Báo cáo khoa học: Analysis of the CK2-dependent phosphorylation of serine 13 in Cdc37 using a phospho-specific antibody and phospho-affinity gel electrophoresis doc

Analysis of the CK2-dependent phosphorylation of
serine 13 in Cdc37 using a phospho-specific antibody
and phospho-affinity gel electrophoresis
Yoshihiko Miyata and Eisuke Nishida
Department of Cell and Developmental Biology, Graduate School of Biostudies, Kyoto University, Japan

Keywords
Cdc37; CK2; gel electrophoresis; Hsp90;
protein kinase
Correspondence
Y. Miyata, Department of Cell &
Developmental Biology, Graduate School of
Biostudies, Kyoto University, Kitashirakawa
Oiwake-cho, Sakyo-ku, Kyoto 606-8502,
Japan
Fax: +81 75 753 4235
Tel: +81 75 753 4231
E-mail:
(Received 22 March 2007, revised 9 August
2007, accepted 4 September 2007)
doi:10.1111/j.1742-4658.2007.06090.x

The CK2-dependent phosphorylation of Ser13 in cell division cycle protein 37 (Cdc37), a kinase-specific heat shock protein 90 (Hsp90) cochaperone, has previously been reported to be essential for the association of
Cdc37 with signaling protein kinases [Bandhakavi S, McCann RO, Hanna
DE & Glover CVC (2003) J Biol Chem 278, 2829–2836; Shao J, Prince T,
Hartson SD & Matts RL (2003) J Biol Chem 278, 38117–38220; Miyata Y
& Nishida E (2004) Mol Cell Biol 24, 4065–4074]. Here we describe a new
phospho-specific antibody against Cdc37 that recognizes recombinant purified Cdc37 only when incubated with CK2 in the presence of Mg2+ and
ATP. The replacement of Ser13 in Cdc37 by nonphosphorylatable amino
acids abolished binding to this antibody. The antibody was specific for
phosphorylated Cdc37 and did not crossreact with other CK2 substrates

such as Hsp90 and FK506-binding protein 52. Using this antibody, we
showed that complexes of Hsp90 with its client signaling kinases, Cdk4,
MOK, v-Src, and Raf1, contained the CK2-phosphorylated form of Cdc37
in vivo. Immunofluorescent staining showed that Hsp90 and the phosphorylated form of Cdc37 accumulated in epidermal growth factor-induced membrane ruffles. We further characterized the phosphorylation of Cdc37 using
phospho-affinity gel electrophoresis. Our analyses demonstrated that the
CK2-dependent phosphorylation of Cdc37 on Ser13 caused a specific gel
mobility shift, and that Cdc37 in the complexes between Hsp90 and its client signaling protein kinases was in the phosphorylated form. Our results
show the physiological importance of CK2-dependent Cdc37 phosphorylation and the usefulness of phospho-affinity gel electrophoresis in protein
phosphorylation analysis.

Protein kinases play pivotal roles in cellular signal
transduction systems. Reversible protein phosphorylation is one of the major mechanisms used to control
the function, localization and stability of proteins
inside cells [1]. Therefore, the analysis of protein kinase
activity and the phosphorylation level of their substrates are important for understanding signal transduction pathways at a molecular level. Many methods

have been described for determining protein kinase
activity and protein phosphorylation levels. In vitro,
phosphorylation reactions can be monitored by incubating a protein kinase and a substrate in the presence
of radioactive ATP ([32P]ATP[cP]), followed by
SDS ⁄ PAGE and autoradiography to quantify radioactivity in the substrate. When a peptide substrate
is used, the amount of radioactivity incorporated into

Abbreviations
Cdc37, cell division cycle protein 37; EGF, epidermal growth factor; ERK, extracellular signal-regulated kinase; FKBP52, FK506-binding
protein 52; GST, glutathione S-transferase; HRP, horseradish peroxidase; Hsp90, heat shock protein 90; TBB, 4,5,6,7tetrabromobenzotriazole.

5690

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

Y. Miyata and E. Nishida

the peptide can be quantified by scintillation counting
after separating the radioactive peptide from free
ATP ⁄ ADP using phosphocellulose filters. Protein
phosphorylation in vivo can be examined in several
ways. The phosphorylation level of a substrate can be
determined by isolating the substrate from radiolabeled
cells or tissues by immunoprecipitation followed by
SDS ⁄ PAGE and autoradiography. For a phosphorylation site with a known sequence, it is possible to
obtain a phospho-specific antibody that reacts with the
substrate only in its phosphorylated form by immunization and affinity purification with the corresponding
phosphopeptide. A phospho-specific antibody can then
be used to directly quantify the site-specific phosphorylation of a substrate in vivo by western blot analysis.
Recently, MS has become a powerful technology for
large-scale detection and quantification of in vivo protein phosphorylation [2,3].
Although the precise molecular mechanism remains
to be elucidated, in some cases protein phosphorylation causes a mobility shift of the protein band on
SDS ⁄ PAGE, often but not always decreasing the
mobility. However, the mobility shifts are generally
not very large, and in most cases protein phosphorylation does not induce a mobility shift at all. Moreover,
achieving optimal band shifts often requires special gel
compositions (such as a low concentration of bis-acrylamide), which can only be determined by somewhat
hit-and-miss experimentation. A more reproducible
and reliable method for discriminating phosphorylated
and nonphosphorylated forms of a broad range of
proteins by gel electrophoresis has long been sought.
Recently, Kinoshita et al. identified alkoxide-bridged

dinuclear metal complexes as novel phosphate-binding
compounds that preferentially capture phosphomonoester dianions bound to serine, threonine and tyrosine
residues in proteins [4]. They also reported that these
compounds could be used to separate phosphorylated
and unphosphorylated proteins in SDS ⁄ PAGE [5].
Protein kinase activity in cells is regulated in many
different ways. Releasing an inhibitory subunit from a
catalytic subunit can activate a kinase. By contrast,
binding an activating regulatory coprotein to an inactive catalytic subunit can activate a kinase. In many
signal-transducing protein kinases, site-specific phosphorylation by an upstream protein kinase (a kinasekinase) activates them. Before these activation steps,
the protein kinases must be in the correct structural
conformation, to be activated by the appropriate stimuli. However, the activation-ready structures of signaling protein kinases are relatively unstable in nature
and require additional proteins called ‘molecular chaperones’ to stabilize them within cells. Among the

Cdc37 phosphorylation by CK2 and signaling kinases

molecular chaperones, heat shock protein 90 (Hsp90)
and cell division cycle protein 37 (Cdc37) have been
shown to be specifically required for the stability and
function of many signaling protein kinases, including
Raf1 [6,7], Cdk4 [8–10], MOK [11], IKK [12], and
v-Src [13]. Hsp90 is an important molecular chaperone
whose ATP-dependent function is essential for the
folding and function of many signaling molecules,
including protein kinases and steroid hormone receptors [14–16]. Cdc37 both acts as a molecular chaperone
by itself and is also required for the efficient recruitment of Hsp90 to protein kinase complexes [17,18].
Therefore, Cdc37 activity is crucial for many signaling
protein kinases to function correctly in vivo.
Protein kinase CK2 is a ubiquitous and highly conserved protein kinase that is known to be involved in
many physiological functions by phosphorylating a

plethora of substrates [19–21]. CK2 is elevated in many
types of tumor, and its overexpression is tumorigenic
in experimental models in animals, suggesting that
CK2 is involved in both cell cycle control and neoplastic cell growth [22]. Although CK2 was one of the earliest protein kinases identified, its regulatory
mechanism remains largely unknown. CK2 is composed of two catalytic subunits (a and ⁄ or a¢) and two
noncatalytic subunits (b). The catalytic subunits of
CK2 are constitutively active, whether or not they are
associated with noncatalytic b-subunits [23], CK2
activity is independent of any known second messengers, and no upstream ‘kinase-kinase’ has been identified that activates CK2 [24].
We and others have previously identified Cdc37 as
a pivotal substrate for CK2 and reported that Cdc37
phosphorylation by CK2 is essential for Cdc37 to act
as a molecular chaperone for many signaling protein
kinases [25–28]. Moreover, the molecular chaperone
functions of Hsp90 and Cdc37 are required for CK2
itself to be optimally active [25,29,30], suggesting that
CK2 and Cdc37 together constitute a positive feedback mechanism to control various signaling protein
kinases [25,31]. Therefore, analyzing the level of
CK2-dependent Cdc37 phosphorylation in vivo should
be crucial for understanding the regulatory mechanisms of Cdc37, CK2, and other signaling protein
kinases.
In this report, we describe a new antibody that recognizes Cdc37 only when phosphorylated by CK2. We
also demonstrate that the phosphorylation of Cdc37
can be analyzed by phospho-affinity gel electrophoresis. Together, the phospho-specific antibody and phospho-affinity gel electrophoresis enabled us to directly
study CK2-dependent Cdc37 phosphorylation in vitro
and in vivo.

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5691

Cdc37 phosphorylation by CK2 and signaling kinases

Y. Miyata and E. Nishida

Results
Phospho-specific antibody against Cdc37
We synthesized the phosphopeptide VWDHIEVpSDDEDETHC, including amino acid residues 6–20 of
mammalian Cdc37, in which Ser13 was phosphorylated. This sequence is in the most N-terminal region
of Cdc37, where the amino acid sequence conservation
between species is highest. In fact, the amino acid
sequence of the peptide we used is identical in human,
chimpanzee, rhesus monkey, cattle, pig, rat, mouse
and chick sequences. The serine at position 13 of
Cdc37, which was included in the synthetic peptide,
has been reported to be phosphorylated by CK2 and
important for the functional regulation of Cdc37 [25–
28]. We raised rabbit antiserum against this phosphopeptide and examined the specificity of the purified
antibody. Purified recombinant Cdc37 was incubated
with the recombinant catalytic subunit of CK2 (CK2a)
or with CK2 holoenzyme (CK2a2b2) or alone, in the
presence or absence of Mg2+ and ⁄ or ATP. As shown
in Fig. 1A, the antibody (anti-[pSer13]-Cdc37 hereafter) recognized Cdc37 only when incubated with either
CK2a or CK2 holoenzyme in the presence of ATP and
Mg2+ (lanes 6 and 9). The absence of either CK2
(lane 3), Mg2+ (lanes 4 and 7) or ATP (lanes 5 and 8)
completely abolished antibody binding (Fig. 1A), indicating that anti-[pSer13]-Cdc37 specifically recognized

the CK2-phosphorylated form of Cdc37. The presence

of equal amounts of Cdc37 in the phosphorylation
mixtures was checked by probing western blots with
an antibody to Cdc37 (Fig. 1B).
The binding of anti-[pSer13]-Cdc37 was completely
abolished when Ser13 in Cdc37 was replaced by a nonphosphorylatable amino acid. Recombinant wild-type
Cdc37 and two Cdc37 mutants in which Ser13 was
replaced by alanine [Cdc37(13SA)] or aspartic acid
[Cdc37(13SD)] were incubated with CK2a or CK2 holoenzyme or alone, in the presence of Mg2+–ATP. Wildtype Cdc37 bound anti-[pSer13]-Cdc37 in the presence
of CK2a or CK2 holoenzyme (Fig. 1C, lanes 4 and 7).
By contrast, neither Cdc37(13SA) nor Cdc37(13SD)
were recognized by anti-[pSer13]-Cdc37, even after incubation with CK2a or CK2 holoenzyme (Fig. 1C,
lanes 5, 6, 8 and 9). The amounts of Cdc37(WT),
Cdc37(13SA) and Cdc37(13SD) in all the incubation
mixtures were approximately the same, as shown by
Coomassie brilliant blue (CBB) staining (Fig. 1D).
These results showed that anti-[pSer13]-Cdc37 recognized Cdc37 only when Ser13 was phosphorylated by
CK2, and that isolated CK2a phosphorylated the same
site in Cdc37 (Ser13) as purified CK2 holoenzyme.
We next examined the specificity of anti-[pSer13]Cdc37 for Cdc37 in comparison to other CK2-phosphorylated proteins. CK2 phosphorylates serine or
threonine residues followed by a stretch of acidic
amino acids [19,21], which constitutes a consensus

Fig. 1. Phospho-specific antibody against an Hsp90 cochaperone, Cdc37. (A, B) Recombinant Cdc37 was incubated at 30 °C for 30 min
alone (lanes 1–3), with CK2a (lanes 4–6), or with purified CK2 holoenzyme (lanes 7–9), in the presence (+) or absence (–) of Mg2+ and ⁄ or
ATP as indicated above the track. Western blots of these mixtures with anti-[pSer13]-Cdc37 (A) or with anti-Cdc37 (B) are shown. The positions of molecular weight markers and Cdc37 are shown. (C, D) Wild-type protein [Cdc37(WT), lanes 1, 4 and 7] as well as two mutant proteins [Cdc37(13SA), lanes 2, 5 and 8, and Cdc37(13SD), lanes 3, 6 and 9] were incubated at 30 °C for 30 min alone (lanes 1–3), with CK2a
(lanes 4–6) or with CK2 holoenzyme (lanes 7–9), and the mixtures were analyzed by western blotting with anti-[pSer13]-Cdc37 (C). CBB
staining is shown in (D). The positions of molecular weight markers and Cdc37 are shown.

5692

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS


Y. Miyata and E. Nishida

phosphorylation sequence for CK2. Thus, the amino
acid sequences surrounding CK2 phosphorylation sites
are similar in most CK2 substrates. We examined two
other molecular chaperones, Hsp90 and FK506-binding protein 52 (FKBP52), which are also known to be
phosphorylated by CK2, to find whether they were recognized by anti-[pSer13]-Cdc37 when phosphorylated.
In fact, the amino acid sequences around the known
CK2 phosphorylation sites in Cdc37 (Ser13) [26,28],
FKBP52 (Thr143) [32], and Hsp90 (Ser231 and
Ser263) [33] are highly homologous (Fig. 2A). Purified
recombinant FKBP52 and Hsp90 were incubated with
or without CK2 in the presence of Mg2+–
[32P]ATP[cP]. Cdc37(WT) and Cdc37(13SA) were
included as positive and negative controls. Analysis of
the phosphorylation mixtures by SDS ⁄ PAGE and
autoradiography clearly showed that Cdc37(WT),
FKBP52 and Hsp90 were heavily phosphorylated by
CK2 in vitro (Fig. 2B). Western blot analysis of the
same phosphorylation mixtures probed with anti[pSer13]-Cdc37 showed that the antibody only recognized CK2-phosphorylated Cdc37 (Fig. 2C, lane 5),
and not CK2-phosphorylated FKBP52 (Fig. 2C,
lane 7) or CK2-phosphorylated Hsp90 (Fig. 2C,
lane 8). The presence of equal amounts of Cdc37,
FKBP52 and Hsp90 in the phosphorylation mixtures
were checked by CBB staining (Fig. 2D). These results
indicated that anti-[pSer13]-Cdc37 specifically recognized the CK2-phosphorylated form of Cdc37, and did
not crossreact with a generic CK2 phosphorylation

consensus sequence.
Phosphorylated Cdc37 associates with signaling
protein kinases

Cdc37 phosphorylation by CK2 and signaling kinases

to Cdc37, we next investigated whether the CK2-phosphorylated form of Cdc37 associated with signaling
protein kinases in vivo. Four typical Hsp90 client kinases, Cdk4, MOK, v-Src, and Raf1, were expressed
in COS7 cells as FLAG-tagged proteins, and kinase–
Hsp90–Cdc37 complexes were immunopurified on
anti-FLAG affinity resin. As controls, two nonclient
kinases for Hsp90, CK1 and DYRK2, were included.

A

B

C

We previously reported that replacing Ser13 in Cdc37
with a nonphosphorylatable amino acid abolished the
binding of Cdc37 to Hsp90 client signaling protein
kinases [26,27]. Using the phospho-specific antibody

Fig. 2. Substrate specificity of the antibody against CK2-phosphorylated Cdc37. (A) Alignment of the amino acid sequences surrounding the CK2 phosphorylation sites of Cdc37 (rat), FKBK52 (rabbit),
and Hsp90 (human). The position of the CK2-catalyzed phosphorylation sites is indicated by an arrow. (B) Recombinant purified
Cdc37(WT) (lanes 1 and 5), Cdc37(13SA) (lanes 2 and 6), FKBP52
(lanes 3 and 7) or Hsp90 (lanes 4 and 8) was incubated alone
(lanes 1–4) or with purified CK2 (lanes 5–8) in the presence of
Mg2+-[32P]ATP[cP], and phosphorylated proteins were visualized by

autoradiography after SDS ⁄ PAGE. (C) The same protein mixtures as
shown in (B) were analyzed by western blotting with anti-[pSer13]Cdc37. (D) The same protein mixtures shown in (B) and (C) were
stained with CBB after SDS ⁄ PAGE. The positions of molecular
weight markers, as well as Hsp90, FKBP52, and Cdc37, are shown.

D

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5693


Cdc37 phosphorylation by CK2 and signaling kinases

Y. Miyata and E. Nishida

The amounts of Hsp90, total Cdc37, CK2-phosphorylated Cdc37 and immunoprecipitated kinases were
examined by western blot analysis, using the corresponding antibodies. As shown in Fig. 3, each kinase,
Cdk4 (lane 4), MOK (lane 5), v-Src (lane 6), and Raf1
(lane 7), associated specifically with Hsp90 (Fig. 3A)
and with Cdc37 (Fig. 3B). Importantly, anti-[pSer13]Cdc37 recognized Cdc37 in all Hsp90–client protein
kinase complexes (Fig. 3C), indicating that CK2-phosphorylated Cdc37 was present in these kinase complexes. The control kinases, CK1 (lane 2) and DYRK2
(lane 3), did not bind to Hsp90 (Fig. 3A), Cdc37
(Fig. 3B), or phospho-Cdc37 (Fig. 3C), although comparable amounts of protein were immunoprecipitated
for each kinase, as shown on western blots obtained
using anti-FLAG (Fig. 3D).
Intracellular distribution of CK2-phosphorylated
Cdc37
As shown above, CK2-phosphorylated Cdc37 associates with signaling protein kinases, so intracellular
regions where signaling protein kinases accumulate

might also be expected to contain high concentrations
of phosphorylated Cdc37. Growth factors such as epidermal growth factor (EGF) and insulin-like growth
factor-I are known to induce membrane ruffling, and
actin cytoskeleton and signaling molecules, such as
Rho family G-proteins and protein kinases, are known
to accumulate in these areas. We therefore examined
the intracellular distribution of CK2-phosphorylated
Cdc37 in EGF-stimulated KB cells, a cell line known
to show prominent membrane ruffling in response to
growth factors [34]. KB cells were serum starved by
incubating them in medium containing only 1% fetal
bovine serum for 4 h and then incubated with or without 30 nm EGF for 5 min. Anti-Hsp90, anti-Cdc37
and anti-[pSer13]-Cdc37 were used to immunofluorescently stain the cells, to examine the intracellular distributions of the proteins. Before EGF stimulation,
Hsp90 was mainly localized in the cytoplasm, whereas
Cdc37 and phospho-Cdc37 were present throughout
the KB cells, both in the cytoplasm and in the nucleus
(Fig. 4A–C). EGF induced membrane ruffling and
Hsp90 was localized to the membrane ruffles (Fig. 4D,
arrowheads) as previously reported [35]. Cdc37 also
accumulated in areas of membrane ruffling in EGFstimulated KB cells (Fig. 4E, arrowheads), as did
CK2-phosphorylated Cdc37 (Fig. 4F, arrowheads).
These results indicated that phosphorylated Cdc37
colocalized with Hsp90 in the growth factor-induced
membrane ruffles, where signaling protein kinases also
accumulated, and are consistent with the results in
5694

A

B

C

D

Fig. 3. Association of Ser13-phosphorylated Cdc37 with various
Hsp90 client protein kinases. COS7 cells were transfected with
empty vector DNA (lane 1, control), or plasmids encoding FLAGtagged protein kinases CK1 (lane 2, control), DYRK2 (lane 3, control), Cdk4 (lane 4), MOK (lane 5), v-Src (lane 6), or Raf1 (lane 7),
and the kinase–chaperone complexes were immunopurified. The
amounts of Hsp90 (A), Cdc37 (B), phosphorylated Cdc37 (C) and
protein kinase (D) in the kinase–chaperone complexes were
assessed by western blotting with anti-Hsp90, anti-Cdc37, anti[pSer13]-Cdc37, and anti-FLAG, respectively. The positions of
molecular weight markers are shown.

Fig. 3, showing that phosphorylated Cdc37 forms
complexes with Hsp90 client signaling protein kinases.
We next investigated whether the accumulation of
phosphorylated Cdc37 in the EGF-induced membrane
ruffles was a result of an increase in the phosphorylation of Cdc37 by activated CK2. Serum-starved KB
cells were treated with EGF for up to 60 min, and
the levels of phosphorylated Cdc37 in total cell
extracts were determined by western blotting using

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS


Y. Miyata and E. Nishida

Cdc37 phosphorylation by CK2 and signaling kinases

A

B

C

D

E

F

Fig. 4. Subcellular localization of phosphorylated Cdc37 in growth factor-induced membrane ruffles. KB cells incubated in low-strength serum
(1%) were untreated (A–C) or treated with 30 nM EGF for 5 min (D–F). The intracellular localization of Hsp90 (A, D), total Cdc37 (B, E) or the
phosphorylated form of Cdc37 (C, F) are shown in the left columns by immunofluorescent microscopy. Corresponding phase contrast
images are shown in the right columns. For each panel, two typical images (top and bottom) from different fields are shown.

anti-[pSer13]-Cdc37 (Fig. 5A). The levels of total
Cdc37 (Fig. 5B), phosphorylated extracellular signalregulated kinase (ERK) (Fig. 5C) and total ERK
(Fig. 5D) were also measured by western blotting using
anti-Cdc37, anti-[pTEpY]-ERK (an antibody specific
for the dually phosphorylated, activated form of
ERK), and anti-ERK, respectively. The results clearly
showed that the levels of phosphorylated Cdc37 as well
as total Cdc37 did not change after EGF stimulation
(Fig. 5A,B). The activation of EGF-induced signaling
pathways under these conditions was confirmed by the
observation that EGF rapidly stimulated the dual
phosphorylation of ERK (Fig. 5C). These results agree
with an earlier report showing that growth factors did

not significantly activate CK2 in cells [36]. In addition,
we noted that anti-[pSer13]-Cdc37 recognized a single
protein band in western blots of whole cell lysates
(Fig. 5A), reinforcing our conclusion that anti[pSer13]-Cdc37 specifically recognizes phospho-Cdc37
but not other CK2 substrates (Fig. 2). Taking these
findings together, we concluded that CK2-phosphorylated Cdc37 accumulates in EGF-induced membrane
ruffles as a result of intracellular redistribution of
phospho-Cdc37 rather than a net increase in the phosphorylation of Cdc37 by CK2 within the cell.

Analysis of Cdc37 phosphorylation by
phospho-affinity gel electrophoresis
We wanted to establish a simple biochemical method
for analyzing the phosphorylation of Cdc37. To this
end, we investigated separating phosphorylated Cdc37
from nonphosphorylated Cdc37 by phospho-affinity
gel electrophoresis, a technique recently developed by
Kinoshita et al. [5]. Purified recombinant Cdc37 was
incubated with CK2a or CK2 holoenzyme or alone, in
the presence or absence of ATP and ⁄ or Mg2+, and the
mixtures were analyzed by phospho-affinity gel electrophoresis. CBB staining of the phospho-affinity gel
showed that the mobility of Cdc37 was markedly
decreased when it had been incubated with CK2a or
CK2 holoenzyme in the presence of both ATP and
Mg2+ (Fig. 6A, lanes 6 and 9). This mobility shift was
not a nonspecific effect due to ATP and ⁄ or Mg2+,
because no shift was observed in the absence of CK2
(Fig. 6A, lanes 1–3). Nor was the mobility shift caused
by the physical association of CK2 and Cdc37, as
CK2 did not induce the Cdc37 mobility shift in the
absence of ATP or Mg2+ (Fig. 6A, lanes 4, 5, 7 and

8). The CK2a- or CK2 holoenzyme-dependent phosphorylation of Cdc37 in the presence of Mg2+–ATP

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5695


A

60

30

15

5

EGF

0

Cdc37 phosphorylation by CK2 and signaling kinases

Y. Miyata and E. Nishida

min

62
47
33

1 2 3 4 5
WB: Anti-[pSer13]-Cdc37

B

62
47
33
1 2 3 4 5
WB: Anti-Cdc37

C

62
47
33
1 2 3 4 5
WB: Anti-[pTEpY]-ERK

D

62
47
33
1 2 3 4 5
WB: Anti-ERK

Fig. 5. Effect of EGF treatment on the phosphorylation of
Cdc37. KB cells incubated in low-strength serum (1%) were
untreated (lane 1) or treated with 30 nM EGF for 5 min (lane 2),

15 min (lane 3), 30 min (lane 4), or 60 min (lane 5). Cell extracts
were prepared, and phosphorylated Cdc37 (A), total Cdc37 (B),
dually phosphorylated and activated ERK (C) and total ERK
(D) were labeled on western blots with the corresponding
antibodies.

was confirmed by labeling with anti-[pSer13]-Cdc37
(Fig. 1A). It should be noted that no band shift was
detected when the protein mixtures were analyzed by
normal SDS ⁄ PAGE (Fig. 1B,D).
We then examined the effect of Cdc37 mutations at
the CK2 phosphorylation site on phospho-affinity gel
electrophoresis mobility. Recombinant purified wildtype Cdc37 protein or Ser13 mutants of Cdc37,
5696

Cdc37(13SA) and Cdc37(13SD) were incubated
with CK2a or CK2 holoenzyme in the presence of
Mg2+–ATP and analyzed by phospho-affinity gel
electrophoresis. In contrast to the CK2a- or CK2
holoenzyme-dependent mobility shift shown by wildtype Cdc37 (Fig. 6B, lanes 4 and 7), the mobilities of
Cdc37(13SA) and Cdc37(13SD) were not affected by
CK2a or CK2 (Fig. 6B, lanes 5, 6, 8 and 9), indicating
that the mobility shift of Cdc37 in phospho-affinity gel
electrophoresis was caused by its CK2-dependent phosphorylation on Ser13. In addition, these results showed
that CK2a phosphorylated Cdc37 only on Ser13, as in
the case of CK2 holoenzyme, even in the absence of
CK2b. We therefore used CK2a to phosphorylate
Cdc37 in vitro in subsequent experiments.
Phospho-affinity gel electrophoresis can be combined
with other detection systems, including autoradiography and western blotting. Cdc37 was incubated with

or without CK2a in the presence of Mg2+–ATP, and
separated by phospho-affinity gels. The phospho-affinity gels were washed with a buffer containing EDTA
to remove Mn2+, and thereby remove the phosphatebinding activity of Phos-tag, so that the phosphoproteins in the gel could be transferred to a membrane.
The membrane was probed with anti-Cdc37 or anti[pSer13]-Cdc37. Anti-Cdc37 labeled both the lowmobility and high-mobility Cdc37 bands (Fig. 7A,
lanes 1 and 2), whereas anti-[pSer13]-Cdc37 recognized
only the low-mobility Cdc37 band (Fig. 7B, lane 2),
indicating that the band with the decreased mobility
represented Cdc37 in the Ser13-phosphorylated form.
When we carried out the phosphorylation reaction in
the presence of [32P]ATP[cP], analyzed the mixtures by
phospho-affinity gel electrophoresis and autoradiographed the gels, radioactivity was detected only in the
low-mobility band using wild-type Cdc37 (Fig. 7C,
lane 3). Neither the mobility shift nor the radioactivity
could be detected when CK2a was omitted (Fig. 7C,D,
lane 2) or when the phosphorylation site mutant
Cdc37(13SA) was used (Fig. 7C,D, lane 4). We therefore concluded that CK2 phosphorylates only one
residue, Ser13, in Cdc37, and that this single phosphorylation causes the Cdc37 mobility shift seen in
phospho-affinity gel electrophoresis.
Analysis of Cdc37 phosphorylation and
dephosphorylation by phospho-affinity gel
electrophoresis
To analyze the time course of Cdc37 phosphorylation
in vitro, mixtures of recombinant Cdc37 and CK2a, in
the presence of Mg2+–ATP, were sampled at different
time points, and the samples were analyzed by

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS


Y. Miyata and E. Nishida

A

B

Fig. 6. Analysis of Cdc37 phosphorylation by phospho-affinity gel
electrophoresis. (A) Recombinant Cdc37 alone (lanes 1–3), with
CK2a (lanes 4–6) or with CK2 holoenzyme (lanes 7–9), in the presence (+) or absence (–) of Mg2+ and ⁄ or ATP as indicated, was incubated at 30 °C for 30 min, analyzed by phospho-affinity gel
electrophoresis, and then stained with CBB. (B) Wild-type protein
(lanes 1, 4 and 7) and two Cdc37 mutant proteins, 13SA (lanes 2, 5
and 8) and 13SD (lanes 3, 6 and 9), were incubated alone
(lanes 1–3), with CK2a (lanes 4–6) or with CK2 holoenzyme
(lanes 7–9) in the presence of Mg2+–ATP for 30 min at 30 °C. The
phosphorylation mixtures were analyzed by phospho-affinity gel
electrophoresis and stained with CBB.

phospho-affinity gel electrophoresis. Phosphorylation
was rapid and detected within 2 min (Fig. 8A, lane 4)
and completed within 6 min at 30 °C (Fig. 8A,

Cdc37 phosphorylation by CK2 and signaling kinases

lanes 4–7). The transition in mobility from the higher
to lower bands was direct, with no bands of intermediate mobility being detected, supporting our previous
conclusion that CK2a phosphorylates only one site in
Cdc37.
We then studied the time course of dephosphorylation by incubating CK2a-phosphorylated Cdc37
(Fig. 8B, lane 2) with (Fig. 8B, lanes 7–10) or without
(Fig. 8B, lanes 3–6) k-phosphatase in the presence of
the CK2 inhibitor 4,5,6,7-tetrabromobenzotriazole

(TBB) and analyzing the products by phospho-affinity
gel electrophoresis and western blotting using antiCdc37. Incubating phospho-Cdc37 with k-phosphatase
induced rapid dephosphorylation, with a high-mobility
Cdc37 band appearing within 3 min (Fig. 8B, lane 8).
Again, the band shift induced was direct, from the
low-mobility to high-mobility bands, with no intermediate bands being detected.
Analysis of signaling kinase–Hsp90–Cdc37 complexes by phospho-affinity gel electrophoresis
Finally, we analyzed complexes between signaling protein kinases and Hsp90–Cdc37 molecular chaperones
by phospho-affinity gel electrophoresis. The Hsp90 client kinases Cdk4, MOK, v-Src and Raf1 were
expressed as FLAG-tagged fusion proteins in COS7
cells, and the kinase–chaperone complexes were immunopurified using anti-FLAG agarose. The immunocomplexes were separated by phospho-affinity gel
electrophoresis, transferred to western blots, and
labeled with anti-Cdc37 or anti-[pSer13]-Cdc37. Specific associations between Cdc37 and Cdk4, MOK,
v-Src and Raf1 were observed (Fig. 9A), consistent
with the result shown in Fig. 3. Interestingly, Cdc37 in
the protein kinase complexes was detected as a single
band in phospho-affinity gels (Fig. 9A), and the band
was recognized by anti-[pSer13]-Cdc37 (Fig. 9B),
indicating for the first time that all of the Cdc37 in the
signaling kinase–Hsp90 complexes was in its Ser13phosphorylated form in vivo.

Discussion
In this study, we have produced a phospho-specific
antibody against Cdc37, which recognized recombinant
purified Cdc37 only when incubated with CK2 in the
presence of Mg2+ and ATP. The specificity of this
antibody was demonstrated by showing that it did not
recognize mutant Cdc37 in which the CK2 phosphorylation site, Ser13, had been replaced with nonphosphorylatable amino acids, and it did not recognize other
CK2-phosphorylated proteins such as Hsp90 and

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5697


Cdc37 phosphorylation by CK2 and signaling kinases

A

Y. Miyata and E. Nishida

B

C

D

Fig. 7. Characterization of Cdc37 phosphorylation by phospho-affinity gel electrophoresis. Recombinant Cdc37 was phosphorylated with
CK2a and analyzed by phospho-affinity gel electrophoresis. (A) Non-phosphorylated (lane 1) and phosphorylated (lane 2) Cdc37 were separated by phospho-affinity gel electrophoresis and analyzed by western blotting with anti-Cdc37, for total Cdc37. (B) The same membrane as
in (A) was stripped and reprobed with anti-[pSer13]-Cdc37. (C) Wild-type protein [Cdc37(WT)] (lane 3) and a mutant [Cdc37(13SA)] (lane 4) of
Cdc37 were phosphorylated with CK2a in the presence of [32P]ATP[cP] and analyzed by phospho-affinity gel electrophoresis followed by
autoradiography. As controls, CK2a alone (lane 1) or Cdc37(WT) alone (lane 2) was incubated under the same conditions. (D) CBB staining of
the same gel as in (C).

FKBP52. Thus, the antibody specifically recognizes
Cdc37 only when it is phosphorylated on Ser13 by
CK2. Using this antibody, we have shown that complexes between Hsp90 and its client signaling kinases
Cdk4, MOK, v-Src and Raf1 contain CK2-phosphorylated Cdc37 in vivo. These results are consistent with
previous reports that CK2 phosphorylates Cdc37 on
Ser13 both in vivo and in vitro, and that this phosphorylation is essential for the binding of Cdc37 to multiple

Hsp90 client protein kinases [25,26,28].
Immunofluorescence staining with the antibody
showed that phosphorylated Cdc37 accumulated in
growth factor-induced membrane ruffles in KB cells.
This accumulation was the result of a redistribution of
phospho-Cdc37 within the cells rather than an upregulation of total Cdc37 phosphorylation by CK2 after
EGF stimulation. Previously, Hsp90 has been reported
to bind to polymerized actin and to accumulate in
growth factor-induced membrane ruffles, where actin
filaments are abundant [35]. In fact, Hsp90 has been
suggested to be involved in the intracellular distribution and trafficking of many signaling molecules by
interacting with its client proteins and the cytoskeletal
architecture [37,38]. Membrane ruffling is one of the
morphological responses rapidly induced in cells by
growth factor stimulation, and the accumulation of a
5698

variety of signaling molecules, including receptor tyrosine kinases and G-proteins, in membrane ruffles has
been reported [39]. This might be why Hsp90 and
phosphorylated Cdc37 accumulate in membrane ruffles, as only the phosphorylated form of Cdc37 is
active in recruiting Hsp90 to its client signaling kinases
[26]. The Hsp90–Cdc37–kinase complexes might then
interact with the actin cytoskeleton via Hsp90 in the
membrane ruffling region. Membrane ruffling has been
related to the metastatic status of tumor cells, and it
has been suggested as an indicator of tumor cell motility and metastatic potential [40]. Therefore, the inhibition of Hsp90, Cdc37 or CK2 might influence, not
only signaling affecting cell growth, but also the metastasis of neoplastic cells.
In this study, we exploited the recently developed
compound Phos-tag, which has specific phosphatebinding activity [4,5], to separate phosphorylated and
nonphosphorylated forms of Cdc37 by phospho-affinity gel electrophoresis. We have demonstrated a specific mobility shift in Cdc37 only when it is incubated

with CK2 in the presence of Mg2+ and ATP. Replacing the CK2 phosphorylation site in Cdc37 with nonphosphorylatable amino acids completely abolished
this mobility shift. The low-mobility Cdc37 band was
both the only band recognized by the phospho-specific

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS


Y. Miyata and E. Nishida

A

B

Fig. 8. Time course of Cdc37 phosphorylation and dephosphorylation analyzed by phospho-affinity gel electrophoresis. (A) Time
course of CK2-dependent Cdc37 phosphorylation was analyzed by
phospho-affinity gel electrophoresis, between 0 min (lane 1) and
10 min (lane 9). (B) Time course of the dephosphorylation of CK2phosphorylated Cdc37 by k-phosphatase was analyzed by phosphoaffinity gel electrophoresis. Lane 1, nonphosphorylated Cdc37;
lane 2, Cdc37 phosphorylated with CK2a. Phosphorylated Cdc37,
as in lane 2, was incubated with (lanes 7–10) or without (lanes 3–6)
k-phosphatase in the presence of the CK2 inhibitor TBB for up to
20 min. The mixtures were analyzed by phospho-affinity gel electrophoresis followed by western blotting with anti-Cdc37. Samples
were analyzed after dephosphorylation for 0 min (lanes 3 and 7),
3 min (lanes 4 and 8), 10 min (lanes 5 and 9), and 20 min (lanes 6
and 10).

antibody against Cdc37 and the only band that
became radioactive after incubation with CK2 in the
presence of [32P]ATP[cP]. Using this new technique,
we have been able to strengthen our previous conclusion that CK2 rapidly phosphorylates Cdc37 at only
one site, Ser13, and to demonstrate for the first time

that only phosphorylated Cdc37 is present in signaling
protein kinase complexes with Hsp90. These results
further support the previous proposal that the phosphorylation of Cdc37 by CK2 is essential for its protein kinase-binding activity [25,26,28]. Future studies
will be needed to elucidate whether the phosphorylation of Cdc37 fluctuates dynamically within cells
according to cellular conditions.

Cdc37 phosphorylation by CK2 and signaling kinases

Phospho-affinity gel electrophoresis is a simple and
easy method with the advantage that it uses no radioactive material. Moreover, it is a technique that could be
used for the analysis of virtually any phosphoprotein, as
a phosphorylation-dependent mobility shift in these
gels could occur for any protein. It should be noted,
however, that large excesses of metal chelators such as
EDTA and chemicals containing phosphate moieties,
including sodium phosphate, b-glycerophosphate or
ATP, may interfere with the assay. In addition, crude
protein mixtures such as tissue extracts that may contain
a large amount of phosphoproteins may not be easy to
analyze directly with this technique. Further improvements in phospho-affinity gel electrophoresis might be
possible and might broaden its applicability.
Previous studies have identified many substrates for
CK2 [21], and demonstrated its involvement in many
different cellular functions, including cell division, cell
survival, and gene expression [19,20,24]. Surprisingly,
however, the regulatory mechanism of CK2 in cells
remains largely unknown. Observations that CK2 activity was enhanced by growth factors have been challenged [36], and this discrepancy arose partly from the
lack of a reliable method for quantifying CK2 activity
in vivo. Our results agree with those of previous studies
[25,26,28] in demonstrating that CK2 phosphorylates

only one site in Cdc37, so the phosphorylation state of
Cdc37 should be an index of CK2 activity both in vivo
and in vitro. The phosphorylation of Cdc37 can now be
readily monitored using the phospho-specific antibody
and phospho-affinity gel electrophoresis, as described
here. Several lines of evidence suggest a critical role for
CK2 in tumorigenesis [22]. For example, CK2 is tumorigenic when overexpressed in a transgenic mouse [41]. As
the CK2–Cdc37 module has been suggested to function
as a master switch with a positive feedback mechanism
for various signaling protein kinases [25,31], the development of a simple method for determining CK2 activity and Cdc37 phosphorylation is likely to be both
biologically and clinically important in the future.

Experimental procedures
Plasmids, proteins and antibodies
Plasmids that encode the FLAG-tagged protein kinases
Cdk4, MOK, v-Src and Raf1 for expression in mammalian
cells have been described previously [26]. cDNA for Zebrafish CK1a was a kind gift from J. E. Allende (Universidad
de Chile), and an EcoRI fragment of the coding region was
inserted into an EcoRI site of pCMV–Tag2B to obtain an
expression plasmid encoding FLAG-tagged CK1. An
expression plasmid for FLAG-tagged DYRK2 used as a

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5699


Cdc37 phosphorylation by CK2 and signaling kinases

A

Y. Miyata and E. Nishida

B

Fig. 9. Analysis of the signaling protein
kinase–chaperone complexes by phosphoaffinity gel electrophoresis. Hsp90 client kinases Cdk4 (lane 2), MOK (lane 3), v-Src
(lane 4) and Raf1 (lane 5) were expressed in
COS7 cells and immunopurified. The protein
kinase complexes were analyzed by phospho-affinity gel electrophoresis and western
blotting with anti-Cdc37 (A) or anti-[pSer13]Cdc37 (B). COS7 cells transfected with
DNA from the empty vector were used as a
control (lane 1). The position of Cdc37 in the
gel is shown on the right.

control will be described elsewhere. A dual expression plasmid (CK2a + Cdc37 ⁄ pETDuet-1b) that encodes His6–
CK2a and Cdc37 has been described previously [27]. A
BamHI fragment encoding full-length human Hsp90a,
described previously [27], was inserted into a BamHI site in
pGEX6P2 to obtain a bacterial expression plasmid for
human Hsp90a. Wild-type [Cdc37(WT)] and mutant
[Cdc37(13SA) and Cdc37(13SD)] recombinant rat Cdc37
proteins were expressed as glutathione S-transferase (GST)
fusion proteins in Escherichia coli and purified as described
previously [26]. Native CK2 holoenzyme (a2b2) was purified
from porcine testes as described previously [30]. The expression and purification of recombinant FKBP52 has been
described previously [32]. EGF was purchased from BD
Biosciences (San Jose, CA, USA) and k-phosphatase from
Upstate Biotechnologies (Lake Placid, NY, USA).
Horseradish peroxidase (HRP)-conjugated anti-Cdc37

(E-4) was purchased from Santa Cruz (Santa Cruz, CA,
USA), anti-FLAG (M2) and anti-FLAG (M2) affinity gel
from Sigma (Saint Louis, MO, USA), anti-ERK2 (clone
D-2) from Santa Cruz, and antibody specific for the dually
phosphorylated (pTEpY) ERK (V803A) from Promega
(Madison, WI, USA). Anti-Hsp90 has been described previously [35]. Antibody specific for phosphorylated Cdc37
was raised in rabbits against a phospho-peptide,
VWDHIEVpSDDEDETHC, corresponding to amino acids
6–20 of Cdc37 with an additional C-terminal cysteine for
conjugation. Rabbit antisera were affinity purified using the
phosphopeptide, and then immunoadsorbed with an affinity
resin conjugated with the nonphosphorylated form of the
peptide to remove antibodies reactive to this nonphosphorylated form. The peptides were synthesized by MBL Co.
(Nagoya, Japan), and immunizations and affinity purification were performed at SCRUM Inc. (Tokyo, Japan).

5700

Expression and purification of Hsp90–client
protein kinase complexes
COS7 cells were cultured and transfected with mammalian expression vectors by electroporation, and cell
extracts were prepared in an extraction buffer (50 mm
Tris ⁄ HCl, 10% glycerol, 100 mm NaF, 50 mm NaCl,
2 mm EDTA, 2 mm sodium orthovanadate, 10 mm
sodium pyrophosphate, 1 mm dithiothreitol, 1% Nonidet
P-40, pH 8.0) as described previously [32]. Extracts containing equal amounts of protein were incubated with
40 lL of anti-FLAG affinity gel for 12 h at 4 °C. The
immunocomplexes were extensively washed with the
extraction buffer, and then shaken with 0.33 mgỈmL)1
(final concentration) 3 · FLAG peptide (Sigma) in 50 lL
of extraction buffer at 4 °C for 2 h. Proteins eluted from

anti-FLAG affinity gel were collected by brief centrifugation (13 000 g for 5 min at 2 °C using an Eppendorf
5415R centrifuge and wide angle rotor), and then treated
in Amicon 0.22 lm Ultrafree centrifugal filtration tubes
(Millipore, Bedford, MA, USA). For phospho-affinity gel
electrophoresis analysis, Tris-buffered saline (50 mm
Tris ⁄ HCl, 150 mm NaCl, pH 7.4) was used for washing
and elution instead of extraction buffer.

Expression and purification of CK2a
Introduction
of
the
dual
expression
plasmid
CK2a + Cdc37 ⁄ pETDuet-1b into the host E .coli BL21
CodonPlus (DE3) RIL (Stratagene, La Jolla, CA, USA)
and induction of the expression of CK2a have been
described previously [27]. E. coli cell pellets were collected
from 500 mL cultures and solubilized in 35 mL of
B-PER bacterial extraction solution (Pierce, Rockford,

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS


Y. Miyata and E. Nishida

IL, USA) supplemented 1 : 100 (v ⁄ v) with an EDTA-free
protease inhibitor cocktail (Sigma). After the addition of
3 m NaCl to adjust the final concentration of NaCl to

700 mm, the mixture was gently agitated for 20 min at
2 °C using an Avanti HP-25. The mixture was centrifuged at 18 000 g for 15 min at 2 °C, and the supernatant was filtered through a MillexHV 0.45 lm filter unit
(Millipore). His6-tagged CK2a was purified on a 10 mL
column of Talon beads (Clontech-Takara BIO, Mountain
View, CA, USA) with a linear gradient of 0–500 mm
imidazole in 50 mm sodium phosphate and 300 mm NaCl
(pH 7.0). Fractions containing CK2a were concentrated
using an Amicon Ultra-15 Centrifugal Filter Device
(Millipore). CK2a was further purified on a Superdex 200
HR10 ⁄ 30 gel filtration column (GE Healthcare Biosciences, Uppsala, Sweden) with a buffer containing 50 mm
Tris, 1 mm EDTA, 1 mm dithiothreitol, and 200 mm
NaCl (pH 7.4). Fractions containing CK2a were further
purified using a ResourceQ column (GE Healthcare Biosciences) and a linear gradient of 0–1000 mm NaCl in
50 mm Tris, 1 mm EDTA, and 1 mm dithiothreitol
(pH 9.0). Purified CK2a was dialyzed in the same buffer
containing 200 mm NaCl. This recombinant CK2a was
more than 98% pure and ran as a single band on
SDS ⁄ PAGE stained with CBB.

Expression and purification of human Hsp90a
The bacterial expression plasmid encoding GST-tagged
human Hsp90a (Hsp90a ⁄ pGEX6P2) was introduced into
a host E. coli strain BL21 CodonPlus (DE3) RIL. After
reaching an A600 nm level of 0.7, 0.1 mm (final concentration) isopropyl-1-thio-b-d-galactopyranoside was added to
induce expression, and the culture was shaken for 4 h at
30 °C. A frozen E. coli cell pellet from a 500 mL culture
was solubilized in 40 mL of B-PER solution (Pierce) supplemented 1 : 100 (v ⁄ v) with a bacterial protease inhibitor
cocktail (Sigma). Purification of GST–Hsp90a and cleavage with PreScission Protease (GE Healthcare Biosciences) to remove the GST moiety from the fusion protein
were performed essentially as described previously [11].
Hsp90a was further purified by ResourceQ column chromatography using a linear gradient of 0–1000 mm NaCl

in 50 mm Tris ⁄ HCl, 1 mm EDTA, and 1 mm dithiothreitol (pH 7.4). This recombinant Hsp90a was more than
95% pure and ran as a single band on SDS ⁄ PAGE
stained with CBB.

Phosphorylation with CK2 and dephosphorylation with k-phosphatase
Recombinant purified Cdc37, FKBP52 or Hsp90 were
shaken with purified native CK2 holoenzyme or with
recombinant purified CK2a at 30 °C for 30 min (or for
the indicated time periods). The standard phosphorylation

Cdc37 phosphorylation by CK2 and signaling kinases

mixture was 85 mm Tris, 6 mm Hepes, 100 mm NaCl,
10 mm MgCl2, 0.2 mm EDTA, 0.2 mm dithiothreitol,
7.5 mm ATP (or 0.75 mm ATP including [32P]ATP[cP] for
radioactive assays), 0.5 mgỈmL)1 substrate protein, and
0.096 mgỈmL)1 CK2 holoenzyme or CK2a (pH 7.5). In
the experiments shown in Figs 7 and 8, the phosphorylation mixture contained 265 mm NaCl. Native CK2, CK2a,
ATP and ⁄ or MgCl2 were omitted from the incubation
mixtures where indicated. For dephosphorylation, k-phosphatase was added and incubated for the indicated time
periods at 37 °C in the presence of a specific CK2 inhibitor, TBB. The final incubation buffer for phosphatase
treatment was 70 mm Tris, 57 mm Hepes, 220 mm NaCl,
8 mm MgCl2, 2 mm MnCl2, 0.26 mm EDTA, 5 mm dithiothreitol, 6.1 mm ATP, 20 lm TBB, 2% dimethylsulfoxide, 0.4 mgỈmL)1 Cdc37, 0.08 mgỈmL)1 CK2a, and
0.4 unitỈlL)1 k-phosphatase. Phosphorylation and dephosphorylation reactions were stopped by adding 1 : 3 (v ⁄ v)
SDS sample buffer followed by incubation at 98 °C for
5 min.

Immunofluorescent staining
KB cells were cultured on glass coverslips in DMEM supplemented with 10% fetal bovine serum. To serum starve
cells, the medium was replaced with DMEM + 1% fetal

bovine serum for 4 h. EGF was added to cultures at a final
concentration of 30 nm, and cells were incubated at 37 °C
for 5 min to induce membrane ruffling. Cells were fixed
with formaldehyde, permeabilized, and stained with antibodies as described previously [27].

Phospho-affinity polyacrylamide gel
electrophoresis
Phos-tag acrylamide was purchased from the Phos-Tag
Consortium (), and 5 mm stock
solutions in water were prepared. Phospho-affinity gel
electrophoresis was performed essentially as described
previously [5]. The final composition of phospho-affinity
SDS-polyacrylamide gels was 7.8% acrylamide, 0.21% bisacrylamide, 375 mm Tris ⁄ HCl, pH 8.8, 0.1% SDS, 0.05%
ammonium persulfate, 0.086% TEMED containing 50 lm
Phos-tag acrylamide, and 100 lm MnCl2. The stacking gel
solution used was 3.9% acrylamide, 0.11% bis-acrylamide,
124 mm Tris ⁄ HCl, pH 6.8, 0.1% SDS, 0.1% ammonium
persulfate, and 0.1% TEMED. The electrophoresis running
buffer used was 25 mm Tris, 192 mm glycine, and 0.1%
SDS (pH 8.3). Heat denaturation with SDS, sample loading
and CBB staining procedures were the same as for ordinary
SDS ⁄ PAGE. The final concentration of NaCl in the samples was adjusted to 265 mm before heat denaturation.
Electrophoresis was stopped when a prestained 40 kDa
protein marker reached the front edge of the gels, to give
clear band separation.

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5701

Cdc37 phosphorylation by CK2 and signaling kinases

Y. Miyata and E. Nishida

Other procedures
Normal SDS ⁄ PAGE was performed with 10% acrylamide
gels. Western blotting was performed with HRP-conjugated
secondary antibodies or peroxidase-conjugated antibodies
and a chemiluminescent detection system as described previously [26]. For western blotting after phospho-affinity gel
electrophoresis, gels were washed with 1 mm EDTA in
transfer buffer (25 mm Tris, 192 mm glycine, 0.02% SDS,
20% methanol) for 15 min and then in transfer buffer without EDTA for 15 min at room temperature before proteins
were transferred to poly(vinylidene difluoride) membranes.
For western blotting with phospho-specific antibodies, the
poly(vinylidene difluoride) membranes were treated with
Blocking One-P Solution (Nacalai Tesque, Kyoto, Japan),
instead of 5% skimmed milk, after transferring proteins to
the membranes.

Acknowledgements
We thank T. Sakabe for excellent technical assistance,
and Dr E. Kinoshita (Hiroshima University) for helpful
suggestions. This work was supported by Grants-in-Aid
for Scientific Research from the Ministry of Education,
Culture, Sports, Science and Technology of Japan.

References
1 Cohen P (2001) The role of protein phosphorylation in
human health and disease. The Sir Hans Krebs Medal

Lecture. Eur J Biochem 268, 5001–5010.
2 Olsen JV, Blagoev B, Gnad F, Macek B, Kumar C,
Mortensen P & Mann M (2006) Global, in vivo, and
site-specific phosphorylation dynamics in signaling networks. Cell 127, 635–648.
3 Rush J, Moritz A, Lee KA, Guo A, Goss VL, Spek EJ,
Zhang H, Zha XM, Polakiewicz RD & Comb MJ
(2005) Immunoaffinity profiling of tyrosine phosphorylation in cancer cells. Nat Biotechnol 23, 94–101.
4 Kinoshita E, Takahashi M, Takeda H, Shiro M &
Koike T (2004) Recognition of phosphate monoester
dianion by an alkoxide-bridged dinuclear zinc(II)
complex. Dalton Trans 1189–1193.
5 Kinoshita E, Kinoshita-Kikuta E, Takiyama K & Koike
T (2006) Phosphate-binding tag, a new tool to visualize
phosphorylated proteins. Mol Cell Proteomics 5, 749–
757.
6 Stancato LF, Chow Y-H, Hutchison KA, Perdew GH,
Jove R & Pratt WB (1993) Raf exists in a native heterocomplex with hsp90 and p50 that can be reconstituted
in a cell-free system. J Biol Chem 268, 21711–21716.
7 Grammatikakis N, Lin JH, Grammatikakis A, Tsichlis
PN & Cochran BH (1999) p50cdc37 acting in concert
with Hsp90 is required for Raf-1 function. Mol Cell Biol
19, 1661–1672.

5702

8 Dai K, Kobayashi R & Beach D (1996) Physical interaction of mammalian CDC37 with CDK4. J Biol Chem
271, 22030–22034.
9 Lamphere L, Fiore F, Xu X, Brizuela L, Keezer S,
Sardet C, Draetta GF & Gyuris J (1997) Interaction
between Cdc37 and Cdk4 in human cells. Oncogene 14,

1999–2004.
10 Stepanova L, Leng X, Parker SB & Harper JW (1996)
Mammalian p50Cdc37 is a protein kinase-targeting subunit of Hsp90 that binds and stabilizes Cdk4. Genes Dev
10, 1491–1502.
11 Miyata Y, Ikawa Y, Shibuya M & Nishida E (2001) Specific association of a set of molecular chaperones including HSP90 and Cdc37 with MOK, a member of the
MAP kinase superfamily. J Biol Chem 276, 21841–21848.
12 Chen G, Cao P & Goeddel DV (2002) TNF-induced
recruitment and activation of the IKK complex require
Cdc37 and Hsp90. Mol Cell 9, 401–410.
13 Perdew GH, Wiegand H, Vanden Heuvel JP, Mitchell C
& Singh SS (1997) A 50 kilodalton protein associated
with raf and pp60v-src protein kinases is a mammalian
homolog of the cell cycle control protein cdc37. Biochemistry 36, 3600–3607.
14 Young JC, Moarefi I & Hartl FU (2001) Hsp90: a specialized but essential protein-folding tool. J Cell Biol
154, 267–273.
15 Prodromou C & Pearl LH (2003) Structure and functional relationships of Hsp90. Curr Cancer Drug Targets
3, 301–323.
16 Picard D (2002) Heat-shock protein 90, a chaperone for
folding and regulation. Cell Mol Life Sci 59, 1640–1648.
17 Pearl LH (2005) Hsp90 and Cdc37 ) a chaperone cancer conspiracy. Curr Opin Genet Dev 15, 55–61.
18 MacLean M & Picard D (2003) Cdc37 goes beyond
Hsp90 and kinases. Cell Stress Chaperones 8, 114–119.
19 Allende JE & Allende CC (1995) Protein kinases. 4.
Protein kinase CK2: an enzyme with multiple substrates
and a puzzling regulation. FASEB J 9, 313–323.
20 Litchfield DW (2003) Protein kinase CK2: structure,
regulation and role in cellular decisions of life and
death. Biochem J 369, 1–15.
21 Meggio F & Pinna LA (2003) One-thousand-and-one
substrates of protein kinase CK2? FASEB J 17, 349–368.

22 Xu X, Landesman-Bollag E, Channavajhala PL &
Seldin DC (1999) Murine protein kinase CK2: gene and
oncogene. Mol Cell Biochem 191, 65–74.
23 Pinna LA (2003) The raison d’etre of constitutively
active protein kinase: the lesson of CK2. Acc Chem Res
36, 378–384.
24 Pinna LA (2002) Protein kinase CK2: a challenge to
canons. J Cell Sci 115, 3873–3878.
25 Bandhakavi S, McCann RO, Hanna DE & Glover CVC
(2003) A positive feedback loop between protein kinase
CKII and Cdc37 promotes the activity of multiple protein kinases. J Biol Chem 278, 2829–2836.

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS


Y. Miyata and E. Nishida

26 Miyata Y & Nishida E (2004) CK2 controls multiple
protein kinases by phosphorylating a kinase-targeting
molecular chaperone Cdc37. Mol Cell Biol 24, 4065–
4074.
27 Miyata Y & Nishida E (2005) CK2 binds, phosphorylates, and regulates its pivotal substrate Cdc37, an
Hsp90-cochaperone. Mol Cell Biochem 274, 171–179.
28 Shao J, Prince T, Hartson SD & Matts RL (2003) Phosphorylation of serine 13 is required for the proper function of the Hsp90 co-chaperone, Cdc37. J Biol Chem
278, 38117–38220.
29 Kimura Y, Rutherford SL, Miyata Y, Yahara I, Freeman BC, Yue L, Morimoto RI & Lindquist S (1997)
Cdc37 is a molecular chaperone with specific functions
in signal transduction. Genes Dev 11, 1775–1785.
30 Miyata Y & Yahara I (1992) The 90-kDa heat shock
protein, HSP90, binds and protects casein kinase II

from self-aggregation and enhances its kinase activity.
J Biol Chem 267, 7042–7047.
31 Miyata Y & Nishida E (2004) Supervision of multiple
signaling protein kinases by the CK2–Cdc37 couple, a
possible novel cancer therapeutic target. Ann NY Acad
Sci 1030, 150–157.
32 Miyata Y, Chambraud B, Radanyi C, Leclerc J,
Lebeau M-C, Renoir J-M, Shirai R, Catelli M-G,
Yahara I & Baulieu E-E (1997) Phosphorylation of the
immunosuppressant FK506-binding protein FKBP52
by casein kinase II (CK2): regulation of HSP90-binding activity of FKBP52. Proc Natl Acad Sci USA 94,
14500–14505.
33 Lees-Miller SP & Anderson CW (1989) Two human 90kDa heat shock proteins are phosphorylated in vivo at

Cdc37 phosphorylation by CK2 and signaling kinases

34

35

36

37

38

39
40

41

conserved serines that are phosphorylated in vitro by
casein kinase II. J Biol Chem 264, 2431–2437.
Kadowaki T, Koyasu S, Nishida E, Sakai H, Takaku
F, Yahara I & Kasuga M (1986) Insulin-like growth
factors, insulin, and epidermal growth factor cause
rapid cytoskeletal reorganization in KB cells. Clarification of the roles of type I insulin-like growth factor
receptors and insulin receptors. J Biol Chem 261,
16141–16147.
Koyasu S, Nishida E, Kadowaki T, Matsuzaki F, Iida
K, Harada F, Kasuga M, Sakai H & Yahara I (1986)
Two mammalian heat shock proteins, HSP90 and
HSP100, are actin-binding proteins. Proc Natl Acad Sci
USA 83, 8054–8058.
Litchfield DW, Dobrowolska G & Krebs EG (1994)
Regulation of casein kinase II by growth factors: a
reevaluation. Cell Mol Biol Res 40, 373–381.
Pratt WB, Silverstein AM & Galigniana MD (1999) A
model for the cytoplasmic trafficking of signalling proteins involving the hsp90-binding immunophilins and
p50cdc37. Cell Signal 11, 839–851.
Pratt WB & Toft DO (2003) Regulation of signaling protein function and trafficking by the hsp90 ⁄ hsp70-based
chaperone machinery. Exp Biol Med 228, 111–133.
Ridley AJ (1994) Membrane ruffling and signal transduction. Bioessays 16, 321–327.
Jiang WG (1995) Membrane ruffling of cancer cells: a
parameter of tumour cell motility and invasion. Eur
J Surg Oncol 21, 307–309.
Seldin DC & Leder P (1995) Casein kinase II a transgene-induced murine lymphoma: relation to theileriosis
in cattle. Science 267, 894–897.

FEBS Journal 274 (2007) 5690–5703 ª 2007 The Authors Journal compilation ª 2007 FEBS

5703