Download A Phosphorylation State-specific Antibody Recognizes Hsp27, a

THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 280, No. 15, Issue of April 15, pp. 15013–15019, 2005
Printed in U.S.A.
A Phosphorylation State-specific Antibody Recognizes Hsp27,
a Novel Substrate of Protein Kinase D*
Received for publication, December 13, 2004, and in revised form, February 8, 2005
Published, JBC Papers in Press, February 22, 2005, DOI 10.1074/jbc.C400575200
Heike Döppler‡§, Peter Storz‡§, Jing Li¶, Michael J. Comb¶, and Alex Toker‡!
From the ‡Department of Pathology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston,
Massachusetts 02215 and ¶Cell Signaling Technology, Beverly, Massachusetts 01915
The use of phosphorylation state-specific antibodies
has revolutionized the field of cellular signaling by Ser/
Thr protein kinases. A more recent application of this
technology is the development of phospho-specific antibodies that specifically recognize the consensus substrate phosphorylated motif of a given protein kinase.
Here, we describe the development and use of such an
antibody which is directed against the optimal phosphorylation motif of protein kinase D (PKD). A degenerate
phosphopeptide library with fixed residues corresponding to the consensus LXR(Q/K/E/M)(M/L/K/E/Q/
A)S*XXXX was used as an antigen to generate an antibody that recognizes this motif. We characterized the
antibody by enzyme-linked immunosorbent assay and
with immobilized peptide arrays and also detected immunoreactive phosphoproteins in HeLa cells stimulated
with agonists known to activate PKD. Silencing PKD
expression using RNA interference validated the specificity of this antibody immunoreactive against putative
substrates. The antibody also detected the PKD substrates RIN1 and HDAC5. Knowledge of the PKD consensus motif also enabled us to identify Ser82 in the human
heat shock protein Hsp27 as a novel substrate for PKD.
We term this antibody anti-PKD pMOTIF and predict
that it will enable the discovery of novel PKD substrate
proteins in cells.
Approximately one-third of all proteins in eukaryotic cells
are phosphorylated on either serine, threonine, or tyrosine, a
reaction that is catalyzed by members of the protein kinase
superfamily (1). Over 500 protein kinases comprise the human
Kinome, and they fall into seven distinct families (2). The
identification of protein substrates of distinct protein kinases
lags significantly behind knowledge of their regulatory mechanisms. However, recent technological advances in the identification of phosphorylation sites, such chemical genetics as well
as mass spectrometry, have enabled progress in this area (3).
In addition, the advent of phosphorylation state-specific antibodies which recognize phosphorylated Ser/Thr residues in a
sequence-specific context has provided much needed insight
into the specific function of many protein kinases (4). A more
recent application of this technique is the development of sub* This work was supported by National Institutes of Health Grant
CA75134 (to A. T.). 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.
§ These authors contributed equally to this work.
! To whom correspondence should be addressed: Dept. of Pathology,
Beth Israel Deaconess Medical Center, 330 Brookline Ave., RN-237,
Boston, MA 02215. Tel.: 617-667-8535; Fax: 617-667-3616; E-mail:
[email protected].
This paper is available on line at http://www.jbc.org
strate-directed phospho-specific antibodies that recognize the
optimal consensus motif of a specific protein kinase or family of
kinases (5). This has been possible through the identification of
consensus phosphorylation motifs by degenerate peptide library approaches, such that the knowledge that many kinases
prefer to phosphorylate, for example, basophilic, acidophilic, or
Ser-Pro-directed motifs has been further refined to obtain a
specific consensus for amino acids surrounding the phosphorylatable Ser or Thr (6). As an example, identification of the
consensus motif phosphorylated by Akt/PKB, an AGC kinase,
enabled the development of an antibody which recognizes this
motif and which was used to identify novel Akt/PKB substrates, such as the TSC2 gene product, tuberin (5, 7, 8).
PKD,1 originally cloned and termed PKC! and identified as
a PKC (protein kinase C) family member, comprises a family of
three closely related isoforms, PKD1, PKD2, and PKD3/PKC".
Based on sequence similarities, PKDs are now grouped into the
CAMK (calcium and calmodulin-dependent kinases) family of
kinases. PKDs regulate a plethora of cellular responses, ranging from cell growth, cell survival, Golgi organization, and
trafficking and immune cell responses in B cells (reviewed in
Ref. 9). Similarly, the regulation of PKD catalytic activity by
cellular location, binding to adapter proteins, and by phosphorylation is also well documented. What has remained elusive,
however, is the identification of specific protein substrates of
PKD, which relay the signal to downstream responses. Notable
exceptions are Kidins220 (kinase D-interacting substrate of
220 kDa), a neuronal PKD substrate protein (10), RIN1, a PKD
substrate involved in the modulation of Ras signaling (11), and
HDAC5 (histone deacetylase 5), a class II deacetylase implicated in suppression of cardiac hypertrophy (12). Clearly, numerous other PKD substrates must exist given the large number of cellular responses attributed to this protein kinase.
We took advantage of the known optimal consensus phosphorylation motif preferred by PKD to develop a substratedirected, phospho-specific antibody that is immunoreactive
against proteins phosphorylated by PKD in cells. Using this
antibody, we have detected multiple phosphoproteins in stimulated cells and have used it to identify a previously unidentified PKD substrate, the heat shock protein Hsp27.
EXPERIMENTAL PROCEDURES
Cell Culture, Antibodies, and cDNA Expression Plasmids—The HeLa
and HEK293E cell lines were purchased from ATCC and maintained in
high glucose Dulbecco’s modified Eagle’s medium supplemented with
10% fetal bovine serum. The anti-PKD (C-20) and anti-Hsp27 antibod1
The abbreviations used are: PKD, protein kinase D; PKC, protein
kinase C; ELISA, enzyme-linked immunosorbent assay; HDAC5, histone deacetylase 5; Hsp27, heat shock protein 27; PBS, phosphatebuffered saline; PDGF, platelet-derived growth factor; PMA, 12-phorbol
13-myristate acetate; RNAi, RNA interference; S*, phosphoserine; T*,
phosphothreonine; GST, glutathione S-transferase.
15013
15014
A PKD Substrate-directed Phospho-antibody
ies were from Santa Cruz Biotechnology (Santa Cruz, CA), anti-FLAG
(M2), and anti-actin from Sigma. Anti-HDAC5, anti-phospho-(Ser/Thr)
Akt/PKB substrate antibody (Akt/PKB pMOTIF), and anti-phosphoHsp27 (pSer82) antibodies were from Cell Signaling Technology
(Beverly, MA). The anti-RIN1 and RIN1 expression plasmids were
generous gifts of J. Colicelli and are published (11). Pervanadate was
prepared as described previously (13). H2O2 (30%) was from Fisher
Scientific. PMA (phorbol 12-myristate 13-acetate), bombesin, and
bradykinin were from Sigma, and PDGF (platelet-derived growth factor) was from R&D Systems (Minneapolis, MN). Recombinant PKD was
expressed in insect cells after infection with baculovirus harboring
hemagglutinin-tagged PKD in pFAST-Bac (Invitrogen) and purified on
a nickel-nitrilotriacetic acid affinity column. The vector-based PKD1
and PKD2 RNAi in the pSUPER expression vector have been described
(14, 15). TransIT HeLa Monster reagent (Mirus, Madison, WI) was used
for all transient transfections according to the manufacturer’s instructions. Cells were stimulated or harvested 24 h after transfection. A
GST-Hsp27 fusion protein was obtained by cloning Hsp27 cDNA into
pGEX-4T1 using the following oligonucleotide primer pair for PCR:
5!-GCGGGATCCATGACCGAGCGCCGCGTCCCCTTC-3! and 5!-GCGCTCGAGTTACTTGGCGGCAGTCTCATCGGA-3!. The S15A and S82A
mutations were introduced using the following oligonucleotide primer
pair: 5!-TCGCTCCTGCGGGGCCCCGCCTGGGACCCCTTCCGCGAC3!and 5!-GTCGCGGAAGGGGTCCCAGGCGGGGCCCCGCAGGAGCGA-3! for S15A and 5!-GCGCTCAGCCGGCAACTCGCCAGCGGGGTCTCGGAGATC-3! and 5!-GATCTCCGAGACCCCGCTGGCGAGTTGCCGGCTGAGCGC-3! for S82A. Mutagenesis was carried out using the
QuikChange strategy (Stratagene), and all constructs were verified by
DNA sequencing.
Antibody Production—The PKD pMOTIF antibody was raised
against a synthetic phosphopeptide antigen CXXXLXR(Q/K/E/M)(M/L/
K/E/Q/A)S*XXXX, where X represents a position in the peptide synthesis where a mixture of all 20 amino acid (excluding C and W) were used,
and where S* represent phosphoserine. The peptide was conjugated to
keyhole limpet hemocyanin and used to immunize rabbits. Phosphopeptide-reactive rabbit antiserum was first purified by protein A chromatography. Further purification was carried out using immunodepletion
by non-phosphopeptide resin chromatography, after which the resulting eluate was chromatographed on a phosphopeptide resin. The antibody specificity of the resulting fractions was tested for specificity
toward optimal PKD substrate sequences by ELISA, immunoblotting,
and peptide arrays (see below).
Peptide Arrays—Covalent membrane-bound phosphopeptide libraries were synthesized directly on nitrocellulose membrane by the
Massachusetts Institute of Technology Biopolymers Laboratory. For
antibody specificity determination, the parental peptide library for the
array was XXLXRXXS*XXXX. The library was arrayed such that each
row represents a fixed position in the indicated library, and this position was systematically fixed with a specific amino acid as indicated in
single amino acid letter code above each column. The first column from
the left is the parental peptide library (Fig. 1C). The purified antibody
fractions eluting from the phosphopeptide chromatography were tested
against this array by incubating the antibody at a dilution of 1:1000
with membranes for 4 h at room temperature in 1% bovine serum
albumin in PBST (PBS " 0.2% Tween 20), followed by three washes
with PBST. Secondary horseradish peroxidase-conjugated antibody was
incubated at a dilution of 1:2000 in PBST for 1 h at room temperature.
After three washes with PBST the signal was revealed by Lumiglo (Cell
Signaling Technology).
Immunoblotting and Immunoprecipitation—Cells were stimulated
or harvested 24 h after transfection and lysed in lysis buffer (50 mM
Tris/HCl, pH 7.4, 1% Triton X-100, 150 mM NaCl, 5 mM EDTA, pH 7.4)
plus protease inhibitor mixture (Sigma-Aldrich). The lysates were used
either for immunoblot analysis or proteins were immunoprecipitated by
a 1-h incubation with the respective antibody (2 !g), followed by a
30-min incubation with protein G-agarose (Amersham Biosciences).
Immune complexes were washed three times with TBS (50 mM Tris/
HCl, pH 7.4, 150 mM NaCl) and resolved by SDS-PAGE or subjected to
kinase assays.
In Vitro Kinase Assays—For kinase assays with GST fusion proteins
or immunoprecipitated substrates, the reaction was carried out by
adding 0.5 !g of purified PKD to 2 !g of purified protein in a volume of
20 !l of kinase buffer. The kinase reaction was started by adding 10 !l
of kinase substrate mix (100 !M ATP (cold assay) or 100 !M ATP and 10
!Ci [#-32P]ATP in kinase buffer) and was carried out for 30 min at room
temperature. To terminate the kinase reaction, SDS sample buffer was
added, and the samples were resolved by SDS-PAGE.
RNA Interference—RNAi plasmids for PKD1 and PKD2 silencing
have been described (14, 15). To silence human Hsp27 the following
oligonucleotide sequences were cloned in pSUPER: 5!-GATCCCCGGATGGCGTGGTGGAGATCTTCAAGAGAGATCTCCACCACGCCATCCTTTTTGGAAA-3! and 5!-AGCTTTTCCAAAAAGGATGGCGTGGTGGAGATCTCTCTTGAAGATCTCCACCACGCCATCCGGG-3!. HEK293E
and HeLa cells were transfected with pSUPER or pSUPER-RNAi using
the TransIT HEK293 or HeLa-Monster reagents, respectively (Mirus).
In all experiments the cells were transfected at 30% confluence. Transfection efficiencies (80 –90%) were controlled using a green fluorescent
protein expression vector. Reduced expression of target proteins was
evaluated by immunoblotting.
ELISA—ELISA was performed according to established protocols
(16). Briefly, 50 !l of 1 !M synthetic phospho- and non-phosphopeptides
were used to coat each well in 96-well plates. Coating was carried out
overnight at 4 °C. Phospho-PKD substrate antibody was used at a 1:1000
dilution. The plates were incubated at 37 °C for 2 h after addition of
primary antibody. An alkaline phosphatase-conjugated goat anti-rabbit
antibody (Cell Signaling Technology) was used as a secondary antibody,
and p-nitrophenyl phosphate (Sigma) was used for color development.
Absorbance at 405 nm was read on an ELISA plate reader.
RESULTS AND DISCUSSION
To develop an antibody-based method for PKD substrate
identification, we first considered the optimal substrate phosphorylation motif of PKD. An oriented degenerate peptide library approach originally devised by Cantley and colleagues
(17) revealed that PKD strongly selects for aliphatic residues
(leucine, valine, and isoleucine) at the #5 position relative to
the phospho-acceptor. In this original study, arginine was
locked at the #3 position to orient the library, but a more
recent application of this technique made use of a positional
scanning peptide library to reveal that, in fact, PKD does
strongly select arginine at #3 (18). We have confirmed these
findings using an immobilized degenerate peptide library array, which was phosphorylated by purified, recombinant PKD.
The library was arrayed such that each row represents a fixed
position in the library, and this position was fixed with each of
the 20 proteogenic amino acids (Fig. 1A). The results confirmed
that PKD strongly prefers aliphatic residues, particularly
leucine, at #5, and arginine at #3. Other selectivities included
basic and aliphatic amino acids at #4, and neutral non-polar
amino acids at #2. Relatively minor selectivity was observed at
"1, although there was a preference for hydrophobic amino
acids (Fig. 1, A and B).
Thus, three independent studies have defined the optimal
consensus phosphorylation motif preferred by PKD. We used
this information to raise an antibody, which we term anti-PKD
pMOTIF, against a synthetic phosphopeptide according to the
PKD motif (Fig. 1B). Epitope mapping of the antibody was first
determined by ELISA with various synthetic phospho- and
non-phosphopeptides. ELISA reactivities relative to the parental phosphopeptide are shown in Table I. Non-phospho control
peptides scored in the 3–5% range of control phosphopeptide,
revealing that the anti-PKD pMOTIF antibody only bound to
phosphopeptides. Strongest antibody binding was detected
when both leucine at #5 and arginine at #3 were present,
consistent with the selectivity of PKD for these residues. If
leucine at #5 was absent, relatively poor binding was observed.
Similarly, peptides with leucine, valine, or isoleucine at #5, but
lacking arginine at #3, revealed no detectable binding above
control.
To further investigate the selectivity of anti-PKD pMOTIF,
we used a membrane-bound peptide array that was incubated
with purified antibody to evaluate relative selectivity at each
individual position (Fig. 1C). In the array, each spot represents
a peptide based on the parental library, with each degenerate
position represented by one amino acid, except the #5 and #3
positions where leucine and arginine were locked, respectively.
In addition, peptides were generated that had either phosphoserine, phosphothreonine, or phosphotyrosine at each of the
A PKD Substrate-directed Phospho-antibody
15015
FIG. 1. Analysis of PKD pMOTIF antibody specificity using phosphopeptide arrays. A, determination of the PKD
optimal phosphorylation motif using a degenerate immobilized peptide array. In
the array, each row represents a fixed
position in the library, and this position
was fixed with each one of the 20 naturally occurring proteogenic amino acids.
The library was incubated with recombinant, purified PKD and [#-32P]ATP,
washed, and exposed to a molecular imager. Each spot was quantitated and is
tabulated such that intensity is expressed
as a percentage of background phosphorylation. B, alignment of the optimal PKD
consensus phosphorylation pMOTIF obtained in three separate studies, as indicated. The relative selectivity of individual amino acids at each position is shown
in ascending order of text size. The blue R
indicates that this arginine was fixed at
p-3 in the degenerate library used in the
Nishikawa et al. study (17). The sequence
of the peptide used for immunization, derived from the PKD consensus motif, is
also shown. X represents any amino acid
(except cysteine and tryptophan), pS represents phosphoserine. C, phosphopeptide array for the PKD pMOTIF antibody.
The immobilized phosphopeptide array
was incubated with PKD pMOTIF antibody. The first column from the left represents the parental peptide library. The
library was arrayed such that each row
represents a fixed position in the library,
and this position was fixed with each indicated amino acid.
fixed positions. The results reveal that anti-PKD pMOTIF has
modest selectivity for amino acids at the #7 and #6 positions,
although methionine and tryptophan showed some selectivity
at #7, whereas histidine, asparagine, and proline were preferred at #6. As expected, only leucine was selected at #5.
Although in the immunizing peptide the #4 position was left
degenerate, the antibody preferentially bound basic amino acids such as arginine and lysine, as well as isoleucine, leucine,
and methionine. This conforms well to aliphatic and basic
amino acids preferred by PKD (Fig. 1C). As expected, peptides
with arginine revealed the strongest binding at #3. At #2,
glutamate, lysine, methionine, and, to a lesser extent, glutamine were preferred. Anti-PKD pMOTIF selected phosphoserine at the phospho-acceptor position, and, to a lesser extent,
phosphothreonine. There was no detectable binding of phosphotyrosine or any other amino acid at this position. Carboxyl-
terminal to the phospho-acceptor, there was modest selectivity
for hydrophobic amino acids at "1 and "2. Taken together,
these results show that anti-PKD pMOTIF binds to phosphopeptides, which accurately reflect the optimal PKD consensus motif.
We next evaluated the ability of this antibody to detect
putative PKD substrates in cells. PKD is activated in response
to a wide variety of agonists that stimulate activation of PKC,
which in turn phosphorylates and activates PKD (19). NIH-3T3
fibroblasts were stimulated with either bombesin, bradykinin,
or PDGF. We also used PMA and pervanadate, which are well
known activators of PKD. In response to all of these agonists,
increased phosphorylation of a number of putative PKD substrates was detected (Fig. 2A). Specifically, three proteins of
$85, 100, and 150 kDa were detected in fibroblasts stimulated
with PDGF, PMA, and pervanadate. A strong immunoreactive
15016
A PKD Substrate-directed Phospho-antibody
TABLE I
Specificity of the PKD pMOTIF antibody using various
phospho- and non-phosphopeptides containing different
versions of the consensus PKD substrate pMOTIF as
determined by ELISA
Reactivity is expressed as a percentage of ELISA reading of each
peptide relative to that of the optimal PKD consensus motif (Fig. 1C).
Uppercase S or T represents non-phospho-Ser/Thr, and lowercase s or t
represents phospho-Ser/Thr.
Peptide no.
Peptide sequence
% pMOTIF
1
2
3
4
5
6
7
8
9
10
11
12
13
CXXXLXR(Q/K/E/M)(M/L/K/E/Q/A)sXXXX
CXXXLXR(Q/K/E/M)(M/L/K/E/Q/A)SXXXX
CXXXLXRXXtXXXX
CXXXLXRXXTXXXX
CXXXXXRRXtXXXX
CXXXXXXLtQ(DE)XXXXX
CXXXRR(S/G)s(K/L/D)XXX
CXXXRR(L/R)s(K/L/D)XXX
CXXXRXRXXtXXXX
CXXRXRX(L/A)s(R/F)XXX
CXXRXRX(R/E)s(R/F)XXX
CXXRXRX(L/A)s(V/A/T)XXX
CXXRXRX(R/E)s(V/A/T)XXX
100
3.8
82
3.7
34.4
4.6
13.7
10.6
28.1
34.3
24.7
37.6
42
LXRXXs/t peptides
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
CGLPRPKsAGTAT
CGLPRPKSAGTAT
CGLYRSPsMPENLNRPRL
CGLYRSPSMPENLNRPRL
CSPALKRSHsDSLD
CSPALKRSHSDSLD
CQRLFRSPsMP
CQRLFRSPSMP
CFLQRYssDPTGAL
CFLQRYSSDPTGAL
CPLGRTQsAPLP
CPLGRTQSAPLP
CPLNRTQsAPLP
CPLNRTQSAPLP
CRRVPFSLLRGPsWDPF
CRRVPFSLLRGPSWDPF
CLMRRNsVTPLAS
CLMRRNSVTPLAS
CRRAHQLFRGFsFVAIT
CRRAHQLFRGFSFVAIT
CLQRRLsVYRQIV
CLQRRLSVYRQIV
CHKVLPRGLsPARQLL
CHKVLPRGLSPARQLL
CGLKRSLsEMEIG
CGLKRSLSEMEIG
CKINRSAsEPSLH
CKINRSASEPSLH
CVPRRKsLVGTPY
CVPRRKSLVGTPY
CTPKKPGLRRRQt
CTPKKPGLRRRQT
CVASMMHRQEtVE
CVASMMHRQETVE
48
49
50
51
52
53
SSRRTtLCGTLD
SSRRTTLCGTLD
CIPTRRHtFRRQNV
CIPTRRHTFRRQNV
CTIPRRNtLPAMDNS
CTIPRRNTLPAMDNS
4.8
3.6
66.6
4
7.6
3.6
72.4
3.4
7.7
4.2
5.1
3.9
26.1
4
59
3.4
26.9
3.7
21.7
4.2
40.7
3.8
3.9
3.5
11.1
3.6
8.9
3.9
25.1
3.9
30.7
3.9
7.4
4
RXXs/t peptides
6.3
3.6
12.2
3.3
3.8
3.5
RXXRXs/t peptides
54
55
56
57
58
59
60
61
62
63
CLKRKRRPtSGLHPED
CLKRKRRPTSGLHPED
CVEMIRRRRPtPAML
CVEMIRRRRPTPAML
SRPRSCtWPLPREI
SRPRSCTWPLPREI
CARGRFAtVVEEL
CARGRFATVVEEL
CIRGRKRtVWGAKQI
CIRERKRTVWGAKQI
29.2
3.1
3.5
2.9
3.6
3
6.1
2.9
22.8
3.2
TABLE I—continued
Peptide no.
64
65
66
67
68
69
Peptide sequence
CTRDRVPtYQYNM
CTRDRVPTYQYNM
CIRDRNGtHLDAGAL
CIRDRNGTHLDAGAL
CMRERLGtGGFGNV
CMRERLGTGGFGNV
% pMOTIF
16.3
3
3.4
3.1
3.5
3.8
V/L/IXXXXs/t peptides
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
VLSPLPsQAMDDLLC
CELQTDGsQASRS
CLEQHGsQPSNSYPSI
CVKYSSsQPEPRTG
CPEVPAERsPRRRSIS
CEEVAAKKsPVKATAP
VIPPHtPVRTVMNTC
GLGPSLtEDQPGPYLAPGLLGSNIHQQGR
CFGLSVQMtEDVY
CLGMTDsEEDLDPM
IGLDCASsEFFK
CSEKLFQGYsFVAPS
CLLTTGGtLKISDL
CLEPQKsLGDEG
CILLSELsRRRIRS
3.1
3.1
3.1
3
3.1
3.1
3
3.4
4.4
3.3
3.1
4.2
3.2
2.9
2.9
band was also detected at $45 kDa. Two immunoreactive
bands at 25 and 27 kDa were also detected in response to all
agonists. These data show that the anti-PKD pMOTIF antibody is immunoreactive against proteins in cells stimulated
with agonists that are known to activate PKD.
Because several other proteins were detected by this antibody even in unstimulated cells, probably reflecting nonspecific
immunoreactivity, we next determined whether any of these
putative substrates could be directly attributed to PKD activity. To achieve this, we reduced endogenous PKD expression in
HEK293 cells with RNAi. Cells were stimulated with either
pervanadate or H2O2, as studies have shown that exposure of
cells to oxidative stress such as H2O2 potently activates PKD
(14, 20). The same range of phosphoproteins revealed by the
anti-PKD pMOTIF antibody in fibroblasts were also detected in
control HEK293 cells (Fig. 2B). In cells transfected with PKD
RNAi, there was a marked reduction in the immunoreactivity
of several proteins, most notably at 27 kDa. Because the antiPKD pMOTIF antibody is phospho-specific we conclude that
these may represent novel PKD substrates.
To further validate the specificity of the PKD pMOTIF antibody, we next compared the immunoreactivity of this antibody
with the Akt/PKB substrate-directed antibody, which recognizes the optimal consensus sequence RXRXXS*/T*$ (5). Stimulation of serum-starved HeLa cells with IGF-1, a potent agonist of the phosphoinositide 3-kinase-Akt/PKB signaling
pathway, resulted in the phosphorylation of a number of proteins detected by the Akt/PKB substrate antibody (Fig. 2C, left
panel). These bands were not detected in cells stimulated with
PMA, which does not activate Akt/PKB. Conversely, PMA, a
potent stimulus for PKD, induced the phosphorylation of a distinct subset of proteins as measured by PKD pMOTIF immunoreactivity (Fig. 2C, right panel). Stimulation of cells with IGF-1,
which does not activate PKD, did not induce the phosphorylation
of the same subset of proteins. Note, however, that both antibodies did detect some proteins in cells stimulated with both agonists, and these are likely to represent neither Akt/PKB nor PKD
substrates. Finally, we determined whether the PKD pMOTIF
antibody is immunoreactive against known PKD substrates. To
this end, RIN1, a neuronal PKD substrate, was detected by PKD
pMOTIF in transfected HeLa cells stimulated with PMA (Fig.
2D). Similarly, endogenous HDAC5, also a known PKD substrate, was detected by this antibody, again in response to PMA
A PKD Substrate-directed Phospho-antibody
15017
FIG. 2. PKD pMOTIF antibody recognizes putative substrates. A, NIH3T3 cells were stimulated with Bombesin
(50 !M, 10 min), bradykinin (50 ng/ml, 10
min), PDGF (50 ng/ml, 10 min), PMA (100
nM, 10 min), or pervanadate (75 !M, 15
min). Cells were lysed, and samples were
immunoblotted against the PKD pMOTIF
antibody. The open arrow indicates a 27kDa protein, and filled arrows indicate
other putative PKD substrates. B,
HEK293E cells were transfected with vector control (pSUPER) or PKD RNAi (pSUPER PKD1/2) for 48 h. Cells were then
stimulated with H2O2 (10 !M, 10 min),
pervanadate (75 !M, 15 min), or PMA
(100 nM, 10 min). Lysates were immunoblotted with the PKD pMOTIF antibody.
The open arrow indicates a putative PKD
substrate of 27 kDa. PKD silencing by
PKD-specific RNAi and equal loading
were determined in control blots using
%-PKD or %-actin. C, HeLa cells were serum-starved for 24 h and then stimulated
with IGF-1 (50 ng/ml) or PMA (100 nM) for
10 min. Lysates were immunoblotted
with anti-Akt/PKB pMOTIF (left panel)
and anti-PKD pMOTIF (right panel).
Open arrows indicate putative Akt/PKB
substrates, and solid arrows indicate putative PKD substrates. D, top panel, HeLa
cells were transfected with RIN1 or vector
control (Ctrl) and then stimulated with
PMA (100 nM, 10 min) and lysates immunoblotted with anti-PKD pMOTIF or antiRIN1. Bottom panel, HeLa cells were
stimulated with PMA (100 nM, 10 min),
and HDAC5 was immunoprecipitated.
Immunoprecipitates were immunoblotted
with anti-PKD pMOTIF, stripped, and reprobed with anti-HDAC5.
stimulation of HeLa cells. Therefore, PKD pMOTIF detects both
putative and known PKD substrate proteins.
To further demonstrate the feasibility of this antibody to
discover novel PKD substrates, we performed a protein BLAST
search on the Swiss-Prot data base using the peptide sequences
from the ELISA screen. The search consistently returned the
heat shock protein Hsp27 with the highest score for a protein
containing the PKD consensus phosphorylation motif. We
noted that one of the most prominent immunoreactive bands in
Fig. 2B migrated at 27 kDa. To validate that this indeed represents Hsp27, we first reduced expression of endogenous
Hsp27 using RNAi, followed by immunoblotting with anti-PKD
pMOTIF. As predicted, immunoreactivity of the 27-kDa band
was significantly reduced following H2O2 stimulation of HeLa
cells (Fig. 3A). Analysis of the Hsp27 amino acid sequence
reveals two optimal putative PKD phosphorylation sites at
Ser15 and Ser82 (Fig. 3B). Shown for comparison is the minimal
PKD consensus phosphorylation sequence and the PKD phosphorylation sites in RIN1 and HDAC5. Next, we evaluated
immunoreactivity of Hsp27 in cells transfected with PKD
RNAi. There was a marked reduction in the phosphorylation of
the 27kDa band as detected by immunoblotting total cell lysates with anti-PKD pMOTIF, whereas total Hsp27 levels were
unaffected (Fig. 3C, left panel). That this band represents
Hsp27 was confirmed by immunoprecipitation of Hsp27, followed by immunoblotting with PKD pMOTIF, and again there
was a reduction in immunoreactivity in cells transfected with
PKD RNAi (Fig. 3C, right panel). Because this antibody is
phospho-specific, we conclude that PKD activation leads to the
phosphorylation of Hsp27, which is detected by PKD pMOTIF.
Next, we investigated on which residue Hsp27 is phosphorylated by PKD. GST-Hsp27 fusion proteins, either wild-type,
Ser15 3 Ala, or Ser82 3 Ala mutations, were incubated with
purified, recombinant PKD in in vitro kinase assays. Both
wild-type and Ser15 3 Ala GST-Hsp27 were efficiently phosphorylated by PKD, whereas the Ser82 3 Ala mutant showed
no detectable phosphorylation (Fig. 3D, left panel). This was
confirmed by immunoblotting separate in vitro kinase assays
either with anti-PKD pMOTIF or with a phospho-antibody
specific to Ser82 in Hsp27. Both antibodies recognized wild-type
Hsp27, wheras Ser82 3 Ala Hsp27 immunoreactivity was reduced (Fig. 3D, right panel). These results demonstrate that: (i)
PKD directly phosphorylates Hsp27 and that (ii) Ser82, and not
Ser15, is the relevant site, at least in vitro. This is also true in
cells, because both wild-type and Ser15 3 Ala Hsp27 are efficiently detected by anti-PKD pMOTIF upon stimulation with
H2O2, whereas no appreciable immunoreactivity was evident
with the Ser82 3 Ala mutant (Fig. 3E, top panel). Again, the
same result was obtained by immunoblotting with anti-pSer82
(Fig. 3E, bottom panel). We therefore conclude that Hsp27 is
phosphorylated at Ser82 by PKD in stimulated cells.
Although much is known about the mechanisms of regula-
15018
A PKD Substrate-directed Phospho-antibody
FIG. 3. Identification of Hsp27 as a PKD substrate. A, HeLa cells were transfected with vector control (pSUPER) or Hsp27 RNAi
(Hsp27-RNAi) for 48 h. Cells were then stimulated with H2O2 (10 !M, 10 min), and lysates were immunoblotted with anti-PKD pMOTIF. Lysates
were also probed with anti-Hsp27 and anti-actin. B, alignment of putative PKD phosphorylation sites in Hsp27 and the PKD minimal substrate
motif. C, HeLa cells were transfected with vector control (pSUPER) or PKD RNAi for 48 h. Cells were then stimulated with H2O2 (10 !M, 10 min)
and lysed. Left panel, lysates were immunoblotted with anti-PKD pMOTIF antibody, stripped, and re-probed with anti-Hsp27. Lysates were also
immunoblotted with anti-PKD and anti-actin for loading control. Right panel, Hsp27 immunoprecipitates were immunoblotted with anti-PKD
pMOTIF, stripped, and reprobed with anti-Hsp27 (ns, nonspecific immunoreactive band). D, baculovirus-expressed, purified PKD was incubated
in an in vitro kinase assay with GST, GST-Hsp27, GST-Hsp27.S15A, or GST-Hsp27.S82A followed by autoradiography (top left panel). Equal
loading of GST and GST fusion proteins was controlled by Coomassie Blue staining (bottom left panel). Baculovirus-expressed, purified PKD was
incubated in a cold in vitro kinase assay with GST, GST-Hsp27, or GST-Hsp27.S82A; resolved by SDS-PAGE; and immunoblotted with the
anti-PKD pMOTIF and anti-pSer82 Hsp27 (right panels) E, FLAG-tagged Hsp27, Hsp27.S15A, Hsp27.S82A, or vector control was overexpressed
in HeLa cells. Cells were stimulated with H2O2 (10 !M, 10 min), and Hsp27 was immunoprecipitated (%-FLAG). Hsp27 phosphorylation was
determined by immunoblotting with anti-PKD pMOTIF (top panel) or with anti-pSer82 Hsp27 (bottom panel). Equal expression of Hsp27 was
determined by reprobing against Hsp27 or anti-FLAG.
tion of PKD and its importance in cell biology, the identification
of specific protein substrates that relay the PKD signal has
remained elusive. Here we have used an antibody-based
method, which we speculate will aid in the identification of
such substrates. We have shown that the anti-PKD pMOTIF
antibody reacts with peptides that conform to the preferred
phosphorylation motif of this kinase and furthermore validate
its use and show that the Hsp27 protein is a previously unidentified in vivo PKD substrate. Phosphorylation of Hsp27 at
Ser15 and Ser82 has previously been demonstrated in response
to treatment of cells with a variety of stresses such as oxidative
stress and heat shock (21, 22), and the MAPKAP kinases 2/3
have been shown to phosphorylate Hsp27 in vitro (21). However, the identity of the physiological kinase(s) for Hsp27 phosphorylation at these residues has not been determined. Using a
combination of the anti-PKD pMOTIF antibody, PKD-specific
RNAi, and in vitro kinase assays, we show that PKD is the
relevant kinase for Hsp27 phosphorylation at Ser82. Hsp27
A PKD Substrate-directed Phospho-antibody
phosphorylation at Ser15 and Ser82 modulates oligomerization
and chaperone function, leading to protection of cells from injury
due stress (23). Because PKD plays a major role in protecting
cells from oxidative stress (14), we further speculate that Hsp27
phosphorylation by PKD may play a key role in this response.
The use of the PKD pMOTIF antibody in combination with
proteome-wide screens should yield much needed information
concerning the identify of additional PKD substrates. Because
other kinases in the human Kinome may reveal optimal phosphorylation motifs similar to that of PKD, it will be important
to perform combinatorial screens with PKD-specific RNAi, as
shown here. Additional in vitro validation by direct phosphorylation of identified putative PKD substrates will also be
important to confirm the newly identified substrate. Given the
present lack of any other rapid, substrate-directed methods to
discover substrates of PKD in cells, this method should be well
suited for analysis of PKD signaling in cells.
Acknowledgments—We thank the members of the joint Friday morning signaling group meeting at Beth Israel Deaconess Medical Center/
Harvard Medical School and members of the Toker laboratory for insightful discussions and advice. We are grateful to John Colicelli for
providing the RIN1 antibody and plasmids.
REFERENCES
1. Cohen, P. (2002) Nat. Cell Biol. 4, E127–E130
2. Manning, G., Whyte, D. B., Martinez, R., Hunter, T., and Sudarsanam, S.
(2002) Science 298, 1912–1934
3. Manning, B. D., and Cantley, L. C. (2002) Science’s STKE 2002, PE49
15019
4. Czernik, A. J., Girault, J. A., Nairn, A. C., Chen, J., Snyder, G., Kebabian, J.,
and Greengard, P. (1991) Methods Enzymol. 201, 264 –283
5. Zhang, H., Zha, X., Tan, Y., Hornbeck, P. V., Mastrangelo, A. J., Alessi, D. R.,
Polakiewicz, R. D., and Comb, M. J. (2002) J. Biol. Chem. 277, 39379 –39387
6. Songyang, Z. (2001) Methods Enzymol. 332, 171–183
7. Obata, T., Yaffe, M. B., Leparc, G. G., Piro, E. T., Maegawa, H., Kashiwagi, A.,
Kikkawa, R., and Cantley, L. C. (2000) J. Biol. Chem. 275, 36108 –36115
8. Manning, B. D., Tee, A. R., Logsdon, M. N., Blenis, J., and Cantley, L. C. (2002)
Mol. Cell 10, 151–162
9. Van Lint, J., Rykx, A., Maeda, Y., Vantus, T., Sturany, S., Malhotra, V.,
Vandenheede, J. R., and Seufferlein, T. (2002) Trends Cell Biol. 12, 193–200
10. Iglesias, T., Cabrera-Poch, N., Mitchell, M. P., Naven, T. J., Rozengurt, E., and
Schiavo, G. (2000) J. Biol. Chem. 275, 40048 – 40056
11. Wang, Y., Waldron, R. T., Dhaka, A., Patel, A., Riley, M. M., Rozengurt, E., and
Colicelli, J. (2002) Mol. Cell. Biol. 22, 916 –926
12. Vega, R. B., Harrison, B. C., Meadows, E., Roberts, C. R., Papst, P. J., Olson,
E. N., and McKinsey, T. A. (2004) Mol. Cell. Biol. 24, 8374 – 8385
13. Storz, P., Doppler, H., Johannes, F. J., and Toker, A. (2003) J. Biol. Chem. 278,
17969 –17976
14. Storz, P., and Toker, A. (2003) EMBO J. 22, 109 –120
15. Storz, P., Doppler, H., and Toker, A. (2004) Mol. Cell. Biol. 24, 2614 –2626
16. Perlmann, H., and Perlmann, P. (1994) Cell Biology: A Laboratory Handbook,
Academic Press, Inc., San Diego, CA
17. Nishikawa, K., Toker, A., Johannes, F. J., Songyang, Z., and Cantley, L. C.
(1997) J. Biol. Chem. 272, 952–960
18. Hutti, J., Jarrell, E. T., Chang, J. D., Abbott, D. W., Storz, P., Toker, A.,
Cantley, L. C., and Turk, B. E. (2004) Nat. Methods 1, 27–29
19. Zugaza, J. L., Sinnett-Smith, J., Van Lint, J., and Rozengurt, E. (1996) EMBO
J. 15, 6220 – 6230
20. Waldron, R. T., and Rozengurt, E. (2000) J. Biol. Chem. 275, 17114 –17121
21. Landry, J., Lambert, H., Zhou, M., Lavoie, J. N., Hickey, E., Weber, L. A., and
Anderson, C. W. (1992) J. Biol. Chem. 267, 794 – 803
22. Lambert, H., Charette, S. J., Bernier, A. F., Guimond, A., and Landry, J. (1999)
J. Biol. Chem. 274, 9378 –9385
23. Rogalla, T., Ehrnsperger, M., Preville, X., Kotlyarov, A., Lutsch, G., Ducasse,
C., Paul, C., Wieske, M., Arrigo, A. P., Buchner, J., and Gaestel, M. (1999)
J. Biol. Chem. 274, 18947–18956