The binding of β-d-glucopyranosyl-thiosemicarbazone derivatives to glycogen phosphorylase: A new class of inhibitors

Article abstract of DOI:10.1016/j.bmc.2010.09.039

Glycogen phosphorylase (GP) is a promising target for the treatment of type 2 diabetes. In the process of structure based drug design for GP, a group of 15 aromatic aldehyde 4-(β-d-glucopyranosyl)thiosemicarbazones have been synthesized and evaluated as inhibitors of rabbit muscle glycogen phosphorylase b (GPb) by kinetic studies. These compounds are competitive inhibitors of GPb with respect to α-d-glucose-1-phosphate with IC₅₀ values ranging from 5.7 to 524.3 μM. In order to elucidate the structural basis of their inhibition, the crystal structures of these compounds in complex with GPb at 1.95-2.23 resolution were determined. The complex structures reveal that the inhibitors are accommodated at the catalytic site with the glucopyranosyl moiety at approximately the same position as α-d-glucose and stabilize the T conformation of the 280s loop. The thiosemicarbazone part of the studied glucosyl thiosemicarbazones possess a moiety derived from substituted benzaldehydes with NO₂, F, Cl, Br, OH, OMe, CF₃, or Me at the ortho-, meta- or para-position of the aromatic ring as well as a moiety derived from 4-pyridinecarboxaldehyde. These fit tightly into the β-pocket, a side channel from the catalytic site with no access to the bulk solvent. The differences in their inhibitory potency can be interpreted in terms of variations in the interactions of the aldehyde-derived moiety with protein residues in the β-pocket. In addition, 14 out of the 15 studied inhibitors were found bound at the new allosteric site of the enzyme.

Full text of DOI:10.1016/j.bmc.2010.09.039

Bioorganic & Medicinal Chemistry 18 (2010) 7911–7922
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
The binding of b-D-glucopyranosyl-thiosemicarbazone derivatives
to glycogen phosphorylase: A new class of inhibitors
Kyra-Melinda Alexacou a,b, Alia-Cristina Tenchiu (Deleanu) a, Evangelia D. Chrysina a,
a
a,
Maria-Despoina Charavgi a, Ioannis D. Kostas a, , Spyros E. Zographos , Nikos G. Oikonomakos
,
⇑
Demetres D. Leonidas a,c,
⇑
a Institute of Organic and Pharmaceutical Chemistry, National Hellenic Research Foundation, 48 Vassileos Constantinou Avenue, 11635 Athens, Greece
b Department of Biology, Chemistry and Pharmacy, Freie Universität Berlin, Takustr. 3, 14195 Berlin, Germany
c Department of Biochemistry and Biotechnology, University of Thessaly, 26 Ploutonos Str., 41221 Larissa, Greece
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 11 June 2010
Revised 9 September 2010
Accepted 16 September 2010
Available online 22 September 2010
Glycogen phosphorylase (GP) is a promising target for the treatment of type 2 diabetes. In the process of
structure based drug design for GP, a group of 15 aromatic aldehyde 4-(b- -glucopyranosyl)thiosemicar-
bazones have been synthesized and evaluated as inhibitors of rabbit muscle glycogen phosphorylase b
(GPb) by kinetic studies. These compounds are competitive inhibitors of GPb with respect to -glu-
cose-1-phosphate with IC50 values ranging from 5.7 to 524.3 M. In order to elucidate the structural basis
D
a
-D
l
of their inhibition, the crystal structures of these compounds in complex with GPb at 1.95–2.23 Å reso-
lution were determined. The complex structures reveal that the inhibitors are accommodated at the cat-
Keywords:
Type 2 diabetes
Glycogen phosphorylase
Glucopyranosyl-thiosemicarbazones
Inhibition
alytic site with the glucopyranosyl moiety at approximately the same position as
a-D-glucose and
stabilize the T conformation of the 280s loop. The thiosemicarbazone part of the studied glucosyl thio-
semicarbazones possess a moiety derived from substituted benzaldehydes with NO2, F, Cl, Br, OH,
OMe, CF3, or Me at the ortho-, meta- or para-position of the aromatic ring as well as a moiety derived from
4-pyridinecarboxaldehyde. These fit tightly into the b-pocket, a side channel from the catalytic site with
no access to the bulk solvent. The differences in their inhibitory potency can be interpreted in terms of
variations in the interactions of the aldehyde-derived moiety with protein residues in the b-pocket. In
addition, 14 out of the 15 studied inhibitors were found bound at the new allosteric site of the enzyme.
Ó 2010 Elsevier Ltd. All rights reserved.
X-ray crystallography
1. Introduction
co-factor pyridoxal-50-phosphate. It is regulated allosterically and
by phosphorylation and it exists in two forms: the unphosphory-
Diabetes type 2 is a complex disease characterized by insulin
resistance and abnormal insulin secretion. It is the most common
type of diabetes. It is estimated that the number of patients suffer-
ing from type 2 diabetes will reach 300 million by the year 2025.1
Glycogen phosphorylase (GP) catalyzes the first step of the intra-
cellular degradation of glycogen, which is an important source of
hepatic glucose production. Thus, inhibition of GP has emerged
as a potential therapeutic strategy for type 2 diabetes.2–4 GP is a
dimer composed of two identical subunits associated with the
lated T state GPb and the phosphorylated R state GPa.4 In the last
decade GP has been the target for inhibitor design5 and many com-
pounds have been reported to bind to four distinct sites, the cata-
lytic, the allosteric, the new allosteric (or indole), and the inhibitor
(or caffeine site).4 The efficacy of such inhibitors on blood glucose
control and hepatic glycogen balance has been confirmed from ani-
mal studies and in vitro cell biology experiments.5–7
The catalytic site of GP is buried at the centre of the monomer,
where domains 1 (residues 1–484) and 2 (residues 485–842) come
together. This site has been extensively investigated with glucose
analogue inhibitors that bind at this site and promote the less ac-
tive T state through stabilisation of the closed position of the 280s
loop (residues 282–287).8 On transition from T state to R state
(activation of the enzyme), the 280s loop becomes disordered
and displaced, opening a channel that allows a crucial residue,
Arg569, to enter the catalytic site. This creates the recognition site
for the substrate phosphate that also allows access of the glycogen
substrate to reach the catalytic site.4 Apart from the catalytic site
GP has an allosteric site, which binds the allosteric activator
Abbreviations: GP, glycogen phosphorylase, 1,4-
-glucosyltransferase (EC 2.4.1.1); GPb, rabbit muscle glycogen phosphorylase b;
PLP, pyridoxal 50-phosphate; glucose,
-glucose; Glc-1-P, -glucose 1-phos-
a-D-glucan:orthophosphate
a
a-
D
a-D
phate; rmsd, root-mean-square distance.
⇑
Corresponding authors. Tel.: +30 210 7273878; fax: +30 210 7273831 (I.D.K.);
tel.: +30 2410 565278; fax: +30 2410 565290 (D.D.L.).
E-mail addresses: ikostas@eie.gr (I.D. Kostas), ddleonidas@bio.uth.gr (D.D.
Leonidas).
Deceased on August 31st 2008.
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.09.039

7912
K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
AMP, situated at the subunit–subunit interface some 30 Å from the
catalytic site. There is also an inhibitor site, located on the surface
of the enzyme some 12 Å from the catalytic site, which binds pur-
ine compounds, nucleosides or nucleotides. The inhibitor site ob-
structs the entrance to the catalytic site tunnel stabilizing the T
state conformation of the enzyme. GP has also a storage site, lo-
cated on the surface of the molecule, some 30 Å from the catalytic
site which might serve as a region through which the mammalian
enzyme is attached to glycogen particles in vivo.4 Finally, a new
allosteric site was discovered from crystallographic studies of the
binding of indole-carboxamide and related synthetic counter-
parts.9,10 The new allosteric (or indole) binding site, located inside
the central cavity of the dimeric enzyme, is formed on association
of the two subunits; the central cavity is partially closed at one end
by residues from the cap’ region (residues 47–78 from the symme-
semicarbazone derivatives. All compounds were found to be com-
petitive inhibitors of GPb with respect to glucose-1-phosphate
(Glc-1-P). We also report here on the crystallographic binding of
these 15 derivatives to GPb, in order to provide rationalizations
for their kinetic properties. Crystallographic data show that these
compounds bind at the catalytic site and at the indole-carboxamide
site of GPb (Fig. 1), with the exception of a molecule with an ortho-
nitro substituted phenyl ring which was found bound only at the
catalytic site. Identification of the structural determinants contrib-
uting to inhibitor binding mode at the catalytic site should provide
better understanding of the mechanism of inhibition of GP, and aid
the design of compounds with improved potency against GP.
2. Results and discussion
try related subunit) and
and their symmetry related equivalents) and at the other end by
a2 helices (Arg33, His34, Arg60 and Asp61
2.1. Organic synthesis
the
a7 tower helices (residues Asn270, Glu273, Ser276 and their
The synthesis and characterization of the aromatic aldehyde
symmetry related equivalents).4
4-(b-
1, was achieved as recently reported by us.16 Condensation of
4-(per-O-acetylated-b- -glucopyranosyl)thiosemicarbazide 1 with
a number of aldehydes according to a known procedure,17,18
afforded aldehyde 4-(per-O-acetylated-b- -glucopyranosyl)thio-
D-glucopyranosyl)thiosemicarbazones 3, outlined in Scheme
We have previously studied the binding of N0-benzoyl-N-b-
D-
glucopyranosyl urea (Bzurea) to GPb and shown that Bzurea is a
potent inhibitor of GPb displaying a Ki value of 4.6
information available from the crystal structure of the GPb-com-
plex at 1.8 Å resolution11 revealed that Bzurea, binds tightly at
the catalytic site and induces significant conformational changes
in the vicinity of the site. In particular the 280s loop (residues
D
l
M.11 Structural
D
semicarbazones 2, the acetyl protective groups of which were
removed by the Zemplén method19 in absolute methanol in the
presence of sodium methoxide. The spectral data indicate the b
configuration, and the E configuration pertaining to the stereo-
chemistry of the C@N bond.16 Compounds 3pF, 3mBr and 3pBr
are new members of this class of bioactive molecules.
282–287) shifts 1.3–3.7 Å (C
a atoms) to accommodate Bzurea. In
this way the inhibitor promotes the less active T state enzyme
through stabilization of the closed position of the 280s loop which
blocks access of the substrate to the catalytic site. The urea and
benzoyl moieties fit tightly into the so-called b-pocket, a side chan-
nel from the catalytic site with no access to the bulk solvent.4 Fur-
thermore, Bzurea binds at the new allosteric site of GPb, which is
highly specific for indole derivatives and currently a target for
the development of hypoglycaemic drugs.8
Thiosemicarbazones are a class of biochemically important
compounds possessing a wide range of biological activity, and
are very promising in the treatment of many diseases.12–15 In a re-
cent paper, we reported a systematic study for the synthesis and
2.2. Enzyme kinetics
The inhibitory efficiency of the thiosemicarbazone derivatives
was tested in kinetic experiments with rabbit muscle glycogen
phosphorylase b (GPb) and their inhibition constants are presented
in Table 1. All compounds displayed competitive inhibition with
respect to the substrate Glc-1-P, at constant concentrations of gly-
cogen (0.2% w/v) and AMP (1 mM) with IC50 values at the
range. The most potent compound is 3pF, with a fluorine atom at
the para-position of the aromatic ring with IC50 = 5.7 M while
the least potent is compound 3pCF3 (IC50 = 524.3 M), with a –
lM
characterization of a novel class of b-D-glucopyranosyl-thiosemi-
carbazone derivatives.16 Using as lead Bzurea, we have now
assessed the inhibitory potential of a series of thiosemicarbazone
l
l
CF3 group at the same position. Comparison of the ligands with a
para substitution (column 1, Table 1) indicates that the inhibitory
potency decreases with the increase in bulkiness of the substituent
at this position. This is also evident when comparing the potency of
compounds 3pOMe (methoxy) and 3pMe (methyl), which have
derivatives of b-D-glucopyranose (Table 1) to GPb. In these
compounds, the urea moiety of Bzurea is replaced by a thiosemi-
carbazone to (i) elongate the linker of the group that binds at the
b-pocket of the catalytic site and (ii) to test the contribution of
the carbonyl oxygen of urea interactions to the inhibitory potency.
In addition, instead of the benzoyl group of Bzurea, 14 compounds
posses a phenyl ring with variety of groups attached at either ortho,
meta or para-position with respect to the thiosemicarbazone moi-
ety while one has a pyridine. This served to investigate the ability
of various groups to enhance the inhibitory potency of these thio-
IC50 values of 406.5 lM and 192.4 lM, respectively or when com-
paring the potency of 3pF with that of 3pMe. The only exception to
this observation is the nitro compound 3pNO2, which is among the
best inhibitors of the enzyme. This exception indicates that the
chemical properties of the substituent also affect significantly the
Scheme 1. Reagents and conditions: (i) ArCH@O, EtOH or MeOH, cat. AcOH, reflux; (ii) NaOMe, MeOH, rt. For compounds 2oNO2, 2pNO2, 2oCl, 2mCl, 2pCl, 2oOH, 2mOH,
2pOH, 2pOMe, 2pCF3, 2pMe, 2pyr and 3oNO2, 3pNO2, 3oCl, 3mCl, 3pCl, 3oOH, 3mOH, 3pOH, 3pOMe, 3pCF3, 3pMe, 3pyr see Ref. 16.

K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
7913
Table 1
Inhibitory efficiency of the thiosemicarbazone derivatives of b-D-glucopyranose and the atom numbering scheme used in crystal structures
O6
OH
O5
O
O4
C6
HOC4
HO
C5
HC8
Ar
C1H
N1 S N2
H
N
NC7
O3 C2OH
C3
N
N3
O2
S1
Ar group
C10
C9
IC50
(
l
M)
Ar group
IC50
(
l
M)
Ar group
IC50 (lM)
C11
O
O7
O7
O
C12
O8O NN4
C10
N4
C11
C12
C13
N
3pNO2
25.7 0.9
3oNO2
484.2 23.3
370.0 9.7
26.6 3.3
O
O8
C13
C14
C9
C14
C10
C9
C11
C12
3pF
F
5.7 0.4
C13
C11
C14
C10
C9
Cl
Cl
C10
C10
C9
C12
C11
C12
C13
C11
C12
C9
Cl
3pCl
28.3 1.7
3mCl
3mBr
23.2 0.5
50.4 1.9
3oCl
C14 C13
C11
C14
C14
C13
C10
C9
Br
C10
C9
C12
C11
C12
Br
3pBr
93.2 5.1
C13
C14
C13
C14
C10 C11
C9
O7
O7
HO
C10
C12
OH
C11
C12
C13
C10
C9
OH
11
C12
C
O7
3pOH
340.5 21.7
3mOH
180.0 7.8
3oOH
C14
C13
C9
C14
C14 C13
C10
C9
C11
C12 C15
O CH3
O7
3pOMe
3pMe
3pCF3
3pyr
406.5 40.6
192.4 5.8
524.3 11.4
200.0 17.0
C14 C13
C10
C11
C12
C
CH3
C15
C13
C11
C12
C14
C10
C9
F1
F
F
F
F2
F3
C13
C11
C14
C10
C9
N4
N
C13 C12
potency of the ligand. Comparison of the position (ortho, meta or
para) of the substituent at the phenyl ring reveals that within the
compounds with chlorine atom (3pCl, 3mCl, and 3oCl) the order
of potency is para ꢀ meta > ortho while for those with a hydroxyl
group (3oOH, 3mOH, and 3pOH) is ortho > meta > para. The
potency of compounds 3pBr and 3mBr with a bromine atom at
Compound 3pF (IC50 = 5.7
lM) is almost equipotent with
Bzurea (Ki = 4.6
l
M),11 while the rest of the inhibitors are poor
inhibitors compared to their lead molecule (Bzurea). However,
compounds 3pNO2, 3mCl, 3pCl, 3mBr and 3oOH are still
considered as potent inhibitors with IC50 values less than
50 lM and could be exploited as improved leads for further
para and meta position does not differ much (IC50 values 93.2
and 50.4 M) while 3oNO2 and 3pNO2 with a nitro group at ortho
and para-position, show large differences in their potency with
IC50 values of 484.2 M and 25.7 M, respectively. Therefore a
l
M
optimization.
l
2.3. Structural studies
l
l
general pattern with respect to the position of the substituent
(i.e., para is better than meta) is not evident from the kinetics
results. However, it seems that the inhibitory potency of the
compounds depends mostly on the chemical and physical
properties (size, charge, polarizability, etc) of the decorating
groups of the phenyl ring. Thus, for halogens the best position is
the meta, for the nitro group the para and for the hydroxyl the ortho
position.
In order to elucidate the structural basis of inhibition we have
determined the crystal structure of GPb in complex with each of
the 15 compounds presented in Table 1. A summary of the data
processing and refinement statistics for the inhibitor complex
structures is given in Supplementary data Tables S1–S3. The
2Fo ꢁ Fc and Fo ꢁ Fc Fourier electron density maps calculated using
the coordinates of the native structure indicated that all com-
pounds except 3oNO2 were bound at both the catalytic site and

7914
K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
lase21,22 where the 280s loop is held in an open conformation (in
the R state GPb the 280s loop is disordered and its position ambig-
uous23) in the vicinity of the catalytic site has shown that the shifts
of the 280s loop are not in the direction of the T to R allosteric tran-
sition. In addition, the new position of the 280s loop imposed by
the binding of the present inhibitors still obstructs the entrance
to the catalytic site of GPb and therefore all fifteen inhibitors can
be classified as T state inhibitors. The mode of binding and the
interactions that the glucopyranose moiety of all fifteen com-
pounds makes with GPb residues at the catalytic site are almost
identical with that of
a-D
-glucose.24,25 The benzaldehyde-derived
thiosemicarbazone moiety is accommodated at the b-pocket of
the catalytic site forming mainly van der Waals interactions with
the protein. The sulfur group of thiosemicarbazone is hydrogen-
bonded either directly to Asp283 OD1 (compounds 3pF, 3oCl and
3pOH) or Leu136N (compounds 3pBr, 3pCl, 3pCF3, 3pOMe and
3pMe), or to both (compounds 3pNO2 and 3pyr) or through a
water molecule (compounds 3mBr, 3mCl, 3oNO2, 3oOH, and
3mOH) to the side chain of Asp283. Apart from these interactions
the thiosemicarbazone moiety is not involved in any other hydro-
gen bond interactions with neighbouring residues. The hydrogen
bond between N1 of N-acetyl-b-
and carbonyl O of His377, an interaction that has been observed
in all b-
-glucopyranosylamine analogues4,8 has not been observed
D-glucopyranosylamine (NAG)
D
in this complex similarly to the Bzurea complex.11 At the new allo-
steric site the glucopyranose moiety is engaged in hydrogen bond
interactions with the side chain atoms of Glu190, and the main
chain nitrogen of Ala192, while four water molecules mediate
hydrogen bonds with side chain atoms of His57, Arg60, Asn187,
and Tyr226. In addition, N1 from the thiosemicarbazone moiety
forms a hydrogen bond with the carbonyl oxygen of Glu190 while
N2 and N3 form hydrogen bonds with the side chain of Arg160. The
interactions between 3pF, 3mBr, 3mCl and 3pNO2 and the protein
are shown in Figure 2.
Figure 1. A schematic diagram of the GPb dimeric molecule viewed down the
molecular dyad. One subunit is colored in dark blue and the other in light blue. The
catalytic and the new allosteric site are marked by bound compound 3pF (in green
and yellow, respectively). Compound 3pF, on binding to the catalytic site, promotes
the less active T state through stabilization of the closed position of the 280 s loop
(shown in pink and thicker) which blocks access for the substrate (glycogen) to the
catalytic site.
at the new allosteric site. 3oNO2 was found bound only at the cat-
alytic site. Electron density maps clearly defined the position of
each atom of the inhibitors within the catalytic and the new allo-
steric site. In crystallographic experiments with a soak of 1 mM
of each inhibitor for 3.5 h, electron density maps suggested strong
binding to the catalytic site, but partial occupancy of the new allo-
steric site by the ligand. It seems therefore that the new allosteric
site is not the primary binding site for these inhibitors. A similar
situation has been also observed with Bzurea.11
The structural results show that all inhibitors are bound with
essentially no disturbance of the overall protein structure. The
superimposition of the structures of native GPb and the
GPb—inhibitor complexes over well defined residues (18–249,
The glucopyranosyl-thiosemicarbazone derivatives studied
have a benzaldehyde-derived group attached at the C8 position
(Table 1), which has either a –Br or a –Cl or a –OH, or a –Me, or
a –OMe or a –NO2, or a –F at ortho, para or meta positions or instead
of a benzaldehyde-derived group there is a pyridinecarboxalde-
hyde-derived moiety. Since the conformation (Fig. 3) and the pro-
tein–ligand interactions pattern of the glucopyranosyl ring and the
thiosemicarbazone linker of the derivatives upon binding to the
catalytic site is very similar, and the same applies for the inhibitors
bound at the new allosteric site, we describe below in more detail
the interactions of the phenyl or pyridyl moiety of compounds
with GPb in groups according to the substituent group of the aro-
matic ring of the inhibitors.
262–312, 326–829) gave rmsd values 0.2–0.4 Å (C
a positions).
However, upon binding to the catalytic site compounds 3pBr,
3pCl, 3pCF3, 3oOH, 3pOMe and 3pMe trigger a significant shift
of the 280s loop with concomitant changes in the adjacent imidaz-
ole of His571. Thus, the range of the rmsd values between the
structures of GPb-inhibitor and native GPb for main chain atoms
is for residues Asn282 (0.4–0.7 Å), Asp283, (0.7–1.0 Å), Asn284
(1.1–1.8 Å), Phe285 (0.8–0.9 Å), Phe286 (0.6–0.7 Å), and Glu287
(0.3–0.4 Å). In the native structure, the side chain atom NE2 of
His571 forms a hydrogen bond with the side chain atom OD2 of
Asp283. In the complex structures with compounds 3pBr, 3pCl,
3pCF3, 3oOH, 3pOMe and 3pMe the side chain of His571 is rotated
2.3.1. The nitro group
Inhibitors 3oNO2 and 3pNO2, 2- and 4-nitro-benzaldehyde
4-(b-D-glucopyranosyl) thiosemicarbazone upon binding to the
catalytic site of GPb take part in 11 and 14 hydrogen bond interac-
tions with the protein, while they exploit 114 and 129 van der
Waals interactions with GPb residues, respectively (Fig. 2). Inter-
estingly structural comparison of the complexes with the native
enzyme reveals that the binding of the two inhibitors does not
cause any significant conformational changes including the 280s
loop. The nitro group in compound 3oNO2 participates in a
hydrogen bond interaction with the side chain of Glu88 (Table
2) and water mediated hydrogen bonds with the main chain
nitrogen of Gly134 and Gly137. In compound 3pNO2 the nitro
group is involved in hydrogen bond interactions with the car-
bonyl oxygen of Tyr280 and Asn282 (Table 2) and a water medi-
ated hydrogen bond with the carbonyl oxygen of Glu287 (Fig. 2).
Kinetic studies revealed that inhibitor 3pNO2 (para-position) is
almost 19 times more potent than 3oNO2 (ortho position). This
(v1 ꢀ 20°) to maintain this hydrogen bond with the side chain of
Asp283at the new position. Furthermore, the conformational
change of the 280s loop results in Asn284 being sandwiched
between the side chains of Phe285 and Tyr613. These phenyl rings
are the key components of the inhibitor site of GPb, a site that
binds caffeine and a number of other fused ring compounds.4,5,8
The insertion of Asn284 leads to the destruction of this site. This
mimics the T to R allosteric transition, which also involves
movement of the 280s loop.20 However, a structural comparison
between the present GPb complexes and maltodextrin phosphory-

K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
7915
Figure 2. Schematic diagram of the molecular interactions between GPb and inhibitors 3pF, 3mBr, 3mCl, and 3pNO2 when bound at the catalytic site (a, c, e, g) and at the
new allosteric site of GPb (b, d, f, h). The side chains of protein residues involved in ligand binding are shown as ball-and-stick models. Hydrogen bond interactions between
the inhibitors, protein residues and water molecules (w) are represented as dotted lines.
difference in their potency can be attributed (i) to the three addi-
tional hydrogen bond interactions of 3pNO2 with respect to those
of 3oNO2, and (ii) to the hydrogen bonds of atom O7 of the nitro
group of 3pNO2 with the main chain oxygen atoms of Tyr280 and
Asn282 from the 280s loop (unlike the nitro group of 3oNO2
which does not form any interactions with residues of the 280s

7916
K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
Figure 3. Structural comparisons of the ligand molecules when bound at the catalytic (left) and at the new allosteric site (right) of GPb are shown in color; 3pF green, 3mBr
yellow, 3mCl cyan, 3pNO2 orange, 3oOH pink, and 3pMe red.
Table 2
Interactions of the aryl ring substituent of the inhibitors with GPb residues at the catalytic site
Code
Hydrogen/halogen bonds
No of Van der
Waals
interactions
Code
Hydrogen/
halogen bonds
No of Van der
Waals
interactions
Code
Hydrogen/
halogen
bonds
No of Van der
Waals
interactions
3pNO2 Tyr280 O; Asn282 O
13
6
11
3oNO2 Glu88 OE1
15
6
3pF
3pCl
Asn282 O
—
3mCl
3mBr
Phe285 O;
Ala383 O
Phe285 O
4
3oCl
Phe285 N
Asp283 O
3pBr
3pOH
Asn282 OD1
Asn282 O
10
0
7
2
3mOH Asn282 O;
3oOH
6
Ala383 O
3pOMe
3pMe
—
—
15
8
3pCF3
Tyr280 O; Asn282 O, OD1; Arg292
NH1, NH2
27
3pyr
Asp339 OD2; Ala383 O
3
loop) that lead to further stabilization of its closed confor-
mation.
catalytic site of GPb, compounds 3mCl, 3mBr and 3oCl, form 8,
8, and 9 hydrogen bonds, and 97, 86 and 100 van der Waals inter-
actions with the protein residues, respectively (Fig. 2). The chlo-
ride group of 3mCl forms two halogen bonds with the carbonyl
oxygen atoms of Phe285 and Ala383 O and the chloro-phenyl ring
engages in 4 van der Waals interactions. The bromide group of
compound 3mBr engages in a halogen bond interaction29 with
the carbonyl oxygen of Phe285 while the bromo-phenyl ring
takes part in 7 van der Waals interactions with protein residues
(Table 2). In the 3oCl complex the chloride group forms a hydro-
gen bond with the main chain nitrogen of Phe285 and the chloro-
phenyl ring participates in 6 van der Waals interactions with GPb
residues (Table 2).
Upon binding to the new allosteric site of GPb, inhibitor 3pNO2
forms three direct and six water mediated hydrogen bonds and 79
van der Waals interactions with GPb residues (Fig. 2). Structural
comparison with Bzurea bound at this site revealed that the gluco-
pyranose moiety shifts to opposite directions by approximately 93°
with respect to the backbone. Inhibitor 3oNO2 was not found
bound at the new allosteric site.
2.3.2. The halogen group
The conformation of the halogen inhibitors when bound at the
catalytic site of GPb are very similar (Fig. 3). 3pF upon binding at
the catalytic site of GPb forms 10 hydrogen bonds and 101 van
der Waals interactions (Fig. 2), while inhibitor 3pCF3 forms 11
hydrogen bonds and 95 van der Waals interactions with protein
residues. Compound 3pCl forms 9 hydrogen bonds, and 103 van
der Waals interactions with the protein residues at the catalytic
site, respectively while 3pBr participates in 9 hydrogen-bonding
interactions and 100 van der Waals interactions with residues of
GPb at the same site.
Fluorine is a poorer hydrogen bond acceptor than oxygen (it
cannot act as a donor)26 and it affects significantly the hydro-
gen-bonding interactions between a ligand and residues of the
catalytic site of GPb27 that cannot donate a proton. In the 3pF
complex, fluorine forms a halogen bond28 with Asn282 O whereas
the fluorine atoms of the trifluoromethyl group of 3pCF3 engage
in hydrogen bond interactions with the side chain atoms of
Arg292, the main chain oxygen and the side chain OD1 of
Asn282, and a water mediated interaction with the side chain
of Glu88. In 3pCl the chloride group is involved only in 11 van
der Waals interactions while in 3pBr bromine is involved in
hydrogen bond interactions with Asn282 OD1 in addition to 10
van der Waals interactions (Table 2). Upon binding to the
The most interesting observation of the binding of the halogen
compounds is that the binding of 3pCF3, 3pCl, and 3pBr causes a
significant conformational change of the GPb residues of the 280s
loop and on average the Ca atoms of residues 282–287 are shifted
by 1.0 Å from their positions at the native structure. This confor-
mational change is less profound in the 3pF, 3mCl, and 3mBr com-
plexes where the average shift of all atoms of residues 282–287 is
0.3 Å while in the 3oCl complex this was not observed. Interest-
ingly, the trifluoromethyl derivative triggers significant conforma-
tional changes to the 280s loop whereas the effect of the binding of
the fluoro-compound is less profound. Therefore, it seems that the
para halogen substituents trigger a most significant conformational
change than those at the meta or ortho position. The exception of
3pF might be attributed to the smaller size of the fluoride group
as compared to those of the trifluoromethyl, chloride or bromide
group.
Kinetic studies revealed that compound 3pF is the most potent
of all 15 compounds presented here, almost 92 times more potent
than compound 3pCF3 and equipotent to Bzurea.11 Structural com-
parison of the GPb–3pF complex with the GPb–Bzurea complex at
the catalytic site (Fig. 4a) showed that the two inhibitors bind in a

K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
7917
similar manner and this provides a structural basis for their equi-
potency. Therefore it seems that replacement of the urea moiety
with a thiosemicarbazone does not affect the potency of the inhib-
itor since Bzurea and 3pF are equipotent.
Kinetic experiments indicated that inhibitors 3pBr and 3mBr
are almost equipotent (Table 1) and this is consistent with struc-
tural studies that show that both inhibitors bind very similarly to
the catalytic site of GPb engaging in approximately the same num-
ber of interactions with protein residues (Table 2). Furthermore, on
binding both compounds 3pBr and 3mBr displace five water mol-
ecules from the active site of the unliganded structure.
Upon binding to the new allosteric site of GPb, inhibitors 3pF
and 3pCF3 form 4 hydrogen bonds each and 111 and 77 van der
Waals interactions, respectively (Fig. 2). The fluoro and fluorom-
ethyl groups do not engage in any hydrogen bond interactions with
protein residues. Structural comparison with the Bzurea binding
Kinetic studies revealed that 3pCl and 3mCl, (para and meta po-
sition) have a significant better inhibitory effect on GPb, than the
3oCl (ortho) derivative with IC50 values 16 and 13 times lower,
respectively. The binding position of the chloro-phenyl ring is dif-
ferent in the three compounds. Its position in compound 3mCl is
1.8 Å apart from its position in compounds 3oCl and 3pCl, and
the angle between the planes of the benzyl ring in the two latter
compounds is ꢂ20°. The chloride group is engaged in halogen bond
interactions with the carbonyl oxygen of Phe285 and Ala383 in
compound 3mCl and with the main chain nitrogen of Phe285 in
compound 3oCl, while in compound 3pCl is involved only in van
der Waals interactions (Table 2). However, the most interesting
observation of the binding of the chloro-compounds is that com-
pounds 3mCl and 3pCl induce the 280s loop conformational
change described above whereas compound 3oCl does not. This
conformational change and the stabilization of the 280s loop seem
to form the structural basis for the much higher potency of com-
pounds 3pCl and 3mCl with respect to 3oCl.
revealed that the b-D-glucopyranose moiety in the two ligands is
shifted to an opposite direction ( Fig. 4b). Upon binding to the
new allosteric site of GPb, compounds 3oCl, 3mCl, and 3pCl form
three hydrogen bonds each and exploit 64, 79, 85 van der Waals
interactions, respectively ( Fig. 2). A structural comparison with
Bzurea which also binds at the same site11 revealed that the
glucopyranose moiety once again shifts to opposite directions by
approximately 93° with respect to the backbone. The binding of
3oCl, 3mCl, and 3pCl at the new allosteric site does not promote
any significant conformational changes apart from the shift of
Figure 4. (a) Stereo diagrams of the structural comparison between GPb in complex with 3pF (cyan) and N0-benzoyl-N-b-
site of GPb and (b) in the vicinity of the new allosteric site (N0-benzoyl-N-b-
-glucopyranosyl urea-GPb complex, magenta) and GPb-3pF complex (subunit 1, yellow and
subunit 2, cyan) and GPb-3pCF3 complex (subunit 1, green and subunit 2 magenta).
D
-glucopyranosyl urea11 (magenta) at the catalytic
D

7918
K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
the side chain of Arg60 which has been observed for all compounds
that bind to this site. Inhibitors 3mBr and 3pBr were also found
bound at the new allosteric site. Compound 3mBr is bound at
the new allosteric site with two alternative conformations in
which the orientation of the bromo-benzyl moiety differs by
180°. Compound 3pBr displays only one conformation at this site
and both 3mBr and 3pBr form 7 hydrogen bonds and 34 van der
Waals interactions with protein residues (Fig. 2). The most signifi-
cant conformational change at this site upon binding of these
inhibitors is the shift of the side chain of Arg60 by ꢂ2.0 Å from
its position at the native GPb structure, which is probably due to
stereochemical impediments imposed by the inhibitor molecule
and has also been observed upon binding of Bzurea to the same
site. A structural comparison with the binding of Bzurea at this
site11 revealed only minor conformational changes.
bazone moiety. On binding to the new allosteric site of GPb, the
inhibitors make a total of four hydrogen bonds and exploit 66
van der Waals interactions. The hydroxyphenyl ring is only in-
volved in van der Waals interactions mainly with Arg60, Val64
and Val400 from the symmetry related protein subunit.
2.3.4. The methoxy/methyl group
The two inhibitors, 4-methoxy-benzaldehyde 4-(b-D-glucopyr-
anosyl)thiosemicarbazone (3pOMe) and the 4-methyl analogue
(3pMe) bind at the catalytic site of GPb in a very similar manner
and conformation. Upon binding to the catalytic site of GPb, each
one form nine hydrogen bonds with protein residues and they ex-
ploit 107 and 101 van der Waals interactions with GPb residues,
respectively (Fig. 5). Upon binding, both inhibitors induce signifi-
cant conformational changes to residues of the 280s loop and the
largest shifts occur in residues Asp283 (1.0 Å at C
a positions) and
2.3.3. The hydroxyl group
Asn284 (1.7 Å at C positions). The methyl group of 3pMe engages
a
The 2-, 3-, and 4-hydroxy-benzaldehyde 4-(b-
D
-glucopyrano-
in 8 van der Waals interactions with GPb residues whereas the
methoxy group of 3pOMe in 15 (Table 2). The methyl derivative
is almost two times more potent than the methoxy one and this
cannot be explained from the structural point of view.
syl)thiosemicarbazones (3oOH, 3mOH and 3pOH) were found
bound at the catalytic and the new allosteric site. The conforma-
tion of the three compounds upon binding at the catalytic site dif-
fers. Thus, in compound 3oOH the torsion angle C8–N3–N2–C7 is
180° so that the conformation about the N2–N3 bond is in trans
geometry while in compounds 3mOH and 3pOH it is ꢁ180° (for
atom numbering, see in Table 1). In addition, the hydroxyphenyl
ring is coplanar with the thiosemicarbazone linker in compounds
3oOH and 3pOH whereas in 3mOH is inclined by ꢂ20°.
Upon binding to the new allosteric site of GPb, inhibitors
3pOMe and 3pMe form three hydrogen bonds each and 79 and
81 van der Waals interactions, respectively. No other profound
conformational changes were noted in the protein environment.
Comparison with Bzurea shows that the thiosemicarbazone moiety
is 180° rotated with respect to the urea moiety, whereas the posi-
tions of the aromatic rings of the three inhibitors superimpose
rather well. As a result the glucopyranose moiety is pointing in
opposite directions in the 3pOMe, 3pMe and Bzurea complexes.
In addition, in inhibitors 3pOMe and 3pMe the aryl ring is coplanar
to the thiosemicarbazone moiety whereas in Bzurea is inclined by
30°.
Upon binding to the catalytic site of GPb, inhibitors 3oOH,
3mOH, and 3pOH form 15, 15 and 13 hydrogen bonds, and exploit
109, 125, and 97 van der Waals interactions, respectively (Fig. 5).
The binding mode of the hydroxyphenyl group differs between
the three inhibitors. The hydroxyphenyl group is bound deep in
the b-pocket involved in van der Waals interactions with protein
residues dominated by polar-non polar contacts with Glu88,
Asn282, Asp283 Phe285, Arg292 and His341. The binding of
3oOH and 3pOH but not 3mOH triggers significant shifts of the
main chain atoms from their position in the native GPb structure.
The major rmsd values of the main chain atoms are for residues
Asn282 (0.7 Å), Asp283 (0.7 Å), Asn284 (1.1 Å), Phe285 (0.9 Å),
and Phe286 (0.6 Å). In the 3pOH complex the hydroxyl group is
hydrogen-bonded with the carbonyl oxygen of Asn282 and it is
not involved in any van der Waals interactions (Table 2). Com-
pound 3mOH displays two alternative conformations for the
hydroxyphenyl group 180° apart and the hydroxyl group takes part
in hydrogen bond interactions with either the carbonyl oxygen of
Ala383 and water mediated interactions with side chain atoms of
Glu88 and Arg292 or the carbonyl oxygen of Asn282 ( Fig. 5,
Table 2). The hydroxyl group in compound 3oOH forms a hydrogen
bond with the carbonyl oxygen of Asp283 (Table 2).
2.3.5. The pyridine group
The binding of 4-pyridinecarboxaldehyde 4-(b-D-glucopyrano-
syl)thiosemicarbazone (3pyr) to the catalytic and to the new allo-
steric site of GPb is not accompanied by any significant
conformational changes of the enzyme including the 280s loop.
Inhibitor 3pyr forms 14 and 4 hydrogen bonds, and 113 and 80
van der Waals interactions at the catalytic and the new allosteric
site of GPb, respectively. Structural comparison of the 3pyr com-
plex with the GPb–Bzurea complex11 but also with the other 14
GPb–inhibitor complexes presented here, reveals that the aromatic
ring of 3pyr is positioned at a different location compared to the
rest of the inhibitors studied. Thus, the N2–N3 bond is rotated by
180° and as a result the pyridine is bound in a new position within
the b-pocket (shifts of 3.3 Å in atom C9 and 7.3 Å in atoms N4–C12
between 3pF and 3pyr). In this new location the pyridine is en-
gaged in hydrogen bond interactions with OD2 of Asp339 and
the carbonyl oxygen of Ala383 and water mediated hydrogen
bonds with Thr340 OG1 and Glu385N (Fig. 5). Furthermore it is in-
volved in van der Waals interactions with atoms from residues
Asp339, His341, Thr378, and Ala383. In the new allosteric site,
3pyr binds in a similar structural mode with the rest of the 14
inhibitors presented here.
Kinetic studies revealed that 3oOH (ortho position) has the best
inhibitory effect on GPb with an IC50 value of 26.6
itors 3mOH and 3pOH (meta and para-position) are less potent
with IC50 values of 180 M and 340.5 M, respectively. This differ-
lM, while inhib-
l
l
ence in their inhibitory potency might be attributed to the number
of the van der Waals interactions of the hydroxyl group with GPb
residues in each inhibitor complex (Table 2). Thus, 3oOH which
displays 6 interactions is more potent than 3mOH (2 interactions)
which in turn is more potent than 3pOH (no interactions).
2.4. Conclusions
On binding to the new allosteric site all three inhibitors have
similar conformation which differs significantly from that ob-
served in the catalytic site. Thus, the torsion angle C9–C8–N1–C1
is ꢁ177° and 23° in the new allosteric site. As a result the hydroxy-
phenyl ring and the glucopyranose ring of the bound inhibitors are
not in line as in the catalytic site but inclined by approximately
30°. However, in both the catalytic and the new allosteric site
the hydroxyphenyl ring is coplanar to the plane of the thiosemicar-
The synthesis and characterization of 15 b-D-glucopyranosyl-
thiosemicarbazones (three of which are new compounds), is
reported. These have been studied with kinetic and X-ray crystal-
lography studies for binding to GPb. These inhibitors were found
to be competitive inhibitors of the enzyme with respect to
Glc-1-P and they all bind to the active site by anchoring their
b-D-glucopyranose moiety at the a-D-glucose binding site and the

K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
7919
Figure 5. Stereo diagrams of the molecular interactions between GPb and inhibitors 3pOH (a), 3mOH (b), 3pOMe (c), and 3pyr (d) when bound at the catalytic site. The side
chains of protein residues involved in ligand binding are shown as ball-and-stick models. Hydrogen bond interactions between the inhibitors, protein residues and water
molecules (w) are represented as dotted lines.
thiosemicarbazone moiety at the hydrophobic b-pocket of the
active site. In addition, these inhibitors, with the exception of
3oNO2, were also found bound at the new allosteric site of GPb.
Binding at this site does not promote any significant conforma-
tional changes except for small shifts in the residues in the vicinity
of the inhibitor (namely Arg60, Val64, Trp189 and Lys191), which

7920
K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
undergo conformational changes in order to accommodate the
ligand. However, it should be noted that the orientation of the
aromatic group is maintained as observed previously in other
compounds bound at this site.8
C, H, N were carried out on a Perkin–Elmer PE 2400 II instru-
ment.
3.1.2. Aldehyde 4-(per-O-acetylated-b-D-glucopyranosyl)-
Comparing the binding of each compound at the catalytic site, it
is interesting to see that there is a pattern emerging based on the
electronegativity of the directing groups. Between the three com-
pounds with the –Cl substituent, the one in the ortho position of
the aromatic ring causes a moderate shift of the 280s loop upon
binding. The one in the meta position causes a significant shift in
the 280s loop, while the one in the para-position causes a major
shift in the 280s loop. However, between the three compounds
with the –OH substituent the opposite effect is observed. The
substituent in the ortho position of the aromatic ring causes the
biggest shift of the 280s loop and the strongest destabilization of
the protein. In the case of the –Br substituent the effect of a
significant shift of the 280s loop is the same in both the meta
and para-positions of the aromatic ring. In addition, there is a sim-
ilarity between the –O–CH3 and –CH3 substituents. Both cause a
significant shift of the 280s loop as mentioned above. Interestingly,
the binding of the inhibitors with the –NO2 triggers a negligible
shift of the 280s loop. Finally, is it interesting to add that among
all these compounds, only 3oOH, with the –OH substituent in the
ortho position, changes the direction of the phenyl ring of Phe285
completely, compared to all the other cases where it is simply
shifted. Out of the three inhibitors with the –OH substituent
3oOH also causes the biggest shift and destabilization of the
280s loop. It also has the lowest Ki value between the three. These
observations will be of value in the design of further potent and
specific inhibitors of the enzyme.
thiosemicarbazones 2. A general synthetic procedure
The experimental part was in accordance to that previously de-
scribed by us.16 A solution of the substituted benzaldehyde
(2.38 mmol) in ethanol (5 mL) or 4-pyridinecarboxaldehyde in
methanol was added dropwise to a solution of 4-(2,3,4,6–tetra-O-
acetyl-b-D-glucopyranosyl)thiosemicarbazide (1) (1 g, 2.38 mmol)
in ethanol (80 mL) (methanol for the pyridine analogue). After
the addition of a catalytic amount of AcOH, the reaction mixture
was refluxed for 4–48 h. The solution was kept in fridge overnight
to form a precipitate; in some cases the solution was concentrated
to the half volume at temperature below 40 °C, and then water was
added to precipitate a solid. The precipitate was filtered off,
washed with cold ethanol and in some cases with ether, and
recrystallized from methanol or ethanol for 2pyr. Analytical data
for 2oNO2, 2pNO2, 2oCl, 2mCl, 2pCl, 2oOH, 2mOH, 2pOH, 2pOMe,
2pCF3, 2pMe and 2pyr have been reported.16
3.1.2.1. 4-Fluoro-benzaldehyde 4-(2,3,4,6-tetra-O-acetyl-b-D-
glucopyranosyl)thiosemicarbazone (2a). Yield 40%; mp 209–
212 °C (MeOH); Rf (Hex/EtOAc 1:1) 0.5; ½a D25
ꢃ
ꢁ72.9 (c 1, CHCl3); IR
(KBr):
m 3306 (N–H), 2975 (C–Harom), 1745 (C@O), 1553 (C@N),
1039 (C@S), 916 (1-C–H), 840 (C@S) cmꢁ1
;
1H NMR (300.13 MHz,
3
DMSO-d6): d 11.93 (s, 1H, N(2)H), 8.80 (d, J1-H,N(4)H 9.0 Hz, 1H,
N(4)H), 8.07 (s, 1H, CH@N), 7.90 (dd, 3J = 5.7 Hz, 3J = 8.7 Hz, 2H,
3
H
arom), 7.27 (t, 3J = 8.7 Hz, 2H, Harom), 5.96 (dd, J1-H,2-H = 9.3 Hz,
1H, 1-H), 5.43–5.27 (m, 2H, 2-H, 4-H), 4.96 (t, 3J = 9.6 Hz, 1H, 3-H),
3
2
0
In conclusion, this study provides evidence that thiosemicarba-
4.22 (dd, J5-H,6-H = 4.8 Hz, J6-H,6 -H = 12.3 Hz, 1H, 6-H), 4.07 (ddd,
3
3
J4-H,5-H = 9.9 Hz, 1H, 5-H), 3.98 (dd, J5-H,6 -H = 1.8 Hz; 1H, 60-H),
0
zone derivatives of b-
cogen phosphorylase with IC50 values in the low
best inhibitor 3pF showed an IC50 of 5.7 M) and could bind tightly
D-glucopyranose are potent inhibitors of gly-
lM range (the
2.00, 1.99, 1.96 and 1.93 (4 ꢄ s, 12H, CH3); 13C NMR (75.47 MHz,
CDCl3): d 178.9 (C@S), 171.1, 170.7, 169.9 and 169.6 (C@O), 165.9
and 162.6 (C–F), 142.5 (CH@N), 129.7, 129.6, 129.1, 129.1, 116.3
and 116.0 (Carom), 82.3 (1-C), 73.5 (3-C), 72.6 (2-C), 70.5 (5-C), 68.3
(4-C), 61.6 (6-C), 20.8, 20.7 and 20.6 (CH3). Anal. Calcd for
l
at both the catalytic and new allosteric site of GPb as indicated by
the high resolution structures of GPb determined in complex with
15 aromatic substituted thiosemicarbazone derivatives of b-D-
glucopyranose.
C22H26FN3O9S (527.52): C, 50.09; H, 4.97; N, 7.97. Found: C, 50.07;
H, 5.29; N, 7.73.
3. Experimental section
3.1. Organic synthesis
3.1.1. General
3.1.2.2. 3-Bromo-benzaldehyde 4-(2,3,4,6-tetra-O-acetyl-b-D-
glucopyranosyl)thiosemicarbazone (2b). Yield 44%; mp 183–
185 °C (MeOH); Rf (Hex/EtOAc 1:1) 0.5; ½a D25
ꢁ72.5 (c 1, CHCl3);
ꢃ
IR (KBr):
m 3326 (N–H), 2942 (C–Harom), 1737 (C@O), 1529 (C@N),
1035 (C@S), 900 (1-C–H), 843 (C@S) cmꢁ1
;
1H NMR (300.13 MHz,
3
4-(2,3,4,6-Tetra-O-acetyl-b-
zide (1)17 was synthesized by treatment of (2,3,4,6-tetra-O-acetyl-
b-
-glucopyranosyl)isothiocyanate30 (prepared by the reaction of
potassium thiocyanate with (per-O-acetylated- -glucopyrano-
syl)bromide;31 the latter compound was prepared from
-glucose
pentaacetate) with dry hydrazine in ethanol. -Glucose pentaace-
D-glucopyranosyl)thiosemicarba-
CDCl3): d 10.19 (s, 1H, N(2)H), 8.34 (d, J1-H,N(4)H = 8.7 Hz 1H,
N(4)H), 7.92 (s, 1H, CH@N), 7.83 (m, 1H,
Harom), 7.60 (d,
3J = 7.8 Hz, 1H, Harom), 7.53 (d, 3J = 7.8 Hz, 1H, Harom), 7.29 (d,
D
3
3J = 7.8 Hz, 1H, Harom), 5.67 (dd, J1-H,2-H = 9.3 Hz, 1H, 1-H), 5.41
a
-D
3
D
(t, J3-H,4-H = 9.6 Hz, 1H, 3-H), 5.21–5.09 (m, 2H, 2-H, 4-H), 4.38
3
2
0
D
(dd, J5-H,6-H = 4.5 Hz, J6-H,6 -H = 12.6 Hz, 1H, 6-H), 4.15 (dd,
3
3
J5-H,6 -H = 1.5 Hz, 1H, 60-H), 3.91 (ddd, J4-H,5-H = 10.2 Hz, 1H, 5-H),
0
tate and aldehydes were commercially available. All reactions were
carried out under an argon atmosphere with purified and dry re-
agents and solvents. Column chromatography was performed
using silica gel and TLC on Kieselgel 60 F254 plates; the plates were
visualized under UV light or by gentle heating after spraying with
an ethanolic solution (5% H2SO4). Optical rotations were measured
on a Perkin–Elmer polarimeter at room temperature. Melting
points were measured on a Nikon 50i Pol microscope equipped
with a Linkam THMS600 hot stage and a TMS94 temperature
controller or on a Büchi melting point apparatus, and were not
corrected. IR spectra were recorded by a Bruker Tensor 27 instru-
ment. NMR spectra were taken by a Varian 300 (300.13 MHz and
75.47 MHz for 1H and 13C, respectively) or 600 (599.83 MHz and
150.84 MHz for 1H and 13C, respectively). Atom numbering used
for the NMR spectra is shown in Figure 6. Elemental analyses for
2.08 and 2.04 and (2 ꢄ s, 12H, CH3); 13C NMR (75.47 MHz, CDCl3):
d = 178.9 (C@S), 171.1, 170.7, 169.9 and 169.5 (C@O), 142.0
(CH@N), 135.0, 133.6, 130.4, 130.3, 126.4 and 123.1 (Carom), 82.3
(1-C), 73.5 (3-C), 72.5 (2-C), 70.5 (5-C), 68.4 (4-C), 61.6 (6-C),
20.8, 20.8, 20.6 and 20.6 (CH3). Anal. Calcd for C22H26BrN3O9Sꢅ3H2O
(642.47): C, 41.13; H, 5.02; N, 6.54. Found: C, 41.78; H, 4.96; N,
6.21.
Figure 6. Atom numbering used for the NMR spectra of b-D-glucopyranosyl-
thiosemicarbazones.

K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
7921
3.1.2.3. 4-Bromo-benzaldehyde 4-(2,3,4,6-tetra-O-acetyl-b-
D
-
3.1.3.3. 4-Bromo-benzaldehyde 4-(b-
D
-glucopyranosyl)thiosem-
glucopyranosyl)thiosemicarbazone (2c)32. Yield 64%; mp 210–
icarbazone (3pBr). Yield 78%; mp 202–206 °C (decomp., MeOH/
212 °C (MeOH); Rf (Hex/EtOAc 1:1) 0.5; ½a D25
ꢃ
ꢁ79.8 (c 1, CHCl3);
H2O 1:1); Rf (CHCl3/MeOH 5:1) 0.46; ½a D25
ꢃ
+86.3 (c 1, MeOH); IR
IR (KBr):
m
3134 (N–H), 2983 (C–Harom), 1741 (C@O), 1545 (C@N),
(KBr): m 3540 (OH), 2897 (C–Harom), 1553 (C@N), 1035 (C@S), 900
1035 (C@S), 920 (1-C–H), 822 (C@S) cmꢁ1
;
1H NMR (300.13 MHz,
(1-C–H), 822 (C@S) cmꢁ1
;
1H NMR (300.13 MHz, DMSO-d6): d
3
3
CDCl3): d = 9.32 (s, 1H, N(2)H), 8.30 (d, J1-H,N(4)H = 8.7 Hz 1H,
N(4)H), 7.69 (s, 1H, CH@N), 7.55–7.62 (m, 4H, Harom), 5.69 (dd,
11.80 (s, 1H, N(2)H), 8.62 (d, J1-H,N(4)H = 8.7 Hz, 1H, N(4)H), 8.06
(s, 1H, CH@N), 7.81 (d, 2H, Harom), 7.61 (d, 2H, Harom), 5.36 (dd,
3J1-H,2-H = 9.0 Hz, 1H, 1-H), 5.02, 4.90, 4.49 (3 ꢄ s, 4H, OH), 3.64
3
3J1-H,2-H = 9.3 Hz, 1H, 1-H), 5.41 (t, J3-H,4-H = 9.6 Hz, 1H, 3-H),
3
2
3
2
0
0
5.19–5.10 (m, 2H, 2-H, 4-H), 4.39 (dd, J5-H,6-H = 4.5 Hz, J6-H,6 -
(dd, J5-H,6-H = 5.4 Hz, J6-H,6 -H = 11.4 Hz, 1H, 6-H), 3.52–3.14 (m,
5H, 2-H, 3-H, 4-H, 5-H, 60-H); 13C NMR (75.47 MHz, DMSO-d6): d
178.7 (C@S), 141.8 (CH@N), 133.3, 131.7, 129.5 and 123.3 (Carom),
84.1 (1-C), 78.7 (3-C), 77.6 (2-C), 72.0 (5-C), 69.9 (4-C), 60.9
(6-C). Anal. Calcd for C14H18BrN3O5S (420.28): C, 40.01; H, 4.32;
N, 10.00. Found: C, 39.78; H, 4.85; N, 10.03.
3
H = 12.6 Hz, 1H, 6-H), 4.11 (dd, J5-H,6 -H = 2.1 Hz, 1H, 60-H), 3.90
0
3
(ddd, J4-H,5-H = 10.2 Hz, 1H, 5-H), 2.09, 2.05, 2.04 and 2.03 and
(4 ꢄ s, 12H, CH3); 13C NMR (75.47 MHz,CDCl3): d = 179.2 (C@S),
171.3, 170.7, 169.9 and 169.6 (C@O), 140.5 (CH@N), 148.8, 134.8,
132.8, 130.0, 125.0 and 122.5 (Carom), 82.4 (1-C), 73.6 (3-C), 72.5
(2-C), 70.7 (5-C), 68.4 (4-C), 61.6 (6-C), 20.8, 20.7 and 20.6 (CH3).
Anal. Calcd for C22H26BrN3O9Sꢅ4H2O (660.49): C, 40.01; H, 5.19;
N, 6.36. Found: C, 40.56; H, 5.68; N, 5.99.
3.2. Enzyme preparation and kinetic experiments
Glycogen phosphorylase (GPb) was isolated from rabbit skeletal
muscle and purified as described previously.33 Kinetic studies were
performed in the direction of glycogen synthesis in the presence of
various concentrations of inhibitors. All kinetic experiments were
carried out in the presence of 30 mM imidazole/HCl buffer
60 mM KCl, 0.6 mM EDTA, and 0.6 mM dithiothreitol (pH 6.8),
0.2% (w/v) glycogen, 1 mM AMP and various concentrations of
Glc-1-P. Experiments were performed in the presence of 1% v/v
DMSO and Km of the enzyme for Glc-1-P was 2.2–3.0 mM. Enzyme
activity was measured at pH 6.8 by the release of inorganic phos-
phate as described previously.34
3.1.3. Synthesis of aldehyde 4-(b-D-glucopyranosyl)thiosemi-
carbazones 3 by deacetylation in compounds 2. A general
procedure
The experimental part was in accordance to that previously de-
scribed by us.16 Sodium methoxide as a powder (7.5 mmol) was
added to a solution of 2 (1.5 mmol) in dry methanol (20 mL). The
reaction mixture was stirred at room temperature for 3 h and kept
in fridge overnight. Then, the solution was neutralized with
Amberlist-15 (or acetic acid for the compounds 3pCl, 3pBr and
3pyr), it was filtered to remove sodium ions, and the solvent was
evaporated by vacuum at a temperature below 40 °C. Purification
was carried out by recrystallization. Analytical data for 3oNO2,
3pNO2, 3oCl, 3mCl, 3pCl, 3oOH, 3mOH, 3pOH, 3pOMe, 3pCF3,
3pMe and 3pyr have been reported.16
3.3. X-ray crystallography
Crystallographic binding studies were performed by diffusion of
a solution of b-D-glucopyranosyl thiosemicarbazones in the crys-
tallization media with 20% v/v of DMSO into preformed GPb crys-
tals, grown in the tetragonal lattice, space group P43212, as
previously described.35 Experimental conditions for the formation
of the GPb-inhibitor complexes were as follows: 20 mM of 3pF
(13 h), or 3pBr (16 h), or 3oCl (11.5 h), or 3mCl (16 h), or 3pCl
(21 h), or 3oOH (21 h), or 3mOH (3.5 h), or 3pOH (21 h), or 3pMe
(3.5 h), or 3pyr (7 h), 10 mM 3pCF3 (7 h), 5 mM 3oNO2 (18 h),
6.7 mM 3pNO2 (3 h), or 3pOMe (3 h), and 10 mM 3mBr (7 h). Dif-
fraction data were collected at EMBL-Hamburg outstation (Beam-
line X13), Daresbury Synchrotron Laboratory, UK (Beamline PX
10.1) and Max-Lab, Lund (Beamline ID9111-2). Crystal orientation,
integration of reflections, inter-frame scaling, partial reflection
summation, data reduction and post-refinement were all per-
formed using the HKL program suite.36
3.1.3.1. 4-Fluoro-benzaldehyde 4-(b-D-glucopyranosyl)thiosem-
icarbazone (3pF). Yield 74%; mp 171–175 °C (decomp., MeOH/
H2O 1:1); Rf (CHCl3/MeOH 5:1) 0.45; ½a D25
ꢃ
+52.9 (c 1.1, MeOH); IR
m 3596 (OH), 2877 (C–Harom), 1545 (C@N), 1022 (C@S), 890
(KBr):
(1-C–H), 830 (C@S) cmꢁ1
;
1H NMR (300.13 MHz, DMSO-d6):
3
d = 11.75 (s, 1H, N(2)H), 8.60 (d, J1-H,N(4)H = 9.0 Hz, 1H, N(4)H),
8.10 (s, 1H, CH@N), 7.96–7.91 (m, 2H, Harom), 7.29–7.23 (m, 2H,
3
H
arom), 5.38 (dd, J1-H,2-H = 9.0 Hz, 1H, 1-H), 5.03–4.94 (m, 3H,
3
2
0
OH), 4.50 (s, 1H, OH), 3.65 (dd, J5-H,6-H = 5.4 Hz, J6-H,6 -H
=
11.4 Hz, 1H, 6-H), 3.53–3.14 (m, 5H, 2-H, 3-H, 4-H, 5-H, 60-H);
13C NMR (75.47 MHz, DMSO-d6): d = 178.6 (C@S), 164.8 and
161.5 (C–F), 141.8 (CH@N), 130.6, 130.6, 129.9, 129.8, 115.9 and
115.6 (Carom), 84.0 (1-C), 78.7 (3-C), 77.6 (2-C), 71.9 (5-C), 69.9
(4-C), 60.8 (6-C). Anal. Calcd for C14H18FN3O5Sꢅ5H2O (449.45): C,
37.41; H, 6.28; N, 9.35. Found: C, 37.87; H, 5.98; N, 9.56.
Crystallographic refinement of the complexes was performed
by maximum-likelihood methods using REFMAC.37 The starting
model employed for the refinement of the complexes was the
structure of the native T state GPb complex determined at 1.9 Å
resolution (Leonidas et al., unpublished results). Inhibitor mole-
cules were modeled using the Dundee PRODRG server (http://dav
apc1.bioch.dundee.ac.uk/programs/prodrg/), they were included
in the refinement procedure during its final stages and they were
fitted to the electron density maps after adjustment of their torsion
angles. Alternate cycles of manual rebuilding with ‘COOT’ and
refinement with REFMAC improved the quality of the models.
The stereochemistry of the protein residues was validated by PRO-
CHECK.38 Hydrogen bonds and van der Waals interactions were
calculated with the program CONTACT as implemented in CCP439
applying a distance cut off 3.3 Å and 4.0 Å, respectively. Protein
3.1.3.2. 3-Bromo-benzaldehyde 4-(b-D-glucopyranosyl)thiosem-
icarbazone (3mBr). Yield 70%; mp 190–192 °C (decomp., MeOH/
H2O 1:1); Rf (CHCl3/MeOH 5:1) 0.51; ½a D25
ꢃ
+49.5 (c 1, MeOH); IR
m 3547 (OH), 2869 (C–Harom), 1533 (C@N), 1031 (C@S),
(KBr):
916 (1-C–H), 839 (C@S) cmꢁ1
;
1H NMR (300.13 MHz, DMSO-d6):
3
d 11.80 (s, 1H, N(2)H), 8.72 (d, J1-H,N(4)H = 9.0 Hz, 1H, N(4)H),
8.18 (s, 1H, Harom), 8.06 (s, 1H, CH@N), 7.76 (d, J = 7.5 Hz, 1H,
H
arom), 7.58 (d, J = 7.8 Hz, 1H, Harom), 7.38 (dd, J = 7.5 Hz, 7.8 Hz,
3
1H, Harom), 5.39 (dd, J1-H,2-H = 9.3 Hz, 1H, 1-H), 5.00, 4.90, 4.49
3
2
0
(3 ꢄ s, 4H, OH), 3.67 (dd, J5-H,6-H = 5.4 Hz, J6-H,6 -H = 11.4 Hz, 1H,
6-H), 3.58–3.14 (m, 5H, 2-H, 3-H, 4-H, 5-H, 60-H); 13C NMR
(75.47 MHz, DMSO-d6): d 178.8 (C@S), 141.4 (CH@N), 136.4,
132.5, 130.7, 129.2, 127.1 and 122.3 (Carom), 84.2 (1-C), 78.7
(3-C), 77.7 (2-C), 71.8 (5-C), 69.9 (4-C), 60.9 (6-C). Anal. Calcd for
structures were superimposed using LSQKAB.39 The figures were
40
prepared with the programs MOLSCRIPT
,
or Bobscript41 and ren-
dered with Raster3D.42 The coordinates of the new structures have
been deposited with the RCSB Protein Data Bank (http://
C
14H18BrN3O5Sꢅ3H2O (474.32): C, 35.45; H, 5.10; N, 8.86. Found:
C, 35.42; H, 4.87; N, 8.45.

7922
K.-M. Alexacou et al. / Bioorg. Med. Chem. 18 (2010) 7911–7922
12. Beraldo, H.; Gambino, D. Mini-Rev. Med. Chem. 2004, 4, 31.
13. Belicchi-Ferrari, M.; Bisceglie, F.; Casoli, C.; Durot, S.; Morgenstern-
Badarau, I.; Pelosi, G.; Pilotti, E.; Pinelli, S.; Tarasconi, P. J. Med. Chem.
2005, 48, 1671.
14. Kovala-Demertzi, D.; Papageorgiou, A.; Papathanasis, L.; Alexandratos, A.;
Dalezis, P.; Miller, J. R.; Demertzis, M. A. Eur. J. Med. Chem. 2009, 44,
1296.
www.rcsb.org/pdb) with codes with codes 3MQF (3pF), 3MTA
(3mBr), 3MT7 (3pBr), 3MS7 (3oCl), 3MTB (3mCl), 3MT8 (3pCl),
3MS4 (3pCF3), 3MSC (3oNO2), 3MT9 (3pNO2), 3NC4 (3oOH),
3MRV (3mOH), 3MTD (3pOH), 3MRX (3pOMe), 3MS2 (3pMe)
and 3MRT (3pyr).
15. Kovala-Demertzi, D.; Yadav, P. N.; Wiecek, J.; Skoulika, S.; Varadinova, T.;
Demertzis, M. A. J. Inorg. Biochem. 2006, 100, 1558.
16. Tenchiu (Deleanu), A.-C.; Kostas, I. D.; Kovala-Demertzi, D.; Terzis, A.
Carbohydr. Res. 2009, 344, 1352.
Acknowledgements
This work was supported by the EU Marie Curie Early Stage
Training (EST) contract no. MEST-CT-020575 and by the Commis-
sion of the European Communities—under the FP7 ‘SP4-Capacities
Coordination and Support Action, Support Actions’ EUROSTRUCT
project (CSA-SA_FP7-REGPOT-2008-1 Grant Agreement No.
230146). This work was also supported by grants from European
Community—Research Infrastructure Action under the FP6 ‘Struc-
turing the European Research Area’ Programme (through the Inte-
grated Infrastructure Initiative ‘Integrating Activity on Synchrotron
and Free Electron Laser Science’) for work at the Synchrotron Radi-
ation Source MAX-lab, Lund, Sweden, and EMBL-Hamburg Outsta-
tion, Germany.
17. Zhang, S. S.; Ye, S. J.; Li, X. M.; Gu, S. S.; Liu, Q. Chem. Res. Chin. U 2005, 21, 545.
18. Ghosh, S.; Misra, A. K.; Bhatia, G.; Khan, M. M.; Khanna, A. K. Bioorg. Med. Chem.
Lett. 2009, 19, 386.
19. Zemplén, G.; Gerecs, A.; Hadácsy, I. Ber. Dtsch. Chem. Ges. 1936, 69, 1827.
20. Johnson, L. N. FASEB J. 1992, 6, 2274.
21. OReilly, M.; Watson, K. A.; Schinzel, R.; Palm, D.; Johnson, L. N. Nat. Struct. Biol.
1997, 4, 405.
22. Watson, K. A.; Schinzel, R.; Palm, D.; Johnson, L. N. EMBO J. 1997, 16, 1.
23. Barford, D.; Johnson, L. N. Nature 1989, 340, 609.
24. Martin, J. L.; Veluraja, K.; Ross, K.; Johnson, L. N.; Fleet, G. W. J.; Ramsden, N. G.;
Bruce, I.; Orchard, M. G.; Oikonomakos, N. G.; Papageorgiou, A. C.; Leonidas, D.
D.; Tsitoura, H. S. Biochemistry (USA) 1991, 30, 10101.
25. Alexacou, K. M.; Hayes, J. M.; Tiraidis, C.; Zographos, S. E.; Leonidas, D. D.;
Chrysina, E. D.; Archontis, G.; Oikonomakos, N. G.; Paul, J. V.; Varghese, B.;
Loganathan, D. Proteins 2008, 71, 1307.
26. Howard, J. A. K.; Hoy, V. J.; OHagan, D.; Smith, G. T. Tetrahedron 1996, 52,
12613.
27. Tsirkone, V. G.; Tsoukala, E.; Lamprakis, C.; Manta, S.; Hayes, J. M.; Skamnaki, V.
T.; Drakou, C.; Zographos, S. E.; Komiotis, D.; Leonidas, D. D. Bioorg. Med. Chem.
2010, 18, 3413.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.09.039.
28. Auffinger, P.; Hays, F. A.; Westhof, E.; Ho, P. S. Proc. Natl. Acad. Sci. U.S.A. 2004,
101, 16789.
29. Politzer, P.; Lane, P.; Concha, M. C.; Ma, Y. G.; Murray, J. S. J. Mol. Model. 2007,
13, 305.
30. Lindhorst, T. K.; Kieburg, C. Synthesis-Stuttgart 1995, 1228.
31. Cateni, F.; Bonivento, P.; Procida, G.; Zacchigna, M.; Favretto, L. G.; Scialino, G.;
Banfi, E. Bioorg. Med. Chem. 2007, 15, 815.
References and notes
1. Green, A.; Christian Hirsch, N.; Pramming, S. K. Diabetes Metab. Res. Rev. 2003,
19, 3.
32. Although it is
a known compound, full analytical data is not accessible:
2. Aiston, S.; Andersen, B.; Agius, L. Diabetes 2003, 52, 1333.
3. Aiston, S.; Coghlan, M. P.; Agius, L. Eur. J. Biochem. 2003, 270, 2773.
4. Oikonomakos, N. G. Curr. Protein Pept. Sci. 2002, 3, 561.
5. Oikonomakos, N. G.; Somsak, L. Curr. Opin. Invest. Drugs 2008, 9, 379.
6. Treadway, J. L.; Mendys, P.; Hoover, D. J. Expert Opin. Invest. Drugs 2001, 10, 439.
7. Baker, D. J.; Timmons, J. A.; Greenhaff, P. L. Diabetes 2005, 54, 2453.
8. Somsak, L.; Czifrak, K.; Toth, M.; Bokor, E.; Chrysina, E. D.; Alexacou, K. M.;
Hayes, J. M.; Tiraidis, C.; Lazoura, E.; Leonidas, D. D.; Zographos, S. E.;
Oikonomakos, N. G. Curr. Med. Chem. 2008, 15, 2933.
9. Oikonomakos, N. G.; Skamnaki, V. T.; Tsitsanou, K. E.; Gavalas, N. G.; Johnson, L.
N. Structure 2000, 8, 575.
10. Rath, V. L.; Ammirati, M.; Danley, D. E.; Ekstrom, J. L.; Gibbs, E. M.; Hynes, T. R.;
Mathiowetz, A. M.; McPherson, R. K.; Olson, T. V.; Treadway, J. L.; Hoover, D. J.
Chem. Biol. 2000, 7, 677.
Cao, K.; Wang, Z.; Zhao, X. Shiyou Daxue Xuebao, Ziran Kexueban 2004, 28,
100.
33. Melpidou, A. E.; Oikonomakos, N. G. FEBS Lett. 1983, 154, 105.
34. Saheki, S.; Takeda, A.; Shimazu, T. Anal. Biochem. 1985, 148, 277.
35. Oikonomakos, N. G.; Melpidou, A. E.; Johnson, L. N. Biochim. Biophys. Acta 1985,
832, 248.
36. Otwinowski, Z.; Minor, W.. In Methods Enzymology; Carter, C. W. J., Sweet, R. M.,
Eds.; Academic Press: New York, 1997; Vol. 276, p 307.
37. Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr., Sect. D 1997, 53,
240.
38. Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M. J. Appl.
Crystallogr. 1993, 26, 283.
39. CCP4 Acta Crystallogr., Sect. D 1994, 50, 760.
40. Kraulis, P. J. J. Appl. Crystallogr. 1991, 24, 946.
41. Esnouf, R. M. J. Mol. Graphics Modell. 1997, 15, 132.
42. Merritt, E. A.; Bacon, D. J. Macromol. Crystallogr. 1997, B277, 505.
11. Oikonomakos, N. G.; Kosmopoulou, M.; Zographos, S. E.; Leonidas, D. D.;
Chrysina, E. D.; Somsak, L.; Nagy, V.; Praly, J. P.; Docsa, T.; Toth, B.; Gergely, P.
Eur. J. Biochem. 2002, 269, 1684.