DNA Microarrays Gene Expression Applications 1st Edition

Visit the link below to download the full version of this book:

https://medipdf.com/product/dna-microarrays-gene-expression-applications-1st-edi
tion/

Click Download Now

B. R. Jordan (Ed.)

DNA Microarrays:
Gene Expression
Applications

With 20 Figures
DR. B.R. JORDAN
Marseille-Genopole
Parc Scientifique de Luminy, Case 901
13288 Marseille, Cedex 9
France

ISBN 978-3-540-41508-4

Library of Congress Cataloging-in-PubIication Data

DNA microarrays: gene expression appIications I Bertrand Jordan, ed. p. cm
Includes bibliographical references and index.
ISBN 978-3-540-41508-4 ISBN 978-3-642-56517-5 (eBook)
DOI 10.1007/978-3-642-56517-5
1. DNA microarrays-Laboratory manuals. 2. Gene expression-Labordtorymanuals
1. Jordan, Bertrand. QP624.S.D726 D634 2001 572.8'65-dc21

This work is subject to copyright. Ali rights reserved, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of
illustrations, recitation, broadcasting, reproduction on microfilm or in any other
way, and storage in data banks. Duplication of this publication or parts thereof is
permitted only under the provisions of the German Copyright Law of September
9, 1965, in its current version, and permission for use must always be obtained from
Springer-Verlag. Violations are liable for prosecution under the German
Copyright Law.

http://www.springer.de

© Springer-Verlag Berlin Heidelberg 2001

Origina11y published by Springer-Verlag Berlin Heidelberg New York in 2001

The use of general descriptive names, registered names, trademarks, etc. in this
publicat ion does not imply, even in the absence of a specific statement, that such
names are exempt from the relevant protective laws and regulations and therefore
free for general use.

Production: PRO BOIT GmbH, 69126 Heidelberg, Germany

Cover design: D&P, 69121 Heidelberg, Germany
Typesetting: AM-productions GmbH, 69168 Wiesloch, Germany
Printed on acid-free paper SPIN 10792625 39/3130Re - 543210
Preface

The aim of this book is to provide in compact form a compre-

hensive and practical survey of expression measurement using
DNA arrays. I have endeavoured to assemble chapters written by
scientists who are actual users of the technology and have had to
cope with the various practical problems involved in setting up
new methods in the laboratory; I believe this is how such a book
can be most useful to its readers.
Chapter 1 provides some background on the history of DNA
array development; Chapters 2, 3 and 4 focus on various types of
micro arrays: glass microarrays are described by experienced
academic users, while the less-known (but in some situations
superior) nylon micro arrays are presented by their developers
together with some information on nylon macroarrays that are
still widely used. Chapter 5 describes in detail the use of oligo-
nucleotide chips in a research laboratory, again with emphasis
on practical aspects. The principles and practice of both data
acquisition and data mining in the field of expression measure-
ment are presented in Chapter 6 by very experienced authors,
together with a wealth ofInternet resources that are particularly
useful in this fast-moving field. A short last chapter (Chapter 7)
attempts to forecast the likely evolution of this field.
All the authors have strived for clarity and insistence on prac-
tical aspects; I can only hope that the result will be satisfactory
for the reader.

Marseille-Genopole, March 2001 B.R.JORDAN

P.S. One vexing point in the terminology of DNA arrays is that the DNA segments
bound to the device are called "target" by some and "probe" by others, while the
reverse applies to the labelled material prepared from the sample. I have not tried
to force a solution to this problem, apart from ensuring consistency within each
chapter.
 VII

Contents

Chapter 1
DNA Arrays for Expression Measurement:
An Historical Perspective
BERTRAND R. JORDAN

1 Introduction.................................... 3
2 The Forerunners: Colony Filters and Dot Blots. . . . . . . 3
3 "High-Density" Filters. . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
4 Colony Filters and Expression Analysis
(Qualitative) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
5 Quantifying Expression with
High-Density Filters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
6 Miniaturisation: cDNA Microarrays . . . . . . . . . . . . . . . . 10
7 Miniaturisation: Oligonucleotide Chips ............ 10
8 Data Acquisition, Storage and Mining .............. 11
9 A (Provisional) Conclusion . . . . . . . . . . . . . . . . . . . . . . . 11
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Chapter 2
Expression Profiling with cDNA Microarrays:
A User's Perspective and Guide
SEAN GRIMMOND and ANDY GREENFIELD

1 Introduction.................................... 15
2 Expression Profiling with cDNA Microarrays:
The Basics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
3 Production of cDNA Microarrays .................. 17
4 Conclusions.................................... 27
5 Protocols....................................... 28
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
VIII Contents

Chapter 3
cDNA Microarrays on Nylon Membranes with
Enzyme Colorimetric Detection
KONAN PECK and YUH -PYNG SHER

1 Introduction .................................... 37
2 Fabrication of Microarrays on
Nylon Membranes ............................... 37
3 Enzyme Colorimetric Detection ................... 40
4 Characteristics of the Microarray/CD System ........ 41
5 Data Processing in the Microarray/CD System ....... 46
6 Protocol 1: Preparation of cDNA Targets for
Microarray Fabrication ........................... 49
7 Protocol 2: cDNA Complex Probe Labeling ......... 50
8 Protocol 3: Array Hybridization and
Color Development .............................. 51
9 Protocol 4: Modified Catalyzed Reporter
Deposition (CARD) Method ...................... 52
References ...................................... 53

Chapter 4
cDNA Macroarrays and Microarrays on
Nylon Membranes with Radioactive Detection
BEATRICE LORIOD, GENEVIEVE VICTORERO and
CATHERINE NGUYEN

1 Introduction.................................... 57
2 Manufacture of Nylon Arrays. . . . . . . . . . . . . . . . . . . . . . 58
3 Labelling and Hybridisation with
Complex Probes from Total RNA .................. 65
4 Signal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
5 Protocol 1: Array Preparation ..................... 73
6 ProtocoI2:RNAPreparation...................... 77
7 Protocol 3: Labelling. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
8 Protocol 4: Hybridisation ......................... 80
9 ProtocolS: Data Processing ....................... 82
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
Contents IX

Chapter 5
Oligonucleotide Chips for Expression Analysis:
Principles and Practical Procedures
PIERRE CASELLAS,ANNICK PELERAUX and SYLVAINE GALIEGUE
1 Introduction.................................... 87
2 Principles...................................... 87
3 Bioinformatics.................................. 96
4 Conclusion..................................... 100
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 101

Chapter 6
Gene Expression Data Mining and Analysis
ALVIS BRAZMA, ALAN ROBINSON and JAAK VILO
1 Introduction.................................... 106
2 From Raw Data to Gene Expression Matrix. . . . . . . . .. 107
3 Gene Expression Matrix Analysis . . . . . . . . . . . . . . . . .. 110
4 Expression Pro filer .............................. 119
5 Applications.................................... 120
6 From Gene Expression Data to
Gene Regulatory Networks ........................ 122
7 Conclusions.................................... 125
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 126

Chapter 7
Future Trends in the Use of DNA Arrays for
Expression Measurement
BERTRAND R. JORDAN
1 Introduction.................................... 133
2 Future Chips: How Complex? . . . . . . . . . . . . . . . . . . ... 133
3 From "Clone-Based" to "Sequence-Based Arrays" .... 134
4 From "Home-Made" Microarrays to
Ready-Made Chips.............................. 136
5 From "Stand Alone" Array to Integrated
"Lab-on-a-Chip" ................................ 136
6 From Fluorescent Labelling to Electrical Detection. .. 137
7 Progress Towards More Sophisticated
Data Interpretation
and (Maybe) a Unified Expression Repository....... 137
8 Expression Measurement Is Here to Stay. . . . . . . . . . .. 138
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 138

Subject Index........................................ 139

CHAPTER 1

DNA Arrays for Expression Measurement:

An Historical Perspective
BERTRAND R. JORDAN!

1 Introduction .................................... 3
2 The Forerunners: Colony Filters and Dot Blots ....... 3
3 "High-Density" Filters ............................ 5
4 Colony Filters and
Expression Analysis (Qualitative) .................. 7
5 Quantifying Expression with
High -Density Filters ............................. 7
6 Miniaturisation: cDNA Microarrays ................ 10
7 Miniaturisation: Oligonucleotide Chips ............ 10
8 Data Acquisition, Storage and Mining .............. 11
9 A (Provisional) Conclusion ....................... 11
References ...................................... 12

1 Marseille-Genopole, Parc Scientifique de Luminy, Case 901, l3288 Marseille,

Cedex 9, France
DNA Arrays for Expression Measurement: An Historical Perspective 3

1
Introduction

Basically, DNA arrays consist of a series of DNA segments regularly arranged on

some kind of support, and the expression measurement involves hybridising
the whole array with a labelled DNA or RNA sample. The essential feature is
parallel processing: in a single experiment, information is obtained on each of
the hundreds or thousands of entities present on the array (e.g. whether or not,
and at which quantitative level, they hybridise with a given nucleic acid
species). It is this parallelism that makes them so important at a time when
many megabases of genome sequence and thousands upon thousands of genes
need to be analysed.

2
The Forerunners: Colony Filters and Dot Blots

This principle of parallel processing was already implemented in the 1970s.

Colony hybridisation (Grunstein and Rogness 1975) was used to search for
specific genes among libraries; dot blots (Kafatos et al. 1979) and slot blots
allowed homology determination or expression analysis on series of samples,
with radioactive labelling in almost all cases (Fig. 1.1). The density of colony
filters could be quite high, up to 10,000 on a Petri dish-sized membrane, but
these colonies were arranged at random on the membrane since they resulted
from plaque lifts or from direct spreading of a transformation reaction onto the
membranes. Thus no permanent record of the colonies existed, and the
experiments were directed at isolating one or several "positive" clones, the rest
being discarded. Dot and slot blots, on the other hand, were done in ordered
format, often with 96-trough devices geared to microtitre plates; the DNA (or
RNA) solutions were passed through the membrane under conditions
conductive to binding, and the resulting microtitre-sized dot blot displaying
up to 96 spots was used for expression or homology analysis (Fig. 1.1).
4 BERTRAND R. JORDAN

Fig. 1.1. a A phage plaque lift on round nitrocellulose membrane (approx. 1,000 plaques) with a few
positive plaques showing up after hybridisation and exposure to X-ray film. Marks on edge of filter are
used for subsequent critical positioning of the autoradiograph to recover "positive" plaque material
from the Petri dish, since plaques are randomly distributed over the surface. Diameter is
approximately 8 cm. b An example of the first implementations of ordered DNA arrays, a dot-blot
performed with a 96-well filtration device and hybridised with a radioactive probe. Example shown
is a screening application with positive/negative samples. Dimensions are ca. 8x 12 cm
DNA Arrays for Expression Measurement: An Historical Perspective 5

3
"High-Density" Filters

A major change in the field was the development in the late 1980s of robotic
devices ("gridding robots") that made it possible to spot bacterial colonies in a
compact and regular pattern. The resulting "high-density filters" contained
approximately 10,000 spots on a 22x22 cm 2 surface, corresponding to a "pitch"
(centre to centre spacing) of approximately 2 mm (Fig. 1.2). Of course, these
colonies had to be ordered previously , i.e. picked from agar plates and
distributed into micro titre plates for storage and filter construction. This was
often done by hand, although "picking robots" were being developed and
reliable commercial models appeared on the market in the early 1990s.
Hans Lehrach's group, at EMBL (Heidelberg, Germany), then at ICRF
(London, UK) and now at the Berlin Max Planck Institute, was the major
proponent of this approach for genome analysis (Hoheisel et al. 1991; Lennon
and Lehrach 1991), ata time when the approach taken in the USA relied almost
exclusively on PCR methods and PCR screening of cleverly arranged pools of
clones (Olson et al. 1989; Green and Olson 1990) - hybridisation being
considered "not robust enough" for systematic genome work. Lehrach and
coworkers introduced the concept of the "reference library" as a defined and
more or less permanent set of clones, each of which has a definite physical
address (e.g. well B12 of plate 312), is present at a known position in a set of
high-density membranes, and is stored with all the information obtained in a
relational database (Nizetic et al. 1991). This made it possible to acquire, store
and correlate data from many different laboratories, as long as they used the
same membranes and sent back their results to the central database, a major
hurdle in practice.
For a while, high-density membranes were mostly used for library access.
Several resource centres accumulated various genomic and cDNA libraries and
distributed them to laboratories in the form of high-density membranes. When
the user had identified a positive spot using a probe for a gene of interest, the
centre then provided the corresponding clone (Fig. 1.2). This approach
worked fairly well, at a time when sequence information was very scarce, and
allowed many users to progress much faster with their projects. As noted before,
information feedback to the resource centre has always proved difficult to
enforce; the data coming from very diverse laboratories also turned out to be
difficult to handle, so that, contrary to early expectations, this approach did not
contribute much to the construction oflarge-scale physical maps.
6 BERTRAND R. JORDAN

Fig. 1.2. An early"high-density filter" used to access genomic or cDNA libraries. In this case, 10,000
bacterial colonies containing DNA segments cloned in cosmids have been spotted and grown on a
22x22 cm nylon membrane. Hybridisation with a probe prepared from the insert of a cDNA clone
reveals one or several positive spots, indicating the corresponding genomic clone; the relevant
cosmid can be ordered from the resource centre that provided the membrane by indicating its
coordinates. The whole grid is visible because of light background hybridisation; successive
hybridisations to search for various genomic clones are performed on the same membrane without
stripping (to avoid loss of material), hence the large number of "positive" spots. This was the first
widely used application of high-density DNA arrays, pioneered by Hans Lehrach's group and
involving only positive/negative scoring, usually by visual inspection
DNA Arrays for Expression Measurement: An Historical Perspective 7

4
Colony Filters and Expression Analysis (Qualitative)

The use of unordered or ordered colony filters containing cDNA clones for
expression analysis began in the early 1980s with a qualitative application,
differential screening. Parallel hybridisation of "identical" membranes
containing clones from conventional or subtracted cDNA libraries with
complex labelled cDNA mixtures prepared from two different samples was
used, alone or in conjunction with other methods, to pinpoint genes whose
expression was different under the two conditions. Using radioactivity and x-
ray film detection, this method was necessarily qualitative; it was also
cumbersome and suffered from a number of technical problems, but was
nevertheless instrumental in isolating several important genes such as the T
cell receptor (Hedrick et al. 1984) or the CTLA (cytotoxic T lymphocyte-
associated transcripts) series of molecules (Brunet at al. 1988).

5
Quantifying Expression with High-Density Filters

The concept of using the newly developed imaging plate systems that were just
beginning to penetrate biological research laboratories (Amemiya and
Miyahara 1988) for quantitative acquisition of these data and more refined
analysis of expression patterns was fairly obvious and discussed as early as
1990, with a first publication reporting actual data in 1992 (Gress et al. 1992),
but the full implementation of the technology took a fair amount of time. This
is not unusual when moving from a qualitative to a quantitative application,
especially as expression measurement with DNA arrays involves the
quantification of very weak signals if data for genes expressed at very low levels
are required. A number of artefacts leading to spurious data (especially with the
unsequenced cDNA libraries of that period) had to be identified and taken care
of (Nguyen et al. 1995), standardisation methods had to be worked out
(Bernard et al. 1996), and the rather primitive and user-unfriendly image
analysis programs of the period had to be improved (Granjeaud et al. 1996) - a
task made easier by the quickly increasing computing power available to
scientists. Automation of PCR and standardisation of plasmid vectors used for
cDNA libraries led to a shift from colony filters to membranes on which
amplified DNA has been deposited, although colony filters are still used in
some cases.
During the first half of the 1990s, a few groups worked out the methods,
published proof-of-principle papers (Nguyen et al. 1995; Zhao et al. 1995;
Pietu et al. 1996) and began accumulating expression data in different systems
(Fig. l.3), leading later to biological results published in the relevant journals.
Some of this work was performed in "gene discovery" mode, i.e. measuring the
8 BERTRAND R. JORDAN

expression level in a number of conditions for a large set of genes (represented

by sequenced or, in some cases, un sequenced cDNAs), in order to pinpoint
genes whose expression patterns suggest their involvement in particular
biological events (see, e.g., Carrier et al. 1999). In this type of approach,
extensive biological knowledge is necessary at the start, to define precisely the
conditions that are likely to yield the most significant information, and later to
choose the "best" genes out of the dozens that display the required expression
profile and to define and perform the additional experiments that will allow
this choice to be done most effectively. In this context, the biological samples
are used to obtain information on the genes and to highlight the most
interesting ones (preferably "new", i.e. not yet described in detail), whose close
study is likely to be the most rewarding with regard to the biological question
approached.
Expression measurement was also performed in an "expression profiling"
mode, in which the set of genes (often more restricted) was chosen a priori and
usually well known, and the objective was to obtain information on the
samples: analysis of the expression profile for a series of tumours, for instance,
in the hope of obtaining prognostic and therapeutic information (see, e.g.,
Bertucci et al. 1999b). In this case the genes are used as tools to derive
information on the samples, exactly the opposite of the previous situation. Of
course, if technical progress, miniaturisation and decreasing costs make it
possible to routinely include all human genes in every experiment, the two
approaches will eventually coalesce.
The high-density filter (under the more trendy name of "macro array") is still
with us: for experiments of moderate scope, it performs quite adequately and
blends well with current laboratory practice and equipment. In fact, a number of
manufacturers market such membranes with sets of cDNA clones (in the form of
PCR products or bacterial colonies) from various organisms, an indication of the
continuing popularity of the method. Even a firm like Incyte, which has heavily
promoted the micro array approach after acquiring a specialist company, Synteni,
now also sells macroarrays.
DNA Arrays for Expression Measurement: An Historical Perspective 9

Fig. 1.3. Typical macroarray on nylon membrane, in this case measuring 8x 12 cm and containing
3,072 bacterial colonies. Bottom hybridisation with a common vector sequence showing all spots
(and allowing measurement of relative amount of DNA in each of them); top hybridisation with
radioactive cDNA made from total mRNA. Hybridisation intensities are acquired and quantified
using an imaging plate system, allowing measurement of relative mRNA abundances