peptide arrays for kinome analysis: new opportunities and remaining challenges

REVIEW

Peptide arrays for kinome analysis: New opportunities

and remaining challenges

Ryan Arsenault1,2, Philip Griebel2,3 and Scott Napper1,2

1 Department of Biochemistry, University of Saskatchewan, Saskatoon, Saskatchewan, Canada2 Vaccine and Infectious Disease Organization, University of Saskatchewan, Saskatoon, Saskatchewan,

Canada3 School of Public Health, University of Saskatchewan, Saskatoon, Saskatchewan, Canada

Received: May 31, 2011

Revised: September 28, 2011

Accepted: October 4, 2011

Phosphorylation is the predominant mechanism of post-translational modification for

regulation of protein function. With central roles in virtually every cellular process, and

strong linkages with many diseases, there is a considerable interest in defining, and ulti-

mately controlling, kinase activities. Investigations of human cellular phosphorylation events,

which includes over 500 different kinases and tens of thousands of phosphorylation targets,

represent a daunting challenge for proteomic researchers and cell biologists alike. As such,

there is a priority to develop tools that enable the evaluation of cellular phosphorylation events

in a high-throughput, and biologically relevant, fashion. Towards this objective, two distinct,

but functionally related, experimental approaches have emerged; phosphoproteome investi-

gations, which focus on the sub-population of proteins which undergo phosphorylation and

kinome analysis, which considers the activities of the kinase enzymes mediating these

phosphorylation events. Within kinome analysis, peptide arrays have demonstrated consid-

erable potential as a cost-effective, high-throughput approach for defining phosphorylation-

mediated signal transduction activity. In particular, a number of recent advances in the

application of peptide arrays for kinome analysis have enabled researchers to tackle

increasingly complex biological problems in a wider range of species. In this review, recent

advances in kinomic analysis utilizing peptides arrays including several of the biological

questions studied by our group, as well as outstanding challenges still facing this technology,

are discussed.

Keywords:

Kinase / Kinome / Peptide array / Phosphoproteome / Phosphorylation /

Protein arrays

1 Background

In the late 1950s, Krebs and Fischer were the first to

describe the role of reversible protein phosphorylation for

the regulation of enzymatic activity [1, 2]. For this pivotal

contribution to science they were awarded the Nobel Prize.

Protein kinases, which catalyze the transfer of the g phos-

phate group from ATP to specific serine, threonine or

tyrosine hydroxyl groups in a target protein substrate, are

now recognized as one of the largest and most important

enzyme classes. Consisting of over 500 members, human

protein kinases are responsible for modifying an estimated

one-third of the human proteome [3, 4] with many members

of the proteome undergoing complex patterns of kinase

modification at multiple sites to generate distinct isoforms

with unique functional characteristics. The presence and

dynamic nature of these phosphorylation isoforms adds a

daunting layer of complexity to characterizing and under-

standing the proteome. While much is unknown of how

Colour Online: See the article online to view Figs. 1, 3–5 in colour.

Abbreviations: Bregs, regulatory B cells; LPS, lipopolysaccharide;

ODN, oligodeoxynucleotide; PP, Peyer’s Patch; TLR, Toll-like

receptor

Correspondence: Dr. Scott Napper, Department of Biochemistry,

University of Saskatchewan, 120 Veterinary Road, University of

Saskatchewan, Saskatoon, Saskatchewan, S7N 5E3 Canada

E-mail: [email protected]

Fax: 11-306-966-7478

& 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.proteomics-journal.com

Proteomics 2011, 11, 4595–4609 4595DOI 10.1002/pmic.201100296

these modifications, even within a static infrastructure of

proteins, achieve complex biological responses, there is a

growing appreciation of the importance of kinases in

controlling cellular responses and of the potential for char-

acterizations of global kinase activity (the kinome) to offer

critical insight into biology.

There is a considerable debate as to the most appropriate

level at which to define cell responses. Transcriptional

analysis, based largely on availability and maturity of the

approach, remains the most widely applied technique for

global analysis of cellular responses. However, due to a

multitude of post-transcriptional regulatory events, there are

concerns that descriptions of patterns of gene expression, no

matter how comprehensive, do not accurately describe or

predict cellular phenotypes. Specifically, a major criticism of

genetic approaches is their inability to consider post-

transcription regulatory events such as gene silencing,

mRNA stability, unique translational efficiencies, protein

turnover, sequestration of enzymes away from substrates,

and activation and deactivation of proteins by any number of

post-translational modifications. Intuitively, characteriza-

tions of host responses that occur closer to the functional

phenotype should have greater potential to circumvent these

complicating factors and offer a clearer picture of cellular

response. From this perspective, protein kinases are at the

core of signal transduction with central roles in regulation of

virtually every aspect of cellular behavior. Through their

ability to modulate protein conformation and functional

characteristics, kinases control diverse processes such as

metabolism, transcription, cell cycle progression, cytoskeletal

rearrangement and cell movement, apoptosis, and differ-

entiation. As such, characterizations of host cellular

responses at the level of phosphorylation-mediated signal

transduction have the potential to offer important, and

predictive, insight in cellular mechanisms of phenotypes.

Investigations of cellular response at the level of protein

phosphorylation are also important and appropriate as the

disruption of kinase-mediated signaling cascades are asso-

ciated with a spectrum of diseases including cancer,

inflammation, neurological disorders and diabetes [5].

Indeed within the human genome, over 250 protein kinase

genes map to disease loci [6]. The involvement of kinases in

disease typically results from improper levels of expression/

localization/activity or mutations in the protein sequences

that alter these activities.

The role of kinases in many diseases, as well as their

regulatory role in many central pathways, makes them

logical targets for drug therapy [7]. Fortuitously, the

conserved catalytic cleft of the kinases is highly attractive for

drug design making the kinases highly ‘‘druggable’’ [8].

Kinase inhibitors have also been proposed as a more precise

mechanism for therapeutic intervention than other strate-

gies such as the targeted down-regulation of particular

genes. Not surprisingly, the central role of kinases in many

diseases, cancer in particular, and the potential to treat

complex phenotypes by targeting specific biomolecules,

have prompted drug companies to invest considerable effort

into the development of kinase inhibitors. There are esti-

mates that approximately half of the current Research and

Development budget of the pharmaceutical industry is

focused on kinases and their inhibitors. Kinases are the

most frequently targeted gene class in cancer therapeutics,

and are second only to G protein-coupled receptors across all

therapeutic areas [7, 8]. In addition to the immediate value

of these emerging molecules as therapeutics, these inhibi-

tors also represent a valuable resource with the potential for

utilization in research for hypothesis validation. Given the

magnitude of effort devoted towards their development it is

certain that additional kinase inhibitors, of greater range

and improved specificity, will be developed.

There are a number of licensed, and soon-to-be licensed,

kinase inhibitors that emphasize the potential of these

targets. Gleevac (imatinib), a potent inhibitor of the consti-

tutively active breakpoint cluster region-Abelson murine

leukemia viral oncogene homolog 1 (BCR-ABL) fusion

protein, is approved for the treatment of leukemia and

gastrointestinal stromal tumors [9, 10]. Other protein kinase

inhibitors, such as the epidermal growth factor receptor

(EGFR) inhibitors (Tarceva, Genetech) and getinib (Iressa

AstraZeneca, London UK), have either received FDA

approval or are in the late stage clinical development to treat

different cancers [11, 12]. The potential therapeutic value of

kinase inhibitors is not limited to the treatment of cancers.

For example, ruboxistaurin to treat diabetic retinopathy,

safingol, a protein kinase C inhibitor, for treatment of

atopical dermatitis and fasudil, which has received approval

in Japan, for treatment of cerebral ischemia. Therapeutic

modulation of kinase activity can also have anti-inflamma-

tory and immunosuppressive effects. For example, two

critical immunosuppressive drugs, cyclosporine A and

rapamycin, function through modulation of the phosphor-

ylation status of the cell; cyclosporine A through inhibition

of a phosphatase [13] and rapamycin through inhibition of a

kinase [14]. Other anti-inflammatory drugs that suppress

tumor necrosis factor (TNF)-a and interleukin (IL)-1bexpression also function through kinase inhibition [15].

1.1 Phosphoproteome and kinome analysis

A number of experimental approaches are available for the

analysis of phosphorylation based cellular signaling, these

can be divided into two groups, phosphoproteome and

kinome analysis [16]. The difference is dependent on

whether the consideration is on the protein kinases that

phosphorylate proteins, the kinome, or the targets of these

enzymes, the phosphoproteome. While these types of

analysis are strongly linked, representing the same biologi-

cal phenomena, and are at times considered interchange-

able, the experimental approaches are distinct. We suggest a

strict delineation where phosphoproteome analysis consid-

ers only the proteins containing phosphoryl groups and the

4596 R. Arsenault et al. Proteomics 2011, 11, 4595–4609


kinome refers to the enzymatic activities of the kinase

compliment of a cell, regardless of substrate target. This

distinction between the two should not be exaggerated,

however, as the phosphoproteome represents the net action

of the kinome, as well as that of active phosphatases.

Phosphoproteome analysis seeks to define the sub-

population of the proteome with respect to the identities and

points of phosphorylation. Phosphoproteome analysis

shares the same challenges that are associated with standard

proteomic analysis but is complicated experimentally by: the

relative scarcity of phosphoproteome members, the dynamic

nature of protein phosphorylation and the proteome size

overwhelming, or suppressing, the ability to detect the

phosphoproteome [17–19]. The small concentration of

phosphoproteins relative to the entire proteome is a signif-

icant challenge in phosphoproteome analysis. While a large

fraction of the proteome undergoes phosphorylation, many

of the proteins involved in signal transduction are expressed

at very low levels. Thus, the proteins of most interest when

studying cellular signaling are those that are the most

difficult to isolate. This is exacerbated by the fact that these

proteins, in addition to being found in low abundance, are

often phosphorylated at sub-stoichiometric levels. This

means that a small fraction of a given signaling protein is

phosphorylated at any one time; only 1–2% of the total

individual protein compliment of a cell is found in the

phosphorylated form [16–19]. When considering all these

experimental limitations, a promising alternative for char-

acterization of cellular phosphorylation is to focus instead

on the kinome. Investigations of enzymatic activities offer

greater potential for targeted, and perhaps mostly impor-

tantly, sensitive, analysis. Specifically, the well-defined and

highly conserved chemistry of enzymatic phosphorylation

permits rapid characterization of kinase activity, provided an

appropriate substrate is available.

1.2 Kinome Analysis through peptide arrays

A central obstacle for global kinome analysis is the nature of

the substrates to be employed. While proteins are the

physiological substrates for the kinases they are problematic

to mass produce and relatively unstable on array format. An

alternative is to use peptides that represent sequences

surrounding a site of phosphorylation. Many protein kina-

ses recognize phosphoacceptor sites determined by residues

surrounding the phosphorylated amino acid, as opposed to

higher order secondary or tertiary structures. Specifically,

the target specificity of many kinases is a function of the

residues in the 14 and �4 flanking positions of the phos-

phoacceptor site [20]. Synthetic peptides modeled on the site

of phosphorylation have been shown to be appropriate

substrates with Vmax and Km values approaching that of the

intact protein [21]. Relative to the complete protein, peptides

are easily synthesized, inexpensive, highly stable and

amenable to array technology [22]. Construction of arrays

representing hundreds to thousands of immobilized

peptides allows the profiling of cellular signaling activities

by determining the activities of hundreds of kinases in a

single experiment.

The basic premise of kinome analysis through peptide

arrays is that peptides representing the phosphorylation

target sites of proteins are synthesized and spotted onto an

array surface [23]. Detailed review of the commonly

employed methodologies for peptide synthesis and array

spotting are available elsewhere [16]. Following production

of the array a sample containing active kinase, or kinases,

such as a cellular lysate, is applied for the array and these

enzymes phosphorylate their respective target sequences

using ATP as the phosphate source. This phosphorylation

event is visualized using one of a number of methods

including phosphorylation-specific antibodies, radioactivity

or phospho-specific stains (Fig. 1). Quantification of the

extent of phosphorylation of a given peptide by lysates

representing different experimental conditions (control

versus for treatment) enable evaluation of relative kinase

activity. It is also possible to infer the extent of cellular

phosphorylation of the protein that is represented by the

peptide.

A large number of peptide arrays for kinome analysis are

commercially available. These arrays range in size from

dozens to thousands of peptides, representing defined

phosphorylation sites. A partial listing of the peptide arrays

that are currently available for kinome analysis is presented

in (Table 1).

2 Success stories

While still an emerging technology, the literature contains

numerous examples of the successful application of peptide

arrays for kinome analysis. The following studies carried out

by our group represent characterizations of cellular

responses to stimuli of a range of biological complexity,

from characterizing responses to defined, singular ligands

to describe changes in phosphorylation-mediated signal

transduction during disease states. These studies were

carried out in bovine and in most cases are related to the

study of immunological signaling. The use of bovine as a

model is important for a number of reasons, the commercial

importance of bovine cannot be overstated and any under-

standing that aids in the health of the bovine population is

extremely important to this industry. The ability to do

infection studies and to collect adequate amounts of mate-

rial, for examples cells and tissue, is key for peptide array

studies where large numbers of cells may be needed.

Immunological studies involving specific immune stimu-

lating ligands are well suited for peptide array work as they

tend to work through specific receptors and activate defined

signaling pathways. These pathways can then be picked out

of a complex signaling network more easily than studies that

have broad cellular effects.

Proteomics 2011, 11, 4595–4609 4597


2.1 Validating and expanding known signaling

pathways to defined ligands

Much of the early confidence in the ability of peptide arrays

to offer accurate insight into cell signaling events stemmed

from investigations with isolated cells in response to a

defined ligand. In particular, validation of the kinome array

technology was first performed by analyzing cellular

responses induced by a specific ligand for which signal-

ing events had been characterized through independent

Figure 1. Methods of detecting phosphorylated peptides: (A) Kinases specific for their target peptide sequence phoshorylate the appro-

priate residue on the array. The g phosphate group of ATP which is the phosphate donor for the kinase is radioactive and its emissions can

be visualized and quantified. (B) Direct detection of phospho-peptides by using a labelled antibody specific for the phosphorylated form of

the peptide. (C) Indirect detection of the phospho-peptide by first using a phospho-specific antibody which is then visualized by with a

labeled anti-immunoglobulin antibody. Indirect labeling can enhance sensitivity by amplifying the signal from the phosphorylation event.

(D) A labeled phosphoamino acid chelator is added after the kinase reaction. Surface plasma resonance or MALDI-TOF MS can then be

used for detection. (E) Fluorescent phosphospecific stain which only binds phosphorylated peptide can be used to directly detect the

phosphorylated residues present on the array following completion of the kinase reaction. Adapted from Jalal et al. (2007) [16].

Table 1. Commercially available peptide arrays

Name Technology Company Number ofsubstrates

Visualizationmethod

PepScan Spotted peptide array PepScanTherapeutics

10 000–100 000 Variable

PepStar Custom spotted peptide array JPT 100–1000 VariablePamChip Immobilized peptide on porous surface Pamgene 50–150 FluorescencePeptide array Unbound 96-well array ChemPep 96 RadiolabelCelluSpots Custom spotted peptide array cellulose

membraneIntavis AG 1–384 Variable

Reverse-phase proteinmicroarray

Sample blotting peptide array TheranosticsHealth

100s Fluorescence



methodologies. In one such example our group conducted

an investigation of signaling in bovine monocytes following

stimulation with ligands for different members of the same

receptor family. The Toll-like receptors (TLRs) are a family

of pathogen recognition receptors that alert the host to the

presence of microbial challenge [24]. Various microbial

associated ligands activate specific TLRs. For example,

lipopolysaccharide (LPS) activates TLR4, whereas microbial

DNA (CpGs) activates TLR9. Outside their academic inter-

est, the responses induced by each of these ligands is also of

considerable medical important with LPS representing a

causative agent of sepsis and CpG’s are promising immu-

nomodulatory molecules.

While activation of distinct members of the TLR family

has been reported to activate a conserved, and well defined,

signaling response, there are also specialized responses

associated with activation of individual family members that

are less well defined. Within this investigation, while

confirming the activation of the characterized TLR pathway,

unique, phosphorylation events which were ligand-specific

and not previously associated to TLR activity were defined

and validated (Fig. 2). For example, array data predicted

Tyr40 phosphorylation of Etk in response to LPS stimula-

tion, but not CpG stimulation. This specific response was

confirmed through the use of a phosphorylation-specific

antibody. Additionally, based on patterns of phosphorylation

of peptides corresponding to the p40phox and p47phox

proteins the arrays predicted activation of the oxidative burst

by CpG, but not LPS. The phenotypic consequences of these

apparent differences in signaling were confirmed through

functional assays. This highlights the ability of kinome

arrays to describe complex phenotypes that are multifaceted

in their regulation through phosphorylation [25].

Kinomic characterization with peptide arrays of human

monocyte responses to activation with LPS have also been

characterized by other labs with a greater emphasis on the

temporal changes that occur following exposure to a ligand

[26]. The results of this study were in close agreement with

our study and confirm the interspecies conservation of TLR

signaling pathway. These results provided novel insight into

the complexity and mechanisms of TLR signaling and

highlight the utility and power of the kinome arrays.

2.2 Understanding signaling responses induced by

phenotypically conserved, but mechanistically

distinct, ligands

Peptide arrays have been successfully applied to elucidate

signaling events induced by individual ligands, which may

share common signaling events despite interacting with

distinct receptors. The cytokines IL-32 and IL-17 are emer-

ging as critical players in the pathophysiology of chronic

inflammation and autoimmune disease. It is speculated that

the molecular mechanisms governing IL-32- and IL-17-

mediated cellular responses are differentially dependent on

the TNF pathway. Peptide-array-based kinome analysis was

used to demonstrate that a peptide target corresponding to

TNF-receptor-1 (TNF-R1) was phosphorylated in the

presence of cytokine IL-32 but not IL-17. This critical finding

was then independently validated through the use of a

monoclonal antibody to specifically block TNF-R1 and

Figure 2. Conserved and

distinct TLR4 and TLR9 signal-

ing in bovine monocytes:

Signaling events indicated by

peptide array kinome analysis

in bovine monocytes following

treatment with LPS or CpG.

Canonical signaling pathways

as well as novel signaling

events were observed. Adapted

from Arsenault et al. (2009) [25].

Proteomics 2011, 11, 4595–4609 4599


suppress IL-32-induced downstream responses. This

approach provided independent confirmation that IL-32-

mediated activity was dependent on TNF-R1. In contrast,

and consistent with the array results, blocking TNF-R1 did

not impact IL-17-induced downstream responses. This study

provided critical insight into the molecular mechanism of

the differential dependence of IL-32 and IL-17 on the TNF-

pathway and is likely to be of immediate therapeutic

importance.

The same investigation of IL-32- and IL-17-induced

signaling responses also identified p300 (transcriptional

co-activator) and death-associated protein kinase-1 (DAPK-1)

as common protein phosphorylation targets for IL-32 and IL-

17. Phosphorylation of p300 and DAPK-1 upon cytokine

stimulation was confirmed by immunoblots. The presence of

these common targets was further supported by additional

results demonstrating similar transcriptional responses, the

direct activation of nuclear factor k-light-chain-enhancer of

activated B cells (NF-kB) and the induction of chemokine

production by both IL-32 and IL-17. Furthermore, knock

down of p300 and DAPK-1 altered downstream responses

induced by IL-32 and IL-17, and impacted certain cellular

responses induced by TNF-a and IL-1b. These findings

allowed the authors to hypothesize that p300 and DAPK-1

represent nodes where the inflammatory networks of IL-32

and IL-17 overlap, and that p300 and DAPK-1 impact both

TNF-dependent and -independent processes. Therefore p300

and DAPK-1 may be viable therapeutic targets for chronic

inflammatory diseases [27]. Importantly, in this example, use

of peptide arrays greatly accelerated discovery in contrast to

using other techniques to investigate hypothesized targets

one at a time.

2.3 Understanding phenotypic variability

The power of kinome analysis, and in particular utilizing

peptide arrays, is well exemplified by efforts to elucidate the

cellular mechanisms of phenotypic differences between

closely related cells. The challenge of these scenarios is that

the cells may appear to present a highly conserved infra-

structure of biomolecules but subtle differences in signal

transduction through this network may dictate unique

functional phenotypes. Intestinal Peyer’s patch (PP) regu-

latory cluster of differentiation (CD)211 B cells (Bregs)

suppress TLR9-induced innate immune responses.

However, it is unknown if TLR9 activation is similarly

regulated in PP Bregs. We investigated the TLR9 agonist,

CpG oligodeoxynucleotides (ODN) induced response in the

PP Bregs [28]. It was observed that PP CD211 Bregs had high

expression of TLR9 mRNA, but would not proliferate upon

stimulation with CpG ODN. Furthermore, while CD211 B

cells from blood secreted IgM or IL-12 following CpG ODN

stimulation, PP CD211 Bregs did not. We hypothesized that

PP CD211 Bregs unresponsiveness resulted from a lack of

ability of CpG to activate TLR9 signaling in this type of cell.

This hypothesis was confirmed when kinome analysis

showed dynamic patterns of phosphorylation in adapter

proteins of TLR such as interleukin-1 receptor-associated

kinase (IRAK)1, mitogen-activated protein kinase kinase

kinase 7 (TAK1), I-k-B kinase (IKK) and NF-B-p65 in blood

CD211 B cells stimulated with CpG, indicating an activa-

tion of the TLR9 pathway. In contrast, CpG ODN stimula-

tion of PP CD211 Bregs displayed phosphorylation events in

TLR9 adapter proteins indicative of TLR9 pathway inacti-

vation. This lack of apparent TLR9-signaling events for

peptides representing proteins which are early inter-

mediates in the TLR-signaling pathway indicated that

signaling is blocked close to the receptor (Fig. 3). These

Figure 3. TLR9 signaling differences in related cell types: There

are numerous phosphorylation differences between PP CD211

Bregs and blood CD211 B cells. The spots indicate the proteins

and residues which were differentially phosphorylated following

stimulation with CpG ODN. CD211 B cells from blood displayed

strong TLR9 responses which CD211 Bregs displayed no TLR9

response. Adapted from Booth et al. (2010) [28].



observations point toward a novel means by which TLR-

expressing cells control TLR responses in cells with regu-

latory functions. Our observations suggest a novel

mechanism by which TLR responses are regulated in TLR-

expressing cells with regulatory functions [28].

2.4 Cancer profiling

Cancer is the prototypical kinase disease and for many

researchers the study of kinases is intimately associated

with the study of cancer. It is very often a disregulation

of a subset or even a single protein kinase that is the cause

of cancer. In fact, kinases and their regulators are

commonly mutated oncogenes and tumor suppressors [29].

Kinase targeting is by any measure a massive portion of

current cancer research. Thus, it is logical that cancer

has had an outsized impact on the field of kinome

analysis and the development of peptide arrays. The power

that peptide arrays have is their ability to screen multiple

(up to 1000 or more) kinase target sites in one experiment.

In the drug discovery field, this is an important considera-

tion and can significantly increase the rate of research. In

addition, the kinome profiling of various tumors and

cancerous cells provides important biological information

on how cancers from individual patients compare

with respect to what they have in common and what defines

their individual biology. Important work in this field has

been carried out on pediatric brain tumor profiling

[30], leukemia profiling [31], chondrosarcoma [32], colon

cancer [33].

The screening of therapeutic agents that inhibit key

phosphorylated signaling molecules is a major focus of

cancer research. However, a critical consideration is that the

animal model or cell culture-based model is accurately

representing the true biology of target cancer cells in the

host. The combination of species-specific kinome arrays and

the Bovine Leukemia Virus (BLV) ovine model system,

allowed analysis of changes in kinase activity that occur

between host and culture in transformed B cells [34]. The

results indicated that the phosphorylation patterns were

significantly altered when cancer cells were removed from

host and passaged for use in tissue culture. These changes

were in pathways that define transformation in this cell type,

which could possibly alter results if one was using such a

host-derived cell line for therapeutic screens. This observa-

tion suggests that the external environment of the cells has a

profound effect on biologically important cell signaling

events. This analysis, using high-throughput kinome tech-

niques, was able to identify key phosphorylation events that

define cancer progression in B cells. In addition, to deter-

mine the alteration of critical signaling events between

primary cancer cells in vivo and cultured cells in vitro. Loss

or gain of key signaling pathways when cells are cultured

has a significant impact on the translation of knowledge

from the bench to the clinic [34].

3 New opportunities for peptide arraykinome analysis

3.1 Species-specific peptide arrays

Until recently the peptide arrays that have been created for

kinome analysis were based on phosphorylation events

characterized from the homologous species. The vast

majority of characterized phosphorylation events are for

human and mouse with only limited information available

for other species. This creates a barrier for creation of arrays

for species where the phosphoproteome has yet to be

defined. However, as the specific sites for protein phos-

phorylation, and their subsequent biological consequences,

are often conserved we previously hypothesized that it would

be possible to predict the sequence contexts of phosphor-

ylation events in proteins of other species based on genomic

information.

To test this hypothesis, our lab interrogated Phosphosite

(www.phosphosite.org) and Phosphobase (phospho.

elm.eu.org) which are publically available online databases

that hold information on manually curated and literature-

based serine, threonine and tyrosine phosphorylation sites.

The information contained within these sites is predomi-

nantly for phosphorylation events from human and mouse.

Search results produce peptide sequences, usually about 15

amino acids in length, which correspond to characterized

phosphorylation sites. Information about the recognizing

kinase, as well as links that describe the biological function

of the specific modification, is listed. The amount of

sequence conservation between bovine and human kinase

phosphorylation target recognition sites was investigated by

searching nearly a thousand peptides of 15 amino acids in

length, representing human phosphorylation sites, against

the NCBI-NR protein database by Blastp program in order to

generate orthologous bovine peptides. The results indicated

that approximately half of the bovine sequences matched

100% to those of human. The majority of the remaining half

had limited sequence differences, usually amounting to only

1 or 2 amino acids. An annotation comparison between the

query and hit sequences was used to confirm that both

referred back to the same protein identity [16]. These results

show that phosphorylation sites are not fully conserved

between human and bovine and these differences would

limit the utility of using human arrays to analyze the bovine

kinome by this method. Interestingly, other labs have

utilized human-based peptides arrays for species as evolu-

tionarily distant as Arabidopsis [35]. This, in our opinion,

speaks more highly of the enthusiasm to apply peptide

kinome to species of interest, and the lack of species-

appropriate tools, than to the value or appropriateness of

attempting to transcend species barriers with generic

peptide arrays.

For our first generation bovine-specific peptide array, 300

peptides were selected from the initial list of 880 peptides.

Target peptides represented proteins involved in a spectrum

Proteomics 2011, 11, 4595–4609 4601


of biological events but with emphasis on proteins, path-

ways and processes associated with innate immunity.

Specific peptides were selected based on biological signifi-

cance and degree of sequence conservation surrounding the

phospho-acceptor. These criteria were predicated on the

assumption that sequence conservation may reflect func-

tional conservation. Notably, many proteins undergo phos-

phorylation at multiple sites to control discrete aspects of

protein function. Therefore, numerous proteins on the array

were represented by multiple peptides, representing distinct

phosphorylation events, to allow more detailed insight into

cellular responses under different treatment conditions.

This has proven a powerful feature of the arrays in providing

a comprehensive assessment of how critical proteins are

regulated by complex and dynamic patterns of phosphor-

ylation at multiple sites. Each peptide was printed in tripli-

cate on the array to provide a measure of technical

reproducibility within each experiment (Fig. 4).

Our first generation arrays, and many of the arrays

reported in the literature, employ identical blocks that are

printed in triplicate. There may be an advantage, however, to

randomly place peptides within replicate blocks to minimize

possible effects resulting from the relative placement of a

peptide within the grid. Peptide synthesis and printing is

performed by a commercial partner (JPT Technologies) and

utilization of the arrays in the lab does not depend on any

specialized equipment that would not be present in labs

conducting transcriptional array investigations.

We have employed a similar bioinformatics approach to

develop a porcine peptide array. A similar degree of conser-

vation was observed when using human phosphoproteome

databases to identify putative porcine phosphorylation sites as

was encountered for the bovine array. Generation of arrays

for species for more evolutionarily distant organisms does

become problematic. For example, peptide sequences corre-

sponding to thousands of characterized phosphorylation

events were investigated using the same protocol to establish

the degree of conservation across several mammalian species

as well as select plants and insects. For organisms such as

plants and insects it is much more problematic to predict

phosphorylation sites based on these linear bioinformatic

approaches (Table 2). For these organisms, greater effort to

take into account unique biology associated with these

species as well as degrees of kinase conservation, their

consensus sequences and phosphorylation targets will also

assist in creating arrays for these species.

Databases are emerging describing the kinase comple-

ment for economically important plant species such as rice

(phylomics.ucdavis.edu/kinase/). Given the economic

importance of these organisms, in particular plants, and the

potential for kinome analysis to offer functional insight into

phenotypes and provide biomarkers for selective breeding,

these efforts are certainly justified. Interestingly, a number

of groups have reported on peptide array kinome analysis of

plants utilizing peptide arrays corresponding to human

phosphorylation sequences.

3.1.1 Non-traditional species

A tremendous advantage of kinome analysis for systems

biology investigations is that kinases are highly conserved

across species. This facilitates extrapolation of research find-

ings to other species and allows researchers to test inhibitors

of human kinases in animal models of disease. The only, but

often overlooked, caveat is the requirement for appropriatemodels of disease. The recent characterization of the mouse

kinome has been heralded for its ability to ‘‘enhance the

Figure 4. Peptide array diagram:

Representation of a custom peptide

array from JPT Peptide Technologies.

Three hundred peptides are printed

within each block which is then prin-

ted in triplicate. Control spots are

printed outside of these blocks to aid

in array orientation and gridding

during analysis. (A) Single block of

peptides. Hollow circles indicate

custom designed peptides, solid black

circles are control spots. (B) Entire

array chip with block printed in tripli-

cate. Adapted from JPT schematic.



exploration of the roles of all kinases in mouse models of

human diseases’’ [36]. While mice offer unprecedented

opportunity for generation of mutants to study kinase func-

tion and consequences of kinase ablation, the value of the

mouse as a model for human disease are far less definitive.

The availability, cost and ease of handling inbred,

genetically defined mouse strains have made this the species

of choice for most animal experimental work. Unfortunately,

these selection criteria do not ensure accurate disease models

and efficient knowledge translation. An ever increasing

number of infection studies, with a variety of different

stimuli, pathogens and treatments, have demonstrated

divergent biological mechanisms, responses and distinct

pathologies when comparing mice and humans [37]. This

has prompted many researchers to invest greater resources

into the development of animal models, which more accu-

rately reproduce relevant human physiology and pathology.

Furthermore, the use of animals in biomedical research

ethically requires selection of species that match the human

disease as closely as possible to minimize the number of

animals required and maximize the potential benefit.

There are a number of factors that limit the capacity of

mouse models to accurately reflect human responses. First,

from our specific interest in infection and immunity, when

vaccinating or challenging an outbred population it has been

frequently observed that there is a binomial distribution

in immune responses referred to as ‘‘low’’ and ‘‘high’’

responders [38]. This reflects a genetic component that

influences the magnitude of immune responses within

individual animals. This individual variation is an essential

consideration when developing therapeutics for humans

and is best represented by outbred species. Second, effective

animal models must share the same physiological char-

acteristics as humans, or at least reflect them as closely as

possible. For example, the skin physiology of the pig is very

similar to humans that render the pig a good model for

studying intracutaneous or topical delivery of compounds.

3.1.2 Interspecies comparative analysis

Efficient translation of research from animal models to

clinical application will require more detailed knowledge

regarding conservation of cell signaling responses in model

species. For comparative kinomic investigations, it is

possible to design arrays reflecting specific biology or

signaling pathways for all species of interest. This can be

achieved through the selection of species-specific peptides

from orthologous proteins when sufficient proteomic

information is available for all species. Alternatively, it is

possible to design arrays using peptides that are absolutely

conserved in sequence for all species of interest. These

arrays could then be used to interrogate samples of interest

from each species. This approach would minimize potential

technical problems associated with the use of different

arrays but it may be necessary to then validate that

conserved peptides are selected from orthologous proteins.

The use of outbred species for kinome analysis also raises

a number of challenges due to genetic and environmental

differences among individual animals. Both these factors

may significantly influence kinase activity before and after

experimental manipulation of isolated cells or the host.

Therefore, experimental design and subsequent data analysis

must accommodate these potential sources of data variation.

Variation in kinase activity among individual animals has

presented a challenge when analyzing data for conserved

responses to a specific pathogen or stimulus. This variation

in kinase activity may also provide an opportunity to further

understand the biological basis for specific differences in

host phenotype or responses. We have initiated kinome

studies in animal populations with defined differences in

their susceptibility to pathogen infection. It is too early in

these studies to determine whether kinome analysis will

provide specific biomarkers of disease resistance and

susceptibility or provide significant insight into the biological

mechanisms mediating these specific phenotypes.

Table 2. Phosphorylation target site differences between species

Sequence difference % of Proteins versus Homo sapiens

Cow (%) Chicken (%) Mouse (%) Pig (%) Rice (%) Fruit fly (%) Saccharoycescerevisiae (%)

0 34.82 12.78 27.29 26.82 0.03 0.54 0.071 15.86 8.56 15.04 11.35 0.03 0.66 0.082 9.66 6.71 9.58 7.07 0.02 0.86 0.163 6.43 5.33 6.39 4.50 0.03 1.00 0.164 4.59 4.62 4.60 3.20 0.04 1.01 0.175 3.09 3.82 3.10 2.41 0.07 1.15 0.196 1.63 2.74 1.61 1.11 0.05 1.10 0.187 0.51 1.33 0.47 0.42 0.02 0.84 0.128 0.08 0.33 0.05 0.08 0.01 0.42 0.03No similarity 23.34 53.75 31.88 43.03 99.69 92.31 98.83

Various species display differences in their phosphorylation target site amino acid sequence. As expected the further one movesphylogenically away from human the greater the number of sequence differences are observed. Human target sequences were obtainedfrom PhosphoSitePlus (www.phosphosite.org) and were blasted against the corresponding animal proteome.

Proteomics 2011, 11, 4595–4609 4603


Transcriptional analysis has been used to compare host

responses to a variety of pathogens and identify pathogen-

specific immune evasion strategies [39]. Validation of sustained

transcriptional responses following infection is difficult and

recent evidence that pathogens produce miRNA [40] further

complicates the interpretation of host transcriptional responses.

We have used kinome arrays biased for innate immune

responses to compare cellular responses to a variety of patho-

gens. These analyses provide the opportunity to not only identify

specific innate immune responses that have been disrupted, but

also identify the specific adapter proteins that have been targeted

by the pathogen. This level of knowledge may be critical when

comparing clinical isolates of pathogen that vary in virulence or

designing attenuated pathogens for use in vaccines.

3.2 Process-specific peptide arrays

An additional advantage of the peptide arrays is that

peptides can be selected to enable focussed investigation of a

particular priority cellular process. Such wilful construction

offers the researcher richer information about signaling

events associated with a particular biological response such

as apoptosis, cell division or, in the case of our bovine array,

innate immunity. Our approach to date has been to devote a

considerable portion (half to two thirds) of the peptides on

the array to central, but generic, signaling pathways. This

enables novel discovery of how a stimulus may impact

cellular responses in a manner which may not be immedi-

ately intuitive. The remaining peptides represent focussed,

hypothesis-driven selections which enable the researcher to

better interrogate specific cellular responses which are of

priority. As the design and production of customized arrays

becomes increasingly routine it is easy to envision kinome

investigations through peptide arrays following an iterative

process of array design with increasing focus and repre-

sentation of particular processes within an investigation.

3.3 Non-radioactive protocols

Many of the initial peptide array kinome protocols utilized32P-g-ATP for detection and quantification of peptide phos-

phorylation. While effective such approaches are disadvan-

taged in their cost, sensitivity and safety. More recent

manuscripts describe the use of phospho-specific fluor-

escent stains that are compatible with the arrays, less costly

and suitable for use with currently available microarray data

scanners. In our hands the use of these stains, in particular

Diamond ProQ phosphostain, offers greater sensitivity and

reproducibility than radioactive protocols [25].

3.4 Flow-through 3D arrays

A variation of the peptide array for kinome analysis discussed

above is the flow-through 3D array [41]. This technology is

epitomized by the PamChip system developed by Pamgene of

the Netherlands. Peptides that correspond to specific kinase

target sites are covalently immobilized on an activated alumi-

num oxide surface. The unique nature of this array is the

surfaces on which the peptides are immobilized are a series of

pores through which the analyte solution can be pumped. As

the solution is pumped back and forth through the pores any

kinase contained within will phoshorylate its corresponding

target peptide. This is visualized using phosphorylation-specific

antibodies bound to a fluorophore. The pumping rate and

number of cycles can be varied and a fluorescence reading is

taken after every pump cycle. Following each pumping cycle

more peptides will be phosphorylated resulting in greater

fluorescent signal. This allows the user to take kinetic as well as

end point measurements of kinase enzyme activity (Fig. 5) [42].

This technique has been applied for a number of important

studies including kinome profiling of pediatric brain tumors

[30], development of protein kinase C (PKC) inhibitors [43] and

leukemia treatment target identification [31].

3.5 Bead-based solution-phase heterogeneous

kinase assays

Solution-phase peptide-based kinase assays are an alter-

native to the arrays based on immobilized peptide on a solid

array surface discussed above. The technique is based on the

covalent linking of peptide substrates onto beads via an

acrylamide linker [44]. Each individual bead will contain a

number of identical peptide substrates linked to its surface.

A mixture of beads containing different peptide substrates

are combined within a single well of a 96-well plate. To these

wells is added cell lysates that contain active protein kinases.

These kinases will then phosphorylate their respective target

peptides. The reaction is then stopped and synthetically

phosphorylated internal standards are added to beads.

Fluorescent antibodies against phospho-peptides are added

to the wells to bind to the phosphorylated peptide. Sylvester

and Kron [44] used Luminex beads and the Luminex flow

system to identify both the beads, thus the peptide substrate,

and the antibody, thus the peptides phosphorylation state,

through a measurement of two channel fluorescence. The

fluorescent of the bead indicated peptide identification and

the fluorescence of the phospho-specific antibody indicated

phosphorylation of the peptide substrate by the active kinase

in the sample. The ability to screen multiple peptides within

a single well of a 96-well plate allows various conditions,

concentrations or treatments to be assayed within a single

experiment. The main advantage of this technique over

solid-supported peptide substrates is that the kinases have

greater access to their target peptides within solution [45].

With a uniform distribution within solution, beads linked to

peptide substrates have a higher reaction rate with kinases

than solid-phase arrays. In addition, beads provide a larger

surface area for reactions than a two-dimensional array

surface and a high loading capacity of peptide on the beads.



4 Remaining challenges

4.1 Statistics and mining

The microarray field has been dominated by genomics with

high-throughput methods based on either oligonucleotide or

cDNA arrays. With the advent of peptide and antibody arrays,

it was assumed that the mathematical manipulations and

statistical tools developed for nucleotide arrays would be

directly applicable to these new technologies. As time has

gone on and peptide/antibody arrays have become more

established and diverse in their applications, it has become

clear that their full potential is not being exploited by these

copied analytical methodologies. There are several reasons

for this deficiency. First, the number of kinase targets or

antibodies bound to the array is usually much lower than the

typical oligonucleotide or cDNA array, potentially up to 20

times less. This means that each individual spot on the array

is of much more value and the indiscriminate discarding of

data points is undesirable. The statistical methods used to

analyze the data need to reflect this difference. In addition,

the two techniques are fundamentally different in a number

of ways. For nucleotide arrays, the hybridization of labeled

cDNA is the means of achieving a signal, but kinome arrays

require the transfer of phosphate groups by enzymatic

activity. This means that there is a scale of response from

zero phosphorylation to many multiples of the control and it

is difficult to determine where a cut off should be set. Also,

the variance in signal is affected by this dynamic range of

phosphorylation making statistical assumptions difficult.

Data normalization are required before statistical tests can be

conducted; however, normalization procedures are based

primarily on nucleotide microarray data and may not be the

most appropriate for kinome data.

Once these difficulties have been dealt with the inter-

pretation of the collected data is another hurdle to overcome.

Figure 5. Flow-through array schematic: The 96-well PamChip array contains a porous metal surface to which peptides are immobilized.

The sample is pumped through and the kinase has the opportunity to phosphorylate its given target sequence. Following each pump

cycle, a fluorescence reading is taken. This allows for a real-time kinetic readout of enzymatic activity. Thus, this technique allows for a

kinetic as well as endpoint measurement of peptide phosphorylation and kinase activity. Adapted from Koppen, et al. (2009) [42] and

Pamgene (www.pamgene.com/index.php/PamChip/).

Proteomics 2011, 11, 4595–4609 4605


With any high-throughput method there will be a wealth of

data generated and one must decide how to mine it for

significant and biologically relevant results. This has been

historically difficult with kinomic array data and many

papers have focused on identifying data which confirm

previous low-throughput methods and then focusing on one

or two novel phosphorylation events that can be confirmed

with Western blots or bioassays. While it is important that

array data agree with the previous reports the main goal of

this method is to discover novel biology and in many cases

uncover targets for therapy. A more systematic data mining

approach is needed to fully explore the value of kinome data.

4.2 Kinase specificity/efficiency

A great unknown that all kinomic researchers are faced with

is the specificity with which an individual kinase recognizes

a spotted target sequence and how efficiently it phosphor-

ylates that sequence. While enzymatic work has been done

showing kinases phosphorylate target peptides with

comparable mechanics to native protein [46, 47], the speci-

ficity and efficiency of each individual kinase or each target

sequence is unknown. This can greatly affect the data

interpretation if one considers the magnitude of phosphor-

ylation on the array as biologically significant. Indeed as

Sikkema et al. [30] wrote in their kinome profiling study,

‘‘ydetermining the sensitivity of a substrate for specific

kinases will prove to be of tremendous value in future

application of the peptide array.’’ They go on to summarize

the largest drawback of the current state of peptide array

kinome research. ‘‘Therefore, generating hypotheses is the

limit of what is possible at the current developmental stage

of peptide microarray technology.’’ The peptide arrays in

their current state simply point us in directions of further

study using other, more established, methods. This is

further discussed in the following section.

4.3 A systematic approach needed

A new approach must be taken in the analysis of data

generated by the kinomic peptide arrays. Instead of simply

being a means of generating hypotheses for future work, the

peptide arrays can be developed into a tool to provide a

better picture of what is going on in the cell under a given

condition. This is currently not possible and will require

tools that synthesize all of the data from the array and

organize it into functional groupings or networks. This

‘‘value-added’’ approach to kinomic peptide array data will

allow researchers to not only simply find a single interesting

phosphorylation event to study further or give a list of

differential phosphorylation events between two treatments/

diseases, but provide biological relevant data. Examples may

include which cell signaling pathways are affected by a given

treatment and at what point is it diverging from control,

what cell function is affected by a disease and how a specific

function is linked to other core functions, such as metabo-

lism, cell cycle and apoptosis.

In addition to the previously mentioned additional

capacities, in the future integration of kinome data with

other high-throughput methods will be needed. As Parikh

and Peppelenbosch [48] wrote in their review of kinome

profiling and cancer ‘‘ythe future lies in moving beyond

the idea of an individual ‘‘omics’’ approach and studying a

biological system in its entirety by combining data from all

different omics technologies, so that disease can be better

modeled and thoroughly understood compared with a single

pathway approach.’’ The synthesis of genomics, tran-

scriptomics, proteomics and kinomics to aid in under-

standing is the important development still to come to move

peptide array kinomics into its own rather than being a

hypothesis generating technique.

4.4 Validation of kinome data

The philosophy of our lab is that the information emerging

from the arrays should be utilized to suggest signal beha-

viors and their associated phenotypes, which can then be

validated through independent experimental approaches.

Such validation could come in the form of targeted phos-

phoproteome investigations with phosphorylation-specific

antibodies. Alternatively, where species-appropriate phos-

phorylation-specific antibodies are not available, it is

necessary to experimentally investigate the phenotypic

events that are suggested by the kinome data such as

apoptosis, oxidative burst and pro/anti-inflammatory

responses. The practice of validating kinome results through

independent approaches also helps to determine the

number of biological, and most importantly technical,

replicates that are appropriate. The number of replicates

should be sufficient to ensure that a reliable and statistically

significant picture of the signal emerges.

4.5 Biological importance

The protocols employed for peptide-array-based kinome

analysis introduce biological artefacts which compromise the

value of the emerging data. In particular, disruption of cells

to release cellular kinases results in an inherent loss of

cellular structure and organization that is undoubtedly

biologically important. Loss of this organization within the

context of a cellular lysate likely impacts the activity of

kinases through the loss of regulating interactions. Similarly,

the peptide arrays may present a kinase the opportunity to

phosphorylate a substrate which it would not normally

encounter in the cell due to compartmentalization or the

absence of co-incidental substrate and kinase expression.

The specificity of kinases depends not only on their ability

to recognize and modify a protein target, but also on their



spatial and temporal access to these targets. The intracellular

environment is highly complex and organized with the

particular and dynamic location of various biomolecules,

including kinases and their substrates, providing an impor-

tant level of control for biological functions. As such the

translocation of a cellular kinase from the cytosol to more

specialized locations can serve to regulate activity. For

example, b-adrenergic receptor kinase translocates from the

cytosol to the cell membrane to phosphorylate and desensi-

tize G-protein-coupled receptors [49–51], while mitogen-

activated protein kinase (MAPK) translocates from the cyto-

sol to the nucleus to regulate gene transcription [52]. This

situation can become even more complex for multipurpose

kinases that regulate distinct and functionally unrelated

processes within in the same cell. For example, PKCdregulates both depolarization events, apoptosis or

cell growth through either translocation to the cell

membrane to inhibit inward recitifying K1 channels or

translocation to the mitochondria to activate ATP-dependent

K1 channels or translocation to the nucleus for regulation of

gene transcription respectively [53–55]. Depending upon the

specific needs of a cell these distinct patterns of translocation

and associated responses can be achieved simultaneously.

Targeted regulation of kinase activity through compart-

mentalization is not limited to dynamic patterns of kinase

localization, but is also achieved through the discrete

formation of factors that regulate kinase activity. For

example, formation of cyclic adenosine monophosphate

(cAMP), a potent activator of protein kinase A, has been

shown to occur within discreet microdomains in cardiac

myocytes in response to b-adrenergic stimulation [49]. This

in turn would be anticipated to result in equally compart-

mentalized activation of protein kinase A (PKA), which was

suggested by the authors of the paper to represent a

mechanism for compartmentalized activation of PKA in

specific sub-cellular locations [56].

The loss of cellular organization likely represents a

greater challenge for kinome rather than phosphoproteome

investigations as kinome analysis depends on measure-

ments of enzymatic activity which are taken post-lysis,

whereas phosphoproteome analysis characterizes events

which occur prior to cell lysis.

5 Concluding remarks

A reliable methodology for characterizing host signal

transduction activity offers tremendous promise to under-

stand both cellular physiology and pathophysiology. This is

not a trivial task giving the complexity and dynamic nature

of these responses. Given the current demonstrated poten-

tial for utilizing the kinases to understand biology as well as

providing targets for therapeutic intervention, these efforts

are well justified. As we continue to realize the potential of

kinase inhibitors tools such as the peptide arrays for kinome

analysis may ultimately lead to better application of a wide

variety of therapies to enhance individual patient medicine

and minimize the risk of adverse reactions.

While much of this initial excitement has been within the

context of cancer the potential to define the cellular

mechanisms of disease through kinome analysis and the

opportunity to treat these diseases with kinases inhibitors is

being appreciated for an increasingly wide spectrum of

diseases. We anticipate that in the future kinome analysis

will have increasingly important roles in the study of

host–pathogen interactions. The interaction between a

particular pathogen and its target cell are often complex and

multifaceted presenting a considerable challenge for any

experimental approach. The effort required to understand

these complex interactions are justified, in particular for

pathogens that establish chronic infections, because

understanding the mechanisms used by a pathogen to

subvert the host response provides critical targets for ther-

apeutic intervention. In particular, many pathogens have

been shown to target host phosphorylation-mediated signal

transduction pathways. Therefore, understanding host

responses at the level of the kinome is highly appropriate.

Interestingly, the pathogenic mechanisms of several patho-

gens have been shown to involve the production eukaryotic-

like kinase effector molecules, which are translocated into

the host cell for direct subversion of host processes [57].

These eukaryotic-like kinases have been identified as highly

attractive therapeutic targets [58].

A large number of practical and philosophical questions

and challenges remain to be addressed for peptide-array

kinome analysis in the coming years. We anticipate that the

development of customized software packages for the analy-

sis and interpretation of kinome data will represent a turning

point in the utility of this approach. That is researchers can

place greater confidence in the biological meaning of

kinomics data with less requirement for development of

specialized expertise for understanding the data there will be

a greater uptake of the technology into non-specialized labs

wishing to embark in this sub-discipline of proteomics.

There also remains a key philosophical question about

the arrays that holds considerable practical importance. Do

they represent kinome arrays or phosphoproteome arrays?

Or both? As the arrays depend upon the activities of the

cellular kinases to phosphorylate the peptide targets this

would seem consistent with a kinome array. However, as the

extent of phosphorylation of the peptide is used as a

surrogate marker to predict the extent of phosphorylation of

the corresponding protein within the cell this is more

consistent with phosphoproteome analysis. This is not a

trivial distinction. For example, in each scenario, one might

consider the number of peptides that would be required for

a comprehensive array. From the perspective of a kinome

array just over 500 peptides, if they represent substrates of

ideally specificity of the corresponding kinases, would offer

a global perspective on the relative activities of each kinase.

From a phosphoproteome perspective, the array would

require peptides to represent each cellular phosphorylation

Proteomics 2011, 11, 4595–4609 4607


event, tens of thousands in number, to obtain a compre-

hensive understanding of phosphorylation-mediated signal

transduction events within the cell. There would obviously

be considerable differences as well in how this data would

be handled, interpreted and validated. Future studies will be

required to further define kinase specificity before the

optimal balance between peptide array design and data

interpretation can be achieved.

The authors acknowledge the financial contributions of theCanada Research Chairs program (Neonatal Mucosal Immu-nology) and the Natural Sciences and Engineering ResearchCouncil.

The authors have declared no conflict of interest.

6 References

[1] Krebs, E. G., Fischer, E. H., Phosphorylase activity of skeletal

muscle extracts. J. Biol. Chem. 1955, 216, 113–120.

[2] Fischer, E. H., Krebs, E. G., Conversion of phosphorylase b

to phosphorylase a in muscle extracts. J. Biol. Chem. 1955,

216, 121–132.

[3] Manning, G., Whyte, D. B., Martinez, R., Hunter, T., Sudar-

sanam, S., The protein kinase compliment of the human

genome. Science 2002, 298, 1912–1934.

[4] Hunter, T., Protein kinases and phosphatases: the yin and

yang of protein phsophorylation and signaling. Cell 1995,

80, 225–236.

[5] Manning, G., Caenepeel, S., Encyclopedia of protein kinases

in human diseases, Catalog Reference Manual, Cell

Signaling Technology, Beverly, MA, 2005.

[6] Knuutila, S., Bjorkqvist, A. M., Autio, K., Tarkkaqnen, M.

et al., DNA copy number amplifications in human

neoplasms: review of comparative genomic hybridization

studies. Am. J. Pathol. 1998, 152, 1107–1123.

[7] Cohen, P., Protein kinases–the major drug targets of the

twenty-first century? Nat. Rev. Drug Discov. 2002, 1, 309–315.

[8] Hopkins, A. L., Groom, C. R., The druggable genome. Nat.

Rev. Drug Discov. 2002, 1, 727–730.

[9] Druker, B., Tamura, S., Buchdunger, E., Ohno, S. et al.,

Effects of a selective inhibitor of the Abl tyrosine kinase on

the growth of Bcr-Abl positive cells. Nat. Med. 1996, 2,

561–566.

[10] Druker, B., Talpaz, M., Resta, D., Peng, B. et al., Efficacy and

safety of a specific inhibitor of the CR-ABL tyrosine kinase in

chronic myeloid leukemia. N. Engl. J. Med. 2001, 2,

561–566.

[11] Noble, M., Endicott, J., Johnson, L., Protein kinase inhibi-

tors: insights into drug design from structure. Science 2004,

303, 1800–1805.

[12] Eglen, R. M., Reisine, T., The current status of drug

discovery against the human kinome. Assay Drug Dev.

Tech. 2009, 7, 22–43.

[13] Liu, J., Farmer, J. D., Lane, W. S., Friedman, J. et al.,

Calcineurin is a common target of cyclophilin–cyclosporin A

and FKBP–FK506 complexes. Cell 199, 166, 807–815.

[14] Heitman, J., Movva, N. R., Hall, M. N., Targets for cell cycle

arrest by the immunosuppressant rapamycin in yeast.

Science 1991, 253, 905–909.

[15] Lee, J. C., Laydon, J. T., McDonnell, P. C., Gallagher, T. F.

et al., A protein kinase involved in the regulation of

inflammatory cytokine biosynthesis. Nature 1994, 372,

739–746..

[16] Jalal, S., Kindrachuk, J., Napper, S., Phosphoproteome and

kinome analysis: unique perspectives on the same problem.

Curr. Anal. Chem. 2007, 3, 1–15.

[17] Mann, M., Ong, S. E., Gronborg, M., Stern, H. et al., Trends

Biotechnol. 2002, 20, 261.

[18] Kalume, D. E., Molina, H. and Pandey, A., Tackling the

phosphoproteome: tools and strategies. Curr. Opin. Chem.

Biol. 2004, 7, 64–69.

[19] Gorg, A., Weiss, W., Dunn, M. J., Current two-dimensional

electrophoresis technology for proteomics. Proteomics

2004, 4, 3665–3685.

[20] Kreegipuu, A., Blom, N., Brunak, S., Jarv, J., Statistical

analysis of protein kinase specificity determinants. FEBS

Lett. 1998, 430, 45–50.

[21] Zhu, H., Klemic, J. F., Chang, S., Bertone, P. et al., Analysis

of yeast protein kinases using protein chips. Nature Genet.

2000, 26, 283–289.

[22] Ouyang, Z., Takats, Z., Blake, T. A., Gologan, B. et al.,

Preparing protein microarrays by soft-landing of mass-

selected ions. Science 2003, 301, 1351–1354.

[23] Houseman, B. T., Huh, J. H., Kron, S. J., Mrksich, M.,

Peptide chips for the quantitative evaluation of protein

kinase activity. Nature Biotech. 2002, 20, 270–274.

[24] Kindrachuk, L., Potter, J., Wilson, H., Griebel, P., Napper, S.,

Activation and regulation of toll-like receptor 9: CpGs and

beyond. Mini. Rev. Med. Chem. 2008, 8, 590–600.

[25] Arsenault, R. J., Jalal, S., Babiuk, L. A., Potter, A. et al.,

Kinome analysis of toll-like receptor signaling in bovine

monocytes. J. Recept. Signal. Transduct. Res. 2009, 29,

299–311.

[26] Diks, S. H., Kok, K., O’Toole, T., Hommes, D. W. et al.,

Kinome profiling for studying lipopolysaccharide signal

transduction in human peripheral blood mononuclear cells.

J. Biol. Chem. 2004, 279, 49206–49213.

[27] Turner-Brannen, E., Choi, K. G., Arsenault, R., El-Gagalwy,

H. et al., Inflammatory cytokines IL-32 and IL-17 have

common signaling intermediates despite differential

dependence on TNF-receptor 1. J. Immunol. 2011, 186,

7127–7135.

[28] Booth, J. S., Arsenault, R., Napper, S., Griebel, P. J. et al.,

TLR9 signaling failure renders Peyer’s patch regulatory B

cells unresponsive to stimulation with CpG oligodeoxy-

nucleotides. J. Innate Immun. 2010, 5, 483–494.

[29] Knight, Z. A., Lin, H., Shokat, K. M., Targeting the cancer

kinome through polypharmacology. Nat. Rev. Cancer. 2010,

10, 130–137.



[30] Sikkema, A. H., Diks, S. H., den Dunnen, W. F. A., ter Elst, A.

et al., Kinome profiling in pediatric brain tumors as a new

approach for target discovery. Cancer Res. 2009, 69,

5987–5995.

[31] ter Elst, A., Diks, S. H., Kampen, K. R., Hoogerbrugge, P. M.

et al., Identification of new possible targets for leukemia

treatment by kinase activity profiling. Leukemia Lymphoma

2011, 52, 122–130.

[32] Schrage, Y. M., Briaire-de Bruijn, I. H., de Miranda, N. F. C.

C., van Oosterwijk, J. et al., Kinome profiling of chon-

drosarcoma reveals SRC-pathway activity and dasatinib as

option for treatment. Cancer Res. 2009, 69, 6216–6222.

[33] Tuynman, J. B., Vermeulen, L., Boon, E. M., Kemper, K.

et al., Cyclooxygenase-2 inhibition inhibits c-Met kinase

activity and Wnt activity in colon cancer. Cancer Res. 2008,

68(4),1213–1220.

[34] Van den Broke, A., Arsenault, R., Cleuter, Y., Dehouck, C.,

Kinome profiling of bovine leukemia virus-indiced ovine

leukemia: an approach for identifying altered signaling

pathways and drugable targets in cancer. JAIDS 2011, 56,

98.

[35] Ritsema, T., Joore, J., van Workum, W., Pieterse, C. J.M.,

Kinome profiling of Arabidopsis using arrays of kinase

consensus substrates. Plant Methods 2007, 3, 3.

[36] Caenepeel, S., Charydczak, G., Sudarsanam, S., Hunter, T.,

Manning, G., The mouse kinome: discovery and compara-

tive genetics of all mouse protein kinases. Proc. Natl. Acad.

Sci. USA 2004, 32, 11707-11712.

[37] Gerdts, V., van den Hurk, S., Griebel, P., Babiuk, L., Use of

animal models in the development of human vaccines.

Future Med. 2007, 2, 667–675.

[38] Wilkie, B. N., Mallard, B. A., Genetic effects on vaccination.

Adv. Vet. Med. 1999, 41, 39–51.

[39] Wilson, K. T., Crabtree, J. E., Immunology of Helicobacter

pylori: insights into the failure of the immune response and

perspectives on vaccine studies. Gastroenterology 2007,

133, 288–308.

[40] He, S., Yang, Z., Skogerbo, G., Ren, F. et al., The properties

and functions of virus encoded microRNA, siRNA, and other

small noncoding RNAs. CRC Crit. Rev. Microbiol. 2008, 34,

175–188.

[41] Hilhorst, R., Houkes, L., van den Berg, A., Ruijtenbeek, R.,

Peptide microarrays for detailed, high-throughput substrate

identification, kinetic characterization, and inhibition

studies on protein kinase A. Anal. Biochem. 2009, 387,

150–161.

[42] Koppen, A., Houtman, R., Pijnenburg, D., Jeninga, E. H.

et al., Nuclear receptor coregulator interaction profiling

identifies TRIP3 as a novel peroxisome proliferator-

activated receptor y cofactor. Mol. Cell. Proteomics 2009, 8,

2212–2226.

[43] Poot, A. J., van Ameijde, J., Slijper, M., van den Berg, A.

et al., Development of selective bisubstrate-based inhibitors

against protein kinase C (PKC) isozymes by using dynamic

peptide microarrays. ChemBioChem 2009, 10, 2042–2051.

[44] Sylvester, J. E., Kron, S. J., A bead-based activity screen for

small-molecule inhibitors of signal transduction in chronic

myelogenous leukemia cells. Mol. Cancer Ther. 2010, 9,

1469–1481.

[45] Wu, D., Sylvester, J. E., Parker, L. L., Zhou, G., Kron, S. J.,

Peptide reporters of kinase activity in whole cell lysates.

Biopolymers 2010, 94, 475–486.

[46] Zetterqvist, O., Ragnarsson, U., Humble, E., Berglund, L.,

Engstrom, L., The minimum substrate of cyclic AMP-

stimulated protein kinase, as studied by synthetic peptides

representing the phosphorylatable site of pyruvate kinase

(type L) of rat liver. Biochem. Biophys. Res. Commun. 1976,

70, 696–703.

[47] Kemp, B. E., Graves, D. J., Benjamini, E., Krebs, E. G., Role

of multiple basic residues in determining the substrate

specificity of cyclic AMP-dependent protein kinase. J. Biol.

Chem. 1977, 252, 4888–4894,

[48] Parikh, K., Peppelenbosch, M. P., Kinome profiling of clin-

ical cancer specimens. Cancer Res. 2010, 70, 2575.

[49] Benovic, J. L., Strasser, R. H., Caron, M. G., Lefkowitz, R. J.,

Beta-adrenergic receptor kinase: identification of a novel

protein kinase that phosphorylates the agonist-occupied

form of the receptor. Proc. Natl. Acad. Sci. USA 1986, 83,

2797–2801.

[50] Strasser, R. H., Benovic, J. L., Caron, M. G., Lefkowitz, R. J.,

Beta-agonist and prostaglandin EI-induced translocation of

the beta-adrenergic receptor kinase: evidence that the

kinase may act on multiple adenylate cyclase-coupled

receptors. Proc. Natl. Acad. Sci. USA 1986, 83, 6362–6366.

[51] Benovic, J. L., Mayor, F., Somers, R. L., Caron, M. G.,

Leftowitz, R. J., Light-dependent phosphorylation of

rhodopsin by beta-adrenergic receptor kinase. Nature 1986,

321, 869–872.

[52] Pouyssegur, J., Lenormand, P., Fidelity and spatio-temporal

control in MAP kinase (ERKs) signalling. Eur. J. Biochem.

2003, 270, 3291–3299.

[53] Perillan, P. R., Chen, M., Potts, E. A., Simard, J. M., Trans-

forming growth factor beta1 regulates Kir2.3 inward recti-

fier K1channels via phospholipase C and protein kinase C

gamma in reactive astrocytes from adult rat brain. J. Biol.

Chem. 2001, 227, 1974–1980.

[54] Majumder, P. K., Mishra, N. C., Sun, X., Bharti, A. et al.,

Targeting of protein kinase C gamma to mitochondria in the

oxidative stress response. Cell Growth Differ. 2001, 12,

465–470.

[55] O’Flaherty, J. T., Chadwell, B. A., Kearns, M. W., Sergeant,

S., Daniel, L. W., Protein kinases C translocation responses

to low concentrations of arachidonic acid. J. Biol. Chem.

2001, 276, 24743–24750.

[56] Skalhegg, B. S., Tasken, K., Specificity in the cAMP/PKA

signalling pathway. Differential expression, regulation, and

subcellular localization of subunits of PKA. Front Biosci.

2000, 5, d679–d693.

[57] Alber, T., Signaling mechanisms of the Mycobacterium

tuberculosis receptor Ser/Thr protein kinases. Curr. Opin.

Struct. Biol. 2009, 6, 650–657.

[58] Schreiber, M., Res, I., Matter, A., Protein kinases as

antibacterial targets. Curr. Opin. Cell Biol. 2009, 2,

325–330.

Proteomics 2011, 11, 4595–4609 4609


peptide arrays for kinome analysis: new opportunities and remaining challenges

Documents