Discovery of cell-active phenyl-imidazole Pin1 inhibitors by structure-guided fragment evolution

Article abstract of DOI:10.1016/j.bmcl.2010.09.063

Pin1 is an emerging oncology target strongly implicated in Ras and ErbB2-mediated tumourigenesis. Pin1 isomerizes bonds linking phospho-serine/threonine moieties to proline enabling it to play a key role in proline-directed kinase signalling. Here we report a novel series of Pin1 inhibitors based on a phenyl imidazole acid core that contains sub-μM inhibitors. Compounds have been identified that block prostate cancer cell growth under conditions where Pin1 is essential.

Full text of DOI:10.1016/j.bmcl.2010.09.063

Bioorganic & Medicinal Chemistry Letters 20 (2010) 6483–6488
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery of cell-active phenyl-imidazole Pin1 inhibitors
by structure-guided fragment evolution
Andrew Potter, Victoria Oldfield , Claire Nunns, Christophe Fromont à, Stuart Ray, Christopher J. Northfield,
Christopher J. Bryant, Simon F. Scrace, David Robinson §, Natalia Matossova, Lisa Baker, Pawel Dokurno,
–
⇑
Allan E. Surgenor, Ben Davis, Christine M. Richardson , James B. Murray, Jonathan D. Moore
Vernalis (R&D) Ltd, Granta Park, Great Abington, Cambridge CB21 6GB, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 28 July 2010
Revised 6 September 2010
Accepted 10 September 2010
Available online 17 September 2010
Pin1 is an emerging oncology target strongly implicated in Ras and ErbB2-mediated tumourigenesis. Pin1
isomerizes bonds linking phospho-serine/threonine moieties to proline enabling it to play a key role in
proline-directed kinase signalling. Here we report a novel series of Pin1 inhibitors based on a phenyl
imidazole acid core that contains sub-
lM inhibitors. Compounds have been identified that block prostate
cancer cell growth under conditions where Pin1 is essential.
Ó 2010 Elsevier Ltd. All rights reserved.
Keywords:
Pin1
Peptidyl prolyl isomerase
Fragments
Cancer
The parvulin-type rotamase, Pin1, catalyzes cis/trans isomerisa-
tion around the peptidyl bond linking phospho-serine/threonine
and proline residues.1 Pin1 therefore regulates the rate of certain
conformational changes induced by phosphorylation/dephospho-
rylation of its substrates. As proline-directed protein kinases and
phosphatases only recognize Ser/Thr-Pro moieties in the trans
conformationas substrates,2–4 Pin1 also regulatessignallingdynam-
ics for pathways involving cyclin-dependent kinases, MAP kinases
and GSK3.5 Pin1 is frequently overexpressed in tumours6 and in
prostate cancer, overexpression is also associated with poor progno-
sis.7 Pin1-deficient mice are viable and their cells resistant to trans-
formation by Ras and ErbB2.8,9 Consequently, there are grounds to
anticipate Pin1 inhibitors will be useful for cancer therapy.
The irreversible Pin1 inhibitor, juglone (1), has been used exten-
sively in studies addressing the role of Pin1 inside cells and the
therapeutic potential of Pin1 inhibition in vivo.10–13 Juglone, how-
ever, lacks selectivity12 and is not a suitable start point for drug
discovery. Steady progress has been made in the identification of
peptidic Pin1 inhibitors,14–16 culminating in cyclic peptides
capable of blocking Pin1 activity inside cells in the sub-l
M range.17
In addition, several cell-active small molecules have been claimed
to act via Pin1 inhibition including several cinnamic acid deriva-
tives18 and a series of aryl indanyl ketones that reduce b-catenin
levels and suppress p53-dependent transcription as would be
expected for a Pin1 inhibitor.19 Unfortunately, crystallographic
information defining the binding mode of these small molecules
has not emerged.
Drug discovery against Pin1 has been reported to be challeng-
ing: the major problem is finding verified hits as start points. Pfizer
have disclosed that they screened 106 compounds versus Pin1
without finding any hits whose binding to target can be verified
by NMR or isothermal calorimetry.20 Taking an alternative ap-
proach to hit identification has proved fruitful: Pfizer designed a
series of phosphate-containing aminophenyl-propanols20 that
were subsequently evolved into a series of phenylalanines that
inhibited Pin1, a property shared with their styryl analogues.21
Vernalis have disclosed a number of benzimidazole- and napthyl-
alanine based inhibitors of Pin1 that were evolved from an indole
carboxylate fragment (2) identified in an NMR-based screen.22 In-
dole and benzimidazole alanines have also been identified as Pin1
inhibitors by Pfizer.23
⇑
Corresponding author.
E-mail address: j.moore@vernalis.com (J.D. Moore).
Present address: Haddow Laboratories, Institute of Cancer Research, 15 Cotswold
Rd, Belmont, Sutton, Surrey SM2 5NG, UK.
Several members of our benzimidazole series of Pin1 inhibitors
had impressive activity (ꢀ100 nM) versus the target in vitro, but
failed to suppress Pin1 activity in cells due to poor permeability.
Activity versus cells could be obtained by reducing the polar
surface area of the molecules by replacing the benzimidazole
moiety with a naphthyl group, but this led to a 10–20-fold loss
à
Present address: School of Pharmacy, University Park, Nottingham, NG7 2RD, UK.
Present address: Division of Biological Chemistry & Drug Discovery, College of Life
§
Sciences, James Black Centre, University of Dundee, Dundee, DD1 5EH, UK.
–
Present address: Biofocus DPI, Chesterford Research Park, Saffron Walden, Essex,
CB10 1XL, UK.
0960-894X/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2010.09.063

6484
A. Potter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6483–6488
in potency.22 For a variety of reasons, the benzimidazole/naphthyl
series were not progressed further and we sought new fragment
hits that might be more suitable for evolution into drugs.
Nine hundred fragments24 were tested for their ability to inhibit
Pin1-mediated isomerisation of the peptide AApSPR-pNA in a pro-
tease coupled assay25,26 at a compound concentration of 2 mM.
Forty putative ‘hits’ were identified, including the indole carboxyl-
ate and benzimidazole propionic acids reported previously.22 Two-
dimensional NMR experiments, however, only verified binding for
2 of the 37 novel ‘hits’ to Pin1 (see Fig. S1 for an example). In the
absence of corroborative data we cannot exclude the possibility
that a given compound inhibits Pin1-mediated peptide isomerisa-
tion via non-specific mechanisms such as forming colloidal
aggregates.27
OH
O
O
N
OH
H
1
2
O
O
O
OH
OH
O
N
O
NH
OH
N
N
H
N
O
4
5
3
From soaking experiments with crystals of Pin1R14A protein28
we obtained a bound structure for one the NMR-verified com-
pounds, the pyridine-pyrazole acid 3 (Pin1 IC50 720 lM) (Fig. 1a).
There are notable similarities between the binding mode of 3
and the indole carboxylate fragment (2) identified previously
(see Fig. S2):22,26 the Lys63-interacting carboxylate moieties are
identically positioned and one of the pyrazole nitrogens of 3 over-
laps with the indole NH of 2 and recapitulates the hydrogen-bond
interaction with Cys113. A variety of near neighbours of 3 were
acquired and assayed for their ability to inhibit Pin1 and soak
into Pin1 crystals. Crystal structure were obtained for several ana-
logues bound to Pin1, including the phenyl furan acid 4 (Pin1
IC50 >1000 lM), in which the heteroatom of the five-membered
ring of the ligand interacts via a hydrogen bond with Ser154, which
is on the opposite site of the Pin1 active site from Cys113 (Fig. 1b).
The starting point for our medicinal chemistry, however, was the
most potent of the near neighbours identified: the phenyl-imidaz-
ole acid, 5 (IC50 180 lM; ligand efficiency (LE) 0.34), which makes
H-bond interactions with both Cys113 and Ser154 via its imidazole
nitrogens (Fig. 1c).
We began by exploiting a literature route29 to synthesise ana-
logues of 5 bearing a methyl moiety at the 5-position of the imid-
azole ring, and investigated SAR around the pendant phenyl ring
(Scheme 1). The commercially available ylide (6) was condensed
with acetyl chloride and the resulting phosphanylidene (7), oxi-
dised to tricarbonyl (8) by Oxone™. The desired imidazoles were
then prepared by refluxing the tricarbonyl (8) and corresponding
aldehyde with ammonium acetate under acidic conditions. Typical
yields of imidazoles were 60–80%. The ester of (9) was then sapon-
ified under basic conditions generating acids (10) in good overall
yield.
Figure 1. Pin1R14A bound crystal structures of (a) 3 (gray: PDB ID: 3ODK) (b) 4
(yellow: PDB ID 2XP3) and (c) 5 (green: PDB ID: 2XP4). Direct H-bonds between
ligand and Pin1 are shown as yellow dashed lines, water molecules are shown in
red. All three ligands occupy the catalytic site of Pin1 and make an H-bond
interaction between one of the ligand’s carboxylate oxygens and Lys63. The
pyridine nitrogen of 3 also interact via H-bonds with both Ser154 and Gln131, while
a pyrazole nitrogen makes a hydrogen bond with Cys113. The furan oxygen of 4
accepts an H-bond from Ser154. Both the imidazole nitrogens of 5 are involved in
H-bond interactions: one with Cys113 and the other with Ser154.
Introduction of the methyl group at the 5-position (10a) of par-
ent fragment 5, led to a small drop in ligand efficiency (LE 0.30)
(see Table 1). Structural data (Fig. S3a) suggested that elaboration
of the phenyl group was only possible at the 3-position, which affor-
ded a vector into a small hydrophobic pocket bounded by the side
chains of Met130, Gln131 and Phe134. We were delighted to see a
>10-fold potency improvement by introduction of a trifluoromethyl
(10b), methyl (10c), nitrile (10d) or chloro (10e) at the 3-position of
the phenyl ring, which increased LE to 0.38 for the chloro
0
substituted compound. LLEAT a normalised measure of ligand-lipo-
philicity efficiency (Paul Mortenson, personal communication)26,30

A. Potter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6483–6488
6485
(b)
.H2O
O
R1
(a)
O
O
O
O
OEt
OEt
OEt
R1
CHO
a
NC
NC
N C
b
O
H
PPh3
PPh3
6
7
8
NH2
11
12
R1
R1
(d)
CN
(c)
N
N
O
CO2H
CO2H
N
d
N
O
OH
c
N
R
R
N
OEt
CN
N
N
CHO
R
14
13
10
9
R1
Scheme 1. Reagents and conditions: (a) AcCl, DIPEA, DCM, rt, 1 h, 98%; (b) oxoneTHF/
H2O (2:1), rt, 18 h, 100%; (c) RCHO, NH4OAc, AcOH, 100 °C, 60–80%; (d) NaOH, EtOH,
reflux, 100%.
R1
O
COCl
N
CO2H
N
N
e
N
H
N
N
R2
N
O
ClOC
Table 1
O
R1
Biological data for 3-substituted phenyl imidazole acid Pin1 inhibitors
O
O
15
16
Scheme 2. Reagents and conditions: (a) diamino-malonitrile, H2SO4, 98%; (b)
nicotinamide, N-chlorosuccinamide, DMF, 50–60%; (c) H2SO4, H2O, reflux 70–80%;
(d) SOCl2, DMF, PhMe, quant; (e) R2NH2, NEt3, DCM, 30–40%.
N
R
N
a
b
R
Pin1 IC50
(lM)
LE (kcal/mol/
heavy atom)
LLEAT (kcal/mol/
and 0.52 lM for 14b). Diacids were chlorinated using thionylchlo-
heavy atom)
ride to dimer (15). Careful treatment of this acyl chloride at low
temperature with a variety of amines followed by preparative
chromatography led to the amido acids (16), in moderate
(30–40%) yield.
10a
10b
10c
10d
10e
10f
H
360
23
34
19
20
0.30
0.31
0.36
0.36
0.38
0.27
0.24
0.20
0.29
0.34
0.27
0.23
CF3
CH3
CN
Cl
Initial results in this series were not encouraging: the methyl
amide (16a) maintained the high LE of 10a, but LE declined for
the larger secondary amides (16b) and (16c) (see Table 2). Impres-
sive gains in affinity were obtained however, for the morpholino
tertiary amide (16d) for which we were able to obtain a crystal
structure (Fig. 2a). To our surprise, the amide substituent was
not positioned in the channel leading to Trp73; instead, the unsat-
urated morpholine ring stacked along the hydrophobic part of the
Arg68 side chain and the acid group bound to both Lys63 and Arg
69 and was orientated towards the Trp73 channel. This prompted
us to further explore tertiary amides.
OCH3
270
a
IC50 values are means of P2 determinations rounded to two significant figures.
The coefficient of variance of this assay was 26% (repeat assays with compound 3
from reference 17 yielded mean IC50 values of 0.313 M with standard deviation
0.081 M; n = 4).
l
l
b
LLEAT is a measure of ligand-lipophilicity efficiency used at Astex Therapeutics
(Paul Mortenson, personal communication) that takes into account molecule size
and is ideal for tracking evolution of fragments to leads (see Supplementary data)26
approached or exceeded the minimum target value of 0.3 for the
chloro and nitrile substituted compounds, 10d and 10e. The crystal
structure of 10e bound to Pin1 (Fig. S3) confirmed that the 5-posi-
tion substituents occupied the hydrophobic pocket as predicted,
with the position of the phenyl and imidazole rings being unaltered
from the parent 10a.26 Larger substituents, (e.g. compound 10f)
were not beneficial nor were alternative substitutions around the
phenyl ring.31
We considered that the vector towards the channel in Pin1
leading to Trp73 supplied by the 5-methyl on the imidazole ring
provided the best opportunity to improve potency (see Fig. S3).
Accessing this channel via amido acids that might make productive
interactions with Arg69 was an attractive option, as the route to
such compounds had already been described in the literature
(Scheme 2).32 The aldehyde (11) was condensed with diamino-
malonitrile in DMSO and concentrated sulphuric acid. The Schiff’s
base produced (12) was then cyclised with nicotinamide and N-
chlorosuccinamide in DMF; the dicyanoimidazole (13) formed in
50–60% yield. Hydrolysis of (13) under acidic conditions led to
the diacid (14).
The N-benzyl N-acetic acid ethyl ester analogue (16f: Pin1 IC50
8 lM) bound to Pin1 in a similar manner to 16d with the carbox-
ylate oriented towards Trp73 and the benzylic moiety packing
along the arginine side chains (Fig. 2b). An interesting additional
interaction was, however, observed between the carbonyl oxygen
of the ester substituent and the secondary amine nitrogen in
Arg69 (Fig. 2b). Unfortunately, 16f did not impair proliferation of
PC3 prostate cancer cells, even under the serum-starved conditions
where Pin1 expression is known to be important.33 Previous expe-
rience suggested increasing the lipophilicity of Pin1 inhibitors
could improve activity vs. cells.22 We chose to synthesise an N
,N0-dibenzyl analogue as modelling suggested the aryl groups
might stack with the side chains of both Arg68 and Arg69. This
hypothesised binding mode could not be verified as no crystal
structure could be obtained for 16g (or any similar compound).
Encouragingly, 16g retained the potency (Pin1 IC50
while gaining the ability to block proliferation of PC3 cells (GI50
21 M). This gain in cell activity, however, was at the expense of
4 lM) of 16f
l
a move into less drug-like chemical space as LLEAT declined from
The diacids 14a and 14b were particularly potent and ligand-
efficient inhibitors of Pin1 (Table 2). The crystal structure obtained
for 14a bound to Pin1 suggested the molecule conserved the bind-
ing mode of the other phenyl-imidazoles (Fig. S4).26 Isothermal
calorimetry confirmed these compounds bound tightly to Pin1
0.20 for 16f to 0.12 for 16g.
Replacing the N-benzyl moiety with more polar side chains
(16h and 16i) yielded compounds with improved ligand efficiency
and reasonable ligand lipophilic efficiency. The N-methyl primary
amide analogue (16i) was particularly notable (IC50 4.1
lM; LE
with 1:1 stoichiometry (ITC-derived Kd values: 1.7 lM for 14a
0.31; LLEAT 0.43) and its crystal structure (Fig. 3) indicated that

6486
A. Potter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6483–6488
Table 2
Biological data for phenyl imidazole Pin1 inhibitors
R1
H
CO2H
N
O
N
R2
R1
R2
Pin1 IC50
(l
M)
LE (kcal/mol/heavy atom)
LLEAT (kcal/mol/heavy atom)
PC3 GI50 (lM)
a
14a
14b
16a
16b
16c
H
F
H
H
H
OH
OH
NHMe
NHEt
NHBn
N
1.46
0.32
80
140
220
0.44
0.46
0.29
0.26
0.20
0.47
0.46
0.34
0.28
0.15
ND
ND
ND
ND
ND
16d
16e
H
H
15
31
0.28
0.23
0.34
0.18
>200
>200
O
NMeBn
N
16f
H
8.0
0.22
0.20
>200
O
O
16g
16h
H
H
N(Bn)2
4.0
9.4
0.23
0.29
0.12
0.37
21
ND
CO2H
CONH2
N
N
16i
16j
H
H
4.1
0.31
0.25
0.43
0.27
>200
>80
H2NOC
N
0.52
N
S
a
IC50 values are means of P2 determinations rounded to two significant figures where appropriate. The coefficient of variance of this assay was 26%.
the ethyl amide substituent packed against the hydrophobic por-
tion of the Arg68 side chain. Substitution of the methyl with a
methyl-benzothiophene moiety resulted in 16j, the most potent
accord with ITC experiments (Kd 1.4
Homologation of one of the N-benzyl moieties to an N-ethylphenyl
lM; stoichiometry 0.9).
resulted in 20, a sub-lM inhibitor of Pin1 that exhibited excellent
Pin1 inhibitor identified in this series (IC50 0.52
lM: LE 0.25). No
permeability in a CaCo-2 assay (35 ꢁ 10ꢂ6/cm)31 and blocked the
crystal structure of 16j bound to Pin1 could be obtained, so it is
not clear whether this molecule retains the binding mode of 16i.
While 16i and 16j had encouraging potency versus Pin1 in vitro,
the ability to block PC3 proliferation had been lost. We therefore
reasoned that synthesising further compounds similar to 16g
might represent our best chance of progress towards identifying
Pin1 inhibitors that are active in cells, even though such molecules
would be highly lipophilic.
We had previously shown that 3-substitution on the phenyl
ring improved potency by up to 10-fold and wondered if incorpo-
ration of a suitable group into this position of our most active ter-
tiary amides would do likewise. We were aware, however, that
consistent SAR may not be observed because the Pin1 enzyme
would only be able to bind one of the slowly interconverting rota-
meric forms of these ligands. The rotameric forms were clearly
present by NMR.31
proliferation of serum-starved PC3 cells with a GI50 of 13 lM. 20
did not inhibit the activity of FKBP12 or cyclophilin A indicating
that it specifically inhibited the PPIase activity of Pin1.
Pin1 inhibitors are anticipated to interfere with multiple cellu-
lar processes including the expression of cyclin D1,8,34 and signal-
ling cascades triggered by insulin.35 We therefore asked whether
20 interfered with any of these processes as doses close to its
GI50. Figure 5 shows that treatment of PC3 cells with 20 phenocop-
ies siRNA-mediated knockdown of Pin1 in reducing cyclin D1 lev-
els in serum-starved cells (EC50 ꢀ10
cyclin E1 expression. In addition 20 impairs p70S6-Kinase phos-
M). Together these
lM) but having no effect on
phorylation in response to insulin (EC50 ꢀ30
l
data provide considerable evidence that the effects on cells of 20
are mediated, at least in part, via inhibition of Pin1.
Fragment-based hit identification combined with structure-
guided hit evolution provide a powerful means to identify novel
chemical series against targets considered challenging from a drug
discovery perspective. Twice using these approaches we have iden-
In the context of 16e, substituting a chlorine at the 3-position of
the phenyl ring adjacent to the imidazole (17) led to a >10-fold
improvement in activity vs. Pin1 (IC50
2
l
M) (Table 3). A crystal
tified inhibitor series that contain examples exhibiting sub-lM
structure for 17 bound to Pin1R14A was obtained (Fig. 4), which
indicated complete conservation of the binding mode observed
for the deschloro compounds 16f and 16i. However, substitution
of chlorine at the 3-position was not always beneficial: incorpora-
tion of the 3-substituent into the morpholino (16d) yielded a com-
pound (18) at best equipotent versus Pin1 (IC50 21 lM). 3-Chloro
substitution of the phenyl in 16g yielded a molecule (19) with
threefold improved potency versus Pin1, albeit without a corre-
affinity for the Pin1 target, a context where high-throughput
screening has reportedly failed.20 Though the approach used here
(screening a fragment library in an enzymatic assay at a high li-
gand concentration) yielded a set of hits that may have been dom-
inated by compounds interfering with Pin1-mediated peptide
isomerization via non-specific mechanisms such as formation of
colloidal aggregates,27 the small number (<40) of hits and solubil-
ity of the library members enabled 2D NMR spectroscopy to pick
out a novel aryl-pyrazole inhibitor that could be crystallised within
Pin1. This initial hit was rapidly evolved to a series of phenyl-
sponding improvement in activity versus PC3 cells. The 1.4
lM
IC50 for 19 in the trypsin-coupled functional assay was in excellent

A. Potter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6483–6488
6487
Table 3
Biological data for 3-chloro-phenyl-imidazole acid Pin1 inhibitors
Cl
H
N
CO2H
O
R
N
a
R
Pin1 IC50
LE (kcal/mol/ LLEAT (kcal/ PC3 GI50
(lM)
heavy atom)) mol/heavy
atom)
(lM)
17
18
2.0
21
0.26
0.24
0.20
0.28
56
N
O
>200
N
N
19
20
1.4
0.23
0.23
0.11
0.13
18
13
N
0.83
a
IC50 values are the mean of at least two determinations and are rounded to two
significant figures where appropriate. The coefficient of variance of this assay was
26%.
Figure 2. Crystal structures of (a) 16d (cyan; PDB ID 2XP8) and (b) 16f (orange;
PDB ID 2XP9) in Pin1R14A. The carboxylate moiety is positioned in the channel
leading towards Trp73. The morpholino group of 16d and the benzylic moiety of 16f
stack along the hydrophobic side chain of Arg68. Water molecules are shown in red.
Figure 4. Crystal structure of compound 17 in Pin1R14A (PDB ID: 2XPB).
Introduction of a chlorine at the 3-position of the phenyl ring did not alter the
binding mode from that observed with 16f (see Fig. 2b). Water molecules are shown
in red.
modes have been reported.20–22 Like previous series of Pin1 inhib-
itors, the imidazole amino acids reported here can be said to have a
tripartite binding mode that comprises the following elements: a
ring system occupying the hydrophobic cavity at the core of the
Pin1 active site; acidic groups that interact with the basic side
chains of the residues (Lys63, Arg68 and Arg69) that accommodate
the phosphate in Pin1’s substrates; and additional moieties that
bind to hydrophobic but solvent-exposed surfaces of Pin1 adjacent
to the active site. The difference is that while the previously dis-
closed series contact a shelf-like surface located beyond Cys113
on the other side of the binding site,20–22 the ligands described in
this report stack along the side chain of Arg68 and possibly
Arg69. This novel binding mode of the imidazole amido acids raises
the possibility of designing molecules that occupy novel
Figure 3. Crystal structure of compound 16i in Pin1R14A (PDB ID: 2XPA). The ethyl
amide of the ligand stacks along the side chain of Arg68, while the N-methyl moiety
points into solvent. Waters are shown in red.
imidazoles that contained members inhibiting Pin1 with sub-lM
IC50 values and that prevented the growth of prostate cancer cells.
The Pin1 inhibitors described in this report do not bind in the
same manner as any of the classes of ligands for which binding

6488
A. Potter et al. / Bioorg. Med. Chem. Lett. 20 (2010) 6483–6488
References and notes
1. Lu, K. P.; Zhou, X. Z. Nat. Rev. Mol. Cell Biol. 2007, 8, 904.
2. Weiwad, M.; Kullertz, G.; Schutkowski, M.; Fischer, G. FEBS Lett. 2000, 478, 39.
3. Brown, N. R.; Noble, M. E.; Endicott, J. A.; Johnson, L. N. Nat. Cell Biol. 1999, 1,
438.
4. Zhou, X. Z.; Kops, O.; Werner, A.; Lu, P. J.; Shen, M.; Stoller, G.; Küllertz, G.;
Stark, M.; Fischer, G.; Lu, K. P. Mol. Cell. 2000, 6, 873.
5. Lu, K. P.; Finn, G.; Lee, T. H.; Nicholson, L. K. Nat. Chem. Biol. 2007, 3, 619.
6. Bao, L.; Kimzey, A.; Sauter, G.; Sowadski, J. M.; Lu, K. P.; Wang, D. G. Am. J.
Pathol. 2004, 164, 1727.
7. Ayala, G.; Wang, D.; Wulf, G.; Frolov, A.; Li, R.; Sowadski, J.; Wheeler, T. M.; Lu,
K. P.; Bao, L. Cancer Res. 2003, 63, 6244.
8. Liou, Y. C.; Ryo, A.; Huang, H. K.; Lu, P. J.; Bronson, R.; Fujimori, F.; Uchida, T.;
Hunter, T.; Lu, K. P. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 1335.
9. Ryo, A.; Liou, Y. C.; Wulf, G.; Nakamura, M.; Lee, S. W.; Lu, K. P. Mol. Cell. Biol.
2002, 22, 5281.
10. Hennig, L.; Christner, C.; Kipping, M.; Schelbert, B.; Rücknagel, K. P.; Grabley, S.;
Küllertz, G.; Fischer, G. Biochemistry 1998, 37, 5953.
11. Rippmann, J. F.; Hobbie, S.; Daiber, C.; Guilliard, B.; Bauer, M.; Birk, J.; Nar, H.;
Garin-Chesa, P.; Rettig, W. J.; Schnapp, A. Cell Growth Differ. 2000, 11, 409.
12. Fila, C.; Metz, C.; van der Sluijs, P. J. Biol. Chem. 2008, 283, 21714.
13. Esnault, S.; Braun, R. K.; Shen, Z. J.; Xiang, Z.; Heninger, E.; Love, R. B.; Sandor,
M.; Malter, J. S. PLoS One 2007, 21, e226.
Figure 5. Compound 20 modulates biomarkers dependent upon Pin1 function. (a)
Transfection of siRNA vs. Pin1 reduces expression of cyclin D1, but cyclin E1 levels
are unaffected. (b) 6 h exposure to 20 reduces cyclin D1 levels, but cyclin E levels
are unaffected. (c) 20 min stimulation with insulin increases phosphorylation of
p70S6 kinase on Thr389; this response is blocked by 60 min pre-treatment with 20.
All data shown is from cells cultured in serum-free conditions for at least 24 h
before cells were harvested.
14. Zhang, Y.; Füssel, S.; Reimer, U.; Schutkowski, M.; Fischer, G. Biochemistry 2002,
41, 11868.
15. Wildemann, D.; Erdmann, F.; Alvarez, B. H.; Stoller, G.; Zhou, X. Z.; Fanghänel,
J.; Schutkowski, M.; Lu, K. P.; Fischer, G. J. Med. Chem. 2006, 49, 2147.
16. Zhao, S.; Etzkorn, F. A. Bioorg. Med. Chem. Lett. 2007, 17, 6615.
17. Liu, T.; Liu, Y.; Kao, H.-Y.; Pei, D. J. Med. Chem. 2010, 53, 2494.
18. Bayer, E.; Thutewohl, M.; Christner, C.; Tradler, T.; Osterkamp, F.; Waldmann,
H.; Bayer, P. Chem. Commun. (Camb). 2005, 516.
19. Daum, S.; Erdmann, F.; Fischer, G.; Féaux de Lacroix, B.; Hessamian-Alinejad,
A.; Houben, S.; Frank, W.; Braun, M. Angew. Chem., Int. Ed. 2006, 45, 7454.
20. Guo, C.; Hou, X.; Dong, L.; Dagostino, E.; Greasley, S.; Ferre, R. A.; Marakovits, J.;
Johnson, M. C.; Matthews, D.; Mroczkowski, B.; Parge, H.; VanArsdale, T.;
Popoff, I.; Margosiak, S.; Thomson, J.; Los, G.; Murray, B. W. Bioorg. Med. Chem.
Lett. 2009, 19, 5613.
combinations of surfaces on Pin1, providing opportunities to
improve potency and/or the physiochemical properties of the
compounds so that they occupy drug-like chemical space.
Improvements in both regards will be necessary to identify Pin1
inhibitors suitable for use as drugs: both of our medicinal chemis-
try programs directed against Pin1 have failed to find compounds
21. Dong, L.; Marakovits, J.; Hou, X.; Guo, C.; Greasley, S.; Dagostino, E.; Ferre, R.-A.;
Johnson, C.; Kraynov, E.; Thomson, J.; Pathak, V.; Murray, B. W. Bioorg. Med.
Chem. Lett. 2010, 20, 2210.
capable of modulating Pin1 activity in cells in the sub-lM range.
The nature of the Pin1 active site makes it difficult to optimise hits
into drug-like molecules: one end is dominated by three positively
charged residues, the other end is a narrow yet shallow hydropho-
bic pocket. The ligands we have identified as Pin1 inhibitors all
contain a carboxylate groups and the most potent compounds have
very low permeability in Caco-2 assays.22,31 The only way we have
found to obtain activity in cells has been to add lipophilic groups,
which has compromised the drug-likeness of our compounds. De-
spite our experiences, fragment based approaches to hit identifica-
tion remain attractive route towards finding Pin1 inhibitors,
because a wide region of chemical space can be sampled with a rel-
atively small compound library. Success, however, is contingent on
having suitable initial hits in the library to be screened.
22. Potter, A. J.; Ray, S.; Gueritz, L.; Nunns, C. L.; Bryant, C. J.; Scrace, S. F.;
Matassova, N.; Baker, L.; Dokurno, P.; Robinson, D. A.; Surgenor, A. E.; Davis, B.;
Murray, J. B.; Richardson, C. M.; Moore, J. D. Bioorg. Med. Chem. Lett. 2010, 20,
586.
23. Do, Q.-Q. T.; Guo, C.; Humphries, P. S.; Marakovits, J. T.; Dong, L.; Hou, X.;
Johnson, M. C. PCT Int. Appl. WO2006040646, 2006.
24. Baurin, N.; Aboul-Ela, F.; Barril, X.; Davis, B.; Drysdale, M.; Dymock, B.; Finch,
H.; Fromont, C.; Richardson, C.; Simmonite, H.; Hubbard, R. E. J. Chem. Inf.
Comput. Sci. 2004, 44, 2157.
25. Fischer, G.; Bang, H.; Berger, E.; Schellenberger, A. Biochim. Biophys. Acta 1984,
791, 87.
26. See Supplementary data.
27. McGovern, S. L.; Helfand, B. T.; Feng, B.; Shoichet, B. K. J. Med. Chem. 2003, 46,
4265.
28. Zhang, Y.; Daum, S.; Wildemann, D.; Zhou, X. Z.; Verdecia, M. A.; Bowman, M.
E.; Lücke, C.; Hunter, T.; Lu, K. P.; Fischer, G.; Noel, J. P. ACS Chem. Biol. 2007, 2,
320.
29. Brackeen, M. F.; Stafford, J. A.; Feldmana, P. L.; Karanewsky, D. S. Tetrahedron
Lett. 1993, 34, 7873.
Acknowledgements
30. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug Disc. 2007, 6, 881.
31. Data not shown.
32. Moriya, O.; Minamide, H.; Urata, Y. Synthesis (Stuttgart) 1984, 12, 1057.
33. Ryo, A.; Uemura, H.; Ishiguro, H.; Saitoh, T.; Yamaguchi, A.; Perrem, K.; Kubota,
Y.; Lu, K. P.; Aoki, I. Clin. Cancer Res. 2005, 11, 7523.
The authors thank Dr. Martin Drysdale for valuable discussions,
Heather Simmonite for expert assistance with compound structure
determination by NMR spectroscopy and Dr. Loic Le Strat for chro-
matographic support.
34. Wulf, G. M.; Ryo, A.; Wulf, G. G.; Lee, S. W.; Niu, T.; Petkova, V.; Lu, K. P. EMBO J.
2001, 20, 3459.
35. Lee, N. Y.; Choi, H. K.; Shim, J. H.; Kang, K. W.; Dong, Z.; Choi, H. S.
Carcinogenesis 2009, 30, 671.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2010.09.063.