Novel thienopyrrole glycogen phosphorylase inhibitors: Synthesis, in vitro SAR and crystallographic studies

Article abstract of DOI:10.1016/j.bmcl.2006.08.047

Two series of novel thienopyrrole inhibitors of recombinant human liver glycogen phosphorylase a (GPa) which are effective in reducing glucose output from rat hepatocytes are described. Representative compounds have been shown to bind at the dimer interface site of the rabbit muscle enzyme by X-ray crystallography.

Full text of DOI:10.1016/j.bmcl.2006.08.047

Bioorganic & Medicinal Chemistry Letters 16 (2006) 5567–5571
Novel thienopyrrole glycogen phosphorylase inhibitors:
Synthesis, in vitro SAR and crystallographic studies
Paul R. O. Whittamore,a,* Matthew S. Addie,a Stuart N. L. Bennett,a Alan M. Birch,a
Michael Butters,b Linda Godfrey,a Peter W. Kenny,a Andrew D. Morley,a
Paul M. Murray,b Nikos G. Oikonomakos,c Ludovic R. Otterbein,a
Andrew D. Pannifer,a, Jeremy S. Parker,b Kristy Readman,a Pawel S. Siedlecki,a
Paul Schofield,a Andy Stocker,a Melvyn J. Taylor,a Linda A. Townsend,a
David P. Whalleya and Jennifer Whitehouseb
aAstraZeneca, Alderley Park, Macclesfield, Cheshire, SK10 4TG, UK
bAstraZeneca, Avlon Works, Bristol BS10 7ZE, UK
cInstitute of Organic and Pharmaceutical Chemistry, The National Hellenic Research Foundation,
48, Vas. Constantinou Ave., 116 35 Athens, Greece
Received 28 February 2006; revised 7 July 2006; accepted 5 August 2006
Available online 30 August 2006
Abstract—Two series of novel thienopyrrole inhibitors of recombinant human liver glycogen phosphorylase a (GPa) which are effec-
tive in reducing glucose output from rat hepatocytes are described. Representative compounds have been shown to bind at the dimer
interface site of the rabbit muscle enzyme by X-ray crystallography.
Ó 2006 Elsevier Ltd. All rights reserved.
Type 2 diabetes is a complex disease caused by defects in
both the action and secretion of insulin, leading to fast-
ing hyperglycaemia and vascular complications. It has
become the fourth leading cause of death and affects
150 million people worldwide, with its prevalence
expected to double by the year 2025.1 Current therapies
do not achieve adequate glycaemic control,2 hence there
is a need for new, effective pharmacological agents.
synthetic ligands.4–7 Of particular interest is an inhibito-
ry allosteric binding site8 at the dimer interface for
which a series of synthetic ligands has been identified,4
including ingliforib (CP368296) which has undergone
phase II clinical evaluation. We report here the identifi-
cation of novel thienopyrrole inhibitors, which we have
shown by X-ray crystallography to bind at the allosteric
site situated at the dimer interface.
Glycogen phosphorylase (GP) catalyses the breakdown
of glycogen to glucose-1-phosphate. In the liver this is
metabolised further to glucose, which is then secreted
into the bloodstream. One approach to reducing hyper-
glycaemia in Type 2 diabetes is to seek to restore normal
net hepatic glucose balance by inhibition of hepatic GP.3
The more active phosphorylated form (GPa) of the en-
zyme is a homodimer; its activity is modulated by phos-
phorylation status and by a number of endogenous and
O
O
Cl
OH
OH
N
H
N
N
H
OH
We prepared a series of inhibitors derived from thie-
no[2,3-b]pyrrole and thieno[3,2-b]pyrrole carboxamides
and compared their activity to the known 5-chloroin-
dole 1 (Table 1). The isolated enzyme and cell assay val-
ues found for 1 match well with those from the
literature.4
Keywords: Glycogen phosphorylase; Diabetes; Thienopyrrole.
*
Corresponding author. Tel.: +44 (0) 1625 513389; fax: +44 (0) 1625
516667; e-mail: paul.whittamore@astrazeneca.com
Present address: Pfizer UK, Sandwich.
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.08.047

5568
P. R. O. Whittamore et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5567–5571
Table 1. Thienopyrrole SAR
Replacement of the chloroindole with the unsubstituted
thieno[2,3-b]pyrrole system 2 led to a novel inhibitor,
albeit with a considerably reduced inhibitory activity–
however, this could be restored by introduction of a
2-chloro or 2-bromo (3 and 4) substituent. This is con-
sistent with the halothienopyrrole moiety binding in a
similar manner to that reported for the chloroindole,8
that is, binding deeply into a hydrophobic pocket at
the interface site bounded by Leu 63, Pro 229 and Trp
67. Interestingly, dibromo substitution (5) led to a
marked reduction in inhibition, possibly due to a severe
steric clash between the additional bromine atom and
R
NMe2
O
Compound
R
Enzyme
inhibitiona
b
IC50 nM
Cellc
b
IC50 lM
H
Cl
N
1
51
2.6
N
H
O
˚
the protein (possibly Val64, which is 3.92 A from the
chloroindole).
O
2
3
2310
17
S
N
H
N
H
SAR in the isomeric thieno[3,2-b]pyrrole series is subtly
different, with very little activity seen for the unsubsti-
tuted compound (6), moderate activity restored by
monohalogenation (7 and 8), 2,3-dibromo (9) now toler-
ated and optimal activity found with 2,3-dichloro substi-
tution (10). The alternative inverted orientation of the
thiophene ring in this series appears to result in the sec-
ond halo atom being directed towards a less sterically
demanding region of the protein. Methylation on either
the pyrrole nitrogen (11) or the carboxamide group (12)
leads to loss of activity, consistent with the key H-bonds
the unsubstituted groups make to the backbone carbon-
yl groups of Asn 187 and Thr 38, respectively.
O
Cl
Br
1.4
1.6
S
N
H
N
H
O
4
5
31
S
S
N
H
N
H
Br
O
20% at 10 lM
Br
N
H
N
H
H
Exploration of a variety of amide N-substituents by
library synthesis led to the discovery of two novel struc-
tural types with promising activity (Table 2). The 3,4-
dihydro-2-quinolone moiety led to potent inhibitor
compounds in both the thieno[3,2-b]pyrrole (14) and
the thieno[2,3-b]pyrrole (15) series, against the enzyme
and also against glucose output from rat hepatocytes.
N
S
6
7
35% at 100 lM
N
O
H
H
N
S
577
Cl
N
H
O
H
Other workers have reported the identification of the
corresponding 5-chloroindole (13) derivative,5 which
we found to show enzyme potency comparable to the
thienopyrroles, but apparently reduced potency in rat
hepatocytes.
N
S
8
390
N
O
H
Cl
S
Br
Cl
H
N
9
23
5
Indan-2-yl was also identified as a modestly active amide
N-substituent (16) in the chlorothieno[2,3-b]pyrrole
series with a trans 1-hydroxy substituent leading to a
slightly more active inhibitor (17). The dichlorothie-
no[3,2-b]pyrrole analogues 18 and 19 showed an
increase in potency more marked than for the previous
amide N-substituents. Halo substitution in the indan
aromatic ring leads to a general reduction in activity
(compounds 20 and 21), consistent with the restricted
space in the aromatic binding pocket.
N
Br
H
O
S
H
N
10
1.0
1.1
N
H
Cl
O
S
S
Br
Br
H
N
11
12
>10,000
>10,000
N
Br
Br
O
Me
Me
N
In order to understand the binding mode of these new
inhibitors, they were soaked into crystals of rabbit mus-
cle GPb (non-phosphorylated form). The complexes
gave well-defined high-resolution structures.10
N
H
O
a Using recombinant human liver GPa: glucose-1-phosphate produc-
tion from glycogen monitored by a multienzyme coupled assay.9
b n = P3 unless otherwise stated.
The structure of 15 (Fig. 1A) showed similarities to
structures previously reported for 5-chloroindole deriva-
tives, with two molecules of inhibitor per molecule of
GPb dimer, and the thieno[2,3-b]pyrrole group binding
c Inhibition of glucose output from primary rat hepatocytes following
glucagon challenge9 determined from seven concentration points
unless otherwise stated.

P. R. O. Whittamore et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5567–5571
5569
Table 2. Amide N-substituents
Compound Structure
Enzyme
Cellc
inhibitiona IC50
b
b
IC50 nM lM
Cl
O
N
7d
N
13
86
N
H
H
H
O
(+/-)
S
Cl
O
N
H
N
H
Cl
14
132
0.56
0.69
NH
O
(+/-)
O
Cl
N
H
S
N
15
16
17
41
806
319
N
H
H
O
(+/-)
Cl
Cl
Figure 1. Inhibitor environment in the crystal structure of rabbit
muscle glycogen phosphorylase complexed with compound 15 (A) or
with compound 17 (B). Residues from one monomer are coloured
pink; residues from the other blue. The figures were prepared using
PyMol (DeLano Scientific).
H
N
S
N
H
O
OH
H
N
S
in the same hydrophobic binding pocket as the indole.
This pocket is formed from Val40 (from the one mono-
mer) and Val64, Arg60, Leu63, Trp67 and Pro229 (from
the other monomer). In addition, the guanidinium
group of Arg60 makes a favourable electrostatic interac-
tion with the indole p-system. The amide carbonyl
group forms a water-bridged H-bond to Asn187 and
the amide NH interacts with the backbone carbonyl of
Thr38, from the second protein subunit.
N
H
3.6
O
(+/-)
S
Cl
Cl
H
N
18
19
20
197
N
Cl
H
O
The structure of compound 17 (Fig. 1b; R,R-diaste-
reoisomer) bound to the enzyme shows many similar
features to those described above, the binding mode
of the thieno[2,3-b]pyrrole-5-carboxamide system being
essentially identical. The aromatic ring of the indan
system binds in the hydrophobic pocket in a similar
way to that described above for the dihydroquinolone.
Again, H-bonding interaction is seen between the two
inhibitor molecules, but in this case it is direct rather
HO
S
H
N
9
1.5
1.3
N
H
Cl
O
(+/-)
HO
S
Cl
H
N
F
49
N
H
Cl
O
(+/-)
CHO
COOMe
a)
S
HO
S
S
Cl
Cl
N3
H
N
Cl
Cl
21
90
N
H
Cl
Cl
O
Cl
(+/-)
b)
a Using recombinant human liver GPa: glucose-1-phosphate produc-
tion from glycogen monitored by a multienzyme coupled assay.9
b n = P3 unless otherwise stated.
S
S
CONR1R2
Cl
COOMe
Cl
N
H
N
H
c Inhibition of glucose output from primary rat hepatocytes following
glucagon challenge9 determined from 7 concentration points unless
otherwise stated.
Cl
Cl
Scheme 1. Reagents and conditions: (a) NaOMe, MeO2CCH2N3,
MeOH; (b) reflux toluene.
d Determined from three concentration points.

5570
P. R. O. Whittamore et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5567–5571
than water mediated, presumably with each hydroxy
group acting both as donor and acceptor. The hydro-
xy groups are seen to make a further, water-bridged
H-bond to His57.
quently inserts into the thiophene CH bond to afford
the desired ring system. As we became more interested
in the dichlorothieno[3,2-b]pyrrole and chlorothie-
no[2,3-b]pyrrole systems as the basis for a lead optimiza-
tion programme, we chose to develop novel alternative
routes12,13 that did not involve the potential explosive
hazards associated with vinyl azides in general. These
are shown in Schemes 2 and 3, and briefly comprise
key steps of Curtius rearrangement, metallation/formy-
lation and then aldol cyclisation.
The general synthetic routes to thieno[2,3-b]pyrrole and
thieno[3,2-b]pyrrole carboxamide inhibitors are shown
in Schemes 1–3.
The initial literature route we employed is illustrated for
the dichlorothieno[3,2-b]pyrrole.11 This involves the
thermal decomposition of an intermediate vinyl azide
(NOT isolated) to a postulated nitrene which subse-
In summary, we have identified series of novel inhibitors
of glycogen phosphorylase and have described SAR,
O
CHO
Cl
CHO
Cl
COOH
O
Cl
HN
a
b
c
S
Cl
Cl
S
S
Cl
S
d
O
O
N
O
O
Cl
N
O
Cl
HN
Cl
f
e
COOMe
CHO
COOMe
Cl
Cl
CHO
S
S
Cl
CHO
S
g
Cl
Cl
Cl
H
H
N
H
h)
i)
N
N
COOMe
Cl
CONR1R2
Cl
COOH
Cl
S
S
S
Scheme 2. Reagents and conditions: (a) AlCl3, Cl2, DCM 20 °C; (b) NaOH, KMnO4, 50 °C; (c) DPPA, Et3N, t-BuOH reflux 12 h; (d) n-BuLi, THF,
DMF, À78 °C to rt; (e) KHCO3, methyl bromoacetate, DMF, 60 °C 3.5 h; (f) AcOH, Ac2O, 120 °C 21 h; (g) K2CO3, DMF, 60 °C 3 h; (h) MeOH,
LiOH, H2O; (i) standard amide coupling conditions.
O
O
a)
b)
O
Cl
COOH
Cl
N
Cl
N
S
S
S
H
H
c)
CHO
CHO
O
d)
O
e)
COOMe
Cl
Cl
N
S
S
N
Cl
N
S
H
H
COOMe
f)
g)
COOH
Cl
CONR1R2
Cl
S
N
S
N
H
H
Scheme 3. Reagents and conditions: (a) DPPA, Et3N, t-BuOH, reflux 12 h; (b) Ac2O, AcOH, 120 °C 4 h; (c) DMF, POCl3, DCM, rt 24 h; (d)
K2CO3, methyl bromoacetate, NMP, t-BuOMe, 40 °C 24 h; (e) DMF, K2CO3, 60 °C 90 min; (f) NaOH, MeOH, H2O, reflux; (g) standard amide
coupling conditions.

P. R. O. Whittamore et al. / Bioorg. Med. Chem. Lett. 16 (2006) 5567–5571
5571
8. Oikonomakos, N. G.; Skamnaki, V. T.; Tsitsanou, K. E.;
Gavalas, N. G.; Johnson, L. N. Structure 2000, 8, 575;
Rath, V. L.; Ammirati, M.; Danley, D. E.; Ekstrom, J. L.;
Gibbs, E. M.; Hynes, T. R.; Mathiowetz, A. M.; McPh-
erson, R. K.; Olson, T. V.; Treadway, J. L.; Hoover, D. J.
Chem. Biol. 2000, 7, 677.
hepatocyte activity, synthetic routes and demonstrated
binding at the dimer interface of the rabbit muscle en-
zyme. Lead optimization in these series and pharmaco-
logical profiling of leading compounds will be reported
in due course.
9. For in vitro assay experimental procedures, see: Bartlett, J.
B.; Freeman, S.; Kenny, P. W.; Morley, A. D.; Whitta-
more, P. R. O. PCT Intl Application WO200220530, 2002;
Chem. Abstr. 2002, 136, 247485.
Acknowledgments
Thanks to Sue Freeman, Simon Poucher, Gemma Con-
vey, Julie Bartlett and Jenny Thomas for the bioscience
testing. Richard Pauptit is thanked for establishing the
crystallography collaboration with the Oikonomakos
group, who have kindly provided protein and crystals.
Thanks to Claire Minshull for the soaking experiments
10. Protein crystals kindly provided by N.G.O. were grown
using published procedures.14 Diffraction data were
collected at SRS, Daresbury, at 100 K for compound
17 and at ESRF, Grenoble, at 100 K for compound 15.
The crystals of the complex with 15 have space group
˚
P43212, unit cell 126.8, 126.8, 115.1 A. 39468 unique
˚
reflections to 2.3 A give 97.7% completeness. The final
R-factor is 18.3% (Rfree using 5% of data is 24.5%).
ˆ
and to Jason Breed, Richard Pauptit, Sian Rowsell and
Julie Tucker for synchrotron data collection and analysis.
2
The mean temperature factor is 31.5 A for protein
˚
2
atoms and 47.8 A for the ligand. Crystals of the
˚
complex with 17 have space group P43212 and unit cell
˚
˚
References and notes
128.1, 128.1, 116.3 A. 98,023 unique reflections to 1.6A
give 74.5% completeness. The final model has an R-
factor of 18.7% (Rfree using 5% of data is 22.6%).
Mean atomic temperature factor for the protein is 18.4
1. Gershell, L. Nat. Rev. Drug Disc. 2005, 4, 367.
2. Krentz, A. J.; Bailey, C. J. Drugs 2005, 65, 385.
3. Treadway, J. L.; Mendys, P.; Hoover, D. J. Exp. Opin.
Invest. Drugs 2001, 10, 439.
4. Hoover, D. J.; Lefkowitz-Snow, S.; Burgess-Henry, J. L.;
Martin, W. H.; Armento, S. J.; Stock, I. A.; McPherson,
R. K.; Genereux, P. E.; Gibbs, E. M.; Treadway, J. L. J.
Med. Chem. 1998, 41, 2934.
5. Rosauer, K. G.; Ogawa, A. K.; Willoughby, C. A.;
Ellsworth, K. P.; Geissler, W. M.; Myers, R. W.; Deng,
Q.; Chapman, K. T.; Harris, G.; Moller, D. E. Bioorg.
Med. Chem. Lett. 2003, 13, 4385.
2
˚
and for the inhibitor is 13.7 A . Data analysis and
structure solution used programs from the CCP4 suite15.
The inhibitors were modelled into electron density using
Quanta2000 (Accelrys). The model was refined using
CNX(Accelrys) and Refmac5.15 Protein Data Bank
deposition codes for the refined structures 15 and 17
are 2GM9 and 2GJ4, respectively.
´
´
11. Eras, J.; Galvez, C.; Garcıa, F. J. Heterocycl. Chem. 1984,
21, 215; Hemetsberger, H.; Knittel, D. Monatsh. Chem.
1972, 103, 194.
12. Scheme 2—for further details, see: Butters, M.; Schofield,
P.; Stocker, A. PCT Intl Application WO2004031193,
2004; Chem. Abstr. 2004, 140, 339190.
13. Scheme 3—for further details, see: Murray, P. M.;
Parker, J. S.; Schofield, P.; Stocker, A. PCT Intl
Application WO2004031194, 2004; Chem. Abstr. 2004,
140, 339191.
14. Oikonomakos, N. G.; Melpidou, A. E.; Johnson, L. N.
Biochim. Biophys. Acta 1985, 832, 248.
15. Collaborative Computing Project, Number 4 Acta Crys-
tallogr. 1994, D50, 760.
6. Kristiansen, M.; Andersen, B.; Iversen, L. F.; Westerg-
aard, N. J. Med. Chem. 2004, 47, 3537; Jakobsen, P.;
Lundbeck, J. M.; Kristiansen, M.; Breinholt, J.; Demuth,
H.; Pawlas, J.; TorresCandela, M. P.; Andersen, B.;
Westergaard, N.; Lundgren, K.; Asano, N. Bioorg. Med.
Chem. 2001, 9, 733.
7. Klabunde, T.; Wendt, K. U.; Kadereit, D.; Brachvogel,
V.; Burger, H.-J.; Herling, A. W.; Oikonomakos, N. G.;
Kosmopoulou, M. N.; Schmoll, D.; Sarubbi, E.; Von
Roedern, E.; Schoenafinger, K.; Defossa, E. J. Med.
Chem. 2005, 48, 6178.