Structure-activity relationships and molecular modelling of 5-arylidene-2,4-thiazolidinediones active as aldose reductase inhibitors

Article abstract of DOI:10.1016/j.bmc.2005.02.026

The structure-activity relationships (SARs) of 5-arylidene-2,4- thiazolidinediones active as aldose reductase inhibitors (ARIs) were extended by varying the substitution pattern on the 5-arylidene moiety and on N-3. In particular, the introduction of an a

Full text of DOI:10.1016/j.bmc.2005.02.026

Bioorganic & Medicinal Chemistry 13 (2005) 2809–2823
Structure–activity relationships and molecular
modelling of 5-arylidene-2,4-thiazolidinediones active
as aldose reductase inhibitors
a
a
a
Rosanna Maccari,a, Rosaria Ottana, Carmela Curinga, Maria Gabriella Vigorita,
*
`
Dietmar Rakowitz,b Theodora Steindlb and Thierry Langerb
a
`
`
Dipartimento Farmaco-chimico, Facolta di Farmacia, Universita di Messina, Viale SS. Annunziata, 98168 Messina, Italy
bInstitute of Pharmacy, Department of Pharmaceutical Chemistry, University of Innsbruck, Innrain 52a, A-6020 Innsbruck, Austria
Received 13 July 2004; revised 4 February 2005; accepted 15 February 2005
Abstract—The structure–activity relationships (SARs) of 5-arylidene-2,4-thiazolidinediones active as aldose reductase inhibitors
(ARIs) were extended by varying the substitution pattern on the 5-arylidene moiety and on N-3. In particular, the introduction
of an additional aromatic ring or an H-bond donor group on the 5-benzylidene ring enhanced ALR2 inhibitory potency. Moreover,
the presence of a carboxylic anionic chain on N-3 was shown to be an important, although not essential, structural requisite to pro-
duce high levels of ALR2 inhibition. The length of this carboxylic chain was critical and acetic acids 4 were the most effective inhibi-
tors among the tested derivatives. Molecular docking simulations into the ALR2 active site accorded with the in vitro inhibition
data. They allowed the rationalization of the observed SARs and provided a pharmacophoric model for this class of ARIs.
Ó 2005 Elsevier Ltd. All rights reserved.
1. Introduction
Therefore, ALR2 has been considered as a target en-
zyme to develop drugs able to prevent the onset and
to check the progression of diabetic complications, even
in the presence of imperfect control of glycaemia.
Aldose reductase (EC 1.1.1.21, ALR2) is an aldo-keto
reductase involved in the development of degenerative
long-term complications of diabetes mellitus. It is the
first enzyme of the polyol pathway and catalyses the
NADPH-dependent reduction of glucose to sorbitol.
An increased flux of glucose through the polyol pathway
occurs under conditions of hyperglycaemia, such as in
diabetes, in tissues possessing insulin-independent glu-
cose transport (retina, lenses, kidney, nerve). It was
demonstrated that the activation of this metabolic path-
way is linked to the onset and progression of chronic
diabetic complications (neuropathy, nephropathy, reti-
nopathy, cataracts), in consequence of the deprivation
of NADPH and NAD+ and intracellular accumulation
of sorbitol with subsequent altered cellular redox poten-
tials and osmotic imbalance.1–5
Over the last three decades numerous ALR2 inhibitors
(ARIs) have been identified, most of which belong to
the carboxylic acids (such as zopolrestat, ponalrestat,
epalrestat) and hydantoins (such as sorbinil) classes of
compounds (Fig. 1).1a,2,6–9 Nevertheless, many of them
have shown to be clinically inadequate because of ad-
verse pharmacokinetics or toxic side-effects.8 At present,
epalrestat is the only ARI available on the market.9,10
Due to the shortage of drugs currently available for the
treatment of diabetic complications, the search for new
ARIs endowed with more favourable biological proper-
ties is still a major pharmaceutical challenge.
We have recently reported several (Z)-5-arylidene-2,4-
thiazolidinediones (1) (Fig. 1) designed in view of the
structural requisites essential for ALR2 inhibition (an
acidic proton, H-bond acceptor groups and a lipophilic
aromatic moiety);9,11–17 in fact, they were shown to
be effective as in vitro ARIs.18 Some N-unsubstituted
Keywords: 2,4-Thiazolidinediones; Aldose reductase; Diabetes melli-
tus; Molecular modelling.
*
Corresponding author. Tel.: +39 90 6766406; fax: +39 90 355613;
e-mail: rmaccari@pharma.unime.it
0968-0896/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2005.02.026

2810
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
F
O
S
O
H
N
H
N
a
O
O
R =
R =
Br
+
Ar CHO
N
S
O
O
N
N
N
R
Ponalrestat
O
F
Ar
H
O
CF3
2 c-k
COOH
N
S
Sorbinil
b
Zopolrestat
O
O
COOCH3
X
O
S
X
COOH
N
c
O
N
R
O
N
COOH
S
N
O
S
O
S
S
Ar
Ar
H
H
H
Epalrestat
R = H, CH2COOCH3, CH2COOH
3 b-k X = CH2
5 a-d X = CH2-CH=CH
4 b-g, j, k X = CH2
6 a, b, d, g X = CH2-CH=CH
1
R'
R' = F, CH3, OCH3, CF3,
OC6H5, CH=NC6H4OH
Scheme 1. Reagents: (a) C5H11N, refluxing C2H5OH; (b) K2CO3,
acetone, BrCH2COOCH3 or BrCH2CH@CHCOOCH3; (c) AcOH,
HCl, reflux.
Figure 1.
derivatives possessed the same inhibitory potency of
sorbinil, the prototype of hydantoins, which was with-
drawn from clinical trials because its in vivo efficacy as
ARI was limited by hypersensitivity reactions.2,6,8 In
fact, 2,4-thiazolidinediones can be considered as hydan-
toin bioisosteres potentially devoid of these adverse ef-
fects, which are related to the hydantoin ring.2,8,18
The N-3 position was maintained unsubstituted or
substituted with an acetic chain (compounds 2, 3, 4,
Scheme 1). Afterwards the N-3 acetic chain was replaced
by the residue of 2-butenoic acid (compounds 5, 6,
Scheme 1). We also decided to synthesize and assay in
our experimental conditions some 2,4-thiazolidinedione
analogues already reported as ARIs (compounds 4b, d,
e, k and 3b, d, e, k)23 with the aim of depicting an exten-
sive SARs picture.
In our compounds the introduction of an acetic chain on
N-3 of the thiazolidinedione ring led to a marked in-
crease in inhibitory potency.18 Moreover, some of the
corresponding methyl esters, although devoid of any
acidic functionality, showed appreciable inhibitory
properties; this finding suggested that the polar N-3 ace-
tate chain was able to reach and bind the polar posi-
tively charged recognition region of the ALR2 active
site formed by Tyr48, His110, Trp111 residues and the
nicotinamide ring of cofactor NADP+.2,8,9,13,14,19 In
any case, the higher inhibitory potency of the analogous
carboxylic acids confirmed the importance of an anionic
head in the interaction with the recognition region of the
enzyme. However, in vivo the ester function could im-
prove the crossing of biomembranes and, analogously,
the N-unsubstituted derivatives might display more
favourable pharmacokinetics due to their pKa values
higher than carboxylic acids.20,21
The ALR2 inhibitory activity of 2,4-thiazolidinediones
2–6 was evaluated in vitro using partially purified
ALR2 from bovine lenses.
The results were rationalized by means of docking simu-
lations into the ALR2 active site.
2. Chemistry
5-Arylidene-2,4-thiazolidinediones 2 were synthesized,
according to a known procedure,18,24 by the condensa-
tion of commercially available 2,4-thiazolidinedione
with suitable benzaldehydes in refluxing ethanol in the
presence of piperidine (Scheme 1).
Esters 3 and 5 were prepared by the reaction between N-
unsubstituted 2,4-thiazolidinediones 2 and methyl bro-
moacetate or 4-bromocrotonate, respectively, using
potassium carbonate as base. Alternatively, acetates 3
could be produced by the reaction of 2 with methyl bro-
moacetate in DMF in the presence of sodium hydride,18
with generally lower yields.
On the basis of these results, in this work we amplified
the SARs of this class of ARIs.
Firstly, we varied the 5-arylidene moiety by synthesiz-
ing: (a) derivatives endowed with a larger lipophilic 5-
arylidene moiety; (b) analogues bearing an H-bond do-
nor group on the 5-benzylidene ring. In fact, it is known
that, except for the polar positively charged recognition
region, the nature of the ALR2 active site is highly
hydrophobic and thus large lipophilic moieties can fa-
vour the interaction of inhibitors with the enzyme.2,13–15,22
Otherwise, the enzyme–inhibitor complex stability could
be increased by means of further interactions through
H-bonds.14,15,18
The synthesis of 5-(3-hydroxybenzylidene)- and 5-(4-
hydroxybenzylidene) substituted acetates 3f and 3g
was carried out by adding methyl bromoacetate drop
wise to a solution of 2f or 2g, respectively, and potas-
sium carbonate in acetone over 1.5 h, in order to prevent
the nucleophilic attack of hydroxyl group to bromo-
acetate; after refluxing of the mixture for 24 h and

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2811
crystallization from methanol, only the desired 5-(3-hydr-
oxybenzylidene)- and 5-(4-hydroxybenzylidene) substi-
tuted acetates were obtained in high yields.
only as Z isomers; in fact, their 1H NMR spectra
showed only one signal attributable to the resonance
of the 5-methylidene proton in the range 7.50–
8.70 ppm. In their 13C NMR spectra, the 5-methylidene
carbon and C5 of the thiazolidinedione ring resonated in
the ranges 130.5–140.7 and 117.5–128.0 ppm,
respectively.
5-(3-Aminobenzylidene)- and 5-(4-aminobenzylidene)-
2,4-thiazolidinediones (2h and 2i) were synthesized start-
ing from 3- and 4-nitrobenzylidene substituted ana-
logues 2j and 2k by the reduction of the nitro group
with tin(II) chloride in refluxing ethanol. Analogously
3-aminobenzylidene and 4-aminobenzylidene substi-
tuted acetates 3h and 3i were obtained starting from ace-
tates 3j and 3k, respectively.
1
In the H and 13C NMR spectra of esters 3, 5 and acids
4, 6 a signal attributable to N–CH2 resonance was diag-
1
nostic: in particular H NMR spectra showed a singlet
at 4.35–4.55 ppm for acetic derivatives 3, 4 and a double
doublet at 4.42–4.52 ppm for 2-butenoic derivatives 5, 6.
In their 13C NMR spectra, besides two signals attribut-
able to the resonances of 2- and 4-carbonylic groups of
the thiazolidinedione ring at 164.5–168.9 ppm, another
signal due to the resonance of the carboxylic carbon ap-
peared in the same range.
The acidic hydrolysis of 3 and 5 provided corresponding
carboxylic acids 4 and 6 (Scheme 1).
The acidic hydrolysis of 5-(4-benzyloxybenzylidene)
substituted methyl esters 3c and 5c led to [5-(4-hydroxy-
benzylidene)-2,4-dioxothiazolidin-3-yl]acetic acid (4g)
and 4-[5-(4-hydroxybenzylidene)-2,4-dioxothiazolidin-3-yl]-
2-butenoic acid (6g), respectively. Therefore [5-(4-ben-
zyloxybenzylidene)-2,4-dioxothiazolidin-3-yl]acetic acid
(4c) was prepared starting from the reaction between
commercially available 2,4-thiazolidinedione and methyl
bromoacetate in acetone in the presence of potassium
carbonate, followed by acidic hydrolysis. (2,4-Dioxo-
thiazolidin-3-yl)acetic acid (8) was then converted to
4c by condensation with 4-benzyloxybenzaldehyde
(Scheme 2).
The structure of acids 4, 6 was also confirmed by means
of their IR spectra, which showed a very broad band in
the region of 3450–2450 cmÀ1 attributable to the stretch-
ing of carboxylic OH.
3. Results and discussion
3.1. Aldose reductase inhibition
Newly synthesized compounds 2–6 were evaluated for
their ability to inhibit the in vitro reduction of D,L-glyc-
eraldehyde by partially purified ALR2 from bovine
lenses, using sorbinil and ponalrestat25 as reference
drugs (Table 1).
Attempts to obtain [5-(3-aminobenzylidene)-2,4-dioxo-
thiazolidin-3-yl]acetic acid and its 4-amino substituted
isomer, through different synthetic pathways (hydrolysis
of corresponding methyl esters 3h and 3i, reduction of
acids 4j and 4k with tin(II) chloride) failed.
Our previous results had indicated that the presence of
an additional aromatic ring on the 5-benzylidene group
increased the ALR2 inhibitory activity of 5-arylidene-
2,4-thiazolidinediones, very likely through further van
der Waals interactions with the hydrophobic region of
the active site of the enzyme.18
Analytical and spectroscopic data (IR, 1H and 13C NMR)
confirmed the structures assigned to compounds 2–6.
Analogously to their previously reported analogues 1,18
5-arylidene-2,4-thiazolidinediones 2–6 were obtained
Thus, starting from the significant activity of [2,4-dioxo-
5-(3-phenoxybenzylidene)thiazolidin-3-yl]acetic acid (4a,
IC50 = 0.13 lM),18 we first synthesized its 4-phenoxy
substituted analogue (4b). The displacement of the 3-
phenoxy group to position 4 of the 5-benzylidene moiety
reduced activity and, in fact, acid 4b was shown to be
about six times less potent than 4a; corresponding
methyl ester 3b was also significantly less active than
3a (Table 1).
O
S
O
S
COOCH3
H
N
a
N
b
O
O
7
O
O
S
The replacement of the 4-phenoxy with the 4-benzyloxy
group enhanced activity leading to a good ALR2 inhib-
itor (4c, IC50 = 0.28 lM). Similar inhibitory effect was
COOH
COOH
c
N
N
S
O
O
shown by 4-phenyl substituted analogue 4d (IC50
=
H
8
0.26 lM), whereas N-unsubstituted derivatives 2c, 2d
and methyl esters 3c, 3d were almost inactive, producing
10–35% inhibition at 50 lM dose (Table 1).
O
4c
The condensation of a further benzene ring to the 5-benz-
ylidene group led to 5-naphthalen-1-ylmethylene substi-
tuted derivatives endowed with different inhibitory
Scheme 2. Reagents: (a) K2CO3, acetone, BrCH2COOCH3; (b) AcOH,
HCl, reflux; (c) 4-benzyloxybenzaldehyde, C5H11N, refluxing
C2H5OH.

2812
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
Table 1. In vitro bovine lenses ALR2 inhibitory activity of 2,4-thiazolidinediones 2–4
O
S
R
N
2 R = H
3 R = CH2COOCH3
O
4 R = CH2COOH
Ar
H
a
a
a
Ar
Compd
IC50
Compd
IC50
Compd
IC50
2ab
6.14 (5.54–6.81)
3ab
1.32 (1.00–1.74)
4ab
0.13 (0.11–0.15)
O
O
2bb
2c
41% (37 lM)
10% (50 lM)
20% (50 lM)
10.7 (9.58–11.97)
3b
3c
3d
3e
33% (50 lM)
12% (50 lM)
35% (50 lM)
12% (50 lM)
4b
4c
4d
4e
0.82 (0.76–0.88)
0.28 (0.12–0.65)
0.26 (0.18–0.37)
0.17 (0.11–0.27)
CH2O
2d
2e
HO
2f
2g
2h
2i
10.7 (9.91–11.65)
8.96 (6.64–12.09)
20.2 (18.4–22.2)
44% (50 lM)
3f
3g
3h
3i
43% (50 lM)
4f
0.66 (0.48–0.91)
0.15 (0.10–0.22)
6.18 (5.74–6.66)
39.1 (38.6–39.6)
25% (50 lM)
4g
HO
H2N
H2N
O2N
2j
21% (12.5 lM)
3j
22% (12.5 lM)
8% (12.5 lM)
4j
0.49 (0.45–0.54)
1.19 (1.08–1.30)
2k
Sorbinil
Ponalrestat
37% (12.5 lM)
1.10 (1.01–1.19)
0.08 (0.073–0.087)
3k
4k
O2N
a IC50 (lM) (95% C.L.) or % inhibition at the given concentration.
b Ref. 18.
potency: (5-naphthalen-1-ylmethylene-2,4-dioxothiazo-
lidin-3-yl)acetic acid (4e) was shown to be a potent
ALR2 inhibitor (IC50 = 0.17 lM); N-unsubstituted ana-
logue 2e displayed activity about 60 times lower than
that of 4e, whereas corresponding methyl ester 3e was
ineffective (Table 1). Compound 4e was shown to be
slightly more potent than (5-naphthalen-2-ylmethylene-
2,4-dioxothiazolidin-3-yl)acetic acid, already known,26
which in our assay displayed IC50 value (0.22 lM) very
close to the reported one (0.27 lM).26
2,4-thiazolidinedione 2g and methyl ester 3g showed
micromolar IC50 values, whereas, as expected, acid
4g exerted significant inhibitory activity (IC50
=
0.15 lM), which was 60 and 40 times higher than 2g
and 3g, respectively. Compound 4g proved to be
the most active among the 5-arylidene-2,4-thiazolidined-
iones reported here, along with the above-mentioned
5-naphthalen-1-ylmethylene
(4e), which exhibited the same inhibitory potency
substituted
analogue
(Table 1).
In the meantime, we synthesized 5-arylidene-2,4-thiazo-
lidinediones bearing a hydroxy group on the 5-arylidene
moiety as H-bond donor group.
The displacement of the hydroxy group to position 3 of
the 5-benzylidene ring decreased inhibitory effectiveness;
also in this case, acid 4f was the most active among the
tested 5-(3-hydroxybenzylidene) substituted analogues,
exhibiting a submicromolar IC50 value, which was 4
times higher than that of 4g.
5-(4-Hydroxybenzylidene) substituted derivatives pro-
duced appreciable ALR2 inhibition: N-unsubstituted

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2813
The isosteric replacement of the hydroxy group with the
amino group, which is a less efficacious H-bond donor
group, led to less active compounds (2, 3 h, i). The de-
crease in inhibitory effectiveness was more marked in
3- and 4-nitro substituted analogues (2–4 j, k), which
were devoid of any H-bond donor group in the 5-arylid-
ene moiety.
chosen to represent the series of N-unsubstituted 5-
arylidene-2,4-thiazolidinediones 2. Acid 6g was the
most active among the tested 2-butenoic acids 6
(Table 2).
In order to find a crystallographic structure of ALR2
suitable for the docking experiments, the Brookhaven
Protein Database was searched and several entries of al-
dose reductase co-crystallized with different inhibitors
were studied in detail. Although biological testing was
performed using the enzyme purified from bovine lens,
the choice of the structure of human aldose reductase
is justified since there is 82% sequence identity and
89% similarity (BLAST analysis, score 556 bits) of the
human enzyme with the bovine one. Finally, the PDB
entry 1EL3, containing the structure of human aldose
reductase complexed with the inhibitor Idd384 (Fig.
2),27 was selected because of its relatively high resolution
On the whole these results confirmed that in 5-arylidene-
2,4-thiazolidinediones, active as ARIs, the presence of
an anionic carboxylic group on N-3 is an important,
even though not essential, structural requisite to pro-
duce high levels of enzyme inhibition. In fact, all acetic
acids 4 exerted high levels of ALR2 inhibition; they
exhibited submicromolar IC50 values and proved to be
more active than sorbinil (with the only exception of
4k), nearly reaching the inhibitory potency of ponalre-
stat (Table 1).
˚
value (1.7 A). Moreover, the enzyme was co-crystallized
The length of the carboxylic chain on N-3 was also
shown to be critical for activity. In fact, the replacement
of the N-3 acetic chain with the residue of 2-butenoic
acid (esters 5 and acids 6) was detrimental for activity
(Table 2). Only 5-(4-hydroxybenzylidene) substituted
derivative 6g exerted appreciable ALR2 inhibition at
micromolar doses (IC50 = 4.17 lM).
with Idd384, which shows a structural resemblance to
our ARIs; in fact, the structures of the selected 2,4-thi-
azolidinediones do not include a benzothiazole ring able
to induce large conformational changes and to open a
new hydrophobic pocket in the ALR2 active site, as in
the case of zopolrestat.8,28 Thus we assumed that our
ARIs and Idd384 might bind to the enzyme in a similar
way. The LigandFit module implemented in the soft-
ware package Cerius2 was used for the docking and
scoring procedure.29 The final docked pose of Idd384
was found to be in good agreement with the orientation
of the ligand in the crystal structure, confirming the abil-
ity of the method to accurately predict the binding
conformation.
3.2. Molecular modelling
Six compounds were selected for computational analy-
sis, namely 2a, 3a, 4a, 4b, 4g and 6g. Acetic acids 4a,
4b, 4g and acetate 3a are among the most active ARIs
tested in our laboratories (Table 1). Compound 2a was
Table 2. In vitro bovine lenses ALR2 inhibitory activity of 2,4-thiazolidinediones 5, 6
O
S
R
N
5 R = CH2CH=CHCOOCH3
6 R = CH2CH=CHCOOH
O
Ar
H
a
a
Ar
Compd
IC50
Compd
IC50
O
5a
25% (50 lM)
6a
19% (50 lM)
57% (50 lM)
O
5b
5c
0% (50 lM)
23% (50 lM)
6b
CH2O
5d
6% (50 lM)
6d
6g
21% (50 lM)
4.17 (3.05–5.71)
HO
Sorbinil
Ponalrestat
a IC50 (lM) (95% C.L.) or % inhibition at the given concentration.
1.10 (1.01–1.19)
0.08 (0.073–0.087)

2814
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
Figure 2. Pharmacophore model representing the interactions of Idd384 with the ALR2 active site.
Information about the interactions of Idd384 within the
ALR2 active site was used in the pharmacophore design
process by means of the Catalyst program.30 The carb-
oxylate group interacts with the positively charged
nitrogen of the NADP+ and additionally contributes
to binding by the formation of hydrogen bonds with
Tyr48. Further hydrogen bonds with Trp111 and
His110 are accepted by the sulfone oxygen atoms. The
lipophilic moieties of the inhibitor interact with the
ALR2 active site through hydrophobic bonds. Thus this
pharmacophore consists of three H-bond acceptor fea-
tures and three hydrophobic areas. In Figure 2, the
superposition of Idd384 with the pharmacophore model
is shown.
and all the conformers that were shown to map the
pharmacophores in a Idd384 like mode by at least two
of the three hypotheses were selected for docking and
imported into the Cerius2 software package. This pre-
selection procedure of conformers within the Catalyst
program proved to be more valuable than flexible ligand
docking, since the algorithm in Cerius2 could not pro-
vide satisfying conformational sampling for this target
resulting in conformations and poses to be considerably
diverse from the known binding mode of Idd384.
The files containing the most suitable 38 conformers
were used for a rigid fit docking procedure followed
by a final energy minimization step in the protein, which
ensured that optimal results were obtained even from
non-ideal conformations. Thereby, 230 different orienta-
tions of the 38 conformers were retrieved. These 230 li-
gand–enzyme complexes were evaluated via the
computation of scores using all the scoring functions
implemented in Cerius2: Ligscore1, Ligscore2, PLP1,
PLP2, JAIN, PMF and LUDI. The concept of consen-
sus scoring was applied, which proved to be a valuable
tool for improving the prediction efficiency of scoring
functions.
The molecular models of the selected 2,4-thiazolidinedi-
ones were built using Catalyst, which was also used to
generate conformational models, feature-based pharma-
cophore hypotheses, and search for conformers that fit
these hypotheses best. This procedure was used in order
to detect beforehand those conformers that closely resem-
ble the protein-bound conformation of Idd384 and most
likely are the binding conformations of our ARIs.
Common pharmacophoric features shared by all se-
lected 5-arylidene-2,4-thiazolidinediones were identified
as follows: two H-bond acceptors representing the carb-
oxylic group and one sulfone oxygen atom of Idd384
together with a hydrophobic region representing the
dimethylphenyl system (Fig. 3). Starting from the initial
model derived from the active protein-bound conforma-
tion of Idd384, three pharmacophore hypotheses were
manually built, which slightly varied in size and location
of the features in order to cover the flexibility of the side
chains within the active site. The selected 2,4-thiazolid-
inediones were fitted on the pharmacophore models
Especially LigScore1 and LigScore2 reflected very well
the ranking order of measured in vitro activities. We
therefore suggest that LigScore1 and LigScore2 were
the scoring functions of choice for this structural class
of ARIs in our experimental conditions. Nevertheless,
it must be mentioned that differences in binding affinities
of the compounds were rather small in some cases and
therefore hardly elusive by means of scoring.
In Figure 4 the superposition of docked structures of
compounds 3a, 4a and Idd384 is shown together with

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2815
Figure 3. Compounds 2a, 3a and 4a (from the top to the bottom) mapped on a common features pharmacophore model (Green: H bond acceptor;
Cyan: lipophilic region).
the most important interacting amino acids within the
binding site. As known, ARL2 inhibition strongly de-
pends on charge–charge interactions and hydrogen
bonds of inhibitors with the recognition region formed
by Tyr48, His110, Trp111 and NADP+ and on aro-
matic–aromatic interactions with the hydrophobic re-
gion of the ALR2 active site.2,8,9,13,14,19 Most of these
bonds as well as new possibilities for interaction were
evidenced when the 5-arylidene-2,4-thiazolidinediones
under study were docked into the active site of the en-
zyme (Figs. 4–6). When measuring the distances between
the atoms belonging to their polar part and the sur-
rounding amino acids, additional and alternative possi-
bilities for the formation of hydrogen bonds could be
observed. Two novel hydrogen bonds for inhibitor 4a
were enabled by Cys298; the median phenyl ring was
centred between the hydrophobic Pro218, Phe122 and
Trp219 (Fig. 5).
Figure 6 shows the image of the docked compound 4g.
This complex revealed Trp20 a new hydrogen bond
partner for one of the cyclic imide carbonyl oxygens;
Cys298 and Tyr48 could interact with the carboxylate
moiety and the phenyl substituent was surrounded by
the hydrophobic side chains of Trp20, Pro218 and
Trp219.
It is worth mentioning that LigandFit keeps the protein
rigid during the docking procedure. Thus, other possible
H-bonds between the enzyme and our inhibitors might
not be evidenced in these experimental conditions and
slightly enlarged distance compared to the boundary

2816
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
Figure 4. Superposition of docked structures of compounds 3a (red) and 4a (orange) with the bound conformation structure of Idd384 (purple)
(relaxed stereo view).
Figure 5. Possibilities of hydrogen bond interactions of docked inhibitor 4a in the ALR2 active site.
values for hydrogen bonds had to be accepted as possi-
ble new interactions.
compensate the obvious lack of charge–charge interac-
tion with NADP+, occurred.
No explanation could be provided by this docking meth-
odology concerning the high activity of ester derivative
3a in comparison with free acid 4a. Although possibili-
ties of hydrogen bonds between the carbonyl oxygen
atoms and the ALR2 active site were observed, no
additional highly favourable interactions, which might
4. Conclusions
The ALR2 inhibitory data reported here confirmed that
5-arylidene-2,4-thiazolidinediones can be considered as
promising ARIs and allowed us to extend their struc-
ture–activity relationships.

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2817
Figure 6. Possibilities of hydrogen bond interactions of docked inhibitor 4g in the ALR2 active site.
The introduction of an additional aromatic ring or an
H-bond donor group on the 5-benzylidene ring was
shown to be very favourable for the ALR2 inhibitory
effectiveness of these compounds.
recorded on a Varian 300 magnetic resonance spectro-
meter (300 MHz for1H and 75 MHz for 13C). Chemical
shifts are given in d units (ppm) relative to the internal
standard Me4Si and refer to CDCl3 or DMSO-d6
solutions. Coupling constants (J) are given in hertz
(Hz). 13C NMR spectra were determined by attached
proton test (APT) experiments and the resonances
were always attributed by proton–carbon heteronu-
clear chemical shift correlation. Mass spectra were
recorded on a Shimadzu GC17A/MSQP5050 spectro-
meter.
The presence of the carboxylic anionic head of the N-3
acetic chain provided the highest activity levels and al-
lowed the identification of some potent inhibitors,
whose effectiveness was higher than that of sorbinil
and similar to that of ponalrestat. In addition, our re-
sults indicated that the substituent on N-3 could be a po-
lar group able to get to and bind the polar recognition
region of ALR2 or a hydrogen atom, although in these
cases the inhibitory potency decreased.
Unless stated otherwise, all materials were obtained
from commercial suppliers and used without further
purification.
The inhibition data obtained were rationalized by means
of molecular docking simulations into the ALR2 active
site, which were in agreement with the above-mentioned
structure–activity relationships. The molecular model-
ling experiments provided a pharmacophoric model
for this class of ARIs.
5.2. General method for the synthesis of 5-arylidene-2,4-
thiazolidinediones 2c–g, j, k
A mixture of 2,4-thiazolidinedione (2.4 g, 20 mmol),
aldehyde (20 mmol), piperidine (1.4 g, 16 mmol) and
EtOH (150 mL) was refluxed for 24 h. The reaction mix-
ture was poured into H2O and acidified with AcOH to
give 2c–g, j, k as solids, which were recrystallized from
methanol.
5. Experimental
5.1. Chemistry
5.2.1. 5-(4-Benzyloxybenzylidene)-2,4-thiazolidinedione
(2c). Yield 96%; mp 199–202 °C; 1H NMR (DMSO-
d6): d 5.19 (s, 2H, CH2); 7.18 (m, 2H, arom); 7.38–7.45
(m, 5H, arom); 7.56 (m, 2H, arom); 7.75 (s, 1H, CH);
12.41 (br s, 1H, NH); 13C NMR (DMSO-d6): d 69.5
(CH2); 120.4 (5-C); 115.7, 127.9, 128.1, 128.6, 131.9
(CH arom); 132.2 (CH methylidene); 125.7, 136.5,
160.1 (Cq arom); 167.5, 168.1 (CO); Anal.
(C17H13NO3S) C, H, N.
Melting points were recorded on a Kofler hot-stage
apparatus and are uncorrected. TLC controls were car-
ried out on precoated silica gel plates (F 254 Merck).
Elemental analyses (C, H, N), determined by means of
a C. Erba mod. 1106 elem. Analyzer, were within
0.4% of theory. IR spectra were obtained with a Per-
kin–Elmer 683 spectrophotometer as Nujol or esachlo-
robutadiene mulls. 1H and 13C NMR spectra were

2818
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
5.2.2. 5-Biphenyl-4-ylmethylene-2,4-thiazolidinedione (2d).
Yield 72%; mp 249 °C; H NMR (DMSO-d6): d 7.41–
7.25 (m, 1H, arom); 7.29 (m, 1H, arom); 7.70 (m, 1H,
arom); 8.13 (s, 1H, CH); 13C NMR (DMSO-d6): d
114.2, 116.5, 118.4, 129.9 (CH arom); 122.5 (5-C);
132.9 (CH methylidene); 133.8, 149.6 (Cq arom);
167.6, 167.8 (CO); Anal. (C10H8N2O2S) C, H, N.
1
7.52 (m, 3H, arom); 7.68–7.76 (m, 4H, arom); 7.87 (m,
2H, arom); 7.84 (s, 1H, CH); 11.77 (br s, 1H, NH);
13C NMR (DMSO-d6): d 123.6 (5-C); 127.1, 127.7,
128.6, 129.5, 131.7 (CH arom); 131.1 (CH methylidene);
132.3, 139.0, 142.1 (Cq arom); 167.7, 168.3 (CO); Anal.
(C16H11NO2S) C, H, N.
5.2.9. 5-(4-Aminobenzylidene)-2,4-thiazolidinedione (2i).
The procedure was the same as reported for compound
2h, starting from 5-(4-nitrobenzylidene)-2,4-thiazolidine-
dione (2k). Yield 38 %; mp 231 °C; 1H NMR
(DMSO-d6): d 6.64 (m, 2H, arom); 7.27 (m, 2H, arom);
7.51 (s, 1H, CH); 13C NMR (DMSO-d6): d 114.6, 132.3
(CH arom); 122.2 (5-C); 132.5 (CH methylidene); 138.2,
151.1 (Cq arom); 167.2, 167.6 (CO); Anal.
(C10H8N2O2S) C, H, N.
5.2.3. 5-Naphthalen-1-ylmethylene-2,4-thiazolidinedione
1
(2e). Yield 88%; mp 202 °C; H NMR (DMSO-d6): d
7.36–7.51 (m, 4H, arom); 7.74 (m, 2H, arom); 7.95 (m,
1H, arom); 8.34 (s, 1H, CH); 11.98 (br s, 1H, NH);
13C NMR (DMSO-d6): d 123.1, 125.3, 126.1, 126.6,
127.2, 128.4, 128.6 (CH arom); 126.9 (5-C); 130.5 (CH
methylidene); 130.1, 130.8, 133.1 (Cq arom); 166.6,
167.8 (CO); Anal. (C14H9NO2S) C, H, N.
5.3. General method for the synthesis of (5-arylidene-2,4-
dioxothiazolidin-3-yl)acetic acids methyl esters 3b–g, j, k
5.2.4.
5-(3-Hydroxybenzylidene)-2,4-thiazolidinedione
1
(2f). Yield 94%; mp 265 °C; H NMR (DMSO-d6): d
6.89 (d, J = 7.8 Hz, 1H, arom); 6.99 (s, 1H, arom);
7.04 (d, J = 8.4 Hz, 1H, arom); 7.35 (dd, J = 8.4 and
7.8 Hz, 1H, arom); 7.70 (s, 1H, CH); 9.96 (br s, 1H,
NH); 13C NMR (DMSO-d6): d 116.0, 117.8, 121.4,
130.5 (CH arom); 123.5 (5-C); 132.0 (CH methylidene);
134.3, 157.9 (Cq arom); 167.5, 168.1 (CO); Anal.
(C10H7NO3S) C, H, N.
A
mixture of 5-arylidene-2,4-thiazolidinedione
2
(10 mmol), methyl bromoacetate (3.06 g, 20 mmol) and
potassium carbonate (2.76 g, 20 mmol) in acetone
(120 mL) was refluxed for 24 h. After cooling, the inor-
ganic salts were filtered off; then the solvent was evapo-
rated under reduced pressure and the solid residue was
recrystallized from methanol providing pure acetate 3.
5.3.1. [2,4-Dioxo-5-(4-phenoxybenzylidene)thiazolidin-3-
yl]acetic acid methyl ester (3b). Yield 72%; mp 113–
115 °C; H NMR (CDCl3): d 3.80 (s, 3H, CH3); 4.51
(s, 2H, CH2); 7.05–7.10 (m, 3H, arom); 7.21–7.25 (m,
2H, arom); 7.42–7.51 (m, 4H, arom); 7.91 (s, 1H, CH);
13C NMR (CDCl3): d 41.8 (CH2); 52.5 (CH3); 119.2
(5-C); 118.3, 119.9, 124.5, 129.9, 132.1 (CH arom);
133.8 (CH methylidene); 127.5, 155.5, 159.9 (Cq arom);
165.4, 166.5, 167.1 (CO); Anal. (C19H15NO5S) C, H, N.
5.2.5.
5-(4-Hydroxybenzylidene)-2,4-thiazolidinedione
1
1
(2g). Yield 68%; mp 252 °C; H NMR (DMSO-d6): d
6.93 (m, 2H, arom); 7.47 (m, 2H, arom); 7.72 (s, 1H,
CH); 10.33 (br s, 1H, NH); 13C NMR (DMSO-d6): d
116.3, 132.2 (CH arom); 123.9 (5-C); 132.3 (CH methy-
lidene); 119.0, 159.9 (Cq arom); 167.4, 168.0 (CO); Anal.
(C10H7NO3S) C, H, N.
5.2.6. 5-(3-Nitrobenzylidene)-2,4-thiazolidinedione (2j).
1
Yield 47%; mp 245 °C; H NMR (DMSO-d6): d 7.83
5.3.2. [5-(4-Benzyloxybenzylidene)-2,4-dioxothiazolidin-
3-yl]acetic acid methyl ester (3c). Yield 88%; mp 123–
125 °C; H NMR (CDCl3): d 3.80 (s, 3H, CH3); 4.51
(s, 2H, CH2); 5.15 (s, 2H, CH2O); 7.07 (m, 2H, arom);
7.42–7.48 (m, 7H, arom); 7.91 (s, 1H, CH); 13C NMR
(CDCl3): d 41.9 (CH2); 52.8 (CH3); 70.2 (CH2O); 118.0
(5-C); 115.6, 127.5, 128.3, 128.7, 132.0 (CH arom);
134.6 (CH methylidene); 125.8, 136.0, 160.8 (Cq arom);
165.7, 166.8, 167.6 (CO); Anal. (C20H17NO5S) C, H, N.
(dd, J = 7.8 and 7.5 Hz, 1H, arom); 7.94 (s, 1H, CH);
8.02 (d, J = 7.5 Hz, 1H, arom); 8.30 (d, J = 7.8 Hz,
1H, arom); 8.45 (s, 1H, arom); 13C NMR (DMSO-d6):
d 124.2, 124.3, 129.3, 135.3 (CH arom); 126.6 (5-C);
130.8 (CH methylidene); 134.7, 148.2 (Cq arom);
166.9, 167.1 (CO); Anal. (C10H6N2O4S) C, H, N.
1
5.2.7. 5-(4-Nitrobenzylidene)-2,4-thiazolidinedione (2k).
1
Yield 58%; mp 280 °C; H NMR (DMSO-d6): d 7.88 (s,
1H, CH); 7.91 (m, 2H, arom); 8.36 (m, 2H, arom); 13C
NMR (DMSO-d6): d 124.2, 129.0 (CH arom); 128.0
(5-C); 130.8 (CH methylidene); 139.3, 147.4 (Cq arom);
166.9, 167.2 (CO); Anal. (C10H6N2O4S) C, H, N.
5.3.3. (5-Biphenyl-4-ylmethylene-2,4-dioxothiazolidin-3-yl)-
tic acid methyl ester (3d). Yield 90%; mp 145 °C; 1H NMR
(CDCl3): d 3.81 (s, 3H, CH3); 4.52 (s, 2H, CH2); 7.41–7.48
(m, 3H, arom); 7.61 (m, 4H, arom); 7.73 (m, 2H, arom);
7.99 (s, 1H, CH); 13C NMR (CDCl3): d 41.9 (CH2); 52.9
(CH3); 120.6 (5-C); 127.1, 127.8, 128.2, 129.0, 130.8 (CH
arom); 134.3 (CH methylidene); 131.9, 139.6, 143.4 (Cq
arom); 165.6, 166.7, 167.4 (CO); Anal. (C19H15NO4S)
C, H, N.
5.2.8. 5-(3-Aminobenzylidene)-2,4-thiazolidinedione (2h).
Tin(II) chloride (6.07 g, 32 mmol), was added to a solu-
tion of 5-(3-nitrobenzylidene)-2,4-thiazolidinedione (2j)
(1.6 g, 6.4 mmol) in ethanol and the mixture was re-
fluxed for 2 h. After evaporation to dryness in vacuo,
a saturated solution of sodium hydrogencarbonate was
added to the crude solid. The mixture was then extracted
with ethyl acetate, the organic layer was dried (Na2SO4)
and the solvent removed under reduced pressure. The
solid was recrystallized from methanol. Yield 42%; mp
5.3.4. (5-Naphthalen-1-ylmethylene-2,4-dioxothiazolidin-
3-yl)acetic acid methyl ester (3e). Yield 52%; mp 124 °C;
1H NMR (CDCl3): d 3.83 (s, 3H, CH3); 4.55 (s, 2H,
CH2); 7.57–7.61 (m, 4H, arom); 7.95 (m, 2H, arom);
8.12 (m, 1H, arom); 8.69 (s, 1H, CH); 13C NMR
(CDCl3): d 41.9 (CH2); 52.9 (CH3); 123.4, 125.3, 126.6,
1
215 °C; H NMR (DMSO-d6): d 7.21 (m, 1H, arom);

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2819
126.8, 127.4, 128.9, 131.3 (CH arom); 124.1 (5-C); 131.9
(CH methylidene); 130.3, 131.7, 133.7 (Cq arom); 165.1,
166.7, 167.8 (CO); Anal. (C17H13NO4S) C, H, N.
3H, arom); 7.16 (m, 1H, arom); 7.78 (s, 1H, CH); 13C
NMR (DMSO-d6): d 41.8 (CH2), 52.5 (CH3), 114.2,
116.7, 118.3, 129.7 (CH arom); 119.3 (5-C); 133.1,
149.3 (Cq arom); 135.1 (CH methylidene); 164.7,
164.9, 167.0 (CO); Anal. (C13H12N2O4S) C, H, N.
5.3.5. [5-(3-Hydroxybenzylidene)-2,4-dioxothiazolidin-3-
yl]acetic acid methyl ester (3f). This was synthesized as
above described, adding methyl bromoacetate drop wise
during 1.5 h to a solution of compound 2f and potas-
5.3.10. [5-(4-Aminobenzylidene)-2,4-dioxothiazolidin-3-yl]-
acetic acid methyl ester (3i). The procedure was the same
as reported for compound 3h, starting from [5-(4-nitro-
benzylidene)-2,4-dioxothiazolidin-3-yl]acetic acid methyl
ester (3k). Yield 47%; mp 244 °C; 1H NMR (DMSO-d6):
d 3.68 (s, 3H, CH3); 4.45 (s, 2H, CH2); 6.23 (br s, 2H,
NH2), 6.66 (m, 2H, arom); 7.33 (m, 2H, arom); 7.76
(s, 1H, CH); 13C NMR (DMSO-d6): d 42.0 (CH2), 52.6
(CH3), 114.0, 133.1 (CH arom); 119.5 (5-C); 134.8,
150.6 (Cq arom); 135.4 (CH methylidene); 167.4,
167.7, 168.9 (CO); Anal. (C13H12N2O4S) C, H, N.
1
sium carbonate in acetone. Yield 86%; mp 198 °C; H
NMR (CDCl3): d 3.82 (s, 3H, CH3); 4.52 (s, 2H,
CH2); 6.94 (m, 1H, arom); 7.05 (m, 1H, arom); 7.28–
7.36 (m, 2H, arom); 7.83 (s, 1H, CH); 13C NMR
(CDCl3): d 41.9 (CH2); 53.0 (CH3); 116.7, 118.1, 122.7,
130.4 (CH arom); 121.8 (5-C); 134.7 (CH methylidene);
134.4, 156.6 (Cq arom); 165.6, 166.9, 167.2 (CO); Anal.
(C13H11NO5S) C, H, N.
5.3.6. [5-(4-Hydroxybenzylidene)-2,4-dioxothiazolidin-3-
yl]acetic acid methyl ester (3g). This was synthesized as
above described, adding methyl bromoacetate drop wise
during 1.5 h to a solution of compound 2g and potas-
5.4. General method for the synthesis of (5-arylidene-2,4-
dioxothiazolidin-3-yl)acetic acids 4b, d–g, j, k
1
sium carbonate in acetone. Yield 78%; mp 212 °C; H
A mixture of acetate 3 (10 mmol), glacial AcOH (40 mL)
and HCl 12 N (10 mL) was refluxed for 2 h. After evap-
oration in vacuo, the residue was refluxed again with
AcOH (40 mL) and HCl (10 mL) for 2 h. After evapora-
tion to dryness in vacuo, the crude solid was washed
with H2O and recrystallized from ethanol providing
pure carboxylic acid 4.
NMR (CDCl3): d 3.81 (s, 3H, CH3); 4.51 (s, 2H,
CH2); 6.93 (m, 2H, arom); 7.37 (m, 2H, arom); 7.82 (s,
1H, CH); 13C NMR (CDCl3): d 41.2 (CH2); 52.1
(CH3); 116.1, 132.0 (CH arom); 123.9 (5-C); 134.6
(CH methylidene); 118.8, 159.8 (Cq arom); 166.3,
167.0, 167.5 (CO); Anal. (C13H11NO5S) C, H, N.
5.3.7. [5-(3-Nitrobenzylidene)-2,4-dioxothiazolidin-3-yl]-
acetic acid methyl ester (3j). Yield 78%; mp 211 °C; H
5.4.1. [2,4-Dioxo-5-(4-phenoxybenzylidene)thiazolidin-3-
yl]acetic acid (4b). Yield 95%; mp 210 °C; 1H NMR
(DMSO-d6): d 4.38 (s, 2H, CH2); 7.16 (m, 4H, arom);
7.28 (m, 1H, arom); 7.49 (m, 2H, arom); 7.70 (m, 2H,
arom); 7.98 (s, 1H, CH); 13C NMR (DMSO-d6): d
42.4 (CH2); 118.9 (5-C); 118.3, 120.0, 124.8, 130.5,
132.7 (CH arom); 133.4 (CH methylidene); 127.5,
155.1, 159.4 (Cq arom); 165.2, 167.1, 168.1 (CO); Anal.
(C18H13NO5S) C, H, N.
1
NMR (DMSO-d6): d 3.81 (s, 3H, CH3); 4.53 (s, 2H,
CH2); 7.71 (dd, J = 8.1 and 7.8 Hz, 1H, arom); 7.83
(d, J = 7.8 Hz, 1H, arom); 7.97 (s, 1H, CH); 8.30
(dd, J = 8.1 and 3.0 Hz, 1H, arom); 8.39 (d,
J = 3.0 Hz, 1H, arom); 13C NMR (DMSO-d6): d 42.2
(CH2); 52.6 (CH3); 123.5 (5-C); 124.8, 124.9, 131.9,
135.4 (CH arom); 130.9 (CH methylidene); 134.4,
148.3 (Cq arom); 164.5, 166.2, 167.0 (CO); Anal.
(C13H10N2O6S) C, H, N.
5.4.2. (5-Biphenyl-4-ylmethylene-2,4-dioxothiazolidin-3-yl)-
1
acetic acid (4d). Yield 85%; mp 228–230 °C; H NMR
5.3.8. [5-(4-Nitrobenzylidene)-2,4-dioxothiazolidin-3-yl]-
acetic acid methyl ester (3k). Yield 82%; mp 255 °C;
1H NMR (CDCl3): d 3.82 (s, 3H, CH3); 4.53 (s, 2H,
CH2); 7.69 (m, 2H, arom); 7.97 (s, 1H, CH); 8.35 (m,
2H, arom); 13C NMR (CDCl3): d 42.1 (CH2); 52.9
(CH3); 121.9 (5-C); 124.4, 131.3 (CH arom); 130.6
(CH methylidene); 139.0, 148.2 (Cq arom); 164.7,
165.8, 167.3 (CO); Anal. (C13H10N2O6S) C, H, N.
(DMSO-d6): d 4.41 (s, 2H, CH2); 7.42–7.52 (m, 4H,
arom); 7.76 (m, 3H, arom); 7.88 (m, 2H, arom); 8.10
(s, 1H, CH); 13C NMR (DMSO-d6): d 42.3 (CH2);
120.4 (5-C); 126.8, 127.5, 128.3, 129.1, 131.0 (CH arom);
133.5 (CH methylidene); 131.8, 138.7, 142.2 (Cq arom);
165.0, 166.8, 168.0 (CO); Anal. (C18H13NO4S) C, H, N.
5.4.3. (5-Naphthalen-1-ylmethylene-2,4-dioxothiazolidin-
1
3-yl)acetic acid (4e). Yield 88%; mp 221 °C; H NMR
5.3.9. [5-(3-Aminobenzylidene)-2,4-dioxothiazolidin-3-yl]-
acetic acid methyl ester (3h). Tin(II) chloride (6.07 g,
32 mmol), was added to a solution of [5-(3-nitrobenzyl-
idene)-2,4-dioxothiazolidin-3-yl]acetic acid methyl ester
(3j) (2.06 g, 6.4 mmol) in ethanol and the mixture was
refluxed for 2 h. After evaporation to dryness in vacuo,
a saturated solution of sodium hydrogencarbonate was
added to the crude solid. The mixture was then extracted
with ethyl acetate, the organic layer was dried (Na2SO4)
and the solvent removed under reduced pressure. The
solid was recrystallized from methanol. Yield 52%; mp
240 °C; 1H NMR (DMSO-d6): d 3.38 (s, 3H, CH3);
4.49 (s, 2H, CH2); 5.39 (br s, 2H, NH2), 6.69–6.78 (m,
(DMSO-d6): d 4.44 (s, 2H, CH2); 7.67–7.76 (m, 4H,
arom); 8.11–8.14 (m, 3H, arom); 8.65 (s, 1H, CH); 13C
NMR (DMSO-d6): d 43.0 (CH2); 124.1, 126.3, 127.2,
127.6, 128.2, 129.5, 131.6 (CH arom); 125.1 (5-C);
131.8 (CH methylidene); 130.7, 131.8, 133.9 (Cq arom);
165.2, 168.0, 168.6 (CO); Anal. (C16H11NO4S) C, H, N.
5.4.4. [5-(3-Hydroxybenzylidene)-2,4-dioxothiazolidin-3-
yl]acetic acid (4f). Yield 80%; mp 225 °C; 1H NMR
(DMSO-d6): d 4.38 (s, 2H, CH2); 6.93 (d, J = 8.1 Hz,
1H, arom); 7.03 (s, 1H, arom); 7.09 (d, J = 7.8 Hz, 1H,
arom); 7.36 (dd, J = 8.1 and 7.8 Hz, 1H, arom);
7.90 (s, 1H, CH); 9.91 (br s, 1H, OH); 13C NMR

2820
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
(DMSO-d6): d 41.7 (CH2); 115.5, 117.6, 120.9, 129.9
(CH arom); 119.9 (5-C); 133.6 (CH methylidene);
133.4, 157.4 (Cq arom); 164.5, 166.4, 167.4 (CO); Anal.
(C12H9NO5S) C, H, N.
133.8 (CH methylidene); 125.5, 136.4, 160.4 (Cq arom);
165.1, 167.0, 167.9 (CO); Anal. (C19H15NO5S) C, H, N.
5.5. General method for the synthesis of methyl 4-(5-
arylidene-2,4-dioxothiazolidin-3-yl)-2-butenoic acids
methyl esters 5a–d
5.4.5. [5-(4-Hydroxybenzylidene)-2,4-dioxothiazolidin-3-
yl]acetic acid (4g). Yield 74%; mp 257 °C; 1H NMR
(DMSO-d6): d 4.36 (s, 2H, CH2); 6.97 (m, 2H, arom);
7.54 (m, 2H, arom); 7.90 (s, 1H, CH); 10.41 (br s, 1H,
OH); 13C NMR (DMSO-d6): d 42.1 (CH2); 116.4,
132.7 (CH arom); 116.1, 160.3 (Cq arom); 123.7 (5-C);
134.3 (CH methylidene); 165.2, 167.0, 167.9 (CO); Anal.
(C12H9NO5S) C, H, N.
A
mixture of 5-arylidene-2,4-thiazolidinedione
2
(10 mmol), methyl 4-bromocrotonate (3.58 g, 20 mmol)
and potassium carbonate (2.76 g, 20 mmol) in acetone
(130 mL) was refluxed for 24 h. After cooling, the inor-
ganic salts were filtered off; then the solvent was evapo-
rated under reduced pressure and the solid residue was
recrystallized from ethanol providing pure 2-butenoate 5.
5.4.6. [5-(3-Nitrobenzylidene)-2,4-dioxothiazolidin-3-yl]-
acetic acid (4j). Yield 80%; mp 235 °C; 1H NMR
(DMSO-d6): d 4.41 (s, 2H, CH2); 7.84 (dd, J = 9.0 and
9.0 Hz, 1H, arom); 8.06 (d, J = 9.0 Hz, 1H, arom);
8.17 (s, 1H, CH); 8.32 (dd, J = 9.0 and 3.0 Hz, 1H,
arom); 8.51 (s, 1H, arom); 13C NMR (DMSO-d6): d
42.4 (CH2); 123.6 (5-C); 124.7, 124.8, 131.6, 135.4 (CH
arom); 130.9 (CH methylidene); 134.4, 148.2 (Cq arom);
164.6, 166.3, 167.8 (CO); Anal. (C12H8N2O6S) C, H, N.
5.5.1. 4-[2,4-Dioxo-5-(3-phenoxybenzylidene)thiazolidin-
3-yl]-2-butenoic acid methyl ester (5a). Yield 75%; mp
107–110 °C; H NMR (CDCl3): d 3.74 (s, 3H, CH3);
1
4.48 (dd, J = 6.0 and 1.5 Hz, 2H, CH2); 5.98 (dt,
J = 15.0 and 1.5 Hz, 1H, CH); 6.85 (dt, J = 15.0 and
6.0 Hz, 1H, CH); 7.05–7.12 (m, 3H, arom); 7.23–7.31
(m, 3H, arom); 7.41–7.48 (m, 3H, arom); 7.86 (s, 1H,
CH methylidene); 13C NMR (CDCl3): d 41.8 (CH2);
51.8 (CH3); 119.2, 119.6, 120.6, 124.2, 124.6, 130.0,
134.0 (CH arom); 121.7 (5-C); 123.6, 130.6 (CH@CH);
139.3 (CH methylidene); 134.6, 156.0, 158.3 (Cq
arom); 165.4, 165.8, 166.5 (CO); Anal. (C21H17NO5S)
C, H, N.
5.4.7. [5-(4-Nitrobenzylidene)-2,4-dioxothiazolidin-3-yl]-
acetic acid (4k). Yield 93%; mp 271 °C; 1H NMR
(DMSO-d6): d 4.44 (s, 2H, CH2); 7.93 (m, 2H, arom);
8.10 (s, 1H, CH); 8.37 (m, 2H, arom); 13C NMR
(DMSO-d6): d 42.4 (CH2); 124.1 (5-C); 124.2, 131.0
(CH arom); 131.3 (CH methylidene); 138.9, 147.7 (Cq
arom); 164.6, 166.3, 167.6 (CO); Anal. (C12H8N2O6S)
C, H, N.
5.5.2. 4-[2,4-Dioxo-5-(4-phenoxybenzylidene)thiazolidin-
3-yl]-2-butenoic acid methyl ester (5b). Yield 72%; mp
1
103–105 °C; H NMR (CDCl3): d 3.74 (s, 3H, CH3);
4.50 (dd, J = 6.0 and 1.5 Hz, 2H, CH2); 5.91 (dt,
J = 15.0 and 1.5 Hz, 1H, CH); 6.81 (dt, J = 15.0 and
6.0 Hz, 1H, CH); 7.05–7.11 (m, 3H, arom); 7.22–7.26
(m, 3H, arom); 7.42–7.50 (m, 3H, arom); 7.90 (s, 1H,
CH methylidene); 13C NMR (CDCl3): d 41.7 (CH2);
51.7 (CH3); 118.2, 120.1, 124.6, 130.0, 134.0 (CH arom);
118.9 (5-C); 123.5, 132.2 (CH@CH); 139.4 (CH methy-
lidene); 127.3, 155.3, 160.0 (Cq arom); 165.6, 165.8,
167.2 (CO); Anal. (C21H17NO5S) C, H, N.
5.4.8. [5-(4-Benzyloxybenzylidene)-2,4-dioxothiazolidin-
3-yl]acetic acid (4c). A mixture of 2,4-thiazolidinedione
(1 g, 8.5 mmol), methyl bromoacetate (2.6 g, 17 mmol)
and potassium carbonate (2.35 g, 17 mmol) was refluxed
for 24 h; then the solvent was evaporated under reduced
pressure. The residue was washed with methanol to pro-
vide (2,4-dioxothiazolidin-3-yl)acetic acid methyl ester
1
(7) as oil (1.29 g, yield 80%); H NMR (CDCl3): d 3.77
(s, 3H, CH3); 4.05 (s, 2H, 5-CH2); 4.36 (s, 2H, N–
CH2). A mixture of 7 (0.8 g, 4.23 mmol) glacial AcOH
(17 mL) and HCl 2 N (4.2 mL) was refluxed for 2 h.
After evaporation in vacuo, the crude mixture was re-
fluxed again with AcOH (17 mL) and HCl (4.2 mL).
After evaporation to dryness in vacuo, the crude oil
was washed with H2O and then with ethanol to provide
pure (2,4-dioxothiazolidin-3-yl)acetic acid (8) as oil
5.5.3. 4-[5-(4-Benzyloxybenzylidene)-2,4-dioxothiazoli-
din-3-yl]-2-butenoic acid methyl ester (5c). Yield 98%;
mp 120–122 °C; H NMR (DMSO-d6): d 3.44 (s, 3H,
1
CH3); 4.41 (dd, J = 6.0 and 3.0 Hz, 2H, CH2); 5.15 (s,
2H, CH2O); 5.87 (dt, J = 15.0 and 3.0 Hz, 1H, CH);
6.82 (dt, J = 15.0 and 6.0 Hz, 1H, CH); 7.15 (m, 2H,
arom); 7.34–7.44 (m, 5H, arom); 7.57 (m, 2H, arom);
7.85 (s, 1H, CH methylidene); 13C NMR (DMSO-d6):
d 42.2 (CH2); 51.8 (CH3); 69.7 (CH2O); 116.0, 126.0,
128.0, 128.3, 128.7 (CH arom); 118.5 (5-C); 121.8,
133.1 (CH@CH); 132.5 (CH methylidene); 125.8,
136.3, 160.8 (Cq arom); 166.2, 167.5, 168.5 (CO); Anal.
(C22H19NO5S) C, H, N.
1
(0.44 g, yield 60%); H NMR (CDCl3): d 4.06 (s, 2H,
5-CH2); 4.42 (s, 2H, N–CH2). Piperidine (0.17 g,
2.03 mmol) and 4-benzyloxybenzaldehyde (0.54 g,
2.54 mmol) were added to a solution of 8 (0.44 g,
2.54 mmol) in ethanol and the mixture was refluxed
for 24 h. The reaction mixture was then poured into
H2O and acidified with AcOH to give 4c as solid, which
was recrystallized from methanol. Yield 41%; mp
262 °C; 1H NMR (DMSO-d6): d 4.40 (s, 2H, CH2);
5.23 (s, 2H, CH2O); 7.22 (m, 2H, arom); 7.39–7.51 (m,
5H, arom); 7.65 (m, 2H, arom); 7.99 (s, 1H, CH); 13C
NMR (DMSO-d6): d 42.1 (CH2); 69.5 (CH2O); 117.5
(5-C); 115.8, 127.8, 128.0, 128.5, 132.4 (CH arom);
5.5.4. 4-(5-Biphenyl-4-ylmethylene-2,4-dioxothiazolidin-
3-yl)-2-butenoic acid methyl ester (5d). Yield 90%; mp
1
158–160 °C; H NMR (CDCl3): d 3.75 (s, 3H, CH3);
4.52 (dd, J = 5.7 and 2.1 Hz, 2H, CH2); 6.00 (dt,
J = 16.2 and 2.1 Hz, 1H, CH); 6.88 (dt, J = 16.2 and
5.7 Hz, 1H, CH); 7.42–7.52 (m, 3H, arom); 7.60–7.66
(m, 4H, arom); 7.75 (m, 2H, arom); 7.98 (s, 1H, CH

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2821
methylidene); 13C NMR (CDCl3): d 41.8 (CH2); 51.8
(CH3); 120.4 (5-C); 127.1, 127.8, 128.2, 129.0, 134.2
(CH arom); 123.6, 130.8 (CH@CH); 139.4 (CH methy-
lidene); 128.9, 131.8, 143.4 (Cq arom); 165.6, 166.9,
167.3 (CO); Anal. (C21H17NO4S) C, H, N.
(CH@CH); 124.6 (5-C); 140.4 (CH methylidene);
117.3, 160.9 (Cq arom); 166.1, 167.0, 167.9 (CO); Anal.
(C14H11NO5S) C, H, N.
5.7. Molecular modelling
5.6. General method for the synthesis of 4-(5-arylidene-
2,4-dioxothiazolidin-3-yl)-2-butenoic acids 6a, b, d, g
All calculations and graphic manipulations were per-
formed on a SGI Octane double processor R10K work-
station by means of the Catalyst 4.7 software package30
for pharmacophore model building and the Cerius2 4.9
software package29 for docking. All the compounds
used in this study were built using the 2D–3D sketcher
of the program Catalyst. A representative family of con-
formations was generated for each molecule using the
poling algorithm and the Ôbest quality conformational
analysisÕ method. The parameter set was employed to
perform all the conformational calculations derived
from the CHARMm force field. Conformational diver-
sity was emphasized by selection of the conformers that
fell within 20 kcal/mol range above the lowest energy
conformation found.
A mixture of methyl ester 5 (10 mmol), glacial AcOH
(40 mL) and HCl 12 N (10 mL) was refluxed for 2 h.
After evaporation to dryness in vacuo, the crude solid
was washed with H2O and recrystallized from methanol
providing pure carboxylic acid 6.
5.6.1. 4-[2,4-Dioxo-5-(3-phenoxybenzylidene)thiazolidin-
3-yl]-2-butenoic acid (6a). Yield 92%; mp 155–157 °C;
1H NMR (DMSO-d6): d 4.40 (dd, J = 5.1 and 2.0 Hz,
2H, CH2); 5.81 (dt, J = 15.6 and 2.0 Hz, 1H, CH);
6.75 (dt, J = 15.6 and 5.1 Hz, 1H, CH); 7.05–7.21 (m,
3H, arom); 7.35–7.44 (m, 4H, arom); 7.50–7.55 (m,
2H, arom); 7.90 (s, 1H, CH methylidene); 13C NMR
(DMSO-d6): d 41.8 (CH2); 118.9, 119.2, 120.3, 124.1,
124.8, 130.2, 140.4 (CH arom); 122.3 (5-C); 123.0,
132.4 (CH@CH); 131.0 (CH methylidene); 134.8,
155.7, 157.6 (Cq arom); 165.0, 166.3, 166.7 (CO); Anal.
(C20H15NO5S) C, H, N.
The docking approach was carried out with the tool
LigandFit implemented in the Cerius2 software pack-
age.29 Calculations were performed using the force field
cff 1.01. The site definition was derived from the bound
˚
ligand including the unoccupied grid points within 2.5 A
˚
from heavy atoms and within 2 A from hydrogen atoms
5.6.2. 4-[2,4-Dioxo-5-(4-phenoxybenzylidene)thiazolidin-
3-yl]-2-butenoic acid (6b). Yield 50%; mp 190–192 °C;
1H NMR (CDCl3): d 4.52 (dd, J = 5.1 and 2.0 Hz, 2H,
CH2); 5.94 (dt, J = 15.9 and 2.0 Hz, 1H, CH); 6.98 (dt,
J = 15.9 and 5.1 Hz, 1H, CH); 7.00–7.05 (m, 3H, arom);
7.15 (m, 2H, arom); 7.29–7.33 (m, 4H, arom); 7.90 (s,
1H, CH methylidene); 13C NMR (CDCl3): d 41.3
(CH2); 117.8, 119.6, 124.2, 129.6, 133.4 (CH arom);
118.5 (5-C); 124.0, 131.8 (CH@CH); 138.5 (CH methy-
lidene); 126.9, 154.8, 159.5 (Cq arom); 165.0, 165.1,
166.6 (CO); Anal. (C20H15NO5S) C, H, N.
of the ligand. The compounds were rigidly docked keep-
ing all parameters of interaction energy minimization,
rigid body minimization and site partition at their de-
fault values. A final energy minimization of the ligands
in the protein followed the fitting process. Prioritization
of the docked ligand orientations was achieved employ-
ing all seven scoring functions in Cerius2 (Ligscore1,
Ligscore2, PLP1, PLP2, JAIN, PMF and LUDI) using
the predefined settings. Finally the scores resulting from
these functions were combined through the computation
of a consensus score value. The investigation of the best
100 poses (according to their consensus score) resulted
in 19 different orientations of the selected ARIs accord-
ing to the degree of superimposition onto Idd384.
5.6.3. 4-(5-Biphenyl-4-ylmethylene-2,4-dioxothiazolidin-
3-yl)-2-butenoic acid (6d). Yield 45%; mp 248–250 °C;
1H NMR (DMSO-d6): d 4.45 (dd, J = 6.0 and 2.0 Hz,
2H, CH2); 5.85 (dt, J = 15.0 and 2.0 Hz, 1H, CH);
6.80 (dt, J = 15.0 and 6.0 Hz, 1H, CH); 7.46–7.56 (m,
5H, arom); 7.76 (m, 2H, arom); 7.88 (m, 2H, arom);
8.01 (s, 1H, CH methylidene); 13C NMR (DMSO-d6):
d 41.9 (CH2); 121.0 (5-C); 126.8, 127.5, 128.3, 129.1,
132.7 (CH arom); 123.0, 130.9 (CH@CH); 140.7 (CH
methylidene); 132.0, 138.7, 142.0 (Cq arom); 165.2,
166.4, 167.0 (CO); Anal. (C20H15NO4S) C, H, N.
5.8. Enzyme section
NADPH, DL-glyceraldehyde and dithiothreitol (DDT)
were purchased from Sigma Chemical Co. DEAE-cellu-
lose (DE-52) was obtained from Whatman. Sorbinil was
a gift from Prof. L. Costantino, University of Modena
(Italy). All other chemicals were commercial samples
of good grade. Calf lenses for the purification of
ALR2 were obtained locally from freshly slaughtered
animals. The enzyme was purified by a chromatographic
procedure as previously described.31 Briefly, ALR2 was
released by carving the capsule and the frozen lenses
were suspended in potassium phosphate buffer pH 7
containing 5 mM DTT and stirred in an ice-cold bath
for 2 h. The suspension was centrifuged at 4000 rpm at
4 °C for 30 min and the supernatant was subjected to
ion exchange chromatography on DE52. Enzyme activ-
ity was assayed spectrophotometrically on a Cecil Super
Aurius CE 3041 spectrophotometer by measuring the
decrease in absorption of NADPH at 340 nm, which
5.6.4. 4-[5-(4-Hydroxybenzylidene)-2,4-dioxothiazolidin-
3-yl]-2-butenoic acid (6g). This was synthesized as above
described, starting from 4-[5-(4-benzyloxybenzylidene)-
2,4-dioxothiazolidin-3-yl]-2-butenoic acid methyl ester
1
(5c). Yield 40%; mp 292–295 °C; H NMR (DMSO-d6):
d 4.42 (dd, J = 5.8 and 1.2 Hz, 2H, CH2); 5.80 (dt,
J = 15.9 and 1.2 Hz, 1H, CH); 6.79 (dt, J = 15.9 and
5.8 Hz, 1H, CH); 6.93 (m, 2H, arom); 7.51 (m, 2H,
arom); 7.87 (s, 1H, CH methylidene); 10.40 (br s, 1H,
OH); 12.40 (br s, 1H, COOH); 13C NMR (DMSO-d6):
d 42.4 (CH2); 117.2, 133.4 (CH arom); 123.6, 134.5

2822
R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
accompanies the oxidation of b-NADPH catalyzed by
ALR2. The assay was performed at 37 °C in a reaction
mixture containing 0.25 M sodium phosphate buffer
pH 6.8, 0.38 M ammonium sulfate, 0.11 mM NADPH
and 4.7 mM DL-glyceraldehyde as substrate in a final
volume of 1.5 mL. All inhibitors were dissolved in
DMSO. The final concentration of DMSO in the reac-
tion mixture was 1%. To correct for the nonenzymatic
oxidation of NADPH, the rate of NADPH oxidation
in the presence of all of the reaction mixture components
except the substrate was subtracted from each experi-
mental rate. Each dose-effect curve was generated using
at least three concentrations of inhibitor causing an inhi-
bition between 20% and 80%. Each concentration was
tested in duplicate, and IC50 values as well as the 95%
confidence limits (95% CL) were obtained by using Cal-
cuSyn software for dose effect analysis.32
Compd
C (%)
H (%)
N (%)
3k
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
48.45
48.17
60.84
60.58
61.78
61.58
63.71
63.49
61.33
61.00
51.61
51.39
51.61
51.80
46.76
46.91
46.76
46.58
63.79
63.51
63.79
63.42
64.53
64.36
66.48
66.09
62.98
63.11
62.98
62.76
65.74
65.95
55.08
54.69
3.13
3.28
3.69
3.82
4.09
4.33
3.86
3.94
3.54
3.78
3.25
3.58
3.25
3.01
2.62
2.84
2.62
2.90
4.33
4.47
4.33
4.51
4.68
4.88
4.52
4.39
3.96
3.90
3.96
4.01
4.14
4.00
3.63
3.75
8.69
8.57
3.94
3.71
3.79
3.68
4.13
4.02
4.47
4.28
5.02
4.86
5.02
5.22
9.09
8.90
9.09
8.98
3.54
3.32
3.54
3.39
3.42
3.21
3.69
3.88
3.67
3.75
3.67
3.42
3.83
3.68
4.59
4.32
4b
4c
4d
4e
4f
4g
4j
Appendix A. Experimental analyses
4k
5a
5b
5c
5d
6a
6b
6d
6g
Compd
C (%)
H (%)
N (%)
2c
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
Calcd
Found
65.58
65.55
68.31
68.13
65.87
65.54
54.29
54.02
54.29
54.46
54.53
54.28
54.53
54.39
48.00
47.78
48.00
48.34
61.78
61.47
62.65
62.98
64.58
64.25
62.37
62.21
53.24
53.14
53.24
53.00
53.42
53.29
53.42
53.48
48.45
48.16
4.21
4.48
3.94
4.14
3.55
3.71
3.19
3.36
3.19
3.11
3.66
3.91
3.66
3.87
2.42
2.65
2.42
2.58
4.09
4.19
4.47
4.61
4.28
4.54
4.00
3.83
3.78
3.97
3.78
4.05
4.14
4.28
4.14
4.16
3.13
3.32
4.50
4.17
4.98
4.69
5.49
5.29
6.33
6.12
6.33
6.07
12.72
12.45
12.72
12.51
11.19
11.00
11.19
10.95
3.79
3.63
3.65
3.46
3.96
3.72
4.28
4.35
4.78
4.54
4.78
4.58
9.58
9.27
9.58
9.36
8.69
8.46
2d
2e
2f
2g
2h
2i
2j
2k
3b
3c
3d
3e
3f
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/j.bmc.
2005.02.026.
References and notes
1. (a) Kador, P. F.; Kinoshita, J. H.; Sharpless, N. E. J. Med.
Chem. 1985, 28, 841; (b)Kador, P. F. Med. Res. Rev. 1988, 8,
325; (c) Yabe-Nishimura, C. Pharmacol. Rev. 1998, 20, 21.
2. Costantino, L.; Rastelli, G.; Cignarella, G.; Vianello, P.;
Barlocco, D. Exp. Opin. Ther. Patents 1997, 7, 843.
3. Wolffenbuttel, B. H.; van Haeften, T. W. Drugs 1995, 50,
263.
4. Porte, D., Jr.; Schwartz, M. W. Science 1996, 272, 699.
5. Suzen, S.; Buyukbingol, E. Curr. Med. Chem. 2003, 10,
1329.
3g
3h
3i
3j
6. Larson, E. R.; Lipinski, C. A.; Sarges, R. Med. Res. Rev.
1988, 8, 159.

R. Maccari et al. / Bioorg. Med. Chem. 13 (2005) 2809–2823
2823
7. Sarges, R. In Advances in Drug Research; Testa, B., Ed.;
Academic: London, 1989; Vol. 18, pp 139–175.
8. Costantino, L.; Rastelli, G.; Vianello, P.; Cignarella, G.;
Barlocco, D. Med. Res. Rev. 1999, 19, 3.
9. Costantino, L.; Rastelli, G.; Gamberoni, M. C.; Barlocco,
D. Exp. Opin. Ther. Patents 2000, 10, 1245.
10. Castan˜er, J.; Prous, J. Drugs Future 1987, 12, 336.
11. Kador, P. F.; Sharpless, N. E. Mol. Pharmacol. 1983, 24,
521.
21. Lipinski, C. A.; Fiese, E. F.; Korst, R. J. Quant. Struct.-
Act. Relat. 1991, 10, 109.
22. Kador, P. F.; Lee, Y. S.; Rodriguez, L.; Sato, S.;
Bartoszko-Malik, A.; Abdel-Ghany, Y. S.; Miller, D. D.
Bioorg. Med. Chem. 1995, 3, 1313.
23. (a) Sugimoto, A.; Matsui, S.; Taira, M. JP08109175;
Chem. Abstr. 1996, 125, 86632; (b) Sugimoto, A.; Matsui,
S.; Taira, M.; Ushijima, T. JP08109174; Chem. Abstr.
1996, 125, 86633.
12. Lee, Y. S.; Pearlstein, R.; Kador, P. F. J. Med. Chem.
1994, 37, 787.
24. Momose, Y.; Meguro, K.; Ikeda, H.; Hatanaka, C.; Oi, S.;
Sohda, T. Chem. Pharm. Bull. 1991, 39, 1440.
25. Prous, J. R. Drugs Future 1987, 12, 347.
26. Fresneau, P.; Cussac, M.; Morand, J.; Szymonski, B.;
Tranqui, D.; Leclerc, G. J. Med. Chem. 1998, 41,
4706.
27. Calderone, V.; Chevrier, B.; Van Zandt, M.; Lamour, V.;
Howard, E.; Poterzman, A.; Barth, P.; Mitschler, A.; Lu,
J.; Dvornik, D. M.; Klebe, G.; Kraemer, O.; Moorman, A.
R.; Moras, D.; Podjarny, A. Acta Crystallogr., Sect. D
2000, 56, 536.
13. Urzhumtsev, A.; Tete-Favier, F.; Mitschler, A.; Barban-
ton, J.; Barth, P.; Urzhumtseva, L.; Biellmann, J. F.;
Podjarny, A.; Moras, D. Structure 1997, 5, 601.
14. El-Kabbani, O.; Wilson, D. K.; Petrash, J. M.; Quiocho,
F. A. Mol. Vis. 1998, 4, 19.
15. Lee, Y. S.; Chen, Z.; Kador, P. F. Bioorg. Med. Chem.
1998, 6, 1811.
16. Hohman, T. C.; El-Kabbani, O.; Malamas, M. S.; Lai, K.;
Putilina, T.; McGowan, M. H.; Wane, Y. Q.; Carper, D.
A. Eur. J. Biochem. 1998, 256, 310.
28. Wilson, D. K.; Tarle, I.; Petrash, J. M.; Quiocho, F. A.
Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 9847.
29. Cerius2, release 4.9, Accelrys Inc, San Diego, CA, USA,
2003.
17. Singh, S. B.; Malamas, M. S.; Hohman, T. C.; Nilakantan,
R.; Carper, D. A.; Kitchen, D. J. Med. Chem. 2000, 43,
1062.
18. Bruno, G.; Costantino, L.; Curinga, C.; Maccari, R.;
30. Catalyst, release 4.7, Accelrys Inc, San Diego, CA, USA,
2003.
`
Monforte, F.; Nicolo, F.; Ottana, R.; Vigorita, M. G.
Bioorg. Med. Chem. 2002, 10, 1077.
`
31. Costantino, L.; Rastelli, G.; Vescovini, K.; Cignarella, G.;
Vianello, P.; Del Corso, A.; Cappiello, M.; Mura, U.;
Barlocco, D. J. Med. Chem. 1996, 39, 4396.
32. Chou, T.-C.; Hayball, M. P. CalcuSyn software version
1.1.1., Biosoft, Cambridge, UK, 1996.
19. Lee, Y. S.; Hodoscek, M.; Kador, P. F.; Sugiyama, K.
Chem. Biol. Interact. 2003, 143–144, 307.
20. Zask, A.; Jirkovsky, I.; Nowicki, J. W.; McCaleb, M. L. J.
Med. Chem. 1990, 33, 1418.