomes more pronounced and most of the Petri dishes melt at 70°C. 4. PDMS surfaces after plasma treatment remain activated (and hydrophilic) only a few minutes because uncured lowmolecular weight PDMS migrates from the bulk to the air– surface interface. 5. Proteins attach covalently to epoxy-coated slides due to the reaction between the protein lysine residues with the 3-GPS epoxy groups at basic pH.
PDMS Microfluidic Capillary Systems for Patterning Proteins
193
6. In the present protocol, BSA has been used as the blocking agent in order to reduce the attachment of the cells on the areas where fibronectin is not present. Alternatively, Pluronic® F127 (Ethylene Oxide/Propylene Oxide Block Copolymer, BASF) can also be used to reduce the attachment of the cells (14). 7. Use tape to remove residual PDMS debris. This is a really effective method to clean the PDMS surface. 8. More advanced CSs with a CRV can be designed in order to prevent the drying of the microchannels on the assay area (4). 9. Incubation time of the samples will depend on the concentration of the analytes to be detected and the affinity of the capture antibodies used. An incubation time of 12 min with continuous flow driven by evaporation was used in order to detect tumor necrosis factor a (TNF-a) )at a concentration of 20 pg/mL in cell culture media (15). 10. If the signal is too weak, the concentration can be varied, although problems due to nonspecific adsorptions of detection antibodies over the substrate may increase as well. 11. When using a fluorescence microarray scanner, we recommended sticking the substrate to a microscope slide or a cover slip, with the assay substrate surface stuck to the glass to facilitate focusing and manipulation of the sample. This necessitates a scanner that can scan through the glass (e.g., LS Reload Tecan Laser Scanner or Agilent Technologies DNA microarray Scanner).
Acknowledgments The authors would like to thank Emmanuel Delamarche and Ute Drechsler (IBM Research Center, Zurich) for the fabrication of the molds. We are also very grateful to Saravanan Sundararajan and Haig Djambazian (Genome Quebec, McGill University, Montreal) for providing the myoblast cells and the DIC microscopy imaging of the cells. M. P. acknowledges financial support of the Spanish Ministry of Science postdoctoral fellows. References 1. Lang, S., von Philipsborn A. C., Bernard, A., Bonhoeffer, F., Bastmeyer, M. (2008) Growth cone response to ephrin gradients produced by microfluidic networks. Anal. Bioanal. Chem. 390, 3, 809–816. 2. Jiang, X., Xu Q., Dertinger, S. K. W., Stroock A. D., Fu, T., Whitesides G. M. A. (2005) General method for patterning gradients of biomolecules on surfaces
using microfluidic networks. Anal. Chem. 77, 2338–2347. 3. Delamarche, E., Bernard, A. (1997) Patterned delivery of immunoglobulins to surfaces using microfluidic networks. Science 276, 779–781. 4. Juncker, D., Schmid H., Drechsler, U, Wolf, H., Wolf, M., Michel, B., Nico de Rooij, Delamarche, E. (2002) Autonomous microfluidic capillary system. Anal. Chem. 74, 6139–6144.
194
Pla-Roca and Juncker
5. Delamarche, E., Juncker, D., Schmid, H. (2005) Microfluidics for processing surfaces and miniaturizing biological assays. Adv. Mater. 17, 2911–2933. 6. Bernard, A., Michel, B., Delamarche, E. (2001) Micromosaic immunoassays. Anal. Chem. 73, 8–12. 7. Benn J. A., Hu, J., Hogan B. J., Fry R. C., Samson L. D., Thorsen T. (2006) Comparative modeling and analysis of microfluidic and conventional DNA microarrays. Anal. Biochem. 348, 284–293. 8. Wolf, M., Juncker, D., Michel, B., Hunziker, P., Delamarche, E. (2004) Simultaneous detection of C-reactive protein and other cardiac markers in human plasma using micromosaic immunoassays and self-regulating microfluidic networks. Biosens. Bioelectron. 19, 1193–1202. 9. Rowe, C. A., Tender, L. M., Feldstein, M. J., et al. (1999) Array biosensor for simultaneous identification of bacterial, viral, and protein analytes. Anal. Chem. 71, 3846–3852. 10. Murphy, B. M., He, X., Dandy, D., Henry, C. S. (2008) Competitive immunoassays for
simultaneous detection of metabolites and proteins using micromosaic patterning. Anal. Chem. 80, 444–450. 11. Nam, Y., Branch, D. W., Wheeler, B. C. (2006) Epoxy-silane linking of biomolecules is simple and effective for patterning neuronal cultures. Biosens. Bioelectron. 22, 589–597. 12. Younan, X., Whitesides, G. M. (1998) Soft lithography. Angew. Chem. Int. Ed. 37, 550–575. 13. http://people.seas.harvard.edu/~jones/lab_ arch/nano_facilities/nano_facilities.html, http://microlab.berkeley.edu/, h ttp://snf. stanford.edu/ 14. Tan, J. L., Liu, W., Nelson, C. M., Raghavan, S., Chen, C. S. (2004) Simple approach to micropattern cells on common culture substrates by tuning substrate wettability. Tissue Eng. 10, 865–872. 15. Cesaro-Tadic, S., Dernick, G., Juncker, D., et al. (2004) High-sensitivity miniaturized immunoassays for tumor necrosis factor alpha using microfluidic systems. Lab on a Chip. 4, 563–569.
Chapter 11 Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays Ji Youn Lee and Alexander Revzin Abstract Photolithographic patterning of proteins on surfaces has been used extensively in the past to define cell adhesion domains with micrometer-scale resolution. However, photolithographic patterning is not amenable to depositing several different proteins on the same surface. We propose to merge photolithography with robotic printing of proteins in order to create arrays of protein spots (~300–500 mm diameters) with encoded micrometer-scale cell adhesive domains. This method for biointerface design can employ standard positive tone resist lithography to create temporary stencils for printing of protein arrays. Alternatively, nonfouling poly(ethylene glycol) hydrogels can be micropatterned on top of protein spots. In both cases, cells become adherent on the underlying protein domains, but on-the-spot distribution of cells is defined by the photolithographic pattern. The ability to define multiple cell–substrate and cell–cell interaction scenarios on the same surface is applicable to high-throughput screening of the microenvironment components required for cellular differentiation, for example, for guiding stem cells toward the desired tissue type. Key words: Photolithography, Photoresist, Poly(ethylene glycol) hydrogel, Protein micropatterning, Cellular micropatterning, Microarraying
1. Introduction Controlling cellular interactions by micropatterning is important in tissue engineering, cell-based biosensors, and cell biology. A number of micropatterning approaches, such as photolithography, direct printing, microfluidics, and soft-lithography, have been employed for constructing well-defined cellular microenvironments (1–6). Photolithography is particularly suited for controlling the dimensions and geometry of surface micropatterns. In photolithography, light is projected through a stencil (photomask) onto a light-sensitive polymer (photoresist) making regions within the Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_11, © Springer Science+Business Media, LLC 2011
195
196
Lee and Revzin
polymer layer insoluble in the developer solution. This approach may be used to design cell-adhesive domains on the substrate with single-cell resolution. Importantly, in addition to standard photoresists (2) that are designed to resist exposure to caustic chemicals used in the process of fabricating semiconductor devices, a photoresist could be a biomaterial that resists adsorption of proteins or cells (7, 8). While photolithographic patterning is suitable for defining lateral dimensions of cell-adhesive protein domains, this methodology is limited in the number of different cell adhesive ligands that can be immobilized on the surface. Typically, a single type of protein is used in conjunction with photolithography. Robotic microarraying is an approach designed to print multiple types of biomolecules and is therefore amenable to high-throughput screening studies (9–12). However, most commercial robots print relatively large 100–500 mm diameter protein spots and therefore do not permit precise control of cell–cell interactions. Here we describe a surface modification method combining photolithography and protein microarrays (see Fig. 1) (12, 13). In this method, a micropatterned photoresist or a PEG hydrogel layer serves as a stencil for printing protein microarrays. In the case of photoresist, micropatterned glass substrates with imprinted protein spots are exposed to an organic solvent, leading to lift-off of the photoresist and retention of 20–100 mm diameter cell-adhesive domains in 300–500 mm diameter protein spots. The PEG hydrogel stencil resists adsorption of proteins and therefore does not require removal. Combining photolithography and protein microarraying
Fig. 1. Schematic to show step-by-step procedure for protein and cellular micropatterning. Left panel shows the procedure combining photoresist lithography with protein microarraying to create micron-size protein microdomains for cellular micropatterning. Right panel shows the procedure combining protein microarraying with PEG hydrogel photolithography to create protein microdomains surrounded by nonfouling PEG microwells. Following cell seeding results in an array of hepatocytes.
Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays
197
offers the flexibility of designing multiple cell–cell and cell–surface interaction scenarios all in one culture dish. We envision using this surface engineering approach for high-throughput screening of stimuli required for tissue-specific differentiation of stem cells.
2. Materials 2.1. Surface Cleaning and Modification
1. Glass slides (75 × 25 mm) (VWR international). 2. Four-inch silicon wafer (Wafer World) 3. Piranha solution (see Note 1) – mixture of three parts sulfuric acid (95% v/v in water) and one part hydrogen peroxide (35% w/v in water) (Sigma-Aldrich). 4. 3-Acryloxypropyl trichlorosilane (Gelest, Inc.) (see Note 2). 5. Anhydrous toluene (Sigma-Aldrich).
2.2. Photoresist Lithography
1. AZ 5214-E positive photoresist (Mays Chemical).
2.3. Poly(ethylene Glycol) Hydrogel Lithography
1. Poly(ethylene glycol) diacrylate (PEG-DA) (MW 575) (SigmaAldrich).
2. AZ 300 MIF developer solution (Mays Chemical).
2. 2,2′-Dimethoxy-2-phenylacetophenone (DMPA) (SigmaAldrich) (see Note 3). 3. Precursor solution: thoroughly mix PEG-DA with 1% (w/v) photoinitiator (DMPA). Store the solution in dark at room temperature (see Note 4).
2.4. Protein Micropatterning
1. Collagen from rat tail (type I) (Sigma-Aldrich), FITC-conjugated collagen (type I) (Invitrogen), fibronecin (Millipore), laminin (Sigma-Aldrich), Amersham CyDye™ Value Packs (monoreactive NHS ester) (GE Healthcare), 10× phosphate-buffered saline (PBS) (without calcium and magnesium, Cambrex), Tween-20 (Sigma-Aldrich). 2. Protein solutions (a) Label fibronectin and laminin with Cy3 and Cy5, respectively, according to the manufacturer’s instructions. (b) For physisorption: Dissolve protein in 1× PBS at 0.1 mg/ ml concentration. (c) For microarraying: Dissolve protein in 1× PBS containing 0.005% Tween-20 at 0.2 mg/ml concentration (see Note 5).
198
Lee and Revzin
3. 1× PBS: Dilute 10× PBS in DI water and sterilize by autoclaving or filtering. 4. Tween-20 stock solution: Prepare 0.1% solution by serial dilution of original solution in 1× PBS (see Note 6). 5. Acetone (Sigma-Aldrich). 2.5. Cell Culture and Patterning
1. Tissue culture flask and serological pipettes (Fisher Scientific). 2. Media preparation for HepG2 cells (ATCC): minimum essential media (MEM) supplemented with 10% FBS, 200 U/ml penicillin, 200 mg/ml streptomycin, 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids (all from Invitrogen). 3. Rat primary hepatocytes: Isolated and purified according to established protocols (14). 4. Media preparation for rat primary hepatocytes: Dulbecco’s modified Eagles’ medium (DMEM) supplemented with 10% FBS, 200 U/ml penicillin, 200 mg/ml streptomycin (all from Invitrogen), 7.5 mg/ml hydrocortisone (Pfizer), 20 ng/ml epidermal growth factor (Sigma-Aldrich), 14 ng/ml glucagon (Eli-Lilly), and 0.5 U/ml insulin (Eli-Lilly). 5. Trypsin (0.05%) (Invitrogen).
3. Methods Here we describe the use of photolithographic and microarraying techniques for protein and cellular micropatterning. Extracellular matrix (ECM) proteins were used for patterning and attaching hepatocytes. Subheading 3.1 describes how to modify the substrate using acrylated silane and render the surface moderately hydrophobic. The modified glass surface was conducive to physisorption of proteins such as collagen (I), but it did not support adhesion of hepatocytes, resulting in a clean background. Two photolithographic techniques (1) photoresist lithography and (2) PEG hydrogel lithography are introduced in Subheadings 3.2 and 3.3, respectively. Positive photoresist is patterned on glass to define regions where protein can deposit. Protein molecules are then physically adsorbed onto the surface, followed by the removal of the temporary photoresist stencil along with the protein layer adsorbed on top. In contrast, PEG hydrogels are not removed after microfabrication but permanently remain on the glass surface providing nonfouling regions interspersed with glass domains. Exposure of hydrogel micropatterns to protein solution results in selective adsorption of protein molecules on exposed glass regions. Subheading 3.4.1 describes the method of physiadsorbing protein molecules to create cell
Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays
199
adhesive sites on glass. An alternative protein deposition method, employing microarraying in combination with photolithographic methods as described in Subheading 3.4.2. Microarraying can be used to print multiple ligands on the same surface and therefore offers the potential for high-throughput screening of cell–surface interactions. While cell micropatterning described in Subheading 3.5 focuses on hepatocytes, we have captured and cultured a variety of cell types on protein microarrays by selecting appropriate celladhesive ligands (8, 13, 15). 3.1. Surface Cleaning and Modification
1. Clean glass slides by immersion in piranha solution for 10 min (see Note 7). Then thoroughly rinse the glass slides with deionized (DI) water, dry under nitrogen, and keep them in class 1,000 cleanroom at room temperature. 2. Prior to silane modification, treat the slides in oxygen plasma chamber (YES-R3) at 300 W for 5 min. This process creates reactive oxygen residues on the surface, which can interact with the chlorine group in silane molecules. 3. For silane modification, prepare glass Petri dishes (Pyrex® Glass Culture Dishes, Fisher Scientific) containing substrates, empty glass dish, pipette, tips, and tweezers. Then fill glass dishes with desired volume of anhydrous toluene. Place all the equipment for the modification in the glove bag (Aldrich® AtmosBag, Sigma-Aldrich), vacuum out the air, and introduce enough nitrogen gas to fill the bag (see Note 8). 4. Add 3-acryloxypropyl trichlorosilane to dish containing glass slides to final concentration of 2 mM (e.g., 5 ml per 10 ml toluene). Then rock the dish to mix well and incubate the substrates for 10 min (see Note 9). 5. After modification, transfer the glass slides to a glass Petri dish containing fresh toluene for washing. Take out the dishes from the glove bag, dry the slides under nitrogen, and then cure the slides at 100°C for 2 h. The modified substrates can be stored in a desiccator until use. 6. Characterize the deposition of silane layer either by contact angle measurement (a) or by ellipsometry (b). (a) Place the modified glass slide on the measuring equipment (Rame-Hart goniometer). Then add a drop of water (~5 ml) on the surface and adjust the stage. Read the angle formed between the liquid and solid interface. Typically contact angle of a silane-modified glass slide is 55–60° (see Note 10). (b) The silane layer can be also investigated using ellipsometry (we employ LSE Stokes ellipsometer, Gaertner Scientific). Cut silicon wafer into small pieces (0.5 × 0.5 in.) and modify the substrate using the same procedures as
200
Lee and Revzin
those described for glass slides. Obtain constants from the clean silicon substrate with the refractive index taken to be 1.45 prior to silanization. Ellipsometry measurements from at least three regions of the same substrate are collected to obtain an average thickness for each sample. Typically, the thickness of an assembled silane is 20–25 Å, pointing to the formation of a monolayer or a double layer of silane molecules on the glass surface. 3.2. Photoresist Lithography
1. Spin-coat AZ 5214-E positive photoresist (Solitec spinner) on a silane-modified glass slide at 800 rpm for 10 s followed by 4,000 rpm for 30 s. 2. Soft-bake the coated slide on a hot plate at 100°C for 85 s. 3. Expose the photoresist layer to UV light (10 mJ/cm2) under photomask for 35 s using a Canon PLA-501F Mask Aligner (see Note 11). 4. Develop the exposed photoresist for 5 min in AZ 300 MIF developer solution, wash briefly with DI water to remove residual developing solution, and then dry under nitrogen (see Note 12). Figure 1 illustrates the process and examples for photoresist micropatterns are shown in Fig. 2a.
Fig. 2. Protein and cellular micropatterns created by protein deposition from solution and following cellular adhesion. (a) Photoresist lithographic method. Left panel shows photoresist stencil containing holes of 100 mm in diameter. Cy3labeled collagen (I) deposition and consequent lift-off of photoresist resulted in well-defined protein micropatterns corresponding photoresist micropatterns. Following seeding of hepatocytes created cell clusters of 100 mm in diameter. (b) Array of 30 × 30 mm PEG wells were created by photolithographic method. Middle panel shows that Cy3-labeled collagen only immobilized in PEG wells. This area served as cell attachment pad and the following seeding of hepatocytes generated high-density array hepatocytes with one to three cells per well (right panel ). Reproduced with permission from Revzin et al. (8). Copyright 2003 American Chemical Society.
Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays
3.3. PEG Hydrogel Lithography
201
1. Spin-coat PEG-DA precursor solution at 600–800 rpm for 5 s onto glass surface containing acrylated silane layer (see Note 13). 2. Place a photomask on top of the liquid layer of precursor solution (see Note 14). 3. Expose the sample through a chrome/soda lime photomask to UV for 1.5 s at 60 mJ/cm2 using OmniCure™ Series 1000 light source or 10 s at 10 mJ/cm2 using Cannon PLA-501F Mask Aligner (see Notes 15 and 16). The regions of PEG-DA exposed to UV light undergo free-radical polymerization and become cross-linked, while unexposed regions are dissolved in DI water after 5 min of development (see Note 17). 4. High-resolution images of PEG microstructures are obtained using a JSM 5600LV scanning electron microscopy (SEM) microscope (JEOL, Inc., Peabody, MA) operating at 10 mV of accelerating voltage. Prior to SEM imaging, coat the samples with ~10 nm of Au-Pd by sputtering.
3.4. Protein Micropatterning 3.4.1. Protein Deposition by Physisorption
1. Add protein solution to the photoresist-patterned or PEG hydrogel-patterned glass slide and then incubate for 30 min at room temperature. 2. In the case of photoresist-patterned substrate, sonicate the sample in acetone for 10 min to remove the photoresist, then rinse with DI water and dry under nitrogen (see Note 18). The procedure is similar to that illustrated in the right panel of Fig. 1 except that the sample is incubated with protein solution instead of arraying protein on top of the photoresist layer. In the case of PEG hydrogel-patterned glass slide, wash the sample with DI water at least three times and dry under nitrogen. 3. Glass slide containing protein micropatterns can be kept at 4°C for at least 1 month without detrimental effects to cell attachment. 4. Analyze the size and shape of the protein micropatterns by fluorescent microscopy (we use LSM 5 Pascal confocal microscope, Carl Zeiss, Inc.) using fluorescently labeled proteins (see Note 19). Examples of protein micropatterns are shown in Fig. 2a, b.
3.4.2. Protein Deposition by Microarraying
1. For contact-printed microarrays, our laboratory uses a GMS 417 robotic arrayer (spots of ca. 150 mm in diameter, Genetic Micro Systems, Inc.) or MicroCaster™ hand-held microarrayer system (spots of ca. 500 mm in diameter, Whatman Schleicher & Schuell) under ambient conditions on either silane-modified or photoresist-patterned glass slide. Examples of protein microarrays can be seen in Fig. 3a.
202
Lee and Revzin
Fig. 3. Merging protein microarraying with PEG hydrogel lithography. (a) Array of FITC-collagen (I) spots (ca. 150 mm in diameter with 375 mm center-to-center distance) was generated by robotic microarrayer. Following seeding of hepatocytes generated array of cell clusters. Right panel shows 3 × 3 protein array containing FITC-collagen (I) (lower row), Cy3-fibronectin (middle row), and Cy5-laminin (upper row ). (b) Superimposed image shows that PEG hydrogel microwell patterns are created on the top of the FITC-collagen array (left panel ). After the seeding of hepatocytes, they preferentially adhered to underlying collagen regions while becoming confined within PEG microwells. Higher magnification image by SEM clearly shows that cell adhesion was limited to collagen microdomains. Reproduced with permission from Revzin et al. (12). Copyright 2004 American Chemical Society.
2. In the case of the combination of PEG hydrogel lithography and microarraying, protein spots are first arrayed before the spin coating of the precursor solution (see Subheading 3.4.1) and then PEG microstructures are created on top of the protein micropatterns. The right panel of Fig. 1 illustrates the process flow combining PEG hydrogel lithography and protein microarraying. An example of a superimposed image of FITC-collagen (I) array and PEG microstructure is shown in Fig. 3b. 3. In the case of the combination of photoresist patterning and microarraying, protein spots are arrayed on top of the photoresist layer as illustrated in the left panel of Fig. 1. An example of protein micropatterning is shown in Fig. 4c. 3.5. Formation of Cellular Micropatterns
1. Prior to cell seeding, sterilize protein-micropatterned glass slides with 70% ethanol, and wash with sterile 1× PBS twice (see Note 20). Place the glass slides in the wells of a conventional six-well plate or culture dish (see Note 21). 2. For the cell seeding, collagen-patterned slides were exposed to 3 ml of hepatocytes (either rat primary hepatocytes or
Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays
203
Fig. 4. Merging protein microarraying with photoresist lithography. (a) Photoresist stencil consisting of 50 mm diameter circles. (b) FITC-collagen (I) spot (ca. 500 mm in diameter). (c) Confocal image of FITC-collagen (I) (middle row) and Alexa555-fibronectin (upper and lower rows) robotically printed over a photoresist layer. Smaller microdomains (50 mm in diameter) defined by the photoresist stencil were created. (d) Array of clusters of hepatocytes was created. Each cluster contains 5–6 cells. Reproduced with permission from Lee et al. (13). Copyright 2008 American Chemical Society.
hepatoma cell line, HepG2 cells) suspension in culture medium at a concentration of 1 × 106 cells/ml (see Note 22). 3. After 1 h of incubation at 37°C, the medium containing unattached cells was removed and surfaces were washed twice with 1× PBS (see Note 23). Then the solution was replaced with fresh cell culture medium. 4. Cell arrays formed on the glass slide were imaged using brightfield microscopy (Zeiss Axiovert 40, Carl Zeiss, Inc.). Examples of cell micropatterns are shown in Figs. 2–4.
4. Notes 1. Add sulfuric acid to a glass or quartz container first and then slowly add hydrogen peroxide to prevent violent boiling or splashing. The mixture reacts violently with organic materials and must be handled with extreme care. The solution is extremely hot (over 100°C). 2. Do not open the bottle of saline in ambient conditions because silane functional groups react with water molecules in the air and cause the quality of the self-assembled monolayer to be poor. Smallest quantity from Gelest, Inc. is 5 g, but special order can be placed to obtain five separate bottles of 1 g of solution. This can help to reduce repetitive opening of the bottle, thus preventing a decrease in the conjugation activity. 3. Photoinitiator should be chosen according to the available UV source and its wavelength. Normally, liquid form of
204
Lee and Revzin
photoinitiator is preferred to powder form due to its better miscibility. 4. Thoroughly mix the solution with magnetic stirrer. For the storage, either use dark container or wrap the container with foil. The solution, when kept at room temperature, maintains its crosslinking ability for at least a few months. Optionally, crosslinking agent can be added to the precursor solution to modulate the rigidity and porosity of PEG hydrogel. 5. Protein and Tween-20 concentrations may be adjusted to control the spot size during contact printing. Typically, 0.1– 0.2 mg/ml solution gives good results and too low protein concentration can cause a “coffee stain” effect. In general, the higher the Tween-20 concentration, the larger the spot. However, a too high concentration of Tween-20 makes the shape irregular. Therefore, the concentration should be optimized carefully. 6. The original Tween-20 solution is very viscous and the working solution should be prepared by serial dilution for accuracy. 7. Special caution is required when handling piranha solution. Full protective clothing, including acid-resistant gloves and apron and a face-shield are required. After the cleaning, allow the piranha solution to cool down and then discard the solution in a proper hazardous waste container. 8. The reaction is performed in a glove bag under nitrogen to eliminate atmospheric moisture. 9. Rock the dish long enough to mix the silane solution. An anhydrous environment is very important for the quality of the silane layer. In order to maintain anhydrous environment, glassware should be baked in oven before use and the silane bottle should be opened in the glove bag that is filled with nitrogen. 10. The moderately hydrophobic surface facilitates protein adsorption and also the surface is resistive to cells that require mediator molecules for adhesion. 11. UV exposure time needs to be optimized based on the guidelines from the company. Wear UV-protective goggles for safety. 12. Developing time needs to be optimized along with UV exposure time. After the development, carefully examine samples under microscope. The edge should be sharp and well defined. 13. The spin rate needs to be adjusted according to the desired thickness of PEG hydrogel layer. A higher spin rate will result in the formation of a thinner PEG hydrogel. More solution
Merging Photolithography and Robotic Protein Printing to Create Cellular Microarrays
205
remains on the edges after the spin coating; so one needs to wait until the height of the solution layer becomes even across the substrate. Alternatively, acoverslip can be used to form a thin layer of PEG precursor solution. When low molecular weight of PEG precursor solution is used, spin coating of solution is not desirable due to its low viscosity; the photoinitiator failed to evenly distribute and resulted in a very thin central area and thicker edge. In this case, a coverslip can be used to form uniform thin layer of PEG solution. Add small amount of PEG precursor solution (5–10 ml per 20 × 25-mm glass piece) to the glass piece and then carefully put a coverslip (22 × 22 mm) on top. The size of a coverslip should be bigger than the glass piece to facilitate the removal of the coverslip after UV exposure. 14. Place one or two coverslips between liquid layer and mask to prevent direct contact of solution with mask. 15. UV exposure time needs to be optimized. 16. When using OmniCure 1000, only small samples can be used because the light is not evenly distributed compared to mask aligner. Alternatively, conventional UV illuminator that is used for visualize staining gels can be used when PEG micropatterning over large area is required. 17. Develop sample long enough to get rid of uncrosslinked PEG precursor molecules. PEG residue on surface can cause cell death in later cell patterning steps. When using a coverslip (see Note 8), it should be removed carefully before development. Wafer tweezers will be suitable to remove the coverslip. Place tweezers between coverslip and hydrogel layer and then slowly separate the coverslip from PEG hydrogel layer by lifting. In addition to crosslinking, acrylated groups in PEG molecules covalently bind to the acrylated groups on the glass substrate. 18. The protein integrity after the acetone sonication is investigated by a series of cell adhesion and cellular function assays (13). In the case of collagen (I), there was no detrimental effect on cell adhesion or protein expression by acetone sonication. 19. Alternatively, immunostaining of micropatterned protein can be used. 20. Ethanol residue is harmful for cells, so the residual ethanol should be thoroughly washed off. Alternatively, the samples can be exposed to UV for 30 min. 21. Typically, small glass-slide pieces (one-fourth of regular glass slides) were used for the experiment as they fit in a six-well plate.
206
Lee and Revzin
22. Typically, hepatocytes attach on collagen patterns and start spreading out in 30 min. Higher concentrations of cells can be used to shorten the cell seeding time. 23. Hepatocytes require cell-adhesive ligands for attachment, so they primarily attach to collagen (I) patterned areas; however, there can be some extent of nonspecific attachment to silanemodified surfaces. Carefully wash surfaces by pipetting to get rid of those cells. References 1. Singhvi, R., et al., (1994) Engineering cell-shape and function. Science. 264, 696–698. 2. Bhatia, S.N., M.L. Yarmush, and M. Toner, (1997) Controlling cell interactions by micropatterning in co-cultures: hepatocytes and 3T3 fibroblasts. J. Biomed. Mater. Res. 34(2), 189–199. 3. Takayama, S., et al., (1999) Patterning cells and their environments using multiple laminar fluid flows in capillary networks. Proc. Natl. Acad. Sci. U S A. 96(10), 5545–5548. 4. Park, T.H. and M.L. Shuler, (2003) Integration of cell culture and microfabrication technology. Biotechnol. Prog. 19(2), 243–253. 5. Yap, F.L. and Y. Zhang, (2007) Protein and cell micropatterning and its integration with micro/nanoparticles assembly. Biosens. Bioelectron. 22(6), 775–788. 6. Co, C.C., Y.C. Wang, and C.C. Ho, (2005) Biocompatible micropatterning of two different cell types. J. Am. Chem. Soc. 127(6), 1598–1599. 7. Revzin, A., et al., (2001) Fabrication of poly(ethylene glycol) hydrogel microstructures using photolithography. Langmuir. 17, 5440–5447. 8. Revzin, A., R.G. Tompkins, and M. Toner, (2003) Surface engineering with poly (ethylene glycol) photolithography to creat
9.
10.
11.
12.
13. 14.
15.
high-density cell arrays on glass. Langmuir. 19, 9855–9862. Schena, M., et al., (1995) Quantitative monitoring of gene-expression patterns with a complementary-DNA microarray. Science. 270(5235), 467–470. Anderson, D.G., S. Levenberg, and R. Langer, (2004) Nanoliter-scale synthesis of arrayed biomaterials and application to human embryonic stem cells. Nat. Biotechnol. 22(7), 863–866. Flaim, C.J., S. Chien, and S.N. Bhatia, (2005) An extracellular matrix microarray for probing cellular differentiation. Nat. Methods. 2(2), 119–125. Revzin, A., et al., (2004) Designing a hepatocellular microenvironment with protein microarraying and poly(ethylene glycol) photolithography. Langmuir. 20(8), 2999–3005. Lee, J.Y., et al., (2008) Use of photolithography to encode cell adhesive domains into protein microarrays. Langmuir. 24(5), 2232–2239. Dunn, J.C.Y., et al., (1989) Hepatocyte function and extracellular matrix geometry: longterm culture in a sandwich configuration. FASEB J. 3, 174–179. Zhu, H., et al., (2008) A miniature cytometry platform for capture and characterization of T-lymphocytes from human blood. Anal. Chim. Acta. 608(2), 186–196.
Chapter 12 Generation of Protein and Cell Microarrays on Functionalized Surfaces Yoo Seong Choi and Chang-Soo Lee Abstract The technique of selective immobilization of biomolecules in defined positions or areas using a simple procedure is essential for various applications such as biosensors, biochips, biomedical microdevices, and tissue engineering. For the generation of biomolecule microarrays, it is necessary to develop a functional surface retaining protein functionality and cell viability, and an efficient patterning tool having flexibility of size and shape. In this chapter, we have presented the simple tools of protein and cell microarray based on functionalized surface such as a spotting method with improvement of protein functionality, a functionalized silicon-based surface using photolithography, and an orthogonally polyelectrolyte-coated surface based on soft-lithography. Key words: Microspotting, Lift-off process, Nonspecific binding, Functionalized surface, Micromolding in capillaries, Layer-by-layer assembly, Polyelectrolyte multilayers
1. Introduction The selective patterning of various biomolecules in well-defined areas and positions is critical for the development of biosensors and biochips ranging from miniaturized screening tools to clinical diagnostics (1–5). Currently, substrates including glass, silicon, gold, and polymer are photochemically or chemically functionalized for the preparation of efficient immobilization of proteins and cells (5–7). In general, biomolecular attachment on any surface is basically attributed to physical adsorption, electrostatic interaction, chemical cross-linking, and biological affinity binding (2, 8–10). The selective immobilization of proteins or cells on well-defined surfaces having an ability of prevention of nonspecific
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_12, © Springer Science+Business Media, LLC 2011
207
208
Choi and Lee
binding is important for the development of successful microarrays (1–8). Thus, a substrate should be composed of a biomoleculesrepellent region (background; nonadhesive area) and biomolecules-adhesive regions (patterns) (8, 11, 12). The main purposes of developing protein microarrays are to determine protein function, DNA–protein and protein–protein interactions, to screen for protein markers, and to diagnose disease progression. To achieve these goals, the intact activity of immobilized proteins needs to be maintained on a solid surface. The immobilization of proteins on microarrays requires special care compared to conventional protein immobilization on bulk solid matrix, as high concentration of protein solution and small sample volume is often used. In addition, surface modification for cell patterning is an emerging tool for elucidation of the effects of the substrate on cellular response because cells interact with a change in their microenvironment such as extracellular matrix (ECM), signal molecules, proteins, physical constraints, including pressure and shear force, and chemical composition of substrates (11–13). This chapter describes three representative methods for the generation of protein and cell microarray. The first method is microspotting for the preparation of protein microarray. On the contrary of conventional microspotting, we have presented the strategy of improvement of protein stability with the addition of protein stabilizer (14, 15). Protein stability in microarrays is improved using protein stabilizers (PEG 200, 30% w/v). Under the optimized conditions, the protein stability is improved over fourfold compared to that without the protein stabilizer. The second method is protein micropatterning on a siliconbased functionalized surface derived from a photolithographic lift-off process. It can provide two-dimensional patterns of spatially hydrophilic regions and hydrophic backgrounds. The background hydrophobic thin film is used to suppress nonspecific protein binding, and the hydrophilic target protein binding region is chemically modified to introduce aldehyde group after removal of the photoresist layer. The difference in surface energy between the hydrophilic pattern and background hydrophobic film can induce easier covalent binding of proteins onto defined hydrophilic areas having physical and chemical constraints. Finally, a method for protein and cell patterning on polyelectrolyte (PEL)-coated surfaces using simple micromolding in capillaries (MIMIC) is described. MIMIC produced two distinctive regions. One contained polyethylene glycol (PEG) region fabricated using photopolymerization or adsorption of PEG polymer that provided physical, chemical, and biological barriers to the nonspecific binding of proteins, bacteria, and cells. The second region was the PEL-coated surface that promoted protein and cell immobilization.
Generation of Protein and Cell Microarrays on Functionalized Surfaces
209
2. Materials 2.1. Microspotting of Proteins Using Covalent Binding on Functionalized Glass Slides
1. SuperAldehyde glass substrates (TeleChem International, CA, USA) (see Note 1). 2. Stealth microarray pins (TeleChem International) and microspotting robotic system (see Note 2). 3. Phosphate-buffered saline (PBS), 1 mM, pH 7.5. 4. Polyethylene glycol 200 (PEG, MW 200) (see Note 3). 5. Bovine serum albumin (Sigma, St. Louis, MO).
2.2. Protein Patterning on the Microfabricated Silicon-Based Surface 2.2.1. Microfabrication of Well-Defined Features Using Lift-Off Process
1. Silicon substrate: silicon nitride (Si3N4) layer is grown on silicon substrate using plasma-enhanced chemical vapor deposition method. 2. CYTOP™: A mixture of cyclized perfluoro polymer (CPFP) and its solvent, CTL-809M (Asahi glass Co., Tokyo, Japan). 3. Sulfuric acid (95–98%). 4. Hydrogen peroxide (30% solution in water). 5. Acetone. 6. Ethanol.
2.2.2. Surface Modification of the Microfabricated Silicon-Based Surface
1. Acetic acid. 2. 3-Aminopropyl triethoxysilane (Aldrich, St. Louis, MO). 3. Glutaraldehyde 25% solution (Grade II, Sigma). 4. PBST buffer: 5 mM phosphate buffer, 0.5% Tween-20. 5. Streptavidin (Sigma). 6. Sulfo-NHS-LC-LC-Biotin (Pierce, IL): The biotinylation is performed according to the recommended protocol from Pierce.
2.3. Protein and Cell Patterning on PolyelectrolyteCoated Surfaces Using Micromolding in Capillaries 2.3.1. Preparation of Polyelectrolyte Solutions
1. Cationic PEL solution is prepared with polyallylamine hydrochloride (PAH, MW 70,000, Sigma): Dissolve PAH (20 mM, pH 9.0) in deionized water based on the repeatingunit molecular weight and remove the impurities and nonsoluble residues through filtering system (pore size ~0.2 mm). Store the filtered solution at room temperature before use. 2. Cationic PEL solution is prepared with polystyrene sulfonate ammonium salt (PSS, MW 200,000, Sigma): Dissolve PSS (60 mM, pH 7.0) in deionized water based on the repeating-unit molecular weight and remove the impurities and nonsoluble residues through filtering system (pore size ~0.2 mm). Store the filtered solution at room temperature before use.
210
Choi and Lee
2.3.2. Fabrication of PEG Microstructures Using MIMIC
1. Polydimethylsiloxane (PDMS) prepolymer and curing agent (Sylgard 184, Dow Corning, USA). 2. Polyethylene glycol-dimethacrylate (PEG-DMA) (Mn = 330) (Aldrich). 3. Photoinitiator: DAROCURE 1173 (Ciba-Geigy, USA).
2.3.3. Preparation of Orthogonally Functionalized Surface Using MIMIC
1. Monomethoxy poly(ethylene glycol) (MPEG, Mn = 2,000 and 5,000) (Aldrich). 2. Poly(ethylene glycol) (PEG, Mn = 2,000) (Aldrich). 3. Stannous octoate (Sn(Oct)2) (Aldrich). 4. l-Lactide (LA) (Polyscience).
2.3.4. Synthesis of Biofouling Repellent PEG–PLA Copolymer
1. Various kinds of PEG–PLA block copolymers having different block lengths and compositions are synthesized by ring opening polymerization of LA in the presence of MPEG as a macroinitiator and stannous octoate as a catalyst. 2. A predetermined amount of MPEG is placed into a two-neck flask and dried under vacuum condition at 80°C. 3. The temperature is increased to 110°C. 4. 0.4 wt% Stannous octoate in toluene is added as a catalyst, with N2 purge. 5. After a desired amount of LA is added to the flask. 6. N2 purging and evacuation are repeated at least three times to make an oxygen- and moisture-free environment. 7. The reaction flask is evacuated and heated up to 140°C, with stirring, for 5 h. 8. The resulting product is dissolved in dichloromethane and precipitated with an excess of cold diethyl ether, followed by drying in a vacuum oven for 24 h. 9. Before the use of the PEG-PLA, the characterizations of synthetic polymer have to be analyzed (see Note 4).
3. Methods 3.1. Microspotting of Proteins Using Covalent Binding on Functionalized Glass Slides
1. Prepare protein solutions in PBS with 30% PEG 200 to improve the protein stability and prevent evaporation of microspots of protein solution. Sample result is shown in Fig. 1. 2. Load the aldehyde modified glass slides and protein solution in the spotting device. 3. Spot nanoliter protein solutions on the glass slide, yielding protein spots of 150–200 mm.
Generation of Protein and Cell Microarrays on Functionalized Surfaces
211
Fig. 1. Phase contrast microscopy image of protein spots with different concentrations of PEG 200. This image was captured after 3 h incubation.
4. Incubate for 3 h at room temperature. 5. Wash the slide twice using PBS buffer. 6. Quench free aldehyde groups on the glass slide in 1% (w/v) bovine serum albumin for 1 h. 7. Incubate the printed slides for 1 h at room temperature in solution of target proteins. 8. Decant the solution containing target proteins and add PBST buffer. 9. Incubate slide for 30 min with gentle shaking. 10. Decant PBST buffer and add PBS buffer. 11. Gently shake for 15 min. 12. Repeat washing step for three times. 13. Dry slide under the nitrogen gas or in a centrifuge equipped with slide racks below 450 g force for 1 min. 14. Analyze slide using a commercial fluorescence scanner. 3.2. Protein Patterning on the Microfabricated Silicon-Based Surface 3.2.1. Microfabrication of Well-Defined Features Using Lift-Off Process
1. All of the microfabrication is performed in the clean room. 2. Place silicon substrates for 30 min in a piranha solution consisting of a 4:1 mixture of 50% aqueous solution of H2SO4 and 30% aqueous solution of H2O2. 3. Rinse the substrates with deionized water, completely (see Note 5).
212
Choi and Lee
4. 3 mm-Thick photoresist (AZ 4330) is spun on the substrates and patterned by the conventional photolithography. 5. Spin-coat subsequent hydrophobic thin film (CYTOP™) on the substrates at a speed of 450 g force. 6. Remove the hydrophobic thin film attached to the patterned photoresist by acetone solution in the ultrasonic bath. 7. Rinse sequentially in acetone for 10 min, ethanol for 15 min, and deionized water for 15 min in the ultrasonic bath. Finally, we have obtained highly ordered micropatterns on silicon-based surfaces with hydrophobic thin film background. 3.2.2. Surface Modification of the Microfabricated Silicon-Based Surface
1. Place the microfabricated substrate for 3 h in a modified piranha solution of a 4:1 mixture of 50% aqueous solution of CH3COOH and 30% aqueous solution of H2O2 to preserve the hydrophobic thin film safely. 2. Dry the substrates on the convectional oven at 110°C for 3 h. 3. Immerse the oxidized substrates in ethanol containing 10% 3-aminopropyl-triethoxysilane at room temperature for 12 h. 4. Wash the substrates three times with 95% ethanol/water, three times with 100% ethanol, and three times with deionized water to remove unbound silane compounds and finally dry them under nitrogen stream. 5. React the amino-silanized substrate with 10% glutaraldehyde in 1 mM PBS (pH 7.5) at 30°C for 1 h. 6. Rinse several times with deionized water. 7. Apply 5 mM PBST buffer (5 mM phosphate buffer, 0.5% Tween-20, pH 7.0) containing 1 mg/ml streptavidin into the substrates at room temperature for 1 h. 8. Rinse several times with PBS. 9. Apply biotin-labeled biomolecules into the microarray (see Note 6). 10. Decant the solution containing biotin-labeled biomolecules and add PBST buffer. 11. Incubate slide for 30 min with gentle shaking. 12. Decant PBST buffer and add PBS buffer. 13. Gently shake for 15 min. 14. Repeat washing step for three times. 15. Dry slide under the nitrogen gas. 16. Analyze silicon chip using a commercial fluorescence microscopy. All process is shown in Fig. 2.
Generation of Protein and Cell Microarrays on Functionalized Surfaces
213
Fig. 2. Schematic illustration of the microfabrication process of CYTOPTM on substrates. At first, photoresist patterns were made by conventional photolithography. CYTOPTM polymer was spun on the substrate. The CYTOPTM film-coated wafer was then baked in a convectional oven at 110°C for 10 min. The baked thin film was stripped from the photoresist features by acetone solvent. (a) Thick PR (photoresist) spin coating and patterning. (b) CYTOPTM film spin coating and baking in a convectional oven. (c) PR stripping for remove CYTOPTM attached to PR. (d) Chemical modification with aminosilane and aldehyde linkage. (e) Protein loading from aqueous buffer.
3.3. Protein and Cell Patterning on PolyelectrolyteCoated Surfaces Using Micromolding in Capillaries 3.3.1. PEL Coating on Glass
1. Place glass slides for 30 min in a piranha solution consisting of a 4:1 mixture of 50% aqueous solution of H2SO4 and 30% aqueous solution of H2O2. 2. Rinse glass slides with deionized water, completely. 3. Place the slides in boiling distilled water for 10 min and dry with nitrogen. 4. Immerse the cleaned glass in PAH solution (20 mM, pH 9.0) for 20 min.
214
Choi and Lee
5. Wash three times with distilled water. 6. Immerse the PAH-coated glass in PSS solution (60 mM, pH 7.0) for 20 min. 7. Wash three times with distilled water. 8. Repeat steps 4–7 until the desired number of PEL layers is assembled on the slide with a positive charged PAH on the outermost layer (see Note 7). 3.3.2. Fabrication of PEG Microstructures Using MIMIC Method
1. Produce a master by conventional photolithography. 2. Mix PDMS prepolymer and curing agent (with 10:1 weight ratio) well, and degas in a vacuum desiccator until all bubbles are sufficiently removed. 3. Clean the master by nitrogen purging, and load the degassed PDMS solution into the master and degas again. 4. Cure the PDMS on the master in 8 h at 65°C. 5. Cool the mold into room temperature and remove PDMS from the master (see Note 8). 6. Clean the PDMS mold by sonication in 100% ethanol for 10 min and dry for 1 h. 7. Place the trimmed PDMS mold on PEL-coated glass to make conformal contact. 8. Pour PEG-DMA containing 0.5% (v/v) photoinitiator at the open ends of micromolds (see Note 9). 9. Cure PEG-DMA with an ultraviolet light (250–400 nm, 100 mJ/cm2) for 15 min. 10. Peel off the PDMS micromold.
3.3.3. Preparation of Orthogonally Functionalized Surface Using MIMIC
1. PDMS (Sylgard 184, Dow Corning) micromolds are fabricated against a complementary relief structure that was prepared by conventional photolithography. 2. The heights of the microstructures in the PDMS are approximately 20 mm. 3. Each PDMS mold is cut such that it forms a network with open ends. 4. The trimmed PDMS mold is placed on PEL-coated glass. 5. When the PEG–PLA copolymer is placed at the open ends of the micromolds, it spontaneously fills the void spaces by capillary action. 6. PEG–PLA is bound with PEL multilayer for 1 h and then the PDMS micromold is peeled off. Sample result is shown in Fig. 3.
Generation of Protein and Cell Microarrays on Functionalized Surfaces
215
Fig. 3. A schematic diagram for protein and cell patterning using PEL coating and MIMIC: (a) surface modification with PEL (PAH/PSS), (b) PDMS micromold placement, (c) filling the micromold with a liquid prepolymer by capillary action, (d) the curing of prepolymer by UV radiation, (e) removing the PDMS micromold, and (f) loading of biomolecules onto the fabricated surface.
4. Notes 1. Most of microarray substrates have been used for DNA microarrays. Ideally, the substrate is able to attach biomolecules with preserving their structure and activity. Due to irregular surface nature of biomolecules, optimal substrates for target biomolecules should be investigated based on chemical inertness, nonspecific binding properties, high and uniform density of the molecular surface with accessible end groups for binding, and compatibility with analytical instrument such as scanner, fluorescence microscopy, or mass spectrometry. Substrates including glass, silicon, gold, and polymer are photochemically or chemically modified with amine, aldehyde, NHS, epoxy, nickel chelate, or carboxy acids for the covalent immobilization of proteins. The modified glass slides are also commercially available from Corning Microarray Technology, Genetix Ltd, Xenopore Corp and
216
Choi and Lee
Cel. In this chapter, the commercial aldehyde-coated glass slides are used for the preparation of protein microarrays because the aldehyde group easily reacts with the primary amine group of proteins by forming Schiff base. 2. There are many commercial robotic systems available for microspotting. They address the printing pin selection and equipment required to successfully operate their microarray system. Some basic troubleshooting and quality control tools to aid in producing high-quality printed microarrays on a consistent basis are summarized by R. A. George (16). In this chapter, a homemade spotting robot system equipped with stealth pin (TeleChem International) is used to deliver nanoliter protein solutions on the aldehyde-coated glass slide, yielding protein spots of 150–200 mm. For the precise spotting, spotting schedule was programmed by computer control. 3. Hydrogels, sucrose, glycerol as well as PEG 200 can be generally used as protein stabilizers during the spotting of protein solution. However, PEG 200 at 30% (w/v) has shown the most efficient stabilizer in our study. 4. The chemical compositions of the obtained polymers are confirmed by proton-nuclear magnetic resonance (1H-NMR), measurements using a JNM-AL400 spectrometer (Jeol Ltd, Akishima, Japan) at 400 MHz in CDCl3. Molecular weights and molecular weight distributions are determined by gel permeation chromatography (GPC) (Agilent 1100 HPLC, Plgel MIXED- D & E columns column, poly(ethylene glycol) standard). The flow rate was 1 ml/min and the temperature (both the column compartment and the flow cell of the refractive index detector) is fixed at 40°C. Tetrahydrofuran (HPLC grade) is used as the eluting solvent. Fourier transform infrared (FTIR) spectrometer (Niclotet Magna 560 spectrometer, USA) is used to investigate the presence of the ester carbonyl group in the synthesized copolymers. 5. Cleaning procedure can also involve detergents, strong oxidizers (oxygen plasma, NH3/H2O2), sonication, acid, or base. 6. When different concentrations of FITC-BSA-biotin solution are incubated with the patterned streptavidin surface to quantify binding of FITC-BSA-biotin concentrations, the fluorescence signal intensity linearly increases with the increase in the concentration of the FITC-BSA-biotin up to 100 ng/ml (5). 7. PEL coating is readily monitored with the change of water contact angle because the layer-by-layer method results in significant variation of contact angle. 8. The heights of the microstructures in the PDMS blocks are about 20–100 mm. 9. The PEG-DMA spontaneously fill the void spaces by capillary action.
Generation of Protein and Cell Microarrays on Functionalized Surfaces
217
References 1. Mooney, J.F., Hunt, A.J., McIntosh, J.R., Liberko, C.A., Walba, D.M., and Rogers, C.T. (1996). Patterning of functional antibodies and other proteins by photolithography of silane monolayers. Proc. Natl. Acad. Sci. USA. 93, 12287–12291. 2. Blawas, A. S., and Reichert, W. M. (1998). Protein patterning. Biomaterials. 19, 595–609. 3. Bouaidat, S., Berendsen, C., Thomsen, P., Petersen, S. G., Wolff, A., and Jonsmann, J. (2004). Micro patterning of cell and protein non-adhesive plasma polymerized coatings for biochip applications. Lab. Chip. 4, 632–637. 4. Barbulovic-Nad, I., Lucente, M., Sun, Y., Zhang, M. J., Wheeler, A. R., and Bussmann, M. (2006). Bio-microarray fabrication techniques – a review. Crit. Rev. Biotechnol. 26, 237–259. 5. Lee, C.S., Lee, S.H., Park, S.S., Kim, Y.K., and Kim, B.G. (2003). Protein patterning on silicon-based surface using background hydrophobic thin film. Biosens. Bioelectron. 18, 437–444. 6. MacBeath, G., and Schreiber, S.L. (2000). Printing proteins as microarrays for highthroughput function determination. Science. 289, 1760–1763. 7. Nicolau, D.V., Taguchi, T., Taniguchi, H., Tanigawa, H., and Yoshikawa, S. (1999). Patterning neuronal and glia cells on lightassisted functionalised photoresists. Biosens. Bioelectron. 14, 317–325. 8. Shim, H.W., Lee, J.H., Hwang, T.S., Rhee, Y.W., Bae, Y.M., Choi, J.S., Han, J., and Lee, C.S. (2007). Patterning of proteins and
9. 10.
11.
12.
13.
14. 15. 16.
cells on functionalized surfaces prepared by polyelectrolyte multilayers and micromolding in capillaries. Biosens. Bioelectron. 22, 3188–3195. Emili, A.Q., and Cagney, G. (2000). Largescale functional analysis using peptide or protein arrays. Nat. Biotechnol. 18, 393–397. Jang, C. H., Tingey, M. L., Korpi, N. L., et al. (2005). Using liquid crystals to report membrane proteins captured by affinity microcontact printing from cell lysates and membrane extracts. J. Am. Chem. Soc. 127, 8912–8913. Suh, K. Y., Khademhosseini, A., Jon, S., and Langer, R. (2006). Direct confinement of individual viruses within polyethylene glycol (PEG) nanowells. Nano Lett. 6, 1196–1201. Lee, J. H., Kim, H. E., Im, J. H., et al. (2008). Preparation of orthogonally functionalized surface using micromolding in capillaries technique for the control of cellular adhesion. Colloids Surf. B Biointerfaces. 64, 126–134. Hasirci, V., and Kenar, H. (2006). Novel surface patterning approaches for tissue engineering and their effect on cell behavior. Nanomedicine. 1, 73–89. Lee, C. S., and Kim, B. G. (2002). Improvement of protein stability in protein microarrays. Biotechnol. Lett. 24, 839–844. Dufva, M., and Christensen, C. B. V. (2005). Diagnostic and analytical applications of protein microarrays. Expert Rev. Proteomics. 2, 41–48. George, R.A. (2006) The Printing Process: Tips on Tips. In Kimmel, A. and Oliver, B., ed. Methods in Enzymology, Vol. 410. Elsevier, Philadelphia, pp. 121–135.
Chapter 13 Microprinting of Liver Micro-organ for Drug Metabolism Study Robert C. Chang, Kamal Emami, Antony Jeevarajan, Honglu Wu, and Wei Sun Abstract In their normal in vivo matrix milieu, tissues assume complex well-organized 3D architectures. Therefore, a primary aim in the tissue engineering design process is to fabricate an optimal analog of the in vivo scenario, in which the precise configuration and composition of cells and bioactive matrix components can establish the well-defined biomimetic microenvironments that promote cell–cell and cell–matrix interactions. With the advent and refinements in microfabricated systems which can present physical and chemical cues to cells in a controllable and reproducible fashion unrealizable with conventional tissue culture, high-fidelity, high-throughput in vitro models are achieved. The convergence of solid freeform fabrication (SFF) technologies, namely microprinting, along with microfabrication techniques, a 3D microprinted micro-organ, can serve as an in vitro platform for cell culture, drug screening, or to elicit further biological insights. This chapter firstly details the principles, methods, and applications that undergird the fabrication process development and adaptation of microfluidic devices for the creation of a drug screening model. This model involves the combinatorial setup of an automated syringe-based, layered direct cell writing microprinting process with soft lithographic micropatterning techniques to fabricate a microscale in vitro device housing a chamber of microprinted 3D micro-organ that biomimics the cell’s natural microenvironment for enhanced performance and functionality. In order to assess the structural formability and biological feasibility of such a micro-organ, 3D cell-encapsulated hydrogelbased tissue constructs are microprinted reproducibly in defined design patterns and biologically characterized for both viability and cell-specific function. Another key facet of the in vivo microenvironment that is recapitulated with the in vitro system is the necessary dynamic perfusion of the 3D microscale liver analog with cells probed for their collective drug metabolic function and suitability as a drug metabolism model. Key words: Microfluidics, Cell printing, Tissue engineering, Solid freeform fabrication, Hydrogels, Pharmacokinetics
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_13, © Springer Science+Business Media, LLC 2011
219
220
Chang et al.
1. Introduction The principle drivers for using a large and expanding array of microfabrication tools for conferring unique properties and parallelization of cell-based biological assays with unique cellular microenvironments include facilitating better diagnostics and improving functionality of chemical and biological sensors (1, 2). In addition, the recent development and confluence of two other specific technologies and research domains has made the engineering of tissue constructs amenable for in vitro applications: synthetic, tunable hydrogels to create 3D cellular microenvironments and microbioreactors to precisely control nutrient transport and fluid shear stress. Furthermore, convergence of all these enabling tools provides investigators with the opportunity to construct and study tissues in vitro with heretofore unseen levels of sophistication. Broadly speaking, tissue engineering approaches exploit living cells in a variety of ways toward the goal of restoring, maintaining, or enhancing tissues and organs (3, 4). One promising approach is the implementation of microprinting technology to deploy materials, cells, and tissues. 1.1. Applications of Microprinting Technology
In order to engineer biological tissues in vitro, cultured cells are coaxed to grow on bioactive bioresorbable tissue scaffolds, i.e., temporary synthetic extracellular matrices that provide the biological, chemical, and mechanical cues to guide the cell’s eventual differentiation and assembly into 3D tissues (5). While regeneration of different tissues and organs is currently undergoing investigation and development, new applications of tissue engineering for designing in vitro physiological models to study disease pathogenesis and for pharmacokinetic study also show great promise (6). The present work explores the development and characterization of an in vitro 3D microfluidic device for simulation of the physiological human response to drug administrations and toxic chemical exposures. A layered fabrication-based microprinting technology summoned herein has several practical applications for cells, materials, and tissues microprinted and cultured on its microscale platform. These include, but are not limited to drug screening, biosensors, and the study of cell–cell or cell–matrix interactions in cell biology in disease pathogenesis.
1.1.1. Drug and Toxicology Screening Models
One possible near-future application of in vitro physiological models established from biology-based engineering analysis is in the area of pharmaceutical drug and toxicology screening for drug discovery and development. Currently, cell-based drug screening is used in pharmaceutical development, followed by several rounds of animal testing. Microprinting technology, however,
Microprinting of Materials, Cells, and Tissues
221
can serve as an effective bridge between cell-based assays and animal studies by enabling tissue-based drug screening, resulting in more accurate assessments of drug action to partially replace animals in preclinical drug testing. Existing in vitro cell culture models with human liver cells have already shown great potential in predicting drug toxicity and metabolism in the pharmaceutical industry (2, 7). For example, a microfabricated array bioreactor has been designed for perfused 3D liver culture at reported near-physiological ranges of perfusate flow rates and fluid shear stresses. Primary rat liver cells cultured within these microchannels for a 2-week duration undergo rearrangement to form tissue-like structures. Others have reported the modeling of a microscale in vitro system to serve as a human surrogate for drug analysis, which adheres stringently to a physiologically based pharmacokinetic (PBPK) model in which disparate organs are serially connected by channels to comprise a fluid circuit of modular systems. This methodology is then used to mechanistically simulate and predict drug biotransformation, distribution, and efficacy in vivo. Therefore, one can imagine the design and fabrication of a stamp-sized animal-ona-chip microfluidic device which, with great fidelity, accurately simulates the process of an experimental drug being broken down by the metabolizing liver, absorbed by the intestines, and held onto by fat (8–10). These models mark the far-reaching potential for in vitro drug screening technology as well as highlight the current availability of crude analogs for the organ functions they are intended to biomimic. The inadequacies of current testing methods include, but are not limited to, the lack of a high-fidelity 3D microenvironment and the ability to reproducibly fabricate 3D tissue constructs. The applied rapid prototyping technology presented is a structurally and biologically viable syringe-based microprinting process for layer-by-layer fabrication of 3D cell-encapsulated hydrogel-based tissue constructs. The research conducted is aimed at the achievement of high-throughput reproducible fabrication of microprinted tissue constructs, maintenance of structural integrity, direct integration with the microfluidic platform, and enhancement of cell viability and control of cellular-level differentiation and tissue-level function. More specifically, the objective is to develop a viable direct cell microprinting process for fabrication of reproducible 3D cell-encapsulated alginate-based tissue engineered constructs within 3D tissue chambers as a drug metabolism model. This strategy can be further exploited to fabricate various 3D in vitro tissue analogs consisting of an array of channels with tissue-embedded chambers representing different mammalian tissues for a multitude of applications for clinical pharmaceutical screening efficacy and toxicity for the agent of interest.
222
Chang et al.
1.1.2. Biosensors
Currently, chemical and biological sensors are delicate, complex devices that do not accurately predict the human response to environmental toxins as they are very limited in range and scope. Furthermore, molecular sensors are capable of detecting only single pathogens or chemicals. The limitations of biological sensors are primarily associated with 2D culture on tissue culture plastic which has been a mainstay in conventional cell culture by dispersing complex tissues into single cells and ignoring higherorder processes. Tissue-based sensors fabricated enlisting the precision approaches and processes of microfabrication and microprinting technologies can result in cell signaling pathway components as a highly discriminating and sensitive detection mechanism for identifying a wide range of pathogens and chemicals. More robust sensors based on 3D tissue and/or organ fragments may then be used to rapidly detect biological or chemical threats under a variety of conditions and scenarios. Tools of microprinting technology may enable the development of increasingly sophisticated tissue analogs capable of processing complex environmental information and potentially serve as smart sensors in public health surveillance (11–13).
1.1.3. In Vitro Models of Disease Pathogenesis
High-throughput technologies based on automation, miniaturization, and multiplexing are now feasible for the systematic study of cells. On the one hand, cell-based models have been the prevailing models in studying cancer and disease pathogenesis verified by testing in animal models of diseases (14–17). On the other hand, microprinting technology introduces 3D human tissue models for probing basic biological insights into cells and tissues as well as understanding human disease processes to curtail the use of animals in research. One application is in the area of studying the behavior of complex, highly specialized cells such as neurons that possess the fundamental constraint of requiring other specialized cells for its continued growth and maturation. Substantial progress has already been made in 2D patterning of biological substrates for controlled cell–material interactions at micrometer scales. A micromechanical device housing 3D microprinted tissue constructs, however, can be developed to enable manipulation of these cells in a 3D microenvironment to help explain the fundamental biological processes: cell–cell and cell– matrix signaling and coordination. Furthermore, rooted within these platforms will be measurable surrogates or indicators called biomarkers charged with assessing the physiological state of condition of the 3D microprinted tissue constructs. One such application is with stem cells, which have naturally been investigated as a candidate cell source for tissue engineering applications. Cellular functions, and stem cell differentiation specifically, however, are influenced not only by cell-autonomous programs but also by microenvironmental
Microprinting of Materials, Cells, and Tissues
223
stimuli, which include neighboring cells, extracellular matrix, soluble factors, and physical factors. In embryonic development, for example, many cell types come into contact with one another and communicate toward establishing their final identities. The microprinter can deploy spatially and temporally specified biomarkers for stem cells on a microscale in vitro platform at interim stages of development to identify stem cells and determine their state of differentiation. 1.2. Microprinting Design and Fabrication 1.2.1. Development of Direct Cell Writing System
A proprietary multi-nozzle direct cell writing system has been developed for the layered freeform fabrication of biopolymerbased 3D tissue scaffolds and cell-embedded tissue constructs (1, 18–22). The microprinting process is designed to operate under room temperature and low pressure conditions that are cell friendly to deploy multiple cell types, bioactive factors, or other biologics in controlled amounts with precise spatial positioning to form predesigned, cell-embedded tissue constructs with easily adaptable patterned architectures. Other layered manufacturing methods utilize harsh solvents, high pressures or temperatures, or postprocessing methods that are not compatible with cells and bioactive materials. Compared with most reported cell dispensing systems that are mostly limited to a single nozzle for cell dispensing or limited to cell printing only, the multi-nozzle microprinting capability enables the simultaneous deposition of cells, growth factors, and scaffold materials to form heterogeneous or functional gradient tissue structures. The direct cell writing system configuration is shown in Fig. 1. A designed CAD model of a tissue construct is processed by the data processing software and translated into a 2D process tool path which is then communicated to the motion control to drive the layered fabrication process. This microprinting technology has been adapted to accommodate a wide range of biocompatible and biodegradable aqueous-based polymeric materials and bio active components where the material delivery system supplies various nozzles with the appropriate cellular components and biopolymer matrix material or biological factor. The system consists of an air pressure supply, a material container or reservoir, and a material delivery tube. Each nozzle system has its independent process and the material parameters are tuned and optimized (e.g., the dispensing air pressure and biopolymer concentration). The system implements multiple nozzles with different types and sizes, thus enabling the microprinting of specified hydrogels with different viscosities for constructing 3D tissue constructs. In the development of the direct cell writing system, several micronozzle systems have been investigated to evaluate their performances and feasibility to print cell-embedded biopolymer solutions for tissue engineered constructs. The specific biopolymer microprinting process detailed in this chapter implements a pneumatic
224
Chang et al.
Fig. 1. Direct cell writing system configuration (source: Chang et al. [19] ).
microvalve which is a typical mechanical valve that opens and closes the valve via an applied air pressure regulated by a controller. The pneumatic microvalve is capable of operating in extrusion mode where the controller applies pressure to open the valve by lifting the piston against the spring that lifts the needle from the needle seat. The biopolymer material is then extruded out of the nozzle tip under an applied pressure that is adjusted through the material delivery system. The process of extrusion is complete when the controller closes the valve by reverting the needle back to the needle seat. Multiple pneumatic valves can then be simultaneously operated for microprinting multicellular multimaterial heterogeneous constructs in the development of a high-fidelity 3D microscale tissue analog. The system can continuously extrude hydrogel strands ranging from 30 to 500 mm in their feature dimensions or discrete hydrogel droplets with picoliter volumes. 1.2.2. Direct Cell Writing for Fabrication of Micro-organ Device
The convergence of several enabling tools, namely, solid freeform fabrication (SFF) and microfluidic techniques, is providing investigators with a novel opportunity to construct and study tissues in vitro with heretofore unseen levels of sophistication. Modifications of these traditional techniques have been put in place for these emerging technologies and scientific disciplines to work in concert for novel applications of tissue- engineered constructs as illustrated in Fig. 2.
Microprinting of Materials, Cells, and Tissues
225
Fig. 2. Convergence of layer-by-layer biofabrication and microfabrication techniques (source: Chang et al. [1] ).
Distinct from 2D cell monolayer standard in conventional cell culture methods, the direct cell writing process is integrated with a microfluidic device to fabricate 3D tissue/organ constructs/chambers shown in Fig. 3. Biological studies reveal a fundamentally different cellular phenotype with reduced tissue-specific gene expression with conventional monolayer in vitro culture techniques (23–26). A 3D tissue model will, in contrast, cultivate improved establishment and maintenance of cell-specific function. To create 3D tissue models in practice, cells are first suspended in aqueous sodium alginate solution and microprinted through the system into a design pattern within a microfluidic device. For the alginate-cell suspension, an air-actuated dispensing nozzle is applied with a prespecified internal diameter tip size. Prior to microprinting, ethanol, filtered air, and deionized water are purged through the system for cleaning and sterilization. The other component in the 3D microfluidic micro-organ device system is the microfluidic device with indented chambers. This includes a glass slide with etched microchannels and a polydimethylsiloxane (PDMS) substrate with an indented chamber to constitute the microfluidic device. Both elements of the device are furnished by NASA-JSC, a collaboration from which this
226
Chang et al.
Fig. 3. Overview of direct cell writing 3D micro-organ approach. (a) Schematic representation; (b) fabrication of 3D tissue constructs (source: Chang et al. [19] ).
project is borne out (20). The feasibility of using a standard soft-lithography technique is explored to fabricate microscale in vitro device with microchannels and tissue chambers to house the microprinted tissue. The preferred material for fabrication of the microscale device for housing the embedded tissue is PDMS. PDMS elastomer soft lithography is combined with the micromolding techniques to fabricate a substrate with microfluidic channels and a chamber to house to printed tissue. The advantages of using PDMS for the microfabrication include the ease of bonding, optical properties, and permeability to gases for biological applications. The glass layer consists of an etched bifurcating channel pattern with inlet and outlet holes drilled into the glass (see Fig. 4). The next step is to directly microprint the cell-encapsulated alginate constructs within the PDMS chamber. The PDMS substrate contains a central 10 mm × 10 mm × 750 mm deep indentation. Prior to integrating the alginate constructs onto the PDMS substrate, both the glass and PDMS layers are surface-modified by an RF plasma cleaner to oxidize the glass and PDMS substrate surface methyl groups to form silanol groups. Therefore, a hydrophilic surface with good wettability and improved traction is conferred upon the microfluidic device. The glass and PDMS surfaces are first treated with air plasma, upon which the direct cell writing system microprints alginate admixed with HepG2 liver cell solution
Microprinting of Materials, Cells, and Tissues
227
Fig. 4. Etched microfluidic channel patterns on glass slide (source: Chang et al. [19]).
Fig. 5. Direct microprinting of sinusoidal design pattern directly into chamber (source: Chang et al. [19] ).
in a sinusoidal flow pattern with 250 mm diameter struts (see Fig. 5) using the air-actuated printing nozzle. Calcium chloride and media are deposited using a separate pneumatic printing nozzle to crosslink the alginate and furnish the cells with nutrient media. The PDMS is then bonded to the glass layer with the bifurcating channels aligned to the chamber. Through this process, the cell-encapsulated alginate construct is directly microprinted within the tissue chamber of the PDMS substrate, thus forming an integrated 3D tissue chamber unit. 1.2.3. Application of Liver Micro-organ as a Drug Metabolism Model
Fabricating 3D microscale tissue analogs in vitro requires appropriate microbioreactors that simulate physiological environments for the creation, mechanical perturbations of, integration, and
228
Chang et al.
Fig. 6. Setup of direct cell written 3D micro-organ microbioreactor for (a) parallel drug flow study with (b) fully drug perfused tissue chambers (Chang et al. [1] ).
testing of cells with their supportive matrix. The experimental apparatus setup of medium/drug flow circulation in the microbioreactor for drug metabolism analysis is shown in Fig. 6a, b. To demonstrate effective drug metabolism in the liver chamber, a nonfluorescent prodrug feeds into the system through an inlet port, is metabolized within the liver chamber, and then emerges at the outlet stream as an effluent fluorescent metabolite for analysis. Results of such analysis may then be used in future studies to apprehend the relative pharmacokinetic efficiency as well as relevancy of the tissue chamber design for human application. The micro-organ device is then connected to the syringe pump for simultaneous infusion and withdrawal at the inlet and outlet port with a constant designated volumetric flow rate. The drug substrate EFC is metabolized by the resident hepatocytes into the drug product 7-hydroxy-4-trifluoromethyl coumarin (HFC). The perfused micro-organs are placed within an incubator along with static, nonperfused controls. The effluent is collected within the withdrawal syringe and assayed with a cytofluorimeter
Microprinting of Materials, Cells, and Tissues
229
for HFC drug metabolic conversion with a drug residence time consistent with that of the static controls. 1.3. Cell Printing Liver Micro-organ Experiments for Drug Metabolism Study
There are several laboratory steps in the microprinting experiments to determine the feasibility of a microprinted micro-organ as a drug and toxicology platform for studying metabolite concentrations: (1) the structural formability of microprinted 3D tissue constructs, (2) the effect of direct cell microprinting on cell viability, and (3) the effect of direct cell microprinting on cell-specific function.
1.3.1. Structural Formability of Microprinted 3D Tissue Constructs
A microscale tissue analog is designed and fabricated for studying drug metabolism and its pharmacokinetic properties via direct writing of cells with material delivery media as a 3D cell-seeded hydrogel-based matrix. By integrating the microprinting system with a CAD environment, reproducible 3D structures are realized within micron-order dimensional specifications. Following iterative cycles of design and printing, the appropriate microprinting process parameters (dispensing pressure of 2.0 psi, nozzle head velocity of 10 mm/s, and an internal nozzle tip diameter of 200 mm) are identified to achieve the design specifications of a three-layered sinusoidal pattern with consistent roadwidths of 250 mm.
1.3.2. Effect of Direct Cell Writing on Cell Viability
Assays for cell viability are conducted to study the effect of the microprinting process on cell survival. The aim of this study is to evaluate the cell viability ratios during the days immediately following the direct cell writing of cell-embedded constructs. The experimental setup for this study involves two experimental test samples and one control. The first experimental group is the direct cell writing of alginate and hepatocytes with subsequent addition of cell medium. Experimental group 2 is represented by the direct cell writing of alginate and hepatocytes with a CaCl2 ionic cross-linking solution. Both the process and material parameters for the experimental groups are held constant. In this study, the process parameters are maintained. The control implemented is manual pipetting of alginate and hepatocytes with cell medium as the cross-linking agent. The same cell density is used in both tests and the control. For this setup, a qualitative assay is carried out with Live/Dead cell assay. Each data point represented samples taken at days 1, 2, and 3 after direct cell writing and subsequent addition of media or CaCl2 ionic cross-linking solution. The rationale for this setup with assay is to compare and isolate the effects of the microprinting process and biomaterial cross-linking agent on the cell viability. As demonstrated in Fig. 7, good initial cell viability using Live/Dead assays is achieved for encapsulated hepatocytes microprinted under biofriendly conditions.
230
Chang et al. Hepatocyte Cell Viability as a Function of Direct Cell Writing Process and Crosslinking 100.0
Percentage Cell Viability
95.0 90.0 2D Monolayer + Media
85.0 80.0
Printed 3D Construct + Media
75.0
Printed 3D Construct + CaCl2 Crosslinking
70.0 65.0 60.0 1
2
3
Time (Days) Fig. 7. Cell viability of microprinted constructs vs. nonmicroprinted control (source: Chang et al. [19] ).
Figure 7 shows the cell viability ratios at days 1, 2, and 3 obtained by assigning a percentage cell viability to each bulk sample via manual visualization and count performed at five different locations therein. The control with no bioprinting and cell medium crosslinker shows 83% cell viability. On day 1, experimental group 1 printed alginate/hepatocytes with cell medium cross-linking agent demonstrated an 82% cell viability, and experimental group 2 microprinted alginate/hepatocytes with CaCl2 crosslinker gave a 79% cell viability. A comparison of experimental group 1 and control indicates that a slight drop in cell viability may be attributed to the microprinting process. Furthermore, a comparison between the experimental groups indicates that a slight drop in cell viability may be attributed to the substitution of a stronger crosslinker (CaCl2) for one having trace amounts of cross-linking ions (cell medium). Overall, the cell viability tests demonstrate that hepatocytes are able to survive through microprinting process with a range of 79–82% viability ratio from days 1 to 3, respectively. A previous paper on the direct cell microprinting system examines the effect of varying process parameters on the viability of HepG2 cells within alginate (18). The results suggest that cells require a time period to recover its function after the depositioninduced senescent state. These results establish the range of permissible process parameters and length of culture necessary for cell recovery after direct cell writing. Another study is carried out to assess cell viability by capturing Live/Dead images comparing static 2D controls vs. 3D cross-linked
Microprinting of Materials, Cells, and Tissues
231
tissue constructs under drug flow perfusion over a 24-h period. For the experimental flow group, the volumetric flow rate is set at 0.015 mL/h with constant process parameters maintained across both experimental groups. This flow study demonstrates improved viability of cells microprinted as a 3D tissue construct subsequently subjected to a 24-h flow study period compared with that of the static 2D controls. In order to characterize the effect of the direct cell writing system on subsequent cell proliferation, a proliferation assay is performed at different process parameters on microprinted cells. This may provide a temporal benchmark for future functional study of microprinted cells that undergo proliferation after undergoing the cell-perturbing processes with the direct cell writing system and subsequent chemical gelation, respectively. 1.3.3. Effect of Direct Cell Writing on Cell-Specific Function
The ability to print 3D structures into a microfluidic system is extremely important as studies indicate how 3D culture can produce appreciable differences in cell behavior (7, 9, 10, 16, 17, 24, 26–30). The aim of this 3-day functional study is to evaluate the implications of a microprinted 3D structure on a measure of hepatocyte-specific function, namely urea synthesis. The experimental setup for this study involves one experimental test sample and one control. The experimental test sample is the microprinting of a 3D hepatocyte/alginate tissue construct. The control used is a 2D hepatocellular monolayer with the same cell number as the experimental test sample. In Fig. 8, urea synthesis is quantitatively detected with a cell-based colorimetric assay, allowing comparison of time-dependent phenotype for microprinted liver cells immobilized within an aqueous gel with controls cultured as a hepatocellular 2D monolayer. This data suggests that hepatocytes encapsulated in a 3D alginate/hepatocyte tissue construct synthesize a higher amount of urea than the same number of hepatocytes cultured as a 2D hepatocyte cell monolayer. This difference is most striking on day 1, when the 3D tissue construct shows a marked urea concentration of 1.90 mg/mL compared with a 2D hepatocyte cell monolayer urea concentration of 0.65 mg/mL. While the marked drop in urea synthesis at day 3 is typical of in vitro culture, the 3D structure confers an ability for increased urea production relative to that in the 2D monolayer case. Therefore, this demonstrates that the microprinting process for microencapsulation of hepatocytes in 3D alginate tissue constructs is compatible with the maintenance of liver-specific function. This improved differentiated function is observed because the direct cell writing of encapsulated cells offers tighter control over the spatial distribution of cells, allowing high cell density or a functional coculture element containing multiple cell types within a 3D tissue construct to be assembled. This creates in vivo-like tissues with a complex set of
232
Chang et al. Hepatocyte Urea Synthesis 2.0
Urea Production (mg/dL)
1.8 1.6
3D Tissue Constructs
1.4
2D Monolayer
1.2 1.0 0.8 0.6 0.4 0.2 0.0 1
2
3
Time (Days) Fig. 8. Results of 3-day urea synthesis study of microprinted alginate tissue constructs of microencapsulated hepatocytes (source: Chang et al. [19] ).
local cues in the form of mechanical forces and biochemical signals, resulting in cell–cell communication, either from direct contact or paracrine signaling that is essential for proper cellular behavior, differentiation, and proliferation, along with the concomitant extracellular matrix and bioactive factors produced by the neighboring cells. Furthermore, optimization of process parameters (e.g., nozzle pressure, motion arm velocity, nozzle tip size, etc.) and material parameters (e.g., biopolymer viscosity, cross-linking agent concentrations, etc.) is conducted to achieve high-fidelity 3D structures and seamless integration onto a microfluidic device.
2. Materials 2.1. Microprinting Machine
1. Drexel’s proprietarily developed multi-nozzle direct cell microprinting machine. 2. Drexel’s proprietarily developed control software. 3. Pneumatic Nozzle (EFD Inc., East Providence, RI). 4. Adaptable nozzle tips, 200, 250 mm inner diameter (EFD Inc.). 5. 15 mL Reservoir tube (EFD Inc.).
2.2. General Cell Culture Material and Supplies
1. Centrifuge. 2. Water bath. 3. 15 mL Test tube. 4. Hemocytometer.
Microprinting of Materials, Cells, and Tissues
233
5. Micropipet and tip, 20, 250, 1,000 mL. 6. Pasteur pipet. 7. Microplate Reader (GENios, TECAN, Durham, NC). 8. Incubator. 2.3. Cell Culture and Chemical Formulation
1. HepG2 liver cells (ATCC, Manassas, MA). 2. Phosphate-buffered solution (VWR International, West Chester, PA). 3. Fetal bovine serum (Sigma, St. Louis, MO). 4. Trypsin (VWR International). 5. Dulbecco’s modified Eagle’s medium (DMEM) (ATCC). 6. Sodium pyruvate. 7. Medium viscosity sodium alginate (Sigma). 8. Calcium chloride (Sigma). 9. Deionized (DI) water. 10. Syringe filters, 0.85, 0.45, and 0.20 mm pore sizes.
2.4. Microfabricated Device Preparation
1. Glass cover slide with etched bifurcated channels (dimensions of the channel pattern are 20.5 mm deep, with branching channel widths ranging from 817 mm down to 8 mm). 2. PDMS substrate (dimensions of the indented tissue chamber are 10 mm × 10 mm × 750 mm). 3. RF Plasma Cleaner (Harrick Plasma, Ithaca, NY). 4. 70% Ethanol. 5. Acetone.
2.5. Dynamic Flow Setup for Drug Metabolism Study
1. Adapter, quick connect female luer to female 10–32 coned, PEEK (Upchurch Scientific, Oak Harbor, WA). 2. Fitting, nanotight, short headless, W/F-142N FER, PEEK, 10–32 (Upchurch Scientific). 3. Adhesive rings, preformed for bonding nanoports to substrate (Upchurch Scientific). 4. Nanoport Assembly, headless, 10–32 coned, 1/16 in OD (Upchurch Scientific). 5. Tubing for flow loops, PEEK, 0.020 IN × 1/16 IN × 5 FT (Upchurch Scientific). 6. Syringe pump, accommodates eight syringes (New Era Pump Systems Inc., Wantagh, NY).
2.6. Assays for Probing Cell Viability and Cell-Specific Function
1. Live–Dead assay (Invitrogen, Carlsbad, CA). 2. Alamar Blue proliferation assay (Invitrogen). 3. QuantiChrome Urea Assay (BioAssay Systems, Hayward, CA).
234
Chang et al.
4. EFC (7-ethoxy-4-trifluoromethylcoumarin) – 25 mg (Invitrogen). 5. HFC (7-hydroxy-4-trifluoromethyl coumarin) – 100 mg (Invitrogen). 6. 384 Well, black walled, clear bottom plates. 7. Microplate filters (520, 535, 595 nm) (GENios).
3. Methods The microprinting protocol presented below is established for a reliable, high-throughput drug screening model. 3.1. Cell-Alginate Solution Preparation
1. To prepare the prepolymer solution, dissolve medium viscosity sodium alginate powder in DI water as a 3.0% (w/v) solution. 2. Mix with magnetic stir bar overnight. 3. Sterilize sodium alginate solution by serial filtration using 0.85, 0.45, and 0.20 mm syringe filters. 4. To prepare the ionic cross-linking solution, dissolve calcium chloride (Sigma) in DI water as a 5.0% (w/v) solution. 5. Prepare HepG2 liver cells by culturing and maintaining in DMEM, supplemented with 10% (w/v) fetal bovine serum, and maintained in the incubator at 5% CO2 and 37°C. 6. Centrifuge prepared HepG2 cell suspension (e.g., 8,000 × g for 5 min) and add 500 mL PBS to break up cell pellet. 7. Resuspend the cells in sodium alginate prepolymer solution with gentle pipetting.
3.2. Preparation of Microfluidic Components for Microprinting
1. Wash glass cover slide successively ×3 with acetone, ethanol, and DI water.
3.3. Plasma Treatment of Microfluidic Components for Microprinting
1. Use tweezers to gently place clean glass cover slide and PDMS substrate onto quartz plate inside plasma chamber.
3.4. Direct Cell Writing into Tissue Chamber of Microfluidic Device
1. Take plasma-treated PDMS layer to the printing system.
2. Leave recently baked PDMS substrate as is (i.e., without washing).
2. Turn on vacuum of plasma chamber for 1 min. 3. Expose glass cover slide and PDMS substrate to RF plasma for 30 s.
2. Align tissue chamber and height. 3. Print alginate + cell suspension into the tissue chamber using toolpath.
Microprinting of Materials, Cells, and Tissues
235
4. Specify and adjust the printing process parameters according to the design specifications. 5. CaCl2 cross-linking agent dispensed at regular intervals using second pneumatic nozzle. 6. Affix glass layer to the PDMS layer to seal the system. 7. Place chips in sterile container/seal container. 3.5. Adaptation of Microprinted Tissue Construct for Pharmacokinetic Study
1. Take sealed container to hood and place into chip holder/ clamps.
3.6. Assessment of Cell Viability for 3D Cell-Alginate Constructs
1. Add 1 mL of Live/Dead assay to each sample.
2. Attach to syringe pump using tubing and nanoports. 3. Place inside incubator.
2. Incubate for 45 min. 3. View under a Leica fluorescence microscope at 40× magni fication 4. In order to quantify the cell viability and assign a cell viability percentage throughout the time course study, each sample is visualized and a live–dead cell count performed at five different locations (three peripheral and two central) for the bulk sample.
3.7. Cell Counting in 3D Cell-Alginate Constructs Using Microplate Reader
1. Use a hemocytometer to obtain the number of cells per milliliter in cell solution. 2. Cell solutions are made at different concentrations ranging from 250,000 to 2,000,000 cells/mL at 250,000 cells/mL intervals. 3. X volume of Almar Blue is added to Y volume of each cell solution concentration in step 2 to give a total volume Z. 4. Each Z volume of different cell concentration/Almar Blue solution is placed into six wells in a 6 × 8 well plate. 5. The well plate is placed in the cytofluorimeter and a reading is given out for each well plate. Measure the concentration of reduced Alamar Blue solution using a microplate reader with an excitation filter wavelength of 535 nm and an emission filter wavelength of 595 nm. 6. A curve is obtained presenting the number of cells vs. the microplate reading. 7. Scaffold is immersed in a known cell solution volume for 1 h in incubator. 8. Scaffold is removed from cell solution and kept in a new culture dish. 9. Almar Blue (fluorescence dye) is added at volume A (depending on scaffold size) to the cells seeded scaffold in the new
236
Chang et al.
culture dish and kept in the incubator for 3 h. This step allows the cells in the scaffold to metabolize the Almar Blue. 10. The Almar Blue solution is removed from the scaffold and the W volume is divided into six wells in a 6 × 8 well plate. Each well contains Z mL of the Almar Blue solution. 11. Fresh medium is added to the scaffold and is replaced back in the incubator 12. The well plate is placed in the microplate reader and a reading is given out for each well plate. 13. The data from step 6 is used to obtain the number of cells in step 12. Steps 8–13 could be repeated over a number of days to obtain the proliferation of cells in the scaffold over a couple of days. 3.8. Assessment of Liver Cell-Specific Urea Synthesis Function
1. 1 mL of QuantiChrome urea assay (BioAssay Systems) solution, a fluorometric indicator of liver cell urea synthetic activity, is added to each printed sample. 2. Replenish media with assay daily. 3. Measure concentration of urea using a microplate reader with an absorbance filter wavelength of 520 nm.
3.9. Drug Metabolism Protocol for Liver Cells
1. Drug substrate EFC = 7-ethoxy-4-trifluoromethyl coumarin – 25 mg of drug comes in a glass amber vial and should be reconstituted in 9.68 mL of DMSO to create a 10 mM stock solution of EFC. 2. Drug product HFC (for standard curve) = 7-hydroxy-4trifluoromethyl coumarin – 100 mg of drug comes in a glass amber vial and is reconstituted in 43.5 mL of DMSO to create a 10 mM stock solution of HFC. 3. Carry out the formation of EFC and HFC stock solutions in the laminar flow hood using good sterile technique and sterile DMSO. Store stock solutions dessicated in the refrigerator. 4. Once the drug and standard stock solutions have been made, they can be diluted to the desired working concentration in cell culture media just prior to an experiment. 120 mM EFC has been found to work quite well for drug metabolism studies in the HepG2 cells. 5. To create 10 mL of a 120 mM EFC test solution: dilute 120 mL of the 10 mM EFC stock solution into 9.88 mL of HepG2 medium and mix well. 6. To create 5 mL of HFC standard solutions ranging from 0.1 to 64 mM: dilute 64 mL of the 10 mM HFC stock solution into 9.936 mL of HepG2 medium containing 120 mM EFC. 7. Serially dilute the 64 mM standard to achieve the desired standards. For example, add 5 mL of 64 mM HFC solution to
Microprinting of Materials, Cells, and Tissues
237
5 mL of 120 mM EFC test solution and mix well. This will be 32 mM HFC standard. Then add 5 mL of the 3 mM HFC standard to 5 mL of 120 mM EFC test solution, and so on. until 0.125 mM HFC standard is reached. Use the 120 mM EFC test solution as the “zero” HFC standard. 8. To start a metabolism experiment: remove medium covering test cells. Replace with desired amount of 120 mM EFC test solution. 9. Measure concentration of urea using a microplate reader with an excitation filter wavelength of 360 nm and an emission filter wavelength of 520 nm. References 1. Chang, R., Nam, J., Holtorf, H., Emami, K., Gonda, S., Jeevarajan, A., Wu, H., Sun, W. (2008) A Case Study on 3D Bioprinted Liver Micro-organ as a Drug Metabolism Model. SME Rapid Technologies/Additive Manufacturing for Medical Applications, Edited by Ola Harrysson, accepted. 2. Powers, M.J., Janigian, D.M., Wack, K.E., Baker, C.S., Stolz, D.B., Griffith, L.G. (2002) Functional behavior of primary rat liver cells in a three-dimensional perfused microarray bioreactor. Tissue Eng 8:499. 3. Langer, R., Vacanti, J.P. (1993) Tissue engineering. Science 260:920. 4. Lysaght, M.J., Reyes, J. (2001) The growth of tissue engineering. Tissue Eng 7:485. 5. Griffith, L.G. (2000) Polymeric biomaterials. Acta Mater 48:263. 6. Griffith, L.G., Naughton, G. (2002) Tissue engineering – current challenges and expanding opportunities in science. Science 295:1009. 7. Powers, M.J., Domansky, K., KaazempurMofrad, M.R., Kalezi, A., Capitano, A., Upadhyaya, A., Kurzawski, P., Wack, K.E., Stolz, D.B., Kamm, R., Griffith, L.G. (2002) A microfabricated array bioreactor for perfused 3D liver culture. Biotech Bioeng 78:257. 8. Ghanem, A., Shuler, M.L. (2000) Combining cell culture analogue reactor designs and PBPK models to probe mechanisms of naphthalene toxicity. Biotechnol Prog 16:334. 9. Shuler, M.L., Ghanem, A., Quick, D., Wong, M.C., Miller, P. (1996) A self-regulating cell culture analog device to mimic animal and human toxicological responses. Biotechnol Bioeng 52:45. 10. Viravaidya, K., Sin A., Shuler M.L. (2004) Development of a microscale cell culture analog
11.
12.
13. 14.
15.
16.
17.
18.
to probe naphthalene toxicity. Biotechnol Prog 20:316. Bratten, C.D.T., Cobbold, P.H., Cooper, J.M. (1998) Single-cell measurements of purine release using a micromachined electroanalytical sensor. Anal Chem 70:1164. Chen, P., Xu, B., Tokranova, N., Feng, X., Castracane, J., Gillis, K.D. (2003) Amperometric detection of quantal catecholamine secretion from individual cells on micromachined silicon chips. Anal Chem 75:518. Meyvantsson, I., Beebe, D.J. (2008) Cell culture models in microfluidic systems. Ann Rev Anal Chem 1:423. Borenstein, J.T., Weinberg, E.J., Orrick, B.K., Sundback, C., Kaazempur-Mofrad, M.R., Vacanti, J.P. (2007) Microfabrication of threedimensional engineered scaffolds. Tissue Eng 13(8):1837. Holmes, T.C., de Lacalle, S., Su, X., Liu, G., Rich, A., and Zhang, S. (2000) Extensive neurite outgrowth and active synapse formation on self-assembling peptide scaffolds. Proc Nat Acad Sci USA 97:6728. Semino, C.E., Merok, J.R., Crane, G.G., Panagiotakos, G., Zhang, S. (2003) Functional differentiation of hepatocytelike spheroid structures from putative liver progenitor cells in three-dimensional peptide scaffolds. Differentiation 71:262. Weaver, V.M., Petersen, O.W., Wang, F., Larabell, C.A., Briand, P., Damsky, C. (1997) Reversion of the malignant phenotype of human breast cells in three-dimensional culture and in vivo by integrin blocking antibodies. J Cell Biol 137(1):231. Chang, R., Nam, J., Sun, W. (2008) Effects of dispensing pressure and nozzle diameter on cell survival from solid freeform fabricationbased direct cell writing. Tissue Eng 14(1):41.
238
Chang et al.
19. Chang, R., Nam, J., Sun, W. (2008) Direct cell writing of 3D micro-organ for in vitro pharmacokinetic model. Tissue Eng C 14(2):157. 20. Chang, R., Nam, J., Holtorf, H., Emami, K., Gonda, S., Jeevarajan, A., Wu, H., Sun, W. (2008) Bioprinting of micro-organ tissue analog for drug metabolism study. 37th COSPAR Scientific Assembly, Montreal, Canada, July 13–20, 2008. 21. Khalil, S., Sun, W. (2007) Biopolymer deposition for freeform fabrication of hydrogel tissue constructs. Mater Sci Eng C 27:478. 22. Khalil, S., Nam, J., Sun, W. (2005) Multinozzle deposition for construction of 3D biopolymer tissue scaffolds. Rapid Prototyping J 11:9. 23. Knight, B., Laukaitis, C., Akhtar, N., Hotchin, N.A., Edlund, M., Horwitz, A.F. (2000) Visualizing muscle cell migration. Curr Biol 10:576. 24. Mooney, D.J., Sano, K., Kaufmann, P.M., Majahod, K., Schloo, B., Vacanti, J.P., Langer, R. (1996) Long-term culture of hepatocytes transplanted on biodegradable polymer sponges. J Biomed Mater Res 37:413.
25. Patz, T.M., Doraiswamy, A., Narayan, R.J. (2006) Three-dimensional direct writing of B35 neuronal cells. J Biomed Mater Res B 78(1):124. 26. Roskelley, C.D., Desprez, P.Y., Bissell, M.J. (1994) Extracellular matrix-dependent tissue specific gene expression in mammary epithelial cells requires both physical and biochemical signal transduction. Proc Nat Acad Sci USA 91:12378. 27. Abbot, A. (2003) Biology’s new dimension. Nature 424:870. 28. Sahai, E., Marshall, C.J. (2003) Differing modes of tumour cell invasion have distinct requirements for Rho/ROCK signalling and extracellular proteolysis. Nat Cell Biol 5(8):711. 29. Wolf, K., Mazo, I., Leung, H., Engelke, K., von Andrian, U.H., Deryugina, E.I. (2003) Compensation mechanism in tumor cell migration: mesenchymal-amoeboid transition after blocking of pericellular proteolysis. J Cell Biol 160(2):267. 30. El-Ali, J., Sorger, P.K. and Jensen, K.F. (2006) Cells on chips. Nature 442:403.
Chapter 14 Microcontact Printing Yunyan Xie and Xingyu Jiang Abstract Microcontact printing (mCP) is a useful technique for transferring certain molecules onto surfaces with high spatial resolution using elastomeric stamps. The stamp for mCP is fabricated by replica molding from a master made by microlithography. After wetting with a type of material as an “ink,” the stamp comes into contact with the substrate. The ink is selectively transferred onto parts of the substrate wherever the stamp makes direct contact, to generate patterns and structures with designated features. Self-assembled monolayers (SAMs) and mCP are useful in many different fields, e.g., in the studies of protein adsorption, cell attachment, and in the construction of sensors. Key words: Microcontact printing, Self-assembled monolayers, PDMS
1. Introduction Microcontact printing (1) is a method that uses an elastomeric stamp to form designed micropatterns of various molecules on the substrates by contact-mediated transfer (Fig. 1). The poly (dimethylsiloxane) (PDMS) stamp for mCP is fabricated by replica molding from a master having relief structures on its surface. The PDMS stamp is coated with one type of molecules; this type of molecules can be transferred onto the surface when the stamp comes into direct contact with the substrate, with high resolution. Patterns can be produced with features as small as ~50 nm using mCP with a PDMS stamp (2). Alkanethiols are commonly used as an ink in mCP to generate patterns. Alkanethiols on gold make up the best characterized self-assembled monolayers (SAMs). In alkanethiols that have the general structure of HS-(CH2)nX, the chemical identities of functional groups at the end of alkyl chains (“X”) give rise to various properties of the substrate when these alkanethiols form SAMs on the gold surface (3). When combining Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_14, © Springer Science+Business Media, LLC 2011
239
240
Xie and Jiang
Fig. 1. Photolithography and microcontact printing: (a) the process of photolithography, generation of small features on surfaces for the master, and replica molding via PDMS (b) the process of microcontact printing of SAMs (reproduced from ref. 14 by permission of Artech House).
SAMs and mCP, we can essentially obtain a surface having patterns of any shape and size with desired surface chemistry (4). As a result, mCP can be used to pattern proteins (5), nucleic acid (6), and cells (7, 22); mCP has also found uses in a variety of applications for surface modification and micro-/nanofabrication (8). For biological applications, besides alkanethiols, some molecules, such as nucleic acids (9), proteins (10), polymers (11), and lipid bilayers (12) can be applied as the ink and be transferred directly onto the substrate using a PDMS stamp. For example, mCP of proteins is a method for delivering biologically active proteins such as enzymes and antibodies to surfaces with submicrometer resolution (8). We mainly describe the printing of SAMs and proteins in this chapter because of their widespread uses.
2. Materials 2.1. Fabrication of Masters 2.1.1. The Design and Generation of the Pattern for the Master
Several software programs are available to choose from to generate the patterns of interest: 1. L-edit software (Tanner EDA, Monrovia, CA). 2. Auto-CAD software (CAD–Art Services, Bandon, OR, http://www.outputcity.com). 3. Macromedia Freehand (Micromedia, San Francisco, CA).
Microcontact Printing 2.1.2. Photolithography
241
1. Photoresist SU-8 (MicroChem, Newton, MA), store away from light, acids, heat, and sources of ignition at a temperature of 4–21°C. 2. SU-8 photoresist developer: 1-methoxy-2-propyl acetate, 98–100% (MicroChem, Newton, MA). 3. Silicon wafer (Siltronix, Neuchatel, CH, Switzerland).
2.1.3. PDMS
1. Poly (dimethylsiloxane) (PDMS) prepolymer and curing agent (Sylgard™ 184 Dow Corning, Midland, MI). 2. Perfluoro-1, 1, 2, 2-tetrahydrooctyltrichlorosilane (Alfa Aesar, Ward Hill, MA).
3. Preparation of Gold Slide 1. Cover glass (Yancheng Jingwei Chemicals Co., JiangSu, China). 2. Titanium (99.99%) and gold (99.99%, both from Johnson Mattey, London, UK). 3.1. Printing of SAMs
1. HS (CH2)15CH3, (Fluka, Gallen, Switzerland), stored at −20°C.
3.2. Incubation of Protein Solution
1. PBS (store at 4°C).
3.3. Cell Culture
1. Dulbecco’s modified essential medium (DMEM, GIBCO, Carlsbad, CA) (store at 4°C).
2. HS (CH2)11(OCH2OCH2) 3OH (C11EG3), (Fluka, Gallen, Switzerland), stored at −20°C. (Degas the solvent with an inert gas, such as argon, prior to preparing the solution. Solutions dissolved in ethanol (store at 4°C) should be used within 2 weeks following preparation.)
2. Fibronectin (BD, Franklin Lakes, NJ) (store at −20°C).
2. Fetal bovine serum (FBS, GIBCO, Carlsbad, CA) (store at −20°C). 3. Penicillin/streptomycin (GIBCO, Carlsbad, CA) (store at −20°C). 4. Trypsin/EDTA (GIBCO, Carlsbad, CA) (store at −20°C). 3.4. Printing of Proteins
1. PDMS stamp. 2. Protein. 3. Clean glass slide.
242
Xie and Jiang
4. Methods Microcontact printing of SAMs of thiols makes use of a PDMS stamp with small features to transfer “ink” molecules such as n-alkanethiols (or other molecules that form SAMs) on a bare surface of a metal (Au, Ag, Cu, Pd, and Pt), metal oxide, glass, or semiconductor (4). A stamp with specific features is prepared by replica molding from a master. The master can be fabricated using microlithographic techniques such as photolithography, micromachining, e-beam writing, or from available relief structures such as diffraction gratings (Fig. 1a) (4). The mixture of PDMS prepolymer is poured over the master having relief structure on its surface, then cured, and peeled off to obtain the stamp for inking (Fig. 1a). The stamp is wetted with the first thiol solution for several seconds and then dried by a stream of air (the “inking” process) (Fig. 1b). The “inked” stamp is brought into contact with the surface of a gold substrate for “printing” between 1 and 60 s; this step produces the first SAM on parts of the surface that came into contact with the stamp (Fig. 1b) (4). Immersion of the substrate in another thiol solution for 2–12 h allows the formation of a second SAM on the substrate (Fig. 1b). Thus, a surface is patterned with designated types of SAMs. The process of printing proteins is similar to mCP of thiols. The stamp needs to be equilibrated in the target solution for the adsorption of sufficient amounts of protein. One of the major problems of mCP of proteins is the loss of the biological function of proteins once printed due to the fact that protein molecules change conformation when adsorbed onto the PDMS surface (15). For efficient transfer of the protein, the receiving surface needs to have properties that make it more favorable for the protein to transfer than to remain on the stamp; the speed at which the stamp is released also affects the efficiency of transfer in the mCP of proteins (13). Here, we use the negative photoresist (SU-8) as an example of photolithography to generate a master on silicon wafer, and the method of printing on a planar surface with a planar stamp to illustrate mCP of SAMs and proteins. We use cell-patterning to introduce one of the most popular applications of mCP, as this process includes most of the steps required in many variants of mCP. 4.1. Preparation of the Stamp of PDMS 4.1.1. The Design and Generation of the Pattern for the Master
1. Use a CAD program such as L-edit to generate the designs for the masters (see Note 1). 2. Produce masks on high-resolution (>3,600 dpi) transparency films by a commercial printer from the CAD files, printed by commercial vendor (CAD_Art Services; or Jing Mei Wei Co. Suzhou, China); or a commercial printer: such as Carl Suss contact printer). Chrome mask can be produced by e-beam writing (see Note 2).
Microcontact Printing 4.1.2. Photolithography
243
1. Pretreat silicon wafers following the recommended product protocol (see Note 3). 2. Spin-coat negative photoresist (SU-8) on a silicon wafer and bake according to the manufacturer’s instructions. 3. Expose the photoresist to UV light through the photomask on a mask aligner. 4. Following exposure, perform a postexposure bake to completely cross-link the exposed portions of the photoresist on the substrate. 5. To develop the features, the substrate can be immersed in or sprayed with the SU-8 Developer to dissolve the unexposed photoresist. 6. Following development, rinse the substrate briefly with isopropyl alcohol and dry it with a gentle stream of air or nitrogen. 7. The resist may be baked between 150 and 200°C on a hot plate or in a convection oven to further crosslink the material. Bake times vary depending on the type of bake process and film thickness. This is an optional step. 8. The surface of the SU-8 master needs to be silanized by exposure to a vapor of perfluoro-1, 1, 2, 2-tetrahydrooctyltrichlorosilane in a vacuum desiccator for ~2 h so that the silane could coat the silicon wafer to prevent the adhesion of the PDMS to the silicon wafer (see Note 4).
4.1.3. Replica Molding
1. Mix the curing agent and silicon rubber base (1:10, see Note 5). 2. Pour the mixture over the master (the final product of 4.1.2.8). 3. Heat the mixture at elevated temperatures (60–80°C) for 2 h (see Note 5). 4. Peel the solidified PDMS stamp off the master and cut to suitable sizes (see Note 6).
4.2. Preparation of Gold-Coated Glass Slide
1. Sonicate the glass in soapy water for 5 min, wash it with distilled water, and dry it under a stream of nitrogen (see Note 7).
4.3. Printing of SAMs
1. Ink a PDMS stamp with desired features in a solution of HS (CH2)15CH3 in ethanol (1–10 mM, see Note 8).
2. Coat glass slides with titanium (5 nm) and gold (20 nm) in an e-beam evaporator (see Note 7).
2. Dry the stamp by a stream of nitrogen or compressed air and bring it in contact with the gold-coated glass surface for 10 s to form the first SAM. Figure 2 illustrates common pitfalls during printing (see Note 8).
244
Xie and Jiang
Fig. 2. Diagram illustrating different types of failure events often encountered during mCP. (a) A PDMS stamp with features of height h, width w, and gap distance d. (b) Lateral collapse occurs when adjacent structures make contact and remain adherent (often takes place when d is too small or the d/h ratio is too small). (c) Collapse happens when the features buckle under the weight of the stamp. (d) Sagging occurs when the roof of the stamp collapses against the substrate (typically when h/d is too small) (reproduced from ref. 15 by permission of the Royal Society of Chemistry).
3. Incubate the gold substrate with HS (CH2)11(OCH2OCH2) OH (C11EG3) (2 mM, in ethanol) for 2–12 h at room 3 temperature (see Note 9). 4.4. Incubation of the Substrate with Protein Solution
1. Wash the gold substrate briefly with 75% alcohol in distilled water and follow by a rinse in PBS.
4.5. Culturing Cells
1. Inoculate cells at a density of 10,000–20,000 per cm2. We typically use DMEM supplemented with 10% FBS and penicillin/streptomycin (1%), and kept it in an atmosphere of 37°C and 5% CO2.
2. Incubate the substrates with fibronectin (25 mg/ml) for 2 h at 37°C.
2. Before patterning, wash the cells with PBS, dissociate the cells from culture plates with trypsin/EDTA, and supplement in DMEM containing 10% serum. After centrifugation at 800 rpm for 2 min, resuspend the cell pellet with suitable density in DMEM containing 10% FBS. 3. Wash the substrate (the final product of Subheading 3.4) with PBS solution (see Note 10). 4. Add the cell to the substrate and culture in an atmosphere of 37°C and 5% CO2. 5. After cell attachment, wash the substrate with DMEM containing 10% FBS to remove the unattached cells (Fig. 3). 4.6. Printing of Proteins
Pattering proteins via mCP is a straightforward technique. It typically does not require the use of a metallic substrate, thus it is more accessible to biological applications. It is not always as reliable as mCP with SAMs; it also depends on the type of proteins printed.
Microcontact Printing
245
Fig. 3. Bovine capillary endothelial cells were allowed to attach to a surface patterned with C11EG3 and HS (CH2)15CH7 (C18) (reproduced from ref. 17 by permission of the American Chemical Society).
1. Prepare the stamp by following the protocol described in Subheading 3.1 2. Expose the PDMS stamp to oxygen plasma (for 60 s at 100 W) (see Note 11) 3. Equilibrate the stamp for <40 min with a solution of the targeted proteins (20–200 mg/ml) at 37°C 4. Rinse and dry the stamp; transfer to a target substrate by printing (see Note 11).
5. Notes 1. Pay particular attention to the size of features when designing them. The elastomeric characteristics of PDMS make it easily deformable to cause defects in the pattern. Microcontact printing will fail if the aspect ratio (h/w, Fig. 2) of the relief features in PDMS is not between 0.2 and 2 (18). If h/w ratio is greater than 2 or w/d ratio is less than 0.5, the structure will collapse during printing (Fig. 2b, c). The adjacent structures would adhere to each other and lead to lateral collapse because of capillary action (Fig. 2b), or they would bend and collapse against the substrate due to pressure (Fig. 2c). On the other hand, if d/h ratio is larger than 20, the sagging of PDMS would take place as a result of compressive forces between the stamp and the substrate (Fig. 2d). 2. Generally, we use a transparency film as the photomask when the feature size is larger than 5 mm. Chrome mask, being much more costly (at least 200 times that of transparency), is only necessary for small (<5 mm) patterns. We can obtain sharp contours of features using chrome masks (23). 3. To obtain maximum process reliability, substrates should be cleaned and dried prior to applying the SU-8 resist. We typically
246
Xie and Jiang
start with cleaning by an organic solvent, or a rinse with dilute acid, and DI water. Sometimes the substrates should be subjected to a piranha etch/clean (H2SO4 and H2O2). To dehydrate the surface, bake at 200°C for 5 min on a hotplate (from the SU-8 resist product protocol). A simple alternative to this tedious process of cleaning is to use oxygen plasma to clean the substrate. 4. This step is critical; skipping this step may result in the adhesion between of PDMS and the silicon wafer and thus lead to the damage of the master. 5. Degas the mixture prior to curing. Sylgard 184 is suitable for replicating master with features >1 mm that are separated by distances ~1–10 times the feature sizes (18). We recommend fabricating with hard stamp to prevent the deformation of the stamp for microcontact printing. A simple method to improve the hardness of the stamp is to increase the ratio between the curing agent and the silicon rubber base to 1:8 or even 1:5; or by applying higher heating temperatures, such as 120°C for 4–8 h. 6. If the stamp is not fabricated from a silicon master, the side that carries the features can be nonplanar. For example, when the center of the surface is more recessed than its edges, it will cause defects on the surface after printing unless additional pressure is applied on the center. Thus, it is necessary to check whether the stamping side is planar before inking. Besides checking with the naked eye, we can use the stamp to ink fluorescent proteins directly on a glass slide to observe whether the fluorescent intensity of pattern is even when we bring pressure on the stamp evenly. 7. The thin layer of titanium improves the adhesion of gold on glass. Evaporation of a layer of titanium or chromium of a similar thickness can also result in improved adhesion of gold on glass (16). Many factors cause structural defects in SAMs (19) such as the cleanliness of the substrate and the methods for preparing the substrates. We recommend the most stringent control in preparing a gold substrate with as few defects as possible. We recommend e-beam evaporators over thermal evaporators or other types of coating devices. 8. The alkanethiols can spread laterally from the regions of direct contact of the stamp to noncontacted areas. At low ink concentration, spreading is independent of contact time; at higher ink concentration, spreading becomes contact time-dependent (20, 21). A concentration of 1–10 mM and inking time of 5–60 s are thus recommended when using HS(CH2)15CH3 for obtaining features of high fidelity. 9. The structure of SAMs formed by mCP of alkanethiol is not equivalent to those formed by adsorption of alkanethiol from
Microcontact Printing
247
solution (19). The structure of SAMs near the edges of printed features is different from that in the center. The molecules in an existing SAM can be exchanged when exposed to solutions containing other thiols. For n-alkanethiols, short chains (n < 12) are more rapidly displaced than long ones (n > 14). 10. Pay attention not to dry the substrate with proteins. A Langmuir–Blodgett (LB) film of proteins is likely to form on the gas–liquid interface in a solution of proteins. In the process of drying, if the LB film deposits on the substrate, the patterns on the surface generated by microcontact printing will be obscured. 11. The surface of PDMS is more hydrophobic than glass and silicon wafer. Protein solutions do not spread evenly (instead, they stand up as drops) if we do not treat the stamp with oxygen plasma first. An important requirement for the successful printing of proteins is to keep the time short between the step of drying the stamp (after its inking) and the step of printing proteins. During the printing operation, the ink concentration, inking time, and contact pressure are vital parameters that affect the quality of the resulting pattern. Additional pressure is sometimes applied to ensure full contact between the stamp and the substrate (24).
Acknowledgments We thank Ms. Wenwen Liu, Mr. Bo Yuan, Mr. Kang Sun, and Mr. Dingbin Liu for the technical assistance. We thank the Chinese Academy of Sciences, the NSFC, MOST, and the Human Frontier Science Program for the financial support. References 1. A. Kumar and G. M. Whitesides (1993) Features of gold having micrometer to centimeter dimensions can be formed through a combination of stamping with an elastomeric stamp and an alkanethiol “ink” followed by chemical etching. Appl. Phys. Lett. 63, 2002. 2. H. W. Li, B. V. O. Muir, G. Fichet and W. T. S. Huck (2003) Nanocontact printing: a route to sub-50-nm-scale chemical and biological patterning. Langmuir 19, 1963. 3. N. L. Abbott, J. P. Folkers, and G. M. Whitesides (1992) Manipulation of the wettability of surfaces on the 0.1- to 1-micrometer scale through micromachining and molecular self-assembly. Science 257, 1380. 4. Y. Xia and G. M. Whitesides (1998) Soft lithography. Annu. Rev. Mater. Sci. 28, 153.
5. G. P. Lopez, H. A. Biebuyck, R. Harter, A. Kumar, and G. M. Whitesides (1993) Fabrication and imaging of two-dimensional patterns of proteins adsorbed on self-assembled monolayers by scanning electron microscopy. J. Am. Chem. Soc. 115, 10774. 6. M. Riepl, K. Enander, B. Liedberg, M. Schaeferling, M. Kruschina, and F. Ortigao (2002) Functionalized surfaces of mixed alkanethiols on gold as a platform for oligonucleotide microarrays. Langmuir 18, 7016. 7. X. Jiang, D. A. Bruzewicz, A. P. Wong, M. Piel, and G. M. Whitesides (2005) Directing cell migration with asymmetric micropatterns. Proc. Natl. Acad. Sci. USA 102, 975. 8. A. Bernard, J. P. Renault, B. Michel, H. R. Bosshard, and E. Delamarche (2000)
248
9.
10.
11.
12.
13.
14.
15. 16.
Xie and Jiang Microcontact printing of proteins. Adv. Mater. 12, 1067. S. A. Lange, V. Benes, D. P. Kern, J. K. H. Horber, and A. Bernard (2004) Microcontact printing of DNA molecules. Anal. Chem. 76, 1641. A. Bernard, E. Delamarche, H. Schmid, B. Michel, H. R. Bosshard, and H. Biebuyck (1998) Printing patterns of proteins. Langmuir 14, 2225. D. Arrington, M. Curry, and S. C. Street (2002) Patterned thin films of polyamidoamine dendrimers formed using microcontact printing. Langmuir 18, 7788. J. S. Hovis and S. G. Boxer (2001) Patterning and composition arrays of supported lipid bilayers by microcontact printing. Langmuir 17, 3400. A. Bernard, D. Fitzli, P. Sonderegger, E. Delamarche, B. Michel, H. R. Bosshard, and H. Biebuyck (2001) Affinity capture of proteins from solution and their dissociation by contact printing. Nat. Biotechnol. 19, 866. X. Jiang (2008) Surface patterning for controlling cell-substrate interaction, in Micro and Nanoengineering of the Cell Microenvironment: Technologies and Applications (A. Khademhosseini, J. Borenstein, M. Toner, S. Takayama, eds), Artech House, Norwood, MA, pp.33–51. S. A. Ruiz and C. S. Chen (2007) Microcontact printing: a tool to pattern. Soft Matter 3, 168–177. J. A. Rogers, Z. Bao, K. Baldwin, A. Dodabalapur, B. Crone, V. R. Raju, V. Kuck, H. Katz, K. Amundson, J. Ewing, and P. Drzaic (2001) Paper-like electronic displays: Large-area rubber stamped plastic sheets of electronics and microencapsulated
17.
18.
19.
20.
21.
22.
23. 24.
electrophoretic inks. Proc. Natl. Acad. Sci. USA 98, 4835. X. Jiang, R. Ferrigno, M. Mrksich, and G. M. Whitesides (2003) Electrochemical desorption of self-assembled monolayers noninvasively releases patterned cells from geometrical confinements. J. Am. Chem. Soc.125, 2366. E. Delamarche, H. Schmid, H. A. Biebuyck, and B. Michel (1997) Stability of molded polydimethylsiloxane microstructures. Adv. Mater. 9, 741. J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, and G. M. Whitesides (2005) Selfassembled monolayers of thiolates on metals as a form of nanotechnology. Chem. Rev. 105, 1103. R. B. A. Sharpe, D. Burdinski, J. Huskens, H. J. W. Zandvliet, D. N. Reinhoudt, and B. Poelsema (2004) Spreading of 16-mercaptohexadecanoic acid in microcontact printing. Langmuir 20, 8646. Y. N. Xia and G. M. Whitesides (1995) Use of controlled reactive spreading of liquid alkanethiol on the surface of gold to modify the size of features produced by microcontact printing. J. Am. Chem. Soc. 117, 3274. Y. Li, B. Yuan, H. Ji, D. Han, S. Chen, F. Tian, and X. Jiang (2007) A method for patterning multiple types of cells by using electrochemical desorption of self-assembled monolayers within microfluidic channels. Angew. Chem. Int. Ed. 46, 1094 K. Sun, Z. Wang, and X. Jiang (2008) Modular microfluidics for gradient generation. Lab Chip 8, 1536. L. Libioulle, A. Bietsch, H. Schmid, B. Michel, and E. Delamarche. (1999) Contact-inking stamps for microcontact printing of alkanethiols on gold. Langmuir 15, 300.
Chapter 15 Micromolding for the Fabrication of Biological Microarrays Ashley L. Galloway, Andrew Murphy, Jason P. Rolland, Kevin P. Herlihy, Robby A. Petros, Mary E. Napier, and Joseph M. DeSimone Abstract The PRINT® (pattern replication in non-wetting templates) process has been developed as a simple, gentle way to pattern films or generate discrete particles in arrays out of either pure biological materials or biomolecules encapsulated within polymeric materials. Patterned films and particle arrays can be fabricated in a wide array of sizes and shapes using Fluorocur® (a UV-curable perfluoropolyether polymer) from the nanometer to micron scale. Key words: PRINT® process, Molding, Biological microarray, Nanoarray fabrication, Proteins, Oligonucleotides, Gene delivery, Nanoparticles
1. Introduction Arrays of proteins and nucleic acids have found many applications from fundamental research to medical devices and drug delivery (1, 2). Proteins and DNA arrays used in high-throughput analyses allow rapid analysis of gene or protein expression either for diagnostics or fundamental biological research (1). Also, patterned 2D and 3D films of biomolecules can be used as substrates for cell growth and tissue engineering (3). Novel techniques developed for material science have now found utility in these life sciences applications with the ability to produce micro- and nanopatterned films and arrays. Beyond patterned surfaces, adaptations of soft lithography have recently been developed that can lead to the formation of discrete particles with controlled size and shape. Even arrays of pure biological molecules or mixtures of biological molecules in polymers can be used to generate discrete, size- and shape-controlled particles for drug delivery (4). Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_15, © Springer Science+Business Media, LLC 2011
249
250
Galloway et al.
1.1. Background of Patterned Surfaces
The majority of soft lithographic methods used for preparing arrays of biological materials employ polydimethylsiloxane (PDMS) molds. Crosslinked PDMS is a good material for these applications due to its low toxicity, high flexibility, gas permeability, and low surface energy. Microcontact printing or microtransfer molding uses patterned PDMS brought into contact with a polymer, protein, or DNA solution followed by transfer onto a suitable substrate (1, 3). The channels or cups of a mold can be filled using capillary forces between the fluid and the mold, which creates a pattern of the molded material on a suitable substrate after evaporation or curing step (3, 5). Also of note are the step and flash techniques using fluorinated glass stamps to create a patterned film of DNA or proteins entrapped in a curable matrix, which is then subjected to oxygen plasma to remove excess polymer between the features (6).
1.2. Breakthrough for Particle Generation
While PDMS molds are utilized with great success in preparing patterned arrays, alternative mold materials, such as crosslinked perfluoropolyethers (PFPE) (Fluorocur® molds, Liquidia Tech nologies, Inc.) can provide distinct benefits in creating patterned films and discrete particle arrays. PFPE materials (7, 8) have three clear-cut advantages over PDMS based materials: (1) PFPE’s have an extremely low surface energy, thus allowing for complete filling of the mold cavities without wetting the land area above the cavities. This results in the formation of discrete particles without an inter-connecting flash layer. (2) PFPEs are non-swelling to organic liquids; therefore, providing greater flexibility in design of particle arrays, enabling one to engineer in surface chemistries, degradation characteristics, and deformability. (3) PFPEs naturally have Teflon™-like characteristics allowing for an easy removal of the particle arrays from the PFPE mold. These charac teristics, along with Fluorocur resin’s ability to replicate features down to the nanometer scale make it an ideal material for use in molding. The PRINT® (pattern replication in non-wetting templates) platform technology utilizes the molding advantages of PFPE to offer precise engineering of feature size, shape, composition, and functionality (4, 9–16) Unlike many other nanoparticle and nanoarray fabrication techniques, the PRINT process is versatile and mild enough to be compatible with virtually any biological materials. In fact, the PRINT technology is so gentle, that it can be used to mold polymer micelles, carbon nanotubes, and even biological material such as adenovirus with 0.4 nm resolution (10). Moreover, the inherent mild conditions of the PRINT process can be leveraged to fabricate nanoarrays of pure biological materials (4) or blends of GRAS materials and proteins or oligonucleotides (9, 14, 17). The beauty of the PRINT technology is its flexibility which allows for the solidification of molded materials by lyophilization,
Micromolding for the Fabrication of Biological Microarrays
251
solvent evaporation, photocuring, thermal curing, or crystallization inside the individual cavities of a Fluorocur mold (4, 9–17). These discrete objects inside a mold can be harvested onto a substrate, yielding a nanoarray of biological or organic materials. The PRINT technology has been used to make discrete particles in an array out of pure biological materials such as insulin and albumin (4). Protein dissolution experiments are identical between native protein and molded protein indicating no crosslinking or aggregation occurring during particle fabrication in the PRINT process (4). Nanoarrays of particles containing biological material as a cargo have been generated using the PRINT process. Particles in these arrays are typically fabricated from biocompatible materials such as poly(ethylene glycol) (PEG), poly(lactic-co-glycolic acid) (PLGA), or poly(vinyl pyrrolidone) (PVP) combined with a bioactive oligonucleotide (9). Oligonucleotides including ssDNA, pDNA, siRNA as well as proteins have been successfully entrapped as a cargo within PRINT particle arrays. Proof of this encapsulation can be observed by fluorescence microscopy when fluorescently labeled proteins or oligonucleotides are molded in these types of polymer matrices. Functional cell-based assays can then be used to test for nucleic acid activity. Moreover, the PRINT process allows for direct conjugation (17, 18) of ligands such as proteins, antibodies, carbohydrates, and peptides to the surface of each individual particle in the array. For direct conjugation, nanoarrays must be designed to incorporate a reactive species, such as a primary amine, on the surface of each discrete particle. Surface primary amines can easily be conjugated to a variety of electrophilic molecules, often sold as kits from biochemical vendors. Alternatively, avidin can be attached to the reactive species on the particle’s surface followed by exposure to any biotinylated species. Potential applications of nanoarrays of precisely molded bioactive materials are virtually endless (17) Surface ligands can be used to illicit an immune response, to probe cell-specific interactions, or study cell adhesion. In fact, these surface-modified nanoarrays can be used to probe everyday biological interactions and recognition events. In healthcare, these biological arrays can be used to develop treatments by pathogen detection and characterization, to evaluate and diagnose disease susceptibility and progression with protein–protein and protein–ligand interactions, and to discover potential therapeutic targets faster and more accurately than present techniques (17). In genetics, arrays of biomolecules or bioactive ligands could be used to determine special biomarkers in serum or urine for personalized medicine; genetic identification, and even forensics. If the arrays are harvested, the resulting monodisperse, discrete particles can be used as non-viral gene delivery vectors or shape-specific biosensors (5).
252
Galloway et al.
2. Materials 2.1. Materials Needed for Mold Fabrication
Fluorocur® resin was received from Liquidia Technologies, Inc. 2,2-diethoxyacetophenone photoinitiator was purchased from Aldrich and the ELC-4001 UV lamp (l = 365 nm at >20 mW/ cm2) was purchased from Electro-lite corporation.
2.2. Particle Components, Solvents, and Reagents for Monomer Synthesis
Major components of particle synthesis include trimethyloylpropane ethoxylate triacry late (MW = 428 g/mol) (Aldrich); poly(ethylene glycol) monomethylether monomethacrylate (MW = 1,000 g/mol) (Polysciences); Poly(ethylene glycol) diacrylate (MW = 400 g/mol) (Aldrich); bis(ethyl methacrylate)disulfide was prepared as previously described (19); 2-aminoethylmethacrylate hydrochloride (AEM. HCl) (Aldrich); and acryloxyethyltrimethylammonium chloride (Aldrich). Components such as fluorescein-o-acrylate (Aldrich) and Polyfluor 570 (Polysciences) are incorporated for a fluorescent tag on the particles. DEAP, or 2,2-diethoxyacetophenone (Aldrich) is used as a photoinitiator. In preparing reactive monomers, 1,1¢ carbonyl diimidazole (Aldrich) and Poly(ethylene glycol) monomethacrylate (MW = 485 g/mol) (Polysciences) are used. Common solvents for particle synthesis include 2-propanol (Acros Organics), DMSO (Acros Organics), PBS (Ambion), Acetonitrile (Fisher Scientific), Methanol (Fisher Scientific), and N,N-dimethylformamide (Acros Organics). Biomolecules used as cargo have included Albumin (Sigma), fluorescein-isothiocyanatelabeled avidin (68 kDa) (Sigma), Cy-3-labeled avidin (68 kDa) (Sigma), and DNA oligonucleotide 18 mer, sequence GCT ATT ACC TTA ACC CAG containing a 3¢ fluorescein label (synthesized at Lineberger Comprehensive Cancer Center Nucleic Acids Core Facility at UNC).
2.3. Harvesting Materials
Filters used to filter solvents or particles include 0.22-mm PTFE filter (Millipore) and 25 mm pore size filters (Fisher Scientific). Particles can be collected on 100 nm pore size PVDF centricon tubes (Millipore). Removing particles from the mold can be accomplished with poly(cyanoacrylate) (Aldrich) or mechanical force with Large Glass Microscope Slides (Fisher Scientific) and solvents such as acetone (Fisher Scientific), chloroform (Fisher Scientific), or water (Ambion).
2.4. Reagents for Surface Treatment of Particles
Particles with reactive end groups can be post-treated with AlexaFluor 488-labeled streptavidin (Invitrogen) or ethanolamine (Aldrich). FITC-biotin (Invitrogen) can then be reacted with avidin on the particle surface.
2.5. Reagents for Particle Degradation
If the particle composition includes disulfides, they can be cleaved with dithiothreitol (Acros).
Micromolding for the Fabrication of Biological Microarrays
253
3. Methods 3.1. Description of the PRINT Process (9)
As mentioned above, the PRINT process (or Particle Replication in Non-Wetting Templates) is the first general method to accurately and gently mold size- and shape-specific features from most any material (9). This process begins with the fabrication of a Fluorocur mold from a patterned surface (Fig. 1), such as a nanopatterned silicon master generated by traditional imprint lithography techniques (9). Fluorocur molds can then be filled with a variety of organic materials comprised solely of biomolecules, mixtures of biomolecules and GRAS materials, or even polymers alone (Fig. 2). This particular example describes the fabrication of nanoarrays comprised of poly(ethylene glycol)-based materials.
Fig. 1. (a) Fluorocur resin is poured over a patterned silicon wafer. (b) The resin is solidified on the wafer. (c) The mold is peeled away revealing a perfect imprint of the patterned silicon wafer. (d) SEM image of a patterned silicon wafer. (e) SEM image of a Fluorocur mold generated from the patterned silicon wafer on the left.
Fig. 2. (a) First, the material to be molded is spread evenly across the mold using a roller. (b) The excess material is removed from the mold surface, leaving filled cavities without any material on the land area of the mold. (c) The liquid in the mold cups is solidified. (d) Discrete particles are harvested from the mold, producing a nanoarray on a sacrificial adhesive layer. (e) The nanoarray can be disrupted by dissolving the adhesive layer, leaving monodisperse nanoparticles in solution.
254
Galloway et al.
3.1.1. Fabrication of a Fluorocur Mold (11)
1. 20 mL of Fluorocur® resin (Liquidia Technologies, Inc.) combined with a photoinitiator, such as 0.1 wt% 2,2-diethoxyacetophenone, is poured over the surface of an 8 in. silicon master containing 200 × 200 nm cylindrical posts and allowed to wet the surface completely. 2. The resin, now completely covering the master, is purged with nitrogen gas for 3 min and then subjected to UV light (l = 365 nm at >20 mW/cm2) while remaining under a nitrogen gas purge for two additional minutes. 3. The cured resin can be easily peeled off of the silicon master, yielding a mold with a perfect imprint of the patterned surface (Fig. 1).
3.1.2. Fabrication of Particles in an Array from a Fluorocur Mold (11, 13, 20)
1. A mixture containing 67% trimethyloylpropane ethoxylate triacrylate (MW = 428 g/mol), 20 wt% poly(ethylene glycol) monomethylether monomethacrylate (MW = 1,000 g/mol), 10 wt% 2-aminoethylmethacrylate hydrochloride (AEM HCl), 2 wt% fluorescein-o-acrylate, and 1 wt% 2,2-diethoxyacetophenone was diluted to 10 wt/vol% solution in 2-propanol. 2. This solution is then deposited onto a mold containing 200 × 200 nm cylindrical cavities, laminated with a polymer sheet, and the polymer sheet is removed to yield filled mold cavities. 3. The filled Fluorocur mold is purged with nitrogen for 2 min followed by exposure to 365 nm UV irradiation at >20 mW/ cm2 for an additional 2 min. 4. After curing, the particles are removed from the mold using a medical adhesive such as poly(cyanoacrylate). Images of nanoparticle arrays (9) of various sizes composed of polyethylene glycol based materials are shown in Fig. 3.
3.2. Methods for Fabrication of Protein Arrays and Particles (4)
1. A 25 wt% of albumin was prepared by dissolving 25 mg lyophilized powder with 75 mL H2O. 2. This solution was spotted directly onto a Fluorocur mold (patterned with 200 × 200 nm cylinders) at the contact point of the patterned molded and an unpatterned polyethyleneterephthalate film affixed at the nip point on a laminator. 3. The solution was laminated between the mold and the PET sheet at a speed 0.25 ft/min and a pressure of 50 psi. The PET and mold were separated at the far side of the nip. 4. The solvent was allowed to evaporate from the filled mold by maintaining the arrays exposure to the atmosphere. 5. To harvest discrete particles from the protein array in the Fluorocur mold, 2 mL of a non-solvent (in this case chloroform) was placed on the mold surface and the particles were removed by slowly scraping the surface with a glass slide.
Micromolding for the Fabrication of Biological Microarrays
255
Fig. 3. PEG-based particles in various sizes and shapes (9). (a) 200 × 200 nm trapezoidal particles. (b) 200 × 800 nm rod-like particles. (c) Conical particles with a 500 nm base and a 50 nm tip. (d) 3-mm-sized arrows.
6. To harvest the discrete protein particles in an array on a glass substrate, a small quantity of cyanoacrylate was laminated between the surface of the filled mold and a glass slide. 7. After the polymerization of cyanoacrylate was completed, the mold was slowly peeled from the surface of the glass yield a patterned array of discrete protein particles. 8. The particles can then be released from the array by dissolving the adhesive (see Fig. 4). 3.3. Fabrication of Organic Nanoparticle Arrays with Encapsulated Protein
Encapsulation of biomolecular cargo such as proteins (9, 14, 17) is straightforward using the PRINT process. In one example, avidin is blended with PEG diacrylate and solidified to yield a protein embedded in a biocompatible polymer matrix. To make this nanoarray (9) follow the steps outlined below: 1. Dissolve 1 mg of Cy-3-labeled avidin (68 kDa) in 1 mL of water. 2. 50 mL of this solution is then mixed with 20 mL of PEG400 diacarylate monomer containing 1% photoinitiator. 3. This mixture is concentrated to remove the water completely. 4. Since the concentration in step 3 produces a cloudy suspension, 20 mL of water, the minimum amount of water necessary to obtain a clear solution, is added back. 5. The PEG solution containing avidin is then thinly spread across a mold containing cones with a 500 nm base and a 50 nm tip.
256
Galloway et al.
Fig. 4. Albumin particles harvested on a cyanoacrylate layer.
Fig. 5. Images of conical poly(ethylene glycol) particles with a 500 nm base and a 50 nm tip. (a) Fluorescence image of Cy-3-labeled avidin encapsulated within the PEG matrix. (b) Fluorescence image of FITC-labeled Biotin associated with the array. (c) Overlap of the two images in (a) and (b) showing that the avidin and biotin are co-localized.
6. The mold is sandwiched against a fluorinated surface and pressure (100 N/cm2) is applied to squeeze out any excess protein/PEG glycol solution. 7. While remaining under pressure, the sample is purged with nitrogen and exposed to UV light at 365 nm for 10 min. 8. Removal of the mold left the avidin-containing nanoarray on the fluorinated substrate (Fig. 5a). 3.3.1. Activity Retained of Avidin Encapsulated in Nanoparticle Arrays (9, 14)
In an effort to demonstrate that the avidin protein maintained its integrity, an array containing Cy-3-labeled avidin-containing poly(ethylene glycol) particles were fabricated from a mixture of PEG acrylates) (3). 1. 70% PEG400 diacrylate and 30% PEG1000 monomethacrylate were mixed with Cy3-labeled avidin and molded as described above (Fig. 5a).
Micromolding for the Fabrication of Biological Microarrays
257
2. Fluorescein-isothiocyanate-labeled biotin was exposed to the nanoarray of avidin-containing poly(ethylene glycol) particles for 30 min and then washed well with water to remove any unbound biotin. 3. As evident in Fig. 5b, c, the biotin binds with the avidin in the nanoarray and is only localized on the nanoparticles not in the area in between particles. 3.4. Conjugation of Avidin to Nanomolded Poly(Ethylene Glycol) Particles (17, 18)
In addition to fabricating nanoarrays out of pure protein or mixtures of protein and polymers, it is also possible to attach proteins to the surface of biocompatible nanoarrays. In order to fabricate these nanoarrays, a suitable biocompatible material should be chosen that contains a reactive end group that can serve as a point of attachment after array fabrication. In this example, arrays of triacrylate-based particles containing a reactive carbonyl imidazole group are fabricated as described below. 1. Poly(ethylene glycol) monomethacrylate (MW = 485 g/mol) is treated with 1,1¢ carbonyl diimidazole to produce the reactive monomer, PEG485 carbonyl imidazole monomethacrylate (18) as shown in Fig. 6. 2. A solution of 59 wt% poly(ethylene) glycol428 triacrylate is mixed with 40 wt% PEG485 carbonyl imidazole monomethacrylate and 1% 2, 2¢-diethoxyacetophenone photoinitiator. 3. This mixture is then spotted onto a Fluorocur mold with 200 × 200 nm cavities and covered with a plastic sheet. 4. The plastic sheet is then removed leaving a mold with filled cavities. 5. Next, the filled mold is subjected to a nitrogen purge for 2 min followed by UV exposure (l = 365 nm, 20 mW/cm2) for an additional 2 min. 6. The discrete particles in the mold can be removed onto a cyanoacrylate harvesting layer by first placing a drop of cyanoacrylate onto a glass slide followed by lamination of the slide to the mold surface. 7. After allowing the cyanoacrylate to polymerize (5 min), the mold is removed leaving an array of nanoparticles on the glass slide.
Fig. 6. Reaction of poly(ethylene glycol) monomethacrylate (MW = 485 g/mol) with carbonyl diimidazole to produce the reactive monomer, PEG485 carbonyl imidazole monomethacrylate.
258
Galloway et al.
8. If desired, the particles can remain in the array for surface functionalization or be released into solution by exposure to acetone. 9. The resulting 200 × 200 nm particles dispersed in DMSO are then exposed to a 0.7 mg/mL solution of Alexa-Fluor 488-labeled streptavidin in PBS (18). 10. After stirring for 14 h at room temperature, the particle solution was treated with ethanolamine to quench any remaining reactive end groups. 11. The particle solution was diluted 3× in deionized water, filtered through a 25 mm pore, and collected on a 100 nm pore size filter membrane. 12. After concentration via centrifugal filtration, the particles are resuspended in fresh water and imaged by fluorescence microscopy to see if the avidin conjugation was successful. 13. Alternatively, the nanoarray could be treated directly with the protein solution for a period of time, and then wash with water to obtain conjugation of the protein to the nanoparticle surface while remaining in the array. 14. If labeling of one side only is desired, it is possible to treat the single exposed side of the nanoarray while the individual particles remain in the Fluorocur mold. Treat the exposed side of the array with the protein solution for a period of time and wash with water to obtain an array of particles with one labeled side. It is of course possible to collect these particles in solution as mentioned earlier in this section. 3.5. Encapsulation of Oligonucleotides in Organic Nanoparticle Arrays
1. A solution of UV-curable monomers and a fluorescently labeled DNA oligonucleotide (18 mer, sequence GCT ATT ACC TTA ACC CAG containing a 3¢ fluorescein label) was prepared by adding 2 mg of DNA in 2 mL of H2O to a mixture of 13.65 mg of bis(ethyl methacrylate)disulfide, 1.53 mg of acryloxyethyltrimethylammonium chloride, 0.075 mg of Polyfluor 570, 0.15 mg 2,2¢-diethoxyacetophenone, 2.34 mg of acetonitrile, 2.34 mg of methanol, 9.5 mg of N,Ndimethylformamide, and 0.4 mg of H2O. 2. The mixture was spotted directly onto a 2 × 2 × 1 mm patterned Fluorocur mold and then covered with a plastic film. 3. The film was removed from the mold to leave filled mold cavities. 4. The filled mold was then subjected to UV light (l = 365 nm) for 2 min under a nitrogen purge that was passed through a gas scrubber filled with N,N-dimethylformamide prior to entering the curing chamber.
Micromolding for the Fabrication of Biological Microarrays
259
Fig. 7. Percent release of oligonucleotide from degradable disulfide PRINT™ particles.
5. The isolated polymeric particles containing DNA can be removed from the array in the mold by placing ~0.4 mL of filtered acetone (0.22-mm PTFE filter) and scrapping the surface of the mold gently with a glass slide. 6. The particle suspension was transferred to a centrifuge tube, the particles were pelleted, the supernatant removed, and the particles were dried under vacuum. 7. The particles (1.44 mg) were then suspended in 1 mL of H2O, vortex rigorously, and pelleted out to purify. 8. The DNA can be released from the particles by treatment with 0.1 M dithiothreitol in PBS with 1–2 h. The rate of release in the absence of reductant is minimal (see Fig. 7).
Acknowledgments The authors would like to acknowledge outstanding scientific collaborations between the Carolina Center of Cancer Nanotechnology Excellence, Liquidia Technologies, and the Chemistry Department at the University of North Carolina, Chapel Hill. Much of this work was carried out by a team of exceptional postdoctoral fellows and graduate students. This work was supported by NIH U54-CA-119343 (the Carolina Center of Cancer Nanotechnology Excellence), NIH F32-CA-123650 (Ruth L. Kirschstein National Research Service Award), PPG P01-GM059299-07 (Pharma codynamics of Genes and Oligonucleotides), STC Program of the NSF (CHE-9876674), the William R. Kenan Professorship at the University of North Carolina at Chapel Hill, and through a supported research agreement with Liquidia Technologies.
260
Galloway et al.
References 1. Voldman, J.; Gray, M. L.; Schmidt, M. A. Microfabrication in Biology and medicine. Annu. Rev. Biomed. Eng. 1999, 1, 401–425. 2. Truskett, V. N.; Watts, M. P. C. Trends in imprint lithography for biological applications. Trends Biotechnol. 2006, 24(7), 312–317. 3. Kane, R. S.; Takayama, S.; Ostuni, E.; Ingber, D. E.; Whitesides, G. M. Patterning proteins and cells using soft lithography. Biomaterials 1999, 20, 2363–2376. 4. Kelly, J. Y.; DeSimone, J. M. Shape-specific, monodisperse nano-molding of protein particles. J. Am. Chem. Soc. 2008, 130(16), 5438–5439. 5. Torres, C. M. S.; Zankovych, S.; Seekamp, J.; Kam, A. P.; Cedeno, C. C.; Hoffman, T.; Ahopelto, J.; Reuther, F.; Pfeiffer, K.; Bleidiessel, G.; Gruetzner, G.; Maximov, M. V.; Heidari, B. Nanoimprint lithography: an alternative nanofabrication approach. Mater. Sci. Eng. C. 2003, 23, 23–31. 6. Glangchai, L. C.; Caldorera-Moore, M.; Shi, L.; Roy, K. Nanoimprint lithography based fabrication of shape-specific enzymaticallytriggered smart nanoparticles. J. Control Release 2008, 125, 263–272. 7. Rolland, J. P.; Hagberg, E. C.; Denison, G. M.; Carter, K. R.; DeSimone, J. M. High resolution soft lithography: Enabling materials for nanotechnologies. Angew Chem. Int. Ed. Engl. 2004, 43(43), 5796–5799. 8. Rolland, J. P.; Van Dam, R. M.; Schorzman, D. A.; Quake, S. R.; DeSimone, J. M. Solventresistant photocurable “liquid teflon” for microfluidic device fabrication. J. Am. Chem. Soc. 2004, 126, 2322–2323. 9. Rolland, J. P.; Maynor, B. W.; Euliss, L. E.; Exner, A. E.; Denison, G. M.; DeSimone, J. M. Direct fabrication and harvesting of monodisperse, shape-specific nanobiomaterials. J. Am. Chem. Soc. 2005, 127(28), 10096–10100. 10. Maynor, B. W.; Larue, I.; Hu, Z.; Rolland, J. P.; Pandya, A.; Fu, Q; Liu, J.; Spontak, R. J.; Sheiko, S. S.; Samulski, R. J.; Samulski, E. T.; DeSimone, J. M. Supramolecular nanomimetics: Replication of micelles, viruses, and other
11.
12. 13.
14.
15.
16.
17. 18.
19.
20.
naturally occurring nanoscale objects. Small 2007, 3(5), 845–849. Gratton, S. E. A., Pohlhaus, P. D.; Lee, J.; Guo, J.; Cho, M. J.; DeSimone, J. M. Nanofabricated particles for engineered drug therapies: A preliminary biodistribution study of PRINT™ nanoparticles. J. Control Release 2007, 121(1–2), 10–18. http://www.chem.unc.edu/people/faculty/ desimone/group/research_print.htm. Gratton, S. E. A.; Ropp, P. A.; Pohlhaus, P. D.; Luft, J. C.; Madden, V. J.; Napier, M. E.; DeSimone, J. M. The effect of particle design on cellular internalization pathways. Proc. Nat. Acad. Sci. USA 2008, 105(33), 11613–11618. Euliss, L. E.; DuPont, J. A.; Gratton, S. E. A.; DeSimone, J. M. Imparting size, shape, and composition control of materials for nanomedicine. Chem. Soc. Rev., 2006, 35, 1095–1104. Herlihy, K. P.; Nunes, J.; DeSimone, J. M. Electrically driven alignment and crystallization of unique anisotropic polymer particles. Langmuir 2008, 24, 8421–8426. Petros, R. A.; Ropp, P. A.; DeSimone, J. M. Reductively labile PRINT particles for the delivery of doxorubicin to HeLa cells. J. Am. Chem. Soc. 2008, 130(15), 5008–5009. PCT# WO2008\045486, Liquidia Technologies, Inc. “Nanoparticle Compositions for the Controlled Delivery of Nucleic Acids.” International Patent application # WO 2008/127455, Liquidia, Technologies, Inc. “Nanoarrays and methods and materials for fabricating same.” Li, Y.; Armes, S. P. Synthesis and chemical degradation of branched vinyl polymers prepared via ATRP: use of a cleavable disulfidebased branching agent. Macromolecules 2005, 38, 8155–8162. Gratton, S. E. A., Williams S. S.; Napier, M.; Pohlhaus, P. D.; Zhou, Z.; Wiles, K. B.; Maynor, B. W.; Shen, C.; Olafsen, T.; Samulski, E. T.; Desimone, J. M. The Pursuit of a scalable nanofabrication platform for use in the material and life science applications. Acc. Chem. Res. 2008, 41(12), 1685–1695.
Chapter 16 Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition Yaseen Elkasabi and Joerg Lahann Abstract This book chapter discusses recent advances in the fabrication of microscale surface patterns using chemical vapor deposition polymerization. Reactive poly(p-xylylene) (PPX) coatings are useful for their ability to immobilize specific biomolecules, as determined by the PPX functional group. PPXs can either be modified postdeposition, or they can be patterned onto a substrate in situ. Specific methods discussed in this progress report include microcontact printing, vapor-assisted micropatterning in replica structures, projection lithography-based patterning, and selective polymer deposition. Key words: Bioarrays, Chemical vapor deposition, Immobilization, Micropatterns, Surface engineering
1. Introduction Controlled surface engineering has been a long-standing challenge in the development of bioarrays. Moreover, miniaturized diagnostic systems, such as micro-total analysis systems (mTAS) (1), cell-based assays (2), microseparators for proteins (3, 4), DNA (5), and polysaccharides (6), often require universally applicable surface engineering protocols. Some general surface modification techniques have proven to be versatile in alleviating adverse biological effects. One technique that is widely used to tailor the interfacial properties of metals, metal oxides, and semiconductor surfaces is the use of self-assembled monolayers (SAMs) (7). Based on the terminal functional groups exposed on the surface of a SAM, the reactivity of the surface can be varied. SAMs have been used for the direct immobilization of DNA, polypeptides, and proteins (8). However, the use of SAMs is limited due to the relative chemical instability of the monolayer and the specificity of
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_16, © Springer Science+Business Media, LLC 2011
261
262
Elkasabi and Lahann
the substrates. In contrast, the above-mentioned applications require robust surface chemistry. Extensive efforts have been made to create topological surface patterns using printing methods, such as dip/pen lithography (9), patterning via scanning probes (10), imprinting lithographies (11, 12), or soft lithography (13, 14). Included within soft lithography are micromoulding in capillaries (15), microcontact printing (16), replica molding (17), microtransfer molding (18), solventassisted micromoulding (19), and capillary force lithography (20). Soft lithographical methods rely on the use of elastomeric stamps or replica structures to transfer material from a solution onto a surface. Patterned substrates created using shadow masks included a range of different materials, such as semiconductors (21–23), organic metals (24), polymers (25), biomaterials (26), or cells (27–29). Surface patterns have also been fabricated using lithographical techniques, on the basis of light (30), X-rays (31), electrons (32), ion beams (33), or atoms (34). Furthermore, patterned substrates can be incorporated into microfluidic systems and subsequently used for high-throughput proteomics applications, pharmaceutical screening of cellular assays, or cell-based biosensors. Methods for creating patterns in microfluidic channels previously depended on patterning of a flat substrate, which is then sealed to the microchannel. Some specific processes utilize microfluidic patterning (35), laminar flow patterning (36–38), robotic spotting (39–41), and jet printing (42, 43), and selective plasma etching (44). These patterning methods have been used to pattern hydrogels (45–47), cells (48, 49), and proteins (36) within microfluidic systems. However, they often have several shortcomings. For example, patterns generated by laminar flow patterning and microfluidic patterning are limited to a relatively narrow range of continuous patterns, which are mainly determined by the flow geometry.
2. Chemical Vapor Deposition In addition to solvent-based methods that are being utilized for biomedical surface modification, solventless surface modification methods, such as chemical vapor deposition (CVD) polymerization, are currently being explored for biomedical devices. Many of the advantages of CVD polymerization are unique when compared to solvent-based coating processes. First, impurities associated with the use of solvents, initiators, or plasticizers are essentially nonexistent. Second, CVD coatings are conformal, allowing for simple and uniform modification of three-dimensional substrate geometries (50). Third, although the initiation step requires high temperatures, initiation takes place away from the substrate,
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
263
and the substrates can be controlled and maintained at room temperature. The control over substrate temperature allows for the deposition of polymers onto delicate substrates, as well as onto mechanically strong materials made of inorganic substances. Several examples of CVD-based polymer coatings have been reported: Frank and coworkers (51, 52) have grafted polypeptide chains onto a surface using CVD. Gleason and coworkers (53, 54) have shown that polymerization initiators can be introduced together with the monomer through basic process modification, thus facilitating the polymerization of monomers which do not contain an initiator. Hot filaments within the deposition chamber can be used for initiation of radical polymerizations, which often yields conformal coatings. A major focus of CVD polymerization has been the polymerization of substituted (2.2]paracyclophanes (PCP) to yield functionalized poly(p-xylylenes) (PPX). This CVD polymerization is adapted from a process first developed by Gorham for parylene coatings (55). In this procedure (Fig. 1), a cyclic dimer is sublimated under vacuum (0.2–0.3 Torr), and transported by a carrier gas through an external heat source (T = 600–800°C). If the temperature is sufficiently high, a homolytic cleavage occurs across both bridge bonds, resulting in two quinodimethane diradicals, serving as an initiation step. The radicals then deposit and polymerize onto a sample that is fixed at a particular temperature (between −40 and 60°C). We have successfully modified PCPs with a wide variety of functional groups (56–60), which can then serve as reactive sites for immobilization of biomolecules. Vaporbased polymerization of PCPs produces a conformal PPX coating with mechanical integrity and low dielectric constants. Such properties are useful attributes for various applications including MEMS devices (61–64). In this chapter, we discuss recently developed methods of fabricating micropatterns onto substrates via CVD of reactive poly(p-xylylenes). For this purpose, the CVD technology can be utilized in one of two general ways (Fig. 2): (1) Deposition of a homogenous polymer coating that is reactive, then chemically pattern the coating after CVD treatment, and (2) fabrication of patterns of CVD polymer during the deposition process in situ. 2.1. Microcontact Printing onto Reactive CVD Coatings
Microcontact printing technology can be used to fabricate micropatterns of immobilized biomolecules onto reactive CVD coatings post-deposition. In this method, a PDMS stamp is cast from a photolithographically produced master made of silicon. Once the PDMS is cured, the biomolecule of interest is dissolved in a buffer, and the resulting solution is inked onto the PDMS pattern. The patterned PDMS substrate is then laid onto the surface, and the biomolecules are allowed to react (Fig. 3a). The PDMS stamp can be removed and reused multiple times.
264
Elkasabi and Lahann
Fig. 1. Functionalized [2.2] paracyclophanes (PCPs) can be polymerized into functionalized poly(p-xylylenes) (PPXs) with tailored reactivity. (a) Reaction Scheme yielding functionalized poly-p-xylylenes; (b) Block diagram describing the chemical vapor deposition polymerization process; (c) Examples of functionalized [2.2] paracyclophanes that have been used for CVD polymerization. Taken from (60).
One recent example (65) exploits the specificity of hydrazides toward aldehydes and ketones (66). Carbonyl-containing surfaces can be modified using dihydrazide homobifunctional linkers to form hydrazone bonds on one side, yielding alkyl hydrazide spacers on the other side, which can react further with formyl-containing groups in saccharides (66). Adipic acid dihydrazide was chosen as the linker due to its intermediate-length spacer arm, which leads to accessible reactive sites for further reaction. A substrate coated with poly(4-formyl-p-xylylene-co-p-xylylene) (formylPPX) was patterned with adipic acid dihydrazide, hence creating hydrazide-activated surfaces suitable for targeting saccharides.
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
265
Fig. 2. Two general protocols for fabrication of micropatterned surfaces are described. (1) CVD process, followed by subsequent patterning of the reactive coating. (2) Patterned deposition of the CVD polymer.
Fig. 3. Microcontact printing process for (a) the immobilization of sugars onto aldehyde-functionalized PPX and (b) click chemistry. Taken from (65) and (69).
266
Elkasabi and Lahann
The hydrazide-modified polymer surface was then reacted with 2-mannobiose, a disaccharide. One mannose group reacted with the hydrazide while leaving the other saccharide group free. Rhodamine-labeled concanavalin A, a mannose-specific lectin that recognizes the free mannose unit (67), was used to investigate saccharide binding. Patterned substrates were visualized using fluorescence microscopy (Fig. 3a, inset). The rhodamine-labeled lectin bound specifically to the disaccharide-presenting surface, which was then immobilized onto a substrate coated with formylPPX and patterned with lines of adipic acid dihydrazide. Immobilization of microscale patterns on formyl-PPX can also be extended towards DNA immobilization (68). Supermolecular nanostamping (SuNS) was used to fabricate DNA nanopatterns immobilized onto formyl-PPX. The patterns can be lines or spots, an important feature for the operation of DNA microarrays. Another example (69) involved the use of poly(4-ethynyl-pxylylene-co-p-xylylene) (ethynyl-PPX), a polymer specifically tailored for use in click chemistry. Its reactivity against azides was studied in order to assess whether the coating can be used for heterogeneous click reactions. Huisgen 1,3-dipolar cycloaddition between ethynylPPX and an azide-containing biotin-based ligands in the presence of copper(II) sulfate and sodium ascorbate was examined (Fig. 3b). This coupling reaction yields triazoles, as described for solventbased systems (70). Sodium ascorbate acts as a reductant, generating CuI ions in situ, which then function as the catalyst (70). Biotin azide was chosen as the representative ligand in this study, because biotin forms a strong noncovalent interaction with streptavidin (which has been widely used for binding biotinylated biomolecules) (56). A thin layer of biotin azide and sodium ascorbate was spread onto a film of ethynyl-PPX and dried using N2. In comparison to the concurrent microcontact printing of catalyst and azide, a twostep approach was found to be superior. A patterned PDMS stamp was inked with a CuSO4 solution and kept in contact with the substrate for 12–18 h. The patterned substrate was then rinsed and incubated with an aqueous solution of rhodamine-labeled streptavidin. The immobilization of biotin azide onto ethynylPPX was assed using fluorescence microscopy. The fluorescence micrograph and ellipsometric thickness map shown in Fig. 3b confirm selective protein coupling in the regions where the CuSO4 solution was microcontact printed, thus demonstrating the spatially directed binding of biotin azide to ethynyl-PPX. Thus, the alkyne groups on the polymer surface are reactive and can be effectively used as anchoring sites for various biomolecules. 2.2. CVD Patterning Within Confined Microgeometries
Even though miniaturized bioanalytical devices contain dimensions of high aspect ratios, the homogeneous modification of their surfaces can be challenging. In an attempt to expand CVD polymerization to the coating of complex microgeometries with
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
267
high aspect ratios, a recent study (71) examined the deposition behavior of functionalized poly(p-xylylenes) within preassembled microfluidic devices. It was demonstrated that CVD polymerization can be used to deposit a range of functionalized poly(p-xylylenes) within confined microgeometries. Seven different poly(p-xylylenes) were deposited via CVD polymerization within both removable and sealed PDMS microchannels (72). A subgroup of five poly(p-xylylenes) had reactive side groups (so-called reactive coatings), while two commercially available poly(p-xylylenes) were included as nonfunctionalized references (ParyleneTM N and C). The PDMS microchannels used in this study were open at both ends and were 75 mm high and 100 mm wide. Both straight (1,600 microns long) and meandering channel (2,800 microns long) layouts with high aspect ratios were studied (Fig. 4). For both straight and meandering microchannels the degree of deposition was constant and did not change with increasing
Fig. 4. Conformal deposition of CVD polymers occurs even within microscale geometries. Facile modification and biofunctionalization of microfluidic channels can be attained. Adapted from (71).
268
Elkasabi and Lahann
film thickness. Homogenous surface coverage of different microgeometries has been demonstrated for these reactive coatings. Deposition within aspect ratios of up to 37 was accomplished, based on optical microscopy and imaging XPS results. In addition to the deposition studies, immobilization studies were conducted using permanently sealed PDMS devices (72) after CVD polymerization. The microchannels were coated with either poly(4-amino-p-xylylene-co-p-xylylene) (amino-PPX) or poly(4-trifluoroacetyl-p-xylylene-co-p-xylylene) (PPX-COCF3) prior to immobilization. While amino-PPX provides primary amino groups for coupling with activated carboxyl groups (amide formation), PPX-COCF3 has keto groups that can react with hydrazines or hydrazides. To assess the chemical activity of both reactive coatings, a PFP-derived biotin ligand and a biotin hydrazide ligand were used to evaluate chemical reactivity of amino-PPX and PPX-COCF3, respectively. These ligands undergo nearly quantitative conversion with amines or ketones; also, the interactions between biotin and streptavidin result in confinement of streptavidin on the biotin-modified surface. For all ligand immobilization reactions, aqueous solutions of the corresponding biotin derivative were filled into the sealed microchannels of either meandering or straight geometry. After thorough rinsing with buffer, microchannels were incubated with rhodamine-labeled streptavidin, then the surfaces were rinsed and visualized by fluorescence microscopy. Figure 4 shows microchannels that were coated with polymer and then subjected to the biotin/strepdavidin protocol. Homogeneous distribution throughout the entire microchannel was observed, indicating that functional groups were available throughout the entire coating area, for both aminoPPX and PPX-COCF3. The deposition of reactive CVD coatings within confined microgeometries bridges a critical technological gap toward surface-modified microfluidic devices for use in “BioMEMS” applications. 2.3. Vapor-Assisted Micropatterning in Replica Structures
A related patterning approach utilizes vapor-assisted micropatterning in replica structures (VAMPIR). In this method, chemical and topological surface microstructures can be obtained by masking certain areas of the substrate during CVD polymerization and then depositing the reactive coatings only within the exposed areas. Although conceptually simple, such an approach toward microstructured surfaces came with some challenges. For instance, in CVD polymerization, polymer deposition is transport-limited, and the feasibility of deposition within replica structures with micronscale capillaries was unclear. However, the properties of polymers deposited are of a greater variety. While stencils and shadow masks have been applied for area-selective deposition using both rigid and elastomeric materials (24, 73, 74), many of those pattern processes are limited to hydrophilic polymers that are soluble in
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
269
polar solvents. However, the solvent-free process described here can be used for both hydrophilic and hydrophobic coatings. In a recent study (75), polydimethylsiloxane (PDMS)based replica structures (or stencils) designed to generate a desired surface pattern were reversibly sealed onto a silicon substrate (Fig. 5). The masked substrate was then placed onto a temperature-controlled stage (15°C) inside of the CVD polymerization chamber. 4-pentafluoropropionyl[2.2]paracyclophane underwent pyrolysis and polymerized into poly(4- pentafluoropropionyl-p-xylylene-co-p-xylylene). After completion
Fig. 5. (a) Process of vapor-assisted microstructuring using replica structures (left column) as well as shadow masks (right column) during CVD polymerization. Fluorescently tagged molecules are immobilized onto (b) poly(4-pentafluoropropionyl-p-xylylene-co-p-xylylene) and (c) poly(p-xylylene-4-methyl-2-bromoisobutyrate-co-p-xylylene). The latter was used to grow poly(OEGMA) within the squares, which inhibited the adsorption of fibrinogen (i ) and attachment of NIH 3T3 fibroblasts (ii ). Adapted from (75) and (76).
270
Elkasabi and Lahann
of the CVD polymerization, the PDMS molds were removed, and hence a chemically and topologically structured surface was created. Surface features are defined by the deposited polymer footprints. In one instance, a substrate was masked with a PDMS membrane, which contained footprints shaped into the letters “UM” (Fig. 5b). Subsequent CVD polymerization resulted in ultra-thin polymer films outside of the masked areas. Imaging X-ray photoelectron spectroscopy confirmed the presence of characteristic elements within their localized regions – i.e., fluorine was found only outside the “UM” footprint boundaries, whereas silicon was found within the footprint. PPX-COC2F5 has keto groups that can react with hydrazines or hydrazides in high yields (66). Biotin hydrazide, a model ligand, was used again for immobilization onto the functionalized polymer. In a subsequent step, the well-known interactions between biotin and streptavidin were used for visualization of surface-immobilized biotin. To examine the immobilization of biotin ligands within the patterns, streptavidin conjugated with CdSe quantum dots (Qdot® 525) were allowed to bind to the biotin-modified surfaces. Binding was homogenous throughout the surface-modified areas. As anticipated, after biotin immobilization, the subsequently self-assembled quantum dots were resolved into a range of different predesigned patterns. In another instance (76), poly(p-xylylene-4-methyl-2bromoisobutyrate-co-p-xylylene) was patterned in the same manner as described for PPX-COC2F5. A PDMS structure with square holes was used during the deposition process. After patterned deposition of poly(p-xylylene-4-methyl-2-bromoisobutyrateco-p-xylylene) onto PMMA surfaces (Fig. 5c), the initiator contained within the functionalized coating was used to perform ATRP of poly(OEGMA). Fluorescently labeled fibrinogen was found to adsorb selectively onto the bare PMMA substrate, whereas the poly(OEGMA)-modified squares inhibited protein adsorption (Fig. 5c.i). The attachment and growth of NIH3T3 fibroblasts followed a similar trend (Fig. 5c.ii). The lower limit of VAMPIR feature sizes was also evaluated. A PDMS replica structure was prepared with varying distances between posts (150, 100, 50, and 25 mm). Thicknesses of the deposited polymer coatings were measured in the center of each region. The thickness decreased from 49.6 nm measured for the area with 150 mm feature sizes, over 42 nm (100 mm) and 28.7 nm (50 mm), to 7.3 nm measured for the areas with 25 mm wide features. A relative coordinate system is used to express the coating thickness distribution for different feature sizes (Fig. 6) (77, 78). Rearrangement of the thickness data in terms of dimensionless thicknesses d(x)/d0 and width (x/b) reveals a surprisingly uniform behavior. d(x)/d0 denotes the ratio of the absolute film thickness at the given point x to that at the open surface, and x/b is the ratio of depth over width of the feature. As indicated in Fig. 6, the
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
271
Fig. 6. Plot of dimensionless thicknesses d(x)/d0 vs. dimensionless width (x/b), where d0 is film thickness (nm) on an open area for an according dimension recorded by using imaging ellipsometry; b is the width (mm) of the dimension. Taken from (75).
dimensionless thicknesses measured for feature sizes ranging from 25 to 200 mm fall onto a single trend line. Process parameters dominate over feature size, as predicted by Tolstopyatov et al. In this study, universal thickness distributions for the deposition of unfunctionalized poly-p-xylylene in microchannels were found (77, 78). Given the theoretical and experimental findings done on vapor deposition, vapor-assisted microstructuring in replica structures (VAMPIR) establishes a simple technique to create both chemical and topological surface patterns. 2.4. Projection Lithography-Based Patterning Using on Photoreactive CVD Polymers
Another method of post-CVD micropatterning involves the projection of ultraviolet light onto photoreactive PPX coatings. Recently, a photodefinable polymer, poly(4-benzoyl-p-xylylene-co-p-xylylene) (benzoyl-PPX), was prepared by CVD polymerization and was used for fabrication of discontinuous surface patterns onto threedimensional microscale objects (79, 80). Due to its structural analogy to benzophenone, the reactive coating provides light-reactive carbonyl groups that are readily activated at wavelengths of ~340 nm. The temporarily generated free radicals spontaneously react with adjunct molecules, mainly via C–H abstraction (81). Suh et al. (82) have demonstrated the ability of benzoyl-PPX to immobilize hydrogel elements, an important requirement in microfabrication processes. Capillary force lithography was combined with photoreactive patterning in order to fabricate an array of immobilized PEG hydrogels. As shown in Fig. 7, microstructured stents and microchannels were recently fabricated by a two-step procedure: (1) coating of the objects with a photodefinable polymer, poly[4-benzoylp-xylylene-co-p-xylylene], via CVD polymerization (82, 83)
272
Elkasabi and Lahann
Fig. 7. Schematic description of the 3D projection lithography technique. The method comprises two process steps: deposition of the photodefinable CVD coating (step 1) and subsequent projection lithographic rendering of the polymercoated colloids (step 2). Inset shows an endovascular stent and a microfluidic pathway that are patterned using projection lithography. Adapted from (79) and (80).
and (2) spatially controlled surface reaction of the photoreactive coatings using a highly parallel projection lithographic patterning step. Once the deposition of the photoreactive coatings on endovascular stents was demonstrated, spatially directed microstructuring became achievable. To obtain spatially controlled surface patches on stents, we selectively illuminated certain areas of previously coated stents with UV radiation at 365–400 nm by using a high-throughput projection technique that has been previously used for in situ synthesis of peptides and DNA on microarrays (84–86). After surface modification via CVD polymerization, coated stents were immersed in an aqueous solution of a four-arm star polyethylene glycol (star-PEO, 10,000 g/mol, 1 weight-%). For patterning, a digital micromirror device (DMD, Texas Instruments) was used as a dynamic mask (87). UV radiation of about 365– 400 nm wavelength was modulated by the dynamic mask. The corresponding patterns were then transferred onto the stents (Fig. 5a). DI-water was used to separate excess PEO. The stents were incubated with protein (Alexa Fluor 546-conjugated fibrinogen, Molecular Probes Inc.) solutions for 5 min. After incubation, phosphate-buffered saline (PBS) and DI-water were used to
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
273
rinse off excess adsorbed proteins. After CVD coating, PDMS microchannels were immersed in an aque-ous solution of biotinPEO-LC-amine (10 mM, Pierce) in PBS (pH 7.4). The corresponding patterns were then transferred onto the microchannels. After rinsing, samples were incubated with rhodamine (TRITC)conjugated streptavidin (50 mg/mL, Pierce) in PBS containing 0.1% (w/v) bovine albumin and Tween 20 (0.02% (v/v)) for 60 min. The surface was rinsed several times with PBS containing 0.1% (w/v) bovine albumin and Tween 20 (0.02% (v/v)). Programmable patterns were created by using a 1,024 × 768-pixel digital micromirror device (DMD) (Fig. 7). While the entire surface of the substrates was coated with the photoreactive coating during CVD polymerization, only the areas illuminated with the DMD underwent photochemical conversion of the carbon–oxygen double bond from the singlet ground state into the corresponding triplet state (88). As seen in Fig. 7, an endovascular stent and a microchannel were coated with the photoreactive benzoyl-PPX polymer and subsequently exposed to the DMD grid UV patterning. The PEO-free areas facilitated adsorption of fibrinogen, while the areas of PEO immobilization did not (inside squares). An identical pattern of immobilized streptavidin (inside the squares) was observed within microchannels. Despite the irregular shape and small dimensions of the objects, homogenous chemical micropatterns were obtained on the stent surface, making possible the progression of advanced surface architectures for medical devices. 2.5. Selective Deposition of Reactive CVD Polymers
While the methods mentioned thus far rely on physical means to obtain spatially controlled surface modification, an even simpler approach would be to selectively inhibit CVD polymerization and deposition based on differences in the substrate chemistry. Jensen and coworkers first reported the selective inhibition of parylene™ N, parylene™ C, as well as poly(p-phenylene vinylene) (PPV) by iron and iron salts (89) and used selective CVD polymerization to create a wide range of patterns (89). It was also shown that, in a similar fashion, several transitional metals, metal salts, and organometallic complexes inhibit the growth of parylene™ N and C (90). Suh et al. have used selective CVD polymerization within submicron scale PDMS channels to yield high aspect ratio structures. Surface-coated PDMS channels of as little as 180 nm in width were obtained by depositing iron on the bottom of microchannels and then selectively depositing polymer only on the channel sidewalls (91). Recently, the first selective CVD polymerization of a functionalized poly-p-xylylene was reported (92). The result is a simple patterning process that relies on selective inhibition of polymer films that can act as chemical anchors for further surface modification via covalent immobilization. The study investigated selective inhibition of CVD polymerization by series of metals. Based on IRRAS spectra, Ti appeared
274
Elkasabi and Lahann
to be the only metal that effectively inhibited the growth of poly(4-vinyl-p-xylylene-co-p-xylylene) (vinyl-PPX) during the CVD polymerization process. Ti also inhibited CVD polymerization of poly(4-chloro-p-xylylene). Next, a Ti-coated silicon wafer was prepared, and circles of Au were deposited onto the Ti-coated silicon using a shadow mask. Following the protocol outlined in Fig. 8a, this bi-metal surface was CVD coated with vinyl-PPX and subsequently subjected to olefin cross-metathesis reaction with fluorescein O-methacrylate. Only the Au islands showed significant fluorescence signals (Fig. 8b). In addition, a strong contrast
Fig. 8. (a) Schematic illustration of the selective deposition of poly(4-vinyl-p-xylylene-co-p-xylylene) on patterned Ti/Au substrates. Au was deposited through a shadow mask onto a Ti-coated silicon wafer followed by polymer deposition via CVD polymerization. Olefin cross-metathesis reaction of fluorescein O-methacrylate was used to probe the selective deposited polymer on Au surface. (b) Fluorescence micrograph reveals that only the Au islands showed appreciable signals of fluorescence. Taken from (92).
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition
275
was observed between Au and the Ti surfaces providing further clarification on the selective inhibition of CVD of vinyl-PPX. Moreover, Ti or Au samples were coated with vinyl-PPX, and cross-metathesis reaction of poly(ethylene glycol) methyl ether methacrylate (PEGMA) was conducted on both samples. For each modification step, IRRAS spectra were recorded. Absorption bands at 2,866, 2,924, and 3,013 cm–1 due to C–H symmetric and asymmetric stretching bands can be clearly detected on the Au surface after deposition of vinyl-PPX. In addition, a strong, sharp band at 1,717 cm–1 indicative of the C=O bond of the ester group, and a strong band at 1,110 cm–1, which is due to C–O–C stretches of the ester group appeared after olefin cross-metathesis reaction of OEGMA. At each modification step, no significant signal was detected on the Ti surface, providing strong evidence that vinyl-PPX was not deposited on Ti, which consequently prevented cross-metathesis reaction of OEGMA. The fact that the selectively deposited reactive coatings are equipped with functional groups for further surface modification provides a simple access route toward micro- and nanostructured surfaces.
3. Conclusions The merger of materials science and biotechnology has fueled the demand for novel types of bioarrays with highly engineered surfaces. Toward this goal, a range of methods for the facile fabrication of surface micropatterns has been developed based on vapor-based reactive coatings. In this chapter, we have discussed the current use of CVD polymerization toward surface engineering of bioarrays and biodiagnostic devices. Specifically, we have outlined the adaptability of the CVD process toward microfabrication of polymer thin films. Two general fabrication methods were discussed. First, one may fabricate a biomolecular micropattern after the reactive poly(p-xylylene) has been coated. Micro contact printing and projection lithography are two approaches used for post-CVD surface modification. In microcontact printing, a patterned PDMS stamp is inked with the molecule of interest, then subsequently laid upon the CVD film. Projection lithography employs micromirrors in order to project UV light onto a photoreactive CVD polymer coatings. Secondly, one may pattern a CVD polymer onto the substrate directly during the deposition process. Patterned replica structures mounted onto a substrate will mask deposition over specified areas. In addition, certain polymers have been known to deposit selectively on different transition metals. Thus, a patterned metal substrate would lead to patterned polymer deposition. Microstructuring of reactive CVD polymer coatings produces robust coatings with excellent
276
Elkasabi and Lahann
biocompatibility. While proof of principles are well established in the aforementioned fabrication methods, their broad integration with in vivo devices has yet to be realized and remains as a major objective of future work. References 1. A. Berg, W. Olthius, P. Bergveld, “Micro Total Analysis Systems” 2000, 1st edition, Kluwer, Dordrecht, The Netherlands, 2000. 2. P. Li, D. J. Harrison, “Transport, manipulation, and reaction of biological cells on-chip using electrokinetic effects” Anal. Chem. 1997, 69, 1564. 3. H. Mao, T. Yang, P. S. Cremer, “Design and characterization of immobilized enzymes in microfluidic systems” Anal. Chem. 2002, 74, 379. 4. S. H. Chen, W. C. Sung, G. B. Lee, Z. Y. Lin, P. W. Chen, P. C. Liao, “A disposable poly(methylmethacrylate)-based microfluidic module for protein identification by nanoelectrospray ionization-tandem mass spectrometry” Electrophoresis 2001, 22, 3972. 5. C. S. Effenhauser, J. M. Bruin, A. Paulus, M. Ehrat, “Integrated capillary electrophoresis on flexible silicone microdevices: analysis of DNA restriction fragments and detection of single DNA molecules on microchips” Anal. Chem. 1997, 69, 3451. 6. J. Monahan, A. A. Gewirth, R. G. Nuzzo, “Indirect fluorescence detection of simple sugars via high-pH electrophoresis in poly (dimethylsiloxane) microfluidic chips” Electro phoresis 2002, 23(14), 2347. 7. J.C. Love, L.A. Estroff, J.K. Kriebel, R.G. Nuzzo, G.M. Whitesides, “Self-assembled monolayers of thiolates on metals as a form of nanotechnology” Chem. Rev. 2005, 105, 4, 1103–1169. 8. D. Falconnet, G. Csucs, H. M. Grandin, M. Textor, “Surface engineering approaches to micropattern surfaces for cell-based assays” Biomaterials 2006, 27, 16, 3044–3063. 9. D. S. Ginger, H. Zhang, C. A. Mirkin, “The evolution of Dip-Pen nanolithography” Angew. Chem. Int. Ed. 2004, 43, 30. 10. S. Kramer, R. R. Fuierer, C. B. Gorman, “Scanning probe lithography using self-assembled monolayers” Chem. Rev. 2003, 103, 4367. 11. S. Y. Chou, P. R. Krauss, P. J. Renstrom, “Imprint of sub-25 nm vias and trenches in polymers” App. Phys. Lett. 1995, 67, 3114. 12. S. Y. Chou, P. R. Krauss, P. J. Renstrom, “Imprint Lithography with 25-Nanometer Resolution” Science 1996, 272, 85.
13. Y. N. Xia, G. M. Whitesides, “Soft lithography” Ann. Rev. Mater. Sci. 1998, 28, 153. 14. A. Kumar, G. M. Whitesides, “Features of gold having micrometer to centimeter dimensions can be formed through a combination of stamping with an elastomeric stamp and an alkanethiol “ink” followed by chemical etching” App. Phys. Lett. 1993, 63, 2002. 15. J. L. Wilbur, A. Kumar, E. Kim, G. M. Whitesides, “Microfabrication by microcontact printing of self-assembled monolayers” Adv. Mater. 1994, 6, 600. 16. Y. N. Xia, E. Kim, X. M. Zhao, J. A. Rogers, M. Prentiss, G. M. Whitesides, “Complex optical surfaces formed by replica molding against elastomeric masters” Science 1996, 273, 347. 17. X. M. Zhao, Y. N. Xia, G. M. Whitesides, “Fabrication of three-dimensional microstructures: Microtransfer molding” Adv. Mater. 1996, 8, 837. 18. E. Kim, Y. N. Xia, G. M. Whitesides, “Polymer microstructures formed by moulding in capillaries” Nature 1995, 376, 581. 19. E. Kim, Y. N. Xia, X. M. Zhao, G. M. Whitesides, “Solvent-assisted microcontact molding: A convenient method for fabricating three-dimensional structures on surfaces of polymers” Adv. Mater. 1997, 9, 651. 20. K. Y. Suh, Y. S. Kim, H. H. Lee, “Capillary force lithography” Adv. Mater. 2001, 13, 1386. 21. W. T. Tsang, A. Y. Cho, “Molecular beam epitaxial writing of patterned GaAs epilayer structures” App. Phys. Lett. 1978, 32, 491. 22. T. Schallenberg, T. Borzenko, G. Schmidt, M. Obert, G. Bacher, C. Schumacher, G. Karczewski, L. W. Molenkamp, S. Rodt, R. Heitz, D. Bimberg, “Controlled self-assembly of semiconductor quantum dots using shadow masks” App. Phys. Lett. 2003, 82, 4349. 23. S. De Vusser, S. Steudel, K. Myny, J. Genoe, P. Heremans, “Integrated shadow mask method for patterning small molecule organic semiconductors” App. Phys. Lett. 2006, 88, 103501. 24. D. C. Duffy, R. J. Jackman, K. M. Vaeth, K. F. Jensen, G. M. Whitesides, “Electroluminescent materials with feature sizes as small as 5 mm using elastomeric membranes as masks for dry lift-off” Adv. Mater. 1999, 11, 546.
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition 25. N. Takano, L. M. Doeswijk, M. A. F. van den Boogaart, J. Auerswald, H. F. Knapp, O. Dubochet, T. Hessler, J. Brugger, “Fabrication of metallic patterns by microstencil lithography on polymer surfaces suitable as microelectrodes in integrated microfluidic systems” J. Micromech. Microeng. 2006, 16, 1606. 26. B. R. Ringeisen, J. Callahan, P. K. Wu, A. Pique, B. Spargo, R. A. McGill, M. Bucaro, H. Kim, D. M. Bubb, D. B. Chrisey, “Novel laser-based deposition of active protein thin films” Langmuir 2001, 17, 3472. 27. E. Ostuni, R. Kane, C. S. Chen, D. E. Ingber, G. M. Whitesides, “Patterning mammalian cells using elastomeric membranes” Langmuir 2000, 16, 7811. 28. R. Pal, K. E. Sung, M. A. Burns, “Microstencils for the patterning of nontraditional materials” Langmuir 2006, 22, 5392. 29. D. G. Castner, B. D. Ratner, “Biomedical surface science: foundations to frontiers” Surf. Sci. 2002, 500, 28. 30. P. Rai-Choudhury, “Handbook of Microlithography, Micromachining, and Microfabrication. Volume 1: Microlithography”, SPIE-The International Society for Optical Engineering, 1997. 31. F. Cerrina, “X-ray imaging: applications to patterning and lithography” J. Phys. D Appl. Phys. 2000, 33, R103. 32. N. Yao, Z. L. Wang, Handbook of Microscopy for Nanotechnology, Springer, Berlin, 2005. 33. A. A. Tseng, “Recent developments in nanofabrication using ion projection lithography” Small 2005, 1, 594. 34. D. Meschede, H. Metcalf, “Atomic nanofabrication: atomic deposition and lithography by laser and magnetic forces” J. Phys. D Appl. Phys. 2003, 36, R17. 35. E. Delamarche, A. Bernard, H. Schmid, A. Bietsch, B. Michel, H. Biebuyck, “Microfluidic networks for chemical patterning of substrates: design and application to bioassays” J. Am. Chem. Soc. 1998, 120, 500–508. 36. S. Takayama, J. C. McDonald, E. Ostuni, M. N. Liang, P. J. A. Kenis, R. F. Ismagilov, G. M. Whitesides, “Patterning cells and their environments using multiple laminar fluid flows in capillary networks” Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 5545–5548. 37. B. Regenberg, U. Kruehne, M. Beyer, L.H. Pedersen, M. Simon, O.R.T. Thomas, J. Nielsen, T. Ahl, “Use of laminar flow patterning for miniaturised biochemical assays” Lab. Chip. 2004, 4, 654–657. 38. S. Takayama, E. Ostuni, P. LeDuc., K. Naruse, D. E. Ingber., G. M. Whitesides “Laminar
39.
40.
41.
42. 43.
44.
45.
46.
47.
48.
49.
50. 51.
277
flows: subcellular positioning of small molecules” Nature 2001, 411, 1016. M. Schena, D. Shalon, R. W. Davis, P. O. Brown, “Quantitative monitoring of gene expression patterns with a complementary DNA microarray” Science 1995, 270, 467–470. R.A. Heller, M. Schena, A. Chai, D. Shalon, T. Bedilion, J. Gilmore, D.E. Woolley, R.W. Davis, “Discovery and analysis of inflammatory disease-related genes using cDNA microarrays” Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 2150–2155. D. Shalon, S. J. Smith, P. O. Brown, “A DNA microarray system for analyzing complex DNA samples using two-color fluorescent probe hybridization” Genome Res. 1996, 6, 639–645. M.F. Lopez, M.G. Pluskal, “Protein microand macroarrays: digitizing the proteome” J. Chromatogr., B 2003, 787, 19–27. A. Roda, M. Guardigli, C. Russo, P. Pasini, M. Baraldini, “Protein microdeposition using a conventional ink-jet printer” Biotechniques 2000, 28, 492–496. A. Khademhosseini, K. Y. Suh, S. Jon, G. Eng, J. Yeh, G. J. Chen, R. Langer, “A soft lithographic approach to fabricate patterned microfluidic channels” Anal. Chem. 2004, 76, 3675–3681. W. Zhan, G. H. Seong, R. M. Crooks, “Hydrogel-based microreactors as a functional component of microfluidic systems” Anal. Chem. 2002, 74, 4647–4652. D. J. Beebe, J. S. Moore, J. M. Bauer, Q. Yu, R. H. Liu, C. Devadoss, B. H. Jo, “Functional hydrogel structures for autonomous flow control inside microfluidic channels” Nature 2000, 404, 588–590. J. Heo, K. J. Thomas, G. H. Seong, R. M. Crooks, “A microfluidic bioreactor based on hydrogel-entrapped E. coli: cell viability, lysis, and intracellular enzyme reactions” Anal. Chem. 2003, 75, 22–26. M. Mrksich, L. E. Dike, J. Tien, D. E. Ingber, G. M. Whitesides, “Using microcontact printing to pattern the attachment of mammalian cells to self-assembled monolayers of alkanethiolates on transparent films of gold and silver” Exp. Cell Res. 1997, 235, 305–313. S.W. Rhee, A. M. Taylor, C. H. Tu, D. H. Cribbs, C. W. Cotman, N. L. Jeon, “Patterned cell culture inside microfluidic devices” Lab. Chip 2005, 5, 102–107. S.D. Senturia. “Microsystem Design”, Kluwer, Norwell, 2000. Y.C. Chang, C.W. Frank, “Vapor depositionpolymerization of -amino acid N-carboxy anhydride on the silicon(100) native oxide surface” Langmuir 1998, 14, 326.
278
Elkasabi and Lahann
52. N. H. Lee, C. W. Frank, “Surface-initiated vapor polymerization of various -amino acids” Langmuir 2003, 19, 1295. 53. Y. Mao, K. K. Gleason, “Hot filament chemical vapor deposition of poly(glycidyl methacrylate) thin films using tert-butyl peroxide as an initiator” Langmuir 2004, 20, 2484. 54. T. P. Martin, K. K. Gleason, “Combinatorial initiated CVD for polymeric thin films” Chem. Vap. Dep. 2006, 12, 685. 55. W.F. Gorham, “A new, general synthetic method for the preparation of Linear poly-p-xylylenes” J. Polym. Sci., Part A-1 1966, 4, 3027. 56. J. Lahann, D. Klee, H. Hocker, “Chemical vapour deposition polymerization of substituted [2.2]paracyclophanes” Macromol. Rapid Commun. 1998, 19, 441. 57. J. Lahann, M. Balcells, T. Rodon, J. Lee, I.S. Choi, K.F. Jensen, R. Langer, “Reactive polymer coatings: a platform for patterning proteins and mammalian cells onto a broad range of materials” Langmuir 2002, 18, 3632. 58. J. Lahann, R. Langer, “Novel poly(p-xylylenes): thin films with tailored chemical and optical properties” Macromolecules 2002, 35, 4380. 59. J. Lahann, “Vapor-based polymer coatings for potential biomedical applications” Polym. Inter. 2006, 55, 1361. 60. Y. Elkasabi, M. Yoshida, H. Nandivada, H. Y. Chen, J. Lahann, “Towards multipotent coatings: chemical vapor deposition and biofunctionalization of carbonyl-substituted copolymers” Macromol. Rapid Comm. 2008, 29, 855–870. 61. M. Morgen, S. H. Rhee, J. H. Zhao, I. Malik, T. Ryan, H. M. Ho, M. A. Plano, P. Ho, “Comparison of crystalline phase transitions in fluorniated vs nonfluorinated parylene thin films” Macromolecules 1999, 32, 7555. 62. J. J. Senkevich, S. B. Desu, V. Simkovic, “Temperature studies of optical birefringence and X-ray diffraction with poly(p-xylylene), poly(chloro-p-xylylene) and poly(tetrafluorop-xylylene) CVD thin films” Polymer 2000, 41, 2379. 63. S. Y. Park, S. N. Chvalun, A. A. Nikolaev, K. A. Mailyan, A. V. Pebalk, I. E. Kardash, “The structure of poly(cyano-p-xylylene)” Polymer 2000, 41, 2937. 64. D. Klee, N. Weiss, J. Lahann, “Vapor-based polymerization of functionalized [2.2]paracyclophanes: a unique approach towards surface-engineered microenvironments” Chapter 18, Modern Cyclophane Chemistry, WileyVCH, Weinheim, 2004, p463. 65. H. Nandivada, H. Y. Chen, J. Lahann, “Vaporbased synthesis of poly[(4-formyl-p-xylylene)co-(p-xylylene)] and its use for biomimetic
66. 67.
68.
69. 70.
71.
72.
73.
74.
75.
76.
77.
78.
surface modifications” Macromol. Rapid Comm. 2005, 26, 1794. G. T. Hermanson, “Bioconjugate Techniques”, 1st edition, Academic, San Diego, CA, 1996. D. N. Moothoo, J. H. Naismith, “A general method for co-crystallization of concanavalin A with carbohydrates” Acta Crystallogr. D Biol. Crystallogr. 1999, D55, 1, 353. S. Thenevet, H.Y. Chen, J. Lahann, F. Stellacci, “A generic approach towards nanostructured surfaces based on supramolecular nanostamping on reactive polymer coatings” Adv. Mater. 2007, 19, 4333. H. Nandivada, H. Y. Chen, L. Bondarenko, J. Lahann, “Reactive polymer coatings that ‘click’ ” Angew. Chem. Int. Ed. 2006, 45, 3360. V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless, “A stepwise huisgen cycloaddition process: copper(I)-catalyzed regioselective “ligation” of azides and terminal alkynes” Angew. Chem. 2002, 114, 2708–2711; Angew. Chem. Int. Ed. 2002, 41, 2596–2599. H.Y. Chen, Y. Elkasabi, J. Lahann, “Surface modification of confined microgeometries via vapor-deposited polymer coatings” J. Am. Chem. Soc. 2006, 128, 374. J. C. McDonald, G. M. Whitesides, “Poly(dimethylsiloxane) as a material for fabricating microfluidic devices” Acc. Chem. Res. 2002, 35, 491–499. H. W. Gu, R. K. Zheng, X. X. Zhang, B. Xu, “Using Soft lithography to pattern highly oriented polyacetylene (HOPA) films via solventless polymerization” Adv. Mater. 2004, 16, 1356. M. Graff, S. K. Mohanty, E. Moss, A. B. Frazier, “Microstenciling: a generic technology for microscale patterning of vapor deposited materials” J. Microelectromech. Syst. 2004, 13, 956. H.Y. Chen, J. Lahann, “Vapor-assisted micropatterning in replica structures: a solventless approach towards topologically and chemically designable surfaces” Adv. Mater. 2007, 19, 3801. X. Jiang, H. Y. Chen, G. Galvan, M. Yoshida, J. Lahann, “Vapor-based initiator coatings for atom transfer radical polymerization” Adv. Funct. Mater. 2008, 18, 27. E. M. Tolstopyatov, “Thickness uniformity of gas-phase coatings in narrow channels: I. Long channels” J. Phys. D Appl. Phys. 2002, 35, 1516. E. M. Tolstopyatov, S. H. Yang, M. C. Kim, “Thickness uniformity of gas-phase coatings in narrow channels: II. One-side confined channels” J. Phys. D Appl. Phys. 2002, 35, 2723.
Progress Report on Microstructured Surfaces Based on Chemical Vapor Deposition 79. H. Y. Chen, J. M. Rouillard, E. Gulari, J. Lahann, “Colloids with high-definition surface structures” Proc. Nat. Acad. Sci. 2007, 104, 11173. 80. H. Y. Chen, J. M. Rouillard, E. Gulari, J. Lahann, “High-precision surface modification of three-dimensional geometries using photodefinable ultra-thin polymer coatings” PMSE Preprints 2006, 95, 125. 81. W. W. Shen, S.G. Boxer, W. Knoll, C.W. Frank, “Polymer-supported lipid bilayers on benzophenone-modified substrates” Biomac romolecules 2001, 2, 70–79. 82. K. Y. Suh, R. Langer, J. Lahann, “A novel photodefinable reactive polymer coating and its use for microfabrication of hydrogel elements” Adv. Mater. 2004, 16, 1401. 83. H. Y. Chen, J. Lahann, “Fabrication of discontinuous surface patterns within microfluidic channels using photodefinable vapor-based polymer coatings” Anal. Chem. 2005, 77, 6909. 84. J. Tian, H Gong, N. Sheng, X. Zhou, E. Gulari, X. Gao, G. Church, “Accurate multiplex gene synthesis from programmable DNA microchips” Nature 2004, 432, 1050. 85. J. P. Pellois, X. Zhou, O. Srivannavit, T. Zhou, E. Gulari, X. Gao, “Individually addressable parallel peptide synthesis on microchips” Nat. Biotech. 2002, 20, 922.
279
86. X. L. Gao, X. C. Zhou, E. Gulari, “Light directed massively parallel on-chip synthesis of peptide arrays with t-Boc chemistry” Proteomics 2003, 3, 2135. 87. X. Gao, E. LeProust, H. Zhang, O. Srivannavit, E. Gulari, P. Yu, C. Nishiguchi, Q. Xiang, X. Zhou, “A flexible light-directed DNA chip synthesis gated by deprotection using solution photogenerated acids” Nucl. Acids Res. 2001, 29, 4744–4750. 88. K. S. Taton, P. E. Guire, “Photoreactive selfassembling polyethers for biomedical coatings” Colloids Surf. B 2002, 24, 123–132. 89. K. M. Vaeth, K. F. Jensen, “Selective growth of Poly(p-phenylene vinylene) prepared by chemical vapor deposition” Adv. Mater. 1999, 11, 814. 90. K. M. Vaeth, K. F. Jensen, “Transition metals for selective chemical vapor deposition of parylene-based polymers” Chem. Mater. 2000, 12, 1305. 91. K. Y. Suh, R. Langer, J. Lahann, “Fabrication of elastomeric stamps with polymer-reinforced sidewalls via chemically selective vapor deposition polymerization of poly(p-xylylene)” App. Phys. Lett. 2003, 83, 4250. 92. H. Y. Chen, J. H. Lai, X. Jiang, J. Lahann, “Substrate-selective chemical vapor deposition of reactive polymer coatings” Adv. Mater. 2008, 20, 3474.
Chapter 17 Methods for Forming Human Microvascular Tubes In Vitro and Measuring Their Macromolecular Permeability Gavrielle M. Price and Joe Tien Abstract This chapter describes a protocol for forming open endothelial tubes in vitro and quantifying their permeability to macromolecules. These tubes consist of confluent monolayers of human microvascular endothelial cells in perfused microfluidic collagen gels. The cylindrical geometry of the tubes mimics the shape of microvessels in vivo; it allows simultaneous and/or repeated measurements of permeability coefficients and detection of focal leaks. We have used these in vitro models to test the effects of agonists on microvascular permeability and are developing arrays of microvascular tubes to enable large-scale testing. Key words: Microvascular tissue engineering, Endothelial cells, Collagen, Permeability, Focal leaks
1. Introduction This chapter describes methods recently developed by our group to quantify the barrier function of engineered human microvascular tubes in vitro (1, 2). It provides step-by-step instructions for forming endothelial tubes within channel-containing collagen gels, measuring their permeabilities to macromolecules, and determining their number of focal leaks. We assume that the reader is familiar with standard cell culture techniques and with our complementary review on methods to form cylindrical channels within collagen (3). Our methods are adapted from those originally designed for use in intact or explanted microvessels from mice, rats, and frogs (4–8). In vivo, macromolecular permeability can be quantified with two metrics: (1) an effective permeability coefficient Pd, which describes the ability of a solute to escape uniformly from the vascular lumen and (2) the density of focal leaks, which are localized regions of high permeability (9). The reflection coefficient s,
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_17, © Springer Science+Business Media, LLC 2011
281
282
Price and Tien
which describes the oncotic contribution of a solute, provides a third measure of barrier function (10, 11); we do not describe how to measure this quantity in our system. Traditionally, the permeability of human endothelium has been examined in planar cell cultures in vitro (12). These cultures lack physiologically relevant shear stress, and the permeability coefficients are averaged over large areas and long times. In contrast, the open cylindrical architecture of our vessels allows constant perfusion, which enables real-time measurement of permeabilities and focal leaks, just as in cannulated microvessels. The small footprint of these tubes may eventually enable the development of microvascular arrays for high-throughput assays of human microvascular barrier function.
2. Materials 2.1. Endothelial Cell Culture
1. Human dermal blood microvascular endothelial cells (BECs; Lonza) or human dermal lymphatic microvascular endothelial cells (LECs; Lonza). 2. Sterile phosphate-buffered saline (PBS; Invitrogen). 3. 60-mm-diameter tissue culture polystyrene dishes (Corning). 4. 0.1% gelatin from pig skin (Sigma) in PBS, filter-sterilized, autoclaved, filter-sterilized again, and stored at 4°C. 5. MCDB 131 medium (Invitrogen), supplemented with 10% fetal bovine serum (FBS; Invitrogen), 1% penicillin–streptomycinglutamine (Invitrogen), 1 mg/mL hydrocortisone (Sigma), 80 mM dibutyryl cyclic AMP (db-cAMP; Sigma), 25 mg/mL endothelial cell growth supplement (ECGS; Biomedical Technologies), 2 U/mL heparin (Sigma), and 0.2 mM L-ascorbic acid 2-phosphate (Sigma). 6. Complete medium (see Subheading 2.1, item 5) supplemented with 3% 70 kD dextran (dex; Sigma) and filter-sterilized (“3% dex media”). 7. Trypsin/EDTA (Invitrogen). 8. Dispase (Invitrogen), optional.
2.2. Formation of Microvascular Tubes
1. Culture media (see Subheading 2.1, item 5). 2. Trypsin/EDTA. 3. 1.7-mL microcentrifuge tubes (Costar). 4. 3% dex media (see Subheading 2.1, item 6). 5. Collagen gels with channels, assembled between silicone housings and glass coverslips (as described in 11.3.1–11.3.4 of (3)). 6. Silicone lids with attached tubing (as described in 11.3.5.2 of (3)).
Permeability of Microvascular Tubes
2.3. Measurement of Permeability
283
1. Alexa Fluor 594-conjugated bovine serum albumin (BSA594; Invitrogen). 2. Alexa Fluor 488-conjugated 10 kD dextran (dex-488; Invitrogen). 3. 3% dex media (see Subheading 2.1, item 6) supplemented with 50 mg/mL BSA-594 and 20 mg/mL dex-488. 4. A microscope environmental chamber (Zeiss) that maintains temperature at 37°C. 5. Image-processing software, such as ImageJ (NIH freeware). 6. Silicone blocks with channels (as described in 11.3.3.1 of (3)), with the same dimensions as the channel-containing gels in Subheading 2.2, item 5. 7. Silicone lids with tubing (see Subheading 2.2, item 6).
3. Methods As shown in Fig. 1, we form human endothelial tubes by seeding BECs or LECs into collagen gels that contain open channels and allowing the seeded cells to form confluent linings on the channels
Fig. 1. Schematic diagram of the formation of a perfused microvascular tube from a channel-containing collagen gel. Left panel, a channel within a collagen gel. The collagen is surrounded by a silicone housing and a glass substrate (3 ). The parameters “d ” and “ℓ ” refer to the diameter and length of the channel, respectively. Middle panel, a channel seeded with endothelial cells (ECs). Right panel, a seeded tube coupled to a lid that establishes high-flow perfusion. The tubing of the lid aligns with the centers of the inlet and outlet wells adjacent to the tube.
284
Price and Tien
under constant perfusion. Once the cells reach confluence, the lumen widens with the inlet and outlet diameters increasing by ~40 and ~10%, respectively (1, 2). Figure 2 illustrates a typical time-course of this expansion. Tubular widening takes place over 2 days, and we usually perform permeability assays on the third day after seeding. The operating principle behind our permeability assay is to measure the rate at which a fluorescently labeled macromolecule (bovine serum albumin or dextran) escapes from the endothelial lumen. This idea has been implemented previously for intact or explanted microvessels (5). Application of this assay to engineered endothelial tubes requires two modifications: First, because our tubes consist solely of an endothelial monolayer, they are not
Fig. 2. Maturation of tubes over a span of 3 days. The top row shows phase-contrast images of a middle segment of an unseeded channel. The second row displays spreading endothelial cells in a channel a few hours after seeding (day 0). The subsequent rows show changes in tubular morphology as the cells grow to confluence: By day 1, endothelial cells have formed a confluent monolayer. By day 2, the tube has widened. By day 3, morphological changes have stabilized.
Permeability of Microvascular Tubes
285
mechanically strong enough to withstand direct cannulation. The introduction of fluorescent molecules thus has a lag time on the order of several minutes, as a dead volume is flushed out of the inlet well and tubing. Second, the perfusing medium contains bleachable solutes (possibly, flavins) that can interfere with imaging, which limits the frequency of measurement. With these caveats in mind, we can obtain Pd for each macromolecule. Manual counts of focal leaks provide complementary data on the nonuniformity of leakage. Calibration of the permeability assay is absolutely essential to obtain meaningful values of Pd. The methods described below confirm that the fluorescence signal is proportional to the concentration of solute, and that light collection efficiency does not depend on the location of fluorophore (e.g., above or below the mid-plane of the tube). We use a standard inverted epifluorescence microscope (Zeiss), with a Plan-Neo 10×/0.30 NA objective, 1,388 × 1,040 resolution AxioCam HRm camera, and flat-field correction software. The fluorescent solutes we use are conjugated to Alexa Fluor dyes, which do not photobleach appreciably with our exposure doses. Our calculation of Pd assumes that the tube is cylindrical in geometry; this assumption is valid within each ~1-mm-wide measuring window. The assay also takes the perfusion rate into account to ensure that the lumenal concentration of fluorescent solute is constant during imaging. Imaging systems that do not satisfy the above requirements (e.g., by using highly bleachable dyes such as fluorescein) require extensive mathematical compensation to obtain accurate values of Pd (6, 13). The equation we use for Pd calculates the effective permeability coefficient, which includes diffusional and convective contributions (10, 14). Since the tubes are enclosed in an impermeable housing (silicone and glass), we assume that transendothelial water flux – and, therefore, convective transport – is negligible. The convective contribution to permeability cannot be ignored if a large water flux exists across the endothelium (e.g., due to a large transendothelial pressure), and a more complex analysis must be used (5, 14, 15). 3.1. Formation of Endothelial Tubes 3.1.1. Endothelial Cell Culture
3.1.2. Seeding Channels in Collagen Gels
1. Coat polystyrene dishes with gelatin for 40 min at room temperature. Wash with sterile water and dry. 2. Plate human BECs or LECs in the gelatin-coated dishes. Routinely passage confluent cultures at a 1:4 dilution using trypsin/EDTA or dispase. A confluent 60-mm-diameter dish typically provides enough cells to seed 3–4 channels. 1. Form collagen gels that contain single open channels, following the directions described in our recent work (1, 3). Condition the gels with 3% dex media for at least 1 h at 37°C
286
Price and Tien
by adding media to the inlet well and removing it from the outlet well. Do not allow any media to overflow on top of the silicone housing as this media will interfere with perfusion under high pressures. 2. Treat a culture of BECs or LECs with trypsin/EDTA, collect the cells in ~1 mL dextran-free media, and pellet them at 200 × g for 2 min (see Note 1). 3. Aspirate the media from the vial of pelleted cells and gently resuspend the cells in 20 mL of 3% dex media. 4. Aspirate media from the inlet and outlet wells of a conditioned collagen gel. 5. Add 1–2 mL of cell suspension to each of the inlet and outlet wells. A steady, dense stream of cells should begin flowing through the collagen channel. 6. Modulate the flow velocity by tilting the dish that contains the collagen gel until the cells are nearly stationary within the channel. Hold the dish steady for 30–40 s to allow cells to settle and adhere to the collagen channel. We typically perform this step while constantly viewing the channel with a 10× objective on an inverted microscope. 7. If the seeding is sparse, level the dish to add more cell suspension to the channel as needed. Seeding should require ≤5 min for a single channel (see Note 2). 8. Gently wash the inlet and outlet wells twice with 3% dex media. Avoid disturbing the collagen, else the gel may detach from the surrounding silicone or glass. 9. Add a large drop of 3% dex media to the inlet well and add a small drop to the outlet. Place the seeded tube in a 5% CO2 incubator at 37°C. The cells should visibly spread within the channel after 15 min (see Note 3). Wait at least 1 h before establishing perfusion under high pressure (see below). 3.1.3. Perfusing Seeded Tubes
1. Assemble perfusion lids and associated tubing and dishes by following step 13.3.5.2 described in our recent review (3). 2. Fill reservoir dishes with 30 mL of warm 3% dex media and aspirate media through the tubing. 3. Place a lid on a seeded channel to establish fluidic contact between the tubing and the inlet and outlet wells (see Note 4). 4. Set the height difference of the reservoirs to ~6 cm. A pressure difference of 6 cm H2O should yield a flow rate of 0.5-1 mL/h for subconfluent tubes and ~1 mL/h for confluent, widened ones. 5. Regenerate the pressure difference by pipetting media from the outlet reservoir to the inlet reservoir every ~12 h (see Note 5). The same media may be reused to perfuse a microvascular tube
Permeability of Microvascular Tubes
287
for 3 days (by which we usually perform the permeability assay), after which the media should be replaced (see Notes 6–11). 3.2. Control Experiments and Calibration of Permeability Assay
All control experiments and calibrations should be performed under conditions as close as possible to those used in the actual permeability assay (e.g., with environmental chamber warmed to 37°C, in the dark, etc.).
3.2.1. Signal Detection
1. Place a drop of BSA-594 or dex-488 between two #1½ glass coverslips. Capture fluorescence images using flat-field correction, and measure the average light intensity (e.g., with ImageJ). Focus at various planes to ensure that the intensity of the collected light does not change with focus; the range of focus should span ~1 mm. 2. Repeat the above step with different solute concentrations to create a standard curve of fluorescence intensity vs. concentration. The solute concentrations used in all subsequent sections should be within the linear range of this curve. We typically find a linear range of 0.2–50 mg/mL for BSA-594 and dex-488.
3.2.2. Determination of Assay Sensitivity
1. Form channels in silicone by following step 11.3.3.1 of our review on microfluidic gels (2). 2. Establish perfusion exactly as described for endothelialized tubes (see Subheading 3.1.3) with fluorescent 3% dex media. 3. Capture several fluorescence images in succession. Since silicone is impermeable to proteins, the intensity of the images should theoretically not vary over time. The range of values indicates the inherent noise of the imaging system (e.g., due to lamp flicker) and sets a bound on the precision of our permeability measurements (see Note 12).
3.2.3. Determination of Lag Time
1. Establish perfusion in a silicone channel as in Subheading 3.2.2, but with nonfluorescent 3% dex media. 2. Supplement the media in the inlet reservoir with 50 mg/mL BSA-594 and 20 mg/mL dex-488. 3. Capture fluorescence images every minute as the fluorescent solute flows through the tubing into the inlet well and through the silicone channel. The lag time is the time at which the intensity of fluorescent solute reaches a maximum (within the sensitivity determined in Subheading 3.2.2). Increasing the length of tubing, increasing the size of the inlet well, or decreasing the perfusion rate will increase the lag time. With our standard setup and perfusion conditions, we observe a lag time of <20 min.
288
Price and Tien
3.2.4. Determination of Imaging Frequency That Minimizes Photobleaching
1. Perfuse a seeded tube as described in Subheading 3.1.3 (do not add any fluorescent solutes to the media).
3.3. Measurement of Diffusional Permeability Coefficient Pd
1. Warm the microscope environmental chamber to 37°C.
2. Capture two consecutive fluorescence images of the tube (see Note 13). If the second image is photobleached (i.e., has a lower average intensity) compared to the first one, repeat the process in another tube with a longer delay between capturing the two images. The recovery time is the time interval that routinely yields the same intensities (discounting noise) in consecutive images. We have found 6 min to be sufficient for recovery from photobleaching with our perfusion conditions.
2. Capture background fluorescence images of a perfused tube (do not add any fluorescent solutes to the media). The image should appear black. 3. Draw a measuring window that extends the width of the collagen gel, with the lumen of the tube in the middle, as shown in Fig. 3a. The average signal in the measuring window is the background fluorescence intensity Ib. 4. Replace the media in the inlet reservoir with fluorescent media and gently pipette to mix the fluorescent media with any residual media. Be careful not to introduce bubbles into the inlet tubing. 5. Place the tube back in the incubator for 10 min (half of the lag time found in Subheading 3.2.3). 6. Transfer the tube back to the environmental chamber for an additional 10 min (the remainder of the lag time). This time allows the tube to recover from any agitation during the transfer. 7. Capture the initial fluorescence images of the tube (see Note 14). We obtain images for BSA-594 and dex-488 in parallel, which allow us to use the same focal leak data for both solutes and decreases the time needed for the permeability assay (see Note 15). We typically image a region ~4.5 mm downstream from the inlet. 8. Wait 6 min (the recovery time found in Subheading 3.2.4) and capture another set of images of the tube. Repeat as desired. 9. Draw a measuring window as shown in Fig. 3b. The positions of the measuring windows with respect to the tube should be exactly the same at both time points. The average initial intensity in the measuring window for a given fluorophore is I1 and the average intensity after 6 min is I2. If the tube has nonuniform leakage (i.e., focal leaks), choose a location for the measuring window as far from the nonuniform regions as possible to minimize their effects on the calculation of Pd (see Fig. 3c).
Permeability of Microvascular Tubes
289
Fig. 3. Quantifying the barrier function. (a) Fluorescence image of the background intensity. The tube is indicated by the dotted lines in the middle of the image, while the collagen borders are represented by the dotted lines at the top and bottom. The average intensity within the measuring window (gray box) is the background intensity Ib (scale from 0 to 16,383 using a 14-bit CCD). (b) Fluorescence images of a tube at 0 and 6 min (solute is BSA-594). The average intensities in the measuring window for each time point are I1 and I2, respectively. The diameter of the tube within the measuring window is d. (c) Fluorescence images of a tube with poor barrier function at 0 and 6 min (solute is BSA-594). The arrows indicate focal leaks along the vessel wall. The measuring window is placed away from leaks.
10. Compute Pd as follows (5): Pd =
1 I 2 − I1 d • I 1 − I b ∆t 4
where ∆t = 6 min is the recovery time (determined in Subheading 3.2.4) and d is the diameter of the tube in the middle of the measuring window. Figure 3b shows a sample calculation of Pd (see Notes 12 and 16–19). 3.4. Quantifying Focal Leaks
1. Maximize the contrast in the images taken for the permeability assay. Examine the images for any nonuniformities (typically, regions of excess intensities). Figure 3c shows a microvascular tube that has focal leaks in both the initial and final images. Since counts of focal leaks can be somewhat subjective, the same experimenter should count the number of leaks for all images. 2. Sum the number of focal leaks in the images at both time points, irrespective of the position of the focal leaks. If the same focal leak is present in the initial and final images, then that leak is counted twice.
290
Price and Tien
3. Compute the number of focal leaks per mm of microvascular tube (FL) as follows (2): FL =
n leaks nimages ⋅ l
where nleaks is the total number of focal leaks summed over the different time points, nimages = 2 is the number of time points imaged, and l is the length of the tube in the viewing window. Figure 3c shows a sample calculation of FL. 3.5. Repeating Permeability Measurements
1. Remove fluorescent media from the inlet reservoir and exhaustively wash the reservoir with 3% dex media without fluorescent solutes. 2. Place the tube back in the incubator for 2–3 h. 3. Remove the media from the outlet reservoir and exhaustively wash the reservoir with 3% dex media without fluorescent solutes. 4. Replace the media in the inlet reservoir, and perfuse the tube for ≥12 h before repeating the permeability assay. A new background measurement must be taken, since residual fluorescent molecules are always present in the collagen gel or in the endothelium.
4. Notes 1. Cells that are overtrypsinized or harvested from a subconfluent dish usually seed poorly. 2. If cells clog the tube during seeding, it is sometimes possible to dislodge the clog by gently tapping the dish. Vigorous tapping can cause the collagen gel to detach from the surrounding silicone and glass. We strongly recommend seeding a fresh channel if a clog forms. It is always possible to reseed a sparsely seeded tube but very difficult to repair an overseeded one. 3. If seeded cells do not adhere to the collagen gel or do not begin to spread within 15 min of seeding, the channel may need to be conditioned with media for a longer period of time. If cells do not adhere to well-conditioned channels, then the particular lot of endothelial cells may be unusable for making tubes. We always use primary cells and have found that the immortalized endothelial cell line HMEC-1 (16) does not form well-behaved tubes. 4. The tubing should never become crimped or have air bubbles or dust in it, since the flow rate will decrease and the barrier function can weaken.
Permeability of Microvascular Tubes
291
5. The media in the inlet reservoir should be renewed regularly so that the driving pressure head does not decrease too much as media flows through the tube. 6. On day 0 (the day of seeding), the endothelial cells should adhere and completely spread out within 1–2 h of seeding as shown in Fig. 2. 7. By day 1, the cells should have grown to confluence within the tube. Gaps should not be visible along the endothelial wall in profile. 8. By day 2, the tube should have an expanded diameter although tubes with weak barrier function may not widen much. Expansion of the tube is not uniform: the region near the inlet expands more than that near the outlet, causing the tube to have a gradually sloping conical geometry. The cells should have a cobblestone appearance without any preferential alignment. 9. By day 3, the tube should have stabilized and usually will not expand further. 10. We have successfully maintained tubes for ≥1 week. Perfusion rates slowly decrease in tubes maintained for longer than 4 days, for reasons we do not yet understand. 11. Perfusing the tube under conditions that greatly weaken barrier function may prevent it from widening or growing to confluence and may cause the profile of the tube to appear wavy. We normally culture tubes under standard conditions for 1–2 days and then switch to the experimental condition (e.g., low flow rates) only after the tubes reach confluence. 12. In our hands, the dynamic range of this permeability assay is 5 × 10−8–5 × 10−6 cm/s. Near the lower limit, lamp flicker limits the sensitivity of the assay. Near the upper limit, the tubes invariably contain many focal leaks, and measurements of background intensity Ib become inaccurate. 13. An acute change in the perfusion rate can alter the barrier function of the tube, so it is important that the height difference between the inlet and outlet reservoirs be kept constant. While transporting the tube (e.g., to and from the incubator, to the microscope, etc.), we always hold the inlet and outlet reservoirs in the same positions with respect to one another and with respect to the tube. 14. Cells may start to grow between the collagen gel and the silicone or glass, causing a decrease in perfusion or a leak of fluorescent material around the gel when performing the permeability assay. This ingrowth can be caused by detachment of the collagen gel and may gradually develop in tubes maintained for ≥1 week.
292
Price and Tien
15. The selectivity of the endothelial barrier can be calculated by taking the ratios of Pd for dex-488 over Pd for BSA-594. The higher the selectivity, the stronger the barrier function. 16. As a general principle, tubes that do not widen will display large permeability coefficients. 17. For LEC tubes perfused under the standard conditions −7 cm s described in Subheading 3.1.3, Pd averages 3.0+−1.0 0.7 × 10 +0.9 −7 for BSA-594 and 6.2−0.8 × 10 cm s for dex-488 (geometric means ± 95% CI) (2). Surprisingly, we have noticed that BEC tubes tend to have slightly higher Pd compared with LEC tubes under identical perfusion conditions. 18. Comparisons of Pd should always use a nonparametric test (e.g., Mann–Whitney U-test), since the data do not follow a normal distribution. 19. Typically, permeability values from ~10 pairs of tubes are needed to detect a twofold difference in Pd with statistical significance of 0.05.
Acknowledgments We thank Bingmei Fu for many helpful discussions. This work was supported by the National Institute of Biomedical Imaging and Bioengineering (award EB005792). References 1. Chrobak, K. M., Potter, D. R., and Tien, J. (2006) Formation of perfused, functional microvascular tubes in vitro. Microvasc. Res. 71, pp. 185–196. 2. Price, G. M., Chrobak, K. M., and Tien, J. (2008) Effect of cyclic AMP on barrier function of human lymphatic microvascular tubes. Microvasc. Res. 76, pp. 46–51. 3. Price, G. M. and Tien, J. (2008) Subtractive methods for forming microfluidic gels of extracellular matrix proteins, in Microdevices in Biology and Medicine (Bhatia, S. N. and Nahmias, Y., eds.), Artech House, Boston, MA, pp. 235–248. 4. Michel, C. C. and Phillips, M. E. (1987) Steady-state fluid filtration at different capillary pressures in perfused frog mesenteric capillaries. J. Physiol. 388, pp. 421–435. 5. Huxley, V. H., Curry, F. E., and Adamson, R. H. (1987) Quantitative fluorescence microscopy on single capillaries: a-lactalbumin transport. Am. J. Physiol. 252, pp. H188–H197.
6. Lichtenbeld, H. C., et al. (1996) Perfusion of single tumor microvessels: application to vascular permeability measurement. Microcircu lation 3, pp. 349–357. 7. Fu, B. M. and Shen, S. (2003) Structural mechanisms of acute VEGF effect on microvessel permeability. Am. J. Physiol. Heart Circ. Physiol. 284, pp. H2124–H2135. 8. Curry, F. E., Huxley, V. H., and Sarelius, I. H. (1983) Techniques in the microcirculation: measurement of permeability, pressure and flow, in Techniques in the Life Sciences; Physiology Section, Vol. P3/I: Techniques in Cardiovascular Physiology – Part I (Linden, R. J., ed.), Elsevier, New York, NY, pp. 1–34. 9. Baxter, L. T., Jain, R. K., and Svensjö, E. (1987) Vascular permeability and interstitial diffusion of macromolecules in the hamster cheek pouch: effects of vasoactive drugs. Microvasc. Res. 34, pp. 336–348. 10. Crone, C. and Levitt, D. G. (1984) Capillary permeability to small solutes, in Handbook of
Permeability of Microvascular Tubes Physiology; Section 2: The Cardiovascular System, Vol. IV: Microcirculation (Renkin, E. M. and Michel, C. C., eds.), American Physiological Society, Bethesda, MD, pp. 411–466. 11. Curry, F. E., Michel, C. C., and Mason, J. C. (1976) Osmotic reflexion coefficients of capillary walls to low molecular weight hydrophilic solutes measured in single perfused capillaries of the frog mesentery. J. Physiol. 261, pp. 319–336. 12. Casnocha, S. A., et al. (1989) Permeability of human endothelial monolayers: effect of vasoactive agonists and cAMP. J. Appl. Physiol. 67, pp. 1997–2005. 13. Yuan, F., et al. (1993) Microvascular permeability of albumin, vascular surface area, and vascular volume measured in human
293
adenocarcinoma LS174T using dorsal chamber in SCID mice. Microvasc. Res. 45, pp. 269–289. 1 4. Curry, F. E. (1984) Mechanics and thermodynamics of transcapillary exchange, in Handbook of Physiology; Section 2: The Cardiovascular System, Vol. IV: Microcirculation (Renkin, E. M. and Michel, C. C., eds.), American Physiological Society, Bethesda, MD, pp. 309–374. 15. Fu, B. M. and Shen, S. (2004) Acute VEGF effect on solute permeability of mammalian microvessels in vivo. Microvasc. Res. 68, pp. 51–62. 16. Ades, E. W., et al. (1992) HMEC-1: establishment of an immortalized human microvascular endothelial cell line. J. Invest. Dermatol. 99, pp. 683–690.
Chapter 18 Microarray Bioinformatics Robert P. Loewe and Peter J. Nelson Abstract Bioinformatics has become an increasingly important tool for molecular biologists, especially for the analysis of microarray data. Microarrays can produce vast amounts of information requiring a series of consecutive analyses to render the data interpretable. The direct output of microarrays cannot be directly interpreted to show differences in settings, conditions of samples, or time points. To make microarray experiments interpretable, it is necessary that a series of algorithms and approaches be applied. After normalization of generated data, which is necessary to make a comparison feasible, significance analysis, clustering of samples and biological compounds of interest and visualization are generally performed. This chapter will focus on providing a basic understanding of the generally approaches and algorithms currently employed in microarray bioinformatics. Key words: Microarray, Bioinformatics, Normalization, Clustering, SAM, RMA, PCA
1. Introduction Microarrays are commonly used to analyze DNA, mRNA, proteins, and other biological compounds (1–9). This general approach is being applied at multiple levels of research and has already led, for example, to a better general understanding of disease processes and drug action. Array technology is being used for the identification of drug targets and even the proteins with which the drugs interact. The ability to use array methods in place of existing technologies allows researchers to perform experiments faster and more cheaply, but this also requires a significant refocusing of efforts on the analysis of the results of the microarray experiments. DNA microarrays are represented by ordered sets of DNA molecules each with a defined sequence. Each individual feature is applied to the array at a defined location and thus the identity Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0_18, © Springer Science+Business Media, LLC 2011
295
296
Loewe and Nelson
of the DNA molecule is fixed. Microarray analysis allows the simultaneous detection of expression of thousands of genes in a single sample. While each sample can be seen as fundamentally similar, in reality they contain fundamental variables, which must be controlled and normalized in order to identify differences in biologic parameters. For example, variations in the printing of the array, hybridization efficiency, gene annotation, general problems with sample preparation are just some of the issues that have to be addressed during analysis. The accurate interpretation of microarray data requires the application of a series of bioinformatics methodologies as well as analysis and visualization of subsequent data. Bioinformatics as it pertains to DNA microarrays can be roughly structured into the following steps: ●
Normalization (essential to make samples comparable)
●
Statistical analysis (finding similarities and differences)
●
Visualization (making it possible to see similarities and differences)
The following chapter will provide an overview of a series of methods and accompanying algorithms used to normalize array data and allowing adjacent analysis of samples.
2. Methods 2.1. Generating Data and Minimal Requirements
Before starting an array experiment care should be taken in the organization of the experiment and preparation of the samples to ensure a robust data set for the subsequent analysis. Samples should be generated in at least biological triplicates (essential for statistical relevance), while the number of technical replicates ultimately depends on the specific platform used (10). Data generation refers to the experimental phase of the microarray workflow and in the case of gene expression, analyses can be summarized in the following steps (11). 1. Slide production 2. Sample preparation and hybridization: (a) RNA isolation from biological samples (b) cDNA synthesis and labeling (c) Hybridization 3. Data capturing: (a) Slide scanning (b) Spot finding (gridding)
Microarray Bioinformatics
297
(c) Spot quantification (segmentation, intensity extraction, and background correction) (d) Data report and export As a first principle, it must be assumed that good laboratory procedures and handling ensure that good quality RNA is isolated from the sample and that efficient complementary DNA (cDNA) are generated (see Note 1). If sample amount is scarce, samples (if they are similar) can be pooled or a preamplification can be performed prior to hybridization. Preamplification is often used for the analysis of rare RNA and DNA samples (e.g., biopsy samples). Importantly, preamplification can often have an effect on reproducibility and the intercomparison of some genes, but good results of the application of this approach have been reported (12–14). The hybridization protocols themselves and reagents used for the hybridization are generally provided by the commercial producers of DNA arrays and include detailed troubleshooting protocols. The quality of subsequent data capture is essential in supplying the researcher with usable information for analyses. Slide scanning can be compared to taking a photo and is – in case of fluorescence – mainly influenced by the exposure time. The resulting data or image is further analyzed to obtain signal intensity for the biological “event.” Spot finding is performed by gridding. This identifies the region on the array where the signal of interest is found. As spots on microarrays are not completely uniform in their shape and signal deviation, a segmentation of the spot is often required to extract the “real” intensity. The values obtained can then be corrected for background signals using spot quantification or, at later stages in data processing, by negative controls provided on the array. Depending upon the specific platform used, different file formats are used to represent the exported raw data. The data can be unique for the individual manufacturer (e.g., CEL files for Affymetrix), images (e.g., JPEG, GPR, etc.), or other similar export options. These data then need to be read-out for the individual signal obtained for each probe. This means that a signal-tonoise ratio has to be employed to gather the “most likely” signal. In some cases, the background signal seen on the glass slide or filter can be used. In other settings, probe controls that do not bind to the sample are used. In the case of images, the probes need to be allocated and annotated to transcripts, DNA sequences, proteins, or other biological compounds. This results in a list of probes and their corresponding measured intensities. The next step in analyses is normalization. MIAME (minimum information about a microarray experiment) was originally established by Brazma et al. (15). This standard for array analysis provides essential details for microarray
298
Loewe and Nelson
experiments that helps make comparison and independent interpretation possible. Data description according to MIAME is often required to publish findings. MIAME can be summarized into six contributing elements: (1) the raw data from all hybridization events (output files) must be provided, (2) the final processed (normalized) data for the set of hybridizations used in the experiment or study (e.g., the gene expression data matrix used to draw the conclusions from the study) is included, (3) the essential sample annotation used including experimental factors and their values (e.g., compound and dose in a dose–response experiment), (4) the experimental design including sample data relationships (e.g., which raw data file relates to which sample, which hybridizations are technical, which are biological replicates), (5) sufficient annotation of the array (e.g., gene identifiers, genomic coordinates, probe oligonucleotide sequences or reference to commercial array catalog number), and (6) the essential laboratory and data processing protocols (e.g., what normalization method has been used to obtain the final processed data). Other standards have been developed by the Microarray Gene Expression Data Society (MGED) (16) or are modified MIAME (e.g., Gene expression omnibus GEO) (17). MGED is an international organization of biologists, computer scientists, and data analysts that is focused on establishing standards for data quality, management, annotation, and exchange within the life sciences and biomedical communities. Established methods that allow the breakdown of data, or meta-analysis, are also often used by microarray databases (18). For the most part, these norms are applied for gene expression profiling, but can also be used on other microarray systems (19). In the following discussion, we will try to outline the various parameters that should be addressed in the analysis of array data. We have also included information about the specific algorithms that are often used for the analysis. While the detail may not be sufficient for a true bioinformatician, biologists concerned with the theoretical basis for these methods, may find them of use. 2.2. Normalization
There are many sources of systematic variation in microarray experiments that affect the measured expression levels. Before any mathematical methods are employed, the data generated need to be normalized. Normalization removes the impact of nonbiological influences on the data (e.g., input amount, background noise, imaging of the microarray, facility where the analysis took place, and others) allowing a more accurate representation of the results. Normalization also facilitates direct comparison between arrays and can even allow, in some circumstances, the comparison of data event between platforms. A series of methods have been developed to make the intercomparison of samples possible. In the past, mismatched probes
Microarray Bioinformatics
299
were used in Affymetrix arrays; however, this approach has been largely abolished. The following discussion concentrates on spikein normalization, the use of housekeeping compounds and other models (e.g., RMA, MBEI, ANOVA). After subtracting the background from each “real” signal, the data points gathered are still in disarray. Using the untouched values for further analysis would result in comparison of dissimilar data (20). Some platforms use dual-color measurement to normalize values. In essence, a color exchange or swap (all samples are used with both dyes; dye effects are canceled out by calculating with both color values) is used to ensure comparability (21–23). Spike-ins are now frequently used to control for processes such as sample isolation, cDNA/cRNA arrays reverse transcription, hybridization, and signal detection. Internal standards allow observation of potential loss of analyte during the workflow and make the inevitable signal diminishing effects computable. After calculating the overall mean internal standard (IS) value of all samples, the fraction of the internal standards per specimen (IS of the sample divided with the mean of all IS) can be used as calibrator (24, 25). Example: The signal intensity of IS was measured for arrays A–D. The following values were observed: IS (A ) = 2; IS (B) = 3; IS (C) = 4; and IS (D ) = 5.
The resulting mean [IS(A–D)] was calculated to be 3.5. The calibrator is therefore: Calibrator (A ) = 2 / 3.5; Calibrator (B) = Calibrator (C) =
3 ; 3.5
4 5 ; Calibrator (D ) = . 3.5 3.5
Each signal of each array is then divided by the appropriate calibrator. The same approach is used in regard to housekeepers, which depending on the platform used, can be represented by unchanged genes (DNA analysis), nonregulated genes (e.g., housekeeping genes for cDNA/cRNA arrays), equally present proteins (protein arrays) or other biological compounds which are not altered by processes changed in the settings of the study. The IS is replaced in the calculation by the observed housekeeper signal and calculated to be the calibrator in the same manner. To identify the least variant set (LVS) genes, Calza et al. concluded that the probe level must be analyzed (26). After RMA (later described in detail), c2 test must be computed: χ 2 = αˆ ′V −1αˆ ,
300
Loewe and Nelson
whereas αˆ is the vector of estimated αˆ i s and V is estimated covariance matrix. The array effect ai includes both the technical artifact ti and the real biological effect bi: α i = ti + bi .
When c2 statistics are compared among the genes, those with the smallest value are more likely to arise from genes with bi = 0. This method allows finding housekeeping genes that are not prior known before normalization. In earlier versions of the microarrays from the Affymetrix platform, Li and Wong introduced the model-based expression index (MBEI) (27). Although not often used now, the algorithm has provided the basis for new approaches (e.g., (28)). The estimation procedure is based on a model of how the probe intensity values respond to changes in the expression levels of the gene. By definition, qi is the expression index for the gene in the ith sample. Assuming that the intensity value of a probe will increase linearly as qi increases, but that the rate of increase will be different for different probes. It is also assumed that within the same probe pair (the PM has a corresponding MM), the PM (perfect match, 100% sequence identity, the oligo aligns completely to the targeted transcript) intensity will increase at a higher rate than the MM (mismatch, is a less than 100% sequence identity, the oligo aligns not completely to the targeted transcript) intensity. Here is an example of a simple model: MMij = v j + qi a j + e , PMij = v j + θiα j + θiφ j + ε .
Here PMij and MMij denote the PM and MM intensity values for the ith array and the jth probe pair for this gene, nj is the baseline response of the jth probe pair due to nonspecific hybridization, aj is the rate of increase of the MM response of the j th probe pair, fj is the additional rate of increase in the corresponding PM response, and e is a generic symbol for a random error. The rates of increase are assumed to be nonnegative. The model for individual probe responses implies an even simpler model for the PM–MM differences: y ij = PMij − MMij = qi fj + eij . An additional method to normalize data employs the application of a convolution model such as found in RMA (robust multiarray average) (29). It is frequently used for Affymetrix arrays and consists of three steps: background correction by convolution,
Microarray Bioinformatics
301
nonlinear normalization on the probe level, and a robust multichip summarization method (hence the name). As described by Harbron et al. (30), RMA consists of three steps. – Background correction: probe level data for each microarray are background corrected individually using a probabilistic model. – Quantile normalization: the background corrected probelevel data from each microarray are normalized to common set of quantiles, derived from background corrected data from all microarrays analyzed. – Expression calculation. Estimated separately for each probeset using median polish on the linear model: log 2 ( N ij ) = Pj + I i + ε ij ,
where Ii is the logarithmic intensity for the ith microarray, Nij is the background corrected, and normalized intensity of the jth probe of the ith microarray, Pj is the effect of the jth probe in the probeset, and eij is an error term. Further details and the fundamental outline of the method are described in ref. 29. Another means to describe RMA is shown in the following example. The observed signal (S) consists of two discrete components, X is the signal from the binding biological compound and Y is the background signal ⇒ S = X + Y; the corrected value for each probe is given as an expected value of X: f (a / b ) E ( XIS ) = a + b , Φ (a / b )
φ represents the computation for the density function. F provides values for the distribution function. The signal is then log 2 transformed and a quantile normalization is carried out. This is similar to what occurs in median polish. The Median polish used in RMA is based on Holder et al. (31) assumptions: Assuming that yij = µ i + α j + ε ij , where as α1 + α 2 + + α n = 0,
mi is the gene expression for the probe set on microarray i, aj values the probe affinity for the jth probe in the probe set. And eij gives the residual for the jth probe on the ith microarray. The magnitude of array specific biases can be dependent on expression levels. To make groups of microarrays comparable the following is computed:
302
Loewe and Nelson
Med ik = median( Ai1k ,…, Ainik ) the median array effect associated with probe set k for all ni arrays in treatment group i. In essence, the total sum of all expression values per array is compared and again a calibrator is used to make the samples comparable. This is only possible for high numbers of probes because in low-density arrays (low number of probes) a higher multitude of dysregulated compounds are present. This means that the majority of probes will show very different values as a specific subset of compounds are analyzed (e.g., for inflammation only compounds linked to the immune system). The values that result can undergo significance testing and visualization. Analysis of variance (ANOVA) can also be used to normalize microarray data (32). ANOVA is performed using a collection of statistical models in which the observed variance is partitioned into components due to different independent variables (see below). The technique is also known as Fisher’s ANOVA and is displayed as a Fisher’s F-distribution value to test for statistical significance. One-way ANOVA is used to test for differences among two or more independent groups, but is typically used to test for differences among at least three groups as a two-group analysis can be examined using a T-test. When means are compared, the T-test and the F-test are equivalent. The relation between ANOVA and t is given by F = t 2. In ANOVA analysis, if ygijk denotes intensity values for the gene g, labeled by the dye j, in variety (or sample) k on the array i, an ANOVA model can be written as:
( )
Log ygijk = µ + A j + D j + Vk + Gg + ( AG )gi + (VG )gk + ( DG )gi + e gijk .
In this model, variations arising from both biological differences and unwanted variations (due to non-biological effects) are taken into consideration. In particular, biological differences between different sample types are accounted for by the gene–variety interaction effect [(VG)gk]; unwanted variations are accounted by array effect (Aj), dye effect (Dj), variety effect (Vk), and gene effect (Gg), as well as gene–array interaction effect [(AG)gi] and gene– dye effect [(DG)gi], respectively. To be more specific: m represents the expression level for the gene. egijk is an independent identically distributed error term with zero mean. The array effect (Aj) accounts for variations between arrays, which arise from differential hybridization efficiencies across arrays. Differences in labeling efficiency between the two fluorescent dyes used in some analyses, represents an important variable. In this instance, a constant adjustment is commonly applied to force the distribution of the log-ratios to have a median of zero for each slide. The dye effect (Dj) accounts for variations between fluorescent
Microarray Bioinformatics
303
dyes, which may arise from the distinct properties of different dyes that cause genes labeled with one dye to have differential intensities with respect to the same genes labeled with a second dye. The variety effect (Vk) accounts for overall differences between different sample types (e.g., there are more mRNAs in sample a than those in sample b), which can be caused by intrinsic biological differences between different samples (e.g., the transcription machinery in sample a is more active than that in sample b) or by pipetting errors in the microarray experiment. The gene effect (Gg) accounts for variations of individual genes spotted on the arrays. The gene–array interaction effect [(AG)gi] accounts for the effect of individual spots on a particular array, which is in essence the spot effect. The gene–dye interaction effect [(DG)gi] accounts for variations caused by gene–dye interaction (e.g., some genes may incorporate one dye more efficiently than another dye). 2.3. Significance Calculation
The significance of the dysregulation present in samples can be addressed by the application of a number of corrective algorithms. Dysregulation is in essence, the difference in expression values of the biological compound relative to the sample groups. For example, in one group, the signal from one of the probes is immensely high, and in another group it is more or less abolished. This discrepancy in signal is deemed to be significant if the values of the different samples of all groups do not converge. That is, there is no overlay of the values between the groups (see Note 2.) As previously discussed, ANOVA (32, 33) is represented by a series of algorithms including ANOVA, F-test, Kruskal–Wallis algorithm, and linear-fixed effects models. The F-test and Kruskal– Wallis are often used to compare a set of conditions. ANOVA is most effective if groups are of the same size. The minimum group size is three. The F-statistic is basically the ratio of two variances ⇒ F = var(1)/var(2). If the variance in both groups is the same, the ratio is one and there is no observed difference. Kruskal– Wallis is similar to the F-test, but it uses the direct expression of values. Each gene is ranked and sorted according to its expression value. The algorithm calculates differences in the rank number for each group and then proceeds as the F-test. The linear model of fixed effect models is the same as the ANOVA normalization. Another method to identify significantly dysregulated genes is referred to as “significance analysis of microarrays” (SAM) (34). d (i) =
x I (i) − xU (i) s (i) + s 0
SAM makes use of nonparametric statistics as array data may not always show a normal distribution. The response variable describes, and groups, the data based on experimental conditions. In SAM, the repeated mapping of the elements of a set to other elements
304
Loewe and Nelson
of the same set is used to determine if the expression of any gene is significantly related to the specific question being asked. The use of this permutation-based analysis takes into account correlations in genes and avoids parametric assumptions about the distribution of individual genes. SAM identifies statistically significant genes by carrying out gene-specific analyses, where the relative difference [d(i)] is compared to its distribution following random permutation of the sample groups (e.g., treatment vs. nontreatment). xI (i ) and xU (i ) are defined as average levels of expression for gene (i) in states I and U. The gene-specific scatter s (i) is standard deviation of repeated measurements: s (i) = a ∑ [xm (i) − x I (i)]2 + ∑ [xn (i) − x U (i)]2 . n m For each d(i), a certain proportion of all genes in the permutation set (control set) will be found to be “significant” by chance, and this parameter is subsequently used to calculate a “false discovery rate” (FDR). This is presented as a q-value for each gene in the final list of significantly dysregulated genes. The q-values are influenced by the variability in the whole data set. This implies that changes in the entire data set composition will affect the d(i) distribution in the permuted control set and thus the q-value assigned to a given gene. SAM can have an advantage over other techniques (such as ANOVA or Bonferroni), which assume an independence or equal variance of genes. Another frequently used method is fold change calculation. The mean values of each compound are measured and then divided by their counterpart in the other group. The resulting values show fold change where everything upregulated above a ratio of 2.0 (depending on the platform) and downregulated below a ratio of 0.5 is defined to be of biological significance. Fold change on its own does not include differences in the distribution of values or standard deviation of each group. Therefore, it should not be employed without additional consideration. Modifications of this method can result in statistically interpretable data (35, 36). 2.4. Clustering and Visualization
Grouping genes having similar expression patterns is referred to as gene clustering. This is a useful tool for extracting and visualizing underlying biological phenomena in gene expression data. Clustering produces the assignment of objects into groups. The data set is partitioned into subsets (clusters), so that the data in each subset share some common trait – often proximity according to some defined distance measure. The most frequently used types of data clustering are hierarchical clustering, k-means and fuzzy c-means. A series of clustering procedures are commonly applied in microarray gene analysis. For the most part, clustering
Microarray Bioinformatics
305
algorithms are useful to visualize the data: depending on the employed method, the detected clusters can be displayed using a heatmap or dendrogram. An additional approach is represented by principle component analysis (PCA). Pathway illustration and geneset imaging can also be utilized (37). Cluster analysis is extensively described in ref. 38. Clustering can be supervised or unsupervised. This results in different outputs. Many different types of unsupervised clustering algorithms are commonly applied for the visualization of data. These include hierarchical clustering, k-means clustering, selforganizing maps, hill climbing, and simulated annealing. All these approaches use the same three basic tasks: 1. Pattern representation – patterns or features in the data 2. Pattern proximity – a measure of the distance or similarity defined on pairs of patterns 3. Pattern grouping – methods and rules used in grouping the patterns Supervised methods make use of additional information that is utilized together with the gene expression data. The choice of analysis methods will influence the results as well as their interpretation; therefore, it is important to be familiar with each method, its scope and limitations. The following section addresses these general methods with special reference to applications for pharmacogenomics. 2.4.1. Distance Measure
An important step in any clustering approach is the selection of a distance measure, which will determine the proximity of the values from multiple elements. This will influence the shape of the clusters, as some elements may be close to one another according to one distance, and further away according to another. Goldstein et al. (39) have summarized various distance measurement algorithms and they include the following: Name d (xi , x j ) p
Canberra
xik − x jk
∑x k =1
ik
+ x jk 1/ 2
2 p Euclidean ∑ (xik − x jk ) k =1 p
Manhattan
∑x k =1
ik
− x jk
Maximum max1≤ k ≤ p xik − x jk
306
Loewe and Nelson
The Euclidean distance (also called distance as the crow flies or two-norm distance) is the most frequently used method to calculate distance for subsequent cluster analysis. 2.4.2. Hierarchical Clustering
Hierarchical clustering builds (agglomerative), or breaks up (divisive), a hierarchy of clusters. The traditional representation of this hierarchy is a tree (called a dendrogram), with individual elements at one end and a single cluster containing every element at the other. Agglomerative algorithms begin at the biological compound level (e.g., gene expression value of a single gene), whereas divisive algorithms begin at the sum of all biological compounds (e.g., whole genome as starting point and following partition into subgroups). Eisen et al. (40) were one of the first to use clustering – in this case, hierarchical clustering – for microarray analysis and used the following assumption. For the gene similarity metric, a form of correlation coefficient can be used. Where Gi equals the (log-tansformed) primary data for gene G in condition i. For any two genes X and Y observed over a series of N conditions, a similarity score can be computed as followed: S (X, Y) =
1 N
X i − X offset Yi − Yoffset , FX FY i =1,N
∑
where ΦG =
∑
i =1, N
(Gi − Goffset )
2
N
.
When Goffset is set to the mean of observations on G, then FG becomes the standard deviation of G. S(X,Y ) is defined as exactly equal to the Pearson correlation coefficient for observations of X and Y. Values of Goffset, which are not the average over observations on G, are used when there is an assumed unchanged or reference state represented by the value of Goffset against changes are to be analyzed; in all the examples represented here, Goffset is set to 0, corresponding to a fluorescence ratio of 1.0. Hierarchical clustering will provide a number of clusters, which may not necessarily coincide with the original number of samples or biological compound subgroups (as predicted by the researcher). It will however show if subgroups are present. 2.4.3. k-Means Clustering
The k-means algorithm assigns each point to the cluster whose center (also called centroid) is nearest. The center is the average of all the points in the cluster – that is, its coordinates are the arithmetic mean for each dimension separately over all the points in the cluster.
Microarray Bioinformatics
307
The algorithm steps are described in ref. 38 are summarized as: – Choose the number of clusters, k. (see Note 3). – Randomly generate k clusters and determine the cluster centers, or directly generate k random points as cluster centers. – Assign each point to the nearest cluster center. – Recompute the new cluster centers. – Repeat the two previous steps until some convergence criterion is met. The main advantage of this algorithm is the simplicity in the programming resulting in relatively high speed, which allows the running of large data sets. Its main disadvantage is that it does not compute the same result for each run since the resulting clusters depend upon the initial random assignments. In addition, the fact that the number of clusters is predetermined makes interpretation, if unforeseeable subgroups are present, difficult. The approach minimizes intracluster variance, but does not ensure that the result has a global minimum of variance. MacLachlan et al. provides a conclusive summary of this approach (41): k is defined as the number of clusters set up initially for the analysis. It seeks to find k clusters that minimize the sum of the squared Euclidean distances between each observation yi and its respective cluster mean; that is, it seeks to minimize the trace of W, tr W, where: k
n
(
W = ∑∑ zij y j − y j i =1 j =1
)(y
j
− yj
)
T
is the pooled within-cluster sums of squares and products matrix and n
yi =
∑y j =1
j
n
∑z j =1
ij
is the sample mean of the ith cluster. Here zij is a zero–one indicator variable that is one or zero, according as yj belongs or does not belong to the ith cluster (i = 1, …, k; j = 1, …, n). It is impossible to consider all partitions of the n observations into g clusters unless n is very small as the number of such partitions with nonempty clusters is the Stirling number of the second kind: n 1 n ( −1)( n −1) n k ∑ i n ! j =1
308
Loewe and Nelson
which can be approximated by kn/k!. In practice, k-means is therefore implemented by iteratively moving points between clusters so as to minimize tr W. In its simplest form, each observation yi is assigned to the cluster with the nearest center (sample mean) and then the center of the cluster is updated before moving on to the next observation. Often the centers are estimated initially by selecting k points at random from the sample to be clustered. In the statistical and pattern recognition literature in recent times, attention has been focused on model-based clustering via mixtures of normal densities. 2.4.4. Fuzzy c-Means Clustering
In fuzzy clustering (FCM) each point has a “degree” of belonging to clusters, as in fuzzy logic, rather than belonging completely to just one cluster (41–43). Thus, points on the edge of a cluster may be in the cluster to a lesser degree than points in the center of cluster. With fuzzy c-means, the centroid of a cluster is the mean of all points, weighted by their degree of belonging to the cluster. The degree of “belonging” is related to the inverse of the distance to the cluster center; then the coefficients are normalized and fuzzyfied with a real parameter m > 1 so that their sum is 1. For m equal to 2, this is equivalent to normalizing the coefficient linearly to make their sum 1. When m is close to 1, then the cluster center closest to the point is given much more weight than the others, and the algorithm is similar to k-means. The fuzzy c-means algorithm is very similar to the k-means algorithm where one can: ●
●
●
Choose a number of clusters. Assign randomly to each point coefficients for being in the clusters. Repeat until the algorithm has converged (that is, the coefficients’ change between two iterations is no more than e, the given sensitivity threshold): –– Compute the centroid for each cluster. –– For each point, compute its coefficients of being in the clusters.
The algorithm minimizes intracluster variance as well, but has the same problems as k-means, the minimum is a local minimum, and the results depend on the initial choice of weights. According to Kim et al. (43), FCM clustering can be represented as follows: n
c
minimize J fcm (W , V ) = ∑∑ (wik ) xi − vk , i =1 k =1
m
2
Microarray Bioinformatics
309
where Jm(W,V) represents the objective function defining the quality of the result obtained for prototypes V and membership W, and m is the degree of fuzzification in the clustering. The membership degrees, wik, are defined such that: c 0 ≤ wik ≤ 1, under the constraint of ∑ k =1 wik = 1, for i = 1, …, n. 2
V = (vk) is the cluster center or prototype and xi − vk is the Euclidean distance between gene i and the prototype of cluster k. FCM can be used especially when clusters are alike and sample subgroups differ only to a lesser degree. Cluster analysis is generally visualized using a graphic plot. The compound regulation (up- and downregulation) in each group is displayed in pseudo-colors (green as downregulation and red as upregulation). This display is often referred to as a heatmap and is frequently used in literature for visualization. 2.4.5. Principal Component Analysis
PCA can be used for dimensionality reduction in a data set by retaining those characteristics of the data set that contribute most to its variance, and by keeping lower-order principal components and ignoring higher-order ones (33, 44–46). Such low-order components often contain the “most important” aspects of the data. However, depending on the application, this may not always be the case. PCA is mathematically defined as an orthogonal linear transformation that transforms the data to a new coordinate system such that the greatest variance by any projection of the data comes to lie on the first coordinate (called the first principal component), the second greatest variance on the second coordinate, and so on. PCA is theoretically the optimum transform for a given data in least square terms. PCA has the distinction of being the optimal linear transformation for keeping the subspace that has largest variance. Each microarray gives one value for each principle component or in general dimension. Values and microarrays are termed eigenvalue and eigenarray after transformation (44). As PCA is a dimension reduction technique, the following applies to the data set (46). From the original set of variables Xj, PCA constructs a new set of uncorrelated and orthogonal variables Pj . They are linear combinations of mean-centered variables X j = X j − X j and are often called the loadings or the principal components. It is assumed that these loadings correspond with the eigenvectors of the sample covariance matrix S=
n 1 ′ xi − x )(xi − x ) ( ∑ (n − 1) i =1
of the data. For each loading vector Pj the corresponding eigenvalue lj of S tells us how much of the variability of the data is explained by Pj through the relation lj = var (Pj ) . Usually, these
310
Loewe and Nelson
loading vectors are sorted in descending order of the eigenvalues. Hence, the first k principle components explain most of the variability of the data. After selecting k, we can project the p-dimensional data points onto the subspace spanned by k loading vectors and compute their coordinates with respect to Pj . This yields the scores ti = P (xi − x )
for each i = 1, …, n which have trivially zero mean, as all points taken together cancel each other out in their variation. With respect to the original coordinate system, the projected data point is computed as the fitted value . xˆ i = x + Pt i
PCA can be employed for example to visualize differences of cell types, diseases and tissues. As microarrays are transformed into one point (eigenarray), it is made simple to display vast sample numbers and group them together. 2.4.6. Other Methods
More recently, pathway analysis (e.g., (47)) and gene set enrichment analysis (GSEA) (48) have been applied for the grouping and characterization of array data output. In this analysis, terms with functional meaning that have previously been attributed to the biological compounds accessed by the microarray are grouped by function. For example, the output of DNA arrays can be grouped or filtered by gene ontology (GO) (http://www.geneontology.org) or by association with KEGG pathways (http://www.genome.jp/ kegg/) (49). GSEA pioneered a variety of methods to search for groups of functionally related compounds with a coordinate overor underexpression across a list of genes ranked by differential expression coming from microarray experiments. For a current status of available methods for this analysis see ref. 50.
2.5. Example
As an example of microarray output data, results of a low-density microarray were generated. The array design included 30 biological compounds tested for regulation, and one calibrator, defined as equal in all samples. Two groups (Healthy vs. Ill) with each three individual samples (biological replicates) were calculated. Values are shown in Table 1. With the use of the Calibrator, the microarrays were normalized (described above). Briefly, the mean was calculated for all Calibrator values. Mean(Calibrator) =
∑ all _ calibrators
Number _ of _ samples
.
311
Microarray Bioinformatics
Table 1 Raw data for example microarray Raw data
Healthy 1
Healthy 2
Healthy 3
Compound 1
6,490
5,480
5,435
196
288
207
Compound 2
7,970
6,630
6,794
201
294
212
Compound 3
7,173
6,011
6,062
72
106
77
Compound 4
10,247
8,397
8,884
705
1,032
743
Compound 5
12,524
10,165
10,974
1,053
1,541
1,110
Compound 6
7,742
6,453
6,585
73
108
78
Compound 7
9,222
7,602
7,943
104
153
110
Compound 8
7,628
6,364
6,480
180
264
190
Compound 9
7,742
6,453
6,585
97
143
103
Compound 10
10,133
8,309
8,780
7,752
10,666
9,892
Compound 11
7,970
6,630
6,794
6,116
8,487
7,804
Compound 12
7,742
6,453
6,585
5,944
8,258
7,584
Compound 13
10,247
8,397
8,884
7,839
10,781
10,002
Compound 14
9,108
7,513
7,839
6,977
9,634
8,903
Compound 15
7,628
6,364
6,480
5,857
8,143
7,474
Compound 16
9,905
8,132
8,571
7,580
10,437
9,672
Compound 17
10,589
8,663
9,198
8,097
11,125
10,332
Compound 18
9,564
7,867
8,257
7,322
10,093
9,343
Compound 19
8,995
7,425
7,734
6,891
9,519
8,793
Compound 20
11,955
9,723
10,452
9,131
12,501
11,651
Compound 21
342
292
282
517
757
546
Compound 22
455
389
376
11,026
16,149
11,631
Compound 23
512
438
423
1,938
2,839
2,048
Compound 24
285
243
235
11,413
16,716
12,035
Compound 25
319
272
263
7,718
11,304
8,133
Compound 26
342
292
282
7,752
11,355
8,173
Compound 27
512
438
423
9,691
14,193
10,215
Compound 28
239
204
198
11,396
16,691
12,018
Compound 29
342
292
282
6,719
9,841
7,086
Compound 30
911
778
753
6,202
9,084
6,540
10,970
14,130
11,950
14,500
10,890
12,500
Calibrator
Ill 1
Ill 2
Ill 3
312
Loewe and Nelson
The factor that let each calibrator result in equaling the mean was applied to all compound values of the corresponding group. e.g., Calibrator (Healthy 1) × X = mean(Calibrator), where X is the normalizing factor. The arising arbitrary numbers are given in Table 2. These data represent the starting point for further analysis as well as visualization of the results. The biological compound expression values are visualized in a heatmap (darker grey upregulation, lighter grey downregulation (in general color coded green); both against the mean of all expression values) as depicted in Fig. 1. For this illustration and for all subsequent output files Genesis by the TU Graz (Austria) was used (http://www.genome. tugraz.at/). In this example, it is apparent that Compounds 1–9 are more represented in samples Healthy 1–Healthy 3, whereas Compounds 22–30 are more frequently represented in Ill 1–Ill 3. Compounds 10–20 appear to be similar regarding the measured values. Compound 21 is increased in the “Ill”-group, but the fold change (twofold) is not high enough to induce a color switch. To find significantly differences in the generated data, hierarchical clustering, k-means clustering and one-way ANOVA analysis was employed. As previously described, the length of the branches of the dendrogram (resulting from hierarchical clustering) shows the distance of similarity. In essence, shorter branches show more similar samples or compounds, whereas longer branches symbolize more differences. The resulting dendrogram and hierarchical expression cluster is provided in Fig. 2. Hierarchical clustering shows us that there are three distinct groups present in the analyzed compounds and two in the samples. Distance to the next branch in the dendrogram gives an idea of the degree of similarity. This type of clustering is unsupervised and divides the data set in vast number of groups. If however the group number is predictable, k-means clustering can be used with k as the number of clusters as in Fig. 3. By having three groups of compounds → k = 3 k-means clustering will give the groups in different clusters. Group 1 (left, upregulated in “Healthy”), Group 2 (middle, upregulated in “Ill”), and Group 3 (right, no difference in regulation) are depictured separately. Although illustration gives us an idea of sample and compound regulation, significance testing is necessary to see the true difference. For this example, one-way ANOVA was used due to low complexity of the data. The arising p values and significant statistics are provided in Table 3. One-way ANOVA shows the significance in a p value. In some cases, the program also answers the significance question directly. For Compounds 1–9 and 20–30, a significant difference between the groups was found.
Microarray Bioinformatics
313
Table 2 Normalized data of example microarray Normalized
Healthy 1
Healthy 2
Healthy 3
Compound 1
5,700
6,200
5,200
228
251
207
25
Compound 2
7,000
7,500
6,500
233
256
212
30
Compound 3
6,300
6,800
5,800
84
92
77
75
Compound 4
9,000
9,500
8,500
818
900
744
11
Compound 5
11,000
11,500
10,500
1,222
1,344
1,111
9
Compound 6
6,800
7,300
6,300
85
94
78
79
Compound 7
8,100
8,600
7,600
121
133
110
67
Compound 8
6,700
7,200
6,200
209
230
190
32
Compound 9
6,800
7,300
6,300
113
125
103
60
Compound 10
8,900
9,400
8,400
9,000
9,300
9,900
1
Compound 11
7,000
7,500
6,500
7,100
7,400
7,810
1
Compound 12
6,800
7,300
6,300
6,900
7,200
7,590
1
Compound 13
9,000
9,500
8,500
9,100
9,400
10,010
1
Compound 14
8,000
8,500
7,500
8,100
8,400
8,910
1
Compound 15
6,700
7,200
6,200
6,800
7,100
7,480
1
Compound 16
8,700
9,200
8,200
8,800
9,100
9,680
1
Compound 17
9,300
9,800
8,800
9,400
9,700
10,340
1
Compound 18
8,400
8,900
7,900
8,500
8,800
9,350
1
Compound 19
7,900
8,400
7,400
8,000
8,300
8,800
1
Compound 20
10,500
11,000
10,000
10,600
10,900
11,660
1
Compound 21
300
330
270
600
660
546
−2
Compound 22
400
440
360
12,800
14,080
11,640
−32
Compound 23
450
495
405
2,250
2,475
2,050
−5
Compound 24
250
275
225
13,250
14,575
12,045
−53
Compound 25
280
308
252
8,960
9,856
8,140
−32
Compound 26
300
330
270
9,000
9,900
8,180
−30
Compound 27
450
495
405
11,250
12,375
10,223
−25
Compound 28
210
231
189
13,230
14,553
12,028
−63
Compound 29
300
330
270
7,800
8,580
7,092
−26
Compound 30
800
880
720
7,200
7,920
6,545
−9
12,490
12,490
12,490
12,490
12,490
12,490
1
Calibrator
Ill 1
Ill 2
Ill 3
FC
314
Loewe and Nelson
Fig. 1. Expression visualization of example microarray.
Multiple examples of bioinformatics are detailed in the literature. A number of web pages are available that focus on providing researchers with the possibly to download expression data (e.g., GEO) (http://www.ncbi.nlm.nih.gov/geo/), programs for bioinformatics (e.g., EBI) (http://www.ebi.ac.uk/) or examples for used algorithms. 2.6. Program Packages
Many software packages are readily available for the researcher who wishes to conduct microarray experimental interpretation. The software generally employs the various bioinformatics algorithms that have been outlined in this chapter. RMA-based characterization can be applied using freely available RMAexpress software (http:// www.rmaexpress.bmbolstad.com). SAM-based analysis software is also available online (http://www.stat.stanford.edu/~tibs/SAM/).
Microarray Bioinformatics
Fig. 2. Hierarchical clustering of example microarray.
Fig. 3. k-Means clustering of an example microarray.
315
316
Loewe and Nelson
Table 3 Significance calculation for one-way ANOVA Unique ID
p Value
Significant
Compound 1
1.90E−03
True
Compound 2
1.67E−04
True
Compound 3
5.87E−04
True
Compound 4
1.01E−02
True
Compound 5
4.45E−01
True
Compound 6
1.74E−03
True
Compound 7
5.26E−06
True
Compound 8
4.82E−04
True
Compound 9
3.41E−05
True
Compound 10
0.020681819
False
Compound 11
0.1011849
False
Compound 12
0.12724395
False
Compound 13
0.01946915
False
Compound 14
0.038977187
False
Compound 15
0.14337784
False
Compound 16
0.023464832
False
Compound 17
0.016402658
False
Compound 18
0.028777085
False
Compound 19
0.04231185
False
Compound 20
0.009301341
True
Compound 21
6.86E−02
True
Compound 22
9.30E−02
True
Compound 23
1.46E+00
True
Compound 24
5.53E−02
True
Compound 25
6.33E−02
True
Compound 26
6.77E−02
True
Compound 27
1.12E+00
True
Compound 28
4.65E−01
True
Compound 29
6.70E−02
True
Compound 30
3.33E+00
True
Microarray Bioinformatics
317
Other software tools such as Genesis enable the user to conduct multiple algorithms (http://www.genome.tugraz.at/), allowing not only clustering (e.g., hierarchical and k-means) and significance analysis (e.g., one-way ANOVA), but also PCA and other algorithms. Some programs are applets and add-ons for R (http:// www.r-project.org/), a programming language with focus on statistical analysis. They include BioConductor (51) – a bioinformatics open software with multiple possible application (52) – and RefPlus – a RMA expanding package (30). GenMAPP (http:// www.genmapp.org/) is an example for pathway analysis. Addi tional available tools with similar purpose are reviewed in ref. 35. JMP genomics by SAS (http://www.jmp.com/software/genomics/) is a widely available commercial software, which employs all basic algorithms and is easy-to-use especially for first time bioinformatics users. Graphic displaying and significance testing can be conducted for multiple types of microarrays. ChipInspector available from Genomatix provides a complete platform for DNA analysis. ChipInspector extracts information from the expression level of single probes from microarrays as opposed to probesets. The input files for ChipInspector supports Affymetrix CEL (cell intensity) files (other platforms are possible). ChipInspector assigns probes directly to transcripts and genes and is able to account for alternative transcripts. The software makes use of up to date genomic knowledge and databases of alternative transcripts and promoters to achieve superior signal– noise ratios in microarray analysis. Importantly, the results from this analysis are directly usable as input to the BiblioSphere Pathway software for pathway and association data. This single probe-based analysis of oligonucleotide-based arrays has demonstrated clear advantages over the common probe set-based approach, most notably in improved sensitivity and specificity of the biological findings (DAVID analysis). This was highlighted by the unique detection of a number of categories directly linked to clinical observations (12). While this analysis technique was found to increase the number of significant features, it also simultaneously reduced false-positive rates by an order of magnitude (12). ChipInspector can be summarized in the following steps: 1. Single probe-transcript annotation (resulting reduced list of probes that have a specificity of 100%) 2. Total intensity normalization (modified for usage on single probe level) 3. SAM (adapted to single-probe application) 4. Output generation of significantly expressed transcripts Multiple other software solutions are available. Some are summarized by Werner (50).
318
Loewe and Nelson
3. Notes 1. Quality of the input material is giving rise to differences of microarray level (53). Quality control is essential to achieve comparability between samples. 2. If differences between groups are more drastic than mere up- or downregulation, a discrepancy between present or absent should be used to analyze data. This is done by binarization, whereas 0 expression of the compound stands for absent and 1 stands for present. This approach is more likely to succeed if the sample groups are immensely different (e.g., comparison of brain and spleen tissue samples). 3. Conservative appointing of k can lead to incoherent results if the possibility of subgroups is given. If diseases are analyzed this is often the case (e.g., subgroups of cancer).
Acknowledgments This work was supported by the Deutsche Forschungsgemeinschaft SFB 571 C2, FP6 EU grant INNOCHEM to PJN and BMBF BioChance to PJN. References 1. Angres B. Cell microarrays. Expert Rev Mol Diagn 2005;5(5):769–79. 2. Chiosis G, Brodsky JL. Small molecule microarrays: from proteins to mammalian cells – are we there yet? Trends Biotechnol 2005; 23(6):271–4. 3. Costa JL, Meijer G, Ylstra B, Caldas C. Array comparative genomic hybridization copy number profiling: a new tool for translational research in solid malignancies. Semin Radiat Oncol 2008;18(2):98–104. 4. Liang PH, Wu CY, Greenberg WA, Wong CH. Glycan arrays: biological and medical applications. Curr Opin Chem Biol 2008; 12(1):86–92. 5. Liu XS. Getting started in tiling microarray analysis. PLoS Comput Biol 2007;3(10): 1842–4. 6. Lopez MF, Pluskal MG. Protein micro- and macroarrays: digitizing the proteome. J Chromatogr B Analyt Technol Biomed Life Sci 2003;787(1):19–27. 7. Stadtherr K, Wolf H, Lindner P. An aptamerbased protein biochip. Anal Chem 2005; 77(11):3437–43.
8. Voduc D, Kenney C, Nielsen TO. Tissue microarrays in clinical oncology. Semin Radiat Oncol 2008;18(2):89–97. 9. Wu P, Castner DG, Grainger DW. Diagnostic devices as biomaterials: a review of nucleic acid and protein microarray surface performance issues. J Biomater Sci Polym Ed 2008;19(6):725–53. 10. Simon R. Microarray-based expression profiling and informatics. Curr Opin Biotechnol 2008;19(1):26–9. 11. Knudsen S. Image analysis. In: Guide to analysis of DNA microarray data; 2004. 12. Cohen CD, Lindenmeyer MT, Eichinger F, et al. Improved elucidation of biological processes linked to diabetic nephropathy by single probe-based microarray data analysis. PLoS One 2008;3(8):e2937. 13. Henger A, Schmid H, Kretzler M. Gene expression analysis of human renal biopsies: recent developments towards molecular diagnosis of kidney disease. Curr Opin Nephrol Hypertens 2004;13(3):313–8. 14. Scherer A, Krause A, Walker JR, et al. Opti mized protocol for linear RNA amplification
Microarray Bioinformatics
15.
16.
17. 18. 19.
20. 21.
22.
23.
24.
25.
26.
27.
and application to gene expression profiling of human renal biopsies. Biotechniques 2003;34(3):546–50, 52–4, 56. Brazma A, Hingamp P, Quackenbush J, et al. Minimum information about a microarray experiment (MIAME)-toward standards for microarray data. Nat Genet 2001;29(4): 365–71. Rogers S, Cambrosio A. Making a new technology work: the standardization and regulation of microarrays. Yale J Biol Med 2007;80(4):165–78. Edgar R, Barrett T. NCBI GEO standards and services for microarray data. Nat Biotechnol 2006;24(12):1471–2. Gardiner-Garden M, Littlejohn TG. A comparison of microarray databases. Brief Bioinform 2001;2(2):143–58. Shi L, Reid LH, Jones WD, et al. The MicroArray Quality Control (MAQC) project shows inter- and intraplatform reproducibility of gene expression measurements. Nat Biotechnol 2006;24(9):1151–61. Durinck S. Pre-processing of microarray data and analysis of differential expression. Methods Mol Biol 2008;452:89–110. Dudoit S, Yang YH, Callow MJ, Speed TP. Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments. Stat Sin 2002;12: 111–39. Ritchie ME, Silver J, Oshlack A, et al. A comparison of background correction methods for two-colour microarrays. Bioinformatics 2007;23(20):2700–7. Yang YH, Dudoit S, Luu P, et al. Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 2002;30(4):e15. de Longueville F, Atienzar FA, Marcq L, et al. Use of a low-density microarray for studying gene expression patterns induced by hepatotoxicants on primary cultures of rat hepatocytes. Toxicol Sci 2003;75(2):378–92. de Longueville F, Surry D, Meneses-Lorente G, et al. Gene expression profiling of drug metabolism and toxicology markers using a low-density DNA microarray. Biochem Pharmacol 2002;64(1):137–49. Calza S, Valentini D, Pawitan Y. Normalization of oligonucleotide arrays based on the leastvariant set of genes. BMC Bioinformatics 2008;9:140. Li C, Wong WH. Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc Natl Acad Sci U S A 2001;98(1):31–6.
319
28. Hochreiter S, Clevert DA, Obermayer K. A new summarization method for Affymetrix probe level data. Bioinformatics 2006;22(8):943–9. 29. Irizarry RA, Hobbs B, Collin F, et al. Exploration, normalization, and summaries of high density oligonucleotide array probe level data. Biostatistics 2003;4(2):249–64. 30. Harbron C, Chang KM, South MC. RefPlus: an R package extending the RMA algorithm. Bioinformatics 2007;23(18):2493–4. 31. Holder D, Raubertas RF, Pikounis VB, Svetnik V, Soper K. Statistical analysis of high density oligonucleotide arrays: A SAFER approach. In: ASA annual meeting. Atlanta, GA; 2001. 32. Kerr MK, Martin M, Churchill GA. Analysis of variance for gene expression microarray data. J Comput Biol 2000;7(6):819–37. 33. de Haan JR, Wehrens R, Bauerschmidt S, Piek E, van Schaik RC, Buydens LM. Interpretation of ANOVA models for microarray data using PCA. Bioinformatics 2007;23(2):184–90. 34. Tusher VG, Tibshirani R, Chu G. Significance analysis of microarrays applied to the ionizing radiation response. Proc Natl Acad Sci U S A 2001;98(9):5116–21. 35. Kadota K, Nakai Y, Shimizu K. A weighted average difference method for detecting differentially expressed genes from microarray data. Algorithms Mol Biol 2008;3:8. 36. Zhao H, Chan KL, Cheng LM, Yan H. Multivariate hierarchical Bayesian model for differential gene expression analysis in microarray experiments. BMC Bioinformatics 2008;9(Suppl 1):S9. 37. Schreiber F. Visualization. Methods Mol Biol 2008;453:441–50. 38. MacQueen J. Some methods for classification and analysis of multivariate observations. In: Fifth Berkeley symposium on mathematical statistics and probability. University of California Press, Berkeley, CA; 1967. p. 281–97. 39. Goldstein DR, Ghosh D, Conlon EM. Statistical issues in the clustering of gene expression data. Stat Sin 2002;12:219–40. 40. Eisen MB, Spellman PT, Brown PO, Botstein D. Cluster analysis and display of genomewide expression patterns. Proc Natl Acad Sci U S A 1998;95(25):14863–8. 41. McLachlan GJ, Bean RW, Ng SK. Clustering. Methods Mol Biol 2008;453:423–39. 42. Chen G, Jaradat SA, Banerjee N, Tanaka TS, Ko MSH, Zhang MQ. Evaluation and comparison of clustering algorithms in analyzing ES cell gene expression data. Stat Sin 2002;12:241–62.
320
Loewe and Nelson
43. Kim SY, Lee JW, Bae JS. Effect of data normalization on fuzzy clustering of DNA microarray data. BMC Bioinformatics 2006;7:134. 44. Alter O, Brown PO, Botstein D. Singular value decomposition for genome-wide expression data processing and modeling. Proc Natl Acad Sci U S A 2000;97(18): 10101–6. 45. Daffertshofer A, Lamoth CJ, Meijer OG, Beek PJ. PCA in studying coordination and variability: a tutorial. Clin Biomech 2004;19(4):415–28. 46. Hubert M, Engelen S. Robust PCA and classification in biosciences. Bioinformatics 2004;20(11):1728–36. 47. D’Souza M, Zhu X, Frisina RD. Novel approach to select genes from RMA normalized microarray data using functional hearing tests in aging mice. J Neurosci Methods 2008;171(2):279–87.
48. Jiang Z, Gentleman R. Extensions to gene set enrichment. Bioinformatics 2007;23(3): 306–13. 49. Dopazo J, Al-Shahrour F. Expression and microarrays. Methods Mol Biol 2008;453: 245–55. 50. Werner T. Bioinformatics applications for pathway analysis of microarray data. Curr Opin Biotechnol 2008;19(1):50–4. 51. Gentleman RC, Carey VJ, Bates DM, et al. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol 2004;5(10):R80. 52. Okoniewski MJ, Miller CJ. Comprehensive analysis of affymetrix exon arrays using BioConductor. PLoS Comput Biol 2008;4(2):e6. 53. Popova T, Mennerich D, Weith A, Quast K. Effect of RNA quality on transcript intensity levels in microarray analysis of human postmortem brain tissues. BMC Genomics 2008;9:91.
Index A aa-dUTP-labeled cDNA target.................................. 11, 19 a-Amperometric approach................................................ 43 Analysis of target labeling reaction............18, 21, 23–24, 28 Applications of microarray technology gene expression..................................................... 56–57 genomic analysis and genotyping......................... 57–59 target sequence for microbial arrays antibiotic resistance genes............................... 59–60 bacterial virulence factors................................ 59–60 ribosomal DNA polymorphisms.......................... 59 Aptamer array fluorescein-labeled RNA/DNA aptamers.................. 37 Photoaptamers............................................................ 37 Aptamer immobilization amine-silane-based..................................................... 39 biotinylated aptamer on streptavidin-coated glass slides....................................................... 44 covalent conjugation of NH-aptamer to amino-silanized surfaces....................... 44–45 covalent conjugation of SH-aptamer to thiol-or maleimide-silanized surfaces.......................... 45 Aptamer modification incorporation of commercial amine-reactive compounds to amine-modified aptamer..................................................... 45–46 incorporation of COOH-bearing electroactive labels to amine-modified aptamer............ 46–47 Aptamers conjugation................................................................. 39 labeling....................................................................... 39 novel aptamers............................................................ 38 precipitation................................................................ 39 production and manipulation............................... 38–39 reconstitute lyophilized aptamers............................... 38 RNA aptamers.................................................39, 47, 52 Aptasensor.................................................................. 49, 51 Arabidopsis thaliana genes................................................... 4 Array printing....................................................8, 63, 70, 97 Array regeneration and reutilization................................. 49
B BAC-based CGH DNA microarray................................... 6 BAC clones......................................................................... 8
BAC DNA......................................................................... 8 Bacteria (BAC) genomic libraries....................................... 8 Bacteria microarray, fabrication...............148–150, 153–155 Bioarrays................................................................. 261, 275 Biological microarray.............................................. 249–259 Biomaterials............................. 148, 161–163, 196, 229, 262 Biosensor................................................................133–145 BLAST............................................................61, 71, 85, 86 BRCA1............................................................................... 6
C Cancer.............................................. 5, 6, 121, 222, 252, 318 Capillary force lithography......................147–159, 262, 271 Capillary pump.................177, 178, 181–184, 189, 190, 191 Carbohydrate..............64, 117, 118, 121–124, 128, 130, 251 cDNA..............................................4, 6, 7, 9, 11, 12, 18–21, 23, 29, 31, 56, 57, 62, 64–66, 97, 297, 299 synthesis............................................................. 21, 296 Cell adhesion..................................................162, 170, 178, 180, 188, 202, 205, 251 Cell-alginate solution..................................................... 234 Cell-based biosensor................................133–145, 195, 262 Cell encapsulation.......................................................... 134 Cell free expression......................................97, 98, 100–102 Cell free protein synthesis.......................................... 96, 97 Cell microarray........................................133–145, 207–216 Cell proliferation............................. 162, 170, 231, 232, 236 Cellular micropatterning.................................195, 196, 198 Cell viability assays..................................142, 170, 233–234 CGH DNA microarray.................................................. 5–7 Chemical vapor deposition (CVD).................209, 261–276 Clustal....................................................................... 62, 71, 86 Clustering.........................................................69, 128, 129, 304–310, 312, 315, 317 Collagen............109, 110, 112, 113, 115, 162, 197, 198, 200, 202, 203, 205, 206, 281–283, 285, 286, 289–291 Combinatorial methods.................................................. 161 Comparative genomic hybridization (CGH)........... 5–9, 15 Contact printing................................................64, 204, 275 Copy number variation (CNV)................................ 4, 5, 30 CS chip........................................................................... 183 Cy dye coupling reaction.................................................. 22 Cy-labeled QC oligonucleotides...................................... 88 Cy3-or Cy5-dUTP............................................................ 9
Ali Khademhosseini et al. (eds.), Biological Microarrays: Methods and Protocols, Methods in Molecular Biology, vol. 671, DOI 10.1007/978-1-59745-551-0, © Springer Science+Business Media, LLC 2011
321
Biological Microarrays: Methods and Protocols 322 Index
D Data analysis.......................... 56–59, 63, 64, 68–70, 92, 130 Deletion.................................................................4, 5, 8, 59 Denhardt’s solution...............................................13, 14, 75 Detection strategy electrochemical aptamer microarrays.......................... 51 fluorescence aptamer microarrays............................... 50 surface plasmon resonance imaging (SPRi)................ 50 Differentiation........................................ 108, 117, 161, 162, 197, 220, 222, 223, 232 Dimethylsulfoxide (DMSO)..........................10, 20, 22, 39, 40, 45–47, 143, 236, 252, 258 Direct cell writing.................... 223–227, 229–232, 234–235 Direct fluorescence detection of fluorescent reporters................................... 35 DNA array to protein array (DAPA)........................ 96–104 DNA microarray........................................ 3–31, 40, 55, 56, 63–71, 85–88, 95–105, 215, 266, 295, 296 DNA QC.................................................. 8–9, 14–16, 27–30 Drop immunoassay................................................. 180, 190
E Electrochemical detection.......................................... 42, 43 Electrodes macroelectrode............................................................ 44 Endothelial cells.............................. 245, 282–285, 290, 291 Endothelial tubes.............................................281, 283–286 Enhancer....................................................................... 4, 76 Escherichia coli (E.coli)....................... 60, 72, 76, 97, 99, 103, 148–149, 151–152, 154, 155
F Fluorescence anisotropy.........................................37, 42, 50 Fluorocur mold................................ 251, 253, 254, 257, 258 Focal leaks....................................... 281, 282, 285, 288–291 Food safety....................................................................... 56 Fragmentation buffer........................................................ 11 FRET................................................................................ 35 Functionalized surface............................................ 207–216
G Gene delivery................................................................. 251 Gene set...................................................................... 5, 310 Gene signature............................................................... 5, 6 Glycan.............................................................. 118, 123, 125 Glycomics........................................................118, 125, 128 Glycosylation.......................................................... 118, 128 Gold coated slides...................................................... 38, 43
H Hereditary etiology............................................................. 4 Hot spots..........................................................................161 Human monogenic disease................................................. 4
Human/mouse Cot-1 DNA............................................. 13 Hybridization........................5, 6, 13–14, 16, 23–30, 55, 56, 58, 62–65, 67–68, 70, 71, 73–78, 83–85, 90–91, 118, 119, 120, 127–128, 180, 296–300, 302 Hydrogel microstructures....................................... 133–145 Hydrogel precursor solution............................138, 141, 145 Hydroxylamine................................................11, 22, 40, 47
I Immobilization............................. 36, 38–41, 43–45, 49, 51, 52, 55, 74, 158, 207, 215, 261, 263, 265, 266, 268, 270, 273 Immunofluorescence staining..................104, 110, 113–114 Insertion...........................................................................4, 59 Intercellular reactions............................................. 142–143
L Labeled cRNA purification with RNeasy MinElute column..................................... 22, 23 LAlign.............................................................. 61, 71, 85, 86 Layer-by-layer assembly.................. 148, 150, 151, 155, 159 Lectin................................................................ 117–131, 266 Lectin microarray manufacture.............................................................. 119 print procedure................................................. 120–125 Lift-off process........................................208, 209, 211–212 Liver....................................................................... 219–237 Loading pads........................... 177, 178, 181–184, 188, 190
M Magnetic bead based strand separation................ 73, 82–83 MDA. See Multiple displacement amplification Micellae.................................................... 118, 125–126, 130 Microarray.........................................3–31, 40, 50, 51, 55–92, 95–105, 107–115, 117–131, 133–145, 147–159, 180, 195–216, 249–259, 266, 272, 295–318 bioinformatics........................................60–63, 295–318 oligoprobe binding...................................................... 64 probe design....................................................61, 63, 88 Microbial pathogens................................................... 55–92 Microchannels................................. 135–137, 140–144, 177, 178, 180–184, 188, 190–193, 221, 225, 226, 262, 267, 268, 271, 273 Microfabrication..................................... 134, 184, 198, 209, 211–213, 220, 222, 225, 226, 271, 275 Microfluidic devices........................ 134, 135, 136, 140–142, 145, 220, 221, 225, 226, 232, 234–235, 267, 268 Microfluidics.....................17, 108, 134–136, 138, 140–142, 145, 177–193, 220, 221, 224–227, 231, 232, 234–235, 262, 267, 268, 272, 287 Micromolding in capillaries.....................208–210, 213–215 Micromosaic immunoassays....................178–181, 183, 191 Micropatterns.......... 109, 113, 134, 136, 148, 151, 195–203, 205, 208, 212, 239, 263, 265, 268–271, 273, 275
Biological Microarrays: Methods and Protocols 323 Index
Microprinting......................................................... 219–237 Microscale.................................147, 159, 220, 221, 223, 224, 226, 227, 229, 266, 267, 271 Microspotting...................................... 74, 90, 208–211, 216 Microstructure........................ 133–145, 151, 168–170, 183, 188, 190, 201, 202, 210, 214, 216, 261–276 Microvascular tissue engineering...................................... 81 Microvascular tubes................................................ 281–292 Miniaturizated immunoassays................................ 177–193 M13 macrophages........................... 149, 151, 155, 157, 158 Molding...................135, 137–140, 144, 148, 150, 156–159, 181, 184–186, 239, 240, 242, 243, 250, 262 mRNA translation.............................................................. 4 Multiple displacement amplification (MDA)..................... 9
N Nanoarray fabrication..............................250, 251, 253, 257 NanoDrop Spectrophotometer Analysis of Target Labeling Reaction.................. 18, 21, 23–24, 28 Nanoparticles...................................... 51, 67, 250, 253–259 Nonspecific binding..................................67, 148, 208, 215 Nucleotide microarray............................................ 6, 55–92
O Oligo Design.............................................56, 60–62, 71, 85 Oligo-dT18........................................................9, 10, 17, 19 Oligonucleotides................6, 38, 55–92, 250, 251, 258–259 Organotypic culture................................................ 107, 221
P Particle generation.................................................. 249–251 Pathogen...................................... 55–92, 117, 133, 222, 251 Pathological state................................................................ 4 Patterning........................................ 108, 109, 111, 134, 140, 148–150, 154–155, 177–193, 196, 198, 202, 205, 207–215, 222, 242, 244, 262, 265–268, 271–273 PBPK. See Physiologically-based pharmacokinetic model PCL:PDLLA. See Poly(e-caprolactone):poly(D,L-lactic acid) PCR. See Polymerase chain reaction PCR amplification asymmetric..........................................65, 66, 77, 79–80 multiplex....................................................65, 71, 77–79 multiplex-assymetric....................................... 77–78, 80 standard................................................................ 77, 78 universal primers........................ 8, 62, 65, 71, 78, 80–81 PDMS. See Polydimethylsiloxane PDMS-CSs. See Polydimethylsiloxane-capillary systems Perfluoropolyethers (PFPE)........................................... 250 Perfused tubes................................................................. 288 Permeability..................................... 159, 226, 250, 281–292 Permeability assay....................................284, 285, 287–291
PFPE. See Perfluoropolyethers Phosphate buffers phosphate elution buffers................................10, 20, 31 phosphate wash buffers....................................10, 19, 31 Photolithography PEG hydrogel lithography................197, 198, 201, 202 photoresist lithography......................196–198, 200, 203 Photoreactive CVD polymers.........................271–273, 275 Photoresist.............................. 135, 137, 184, 195–203, 208, 212, 213, 241–243 Physiologically-based pharmacokinetic model (PBPK).............................................. 221 Polydimethylsiloxane (PDMS)...............109–112, 114–115, 135–137, 140, 141, 144, 150, 154, 156–159, 177–193, 210, 214–216, 225–227, 233–235, 239–247, 250, 263, 266–270, 273, 275 Polydimethylsiloxane-capillary systems (PDMS-CSs)................................178, 180–190 Polyelectrolyte (PEL) multilayers...................148, 150, 151, 155–159 Poly(ethylene glycol) hydrogel........................135–137, 144, 196, 198, 201, 202, 204, 205, 271 Polymerase chain reaction (PCR)................... 4, 7, 8, 21, 56, 58, 60–62, 64–66, 70–74, 77–85, 87, 88, 90, 96–102, 104–105 primer design........................... 60–62, 65, 78, 85, 87, 88 Polymer scaffold..................................................... 161–172 Polymorphism.............................................................. 4, 59 Poly(e-caprolactone):poly(D,L-lactic acid) (PCL:PDLLA)............................................. 161 Poly-(TMSMA-r-PEGMA).......... 149, 153–155, 158, 159 Precipitation labeled-DNA.............................................................. 29 labeled targets............................................................. 29 microarray slide..................................................... 27, 29 Predictive factor.................................................................. 5 Prehybridization solution......................................14, 26, 29 Primer extension (PE) reaction.................65, 72–74, 83–84 Primers Cy5-T7-for....................................................98, 99, 102 GENE-for............................................................ 97, 98 GENE-rev.....................................................97, 98, 101 LTT-for.............................................................. 98, 101 LTT-rev.................................................98, 99, 101, 102 NH2-LTT-rev................................................98, 99, 102 T7 domain.................................................................. 99 T7-for.............................................97, 98, 99, 101, 102 T7-rev............................................................. 97, 98, 101 PRINT process........................................250, 251, 253–255 Prognostic factor............................................................. 5, 6 Promoter............................ 4, 21, 52, 66, 73, 78, 83, 99, 100, 182, 185, 317 Protein array................ 95–97, 103–104, 202, 254–255, 299 Protein immobilization............................................. 96, 208 Protein microarray......95–105, 196, 199, 201–203, 208, 216
Biological Microarrays: Methods and Protocols 324 Index
Protein micropatterning..................197–198, 200–202, 208 Protein patterning...........................................209, 211–213 Proteins........................ 3, 4, 8, 16, 35, 37, 41, 47, 52, 53, 64, 71, 77, 95–105, 108–111, 113, 117–119, 126–128, 134, 137, 144, 150–151, 157, 158, 162, 177–193, 195–216, 240–242, 244–247, 249–251, 254–258, 261, 262, 266, 270, 272, 273, 287, 295, 297, 299
R Reactive CVD polymers......................................... 273–275 Reagentless detection................................................. 37, 42 Regulatory regions of genes................................................ 4 RNA annealing................................................................. 21 RNA hydrolysis................................................................ 19 RNA QC.......................................................7, 9–14, 16–27 RNAse H............................................................................ 9 RNAseOne................................................................... 9, 18 RNAsin..................................................................................9 RNeasy MinElute column.......................................... 22, 23
S Salt leaching....................................................165, 167–171 Sandwich assay formats.........................................36, 41, 50 Scaffold arrays........................................................ 161–172 Scaffold porosity..................................................... 168–170 Scanning...................... 13, 16, 24, 27, 28, 30, 31, 56, 57, 58, 59, 68–70, 92, 119, 120, 127–128, 154, 155, 158, 201, 262, 296, 297 Sequence alignment multiple sequence alignment................................ 61, 85 pairwise local alignment............................................. 85 Silanization........................................................43, 114, 200 Single-stranded DNA (ssDNA)................. 8, 23, 24, 64, 65, 71–74, 77, 78, 81–85, 90, 91, 158, 251 Single-stranded RNA (ssRNA)........... 72–74, 78, 81–84, 90 ssDNA. See Single-stranded DNA
ssRNA. See Single-stranded RNA Suboptimal fluorophore-labeled target............................. 41 Superscript II.....................................................9, 10, 18, 19 Supervised analysis methods.................................... 69, 305 Surface engineering.........................................197, 261, 275 Surface plasmon resonance (SPR) imaging..........37, 38, 42, 50–51
T Target amplification........................................56, 70, 78–84 Target labeling, with an amine-reactive fluorophore.. 47–48 Target purification...................................................... 20–21 Three-dimensional tissue microarrays.................... 107–115 Tissue engineering........................... 134, 195, 220, 222, 249 Tissue printing...........64, 220–224, 226, 229, 231, 232, 235 Toxicity............................................................. 134, 221, 250 Two-color cDNA microarray............................................. 6 Two-dimensional tissue microarrays................109, 111–112
U Universal buffer................................................................ 37 Unsupervised analysis methods........................................ 69
V VAMPIR. See Vapor assisted micropatterning in replica structures Vapor assisted micropatterning in replica structures (VAMPIR)........................................... 268–272 Virulence factors........................................59–60, 71, 79, 80 Virus microarray, fabrication...................148–151, 153–158 Voltametric approach.................................................. 43, 51
W WGA. See Whole genome amplification Whole genome amplification (WGA)....................................8–9, 27, 124, 126