CDC25A-inhibitory RE derivatives bind to pocket adjacent to the catalytic site

Article abstract of DOI:10.1039/c3mb00003f

RE derivatives, which are cell-permeable and non-electrophilic dual-specificity protein phosphatase inhibitors developed in our laboratory, inhibit CDC25A/B non-competitively, as determined by means of kinetic experiments. To identify the binding site of RE derivatives, we designed and synthesized the new probe molecule RE142, having a Michael acceptor functionality for covalent bond formation with the enzyme, a biotin tag to enable enrichment of probe-bound peptide(s), and a chemically cleavable linker to facilitate release of probe-bound peptides from avidin beads. LC-MS analysis indicated that RE142 binds to one of the residues Cys384-Tyr386 of CDC25A, within a pocket adjacent to the catalytic site.

Full text of DOI:10.1039/c3mb00003f

View Article Online
View Journal
Molecular
Biosystems
Accepted Manuscript
This is an Accepted Manuscript, which has been through the RSC Publishing peer
review process and has been accepted for publication.
Accepted Manuscripts are published online shortly after acceptance, which is prior
to technical editing, formatting and proof reading. This free service from RSC
Publishing allows authors to make their results available to the community, in
citable form, before publication of the edited article. This Accepted Manuscript will
be replaced by the edited and formatted Advance Article as soon as this is available.
www.molecularbiosystems.org
Volume 6 | Number 1 | January 2010 | Pages 1–276
To cite this manuscript please use its permanent Digital Object Identifier (DOI®),
which is identical for all formats of publication.
More information about Accepted Manuscripts can be found in the
Information for Authors.
Please note that technical editing may introduce minor changes to the text and/or
graphics contained in the manuscript submitted by the author(s) which may alter
content, and that the standard Terms & Conditions and the ethical guidelines
that apply to the journal are still applicable. In no event shall the RSC be held
responsible for any errors or omissions in these Accepted Manuscript manuscripts or
any consequences arising from the use of any information contained in them.
ISSN 1742-206X
METHOD
Thomas Kodadek et al.
Rapid identification of orexin receptor
PAPER
Dieter Willbold et al.
Competitively selected protein ligands binding ligands using cell-based
pay their increase in specificity by a
decrease in affinity
screening accelerated with magnetic
beads
1742-206X(2010)6:1;1-G
www.rsc.org/molecularbiosystems
Registered Charity Number 207890

Page 1 of 9
Molecular BioSystems
View Article Online
CDC25s inhibitors, RE derivatives, binds to pocket adjacent to catalytic active site, which was
revealed by LC-MS analysis using originally developed chemical probes.

View Article Online
Molecular BioSystems
Page 2 of 9
►
Molecular BioSystems
CDC25A-inhibitory RE Derivatives Bind to Pocket Adjacent to the
Catalytic Site
Ayako Tsuchiya,a Miwako Asanuma,b Go Hirai,a Kana Oonuma,a Muhammad Muddassar,a Eri
Nishizawa,a Yusuke Koyama,a Yuko Otani,a Kam Y. J. Zhang,a and Mikiko Sodeoka*a,b
5
Received (in XXX, XXX) Xth XXXXXXXXX 20XX, Accepted Xth XXXXXXXXX 20XX
DOI: 10.1039/b000000x
RE derivatives, which are cell-permeable and non-electrophilic dual-specificity protein phosphatase
inhibitors developed in our laboratory, inhibit CDC25A/B non-competitively, as determined by means of
kinetic experiments. To identify the binding site of RE derivatives, we designed and synthesized the new
10 probe molecule RE142, having a Michael acceptor functionality for covalent bond formation with the
enzyme, a biotin tag to enable enrichment of probe-bound peptide(s), and a chemically cleavable linker to
facilitate release of probe-bound peptides from avidin beads. LC-MS analysis indicated that RE142 binds
to one of the residues Cys384-Tyr386 of CDC25A, within a pocket adjacent to the catalytic site.
potent anti-proliferative activity towards various cancer cell lines.
However, isoform-selective inhibitors are also required as tools
for chemical biology research, e.g., for analysing the
50 characteristics of the individual CDC25 isoforms.6
Introduction
15 Cell division cycle 25 molecules (CDC25s) are members of the
dual-specificity protein phosphatase (DSP) family of enzymes
that dephosphorylate both phospho-serine/threonine and
phospho-tyrosine residues in the same protein. CDC25s
dephosphorylate conserved phospho-tyrosine and phospho-
20 threonine residues on cyclin-dependent kinases (CDKs), thereby
activating the kinases and promoting cell cycle progression and
transition.1 In addition to these functions, CDC25s play an
important role as checkpoint regulators for handling DNA
damage caused by UV light, ionizing irradiation, or chemicals.
25 Therefore, genomic instability is likely to be triggered by
misregulation of CDC25s. Mammalian cells express three
CDC25 isoforms, CDC25A, CDC25B, and CDC25C. It is
believed that all three isoforms contribute to regulation of the
activities of CDKs during all phases of the cell cycle, though
30 specific functions of each enzyme at different phases have been
reported.2
Since CDC25s are associated with oncogenic transformation
and are overexpressed in various cancer cells,3 many medicinal
chemists have focused on the development of CDC25 inhibitors
35 since the late 1990s. Many inhibitors have been reported, as
reviewed recently,4 but most of them have a quinone structure
and show non-selective inhibitory activity towards all isoforms of
CDC25s. Nevertheless, these non-selective inhibitors are
considered to be promising lead compounds for cancer therapy,
40 because mice lacking CDC25B and CDC25C survive and exhibit
normal cell cycle and checkpoint responses, suggesting that other
phosphatases, including CDC25A, are able to functionally
compensate for loss of CDC25B and CDC25C.5 Therefore,
subtype non-selective inhibitors may be appropriate for cancer
45 therapy. Indeed, these quinoid-type inhibitors generally show
powerful inhibitory activity at the enzyme level and exhibit
But, it would be difficult to create isoform-selective inhibitors
based on quinoid compounds, because of their property of
generating reactive oxygen species (ROS) within cells.7 The
active site of the CDC25s has a characteristic P-loop structure
55 with the consensus sequence (H/V)CX5R of all DSPs and protein
tyrosine phosphatases (PTPs).8 The highly conserved cysteine
and arginine residues have crucial roles in the catalytic de-
phosphorylation mechanism. The cysteine residue, which acts as
a nucleophile, triggering hydrolysis of the phosphate ester, is
60 especially reactive compared to other cysteine residues due to the
low pKa value of its thiol moiety. Thus, it may be especially
susceptible to ROS. There are thought to be three possible
mechanisms through which CDC25s inhibitors inhibit CDC25s
and other phosphatases, i.e., reversible interaction of the inhibitor
65 with the active site of phosphatases,9 covalent-bond formation
with a nucleophilic amino acid residue around the active site,10
and oxidation of the critical cysteine residue by ROS generated
by the inhibitor.7 The former two mechanisms are widely
encountered, and indeed, some inhibitors of CDC25s have been
70 shown to exhibit competitive or competitive/non-competitive
mixed-mode inhibition.6c,11 On the other hand, some
phosphatases utilize redox reactions as a mechanism of regulation
of their enzymatic activity.12 Therefore, it is plausible that the
quinone-type inhibitors of CDC25s may act via ROS-mediated
75 oxidation of the susceptible cysteine residue; this would be
consistent with their non-selective inhibition of CDC25 isoforms.
Moreover, ROS may oxidize other phosphatases, as well as
unrelated cysteine-based enzymes, and therefore quinone-type
agents could potentially trigger a range of unrelated events in
80 cells.
This journal is © The Royal Society of Chemistry [year]
[journal], [year], [vol], 00–00 | 1

View Article Online
Page 3 of 9
Molecular BioSystems
Recently, we succeeded in developing a series of inhibitors
complex in a similar manner and inhibit the de-phosphorylation.
Most CDC25 inhibitors are believed to bind to the active site. In
contrast, our kinetic data suggested that RE derivatives showed
that do not generate ROS in cells. These inhibitors were found in
a focused library of enamine derivatives of a natural product, RK-
682 (RE derivatives).13 The RE derivatives have better cell 60 inhibition without disturbing the binding of a small-molecular
5
permeability14 than the original compound, RK-682,15 and its
derivatives.16 Furthermore, most of the RE derivatives showed
strong inhibitory selectivity for VHR, CDC25A, and CDC25B,
and did not inhibit CDC25C, MKPs, PTP1B, or CD45. Although
DSPs and PTPs share the same catalytic mechanism, their active
substrate, 3-O-methylfluorescein phosphate (OMFP).
10 site structures are different. DSPs such as VHR have a shallower
binding pocket for substrate phosphate ester than that of PTPs
such as PTP1B.17 In particular, there is no obvious active site
pocket in CDC25s,18 and hence, development of isoform- and/or
enzyme-specific inhibitors is quite challenging. Nevertheless, our
15 RE derivatives, RE1 (1) and RE44 (2), showed complete
selectivity for CDC25A/B over CDC25C as well as MAPKs and
PTPs. Furthermore, some of the derivatives showed partial
selectivity for VHR, CDC25A, and CDC25B. To create inhibitors
with strict isoform selectivity, which would be useful for
20 chemical biology research to clarify the role of individual CDC25
isoforms, it is essential to understand the inhibition mechanism at
the molecular level. Here, we report the results of kinetic studies
and binding site analysis using our original biotin-conjugated
probes. Based on these findings, we propose a unique inhibition
25 mechanism of RE derivatives for CDC25A, via binding at pocket
1 adjacent to the active site.
Fig. 1 Structures of RE derivatives (1 and 2).
Results and Discussion
Enzyme kinetic studies of RE1 (1) and RE44 (2)
We have reported that benzyl-substituted RE114 (1) showed
30 inhibitory activity towards CDC25A (IC50 = 16.6 M), CDC25B
(IC50 = 8.4 M), and VHR (IC50 = 11.4 M), and that o-
hydroxybenzyl-substituted RE4413 (2) showed selectivity for
CDC25B (IC50 = 4.3 M) over CDC25A (IC50 = 13.5 M) and
VHR (IC50 = 24.9 M) (Fig. 1). RE derivatives were designed as
35 neutral phosphate-mimicking molecules, in which the polarized
core enamine functionality conjugated with two carbonyl groups
was expected to have similar properties to substrate phosphate
ester and to bind to the active site P-loop structure of DSPs.
Molecular modelling of RE1 (1) with VHR suggested that the
40 core structure could bind to the active site, and the results of a
preliminary structure-activity relationship study focusing on the 65 Fig. 2 Lineweaver-Burk plot of the effect of RE derivatives on CDC25s
activities. A) Effects of RE1 (1) on the activity of CDC25A, B) Effects of
RE44 (2) on the activity of CDC25B.
long hydrophobic alkyl chain and primary alcohol were
consistent with this model. Unfortunately, however, attempts to
generate a binding model of RE derivatives with CDC25s using
As mentioned above, RE derivatives show high inhibitory
45 Discovery Studio software failed, probably due to the
selectivity for CDC25A/B, and therefore identification of the
shallowness of the binding pocket on CDC25s. Therefore, for
70 binding site of these derivatives would be helpful to understand
further refinement of the structure, aimed at improving the
the origin of the observed selectivity. Thus, we next planned to
inhibitory potency and selectivity of the ligands, we decided that
identify the binding site by installing a covalent bond-forming
it was necessary to obtain experimental information about the
functionality in the RE derivatives.
50 binding mode of RE derivatives with CDC25s. First, to examine
Design and synthesis of modified RE derivatives
whether RE derivatives compete with substrate, we carried out
enzyme kinetic experiments to establish the character of the
inhibition by RE derivatives. Unexpectedly, as shown in Fig. 2,
Lineweaver-Burk plots of RE1 (1) for CDC25A and RE44 (2) for
55 CDC25B indicated that inhibition was non-competitive, i.e., the
inhibitor should bind to both enzyme and substrate-enzyme
75 Here, we focused on identification of the binding site on
CDC25A and carefully examined its crystal structure18a in
comparison with that of CDC25B18b (Figure 3A and 3B). As
already reported, the crystal structure of CDC25B complexed
with sulfate anions has two possible surfaces available for contact

View Article Online
Molecular BioSystems
Page 4 of 9
with small molecules, in addition to the catalytic site. One of
them is a pocket adjacent to the catalytic site, the so-called
“swimming pool” (pocket 1), and the other is the second sulfate
binding site (pocket 2).19 Pocket 1 is surrounded by positively
constructed based on the crystal structure of CDC25B, indicated
the existence of a similar pocket (pocket 1), as shown in Fig. 3C
and 3D. Therefore, pocket 1 seems to be a good candidate for the
binding site of RE derivatives. The other possible RE-binding site
5
charged groups and is constituted by residues 442-448, 479, and 25 is the second sulfate-binding site, which forms a shallow but wide
531-550 (C-terminus of CDC25B).18b In the crystal structure of
CDC25A18a (Figure 3B) this pocket is not obvious, probably due
to the disordered C-terminal structure and the difference in the
position of the side chain of R436, which is a part of the catalytic
pocket (pocket 2) in both CDC25A and CDC25B (Fig. 3A~D). If
the inhibitor binds to one of these pockets, covalent modification
of CDC25s with an RE derivative might be possible by
introducing an electrophilic functionality into the RE derivative,
10 site loop (Fig. 3B). The position of the side chain of R436 on 30 because nucleophilic cysteine residues (C384 and C441 in
CDC25A was unusual compared to that in the case of CDC25B
(R479) and other DSPs/PTPs, and it covers the neighboring
binding pocket (pocket 1 in Figure 3B). Therefore, we speculated
that pocket 1 might be formed by a change in the orientation of
CDC25A, C426 and C484 in CDC25B) exist in, or at least in
close proximity to, the binding pockets.
The primary alcohol of RE1 (1) was found not to be essential
for CDC25-inhibitory activity, since compound 314 (Scheme 1)
15 R436 in solution. It is likely that these differences in crystal 35 showed comparable inhibitory activity to RE1 (1), and its IC50
structure between CDC25A and CDC25B are induced by sulfate
ion. In fact, interaction of the R479 side chain in the crystal
structures of most other DSPs/PTPs in complexes with oxyanions
such as phosphate and sulfate have similar structures to CDC25B,
value for CDC25A was 18.5 M. Thus, we designed the new RE
derivatives 4~9 (Scheme 1) possessing leaving groups such as
methanesulfonate (mesylate, OMs) and halide instead of the
hydroxyl group of 1, or a Michael acceptor structure (,-
20 and the homology model of CDC25A incorporating sulfate, 40 unsaturated ketone).
Fig.3 (A) Crystal structure of CDC25B (PDB ID: 1QB0); (B) Crystal structure of CDC25A (PDB ID: 1C25); (C) Simulated homology model of CDC25A
derived from the crystal structure of CDC25B (PDB ID: 1QB0); (D) Surface model of the CDC25A homology model.
50 be purified by column chromatography without difficulty. In
contrast, the corresponding bromine and chlorine derivatives (X =
Synthesis of new RE derivatives (4~9)
45 The new RE derivatives were synthesized as illustrated in
Scheme 1. The primary alcohol of 1 was converted to mesylate to
afford 4. Treatment of 4 with DBU in toluene gave a potential
Michael acceptor, ,-unsaturated ketone derivative 5. Both 4
and 5 was found to be stable under neutral conditions, and could
Br, Cl) were unstable and gradually decomposed in solution.
Alkyne-modified RE derivatives were also prepared in order to
analyse covalent bond formation of these molecules with
55 CDC25A utilizing click chemistry” (Huisgen reaction).20

View Article Online
Page 5 of 9
Molecular BioSystems
Carboxylic acid 10 with a terminal alkyne moiety21 was
connected to the C3 position of tetronic acid derivative 11 by
treatment with DCC (1,3-dicyclohexylcarbodimide) and DMAP
(4-dimethylaminopyridine) to afford the corresponding 3-
bond with the catalytic domain of CDC25A (catCDC25A), we
first tried MALDI-TOF/MS analysis of catCDC25A treated with
or without simple derivatives 4 and 5. In the case of mesylate 4, a
mass peak indicating covalent bond formation with 4 was not
5
acyltetronic acid derivative. Treatment of the crude product with 30 observed. In contrast, a small new mass peak corresponding to
HF-pyridine complex in THF gave 12. Condensation of 12 with
benzylamine in the presence of small amount of p-
catCDC25A modified by the Michael acceptor 5 was detected
(Figure S1).22
a
toluenesulfonic acid (p-TsOH) provided enamine 6. Mesylate 7
and ,-unsaturated ketone derivative 8 were prepared similarly
However, it proved difficult to identify the binding site using
this sample, due to the very low labelling efficiency. Namely,
10 from 6. Furthermore, the fluorine-substituted derivative 9 (X = F) 35 peptide fragments generated by trypsin or endoproteinase lysC
was also synthesized by treatment of 6 with DAST (N,N-
diethylaminosulfur trifluoride), and this molecule was found to be
stable.
digestion of catCDC25A modified with 5 were analysed by LC-
MS and compared to those of the control catCDC25A, but no
pronounced difference between the control and RE-modified
catCDC25A was observed, and no RE-modified peptide fragment
40 was detected.
Fig. 4 Inhibitory activities of RE derivatives towards CDC25s, and
covalent bond-forming properties with GST-CDC25A. A) Inhibitory
activities of RE derivatives towards GST-CDC25s; B) Western blotting
45 analysis of GST and GST-CDC25A after treatment with alkyne-modified
RE derivatives 7~9 followed by conjugation with biotin-azide (13) using
Cu(I)-mediated click chemistry. Alkyne-modified compounds were
removed on a Micro Bio-Spin Chromatography 6 column (Bio-Rad)
before click chemistry.
50
Therefore, to further confirm covalent bond formation, we next
used the alkyne-modified RE derivatives 7~9. After treatment of
GST-CDC25A protein in Tris-based buffer with each compound,
a biotinyl group was introduced into the alkynyl tail of the RE
derivative by Cu(I)-mediated azide-alkyne Huisgen reaction20b
15 Scheme 1 (A) Structures of RE derivatives (3~9); (B) Synthesis of new
RE derivatives (4~9); (C) Structure of biotin-azide (13).
Confirmation of covalent bond formation of RE derivatives
with CDC25A by MALDI-TOF/MS analysis and western
blotting
55 with biotin-azide probe 13 (Scheme 1). The GST-CDC25A
protein and the free RE derivative were separated by SDS-PAGE,
and the protein was blotted onto a membrane. The biotinylated
RE derivative-modified protein should be detectable by
immunostaining using anti-biotin antibody. However, strong
60 staining was observed not only for GST-CDC25A, but also for
the control GST protein treated with the compounds, indicating
non-specific labelling of the protein. We speculated that Huisgen
reaction and/or sample preparation for SDS-PAGE in the
20 With these molecules in hand, we first evaluated their inhibitory
activity towards CDC25s. As shown in Fig. 4A, all compounds
showed potent inhibitory activity for GST-CDC25A, and, as we
had hoped, the nature of the C5 substituent on RE compounds did
not affect the potency.
25
To examine whether these derivatives can form a covalent

View Article Online
Molecular BioSystems
Page 6 of 9
presence of residual free RE derivative may have caused the non- 25 utilizing the specific interaction of a tag with a tag-binding
specific binding to GST (Figure S2),22 and therefore, remaining
small molecules were removed by gel filtration before the
Huisgen reaction. The flow-through fractions from the gel
filtration column (Fr. No. 1~4) were subjected to the Huisgen
protein, such as biotin-avidin, is well established as a powerful
tool to enrich target molecules. But, to identify the binding site of
a small molecule on a target protein, especially if the small
molecule-modified protein is a minor component in a mixture of
5
reaction and analysed by western blotting. As shown in Fig. 4B, 30 unmodified or unrelated proteins, enrichment of the modified
selective and significant covalent bond formation of GST-
CDC25A with Michael acceptor 8 was confirmed, and no
labelling of GST protein was observed. It is noteworthy that a
protein/peptide is essential. In the case of the biotin-avidin system,
the strength of the interaction sometimes makes elution of the
desired proteins/peptides difficult and results in low yield. To
overcome these issues, a chemically cleavable linker, which
10 faint band of GST-CDC25A labelled with mesylate 7 was also
observed, suggesting that mesylate 7 could also form a covalent 35 anchors the small molecule and tag and facilitates isolation of the
bond with CDC25A, although the efficiency was very low. In
contrast, no band was detected for the fluorine derivative 9. These
results indicated that a Michael acceptor-type molecule could
15 form a covalent bond more efficiently than electrophiles in the
objective protein/peptide, can be used.23 Therefore, we designed a
new probe, CL1-biotin (18), equipped with biotin tag,
chemically cleavable linker, and an azide group for coupling to
the alkynylated molecule 8, for identification of the binding site.
a
SN2-type reaction mode. Based on these experimental results, we 40 As a chemically cleavable linker, we employed diazobenzene,
designed
a
new Michael acceptor-type RE derivative for
which is cleaved by reducing salt Na2S2O4.24 Since we assumed
that low solubility of protein/peptide modified by 5 (having a
long hydrophobic alkyl chain) might also be an issue, the above
parts were connected with a poly-ethylene glycol (PEG) spacer.
45 We expected that introduction of hydrophilic functionality into
the new RE probe would improve the solubility of the peptide
fragment linked to RE derivative in aqueous buffer solution.
identification of the binding site on CDC25A.
Design and synthesis of multi-functional probe RE142 (19)
20 with biotin moiety and chemically cleavable diazobenzene
linker
As mentioned above, it was difficult to detect a very small
quantity of modified peptide in the presence of a large amount of
unmodified peptides by mass analysis. Affinity purification
Scheme 2 Structures and synthesis of CL1-biotin (18) and RE142 (19).
50 Synthesis of CL1-biotin (18) was started from Bogyo’s
intermediate (14, Scheme 2).24c Condensation of 14 with
unexpectedly strong band was observed for the control GST. It
was found that a significant level of biotin labelling of GST
commercially available N3-PEG-NH2 (15) provided 16, although 60 occurred under Huisgen reaction conditions even in the absence
the chemical yield was not satisfactory. After removal of the
Fmoc group, introduction of commercially available biotin-PEG-
55 COOH (17) afforded 18. However, when we performed Western
blotting experiments similar to Figure 4B using the new CL1-
of RE derivatives. We speculate that some reactive species may
be generated by Cu(I)-mediated oxidation of the diazo-phenol
moiety of 18 and this species may react with surface amino acid
residue(s) on GST. Therefore, we decided to connect 8 and 18
biotin probe 18 instead of the simple biotin-azide probe 13, an 65 before the reaction with CDC25A. The Huisgen reaction of 18

View Article Online
Page 7 of 9
Molecular BioSystems
with 8 in t-BuOH-water proceeded without difficulty to give the 35 catCDC25A
was
observed
(Fig.
5D).
(3-[(3-
new probe molecule for binding site analysis, RE142 (19), in
59% yield.
Cholamidopropyl)dimethylammonio]propanesulfonate) (CHAPS)
was also found to be useful as a detergent for improving the
solubility of 19 and did not disturb the binding of 19 with
catCDC25A (data not shown). Overall, these results indicated
40 that RE142 (19) is suitable for analysis of the binding site of RE
derivatives on CDC25A.
Fig. 6 LC-MS analysis of digested catCDC25A after treatment with
RE142 followed by cleavage of the diazobenzene linker with Na2S2O4. A)
45 Observed mass spectrum of cleaved RE142 (cRE142)-modified peptide;
B) Simulated mass spectrum of cRE142-modified peptide; C) Structure of
cRE142-modified peptide.
5 Fig. 5 Confirmation of covalent-bond formation of RE142 (19) with
catCDC25A. A) MALDI-TOF/MS analysis of catCDC25A (calculated
average M.W. = 24,204); B) MALDI-TOF/MS analysis of catCDC25A
after treatment with compound 19 (6.6 mM, calculated M.W. = 1476.81);
C) Western blotting analysis of catCDC25A after treatment with CL1-
10 biotin (18) or RE142 (19) in Tris-based buffer; D) Western blotting
analysis of catCDC25A after treatment with CL1-biotin (17) or RE142
(19) in phosphate-based buffer.
Identification of RE-modified peptide by LC-MS analysis
As shown in Figure 5, we confirmed the generation of RE142-
50 modified catCDC25A protein by MALDI-TOF/MS and western
blotting analysis, but the efficiency of covalent bond formation
was still low, and enrichment of the modified peptide was
required. After treatment of catCDC25A with RE142 (19, 1 mM)
in the presence of 2% CHAPS, proteins were precipitated with
55 TCA/acetone to remove excess RE142 (19) and CHAPS. The
resulting protein precipitate was dissolved in 7 M guanidine-HCl,
and treated with DTT and iodoacetamide (IAA) to reduce and cap
the thiol groups of cysteine residues. Then it was digested
sequentially with endoproteinase lysine-C (Lys-C) and trypsin to
60 afford a peptide mixture including the RE142 (19)-modified
peptide.22 Enrichment of the modified peptide was conducted
with NeutrAvidin beads at 37 °C. The beads were washed several
times, then treated with cleavage buffer containing 0.5 M
Na2S2O4 and 0.005% SDS to elute the peptide. These treatments
65 reduced the diazobenzene group of the RE-142-labeled peptide
with release of cleaved RE142 (cRE142) having the structure
shown in Figure 6C. NanoLC-MS analysis of the eluted peptides
confirmed the formation of the cRE142-modified peptide
corresponding to 378EFVIIDCRYPYEYEGGHIK396 + cRE142.
70 The observed mass numbers were well matched with the
calculated values, as shown in Fig. 6A and 6B (the error was just
0.15 ppm). These results strongly indicate that the observed ion at
Covalent bond formation of RE142 (19) with CDC25A
The newly synthesized RE142 (19) retained moderate inhibitory
15 activity towards GST-CDC25A (IC50 = 59.2 M) and GST-
CDC25B (IC50 = 56.3 M) but GST-CDC25C (IC50 = >100 M),
although its potency was about 10 times less than that of 5 or 8.
MALDI-TOF/MS analysis of catCDC25A after treatment with 19
(6.6 mM, molar ratio 1:200) resulted in the appearance of a new
20 small peak corresponding to catCDC25A protein with one
molecule of 19 (Fig. 5A and 5B). Since 19 has a biotinyl group,
covalent bond formation was directly detectable by Western
blotting analysis without further transformation. As shown in
Figure 5C, mixing of CL1-biotin (18) with catCDC25A in the
25 absence of Cu(I) salt did not give any biotin-labelled protein. In
contrast, the protein band corresponding to biotin-labelled
catCDC25A was observed when catCDC25Awas treated with
RE142 (19) in Tris-based buffer (pH = 8.2, Fig. 5C), but its
formation showed poor dose-dependency. Since the use of a high
30 concentration of 19 in Tris-buffer resulted in formation of a
suspension, we suspected that the solubility of 19 was insufficient,
and so we examined several other conditions. We found that
phosphate-based buffer (pH = 8.0) gave a clear solution even
with 100 M 19, and in this case, dose-dependent labelling of

View Article Online
Molecular BioSystems
Page 8 of 9
m/z 780 ([M+4H]4+) is that of the cRE142-modified peptide.
state by affecting the side chain conformation of the critical R436,
Further MS/MS analysis indicated that RE142 was bound to one 45 thereby inhibiting the hydrolysis.
of residues C384, R385, or Y386 (Figure S5 and Table S2).22
Interestingly, these amino acid residues are located at binding
pocket 1, shown in Figure 3. Although we cannot exclude the
possibility that RE142 also binds to the other amino acid residues
of cat CDC25A, this was the only modified peptide fragment that
we found. These results strongly suggest that RE142 binds to
pocket 1.
Conclusions
5
We designed a series of RE derivatives from natural product RK-
682 as phosphate-mimicking molecules. Initially, we considered
that, like other phosphatase inhibitors, RE derivatives would bind
50 to the active site of CDC25A and CDC25B in a competitive
manner to substrate phosphate ester. However, kinetic
experiments showed that the inhibition was non-competitive.
10 Plausible mechanism of inhibition of CDC25s by RE
derivatives
Binding site analysis using
a covalent bond-forming RE
derivative revealed that the RE derivative binds to a pocket
55 adjacent to the catalytic site. Although this binding pocket has
been already identified in the literature as a promising target for
drug development, to our knowledge, the present RE derivatives
are the first class of compounds that have been shown to bind to
this pocket. Although we have not yet analysed the binding site of
60 RE derivatives on CDC25B, a similar binding mechanism might
be anticipated, based on the observation of non-competitive
inhibition in the kinetic study of RE44 (2, Fig. 2B). Although the
molecular basis on the observed selectivity of the RE derivatives
is still unclear partially due to the lack of the structural
65 information of CDC25C with high resolution, our finding that RE
derivatives bind at pocket 1 adjacent to the active site represents
the first experimental evidence supporting the idea that a strategy
of targeting this pocket would be effective for developing
inhibitors of the biologically and pharmaceutically important
70 CDC25s.
Although we could not determine which amino acid residue
was modified by RE142 among the above three residues, R385 is
an unlikely candidate, because it is buried in the folded protein
15 (Fig. 7) and the nucleophilicity of the guanidine residue is
generally low. Y386 is located at the surface near pocket 1, and a
phenol group can react with the Michael acceptor. However, it is
well known that a thiol group is a much better nucleophile than a
phenolic hydroxyl group as a Michael donor, and C384 seems to
20 be the most plausible candidate, based on the chemical reactivity.
Although the conformation of the side chain residue of C384 in
the reported crystal structure is poorly placed to react directly
with RE142, a molecular modelling study22 suggested that
covalent bond formation of the thiol group with the exo-
25 methylene group of 4 is possible. Based on these considerations
we think that C384 is the most plausible candidate.
Notes and references
a RIKEN, 2-1 Hirosawa, Wako-shi, Saitama 351-0198, Japan. Fax: 81 48
462 4666; E-mail: sodeoka@riken.jp
b ERATO-JST, 2-1 Hirosawa, Wako-shi, Saitama 351-0198, Japan.
75 † Electronic Supplementary Information (ESI) available: [Experimental
procedures, spectrum data of compounds, and supportive figures]. See
DOI: 10.1039/b000000x/
1
2
M. Malumbres and M. Barbacid, Nat. Rev. Cancer, 2009, 9, 153-166.
a) L. Busino, M. Chiesa, G. F. Draetta and M. Donzelli, Oncogene,
2004, 23, 2050-2056; b) K. Kristjansdottir and J. Rudolph, Chem.
Biol., 2004, 11, 1043-1051; c) R. Boutros, C. Dozier and B.
Ducommun, Curr. Opin. Cell. Biol., 2006, 18, 185-191; d) R.
Boutros, V. Lobjois and B. Ducommun, Nat. Rev. Cancer, 2007, 7,
495-507; e) J. Rudolph, Biochemistry, 2007, 46, 3595-3604; f) D.
Ray and H. Kiyokawa, Cancer Res., 2008, 68, 1251-1253.
80
85
90
Fig. 7 Plausible binding site of RE derivatives on CDC25A.
3
a) J. Z. Shou, N. Hu, M. Takikita, M. J. Roth, L. L. Johnson, C.
Giffen, Q. H. Wang, C. Wang, Y. Wang, H. Su, L. H. Kong, M. R.
Emmert-Buck, A. M. Goldstein, S. M. Hewitt and P. R. Taylor,
Cancer Epidemiol. Biomarkers Prev., 2008, 17, 1424-1435; b) M. G.
Cangi, S. Piccinin, L. Pecciarini, A. Talarico, E. Dal Cin, S. Grassi,
A. Grizzo, R. Maestro and C. Doglioni, Int. J. Cancer, 2008, 123,
1466-1471; c) Z. Wang, C. G. Trope, V. A. Flørenes, Z. Suo, J. M.
Nesland and R. Holm, BMC Cancer, 2010, 10, 233-243; d)
reviewed in ref 2b.
As discussed previously, pocket 1 in the crystal structure of
30 CDC25A without sulfate anion is covered by R436 (Fig. 3A). In
contrast, in the model of the sulfate complex (Fig. 3B), R436 flips
and faces the P-loop, forming multiple hydrogen bonds to the
sulfate oxygen atoms, and pocket 1 is opened. It is likely that the
side chain of R436 is flexible in the absence of substrate. If this is
35 the case, RE derivatives could bind to pocket 1 either in the
absence or presence of substrate phosphate ester (Fig. 7). The
observed non-competitive inhibition mode is reasonable, if RE
derivatives bind at pocket 1. For the hydrolysis of the phosphate
ester, nucleophilic attack of the thiolate residue of C430 and
40 stabilization of the resulting meta-phosphate-like transition state
by hydrogen bonding from the R436 are proposed to be
critical.2e,25 Therefore it is possible that the binding of RE
derivatives to pocket 1 disturbs the stabilization of the transition
95 4 a) M. A. Lyon, A. P. Ducruet, P. Wipf and J. S. Lazo, Nat. Rev. Drug
Discov., 2002, 1, 961-976; b) Z. Y. Zhang, Annu. Rev. Pharmacol.
Toxicol., 2002, 42, 209-234; c) L. Bialy and H. Waldmann, Angew.
Chem. Int. Ed., 2005, 44, 3814-3839; d) M. O. Contour-Galcera, A.
Sidhu, G. Prévost, D. Bigg and B. Ducommun, Pharmacol. Ther.,
100
2007, 115, 1-12; e) M. C. Brezak, P. G. Kasprzyk, M. O. Galcera, O.
Lavergne and G. P. Prévost, Anticancer Agents Med. Chem., 2008, 8,
857-862; f) A. Lavecchia, C. D. Giovanni and E. Novellino, Mini.
Rev. Med. Chem., 2012, 12, 62-73.

View Article Online
Page 9 of 9
Molecular BioSystems
5
6
A. M. Ferguson, L. S. White, P. J. Donovan and H. Piwnica-Worms,
Mol. Cell. Biol., 2005, 25, 2853-2860.
17 J. Yuvaniyama, J. M. Denu, J. E. Dixon and M. A. Saper, Science,
1996, 272, 1328-1331.
Some isoform-selective inhibitors have been reported. As
representative examples, see: a) Y. Aoyagi, N. Masuko, S. Ohkubo,
M. Kitade, K. Nagai, S. Okazaki, K. Wierzba, T. Terada, Y.
Sugimoto and Y. Yamada, Cancer Sci., 2005, 96, 614-619; b) S.
Kar, M. Wang, S. W. Ham and B. I. Carr, Cancer Biol. Ther., 2006,
5, 1340-1347; c) S. Kar, M. Wang, W. Yao, C. J. Michejda and B. I.
Carr, Mol. Cancer Ther., 2006, 5, 1511-1519.
18 a) E. B. Fauman, J. P. Cogswell, B. Lovejoy, W. J. Rocque, W.
Holmes, V. G. Montana, H. Piwnica-Worms, M. J. Rink and M. A.
Saper, Cell, 1998, 93, 617-625; b) R. A. Reynolds, A. W. Yem, C. L.
Wolfe, M. R. Deibel, Jr., C. G. Chidester and K. D. Watenpaugh, J.
Mol. Biol., 1999, 293, 559-568.
5
75
80
85
90
95
19 J. Rudolph, Mol. Pharmacol., 2004, 66, 780-782.
20 a) H. C. Kolb, M. G. Finn and K. B. Sharpless, Angew. Chem. Int.
Ed., 2001, 40, 2004-2021; b) M. G. Finn and V. V. Fokin, in
Catalysis without Precious Metals, Wiley-VCH Verlag GmbH & Co.
KGaA, 2010, pp. 235-260.
21 a) W. P. Heal, B. Jovanovic, S. Bessin, M. H. Wright, A. I. Magee
and E. W. Tate, Chem. Commun., 2011, 47, 4081-4083; b) A.
Lumbroso, N. Abermil and B. Breit, Chem. Sci., 2012, 3, 789-793.
22 See Electronic Supplementary Information (ESI).
23 For a representative article, see: J. Szychowski, A. Mahdavi, J. J. L.
Hodas, J. D. Bagert, J. T. Ngo, P. Landgraf, D. C. Dieterich, E. M.
Schuman and D. A. Tirrell, J. Am. Chem. Soc., 2010, 132, 18351-
18360.
24 a) C. L. Jaffe, H. Lis and N. Sharon, Biochemistry, 1980, 19, 4423-
4429; b) J. B. Denny and G. Blobel, Proc. Natl. Acad. Sci. USA,
1984, 81, 5286-5290; c) S. H. L. Verhelst, M. Fonovic and M.
Bogyo, Angew. Chem., Int. Ed., 2007, 46, 1284-1286; d) M.
Fonović, S. H. L. Verhelst, M. T. Sorum and M. Bogyo, Mol. Cell.
Proteomics, 2007, 6, 1761-1770; e) Y. Y. Yang, M. Grammel, A. S.
Raghavan, G. Charron and H. C. Hang, Chem. Biol., 2010, 17,
1212-1222; f) P. F. Hughes, J. J. Barrott, D. A. Carlson, D. R.
Loiselle, B. L. Speer, K. Bodoor, L. A. Rund and T. A. J. Haystead,
Bioorg. Med. Chem., 2012, 20, 3298-3305.
10 7 M. Brisson, T. Nguyen, P. Wipf, B. Joo, B. W. Day, J. S. Skoko, E.
M. Schreiber, C. Foster, P. Bansal and J. S. Lazo, Mol. Pharmacol.,
2005, 68, 1810-1820.
8
A. Alonso, J. Sasin, N. Bottini, I. Friedberg, I. Friedberg, A.
Osterman, A. Godzik, T. Hunter, J. Dixon and T. Mustelin, Cell,
2004, 117, 699-711.
15
9
a) J. S. Lazo, K. Nemoto, K. E. Pestell, K. Cooley, E. C. Southwick,
D. A. Mitchell, W. Furey, R. Gussio, D. W. Zaharevitz, B. Joo and
P. Wipf, Mol. Pharmacol., 2002, 61, 720-728; b) A. Lavecchia, S.
Cosconati, V. Limongelli and E. Novellino, ChemMedChem, 2006,
1, 540-550; c) S. Kolb, O. Mondésert, M. L. Goddard, D. Jullien, B.
O. Villoutreix, B. Ducommun, C. Garbay and E. Braud,
ChemMedChem, 2009, 4, 633-648; d) H. Park, M. Li, J. Choi, H.
Cho and S. W. Ham, Bioorg. Med. Chem. Lett., 2009, 19, 4372-
4375.
20
25 10 L. Pu, A. A. Amoscato, M. E. Bier and J. S. Lazo, J. Biol. Chem.,
2002, 277, 46877-46885.
11 a) R. L. Rice, J. M. Rusnak, F. Yokokawa, S. Yokokawa, D. J.
Messner, A. L. Boynton, P. Wipf and J. S. Lazo, Biochemistry, 1997,
36, 15965-15974; b) A. P. Ducruet, R. L. Rice, K. Tamura, F.
30
35
40
45
50
55
60
65
70
Yokokawa, S. Yokokawa, P. Wipf and J. S. Lazo, Bioorg. Med. 100
Chem., 2000, 8, 1451-1466; c) K. Tamura, E. C. Southwick, J.
Kerns, K. Rosi, B. I. Carr, C. Wilcox and J. S. Lazo, Cancer Res.,
2000, 60, 1317-1325; d) J. S. Lazo, D. C. Aslan, E. C. Southwick, K.
A. Cooley, A. P. Ducruet, B. Joo, A. Vogt and P. Wipf, J. Med.
Chem., 2001, 44, 4042-4049; e) M. Brisson, T. Nguyen, A. Vogt, J.
Yalowich, A. Giorgianni, D. Tobi, I. Bahar, C. R. Stephenson, P.
Wipf and J. S. Lazo, Mol. Pharmacol., 2004, 66, 824-833; f) K. R.
Kim, J. L. Kwon, J. S. Kim, Z. No, H. R. Kim and H. G. Cheon, Eur.
J. Pharmacol., 2005, 528, 37-42; g) L. Brault, M. Denancé, E.
Banaszak, S. El Maadidi, E. Battaglia, D. Bagrel and M. Samadi,
Eur. J. Med. Chem., 2007, 42, 243-247; h) E. Braud, M. L. Goddard,
S. Kolb, M. P. Brun, O. Mondesert, M. Quaranta, N. Gresh, B.
Ducommun and C. Garbay, Bioorg. Med. Chem., 2008, 16, 9040-
9049; i) X. Feng, L. N. Wang, Y. Y. Zhou, H. P. Yu, Q. Shen, Y.
Zang, Y. B. Zhou, J. Y. Li, H. X. Zhang and J. Li, Acta Pharmacol.
Sin., 2008, 29, 1268-1274; j) H. Park, Y. J. Bahn, S. K. Jung, D. G.
Jeong, S. H. Lee, I. Seo, T. S. Yoon, S. J. Kim and S. E. Ryu, J.
Med. Chem., 2008, 51, 5533-5541; k) H. J. Chen, Y. Liu, L. N.
Wang, Q. Shen, J. Li and F. J. Nan, Bioorg. Med. Chem. Lett., 2010,
20, 2876-2879; l) A. Lavecchia, C. Di Giovanni, A. Pesapane, N.
Montuori, P. Ragno, N. M. Martucci, M. Masullo, E. De Vendittis
and E. Novellino, J. Med. Chem., 2012, 55, 4142-4158.
25 Z.-Y. Zhang, Acc. Chem. Res., 2003, 36, 385-392.
12 a) A. Salmeen, J. N. Andersen, M. P. Myers, T. C. Meng, J. A. Hinks,
N. K. Tonks and D. Barford, Nature, 2003, 423, 769-773; b) R. L.
van Montfort, M. Congreve, D. Tisi, R. Carr and H. Jhoti, Nature,
2003, 423, 773-777; c) J. Sohn and J. Rudolph, Biochemistry, 2003,
42, 10060-10070; d) S. H. Cho, C. H. Lee, Y. Ahn, H. Kim, H. Kim,
C. Y. Ahn, K. S. Yang and S. R. Lee, FEBS Lett., 2004, 560, 7-13;
e) G. Buhrman, B. Parker, J. Sohn, J. Rudolph and C. Mattos,
Biochemistry, 2005, 44, 5307-5316.
13 A. Tsuchiya, G. Hirai, Y. Koyama, K. Oonuma, Y. Otani, H. Osada
and M. Sodeoka, ACS Med. Chem. Lett., 2012, 3, 294-298.
14 G. Hirai, A. Tsuchiya, Y. Koyama, Y. Otani, K. Oonuma, K. Dodo, S.
Simizu, H. Osada and M. Sodeoka, ChemMedChem, 2011, 6, 617-
622.
15 T. Hamaguchi, T. Sudo and H. Osada, FEBS Lett., 1995, 372, 54-58.
16 a) M. Sodeoka, R. Sampe, S. Kojima, Y. Baba, T. Usui, K. Ueda and
H. Osada, J. Med. Chem., 2001, 44, 3216-3222; b) T. Usui, S.
Kojima, S. Kidokoro, K. Ueda, H. Osada and M. Sodeoka, Chem.
Biol., 2001, 8, 1209-1220.