Thiophene inhibitors of PDE4: Crystal structures show a second binding mode at the catalytic domain of PDE4D2

Article abstract of DOI:10.1016/j.bmcl.2011.09.109

PDE4 inhibitors have been identified as therapeutic targets for a variety of conditions, particularly inflammatory diseases. We have serendipitously identified a novel class of phosphodiesterase 4 (PDE4) inhibitor during a study to discover antagonists of the parathyroid hormone receptor. X-ray crystallographic studies of PDE4D2 complexed to four potent inhibitors reveal the atomic details of how they inhibit the enzyme and a notable contrast to another recently reported thiophene-based inhibitor.

Full text of DOI:10.1016/j.bmcl.2011.09.109

Bioorganic & Medicinal Chemistry Letters 21 (2011) 7089–7093
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Thiophene inhibitors of PDE4: Crystal structures show a second binding mode
at the catalytic domain of PDE4D2
Jacob L. Nankervis a, , Susanne C. Feil b, , Nancy C. Hancock b, Zhaohua Zheng a, Hooi-Ling Ng b,
Craig J. Morton b, Jessica K. Holien b, Patricia W.M. Ho c, Mark M. Frazzetto a, Ian G. Jennings a,
b,d,
David T. Manallack a, T. John Martin c, Philip E. Thompson a, , Michael W. Parker
⇑
⇑
a Medicinal Chemistry & Drug Action, Monash Institute of Pharmaceutical Sciences, Monash University (Parkville campus), 381 Royal Parade, Parkville, Victoria 3052, Australia
b Biota Structural Biology Laboratory, St. Vincent’s Institute of Medical Research, 9 Princes St., Fitzroy, Victoria 3065, Australia
c Bone Cell Biology and Disease Unit, St. Vincent’s Institute of Medical Research, 9 Princes St., Fitzroy, Victoria 3065, Australia
d Department of Biochemistry and Molecular Biology, Bio21 Molecular Science and Biotechnology Institute, The University of Melbourne, Parkville, Victoria 3010, Australia
a r t i c l e i n f o
a b s t r a c t
Article history:
PDE4 inhibitors have been identified as therapeutic targets for a variety of conditions, particularly inflam-
matory diseases. We have serendipitously identified a novel class of phosphodiesterase 4 (PDE4) inhibitor
during a study to discover antagonists of the parathyroid hormone receptor. X-ray crystallographic stud-
ies of PDE4D2 complexed to four potent inhibitors reveal the atomic details of how they inhibit the
enzyme and a notable contrast to another recently reported thiophene-based inhibitor.
Ó 2011 Elsevier Ltd. All rights reserved.
Received 11 August 2011
Revised 21 September 2011
Accepted 21 September 2011
Available online 5 October 2011
Keywords:
Phosphodiesterase inhibitors
PDE4
Thiophene
Crystallography
Gewald reaction
Cyclic 30,50-Nucleotide phosphodiesterase family
4
(PDE4)
the PTH1 receptor might interact with the hormone.6,7 Thus we
used the structure to search a virtual library of 2 million compounds
for small molecule antagonists that might be developed into drugs
to treat humoral hypercalcemia of malignancy and bone metastasis
formation. The assays utilized a cre-luciferase reporter assay in a
PTH/PTHrP-responsive osteosarcoma cell line, UMR106, which
effectively measures the increase in cAMP in response to the GPCR
agonist. In this assay, the effect of ligands in regulating the cre-lucif-
erase response of cells to treatment with PTHrP was monitored. A
large number of compounds yielded marked activation of the lucif-
erase activity and amplified the response of PTHrP itself. One iden-
tified possible mechanism for this increase was inhibition of cellular
PDE activity that would suppress cAMP breakdown, so these com-
pounds were then assessed for PDE inhibition using lysates from
the same cell line (Fig. 1a). While around 15 compounds showed
some level of PDE inhibition, a series of six thiophene derivatives
(1–6, Fig. 1b) from the compound library were identified as PDE
inhibitors, and showed comparable inhibition to rolipram
inhibitors have been identified as therapeutic targets in a variety
of conditions, particularly inflammatory diseases.1 While numer-
ous compounds have not advanced past the clinic, Roflumilast
(DaxasÒ, Nycomed), was recently launched in Europe and is due
for release (as Daliresp™, Forest Laboratories) in North America
in late 2011 for the treatment of chronic obstructive pulmonary
disease (COPD).2 The side effect profiles of many PDE4 inhibitors
related to the archetypal compound rolipram have triggered inter-
est in the development of novel chemotypes that may exhibit an
improved therapeutic window.1,3 In addition,
a number of
structures have been obtained of PDE4-ligand co-crystals,
providing the opportunity for structure-based optimization, with
respect to sub-type selectivity, ADME, toxicity and other important
pharmaceutical properties.4,5
We report on thiophene-based PDE4 inhibitors, identified
serendipitously while searching for antagonists of parathyroid hor-
mone-related protein (PTHrP) interaction with its receptor. After
determining the crystal structure of PTHrP bound to a neutralizing
antibody, we reasoned that the antibody might be mimicking how
(IC50 = 0.6 l
M), implicating PDE4 as the responsible isoform.8 The
compounds also inhibited PDE4 isolated from U937 cells (data not
shown). Compound 5 was the most potent and found to have an
IC50 of approximately 100 nM in these cell lines (Table 1).
⇑
Corresponding authors.
E-mail addresses: phil.thompson@vcp.monash.edu.au (P.E. Thompson), mpar-
ker@svi.edu.au (M.W. Parker).
We sought to determine the elements contributing to inhibitory
potency towards PDE4 by preparing
a range of analogues.
Our general synthetic route, shown in Scheme 1, gave us the
These two authors contributed equally to this work.
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.09.109

7090
J. L. Nankervis et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7089–7093
Table 1
a
PDE4 inhibition by thiophene derivatives
R1
NH
R2
A
S
R1
R2
Inhibition IC50
at 1
(%)
a
Number A
l (lM)
Ma
2
5
CH2
N–CO2Et
CO2Et PhCH2CO
CO2Et PhCH2CO
68
86 (85)
1.5
0.10
(0.10)
10a
10b
10c
10d
CH2
CH2
N–CO2Et
N–CO2Et
CONH2 PhCH2CO
(15)
(24)
72 (73)
98 (70)
CN
PhCH2CO
CO2Et 2-Pyridinylacetyl
CO2Et 2-Thiophenylacetyl
(0.08)
0.022
(0.182)
10e
10f
10g
10h
10i
10j
10k
10l
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
N–CO2Et
CONH2 PhCH2CO
CO2Et PhCO
CO2Et PhCH2
CO2Et 4-Ms-PhCH2CO
CO2Et 2-MeO-PhCH2CO
CO2Et 3-MeO-PhCH2CO
CO2Et 4-MeO-PhCH2CO
CO2Et PhCH2CH2CO
CO2Et 3-Pyridinylacetyl
CO2-c- PhCH2CO
ꢀ1 (18)
(14)
(33)
(ꢀ4)
(37)
(9)
23 (26)
(22)
b
O
O
O
OEt
NH
OEt
NH
OEt
NH
10m
10n
ꢀ20
61
O
S
S
O
S
O
S
O
hexyl
10o
N–CO2Et
CO2-c- 2-Thiophenylacetyl
pentyl
CO2Et PhCH2CO
CO2Et PhCH2CO
CO2Et PhCH2CO
95
0.290
(1.0)
3
2
N–CO2tBu
N–H
N–Ac
N–
38
1
10p
11a
12a
12b
(25)
(50)
65
O
O
O
OEt
NH
OEt
NH
OEt
NH
CO2-c- 2-Pyridinylacetyl
C(O)NHCH2Ph pentyl
N–C(O)NHEt CO2-c- 2-Thiophenylacetyl
pentyl
12c
12d
88
98
N
S
S
O
O
S
EtO2C
O
N–C(O)NHEt CO2Et PhCH2CO
0.027
a
Values from assay of rhPDE4D2. Values from UMR106 lysates are given in
5
4
6
O
parentheses (see Supplementary data for methods).
S
O
N
O
O
R1
R1
Figure 1. (a) Plot of activities for screening compounds in cre-luc reporter assays
and PDE inhibition assay in UMR106 cell lines. Methods described in
Supplementary data. (b) Chemical structures of thiophene based PDE4 inhibitors
1–6 identified by screen.
R1
ii or iii
i
NH2
NH
R2
A
CN
A
A
7
S
S
8
9
10
Scheme 1. Reagents and conditions: (i) S8, morpholine; (ii) R2-X; (iii) RCHO,
NaCNBH3.
opportunity to examine the role of the fused ring, the arylaceta-
mide and the ester functions of the thiophene compounds. Com-
pounds 2, 59 and analogues 10a–t could be prepared by varying
the acylating reagent, the ketone or the malonitrile ester derivative
used (Scheme 1, Table 1, see also Supplementary data).
Where N-Boc-piperid-4-one was the starting ketone, the Boc-
protected product could be readily deprotected with trifluoroacetic
acid to yield the secondary amines 11 and then further derivatized
either as carbamate, acetyl or urea derivatives 12a–d (Scheme 2,
Table 1). For a full description of the compounds prepared, see
the Supplementary data.
R1
R1
R1
ii, iii or iv
i
NH
R2
NH
R2
NH
R2
N
N
A
S
Boc
S
S
H
12
10
11
Scheme 2. Reagents and conditions: (i) TFA/DCM, rt, 1 h; (ii) Ac2O, (iii) Cl-CO2R3;
(iv) R3-NCO.
The compounds were screened in isolated enzyme assays using
the scintillation proximity assay (SPA) format with recombinant
human (rh)PDE4D2 overexpressed from E. coli or against PDE iso-
lated from UMR106 cell lysates. We commenced our studies by
examining the activity of our inhibitors in the UMR106 osteosar-
coma cell line. These cells have been shown to be predominantly
PDE4 isoforms.8 When compared to the results using the recombi-
nant PDE4D2, we did see a change in the rank order of potency of
some compounds. This is not unexpected as inhibitory potency
may be a function of isoform, splice variant and in the case of re-
combinant forms, the presence of regulatory domains.
The PDE4 assays (Table 1) identified the three key elements for
potent inhibition: an ester at the 3-position (R1), an aryl- or hetero-
arylacetamide (not substituted) at the 2-position (R2) and a
piperidinyl carbamate or urea on the fused ring (A). In general,
compounds with fused piperidine rings (such as 5) were preferred
over the carbocycles (such as 2). Esters substituents (methyl to
cyclohexyl) were active, while amides and nitriles were not. The
phenylacetyl group could not be substituted, lengthened or short-
ened without diminished activity, but compounds with heteroaryl

J. L. Nankervis et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7089–7093
7091
Figure 2. Dose dependent activation of luciferase activity by 10o in transfected
UMR106 cells. (a) Luminescence (relative light units) due to treatment of cells with
10o alone at 10–10000 nM. (b) Treatment of cells with 1 nM PTHrP plus 10o at
10–10000 nM. Methods described in Supplementary data.
Figure 3. Crystal structure of the catalytic domain of PDE4D2. (a) The tetramer.
Monomers A, B, C and D are colored in yellow, green, red and blue, respectively. Zn
ions are shown as magenta spheres. (b) The monomer. Helices 1–7 are colored in
blue, helices 8–11 are colored in green and helices 12–16 are in colored yellow. Zn
ions are in red.
groups, such as 2-thiophenacetyl (10d) and 2-pyridinylacetyl (10c)
showed strong PDE4 inhibition. Finally, the carbamate could be
substituted by ureas, (12c, 12d) but other changes were unfavor-
able (12a).
The cellular activity of the compounds was demonstrated by the
use of the reporter assay described above. For example, 10o
showed potent dose-dependent amplification of the luminescent
readout both from basal and PTHrP-stimulated cells. (Fig. 2) The
other compounds of the series also showed cellular effects consis-
tent with the observed PDE4 inhibition. These data yielded a pic-
ture of a molecular interaction of ligands with the enzyme that
was well optimized. A binding model was proposed analogous to
ethyl ester derivatives described by Card et al.10 where the invari-
ant, purine-scanning glutamine of PDE4 interacted with the ethyl
ester group, and the thiophene was acting as an isostere for the
purine ring of cAMP. Beyond that the binding pose at the active site
was not obvious.
in complex with four thiophene inhibitors; 10d, 10o, 12c, 12d. The
overall structures of the complexes were very similar to the struc-
ture of apo-PDE4D2, with no overall significant conformational
changes or any significant perturbations of residues in the binding
pocket due to the binding of the ligands. Superposition of the tet-
ramers of the 10d, 10o, 12c, 12d complex structures onto the apo
structure, gives a rmsd between the alpha carbons of 0.6, 0.4, 0.3,
and 0.3 Å, respectively. There is some variation in the orientation
of the side-chain of the active site residue Met 357 in all structures.
This amino acid appears to be flexible and does not have electron
density present in all monomers. The PDE4 inhibitors were located
in the nucleotide binding pocket and have fundamentally identical
binding poses (Fig. 4). Overall there are between 60 and 72 van der
Waals contacts between the protein and the ligands. The best
inhibitors of the PDE4D2, 12d and 10d, show the most clearly de-
fined electron density.
The catalytic domain of PDE4D2 was expressed, purified and
crystallized using modifications of the methods published by Huai
et al.11 The crystal structure of apo-PDE4D2, determined at 2.1 Å
resolution, is composed of four monomers assembled into a tetra-
mer in the asymmetric unit (Fig. 3a). The monomer structures were
largely identical to those previously described with 16
divided into three sub-domains.11 The first sub-domain is composed
of -helices H1–H7, the second is formed by -helices H8–H11 and
the third sub-domain consists of -helices H12–H16 (Fig. 3b). The
intersection of the three sub-domains forms the binding site for
substrates and inhibitors.4,11,12 Two Zn ions were found within
the active site. We were also successful in crystallizing the enzyme
a
-helices
As anticipated, the carbonyl of the ester group for each of the
inhibitors forms the key interaction with Gln 369, and the ethyl
or cyclopentyl groups are located in a small hydrophobic pocket
lined by Met 337 and Met 357. The ester carbonyl, acetamide
and thiophene form a planar surface for a stacking interaction with
Phe 372 (Fig. 4a–d). The orientation of the aryl ring varies slightly
in the four structures to accommodate differences in ligand
a
a
a

7092
J. L. Nankervis et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7089–7093
a
c
b
d
O
O
O
O
NH
O
N
NH
S
O
S
O
N
S
O
S
O
O
10d
10o
O
O
O
O
H
N
NH
H
N
NH
N
S
N
O
S
O
S
O
O
12 d
12c
Figure 4. Structures and binding mode of each compound in the active site of PDE4D2. (a) Compound 10d, (b) Compound 10o, (c) Compound 12c and (d) Compound 12d.
Water molecules are shown in grey. PDB codes: Apo form: 3SL3; 10d: 3SL4; 10o: 3SL8; 12c: 3SL6; 12d: 3SL5.
binding. The carbonyl of the 20-acetamide of the thiophene forms
water-mediated hydrogen bonds to the side-chain of Tyr 159 in
all cases and water-mediated hydrogen bonds to the main- and/
or side-chain oxygens of Asp 318, depending on the ligand and
monomer subunit. The arylacetamide group is positioned deep in
the binding pocket, with the 20-pendant aryl ring folding back to-
wards the ester group, filling a space in the nucleotide binding
pocket not previously occupied by any reported ligand in a PDE4
crystal structure.
The influence of the carbamate and the urea substituents, which
are important to inhibitory potency, is not obvious from the four
crystal structures. Indeed, the density for the ethyl group is poorly
defined in the carbamate of both 10d and 10o and the urea deriv-
atives, 12c and 12d. Depending on the ligand and monomer unit,
the carbonyl can project either towards Ile 336 (10d), or Ile 376
(12c, 12d), whereas in other (12c) monomer subunits it is orien-
tated towards residues Phe 340 and Cys 358.
The structure–activity relationships were in excellent agree-
ment with the crystal structures in several ways. Firstly, the impor-
tance of the ester function to activity is apparent as it makes a
pivotal hydrogen bond with the carboxamide of the conserved
Gln 369, which has been a hallmark of numerous inhibitors such
as rolipram and roflumilast and also the pyrazole esters described
by Card et al.10 Interestingly, the ortho placement of the ester and
Figure 5. Thiophene complex crystal structures. (a) Structure of reported tetra-
the acetamide on the thiophene ring makes a planar arrangement
that stacks against the ring of the phenylalanine. The ethyl or
cyclopentyl groups of the ester fill the space occupied by the cyclo-
pentyl ring of rolipram. The carbonyl of the acetamide makes
hydrogen bond networks via water molecules to an aspartyl resi-
due, Asp 318.
Second, the specific requirement of the arylacetamide is ex-
plained by its folded conformation in the binding pocket, which
would not be expected to tolerate shortening to an arylamide or
lengthening to a phenethylcarboxamide chain. The poor activity
of substituted-phenyl derivatives is also explained in this context
as there is very little available space. The location of the
hydrobenzothiophene inhibitor, 1314 (b) superposition of PDE4B/D thiophene
complex crystal structures: yellow 10d, magenta: 13 PDB:3HMV.
arylacetamide in the binding pocket is also of interest. Comparison
to all reported PDE4 structures described to date shows that this is
an unprecedented binding motif for PDE4 inhibitors.
What is less well explained by the crystal structures is the im-
proved potency of the ethyl carbamate and ethyl urea derivatives
relative to the carbocycle in 2. The generally poor density and var-
iable positioning of these groups suggests flexibility at this region.
The improved potency may best be explained as a presentation of a

J. L. Nankervis et al. / Bioorg. Med. Chem. Lett. 21 (2011) 7089–7093
7093
more hydrophilic region to the aqueous environment at the bind-
ing pocket entrance.
to T.J.M. and M.W.P, the CASS Foundation (P.E.T and T.J.M.) and
funding from the Victorian Government Operational Infrastructure
Support Scheme to St Vincent’s Institute. J.L.N. was supported by
an Australian Post-graduate Award. S.C.F. was supported by a
NHMRC Industry Fellowship. M.W.P. is an Australian Research
Council Federation Fellow and NHMRC Honorary Fellow.
Although the thiophene class and particularly the 2-amino-
3-carboxythiophenes have been shown to display a number of
biological activities,13 our identification of PDE4 inhibition by this
thiophene class coincided with workers at Glaxo who also
published a series of analogous thiophene derivatives, but interest-
ingly the compounds showed markedly different SAR to our own
series and in particular were poor inhibitors of PDE4D relative to
PDE4B.14 Consistent with that, the crystal structure of PDE4B com-
plexed to the GSK inhibitor 13 shows that the thiophene molecule
binds in a completely different orientation, (Fig. 5), with virtually
no overlap of the compounds in the active site. So while the struc-
tural analogy of the inhibitor is obvious, they share nothing in
common from the view of the PDE4 catalytic site.
Supplementary data
Supplementary data (selected enzyme inhibition curves, crys-
tallographic data and experimental procedures) associated with
this article can be found, in the online version, at doi:10.1016/
j.bmcl.2011.09.109.
References and notes
In conclusion, here we have shown the structure–activity rela-
tionships (SAR) and crystallographically determined binding poses
for a series of novel PDE4 inhibitors, one of which was discovered
serendipitously via a cre-luciferase based reporter assay. The class
falls into the category of non-rolipram related PDE4 inhibitors,
which may be of some significance given the perceived association
between the rolipram class and a narrow therapeutic window. The
compounds show an interesting and in some ways unprecedented
binding interaction as observed crystallographically, that is sup-
ported by SAR studies. We hope to use this information further
in the development of clinically or pharmacologically useful
PDE4 inhibitors.
1. Kodimuthali, A.; Jabaris, S. S.; Pal, M. J. Med. Chem. 2008, 51, 5471.
2. Hatzelmann, A.; Schudt, C. J. Pharmacol. Exp. Ther. 2001, 297, 267.
3. Giembycz, M. A. Curr. Opin. Pharmacol. 2005, 5, 238.
4. Burgin, A. B.; Magnusson, O. T.; Singh, J.; Witte, P.; Staker, B. L.; Bjornsson, J. M.;
Thorsteinsdottir, M.; Hrafnsdottir, S.; Hagen, T.; Kiselyov, A. S.; Stewart, L. J.;
Gurney, M. E. Nat. Biotechnol. 2010, 28, 63.
5. Ke, H.; Wang, H. Curr. Top. Med. Chem. 2007, 7, 391.
6. McKinstry, W. J.; Polekhina, G.; Diefenbach-Jagger, H.; Ho, P. W.; Sato, K.;
Onuma, E.; Gillespie, M. T.; Martin, T. J.; Parker, M. W. J. Biol. Chem. 2009, 284,
15557.
7. McKinstry, W. J.; Polekhina, G.; Diefenbach-Jagger, H.; Sato, K.; Onuma, E.;
Gillespie, M. T.; Martin, T. J.; Parker, M. W. Acta Crystallogr., Sect. F: Struct. Biol.
Cryst. Commun. 2009, 65, 336.
8. Ahlstrom, M.; Lamberg-Allardt, C. Biochem. Pharmacol. 1999, 58, 1335.
9. Castanedo, G. M.; Sutherlin, D. P. Tetrahedron Lett. 2001, 42, 7181.
10. Card, G. L.; Blasdel, L.; England, B. P.; Zhang, C.; Suzuki, Y.; Gillette, S.; Fong, D.;
Ibrahim, P. N.; Artis, D. R.; Bollag, G.; Milburn, M. V.; Kim, S. H.; Schlessinger, J.;
Zhang, K. Y. Nat. Biotechnol. 2005, 23, 201.
Acknowledgments
11. Huai, Q.; Colicelli, J.; Ke, H. Biochemistry 2003, 42, 13220.
12. Wang, H.; Peng, M. S.; Chen, Y.; Geng, J.; Robinson, H.; Houslay, M. D.; Cai, J.;
Ke, H. Biochem. J. 2007, 408, 193.
This research was undertaken on the MX1 beamline at the
Australian Synchrotron, Victoria, Australia. We thank the beamline
scientists at the Australian Synchrotron for their help during data
collection. This work was partly supported by grants from the Na-
tional Health and Medical Research Council of Australia (NHMRC)
13. Huang, Y.; Doemling, A. Mol. Diversity 2011, 15, 3.
14. Kranz, M.; Wall, M.; Evans, B.; Miah, A.; Ballantine, S.; Delves, C.; Dombroski,
B.; Gross, J.; Schneck, J.; Villa, J. P.; Neu, M.; Somers, D. O. Bioorg. Med. Chem.
2009, 17, 5336.