Content uploaded by Justin Lee
Author content
All content in this area was uploaded by Justin Lee on Aug 01, 2014
Content may be subject to copyright.
High Throughput Identification of Potential
Arabidopsis Mitogen-activated Protein
Kinases Substrates*□
S
Tanja Feilner‡§, Claus Hultschig‡, Justin Lee¶, Svenja Meyer储, Richard G. H. Immink**,
Andrea Koenig‡, Alexandra Possling‡ ‡‡, Harald Seitz‡, Allan Beveridge‡§,
Dierk Scheel¶, Dolores J. Cahill‡§, Hans Lehrach‡, Ju¨ rgen Kreutzberger‡,
and Birgit Kerstenत
Mitogen-activated protein kinase (MAPK) cascades are
universal and highly conserved signal transduction mod-
ules in eucaryotes, including plants. These protein phos-
phorylation cascades link extracellular stimuli to a wide
range of cellular responses. However, the underlying
mechanisms are so far unknown as information about
phosphorylation substrates of plant MAPKs is lacking. In
this study we addressed the challenging task of identify-
ing potential substrates for Arabidopsis thaliana mitogen-
activated protein kinases MPK3 and MPK6, which are
activated by many environmental stress factors. For this
purpose, we developed a novel protein microarray-based
proteomic method allowing high throughput study of pro-
tein phosphorylation. We generated protein microarrays
including 1,690 Arabidopsis proteins, which were ob-
tained from the expression of an almost nonredundant
uniclone set derived from an inflorescence meristem
cDNA expression library. Microarrays were incubated
with MAPKs in the presence of radioactive ATP. Using a
threshold-based quantification method to evaluate the
microarray results, we were able to identify 48 potential
substrates of MPK3 and 39 of MPK6. 26 of them are
common for both kinases. One of the identified MPK6
substrates, 1-aminocyclopropane-1-carboxylic acid syn-
thase-6, was just recently shown as the first plant MAPK
substrate in vivo, demonstrating the potential of our
method to identify substrates with physiological rele-
vance. Furthermore we revealed transcription factors,
transcription regulators, splicing factors, receptors, his-
tones, and others as candidate substrates indicating that
regulation in response to MAPK signaling is very complex
and not restricted to the transcriptional level. Nearly all of
the 48 potential MPK3 substrates were confirmed by other
in vitro methods. As a whole, our approach makes it
possible to shortlist candidate substrates of mitogen-ac-
tivated protein kinases as well as those of other protein
kinases for further analysis. Follow-up in vivo experiments
are essential to evaluate their physiological relevance.
Molecular & Cellular Proteomics 4:1558 –1568, 2005.
Upon completion of the annotation of several genomes,
including the Arabidopsis genome (1), proteomics will
greatly contribute to our understanding of gene function by
systematic high throughput protein investigation (2–5). Pro-
tein microarrays have become a crucial tool in this field
because they allow parallel, fast, and easy analysis of up to
thousands of addressable proteins, which are printed in a
systematic order with high density on coated glass slides
(6 –11). Analytical protein microarrays consisting of antibod-
ies or protein antigens are increasingly used to profile pro-
teins or antibodies, respectively, in crude protein samples of
interest (6, 12). Using functional protein microarrays, pro-
teins can be directly screened in vitro for a large variety of
activities, including protein-protein (13, 14), protein-DNA
(15, 16), protein-lipid (13), and protein-drug interactions (17,
18), under a wide range of different conditions. Furthermore
previous studies have demonstrated the suitability of pro-
tein microarrays to study protein phosphorylation by ki-
nases, however, only in a low or medium throughput manner
(17, 19, 20). Zhu et al. (19) e.g. used protein microarrays
bearing microwells in which 17 different protein substrates
were covalently immobilized to analyze protein kinases from
Saccharomyces cerevisiae.
In our group, we have developed the first plant (20, 21)
and bacterial microarrays (22) as well as a nonredundant
human protein microarray (23) for analytical and functional
studies. From this technical background, we aimed, in this
study, to develop a protein microarray-based method for high
throughput identification of potential protein kinase substrates
From the ‡Department of Vertebrate Genomics, Max Planck Insti-
tute for Molecular Genetics (MPI-MG), Ihnestr. 73, 14195 Berlin, Ger-
many, ¶Department of Stress and Developmental Biology, Leibniz
Institute of Plant Biochemistry, Weinberg 3, 06120 Halle, Germany,
储RZPD German Resource Centre for Genome Research GmbH,
Heubnerweg 6, 14059 Berlin, Germany, gabi.rzpd.de/, and **Business
Unit Bioscience, Plant Research International, P. O. Box 16, 6700 AA
Wageningen, The Netherlands
Received, January 6, 2005, and in revised form, June 1, 2005
Published, MCP Papers in Press, July 11, 2005, DOI 10.1074/
mcp.M500007-MCP200
Research
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc.1558 Molecular & Cellular Proteomics 4.10
This paper is available on line at http://www.mcponline.org
and to apply this method to reveal novel substrates of Arabi-
dopsis MAPKs.
1
MAPKs are the terminal components of the “three-kinase”
modules of MAPK cascades. These kinases are activated by
phosphorylation: a MAPK kinase kinase (MAPKKK), which is a
serine/threonine protein kinase, phosphorylates the subse-
quent dual specific MAPK kinase (MAPKK), which in turn
activates the MAPK by phosphorylation of a threonine as well
as tyrosine residue in the “activation loop” (24). Both residues
are separated by one amino acid in this loop (Thr-X-Tyr) (25).
Downstream of activated MAPKs, which are described as
serine/threonine kinases, phosphorylation events occur and
may influence the regulation of genes (26). However, the
phosphorylation substrates of the activated MAPKs are, es-
pecially in plants, widely unexplored.
MAPK cascades are universal and highly conserved signal
transduction modules in eucaryotes, including yeasts, ani-
mals, and plants, and mediate the intracellular transmission
and amplification of extracellular stimuli, resulting in the in-
duction of appropriate biochemical and physiological cellular
responses (24, 27, 28). Activation of MAPK cascades is an
important mechanism for stress adaptation by control of gene
expression. In plants, MAPK signaling has been implicated in
abiotic as well as biotic stress situations and is associated
with various physiological, developmental, and hormonal re-
sponses (29, 30). In certain plant species, this activation is an
important component in host and non-host resistance against
several pathogens and exhibits strong similarity to the innate
immune protection systems of mammals and Drosophila (24,
31). Roles for MAPK activation in triggering “early” defense
gene expression have been demonstrated in Arabidopsis (32),
parsley (33), and tobacco (25) in response to pathogens or
pathogen-derived elicitors. In Arabidopsis, a complete signal-
ing cascade following perception of a bacterial flagellin has
been elucidated recently (32). Downstream of the flagellin
receptor, a leucine-rich repeat (LRR) receptor kinase, the
cascade consists of MEKK1 (MAPKKK), MKK4/MKK5 (MAP-
KKs), and MPK3/6 (MAPKs). Signaling via this cascade results
in the up-regulation of WRKY22/WRKY29 transcription factor
gene expression. Despite this effort, nothing is known about
the phosphorylation events and their influence on gene ex-
pression downstream of activated MAPKs. Therefore, in this
study we addressed the challenging task of identifying the
protein substrates of two activated Arabidopsis MAPKs
(MPK3 and MPK6) using a novel protein microarray technique
as a powerful high throughput platform.
EXPERIMENTAL PROCEDURES
Plant Material and cDNA Library Generation—Arabidopsis thaliana
plants (ecotype Columbia-0 (Col0)) were grown under standard
greenhouse conditions (22 °C and 16 h of light). Shortly after bolting,
tops of young inflorescences containing the inflorescence meristems,
floral meristems, and young closed floral buds were harvested. Total
RNA was isolated from this material as has been described by Ver-
woerd et al. (34) followed by mRNA purification with oligo(dT)-cellu-
lose columns according to the manufacturer’s instructions (Amer-
sham Biosciences).
The SuperScript
TM
plasmid system for cDNA synthesis and cloning
(Invitrogen) was used for cDNA library construction following the
manufacturer’s protocol. The cDNA fragments were directionally li-
gated into the SalI/NotI cloning sites of the vector pQE30NASTattB
(GenBank
TM
accession number AY386205) and transformed into
competent Escherichia coli SCS1/pSE11 cells (35). Transformants
were picked and kept in 384-well microtiter plates.
Selection and Rearray of Putative Expression Clones—To select
putative expression clones, protein filters were generated from the
arrayed cDNA library and screened with an anti-RGS-His
6
antibody
(Qiagen, Hilden, Germany) as described in detail previously (20, 35).
Sequence Analysis and Selection of Uniclones—cDNAs were se-
quenced from the 5⬘-end by AGOWA GmbH (Berlin, Germany) using
the pQE65 vector primer. Raw sequence trace files were passed
through the PHRED base calling program (36, 37), and sequence
bases with confidence values ⬍15 were masked. Cloning vector
sequence removal was performed using pregap4 (Staden Package
software, version 2001 (38)). Sequences with a minimum of one of the
following characteristics were eliminated: (a) shorter than 100 nucle-
otides after vector and quality clipping and (b) sequences highly
similar to E. coli DNA (e-value ⬎e
⫺20
).
The remaining sequences were (i) analyzed by BLASTX 2.2.6 (39)
with the “MIPS Arabidopsis Protein Database” as a reference (Munich
Information Center for Protein Sequences, Munich, Germany,
mips.gsf.de/), (ii) translated to the corresponding peptide sequences
in the first forward frame using Emboss Transeq (40), and (iii) clus-
tered using d2_cluster with default parameters (41) to screen for
redundancy of the clone set. Using d2_cluster software sequences
with a minimal overlap of 100 bp and at least 90% sequence identity
are placed in the same cluster. Before clustering, the sequences were
screened for interspersed repeats and low complexity DNA se-
quences using CrossMatch and RepBase database for repetitive
Arabidopsis DNA sequence elements (42). The cDNA sequences were
clustered together with the complete annotated Arabidopsis gene
sequences lacking intron sequences and 5⬘- and 3⬘-untranslated
sequences.
The Emboss Transeq output was screened for clones having a stop
codon within the first 70 triplet codes, and the putative reading frames
and the existence of 5⬘-UTRs were calculated from the BLASTX
reports using scripts written in Perl (www.perl.com/). Clones identified
as singletons by cluster analysis and at least one representative for
each cluster were accepted for the uniclone set. For most of the
clusters one representative each was selected. Selection criteria
were: (i) full-length clone if available, (ii) smallest 5⬘-UTRs, and (iii)
highest similarity to the corresponding gene. More than one sequence
per cluster was selected for clusters that show a discrepancy be-
tween cluster and BLAST analysis, i.e. sequences composing one
cluster with Arabidopsis Genome Initiative (AGI) gene codes, but
1
The abbreviations used are: MAPK, mitogen-activated protein
kinase; ACS, 1-aminocyclopropane-1-carboxylic acid synthase; AGI,
Arabidopsis Genome Initiative; ATM1, A. thaliana meristem 1; CBB,
Coomassie Brilliant Blue; Cy3, indocarbocyanine; FAST, fluorescence
array surface technology; LRR, leucine-rich repeat; LR reaction, re-
combination reaction in the GATEWAY cloning system (Invitrogen) to
create expression clones, recombination of attL and attR sites;
MAPKK, mitogen-activated protein kinase kinase; MAPKKK, mito-
gen-activated protein kinase kinase kinase; MBP, myelin basic pro-
tein; Ni-NTA, nickel nitriloacetic acid; PKA, protein kinase A; ERK,
extracellular signal-regulated kinase; MEK, mitogen-activated protein
kinase/extracellular signal-regulated kinase kinase; MKK, MAPK ki-
nase; MEKK, MAPK/ERK kinase kinase.
High Throughput Identification of Arabidopsis MAPK Substrates
Molecular & Cellular Proteomics 4.10 1559
BLAST found the highest similarity to an Arabidopsis gene not in the
same cluster.
Recombinational Cloning in 96-well Format—cDNA inserts were
transferred from GATEWAY
TM
entry vectors into E. coli destination
vector pQE30NASTDV by LR reaction (attL ⫻attR recombination
reaction in the GATEWAY cloning technology; for further details about
this reaction please see the Invitrogen website). pQE-30NAST-DV is a
derivative of pQE-30NAST (GenBank
TM
accession number AY386205
(pQE-30NAST-attB)) that has been modified to a GATEWAY destina-
tion vector by Invitrogen. LR reactions were carried out in 96-well
format according to the manufacturer’s instructions with the excep-
tion that total volume and enzyme concentration were reduced by
half. Transformation of E. coli SCS1/pSE111 cells with 1
l LR reac-
tion mixture was performed by heat shock in 96-well format using 30
l of cell suspension/well. Two clones per transformation were
screened with colony PCR using vector primers. Recombinant clones
were grown in 96-well microtiter plates. After adding glycerol to an
end concentration of 20% (v/v), clones were stored at ⫺80 °C.
Protein Expression and Purification in 96-well Format—Proteins
were expressed in 1-ml cultures and purified (via metal chelate affinity
chromatography) after lysis in denaturing lysis buffer (100 mM
NaH
2
PO
4
,10mMTris-HCl, 6 Mguanidine hydrochloride, pH 8) as
described previously (20). Purified proteins were analyzed visually for
purity using a 15% polyacrylamide gel. Protein concentrations were
determined by Bradford assay (43).
Purification and Activation of MPK3 and MPK6 —The ORFs of
MPK3 (At3g45640) and MPK6 (At2g43790) were amplified from first
strand cDNA with modified gene-specific primers and cloned into
pGEX4T-1 as BamHI/XhoI or BamHI/NotI fragments, respectively.
Recombinant proteins were expressed in E. coli cells and purified by
glutathione-agarose chromatography according to the instructions of
the supplier (Sigma-Aldrich Chemie). The purified proteins were dia-
lyzed against two changes of water (4 °C, overnight). Myelin basic
protein (MBP; Sigma-Aldrich Chemie) was used as an artificial sub-
strate to evaluate the activity of the recombinant proteins. The inclu-
sion of 1 mMMnCl
2
in addition to 5 mMMgCl
2
in microarray-based
kinase assays (see reaction details below) was found to autoactivate
the recombinant MPK3 to activity levels comparable to those ex-
tracted from elicited plant material.
Generation of Protein Microarrays—The purified His-tagged pro-
teins were positioned on FAST
TM
slides (Schleicher & Schuell; Batch
Numbers AOBZ001, AOBZ033, AOBZ035, and AOBZ645) alongside
with positive and negative controls in an 11 ⫻11 spotting pattern in
two identical fields at a relative humidity of 75% and at 20 °C. The
protein-containing plates were cooled to 8 °C during arraying. The
proteins were printed with an in-house modified and extended Ge-
netix QArray microarrayer (Genetix LTD, New Milton, UK) equipped
with an optimized 4 ⫻4 print head of Genetix stainless steel solid pins
(X2777, tip diameter of 150
m). Pins were inked in 20
l of protein
solution for 50 ms and stamped at one position once for 100 ms using
a soft touch and a spotting distance of 1 mm. After every tenth
transfer and before addressing a new inking position the solid pins
were washed twice for 1 s with deionized water, washed once for 3 s
with 80% technical ethanol (20% bidistilled water), and dried for 3 s
with oil-free, pressured air supplied at 1 bar. A mixture of MBP (end
concentration of 2,000 ng/
l; Sigma-Aldrich Chemie) and mouse
anti-RGS-His
6
antibody (end dilution of 1:10; Qiagen) in deionized
water were chosen as positive controls for the phosphorylation ex-
periments as well as for the detection of the proteins on the array
surface. These positive controls were arranged as guide dots six
times in each block of the array in an asymmetric fashion (schematic
representation of the guide dots is given in Fig. 1) to assure correct
orientation of each image for analysis and to justify the grids used for
quantification. All purified proteins were spotted once in each of the
identical fields in total twice on each array. In addition one spot
position in each block was not addressed. Further controls (deionized
water, PBS, 20 pmol/
l BSA, mouse anti-RGS-His
6
antibody (Qiagen)
diluted 1:10 in PBS, and rabbit anti-mouse IgG3-Cy3 (Dianova, Ham-
burg, Germany) diluted 1:25 in PBS) were spotted three times per field
(six times on each array). After completion of the spotting run, a
Genepix array list file, describing the positioning of all samples on the
microarray, was generated.
Immunoscreening of Protein Microarrays—The microarrays were
washed for1hatroom temperature with TBST (TBS, 0.1% (v/v)
Tween 20) and than blocked for1hatroom temperature with 2%
(w/v) BSA, TBST. Mouse anti-RGS-His
6
antibody (Qiagen) was diluted
1:2,000 in blocking solution and then applied onto the arrays for 1 h
at room temperature followed by two 10-min wash steps with TBST.
The microarrays were further incubated for1hatroom temperature
with the respective Cy3-labeled secondary antibody (rabbit anti-
mouse IgG conjugate; Dianova), which was applied with a 1:800
dilution in blocking solution. Then three wash steps of 30 min each
were performed in TBST. All antibody incubation steps were carried
out using a 200-
l volume underneath a coverslip in the dark. Mi-
croarrays were dried. Fluorescence signals were detected with an
Affymetrix 428 microarray scanner (MWG-Biotech AG, Ebersberg,
Germany) at 532 nm.
Microarray-based Kinase Assay—Protein microarrays, which were
spotted the previous day, were washed in TBST for1hatroom
temperature with vigorous shaking to remove urea from the microar-
rays. Microarrays were blocked for1hatroom temperature with 2%
(w/v) BSA, TBST. All kinase incubations were then performed in the
presence of [
␥
-
33
P]ATP (25
Ci/ml; Amersham Biosciences). Protein
kinase A (PKA) reactions were carried out with 12.5
g/ml PKA from
mouse (New England Biolabs GmbH, Schwalbach, Germany) in PKA
buffer (50 mMTris-HCl, pH 7.5, 10 mMMgCl
2
) for 20 min. In the case
of MAPKs the microarrays were incubated with 100 ng/
l MPK3 or
200 ng/
l MPK6, respectively, in MAPK buffer (25 mMTris-HCl, pH
7.5, 1 mMEGTA, 1 mMDTT, 5 mMMgCl
2
,1mMMnCl
2
,20
MATP) for
30 min. After incubation the microarrays were washed as follows:
twice for 15 min in 2⫻PBST (PBS, 0.1% (v/v) Tween 20), twice for 15
min in 1⫻PBST, once for 30 min in 1⫻PBST, once for 30 min in 0.5⫻
PBST, once for 30 min in 0.1⫻PBST, and once for 30 min in 0.1⫻
PBS.
Microarrays were dried and transferred into an x-ray cassette (Hy-
percassette, Amersham Biosciences). The enzymatic phosphoryla-
tion of the immobilized proteins was detected on the dried and
Saran-covered microarrays with BAS-SR0813 imaging screens (Fuji-
film). After exposition (6 –48 h, depending on the specific activity of
the kinase under investigation) the screens were read out with an
FLA-8,000 microarray scanner (Fujifilm) at 635 nm and a spatial
resolution of 10
m.
Evaluation of Radioactive Signals and Selection of Potential Tar-
gets—The resulting images of the screens were opened in Aida Array
Metrix version 3.45 (Raytest, Straubenhardt, Germany; www.ray-
test.de). Subsequently the Genepix array list file, describing the po-
sitions of the samples on the array, was imported. The illustrated grid
of the protein samples on the array was arranged on the images
according to the positions of guide dots. The spotting pattern and the
relative arrangement of identical fields were entered according to the
user’s manual of the software package. These settings were saved as
a template and used for analyzing all images. Subsequently the
automatic grid-positioning feature was used, and the intensity of all
spots was determined using a spot diameter of 320
m. The average
background of block spot was determined with the “mode of non-
spot” evaluation mode. According to the manual the spot diameter,
which is not considered for blockwise background correction, was
enlarged by 4
m. Subsequently the duplicate correlation of all spots
High Throughput Identification of Arabidopsis MAPK Substrates
1560 Molecular & Cellular Proteomics 4.10
was determined (Fig. 3). To avoid any artifacts resulting from the
precipitation of labeled kinases, only spots deviating less than 25%
from the average intensity of both spots (Fig. 3, region inside both
green outer lines) were considered in the subsequent analysis. Sig-
nals (spots) deviating more than 25% from the average signal inten-
sity were excluded from the subsequent analysis by marking them in
the Aida Array Metrix software package. The reproducibility of the
quantification of different microarrays was verified in Aida Array Com-
pare version 3.53. Spots showing a deviation of the background
corrected signals assigned to the same protein of more than a factor
of 2 between two arrays were not included in the target identification.
Finally only those spots exceeding the average signal intensity of the
background by at least 10 times the deviation of the background
signals were considered to identify potential targets of the kinases.
On-blot Phosphorylation Assay—Proteins were separated using
15% SDS-PAGE followed by blotting the proteins on PVDF mem-
branes and then Coomassie Brilliant Blue (CBB) staining of the gels.
On-blot phosphorylations were carried out using reaction conditions
as described for the microarray-based kinase assay in an appropriate
volume. As positive controls MBP in different dilution steps (1:2
dilution steps from 2,000 –125 ng/
l MBP) were used. For negative
controls, proteins that were not identified as potential substrates in
the microarray assay were used (proteins expressed from clones
311_A05, 312_H05, and 313_K06).
Testing of Protein Solubility in 96-well Format—Protein expression
and purification (via metal chelate affinity chromatography) were per-
formed manually as described by Bussow et al. (44) with the excep-
tion that 50 mMTris, pH 8.0 was substituted by 50 mMHepes, pH 8.0
for cell lysis under native conditions. Aliquots of the following samples
were collected for SDS-PAGE: (i) lysates, (ii) soluble cellular proteins
(supernatant after centrifugation of the lysates), and (iii) purified pro-
tein (eluates after protein purification).
Kinase Assay in Solution Using Refolded Proteins—Purification of
the potential targets for verification was performed as described
above (see “Protein Expression and Purification in 96-well Format”)
except for a scale-up of the bacteria culture volume to 5 ml and of
Ni-NTA-agarose (Qiagen) to 100
l. After the last wash step of dena-
turing purification, the pH of the solution was adjusted by adding 33
lof1MTris (pH 7.5) to the ⬃100
l of remaining wash buffer.
Consecutive dilutions by adding 130
l, 260
l, and 1 ml of refolding
buffer (10 mMTris, pH 7.5, containing 1 mMPMSF) were performed
within a period of 30 min. The urea concentration was thus reduced to
0.5 Min the final step. After centrifugation, the buffer was removed, 1
ml of refolding buffer was added, and the samples were left on ice for
another 2 h. 10
l of the refolded proteins still attached to the Ni-NTA
beads were then used for in-solution kinase assays as described
previously (33).
RESULTS
Construction and Rearray of an Ordered Arabidopsis cDNA
Expression Library—As a source of recombinant Arabidopsis
proteins, we constructed a cDNA library specific for the inflo-
rescence meristem of Arabidopsis in the E. coli expression
vector pQE30NASTattB (GenBank
TM
accession AY386205).
This vector allows isopropyl thiogalactoside-inducible expres-
sion of RGS-His
6
-tagged proteins. After transformation of
E. coli cells with the library, recombinant clones were arrayed
into 384-well microtiter plates. The library consists of 40,000
clones with an average insert size of 1.06 kb. The arrayed
library was gridded onto high density filter membranes and
screened with an anti-RGS-His
6
antibody for putative expres-
sion clones (4,999 clones), which afterward were rearrayed
robotically into a sublibrary named ATM1.
Sequence Analysis of the Library and Generation of a Uni-
clone Set—To create a uniclone set from this sublibrary, all of
the 4,999 clones were sequenced from the 5⬘-end. Sequence
data were passed through several analysis steps after vector
clipping. 4,398 sequences remained after elimination of low
quality sequences and sequences similar to E. coli. All se-
quence data can be accessed via GenBank
TM
by accession
numbers CK117511 to CK122014.
BLASTX analysis of the sequences revealed that 61.8% of
the clone inserts are in the frame of the His tag, which is
comparable to the rearrayed hEx1 library (55%) (23). Further-
more it was calculated that 39.4% of the clones from the
ATM1 library contain the full-length coding sequence. The
d2_cluster analysis computed 2,029 different sequence clus-
ters of which 32% are singletons, whereas 68% of the se-
quences are members of 637 different clusters. 85% of these
clusters had a size of less than six sequences. All sequence
reads were translated to amino acid sequences in the frame of
the His tag, and 1,442 clones were excluded from the uni-
clone set because a stop codon was detected within the first
70 triplet codes of the insert DNA. This ensured that only such
clones were used that express proteins with a size of at least
⬃8 kDa (23). 318 sequences were eliminated because their
insert was calculated to be in the wrong frame by BLASTX
analysis, and another 21 sequences were sorted out due to a
lack of similarity to any Arabidopsis gene.
Finally 2,615 sequences passed all the preceding steps.
987 sequences computed to be singletons were picked out
directly for the uniclone set. From the remaining 1,628 se-
quences belonging to 404 different clusters, 511 additional
sequences were selected (one sequence from 384 clusters,
and more than one from 20 clusters). These selections re-
sulted in a uniclone set of altogether 1,498 clones. Sequence
analysis results for the whole sequence dataset are provided
in Supplemental Table 1 or in Supplemental Table 2 for the
uniclone set.
Extension of the Uniclone Set—The uniclone set of 1,498
clones was extended by 192 full-length cDNA clones. 96 of
them were described previously (21). The other 96 clones are
transcription factor expression clones, which were generated
from GATEWAY entry clones by LR reaction in this study.
These 192 full-length clones together with the 1,498 uniclones
compose the extended uniclone set of 1,690 clones.
Phosphorylation Studies with MPK3, MPK6, and PKA Using
Arabidopsis Protein Microarrays—All 1,690 clones of the ex-
tended uniclone set were expressed in parallel in 96-well
format. Proteins were purified by nickel chelate affinity chro-
matography under denaturing conditions. To control the qual-
ity of the purification 96 randomly chosen proteins were sep-
arated using 15% SDS-PAGE followed by CBB staining. 83%
of the proteins were detected with a size range of 14 –50 kDa
(data not shown).
High Throughput Identification of Arabidopsis MAPK Substrates
Molecular & Cellular Proteomics 4.10 1561
To generate Arabidopsis protein microarrays, the 1,690 pu-
rified proteins of the extended uniclone set and controls were
arrayed on FAST slides. The proteins were spotted once in
each of the two identically fields. Each field consists of 16
blocks, and the proteins were arranged in an 11 ⫻11 spotting
pattern in each block (see Fig. 1). The microarrays were
screened with an anti-RGS-His
6
antibody (Fig. 2,aand b)to
detect recombinant proteins. Furthermore the microarrays
were used for the identification of phosphorylation substrates
of PKA (Fig. 2c), MPK3 (Fig. 2d), or MPK6 (Fig. 2e), each in the
presence of radioactive ATP. Mouse PKA was included as an
example of a kinase from a different family to validate the
specificity of our method. In addition to other controls, we
used a mixture of MBP and anti-RGS-His
6
antibody as posi-
tive control and guide dot for both immunoscreening and
phosphorylation studies. This control was spotted six times in
each block in an asymmetric fashion as detailed in Fig. 1. This
arrangement facilitates the grid adjustment for subsequent
quantification. One position of every block was not addressed
(Fig. 1).
For nearly all (95%) of the spotted proteins of the extended
uniclone set, a signal was detected in the immunoscreen (Fig.
2a). The incubation with the two MAPKs resulted in signals at
distinct positions including the MBP spotting positions (Fig. 2,
dand e). Whereas these MAPK images are very similar, incu-
bation with PKA resulted in a clearly different phosphorylation
pattern (Fig. 2c). No signals were detectable in control exper-
iments with radioactive ATP incubation without any kinase
(data not shown), thus excluding that the signals observed
were due to direct ATP binding or autophosphorylation of
proteins with kinase domains.
To select potential substrates of Arabidopsis MPK3 and
MPK6 based on a quantitative system, we performed two
independent experiments for both kinases including two mi-
croarrays each. The recombinant MPK6 was comparatively
less active, probably as a result of poor folding in the E. coli
expression system, and hence was used at a higher concen-
tration for the screen (see “Experimental Procedures”). Signal
intensities were determined for every microarray with a high
field-to-field correlation as demonstrated for one MPK3 mi-
croarray in Fig. 3. Only proteins that were identified as targets
in both experiments according to our quantitative threshold-
based criteria (see “Experimental Procedures”) were defined
as potential targets, resulting in a list of 48 or 39 for MPK3
(Table I) or MPK6 (Table II), respectively. 26 of them are
common for both kinases as indicated with grey background
in the tables. For comparison only, 35 substrates for mouse
PKA have been identified in one experiment with two microar-
rays, and only a very low number of these targets overlap with
the identified MAPK targets (three for MPK3, and four for
MPK6) (data not shown).
To verify the potential substrates by a simple in vitro
method, all 48 MPK3 substrates were purified, separated by
SDS-PAGE, transferred onto PVDF membranes, and then
incubated with MPK3 using the same assay conditions as in
the microarray-based kinase assay (on-blot phosphorylation).
Nearly all of the 48 potential MPK3 substrates with the ex-
ception of four samples (numbers 44, 45, 47, and 48 in Table
FIG.1.Spotting pattern of Arabidopsis protein microarrays. Two
identical fields were spotted per microarray. Each field consists of 16
blocks with an 11 ⫻11 pattern. Positions of the guide dots are
marked by filled circles. Guide dots (mixture of MBP (2,000 ng/
l) and
mouse anti-RGS-His
6
antibody (1:10)) were spotted in every block six
times in an asymmetric fashion to assure correct orientation of each
image for analysis and to arrange the grids used for quantification.
One position per block was not addressed as indicated by a hyphen.
FIG.2.Immunoscreening (aand b) and phosphorylation studies
(c– e)onArabidopsis protein microarrays. All proteins from the
extended Arabidopsis uniclone set were immobilized on FAST slides
and screened with an anti-RGS-His
6
antibody (a, whole microarray; b,
enlargement of blocks 3 and 7 highlighted in a). Phosphorylation
studies were performed on Arabidopsis protein microarrays with PKA
(c), MPK3 (d), and MPK6 (e). The section of the microarray shown in
bis given in c– e for the phosphorylation experiments using the same
dynamic range (0 –40,000) for all kinases. The PKA microarray (c) was
exposed for 12 h, the MPK3 microarray (d) was exposed for 6 h, and
the MPK6 microarray (e) was exposed for 48 h to the imaging screen.
High Throughput Identification of Arabidopsis MAPK Substrates
1562 Molecular & Cellular Proteomics 4.10
I) were phosphorylated at the membrane, whereas the nega-
tive controls were not detectable (data not shown). To further
verify the targets in an independent in vitro assay using “na-
tive” proteins we first checked the solubility of all MPK3 and
MPK6 substrate candidates in 96-well format. The tested
proteins were not or only poorly soluble after expression in
E. coli and lysis with lysozyme and therefore could not be
purified under native conditions in our experiment (data not
shown). We next attempted to purify the proteins, refold them
while still attached via the polyhistidine tag to Ni-NTA beads,
and subject them to in vitro kinase assays in solution. This
approach is analogous to immunoprecipitation-coupled ki-
nase assays that we routinely perform (33, 45) except that, in
this case, the substrate and not the kinase is attached to the
beads. As seen in Fig. 4, almost all samples with the excep-
tion of six (numbers 15, 27, 28, 44, 47, and 48 in Table I) show
enhanced phosphorylation in the presence of MPK3.
DISCUSSION
Here a new proteomic method allowing high throughput
identification of potential phosphorylation substrates of pro-
tein kinases has been developed using protein microarray
technology. The application of this method for analyzing two
Arabidopsis mitogen-activated protein kinases (MPK3 and
MPK6) led to the selection of several novel substrate candi-
dates of a set of ⬃1,700 purified Arabidopsis proteins. The
results deliver new insights into MAPK downstream signaling.
The general feasibility of protein microarrays to detect pro-
tein phosphorylation by kinases has been demonstrated pre-
viously using different surfaces for protein immobilization:
BSA-N-hydroxysuccinimide monolayers to which substrates
were covalently attached (17), protein microarrays bearing
polydimethylsiloxane-coated microwells (19), or FAST slides,
which are covered with a nitrocellulose-derived polymer for
non-covalent protein attachment (20). However, in these pre-
vious studies, substrate identification was carried out only in
a low/medium throughput manner with regard to the number
of proteins analyzed for phosphorylation on one microarray
(17, 19, 20) and for most instances without quantification of
the microarray results (17, 20). Therefore, in this study we
aimed to develop a high throughput method. Furthermore the
method should allow the identification of substrates based on
quantitative criteria. The combination of an 11 ⫻11 spotting
pattern with a microarray design using duplicates of every
protein on one array (Fig. 1) allows us to immobilize theoret-
ically 1,936 different samples in two identical fields. With this
we have developed a high density protein microarray-based
kinase assay. Compared with other currently available mi-
croarray methods (17, 19, 20) our assay provides the highest
number of proteins that may be analyzed for phosphorylation
in parallel. The efficiency of protein expression and purifica-
tion methods as well as of the technology used for protein
transfer to the microarray surface has been demonstrated by
identifying nearly all of the recombinant proteins after screen-
ing the microarrays with an anti-RGS-His
6
antibody (Fig. 2a).
To detect the phosphorylation of immobilized proteins ra-
dioactively labeled ATP has been used. Initial studies were
performed in which fluorescent dyes were applied for the
identification of phosphorylated amino acids (46). Neverthe-
less the radioactive-based detection is still the most sensi-
tive and robust detection method. Using phosphorimaging
we were able to yield radioactive signals in the 11 ⫻11
pattern with sufficient resolution and a high field-to-field
reproducibility (Fig. 3).
To manage substrate identification with a high significance,
we performed two independent phosphorylation experiments
for every kinase with two microarrays each. Only proteins that
were detected in both experiments, taking into account the
threshold-based quantitative criteria, were defined as poten-
tial targets thus reducing the number of false-positive results.
We identified 48 potential substrates for MPK3 (Table I) and
39 for MPK6 (Table II). As expected, a large number of them
(26 substrates) are common for both kinases. The lower num-
ber of MPK6 compared with MPK3 substrates probably re-
flects a lower representation of the MPK6 substrates in our set
or a lower specific activity of the recombinant MPK6 used in
the screen. To improve the significance of the method it would
be recommendable for future studies to normalize the radio-
FIG.3. Duplicate correlation of signals of spotted proteins
phosphorylated with MPK3 on one microarray. Signal intensities of
identical spots were correlated using the program Aida Array Metrix.
Signals of the phosphorylated spots in field 2 (y axis) are plotted
against the signal intensities of the corresponding proteins in field 1 (x
axis). Each spot of the field-to-field correlation represents the signal
intensity of one protein in these fields. Only duplicates deviating less
than 25% from the average intensity of both spots were considered
for subsequent analysis (region inside of both green outer lines). The
red line shows the ideal distribution of theoretical 100% reproducible
replicates. The regression line of experimental replicates is given in
green (inner line). PSL, photo-stimulated luminescence; r, regression
coefficient.
High Throughput Identification of Arabidopsis MAPK Substrates
Molecular & Cellular Proteomics 4.10 1563
active signals with respect to the concentrations of the differ-
ent transferred proteins.
We demonstrated the suitability of our quantitative criteria
for substrate selection exemplarily for one kinase (MPK3) by
verification of nearly all MPK3 substrates using two different in
vitro methods. 92% of the targets were confirmed with on-
blot phosphorylation (data not shown), and 88% of the sub-
strate candidates were verified by independent in vitro phos-
phorylation of refolded proteins in solution (Fig. 4). The failure
to confirm some of the potential substrates may be due to low
amounts of purified proteins or because of the loss of a
potential linear phosphorylation site by refolding of the protein
in the assay in solution.
The high specificity of the method has been proven com-
paring the phosphorylation pattern of kinases from the same
as well as from different families (Fig. 2). As expected, kinases
from different families yield clearly different patterns, resulting
in a low number of common substrates that we identified for
TABLE I
Potential MPK3 substrates identified in this study
Proteins with gray background have been identified as substrates for both MPK3 and MPK6. For substrates, which are marked in bold, or
for their mammalian homologues, the involvement in signaling pathways has been described as reviewed by Yang et al. (28) and detailed in
the text for some of them. Detailed clone information is available in supplemental Table 2. Nr., consecutive clone numbers; MIPS, Munich
Information Centre for Protein Sequences; TIGR, The Institute for Genomic Research.
High Throughput Identification of Arabidopsis MAPK Substrates
1564 Molecular & Cellular Proteomics 4.10
these kinases (e.g. three common substrates for MPK3 and
PKA). The relatively high number of mouse PKA substrates
identified within our set of Arabidopsis proteins (35 sub-
strates, data not shown) may be due to the fact that several
plant protein kinases belong to the same group as PKA, the
AGC group (named after PKA, PKG, and PKC; see www.nih.
go.jp/mirror/Kinases/pkr/pk_catalytic/pk_hanks_seq_align_
long.html), and are sharing specific sequence motifs such as
the FXXF hydrophobic motif in the C terminus (47).
The method described here is a valuable supplementation
in the spectrum of existing in vitro methods for substrate
identification such as solid-phase phosphorylation screening
of
phage cDNA expression libraries described by Fukunaga
and Hunter (48) or the method developed by Shokat and
colleagues (49 –51). The latter method uses kinases with mod-
ified ATP binding pockets that are accepting unnatural ATP
analogues to display direct substrates of modified kinases in
crude mixtures or cell lysates. In contrast to this method, no
modification of the kinases is needed for our microarray-
based screening. Compared with the screening of phage
expression libraries, our method needs considerably smaller
volumes of active kinase (200
l for one incubation are suffi-
TABLE II
Potential MPK6 substrates identified in this study
See legend of Table I.
High Throughput Identification of Arabidopsis MAPK Substrates
Molecular & Cellular Proteomics 4.10 1565
cient using our method; this is 1,000 times less than for the
screening of phage expression libraries (52)). The screening of
proteins obtained from characterized arrayed clones in our
study allows direct identification of substrate candidates be-
cause a positive result is directly linked to the sequence of the
respective expression clone. Further verification experiments
may be performed with proteins expressed from the same
clone used for the screening experiments.
In general, our assay represents a rapid in vitro screening
for potential substrates of protein kinases. We may obtain
some false-positives because we use denatured proteins as
targets. In the respective native-folded protein potential phos-
phorylation sites are possibly not accessible for the kinase.
However, the confirmation experiment with the refolded pro-
teins points to a low number of false-positives due to this
reason. Furthermore it is possible that the substrate and the
kinase can interact but will never associate in vivo,e.g. be-
cause they are localized in different cellular compartments.
For that reason, potential substrates identified with our
method have to be verified in subsequent in vivo experiments;
e.g. protein-protein interaction studies in planta by innovative
microspectroscopic approaches such as fluorescence reso-
nance energy transfer and fluorescence lifetime imaging mi-
croscopy (53). To gain insights into the sites of phosphoryla-
tion, antibodies against phosphorylated protein epitopes may
be used for detection (54) on protein microarrays. Further-
more peptide arrays (54, 55) or MS-based methods (56, 57)
may be applied in this respect.
An excellent demonstration for the suitability of our method
to find substrates with biological relevance is the identification
of 1-aminocyclopropane-1-carboxylic acid synthase-6 (ACS-6)
as an MPK6 substrate. ACS, the rate-limiting enzyme of ethyl-
ene biosynthesis, has been described very recently as the first
plant MAPK substrate in vivo (58). Liu et al. (58) found that
selected isoforms of ACS including ACS-6, which we identified
in our study (Table II, number 16), are substrates of MPK6.
Phosphorylation of ACS led to the accumulation of the protein
and to ethylene production (58). Furthermore it has been
speculated that plant stress-responsive MAPKs may phos-
phorylate transcription factors or transcription factor-regulat-
ing proteins (24, 32) similar to their mammalian orthologs (28).
Several reports support this assumption, such as the in-
creased nuclear localization of MAPKs in parsley cells after
MAPK activation (45) or the up-regulation of WRKY22/
WRKY29 transcription factor expression upon activation of
the flagellin MAPK cascade (32). Asai et al. (32) suggested that
not the WRKY transcription factors themselves but a specific
WRKY inhibitor may be phosphorylated and inactivated by
MAPKs. In agreement with these findings, we identified sev-
eral transcription factors (Table I, numbers 3, 46, and 47;
Table II, numbers 2, 22, 32, and 39) as well as a transcription
regulator (Table II, number 14) as potential MAPK substrates
in this study. Furthermore the phosphorylation of histones
(Table I, numbers 6, 22, 28, 29, and 39; Table II, numbers 25
and 31) could be involved in regulating gene transcription as
it has been shown that the tail domain of histones regulates
chromatin structure and hence gene transcription (59).
Several substrates, which were identified in this study, or
their mammalian homologues have been described previously
to be involved in signal transduction (28) as indicated in bold
in Tables I and II. The identified LRR family protein (Table I,
number 13) may participate as a receptor in elicitor-induced
MAPK cascades due to its LRR domain (32). Flowering locus
T protein (Table I, number 27; Table II, number 23) is a putative
membrane-associated protein with homology to human phos-
phatidylethanolamine-binding protein (60). This protein is
identical to Raf kinase inhibitor protein that is involved in
regulation of RAF/MEK/ERK signaling pathway (61).
The calmodulin-binding family protein (Table I, number 17)
may be a component of the MAPK cascade because a calm-
FIG.4.Verification of potential MPK3 targets using refolded proteins. The 48 potential MPK3 targets were purified and refolded while
still attached to Ni-NTA beads. 10
l of the protein slurry were used for phosphorylation assays in the absence or presence (shown in every
second lane) of active recombinant MPK3, separated by SDS-PAGE, stained with CBB, and exposed for autoradiography (Autorad.). The
positions of the autophosphorylated MPK3 and an unspecific radioactive band across all samples are indicated on the left by an asterisk and
an empty circle, respectively. GST and MBP were used as the respective negative and positive controls for the reactions. Possible
contaminants from the Ni-NTA bead purification and refolding steps were also incubated with MBP (MBP⫹) to show the slight inhibitory effects
on the assay. M, prestained protein markers (MBI Fermentas, St. Leon-Rot, Germany).
High Throughput Identification of Arabidopsis MAPK Substrates
1566 Molecular & Cellular Proteomics 4.10
odulin-binding protein has been described previously as a
negative regulator for stress tolerance to sodium and osmotic
stress (62). A further target that has been identified for both
MPKs is a casein kinase (Table I, number 32; Table II, number
27). In HeLa cells, a direct interaction of p38 MAPK and casein
kinase 2 was observed, and also a stress-induced activation
of casein kinase 2 by this MAPK was shown (63). All these
results support the assumption that regulation in response to
MAPK signaling is very complex and not restricted to the
transcriptional level (28).
In conclusion, we established a powerful generic test sys-
tem for the in vitro identification of potential protein kinase
substrates by high density protein arrays followed by inde-
pendent verification with refolded proteins. The application of
this test system for plant MAPKs resulted in a short list of
candidates for further analysis. Follow-up experiments such
as in vivo verification and the mapping of phosphorylation
sites in substrates are essential to evaluate the physiological
relevance of the targets in MAPK signaling.
Acknowledgments—We thank Thomas Przewieslik and Thomas
Nitsche (both MPI-MG, Berlin) for technical assistance and Dr. Ralf
Stracke (Max Planck Institute for Plant Breeding Research, Cologne)
for providing 96 Arabidopsis entry clones. We are grateful to Jasmin
Bastian for reading the manuscript.
* This work was funded by the Federal Ministry of Education and
Research (Grants 0312274, 0312272, and 031U102D) and the Max
Planck Society. The costs of publication of this article were defrayed
in part by the payment of page charges. This article must therefore be
hereby marked “advertisement” in accordance with 18 U.S.C. Section
1734 solely to indicate this fact.
The nucleotide sequence(s) reported in the paper has been sub-
mitted to GenBank
TM
EBI Data Bank with accession number(s)
CK117511 to CK122014.
□SThe on-line version of this article (available at http://www.
mcponline.org) contains supplemental material.
§ Present address: Centre for Human Proteomics, Royal College of
Surgeons in Ireland, 121 St. Stephens Green, Dublin 2, Ireland.
‡‡ Present address: Free University of Berlin, Ko¨ nigin-Luise-Str.
12-16, 14195 Berlin, Germany.
§§ To whom correspondence should be addressed: Dept. Neuro-
proteomics, Max Delbru¨ ck Center for Molecular Medicine, Robert-
Ro¨ ssle-Str. 10, 13092 Berlin, Germany. Tel.: 49-30-9406-2636; Fax:
49-30-9406-2629; E-mail: b.kersten@mdc-berlin.de.
REFERENCES
1. The Arabidopsis Genome Initiative (2000) Analysis of the genome sequence
of the flowering plant Arabidopsis thaliana. Nature 408, 796 – 815
2. Phizicky, E., Bastiaens, P. I., Zhu, H., Snyder, M., and Fields, S. (2003)
Protein analysis on a proteomic scale. Nature 422, 208 –215
3. Kersten, B., Bu¨ rkle, L., Kuhn, E. J., Giavalisco, P., Konthur, Z., Lueking, A.,
Walter, G., Eickhoff, H., and Schneider, U. (2002) Large-scale plant
proteomics. Plant Mol. Biol. 48, 133–141
4. Kersten, B., Feilner, T., Angenendt, P., Giavalisco, P., Brenner, W., and
Bu¨ rkle, L. (2004) Proteomic approaches in plant biology. Curr. Proteom-
ics 1, 131–144
5. Zhu, H., Bilgin, M., and Snyder, M. (2003) Proteomics Annu. Rev. Biochem.
72, 783– 812
6. Zhu, H., and Snyder, M. (2003) Protein chip technology. Curr. Opin. Chem.
Biol. 7, 55– 63
7. Feilner, T., Kreutzberger, J., Niemann, B., Kramer, A., Possling, A., Seitz, H.,
and Kersten, B. (2004) Proteomic studies using microarrays. Curr. Pro-
teomics 1, 283–295
8. Labaer, J., and Ramachandran, N. (2005) Protein microarrays as tools for
functional proteomics. Curr. Opin. Chem. Biol. 9, 14 –19
9. Poetz, O., Schwenk, J. M., Kramer, S., Stoll, D., Templin, M. F., and Joos,
T. O. (2005) Protein microarrays: catching the proteome. Mech. Ageing
Dev. 126, 161–170
10. Glokler, J., and Angenendt, P. (2003) Protein and antibody microarray
technology. J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 797,
229 –240
11. Angenendt, P. (2005) Progress in protein and antibody microarray technol-
ogy. Drug Discov. Today 10, 503–511
12. Michaud, G. A., Salcius, M., Zhou, F., Bangham, R., Bonin, J., Guo, H.,
Snyder, M., Predki, P. F., and Schweitzer, B. I. (2003) Analyzing antibody
specificity with whole proteome microarrays. Nat. Biotechnol. 21,
1509 –1512
13. Zhu, H., Bilgin, M., Bangham, R., Hall, D., Casamayor, A., Bertone, P., Lan,
N., Jansen, R., Bidlingmaier, S., Houfek, T., Mitchell, T., Miller, P., Dean,
R. A., Gerstein, M., and Snyder, M. (2001) Global analysis of protein
activities using proteome chips. Science 293, 2101–2105
14. Ramachandran, N., Hainsworth, E., Bhullar, B., Eisenstein, S., Rosen, B.,
Lau, A. Y., Walter, J. C., and LaBaer, J. (2004) Self-assembling protein
microarrays. Science 305, 86 –90
15. Snapyan, M., Lecocq, M., Guevel, L., Arnaud, M. C., Ghochikyan, A., and
Sakanyan, V. (2003) Dissecting DNA-protein and protein-protein interac-
tions involved in bacterial transcriptional regulation by a sensitive protein
array method combining a near-infrared fluorescence detection. Pro-
teomics 3, 647– 657
16. Kersten, B., Possling, A., Blaesing, F., Mirgorodskaya, E., Gobom, J., and
Seitz, H. (2004) Protein microarray technology and ultraviolet crosslink-
ing combined with mass spectrometry for the analysis of protein-DNA
interactions. Anal. Biochem. 331, 303–313
17. MacBeath, G., and Schreiber, S. L. (2000) Printing proteins as microarrays
for high-throughput function determination. Science 289, 1760 –1763
18. Zhou, F. X., Bonin, J., and Predki, P. F. (2004) Development of functional
protein microarrays for drug discovery: progress and challenges. Comb.
Chem. High Throughput Screen. 7, 539 –546
19. Zhu, H., Klemic, J. F., Chang, S., Bertone, P., Casamayor, A., Klemic, K. G.,
Smith, D., Gerstein, M., Reed, M. A., and Snyder, M. (2000) Analysis of
yeast protein kinases using protein chips. Nat. Genet. 26, 283–289
20. Kramer, A., Feilner, T., Possling, A., Radchuk, V., Weschke, W., Bu¨ rkle, L.,
and Kersten, B. (2004) Identification of barley CK2
␣
targets by using the
protein microarray technology. Phytochemistry 65, 1777–1784
21. Kersten, B., Feilner, T., Kramer, A., Wehrmeyer, S., Possling, A., Witt, I.,
Zanor, M. I., Stracke, R., Luecking, A., Kreuzberger, J., Lehrach, H., and
Cahill, D. J. (2003) Generation of Arabidopsis protein chips for antibody
and serum screening. Plant Mol. Biol. 52, 999 –1010
22. Steller, S., Angenendt, P., Cahill, D. J., Heuberger, S., Lehrach, H., and
Kreutzberger, J. (2005) Bacterial protein microarrays for identification of
new potential diagnostic markers for Neisseria meningitidis infections.
Proteomics 5, 2048 –2065
23. Lueking, A., Possling, A., Huber, O., Beveridge, A., Horn, M., Eickhoff, H.,
Schuchardt, J., Lehrach, H., and Cahill, D. J. (2003) A nonredundant
human protein chip for antibody screening and serum profiling. Mol. Cell.
Proteomics 2, 1342–1349
24. MAPK Group (2002) Mitogen-activated protein kinase cascades in plants: a
new nomenclature. Trends Plant Sci. 7, 301–308
25. Yang, K. Y., Liu, Y., and Zhang, S. (2001) Activation of a mitogen-activated
protein kinase pathway is involved in disease resistance in tobacco.
Proc. Natl. Acad. Sci. U. S. A. 98, 741–746
26. Peck, S. C. (2003) Early phosphorylation events in biotic stress. Curr. Opin.
Plant Biol. 6, 334 –338
27. Agrawal, G. K., Iwahashi, H., and Rakwal, R. (2003) Rice MAPKs. Biochem.
Biophys. Res. Commun. 302, 171–180
28. Yang, S. H., Sharrocks, A. D., and Whitmarsh, A. J. (2003) Transcriptional
regulation by the MAP kinase signaling cascades. Gene (Amst.)320,
3–21
29. Zwerger, K., and Hirt, H. (2001) Recent advances in plant MAP kinase
signalling. Biol. Chem. 382, 1123–1131
30. Jonak, C., Okresz, L., Bogre, L., and Hirt, H. (2002) Complexity, cross talk
and integration of plant MAP kinase signalling. Curr. Opin. Plant Biol. 5,
High Throughput Identification of Arabidopsis MAPK Substrates
Molecular & Cellular Proteomics 4.10 1567
415– 424
31. Nu¨ rnberger, T., and Scheel, D. (2001) Signal transmission in the plant
immune response. Trends Plant Sci. 6, 372–379
32. Asai, T., Tena, G., Plotnikova, J., Willmann, M. R., Chiu, W. L., Gomez-
Gomez, L., Boller, T., Ausubel, F. M., and Sheen, J. (2002) MAP kinase
signalling cascade in Arabidopsis innate immunity. Nature 415, 977–983
33. Kroj, T., Rudd, J. J., Nurnberger, T., Gabler, Y., Lee, J., and Scheel, D.
(2003) Mitogen-activated protein kinases play an essential role in oxida-
tive burst-independent expression of pathogenesis-related genes in
parsley. J. Biol. Chem. 278, 2256 –2264
34. Verwoerd, T. C., Dekker, B. M., and Hoekema, A. (1989) A small-scale
procedure for the rapid isolation of plant RNAs. Nucleic Acids Res. 17,
2362
35. Bussow, K., Cahill, D., Nietfeld, W., Bancroft, D., Scherzinger, E., Lehrach,
H., and Walter, G. (1998) A method for global protein expression and
antibody screening on high-density filters of an arrayed cDNA library.
Nucleic Acids Res. 26, 5007–5008
36. Ewing, B., and Green, P. (1998) Base-calling of automated sequencer
traces using phred. II. Error probabilities. Genome Res. 8, 186 –194
37. Ewing, B., Hillier, L., Wendl, M. C., and Green, P. (1998) Base-calling of
automated sequencer traces using phred. I. Accuracy assessment. Ge-
nome Res. 8, 175–185
38. Staden, R., Beal, K. F., and Bonfield, J. K. (2000) The Staden package,
1998. Methods Mol. Biol. 132, 115–130
39. Altschul, S. F., Madden, T. L., Schaffer, A. A., Zhang, J., Zhang, Z., Miller,
W., and Lipman, D. J. (1997) Gapped BLAST and PSI-BLAST: a new
generation of protein database search programs. Nucleic Acids Res. 25,
3389 –3402
40. Mullan, L. J., and Bleasby, A. J. (2002) Short EMBOSS User Guide. Euro-
pean Molecular Biology Open Software Suite. Brief. Bioinform. 3, 92–94
41. Burke, J., Davison, D., and Hide, W. (1999) d2_cluster: a validated method
for clustering EST and full-length cDNA sequences. Genome Res. 9,
1135–1142
42. Jurka, J. (2000) Repbase update: a database and an electronic journal of
repetitive elements. Trends Genet. 16, 418 – 420
43. Bradford, M. M. (1976) A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye
binding. Anal. Biochem. 72, 248 –254
44. Bussow, K., Quedenau, C., Sievert, V., Tischer, J., Scheich, C., Seitz, H.,
Hieke, B., Niesen, F. H., Gotz, F., Harttig, U., and Lehrach, H. (2004) A
catalog of human cDNA expression clones and its application to struc-
tural genomics. Genome Biol. 5, R71
45. Lee, J., Rudd, J. J., Macioszek, V. K., and Scheel, D. (2004) Dynamic
changes in the localization of MAPK cascade components controlling
pathogenesis-related (PR) gene expression during innate immunity in
parsley. J. Biol. Chem. 279, 22440 –22448
46. Martin, K., Steinberg, T. H., Goodman, T., Schulenberg, B., Kilgore, J. A.,
Gee, K. R., Beechem, J. M., and Patton, W. F. (2003) Strategies and
solid-phase formats for the analysis of protein and peptide phosphoryl-
ation employing a novel fluorescent phosphorylation sensor dye. Comb.
Chem. High Throughput Screen. 6, 331–339
47. Bogre, L., Okresz, L., Henriques, R., and Anthony, R. G. (2003) Growth
signalling pathways in Arabidopsis and the AGC protein kinases. Trends
Plant Sci. 8, 424 – 431
48. Fukunaga, R., and Hunter, T. (1997) MNK1, a new MAP kinase-activated
protein kinase, isolated by a novel expression screening method for
identifying protein kinase substrates. EMBO J. 16, 1921–1933
49. Shah, K., Liu, Y., Deirmengian, C., and Shokat, K. M. (1997) Engineering
unnatural nucleotide specificity for Rous sarcoma virus tyrosine kinase to
uniquely label its direct substrates. Proc. Natl. Acad. Sci. U. S. A. 94,
3565–3570
50. Liu, Y., Shah, K., Yang, F., Witucki, L., and Shokat, K. M. (1998) Engineering
Src family protein kinases with unnatural nucleotide specificity. Chem.
Biol. 5, 91–101
51. Eblen, S. T., Kumar, N. V., Shah, K., Henderson, M. J., Watts, C. K., Shokat,
K. M., and Weber, M. J. (2003) Identification of novel ERK2 substrates
through use of an engineered kinase and ATP analogs. J. Biol. Chem.
278, 14926 –14935
52. Fukunaga, R., and Hunter, T. (2004) Methods in Molecular Biology: MAP
Kinase Signaling Protocols, Vol. 250, pp. 211–236, Humana Press Inc.,
Totowa, NJ
53. Hink, M. A., Bisselin, T., and Visser, A. J. (2002) Imaging protein-protein
interactions in living cells. Plant Mol. Biol. 50, 871– 883
54. Lesaicherre, M. L., Uttamchandani, M., Chen, G. Y., and Yao, S. Q. (2002)
Antibody-based fluorescence detection of kinase activity on a peptide
array. Bioorg. Med. Chem. Lett. 12, 2085–2088
55. Houseman, B. T., Huh, J. H., Kron, S. J., and Mrksich, M. (2002) Peptide
chips for the quantitative evaluation of protein kinase activity. Nat. Bio-
technol. 20, 270 –274
56. Glinski, M., Romeis, T., Witte, C. P., Wienkoop, S., and Weckwerth, W.
(2003) Stable isotope labeling of phosphopeptides for multiparallel ki-
nase target analysis and identification of phosphorylation sites. Rapid
Commun. Mass Spectrom. 17, 1579 –1584
57. Ballif, B. A., Villen, J., Beausoleil, S. A., Schwartz, D., and Gygi, S. P. (2004)
Phosphoproteomic analysis of the developing mouse brain. Mol. Cell.
Proteomics 3, 1093–1101
58. Liu, Y., and Zhang, S. (2004) Phosphorylation of 1-aminocyclopropane-1-
carboxylic acid synthase by MPK6, a stress-responsive mitogen-acti-
vated protein kinase, induces ethylene biosynthesis in Arabidopsis. Plant
Cell 16, 3386 –3399
59. Loury, R., and Sassone-Corsi, P. (2003) Histone phosphorylation: how to
proceed. Methods 31, 40–48
60. Kardailsky, I., Shukla, V. K., Ahn, J. H., Dagenais, N., Christensen, S. K.,
Nguyen, J. T., Chory, J., Harrison, M. J., and Weigel, D. (1999) Activation
tagging of the floral inducer FT. Science 286, 1962–1965
61. Yeung, K., Seitz, T., Li, S., Janosch, P., McFerran, B., Kaiser, C., Fee, F.,
Katsanakis, K. D., Rose, D. W., Mischak, H., Sedivy, J. M., and Kolch, W.
(1999) Suppression of Raf-1 kinase activity and MAP kinase signalling by
RKIP. Nature 401, 173–177
62. Perruc, E., Charpenteau, M., Ramirez, B. C., Jauneau, A., Galaud, J. P.,
Ranjeva, R., and Ranty, B. (2004) A novel calmodulin-binding protein
functions as a negative regulator of osmotic stress tolerance in Arabi-
dopsis thaliana seedlings. Plant J. 38, 410 – 420
63. Sayed, M., Kim, S. O., Salh, B. S., Issinger, O. G., and Pelech, S. L. (2000)
Stress-induced activation of protein kinase CK2 by direct interaction with
p38 mitogen-activated protein kinase. J. Biol. Chem. 275, 16569 –16573
High Throughput Identification of Arabidopsis MAPK Substrates
1568 Molecular & Cellular Proteomics 4.10
