See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/300275371

Protein Microarrays: Overview, Applications and Challenges

Chapter · July 2014
DOI: 10.1007/978-94-017-9202-8_8
13 April 2016. The user has requested enhancement of the downloaded file.
READS
CITATIONS 3,224
8

9 authors, including:
Paula Díez
L. Lourido Leiden University Medical Centre 54 PUBLICATIONS 739
Conselleria de Sanidade
CITATIONS
90 PUBLICATIONS 663 CITATIONS
SEE PROFILE
SEE PROFILE

María González
Noelia Dasilva-Freire University of Salamanca
University of Salamanca
57 PUBLICATIONS 817 CITATIONS
16 PUBLICATIONS 347 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Paula Díez on

Translational Bioinformatics 6
Series Editor: Xiangdong Wang, MD, PhD, Prof
György Marko-Varga Editor

Genomic
s and
Proteomic
s for
Clinical
Discovery
and
Developm
ent
Translational Bioinformatics

Series editor

Xiangdong Wang, MD, Ph.D.

Professor of Clinical Bioinformatics, Lund University, Sweden
Professor of Medicine, Fudan University, China
György Marko-Varga
Editor

Genomics and Proteomics

for Clinical Discovery
and Development
Editor
György Marko-Varga
Clinical Protein Science and Imaging
Biomedical Center
Biomedical Engineering
Lund University
Lund , Sweden
Center of Excellence in Biological
and Medical Mass Spectrometry
Biomedical Center D13
Lund University
Lund , Sweden
First Department of Surgery
Tokyo Medical University
Tokyo , Japan
ISSN 2213-2775 ISSN 2213-2783 (electronic)
ISBN 978-94-017-9201-1 ISBN 978-94-017-9202-8 (eBook) DOI
10.1007/978-94-017-9202-8
Springer Dordrecht Heidelberg New York London

Library of Congress Control Number: 2014946410

© Springer Science+Business Media Dordrecht 2014

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifi cally the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfi lms or in any other physical way, and transmission or
information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed. Exempted from this legal reservation are brief
excerpts in connection with reviews or scholarly analysis or material supplied specifi cally for the
purpose of being entered and executed on a computer system, for exclusive use by the purchaser of the
work. Duplication of this publication or parts thereof is permitted only under the provisions of the
Copyright Law of the Publisher’s location, in its current version, and permission for use must always be
obtained from Springer. Permissions for use may be obtained through RightsLink at the Copyright
Clearance Center. Violations are liable to prosecution under the respective Copyright Law.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specifi c statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
While the advice and information in this book are believed to be true and accurate at the date of
publication, 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

Springer is part of Springer Science+Business Media (www.springer.com)

Chapter 8
Protein Microarrays: Overview,
Applications and Challenges

Lucia Lourido , Paula Diez , Noelia Dasilva , Maria

Gonzalez-Gonzalez , Cristina Ruiz-Romero , Francisco Blanco ,
Alberto Orfao ,
Joshua LaBaer , and Manuel Fuentes

Abstract Microarrays technology represents a new tool to address high-

throughput biological studies. Various types of protein microarrays based on the
application, format and content fi nd use in basic research, drug and biomarker
discovery, as well as for the characterization of proteins. In this chapter, we
discuss advantages and limitations of protein microarrays, their features and recent
applications. We also consider the different methods to build protein microarrays
and the recent advances in cell free protein expression systems to construct in situ
protein microarrays. Finally, we describe four types of self-assembled protein
microarrays: PISA (protein array to protein Array); DAPA (DNA to Protein
Array); PuCa ( in situ puromycin array) and NAPPA (Nucleic Acids Programmable
Protein Arrays) and the recent applications of this latter in situ protein array.

Keywords Functional proteomics • Protein microarray • Format • Content •

Protein array applications • Cell-free protein synthesis • in situ protein microarrays

L. Lourido • C. Ruiz-Romero • F. Blanco

Instituto de Investigación BioMedica A Coruña ,
As Xubias de Arriba 84 , 15006 A Coruña , Spain
P. Diez • N. Dasilva • M. Gonzalez-Gonzalez • A. Orfao
Departamento de Medicina, Servicio General de Citometría,
Centro de Investigación del Cáncer, IBSAL , Universidad de Salamanca-CSIC
, Campus Miguel de Unamuno S/N , 37007 Salamanca , Spain
J. LaBaer , MD, Ph.D. (*)
Personalized Diagnostic Laboratory, Biodesign Institute ,
Arizona State University , Tempe , AZ , USA
e-mail: Joshua.labaer@asu.edu
M. Fuentes , Ph.D. (*)
Departamento de Medicina, Servicio General de Citometría,
Centro de Investigación del Cáncer, IBSAL , Universidad de Salamanca-CSIC
, Campus Miguel de Unamuno S/N , 37007 Salamanca , Spain
Proteomics Unit, Centro de Investigación del Cáncer, IBSAL , Universidad de Salamanca-
CSIC , Campus Miguel de Unamuno s/n , 37007 Salamanca , Spain
e-mail: mfuentes@usal.es

© Springer Science+Business Media Dordrecht 2014 147 G. Marko-Varga (ed.), Genomics and
Proteomics for Clinical Discovery and Development, Translational Bioinformatics 6, DOI
10.1007/978-94-017-9202-8_8
148 L. Lourido et al.

8.1 Introduction

DNA arrays and next generation DNA sequencing technologies have found wide
use in the detection of nucleic acids, revealing information about the transcrip
tional states of biological samples in a massive scale; however, gene expression
provides only general and limited information about the function of the gene prod
ucts. Moreover, nucleic acid measurements provide no information about the regu
lation of protein activity, which is markedly affected by posttranslational modifi
cation (PTM). A gene’s function is directly manifested by the activity of its
translated protein. Therefore, the detailed analysis of protein function provides a
better knowledge related to the biological state of the cells (Gygi et al. 1999 ;
Bertone and Snyder 2005 ).
Despite the cell biology knowledge achieved from decades of molecular
biology and genetics, only a small portion of the human protein complement is
understood at the biochemical level. As a refl ection of the new era of research at
scale, pro teomics – the large scale analysis of proteins – is maturing in bringing
methodology to identify, quantify and characterize the functions of all the proteins
involved in biological processes (Bertone and Snyder 2005 ; Yu et al. 2011 ).
 The complexity of the human proteome requires high-throughput (HT)
approaches to defi ne and study the human proteome profi le. During last decade,
protein microarrays have emerged as a useful tool for the analysis of the proteome
at scale. Currently, protein microarrays have been successfully applied in the study
of biomarkers, post-translational modifi cation of proteins, and various types of
interactions with proteins. Protein microarrays have shed light on the biological
roles of proteins involved in disease (Merbl and Kirschner 2011 ; Hanash 2003 ;
Dasilva et al. 2012 ). In this chapter, we review protein microarray technology,
including the classifi cation of protein arrays, their recent applications and chal
lenges of this technology to address the study of human proteome.

8.2 Protein Microarrays

A classical proteomics approach involves the identifi cation of individual proteins

in a protein mixture (e.g., cell or tissue lysate) with some characterization of the
quan tity of each protein species. Separating the sample into fractions to simplify
its com plexity often precedes this type of analysis. Protein separation may be
accomplished with 2D gel electrophoresis (2D-GE) or multi-dimensional liquid
chromatography, and the subsequent identifi cation of protein is done by mass
spectrometry (MS) (Bertone and Snyder 2005 ; Gonzalez-Gonzalez et al. 2012 ).
However, although recent advances have improved the sensitivity and the
reproducibility of these tech niques, they are not readily implemented in a HT
format (Bertone and Snyder 2005 ; Gonzalez-Gonzalez et al. 2012). Moreover, the
majority of the available methods to
149
8 Protein Microarrays: Overview, Applications and Challenges

Table 8.1 Advantages and limitations of protein microarrays

Advantages Limitations
Protein arrays allow monitoring several studies
proteins in the same assay (HT technology) Protein arrays require validation experiments
because of false positives can be detected
Wide range of applications: serum screening, The highest array reported until date included
biomarker discovery and functional proteomic only 9000 different proteins
Easy control of experimental conditions Whole eukaryotic protein arrays still have not been
reported
Low sample consumption Diffi culty to control post-transcriptional modifi cations
Fast Arrayed proteins may not be functional on the surface
Very sensible comparing with other HT technologies Lack of standard protocols

study proteins require denaturing the sample and thus functional characterization
is not possible (Bertone and Snyder 2005 ; Gonzalez-Gonzalez et al. 2012 ). In
contrast with other proteomic strategies, protein microarrays avoid pre
fractionation of the sample. Thus, complex and non-fractionated proteome mix
tures, such as serum, plasma, urine and tissue extracts, can be directly used for
experimentation (Hanash 2003 ; Hanash et al. 2008 ) (Table 8.1 ). For this reason,
among others, protein microarrays offer a powerful technology for functional
proteomics analysis in HT format.
Microarray technologies, like DNA arrays, utilize densely-printed micro-spots
of capture ligands immobilized onto a solid support that are exposed to samples
containing corresponding binding molecules (often referred to as queries), allowing
the simultaneous analysis of thousands of capture targets within the same assay
(LaBaer and Ramachandran 2005 ). Roger Ekins and co-coworkers described these
binding events based on miniaturization as the key parameter. They predict that a
system that uses small amounts of capture molecules and a small amount of sample
can be more sensitive than a system using a hundred times more material. This is
true if K < 0.1 where K is the affi nity constant between ligand and target. The cap
ture ligand is presented in a confi ned area of the array, reducing its diffusion. The
binding event with its specifi c target takes place with the highest possible capture
molecule concentration and therefore, the highest signal intensities and optimal
signal-to noise ratios can be achieved in these small spots (Ekins et al. 1990 ; Ekins
and Chu 1992 ). An immunoassay in an array format displays sensitivities in the
pM to fM range, enabling testing low-abundant (pg/mL) analytes in crude
proteomes with a small volume of sample. In many cases, the sample to test is
minimal so protein microarrays show a relevant advantage in clinical applications.
Thus, protein array technology addresses the necessity of having a multiplex
and highly sensitive protein assay capable of handling and resolving complex
proteomes with limited available sample (Borrebaeck and Wingren 2009 ;
Matarraz et al. 2011 ).
150 L. Lourido et al.

8.3 Array Format

Multiplex protein arrays can be prepared in two major formats in which the
miniaturized assay is performed: planar arrays (such as on glass slides) and bead
arrays, on which the proteins are attached to addressable beads. Here, we briefl y
have described some features of both formats.

8.3.1 Planar Array

Two dimensional planar multiplexed assays consist of high-density microspots of

ligand (protein/peptide/aptamer/tissue) (<250 μm diameter) immobilized onto a
solid support and separated by the minimal distance of 300 μm, which allow a den
sity of >1,000 spots/cm 2 (Bertone and Snyder 2005 ; Matarraz et al. 2011 ;
Ellington et al. 2010 ).

8.3.1.1 Array Chemistries

In planar microarrays, there are some parameters that infl uence the robustness of
the assay performance: spot size and morphology, total ligand binding capacity,
back ground signal, limit of detection and spot reproducibility (Dasilva et al. 2012 ;
Ellington et al. 2010 ).
The greatest challenges in protein immobilization technology are the retention
of natural protein folding, functionality and capacity. As binding proteins to a
surface can alter these parameters, the fi rst key step for success in a planar protein
array is to defi ne the optimal surface and protein immobilization strategy
(Gonzalez
Gonzalez et al. 2012 ; Rusmini et al. 2007 ). Many materials and surface
chemistries have been used for building planar arrays, ranging from PVDF
membranes to glass or gold slides. Glass slides treated with organosilanes are very
commonly used and they are considered suitable substrates for protein
immobilization.
The most often used strategy for protein immobilization is the use of a covalent
attachment, using a wide variety of chemically activated surfaces (e.g. amine, alde
hyde, etc) (Table 8.2 ). The abundance of lysines present on the exterior of the
proteins makes the amine chemistry one of the most popular strategies to immobi
lize proteins.
On one side, N-hydroxysuccinimidine ester (NHS) is the most common agent
to establish strong bonds with protein amine groups and its use was demonstrated
by Patel et al. Aldehyde-glass slides have been also demonstrated by MacBeath
and Schreiber to be a feasible strategy to immobilize large proteins like bovine
serum albumin (BSA). On the other side, bioaffi nity immobilization by complex
avidine
biotin also offers another option to achieve a correct protein attachment. This tech
nique permits an oriented immobilization as well as it is a reversible method,
which allows repeating the use of the same surface.
151
8 Protein Microarrays: Overview, Applications and Challenges

Table 8.2 Available

functionalities in planar arrays according to Side groups Surfaces -NH 2 Carboxylic acid
different functional groups at protein surface Active ester

Amino acids
-Lys, hydroxi-Lys

Epoxy
Aldehyde
Cys -SH Maleimide Pyridyl disulfi de
Vinyl sulfone
-COOH Amine -OH Epoxy
Asp,Glu

Ser,Thr

In some cases, the immobilization of proteins is directed by physical

interactions on hydrophobic (nitrocellulose, polystyrene) or positively charged
(polylysine, aminosilane) surfaces. This is a weak bond which occurs by
adsorption. As a result, proteins can be randomly oriented onto the heterogeneous
surface and this fact may lead to the loss of the functional protein sites, which are
inaccessible (Rusmini et al. 2007 ).

8.3.1.2 Array Printing

Another important variable to build planar arrays is the selection of the printing
method. Briefl y, the printing methods can include contact and non-contact
printing. The size, morphology and reproducibility of the spots on the surface will
depend on the deposition method selected.
In the contact printing, the tiny pins transfer nanoliter volumes of the printing
mix on the surface, with the fi nal transfer volume depending on the size of the pin
and the length of time that pins contact the surface. Both solid and quill type pins
are available for printers, however with the viscous nature of printing mixes that
contain proteins, the quill type pins often do not deliver reproducible volumes.
Alternatively, non-contact deposition technologies utilize capillaries or inkjet tech
nology to deposit picoliter droplets onto the surfaces. In theory, this method
decreases spot-to spot variability and the variability between batches (Ellington et
al. 2010 ; Glokler and Angenendt 2003 ).

8.3.1.3 Assay Execution

There are several factors, which must to be considered before executing protein
microarrays arrays to evaluate the printing reproducibility and the consequent
assay. On one hand, spotting buffer composition can infl uence the protein
stability, the protein binding capacity to the surface and therefore, the quality of
the spots produced. For this, there are many different buffers with different pH,
which can be
152 L. Lourido et al.

used for arraying (carbonate, PBS, citrate, acetate buffer…) samples on the surface.
The choice will depend on the nature of the printed analyte (Kusnezow et al. 2003
). On the other hand the morphology of the spots will depend on sample viscosity
and printing humidity. Sometimes, protein microarray production runs take a long
time, depending on the amount of features on the microarray and batch size. In
this sense, sample evaporation could lead to a gradient of concentration during the
print run or in the worst case to blackout of print head nozzles due to salt out
effects. Moreover, a higher viscosity reduces the time of sample drying. To achieve
a highly reproducible microarray quality it is of prime interest to reduce this evap
oration to the minimum (Gutmann et al. 2005 ). For this aim, the humidity along
the printing must be checked according to the features of the analyte to print. In
some cases, some hygroscopic additives like DMSO or glycerol can be added to
prevent the sample evaporation and improve the stability of the samples (McQuain
et al. 2003 ).
Finally, after arraying, the use of effi cient blocking of reactive surface groups
is critical for a reduced the unspecifi c binding to the surface (background). In this
sense, classical blocking buffers as bovine serum albumin (BSA) at different con
centrations or milk powder are very used in protein arrays to reduce the
background (Kusnezow et al. 2003 ).

8.3.1.4 Assay Detection

The detection of interactions depends on the kind of assay performed, and may
employ fl uorescence, chemoluminiscence, radioisotope labelling or label-free
methods, such as surface plasmon resonance or imaging atomic force microscopy
(Bertone and Snyder 2005 ; Gonzalez-Gonzalez et al. 2012 ).
Fluorescence compounds are often the most useful reporter applied to detect
the protein-protein interactions. The resulting signal confers high sensitivity and
wide dynamic range (approximately 5 logs). Suitable fl uorescence readout
systems, such as high-density microarray scanners, can be used to detect when a fl
uorescent query is stably bound to a particular feature on the array. Signal quantifi
cation is then analyzed by specifi c software. Currently, the automation of these
methods has increased the throughput of the planar arrays.

8.3.2 Beads Arrays

In this section, the characteristics and applications of beads arrays will be

discussed as well as the differences between beads and planar arrays.
The diameter of beads typically varies from 0.02 to 0.6 μm. The beads with a
diameter >0.1 μm are called microspheres and those <0.1 μm are referred to as
nanoparticles (Casado-Vela et al. 2013 ). This size of the beads may impact the
num ber of analytes immobilized on the surface.
8 Protein Microarrays: Overview, Applications and Challenges
153

A
A

B
Flow
+ sample Detector
antibodies
citometry
B
+

Fig. 8.1 Schematic view of suspension arrays resulting of the combination of two different fl
uorescent dyes ( A , B ). Assays are analyzed by coding attributes, and fl ow cytometry is used to
detect assay specifi c fl uorescence signal

Target proteins are immobilized on the bead surfaces, often using chemistries
similar to planar arrays. In order to test multiple proteins simultaneously, the beads
used for each unique protein must be addressable by some kind of bar-coding.
Most often, this is accomplished with the use of coloured coding, using different
colours and multiple intensities of each colour. Beads are fi lled of one or more fl
uorescent dyes and the surface is functionalized and coated with a capture
molecule to bind in an effi cient way the specifi c analytes in a biological sample
(Casado-Vela et al. 2013 ; Kellar et al. 2001 , 2006 ) (Fig. 8.1 ).
The immunodetection is accomplished by fl ow cytometry in which one or
more lasers excites the internal dye(s) of the bead and a detector captures the
colour pro fi le, thus reading the “bar-code” on the bead and identifying the
corresponding tar get protein. A second laser and detector excite and read the fl
uorescent dye linked to the query molecule. These captures or interaction events
are assigned according to fl ow cytometry principles: assay-specifi c beads are
distinguished by either light scatter or internal fl uorescent ratio, and
analyte-dependent signal is generated by the fl uorescence generated by the
capture event (Ellington et al. 2010 ).
This technique is based on fl uorescence cell-sorting which has been used for
more than 20 years in the clinical fi eld (Casado-Vela et al. 2013 ). However, multi
plex fl uorescence bead assays were fi rst reported in 2001 to identify and quantify
cytokines in serum samples. In these naïve approaches relatively few simultaneous
events were described.
Currently, uniquely bar-coded fl uorescence microspheres available for
Luminex Corporation allow more than 500 analytes per assay. Although the
number of detected proteins in planar arrays is signifi cantly higher than beads
arrays (thou sands vs. hundreds), the fl exibility of the bead-based array format in
some settings adds an important dimension to HT analysis.
In relation to sensitivity, the lower detection limit reported for bead fl
uorescence arrays is 1.2 pg/mL with a dynamic range of up to 55-fold reported by
Won using fl uorescence beads. This is 10-fold change more than it is reported
from LC mass spectrometry analysis with a complex protein sample (Casado-Vela
et al. 2013 ; Ellington et al. 2010 ). A advantage regarding to planar array is a
better feasibility and accuracy of the detection since, thanks to fl ow cytometry,
multiple independent measurements may be achieve within each microsphere
population.
154 L. Lourido et al.

Therefore, beads protein microarrays combine the simplicity of immunoassay

with multiplexing capability and sensitivity of protein microarrays. Almost all of
the work reported thus far in the fi eld use bead -based arrays to identify and
quantify particular analytes in serum, blood or other biological fl uids (Schwenk et
al. 2008 ). These assays, like planar assays, require a minimal quantity of the
biological sample to carry out the analysis. In such studies, the color-coded beads
arrays are coated with antibodies with specifi cities against particular targets. In
this assay, the biologi
cal sample is labelled with fl uorescent dye by randomly chemically linking to all
proteins in the sample with amino coupling chemistry. Any dye labelled protein
binding to beads coated with a specifi c antibody will be read out as a signal
attached to beads with the corresponding color code. For these assays, the
accuracy and the reproducibility of the results rely on the specifi city and the
quality of the antibodies employed and the labeling effi ciency of the proteins in
the sample. A signifi cant concern is that many antibodies exhibit cross-reactivity,
which will give false read
outs (Poetz et al. 2005 ; Schwenk et al. 2007 ). For this reason, many efforts must
be made to demand for application-specifi c antibody validation (Stoevesandt and
Taussig 2007 ).
Gevaryas et al. perform a quality control analysis between planar and bead cap
ture protein array measuring the immune response profi les in blood samples from
two large clinical studies on prostate cancer.
First, they studied the reproducibility of these protein arrays with the same sam
ples. Then, they compared the results obtained for the same targets and samples
using both platform arrays.
On one hand, their results show that both approaches quantify changes ranking
4–5-fold in the composition of the samples. On the other hand, the planar array and
beads arrays that they used had good reproducibility (R 2 = 0.77 and 0.75 respec
tively) but they did not agree in around 50 % of the 57 selected proteins for this
study. Therefore, they concluded that these assays highlight the need to check care
fully the quality control of these assays to obtain reproducibility (Ghevaria et al.
2012 ).
In a recent work, Teilacker et al . described a combination platform between
pla nar and microspheres arrays where four antigens were interrogated as a model
sys tem for multiplexed protein detection.
They used populations of fl uorescent encoded microbeads conjugated with bio
tinylated capture antibodies and then immobilized in a fl ow cell. First, the
biotinylated capture antibody was conjugated to the encoded micro bead and
covalently coupled to fl ow cell. The fl uidic device was consisted in car boxilate
glass coverslip, a thin double-coated adhesive silicone gasket with 12 cutouts for
the fl ow channels, and an aluminium plate with ports for tubing connec tion to a
syringe pump.
Then four antigen solutions were spiked in serum or BSA and introduced into
each channel. Finally, fl uorescence imaging of the encoded microbeads was per
formed on an epifl uorescence microscope.
They showed that the sensitivity of this method was comparable to the
sensitivity obtained by enzyme-linked immunosorbent assay (pg/mL) using only 5
μL of the
8 Protein Microarrays: Overview, Applications and Challenges
155

sample for each fl ow channel. With this approach, they also reduced the average
area of the spots and, therefore, a subsequent reduction of sample volume and
reagents. This work leads to a great achievement towards the miniaturization
because they managed to print 250 different analytes in the same area where
typically, in planar arrays, a unique spot is printed (Theilacker 2011 ).

8.4 Content

The content of protein microarrays could have a wide diversity, from antibodies
and cellular lysates to recombinant proteins. In fact, some authors classifi ed
protein arrays according to the content: Assembled arrays or self-assembled
arrays. Here, it will be briefl y reviewed only a few aspects because mostly of the
differences with other classifi cations are based on nomenclature instead of
methodological aspects.

8.4.1 Assembled Arrays

In these kinds of arrays, the target proteins in the array are typically antibodies,
puri fi ed proteins or lysates, which are immobilized onto a functionalized surface.

8.4.1.1 Antibody Arrays

Antibody arrays are generated by printing analytes specifi c reagents (ASR) onto
the array surface (either planar or beads; Fig. 8.2 ). Thus, these arrays specifi cally
target those analytes for which there are antibodies printed on the arrays (LaBaer
and Ramachandran 2005 ; Matarraz et al. 2011 ). These arrays are normally used
to

ab c Direct labeling Sandwich labelling Reverse phase array

 Fig. 8.2 Schematic view of different types of planar assembled arrays. ( a ) A targeted and
competitive assay where proteins are directly labelled with a fl uorophore; ( b ) A targeted assay
with capture antibodies bind unlabeled proteins and these are detected by other labelled antibody;
( c ) Reverse phase array where protein mixtures are directly attached on the surface array and
detected by other proteins or labelled antibodies
156 L. Lourido et al.

identify and quantify the presence of multiple different proteins simultaneously.

They are analytical arrays whose principal application is the detection of differen
tially expressed proteins and their abundance in different samples. Analytical
arrays are commonly used to identify biomarkers, which are biometric
measurements, including molecular signatures, that predict a biological or clinical
condition (e.g., healthy/pathologic), often with potential diagnostic or prognostic
value (Borrebaeck and Wingren 2009 ).
As noted above, the accuracy of such arrays depend highly on the specifi city
and affi nity of the antibodies. Thus, monoclonal antibodies are commonly used in
this setting. However, the specifi city of monoclonal antibodies can vary signifi
cantly; many will bind non-targets ( 27 ). Moreover, producing and qualifying
monoclonal antibodies is expensive and slow, and there are many analytes for
which specifi c monoclonal antibodies cannot be found.
Recently, there has been an increased drive towards developing analyte-specifi c
reagents alternatives to monoclonal antibodies as recombinant antibodies.
Recombinant antibodies are produced in vitro, in a method that does not require
the use of animals, potentially providing a less expensive method to obtain
antibodies. Recombinant anti
bodies are cloned genes encoding fragments of antibodies, which maintain the
recogni tion capacity for the antigen. These fragments are often expressed as a
single chain fragment variable (scFv) or antigen binding-fragment (Fab) (Dahan et
al. 2007 ).
Phage display is a widely used technology to screen recombinant antibody
librar ies for specifi c molecules that recognize a desired antigen. This molecular
technique takes advantage of the replication system of phages to produce different
variants of the same protein or peptide of interest. The nucleotide sequence
encoding the scFv or Fab is inserted into the phage genome as a fusion to a gene
encoding a phage coat protein. This fusion ensures the display of recombinant
antibody proteins at the surface of the mature phage. Then the antibody-displaying
phage are exposed to the desired target antigen, which is immobilized on a surface
(Dahan et al. 2007 ). This allows fractionation of the phage bearing antibodies that
bind the antigen from those that do not. An advantage of this approach is that the
genes encoding the successful binders can be recovered from the phage and used
to produce recombinant antibod
ies is a wide variety of protein expression systems. However, the selection process
for fi nding high affi nity binders may require several selections cycles to achieve
the needed enrichment and sometimes requires the introduction of random
mutations, followed by selection to “mature” the binders to achieve acceptable
specifi city.
Carlsson et al. produced capture protein microarrays using recombinant scFv
antibodies, and demonstrated their use for detecting changes in the levels of several
interleukins and complement proteins among 40 samples from metastatic breast
cancer serum and healthy donors. This same group also used recombinant antibody
arrays to classify serums from systemic lupus erythematosus (SLE) and systemic
sclerosis (SSc) patients (Carlsson et al. 2011 ).
Due to the challenges associated with producing both recombinant and monoclo
nal antibodies, there has been a renewed interest in polyclonal antibody reagents.
Most investigators agree that polyclonal reagents, produced by the traditional
method of inoculating rabbits with whole protein in adjuvant, produce reagents that
8 Protein Microarrays: Overview, Applications and Challenges
157

have too much cross reactivity for protein microarrays, and, in any case, cannot be
adequately characterized as reagents.
Recently, Larsson et al. have developed a multiplex immunization strategy for
generating something that they refer to as monospecifi c antibodies (msAbs). First,
they select the Protein Epitope Signature Tags (PrEST), which are 100–150 poly
peptide sequences predicted to be unique in the proteome, the Protein. The lack of
homology to other proteins and the shorter size of these PrEST is believed to con
tribute to more specifi city in the resulting reagents (Larsson et al. 2006 ). Then,
these fragments are cloned with N-terminal fusion, expressed in E. Coli and purifi
ed. Multiple purifi ed PrESTs are mixed at equal ratios to create a multiplex
antigen to immunize animals and create the polyclonal antiserum. Then, the
antiserum is pro cessed and purifi ed over individual PrESTs to recover the
PrEST-specifi c antibody. This strategy allows a reduction of the cost and the
number of animals needed for HT antibody production. One potential limitation of
the PrEST approach is that, unlike monoclonal and recombinant antibodies, the fi
nal reagent is not renewable.
To prove their strategy, in 2009, Larsson et al. selected two non-overlapping
PrEST of Cytokeratin-17, as a model system to study the specifi city and cross reac
tivity of fi ve antibodies generated towards each PrEST.
Using planar arrays, all antibodies recognized their respective antigen but one
of them showed signifi cant binding to a third Cytokeratin-17 PrEST included on
the array and overlapping amino acid sequence in both two PrEST investigated.
By beads arrays, they found that these antibodies recognised same parts of the
C-terminus of each PrEST.
The resulting affi nity-purifi ed antibodies were also analyzed Western blotting,
immunohistochemistry and immunofl uorescence.
In these assays, except for differences in staining intensity, a similar result was
obtained for all antibodies in the respective PrEST-group.
These data suggest that for targets where it is diffi cult to fi nd unique sequence
regions it is possible to instead raise family specifi c antibodies recognizing a defi
ned group of proteins rather than a specifi c target.
The production of affi nity reagents by any method results in products falling
into a wide range of affi nities and specifi cities; therefore, it is essential to
characterize each ASR with other techniques (western blot, immunoprecipitation
followed by mass spectrometry, etc.) before use in protein arrays (LaBaer and
Ramachandran 2005 ).

Assembled Array Signal Detection

Brief mention should be made of the methods for detecting the binding of analytes
to captures. As mentioned above, analytes in a sample are commonly labeled
directly with a fl uorescent (or other) marker molecule. If one possesses more than
one affi nity reagent to a target of interest, and the reagents recognize different non
competing epitopes, then an indirect sandwich assay can be used to detect the ana
lyte. Moreover, there is in increasing interest in developing detection methods that
do not rely upon any labels at all.
158
L. Lourido et al.

The direct label is the simplest and most direct approach because nearly all proteins
in a sample can be tested simultaneously; providing they are adequately labeled by
the marker and there are corresponding capture reagents on the array. Moreover,
direct labeling also allows for direct comparisons of multiple samples on the same
array by labeling each sample with a distinguishable marker. For example, one
might compare two time points or biological conditions by labeling the samples
with different color labels. However, the specifi city of detection using direct
labeling relies entirely on the corresponding ASR for each analyte, which may
lead to false signals in some cases.
With indirect labeling, the marker is attached to the second affi nity reagent or
to a secondary reagent that recognizes it (e.g., Alexa388-labeled anti-mouse IgG
rec ognition of an analyte specifi c monoclonal antibody). Consequently, the
indirect assay is more specifi c than direct labelling because it requires two ASRs
to recog nize the specifi c analyte to observe signal. However, this assay has two
main draw backs: fi nding a matched pair of antibodies that work well together can
be very challenging, and there appears to be a practical limit of less than 40
analytes that can be measured simultaneously using antibody pairs. Interference
and cross-reactivity become an increasing problem as the number of antibodies
added to a common detection mix increases. Huang et al. used a sandwich assay to
measure the levels of 24 cytokines in two biological conditions. There is a vast
collection of antibodies that recognize cytokines with well-established target
epitopes, but for most other antigens, it is diffi cult to fi nd compatible antibody
pairs to carry out a sandwich assay (Gonzalez-Gonzalez et al. 2012 ; Huang et al.
2001 ).
Antibody microarrays are well suited as screening tools for discovering disease
specifi c biomarkers owing to their potential to measure thousands of proteins in
rapid, low volume assays. A number of reports have applied protein microarrays to
biomarker discovery in cancer.
Recently, Gao et al. measured the abundance of eighty-four proteins in a two
color assay to compare chronic obstructive pulmonary disease (COPD), newly diag
nosed subjects with lung cancer and healthy controls serums (Gao et al. 2005 )
being each sample compared to a pooled reference sample (consisting of a mixture
of all of the sera).
The values determined were the normalized average of base-2 logarithms of the
intensity arising from the individual sample divided by the intensity arising from
the pooled sample, which was measured as Cy3 and Cy5 fl uorescence,
respectively. With this approach, they found that using an analysis variance model,
7 antibodies showed signifi cant differences between both lung tumor patients vs.
normal con
trols and lung tumor patients vs. COPD patients (Gao et al. 2005 ). Wittekind &
colleagues also reported a study where proteins from 30 normal and hepatocellular
liver were differently labelled with Cy3 and Cy5 and hybridized on a nitrocellulose
protein microarray made up 83 different antibodies. Proteins of each condition (1
mg/mL) were labeled with NHS-ester activated Cy3 or Cy5 and mixture of equal
concentrations to incubate on the array. The ratios between dyes were
determinated for the individual proteins. To determine which proteins were found
to be differentially expressed, a cutoff level was fi xed using a hierarchical model.
8 Protein Microarrays: Overview, Applications and Challenges
159

Using a stringent interval of 0.4–1.9 corresponding to 2.5 SDs, fi ve proteins

were classifi ed as differentially up- regulated: IGFII, ADAM9, STAT3, SOCS3,
and cyclin D1 and four proteins (collagen I, SMAD4, FHIT, and SOCS1) as
differen tially down-regulated. These data were confi rmed using western blot
analysis of selected proteins using identical antibodies.
To test the sensibility of the array, purifi ed proteins were spotted demonstrated
high sensitivity and specifi city of the protein microarray system, with a detection
limit of 6.25 pg of spotted proteins.
To test the specifi city of the microarray data, a fl uorescent dye reversal experi
ment was performed. HCC and normal tissue were labelled with Cy3 and Cy5,
respectively. Similar profi les of up- or down-regulated proteins were obtained, irre
spective of the dye used (Tannapfel et al. 2003 ).
Amonkar et al. described the development and preliminary evaluation of a mul
tianalyte profi le that can classify women suspected of having ovarian cancer, into
those with and without ovarian cancer.
In this work, sera from 176 cases representing all stages (I,II,III,IV) of
epithelial ovarian cancer and 187 controls from women presenting, the most
common benign ovarian conditions. These samples were assayed in a protein array
consisted in a panel of 104 antigens, 44 autoimmune and 56 infectious disease
markers
Analytes were quantifi ed by reference to 8-point calibration curves and
machine performance was verifi ed using three quality control (QC) samples for
each analyte. Almost all the samples were analyzed in two rounds and the QC
samples generally had coeffi cients of variance below 15 %.
Then, they built many analyte classifi ers and selected the best model using
boot strap performance. Finally, using a testing set of 245 samples, they built an
11- analyte classifi er had 91.3 % sensitivity and 88.5 % specifi city.
Sreekumar et al. used antibody arrays made up 146 distinct antibodies to
monitor changes in the levels of proteins in colon carcinoma cells after
quimiotherapy (Sreekumar et al. 2001 ).
In this work, control and stimulated cells were labeled separately using either
Cy5 or Cy3 dyes and incubated on the array. Cy3:Cy5 ratios were determined for
the individual proteins and a cutoff of 1.15, value was used as a criterion to defi ne
proteins considered differentially expressed and a P for each of the differentially
regulated proteins was calculated using an unpaired t test
The validation of protein microarray data was done by fl uorescent dye-reversal
and immunoblot analysis to the selected proteins.
Belov et al. developed an nitrocellulose antibody microarray to immunopheno
type leukaemia and lymphomas according to the abundance of a panel of 60 anti
gens or cluster of differentiation (CD) characterized at the surfaces of lymphocytes
(Belov et al. 2003 ).
They estimated the average number of cells bound per dot (determined micro
scopically), which correlated well with average binding density values from the
image analysis software, and generally correlates with results from fl ow
cytometry.
They concluded that the CD antibody microarray enables rapid, concurrent
screening of leukocyte suspensions for expression of many CD antigens. And that,
in contrast to fl ow cytometry, with this platform, cells captured on the CD
antibody
160 L. Lourido et al.

microarray can be imaged directly with an optical scanner without staining or

labeling, and image analysis software gives immediate results. In addition to
protein based affi nity reagents, there is an increasing interest in the use of nucleic
acid-based affi nity reagents. The most common of these are referred to as
aptamers and they are constructed from single stranded DNA or RNA, often with
modifi ed nucleotides, which fold into conformations that bind specifi cally to
antigens. They are chemically synthesized and in vitro selected by cycles of
binding amplifi cation with a SELEX technique (reviewed in (Ellington et al. 2010
)). In 2010, Gold et al. created a new class of aptamers with modifi ed nucleotides,
the Slow Off-rate Modifi ed Aptamer (SOMAmer). With the incorporation of these
quemically-modifi ed nucleotides (SOMAmers) into SELEX experiments, they
measured 813 proteins with low limits of detection (1 pM median), 7 logs of
overall dynamic range (100 fM–1 mM), and 5 % median coeffi cient of variation.
Briefl y, in SOMAMER technologie, the sample is incubated with a mixture of
SOMAmers each containing a biotin, a photocleavable group, and a fl uorescent
tag at 5′-end followed by capture of all SOMAmer-protein complexes on
streptavidin beads. After stringent washing of the beads to remove unbound
proteins and label ing of bead-associated proteins with biotin under controlled
conditions, the com plexes are released from the beads back into solution by UV
light irradiation and diluted into a high concentration of dextran sulfate, an anionic
competitor. The bio tin that was originally part of the SOMAmer remains on beads.
The anionic com petitor coupled with dilution selectively disrupts non-cognate
complexes and because only the proteins now contain biotin, the complexes are
re-captured on a second set of beads from which unbound SOMAmers are
removed by a second stringent washing. The SOMAmers that remain attached to
beads are eluted under high pH-denaturing conditions and hybridized to
sequence-specifi c complementary probes printed on a standard DNA microarray.
In this work, they demonstrated the specifi city of SOMAmers for the proteins
they were selected against but they need to validate and standardize SOMAmer
based measurements and expand these studies to understand the specifi city of
SOMAmers for close homologues and alternate forms, such as the products of alter
native splicing, post-translational modifi cations, and proteolytic cleavage.
With this method, De Groote et al. identifi ed several proteins that exhibit
signifi - cant expression differences during the intensive phase of tuberculosis
therapy. However, these fi ndings require future testing in properly designed
validation stud ies using independent sample test sets with proper disease controls.

8.4.2 Reverse-Phase Arrays

The concept of ‘Reverse-phase’ protein microArray (RPA) assays is essentially

the inverse of the capture arrays. Instead of testing each sample (e.g., clinical
sample) for many possible analytes, each RPA tests many samples for the
abundance of a specifi c analyte. Essentially, it is a sophisticated micro-scale
version of a dot blot. Complex protein mixtures (such as cellular or tissue lysates,
or biological fl uids
8 Protein Microarrays: Overview, Applications and Challenges
161
such as serum, etc.) are printed to a substrate in a defi ned array pattern and then
probed with antibodies or other affi nity reagents that are highly specifi c for
analytes of interest. The most critical aspect of RPAs is the extensive validation of
the anti bodies to ensure that they do not have any cross-reactivity with other
proteins that may be present in the lysate. RPAs can generate 1,000 times more
data using 10,000 times less sample volume than an ordinary western blot (Fig.
8.2 ). Large-scale sam ple collection is a labor-intensive and time-consuming
process; however, the infor mation yielded from RPA assays, enables researchers to
evaluate theoretical protein pathways experimentally in a HT format
(Gonzalez-Gonzalez et al. 2012 ; Spurrier et al. 2008 ).
One of the limitations for this technique is the dynamic range because the
ability to detect low abundance proteins in complex mixtures is a challenge for this
kind of array (Dasilva et al. 2012 ). Several techniques have been developed for
the detec tion of these low abundance proteins. As in the analytical arrays, fl
uorescence is commonly used for standard signal readout. During analysis, signal
intensity among spots is often normalized with the total protein content per spot
which can be mea sured with different dyes (Sypro Ruby, colloidal gold, etc). The
dye choice to quan tify the protein content will depend on type of sample being
investigated, the sensitivity of the stain, the material of the surface and of the
instrument detectors (Gallagher et al. 2011 ).
Paweletz et al. immobilized protein lysates from microdissected histologically
normal prostate epithelium, prostate intra-epithelium neoplasia (PIN) progression
and invasive neoplasia. Using a Wilcoxon test for the statistical analysis, they
observed a signifi cantly increased (p value < 0.03) in the phosphorylation of AKT
protein which was associated with cancer progression in the epithelium transition
along the disease as well as a decreased phosphorylation of ERK (p.value < 0.01)
for the three comparisons. (Paweletz et al. 2001 ). They also validated these results
by Western Blot assay.
Cid et al. were the fi rst to use RPA to study pathogen-host protein interactions.
In this work, they studied the presence of posttranscriptional modifi cations in
effector proteins, T3SS proteins, from different mutants of Salmonella
typhimurium when they infected in vitro HeLa to understand signaling events that
take place along Salmonella infection and the intracellular survival of this bacteria
into the cells.
Lysate collection representing all infection conditions were printed and using
several validated antibodies against phospo-epitopes, they show a comparative
results among the different assays according to abundance proteins or posttranscrip
tional modifi cation (Molero et al. 2009 ).

8.4.3 Self-Assembled Protein Microarrays

These arrays focus mainly on identifying and characterizing the specifi c function
of proteins, as well as their interactions with other molecules (including proteins,
pep tides, small molecules/drugs, enzyme-substrates or nucleic acids) (LaBaer and
Ramachandran 2005 ). These functional protein arrays also allow the detection and
162 L. Lourido et al.

identifi cation of post-translational modifi cations (PTMs), such as glycosylation,

phosphorylation and acetylation, which typically modulate the protein’s function,
regulation and/or turnover (Casado-Vela et al. 2013 ).
 The fi rst critical step to build protein microarrays is to display proteins on a
solid surface for the detection of their biochemical activities in a multiplex
manner. Notably, the intrinsic properties of proteins, particularly their highly
variable bio chemical properties, make building protein arrays more challenging
than building nucleic acid arrays, which have very consistent chemical properties
(Bertone and Snyder 2005 ; Templin et al. 2002 ; Rusmini et al. 2007 ). Briefl y,
some of protein properties which must be accommodated when building protein
arrays include: (i) Wide variety of chemistries, affi nities and specifi cities; (ii)
different oligomer ization state from monomer to multimers; (iii) different PTMs;
(iv) varied protein stability, which is frequently altered when the protein is
deposited or immobilized onto a surface; (v) protein production and purifi cation
in high-throughput manner with high yield could be also challenging
(Gonzalez-Gonzalez et al. 2012 ; LaBaer and Ramachandran 2005 ).
The cell-based expression system and the purifi cation to generate large
quantities of proteins is usually a very tedious task and do not guarantee the
functional integ rity of the protein. This issue represents a bottleneck in the HT
functional proteomic studies. Nowadays, it is possible achieve these drawbacks
building arrays of full length, functional proteins from a library of clones expressed
in situ .
In the self-assembled protein microarrays, the protein are synthesized from
their corresponding messenger ribonucleic acid (mRNA) or complementary
deoxyribo nucleic acid (DNA) directly on the surface of the array and the
immobilization of the nascent protein is coupled in the same step in a fast manner.
On the research side, self-assembled arrays offer the detection of multiple pro
tein interactions with low reagent consumption in a fast and low cost fashion. On
the translational side, the discovery of these interactions will foster the
development of new pharmaceutical targets, diagnostics and therapeutics. Thus,
this technology is an attractive point of sight for the pharmaceutical industry
(Dasilva et al. 2012 ; LaBaer and Ramachandran 2005 ).

8.4.3.1 In Situ Protein Expression Systems

From the discovery of in situ protein expression systems forty years ago by
Nirenberg and Matthai, they have been broadly utilized in the scientifi c
community to solve the issues of in vivo protein production. A main advantage
that these sys tems have over in vivo protein synthesis is that the environmental
conditions can be adjusted easily. Strategies to improve protein folding and
posttranslational process ing include the addition of a variety of reagents and
folding catalysts to the reaction. But, above all, the main goal for cell-free
translation systems is to synthesize biologically active proteins (Casado-Vela et al.
2013 ).
These in vitro expression systems exploit the ability to translate proteins using
properly prepared lysates from a number of different organisms (both prokaryotic
8 Protein Microarrays: Overview, Applications and Challenges
163

and eukaryotic), which provide the ribosomal machinery, accessory enzymes,

tRNAs, amino acids and an appropriate energy source. Moreover, these lysates can
be coupled with specialized RNA polymerases and the appropriate nucleotides to
allow simultaneous transcription, thereby allowing full transcription and translation
of proteins from exogenously added cDNA templates.
Prokaryotic cell-free expression systems can produce up to mg quantities of
protein (Hunt 2005 ; Kigawa et al. 1999 ; Mijakovic and Macek 2012 ; Plotkin and
Kudla 2011 ).
They are reasonably tolerant to additives (cofactors, protease inhibitors or
energy sources) (Casado-Vela et al. 2013 ). However, some of the same limitations
that bac teria have producing eukaryotic proteins, such as marked decrease in
success for proteins >65 kDa, also plague cell free systems from bacteria
(Casado-Vela et al. 2013 ). Typically, bacterial cell-free systems do not produce
posttranslational modi fi cations (PTMs), which can be either useful or not
depending on the application. It is worth noting the recent introduction of highly
characterized cell free systems from bacteria. These systems are produced entirely
from purifi ed recombinant pro teins and ribosomal RNA. In this manner, they are
highly characterized and some applications might benefi t from using a system
where every component is known and there is no risk of contaminants from a
crude lysate (Ref).
Cell-free eukaryotic expression systems include wheat germ, insect, rabbit
retic ulocyte and human lysates.
Although rabbit reticulocyte produces less protein than other expression
systems (0.2 μg/10 μL reaction) and is very expensive, it is commonly used for
functional proteomic studies because it is very fast (2 h approx. in protein
production), it has a very high success rate for most mammalian proteins,
including large and membrane proteins, and it does support limited PTMs
(Casado-Vela et al. 2013 ). Rabbit reticu
locyte lysate also contains most chaperone proteins, so the there is a high
likelihood that translated proteins will fold naturally and even display activity if
they can act monomerically (Casado-Vela et al. 2013 ).
As a result of the source and method of production, each batch of rabbit
reticulo cyte lysate derives from a single animal and is by necessity a
non-renewable resource. An additional disadvantage of this product is that there is
a high degree of variability from one batch of lysate to the next, both in the yield
of protein produced and sometimes in the presence of other factors that can affect
downstream applica tions. Laboratories might have to test a dozen different batches
to fi nd one with the right activity profi le needed for their experiments.
The recent development of cell free lysates from human cells addresses some of
these concerns (Casado-Vela et al. 2013 ). These lysates are produced from a cul
tured human cell line, which allows for much greater control of growth conditions
and signifi cantly minimizes the batch-to-batch variation. Avoiding the presence of
animal serum also reduces the likelihood of contaminants that can interfere with
some assays. The yields of protein for these lysates are quite favorable and in the
case of human protein production, the use of human ribosomes and chaperone
proteins gives the greatest possible chance for natural protein folding (Casado Vela
et al. 2013 ).
164 Arrays 8.5.1 PISA

8.5 Types of In Situ Protein L. Lourido et al.

PISA (protein in situ array) was introduced by He and Taussing in 2001 and it
was the fi rst well known cell free based in situ protein array. This method uses
PCR- amplifi ed DNA as template. The use of PCR product obviates the need to
clone the open reading frame into a plasmid, but for large-scale array production,
the repetitive use of PCR to produce the template for making arrays can become
both costly and lead to losses in fi delity. The DNA encoding the protein of inter
est contains a T7 promoter or another strong transcriptional promoter and an in
frame N- or C-terminal tag sequence for protein capture onto the surface. The tags
are typically short peptides so that their sequences can be incorporated into the
PCR primers. In order to perform PISA; the wells of a microtiter plate are
pre-coated with a tag- capturing agent. After transcription and translation, the
expressed proteins bind onto the surface through the specifi c tag. In the fi rst
PISA, He and Tau used DNA templates to express human anti-progesterone anti
body and luciferase with 6X histidine tags. These were captured into a microtiter
plate with 24 wells coated with nickel nitrilo-triacetic acid (Ni-NTA) and
Ni-NTA-coated magnetic beads respectively. They checked that small quantities of
these proteins can be expressed and immobilized onto the surface and that they
conserved their functional features. Also, they confi rmed its HT applications, such
as the generation of protein arrays for non-available cloned genes or for proteins
without functionally production in heterologous expression systems. They
suggested the combination with in vitro display methodologies for HT identifi
cation. PISA opened the door to cell free production of protein arrays. It
demonstrated that multiple proteins could be produced without the need to use
cells for expression followed by lysis and purifi cation to make the proteins
(Gonzalez-Gonzalez et al. 2012 ). In 2006 , Angenendt reported a PISA where 384
different proteins could be expressed from very small quantities of template
(Angenendt et al. 2006 ; Casado-Vela et al. 2013 ) (Fig . 8.3 ).

8.5.2 DAPA

This innovative technique also was developed by in He et al. in 2008 . DAPA

(DNA Array to Protein array) is a technique derived from PISA but it allows for
the repeated use of the same DNA template slide for printing up to 20 copies of
the same protein array and DNA could be reused after prolonged periods of time.
DAPA starts by spotting the PCR amplifi ed DNA fragments encoding the tagged
protein on one slide. This slide is sandwiched with another Ni-NTA slide where a
tag- capturing agent immobilizes the expressed protein. A permeable membrane
with the cell-free
8 Protein Microarrays: Overview, Applications and Challenges
165

mRNA
Nascent protein

tag
cDNA Captured protein
A) Protein in situ array

Add cell free protein expression Capture de novo protein NI-NTA

system
array (PISA)

Nascent
mRNA protein
Captured protein
slide

B) DNA to Protein Arrays

(DAPA)
membrane
Capture de novo protein

Add cell free protein Captured protein

expression system
 Nascent
mRNA
Puromycin protein

C) In situ puromycin nucleotide

RNA degradation o
array(PuCa) Olig Puromycincapture

Anti-tag protein mRNA

Add cell free protein
expression system

Capture de novo protein

D) Nucleic acid
programmable protein
array (NAPPA)
Captured protein tag

biotin Add cell free protein

expression system
cDNA avidin Nascent

Fig. 8.3 Schematic view of four methods coupling cell-free protein synthesis to protein binding
on the surface arrays. ( A ) protein in situ arrays(PISA); ( B ) DNA to Protein Array(DAPA); ( C )
puromycin capture protein array (PuCa) which uses mRNA as template; ( D ) nucleic acid
program mable protein arrays (NAPPA). Figure adapted from Casado-Vela et al. 2013

lysate, which allows coupled transcription and translation is placed between the
two slides. Proteins synthesized from immobilized DNA spots diffuse through the
membrane and are bound to the surface of the capture ligand on the other slide.
Even though it is used, the DAPA requires long time to express proteins and this
technique is limited by the possibilities of protein diffusion during membrane
penetration, especially regarding larger multimeric proteins (Gonzalez-Gonzalez et
al. 2012 ; He et al. 2008 ).

8.5.3 PuCA

Puromycin capture protein arrays (PuCA) are cell-free expression protein arrays
based on the affi nity of puromycin by just-in time expressed peptide/protein. Tao
and Zhu developed it in 2006. With this method DNA is transcribed into mRNA in
vitro . The mRNA 3′ end is attached with single stranded oligonucleotides
166 L. Lourido et al.

(ssDNA) which is complementary with another ssDNA bearing biotin and

puromycin and this complex together is layered onto the chip surface. Biotin
serves to immobilize the mRNA onto the coated streptavidin surface and
puromycin serves as an anchor to bind the nascent protein translated when
cell-free expression system is added to the array (Gonzalez-Gonzalez et al. 2012 ;
Tao and Zhu 2006 ).

8.5.4 NAPPA

Although useful in research, these mentioned strategies have only been tested
with relatively small numbers of proteins compared with printing purifi ed pro teins
 and have yet to demonstrate the robust ability to produce the high content needed
to justify protein microarrays as a routine proteomics tool (LaBaer and
Ramachandran 2005 ).
In 2004, LaBaer’s lab developed a high-density self-assembled protein microar
ray called nucleic acid programmable protein array (NAPPA)
(Ramachandran,Science). It is based on cDNA templates cloned into expression
plasmids, typically using the Gateway technology, which add a transcriptional
promoter and also adds an in frame polypeptide capture tag. The requirement to
clone the cDNAs into a special ized vector requires a much greater upfront
investment compared with PCR. However, there are several advantages: (1) once
the clone is produced as a glycerol stock it becomes a indefi nitely renewable
resource that can be shared with other labs; (2) if the clone is carefully sequence
verifi ed, then the resource will have long term sequence fi delity; (3) the use of
plasmids removes some of the length constraints on the epitope tags, so that
functional protein tags can be used. In most applications of NAPPA the proteins
are fused with glutatione-S-transferase (GST); however, other tags such as fl ag,
HA, c-myc, and Halo tag have been used in specifi c applications. High quality
supercoiled plasmid DNA is purifi ed from bacteria cultures and printed onto an
activated ester surface along with a homo-bifunctional crosslinker, bovine serum
albumin (BSA) and anti-GST antibody. BSA effi ciently increased the DNA
binding and reduces the unspecifi c interactions and anti-GST attaches the protein
expressed (Ramachandran, Science). When cell-free expression system is added to
the array, a coupled transcription/translation reaction is produced and the nascent
protein is linked to the capture agent tag the C-terminal end assuring the complete
translation of the protein (Ramachandran et al. 2008a ).
In an updated method for NAPPA, LaBaer and colleagues built an array of
1,000 human genes available through the DNASU repository (http://DNASU.org)
and demonstrated that 96 % of the genes showed detectable protein signal,
including both soluble and membrane proteins (Ramachandran Nature Methods).
They con
cluded that the protein size had only a modest effect with 98 % of proteins <50
kDa showing good display levels, whereas proteins >100 KDa showing success
around 88 %. With this report they concluded that this method enables various
experimental approaches to study protein function in HT (Ramachandran et al.
2008a ).

View publication stats