In silico scaffold evaluation and solid phase approach to identify new gelatinase inhibitors

Article abstract of DOI:10.1016/j.bmc.2012.02.010

Among matrix metalloproteinases (MMPs), gelatinases MMP-2 (gelatinase A) and MMP-9 (gelatinase B) play a key role in a number of physiological processes such as tissue repair and fibrosis. Many evidences point out their involvement in a series of patholog

Full text of DOI:10.1016/j.bmc.2012.02.010

Bioorganic & Medicinal Chemistry 20 (2012) 2323–2337
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
In silico scaffold evaluation and solid phase approach to identify new
gelatinase inhibitors
Alessandra Topai a, , Perla Breccia b, Franco Minissi a, Alessandro Padova c, Stefano Marini d,
⇑
Ilaria Cerbara a
a Colosseum Combinatorial Chemistry Centre for Technology (C4T S.C.a r.l.), Via della Ricerca Scientifica s.n.c., I-00133 Rome, Italy
b BioFocus, Chesterford Research Park, UK-CB10 1XL Saffron Walden, United Kingdom
c Molecular Informatics Dept, Siena Biotech S.p.A., Strada del Petriccio e Belriguardo, 35, I-53100 Siena, Italy
d Department of Experimental Medicine and Biochemical Sciences, University of Roma Tor Vergata, Via Montpellier 1, I-00133 Roma, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Among matrix metalloproteinases (MMPs), gelatinases MMP-2 (gelatinase A) and MMP-9 (gelatinase B)
play a key role in a number of physiological processes such as tissue repair and fibrosis. Many evidences
point out their involvement in a series of pathological events, such as arthritis, multiple sclerosis, cardio-
vascular diseases, inflammatory processes and tumor progression by degradation of the extracellular
matrix. To date, the identification of non-specific MMP inhibitors has made difficult the selective target-
ing of gelatinases. In this work we report the identification, design and synthesis of new gelatinase inhib-
itors with appropriate drug-like properties and good profile in terms of affinity and selectivity. By a
detailed in silico protocol and innovative and versatile solid phase approaches, a series of 4-thiazolydi-
nyl-N-hydroxycarboxyamide derivatives were identified. In particular, compounds 9a and 10a showed
a potent inhibitory activity against gelatinase B and good selectivity over the other MMP considered in
this study. The identified compounds could represent novel potential candidates as therapeutic agents.
Ó 2012 Elsevier Ltd. All rights reserved.
Received 13 October 2011
Revised 2 February 2012
Accepted 3 February 2012
Available online 13 February 2012
Keywords:
Matrix metalloproteinases
Gelatinase inhibitors
Solid phase synthesis
Pyroglutammate
Thioproline
1. Introduction
inhibition of gelatinases may provide confidence for a successful
clinical trial, targeting acute cardiovascular diseases and tumor
Matrix metalloproteinases (MMPs) are a family of approximately
27 zinc-dependent endopeptidases, involved in the proteolytic
processing of several components of the extracellular matrix (ECM),
such as collagen, proteoglycans and fibronectin.
In the last decades, considerable evidences have highlighted the
involvement of abnormal MMP expression in several pathological
conditions. In particular, MMP-2 and MMP-9, main enzymes able
to degrade type IV collagen and gelatin, have been demonstrated
to be directly involved in endothelial cell migration and vascular
remodeling during angiogenesis, in tumor cell growth and progres-
sion during metastasis.
In the last twenty years, the blockage of the extracellular matrix
degrading activities of MMPs has been considered an attractive
strategy to treat a broad spectrum of diseases including cancer,
neurological diseases, arthritis and cardiovascular pathologies.1–3
Moreover, although several potent MMPIs have been discov-
ered,4–8 the serious dose-limiting side effects occurring in clinical
trials,9–11 probably due to the non-selective inhibitory activity,
have underlined the importance to identify novel inhibitors able
to discriminate among MMPs.12,13 In this scenario, the selective
progression.14–16 In particular MMP-2 is considered to be a better
candidate target for anticancer drugs than MMP-9, whose inhibi-
tion is related to greater drug side effects.17,18
In this paper we describe the identification of novel and selec-
tive gelatinase inhibitors by an in silico rational drug design ap-
proach and a versatile solid phase synthetic strategy. Starting
from the MMP-2 complexes available in the Protein Data Bank
(PDB),19 an active site mapping of gelatinase A was performed to
generate the pharmacophore model used for the in silico screening
of our in-house small molecule database. The screening was then
followed by rational compound selection based on criteria, such
as physico-chemical properties, ADMET predictions and synthetic
feasibility. The computational approach led to the identification
of the pyroglutamic moiety (Fig. 1) as a promising scaffold for the
design and optimization of novel gelatinase inhibitors. By rational
structure optimization, 4-thiazolydinyl-N-hydroxycarboxyamide
⇑
Corresponding author. Tel.: +39 0672594029; fax: +39 0672594031.
E-mail address: alessandra.topai@uniroma2.it (A. Topai).
Figure 1. The pyroglutamic scaffold.
0968-0896/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2012.02.010

2324
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
derivatives were identified as novel potential class of gelatinase
inhibitors.
2. Results and discussion
2.1. MMP-2 binding site mapping and pharmacophore
generation
The design of new gelatinase inhibitors started with the recog-
nition of the representative examples of gelatinase complexes in
the PDB. Focusing on MMP-2 structures in complex with inhibitors
of known potency, the 11 NMR solution structures of MMP-2 cat-
alytic domain (cd) (pdb code: 1hov) complexed with SC-74020
(MMP-2 IC50 less than 0.1 nM) and the crystal structure of
cdMMP-2 (pdb code: 1qib) were selected for this work.
Overlay studies were carried out in order to identify inhibitor
key features, required for protein recognition and thus significant
for the pharmacophore model generation. The alignment, per-
formed with a combinatorial extension (CE) algorithm,20 showed
a good overlay between the 11 NMR complexes and X-ray struc-
ture (RMSD in the range of 1.6–1.9 Å (Fig. 2a). Moreover, no signif-
icant difference in the binding mode of SC-74020, retrieved from
the NMR complex with the lower RMSD with respect to 1qib struc-
ture (structure No. 7), was observed when the ligand was docked
into the catalytic site of MMP-2 X-ray structure. As shown in Figure
2b, the hydroxamate zinc binding group (ZBG) is coordinated to
Zn2+ ion, together with the catalytic histidine, while the P01 group
is involved in hydrophobic interactions with the amino acids of
S01 subsite, comprising a structurally stable hydrophobic pocket
able to bind drug molecules at that position. According to the ob-
served results, the more accurate MMP-2 X-ray structure was se-
lected for further docking simulations.
Figure 3. (a) Key pharmacophoric features defined on SC-74020 NMR conformation
(pdb code: 1hov): F1 identified the zinc binding group, and the F2 corresponded to
the P01 aromatic ring group, labeled as yellow and green spheres, respectively. (b)
The excluded volume defined by the protein residues with at least one atom within
4.5 Å from any atom of SC-74020.
deeper modifications, no further features were included to de-
scribe the long P01 group.
2.2. 2D- and 3D-database generation
By using MOE package software21 (see Section 5 for details), the
selected SC-74020 NMR conformation was used to develop a phar-
macophore model to mine novel and potential gelatinase inhibi-
tors from our in-house virtual database, named MoDa. As shown
in Figure 3, we considered two key pharmacophoric features: the
ZBG and the P01 moiety, which represent the two known binding
and stabilizing interactions with the molecular target. In order to
identify new scaffolds to be used for further investigations and
About 2 million unique commercially available small molecules
have been collected in a proprietary 2D small molecule database,
namely MoDa (Molecular Database). By using MOE, drug-like fil-
ters were applied to remove non drug-like, toxic or reactive mole-
cules.22–24 Finally a number of around 600,000 unique structures
were selected and for each molecule the 3D coordinates were de-
rived. After energy minimization, for each 3D structure, a set of
representative conformers was generated by molecular dynamic
simulations. Low-energy conformations (65 kcal/mol from the
global minimum) were included in the conformer database with
a limit of 70 conformers per molecule, for the most flexible
structures.
2.3. Pharmacophore based virtual screening
The derived 3D database was screened in MOE against the phar-
macophore model. Hit compounds were ranked according to the
score value defined by the RMSD of the aligned pharmacophoric
points onto the pharmacophore model: no overlap on the excluded
volume was permitted and, finally, about 200,000 molecules were
identified. In order to select a more restricted number of com-
pounds to be assayed against gelatinases, the selection was limited
to a set of compounds bearing classical zinc binding groups (ZBG)
such as carboxylic and hydroxamic acids. A molecular weight limit
of 300 Da was imposed and a set of 150 compounds was obtained.
Virtual hits were then visually inspected and selected following
two criteria: synthetic feasibility and easiness to perform chemical
modifications. Among the hit list, 15 compounds were selected,
purchased, and assayed. The virtual screening flowchart is summa-
rized in Figure 4, while the most representative compounds are de-
picted in Figure 5.
Figure 2. (a) Superimposition of NMR complex and X-ray MMP-2 structures,
represented by blue and green cartoons, respectively. S01 pocket is shown in pink
while zinc ion is rendered as a light-gray sphere. The figure shows the NMR solution
(No. 7 of 1hov.pdb) with lower RMSD compared to the MMP-2 X-ray coordinates,
with SC-74020 rendered in yellow ball and stick figures. (b) Docking result of SC-
74020 into the binding site of MMP-2 X-ray: the best scored pose and the NMR
conformation are labeled as gray and yellow ball and stick figures, respectively.
Among the screened compounds, the pyroglutamic acid deriva-
tive (S)-10b showed an in vitro MMP-2 inhibitory activity in the
micromolar range (IC50 = 315 lM), prompting us to consider it

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2325
Figure 4. Virtual screening pharmacophore based flow-chart.
O
OH
O
O
N
OH
S
OH
O
S
O
N
Cl
O
HO
O
O
Cl
N
O
OH
NH2
N
S
N
Cl
O
OH
N
HO
O
N
O
O
(S)-1'b
Figure 5. Selected hits from the in silico screening of MoDa database.
as a promising chemical scaffold for a new class of gelatinase
inhibitors.
moiety, keeping the ZBG and P01 groups unchanged (Fig. 6) and eval-
uating chemical modifications on the five-membered ring and the
methylene group. The pyroglutamate ring was replaced with com-
mercially available proline (2a), thiazolidine (3a) and hydroxypro-
line (4a). Thiazolidinone derivative 5a was synthesized to
investigate the influence of both the carbonyl group and the sulfur
atom on the ring. Moreover, we synthesized the sulfonamide deriv-
atives of pyroglutamate (6a), proline (7a), and thiazolidine (8a)
since several literature data reported aryl-sulfonamide substituents
to be the most favorable ones.25
All novel compounds were synthesized (Scheme 3–6) and
tested against gelatinases (Table 2). SAR analysis suggested that
the introduction of a polar group, such as a carbonyl moiety at Y,
maintaining a methylene group at Z position, strongly increases
the inhibitory potency (i.e., 1a vs 2a; 5a vs 3a). Nevertheless, when
the methylene group in Z is replaced with the SO2 function, the
influence of the carbonyl moiety in Y is negligible (i.e., 6a vs 7a).
In contrast, the SO2 group in Z enhances the positive effect of the
substitution in the X position with a sulfur atom (compare com-
pounds 2a/3a, 1a/5a and 7a/8a). We identified compound 8a as
lead compound, with an IC50 value of 200 nM and of 230 nM to-
wards MMP-2 and MMP-9, respectively. In order to get a more de-
tailed representation of the recognition features, the compound
was docked into the binding site of MMP-2 crystal structure.
2.4. Chemical modification on selected scaffold
A first series of chemical modifications were performed on (S)-
10b template in order to investigate the influence on gelatinase
inhibitory activity of (1) halogen position on the aromatic ring,
(2) benzene ring replacement with a well-fitting P01 substituent
(biphenyl moiety), (3) introduction of a non-carboxylic acid ZBG
(hydroxamic acid or amide group) and (4) chirality. These com-
pounds were synthesized according to Schemes 1 and 2 and bio-
logically screened against MMP-2 and MMP-9 (Table 1).
The biological data confirmed the expected results in terms of
best ZBG and P01 residues, while no significant effect of stereochem-
istry at C5 position was recorded. The in vitro screening led us to
the identification of several hits and one of which, (S)-1a, was se-
lected as the most versatile and promising hit, with a IC50 equal
to 9 and 144
lM on MMP-2 and MMP-9, respectively.
2.5. Hit to lead
In order to improve gelatinase inhibitor potency, different chem-
ical modifications were performed on the (S)-1a pyroglutamate

2326
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
O
O
O
OH
O
O
O
OH
O
+
a
HN
HN
Fmoc
Fmoc
12 (R/S)
NovaSyn TG-OH
11 (R/S)
b
O
OH
O
N
H
N
R
O
O
R
O
O
d
O
N
1' a-c (R/S)
O
R
c
e
O
14 a-c (R/S)
H
N
O
N
OH
O
R
13 a-c (R/S)
1 a-c (R/S)
Scheme 1. Reagents and conditions: (a) (1) DIC, dry DCM, rt, 1 h; (2) DMAP, dry DMF, rt, 2 h; (b) (1) 20% piperidine in DMF, 20 min, rt; (2) R-CHO, CH3COOH, NaCNBH3, in
TMOF/MeOH, rt, 1 h; (c) (1) PhSiH3, Pd[P(Ph)3]4, dry DCM (2 ꢀ 5 min); (2) HATU, DIEA, dry DMF, rt, 3 h; (d) NaOH 0.5%, rt, 1.5 h; (e) NH2OH 50% wt in H2O, THF, rt, 5 h.
OH
O
O
NH
O
O
NH2
O
O
HN
+
a
HN
Fmoc
Fmoc
11 (R/S)
15 (R/S)
b
H
N
O
R
O
NH
O
O
NH2
N
N
O
R
O
O
d
c
R
NH
17 a-c (R/S)
16 a-c (R/S)
1" a-c (R/S)
Scheme 2. Reagents and conditions: (a) HBTU, HOBT, NMM, dry DMF, rt, 2 h; (b) (1) 20% piperidine in DMF, 20 min, rt; (2) R-CHO, CH3COOH, NaCNBH3, in TMOF/MeOH, rt,
1 h; (c) (1) PhSiH3, Pd[P(Ph)3]4, dry DCM (2 ꢀ 5 min); (2) HATU, DIEA, dry DMF, rt, 3 h; (d) TFA/TIS (95:5), rt, 1 h.
The results revealed that compound 8a adopts the expected bind-
ing mode with P01 moiety involved in hydrophobic interactions
with S01 residues, giving high docking score in according with
experimental data (data not shown).
activity on both gelatinases (i.e., 8h and 8i), while, as expected, a
relevant 20-fold improvement occurs upon substitution of the
biphenyl ring with a para-substituted diaryl ether (see compounds
8j, 8k, 8l in Table 3).
Subsequently, the accomplishment of a further improvement in
inhibitory activity and a good degree of selectivity between the two
gelatinases were achieved through substitution of –S– by –S(O)n–
(where n = 1 or 2, see compound 9a and 10a in Table 3).
To underline the selectivity profiles over other MMPs, the most
promising compounds were tested against a broader panel of
MMPs (Table 4).
Compounds 8k and 8l exhibited a relevant rate of discrimina-
tion with respect to MMP-1 and MMP-3. Moreover, compounds
9a and 10a showed a significant MMP-2 selectivity even over
MMP-9. Unfortunately, more the activity increases, more the selec-
tivity fades out. In addition, data collected for compound 10k
(Table 4) show that when gelatinase activity is in the low nanomo-
2.6. 8a scaffold optimization
From compound 8a further modifications were performed on P10
group. Several 8a analogues bearing different bulky groups in R po-
sition were synthesized, most of them being related to known and
potent MMP inhibitors. Furthermore, the effects on the biological
activity of the –S– substitution on the five-membered ring with
sulfoxide and sulfone moiety were evaluated.
All the compounds were synthesized and tested against MMP-2
and MMP-9 (Table 3). According to the biological data, the substi-
tution of the biphenyl moiety with more rigid groups, such as
naphthalene or quinoline, dramatically decreases the inhibitor

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2327
Table 1
Gelatinase A and B inhibitory activity of (R)- and (S)-1 derivatives
O
N
X
O
X
N
R
R
R
X
Cpd
IC50 (M)a
Cpd
IC50 (M)a
MMP-2b
MMP-9b
MMP-2b
MMP-9b
CONHOH
(R)-1a
3
0.3
30
3
(S)-1a
9
1
144 10
COOH
CONH2
(R)-10a
(R)-100a
290
>500
ND
ND
(S)-10a
(S)-100a
340
>500
ND
ND
CONHOH
(R)-1b
>500
>500
(S)-1b
>500
ND
Cl
COOH
CONH2
(R)-10b
(R)-100b
>500
>500
ND
ND
(S)-10b
(S)-100b
315
>500
ND
ND
CONHOH
(R)-1c
400 20
430
(S)-1c
>500
ND
Cl
COOH
CONH2
(R)-10c
(R)-100c
>500
>500
ND
ND
(S)-10c
(S)-100c
383
>500
ND
ND
a
ND: not determined.
MMP2: full length; MMP9: catalytic domain.
b
HO
S
N
S
O
N
N
N
2a-(S)
3a-(R)
4a-(S)
5a-(R)
H
N
OH
O
N
O
S
O
N
SO2
N
SO2
N
SO2
1a-(S)
6a-(S)
7a-(S)
8a-(R)
Figure 6. (S)-1a scaffold modifications.
lar range, high activity is recorded also towards other MMP
members.
respectively). The lack of this additional favorable interaction is
probably due to the presence of a larger S01 pocket that forces the
molecules to plunge in deeply.
2.7. In silico investigation of MMP-2 selectivity over MMP-9
3. Solid phase synthesis
To investigate the different activity of 9a and 10a towards the
two gelatinases, docking analysis on MMP-9 and MMP-2 binding
sites was performed (see Section 5 for details). The comparison
of the best scored solutions of both molecules highlighted a differ-
ent binding mode in the two protein binding sites. For instance,
visual inspection of 10a binding mode (Fig. 7) revealed that one
oxygen atom of the sulfonamide group can form an hydrogen bond
with the NH group of MMP-9 Gln402 side-chain while in MMP-2
the hydroxamic NH group is involved in a hydrogen bond with
the backbone of Ala165. The most remarkable difference in the
binding modes of the two inhibitors relies on a significant change
of overall shapes of the two heterocyclic rings in the two gelatinas-
es. In MMP-2 binding site, one oxygen of the endocyclic SO2/SO is
involved in an hydrogen bond with the NH group of Ala164 back-
bone, while in MMP-9 the same group does not give further inter-
actions, leading to a lower docking score (10a: 82.01 and 69.70,
A versatile solid phase strategy for the synthesis of scaffolds 1–8
was developed.
In order to obtain compounds 1–10–100a–c we performed a new
synthetic solid phase approach using either a NovaSyn Tentagel re-
sin or a Rink amide resin depending on whether the final cleavage
should release a carboxylic/hydroxamic or an amidic moiety
(Scheme 1 and 2). Both enantiomers of Fmoc-protected amino acid
11 were commercially available in pure enantiomeric form. Attach-
ment to the resin followed by deprotection of the Fmoc group and
subsequent reductive amination with appropriate alkylating re-
agents provided N-alkylated compounds 13–16a–c. Removal of
the protective group easily allowed condensation of the pyroglu-
tamic ring to obtain 14–17a–c. Final cleavage from the resin gave
compounds 1–10–100a–c with good purity degree (>90% by HPLC–
MS).

2328
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
X
N
HO
+
Cl
ONHFmoc
Cl
Fmoc-NHOH
18
+
a
Cl
Fmoc
O
20 (X=CH2)
21 (X=S)
22 (X=CH-Ot-Bu)
19
b
O
N
HO HN
Fmoc
X
N
X
N
X
NH
NH
Cl
O
O
O
d
O
Cl
c
23 (X=CH2)
24 (X=S)
25 (X=CH-Ot-Bu)
2a (X=CH2)
3a (X=S)
4a (X=CH-Ot-Bu)
23 (X=CH2)
24 (X=S)
25 (X=CH-Ot-Bu)
Scheme 3. Reagents and conditions: (a) DIPEA, DCM, rt, 60 h; (b) (1) 20% piperidine in DMF, 20 min, rt; (2) HATU, DIPEA, DMF/DCM (1:1), rt, 12 h; (c) (1) 20% piperidine in
DMF, 20 min, rt; (2) 40-(Ph)-PhCHO, CH3COOH, NaCNBH3, in TMOF/MeOH, rt, 1 h; (d) AcOH/TFE/DCM (1:1:8), rt, 1 h.
Tr
S
Tr
S
Tr
S
Fmoc
HN
+
HO
Fmoc
O
O
a
N
H
b
HN
OH
29
O
O
O
31
30
c
S
H
N
O
S
d
N
OH
O
O
N
O
O
5a
32
Scheme 4. Reagents and conditions: (a) HATU, DIEA, DMAP, dry DMF/DCM, rt, 6 h; (b) (1) 20% piperidine in DMF, 20 min, rt; (2) R-CHO, CH3COOH, NaCNBH3, in TMOF/MeOH,
rt, 1 h; (c) (1) TFA/TIS/DCM (10:5:85); (2) 10% Et3N in DMF, 5 min, rt; (3) CDI, dry DCM, rt, 6 h; (d) NH2OH 50% wt in H2O, THF, rt, 5 h.
To obtain compounds 2a, 3a and 4a Fmoc protected hydroxyl-
amine was tethered to 2-chloro-chlorotrytil resin, which proved
to be suitable to ensure the final release of hydroxamic derivatives
(Scheme 3). After removal of the Fmoc group, coupling of
compound 19 with Fmoc-protected proline/thioproline/hydroxy-
proline-carboxylic acids 20–22 gave compounds 23–25. Fmoc
group was then cleaved and subsequent reductive amination with
p-phenylbenzaldehyde and cleavage from the resin released 2a, 3a
and 4a in good yields.
An innovative strategy was developed to obtain the oxothiazoli-
dine hydroxamate derivative 5a (Scheme 4). The compound was
conveniently synthesized starting from Fmoc protected (R)-2-amino-
3-(tritylthio)propanoic acid 29. The substrate was anchored on
NovaSyn Tentagel resin and after Fmoc cleavage reductive
amination with p-phenylbenzaldehyde was performed. Compound
31 was then deprotected, cyclized, and cleaved from the resin to give
compound 5a.
An original pattern was then envisaged to synthesize compound
6, which maintained the same ZBG and P01 substituent as 5a, chang-
ing the thiazolidinone ring into a pyrrolidonone (Scheme 5). Car-
boxylic acid (S)-11 was converted to hydroxamic acid (S)-33 and
anchored on 2-chloro-chlorotrytil resin to obtain (S)-34. Then,
Fmoc deprotection followed by sulfonamide synthesis gave (S)-
35. Ring closing reaction occurred using the same conditions ap-
plied to synthesize derivatives 13 and 16 after acid deprotection
(Scheme 1 and 2). Final cleavage from the resin using TFA produced
6 with high purity.
Compounds 7a, 8a, 8h–l, 9a, 10a, 10k were obtained only from
two building blocks 20 and 21 (Scheme 6). NovaSyn Tentagel resin
was chosen to tether the two acids. After Fmoc cleavage the amino
group was reacted with several sulfonyl chlorides to obtain sulfon-
amides 39a and 40a, h–l.
Compounds 39a and 40h–l were immediately released from the
resin to give 7a and 8h–l, while 40a was split into two different

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2329
HO
NH
O
O
OH
O
O
O
O
Cl
+
a
HN
Cl
HN
Fmoc
33
Fmoc
11 (S)
b
NH
35
O
O
NH
O
O
O
O
S
O
c
O
HN
HN
Fmoc
34
d
H
N
H
O
O
N
N
N
S
OH
O
O
S
O
O
e
O
O
O
6a
36
Scheme 5. Reagents and conditions: (a) (1) SOCl2, dioxane, 50 °C, 2.5 h; (2) NH2OH, DIEA, dioxane/H2O, rt, o.n. (b) (1) DIEA, DMF/DCM (1:10) rt, 60 h; (2) DCM/MeOH/DIEA
(17:2:1), rt, 3 ꢀ 15 min; (c) (1) 20% piperidine in DMF, 20 min, rt; (2) DIEA, R-SO2Cl, dry DCM, rt, o.n.; (d) (1) PhSiH3, Pd[P(Ph)3]4, dry DCM (2 ꢀ 5 min); (2) HATU, DIEA, dry
DMF, rt, 3 h; (e) TFA 5%, DCM, rt, 2 h.
batches. The first batch was used to give 8a, the second one was
oxidized with MCPBA, so to introduce a sulfone moiety on the ring
to give 41a. Compound 41a was again split in two in order to
obtain 9a and perform the oxidation of sulfone to sulfoxide to
finally release 10a.
5.1.2. MoDa-3D database
By using MOE, the compounds were selected from the MoDa-2D
database according to the drug-like criteria, by property descriptor
calculation, which include: number of violations of the Lipinsky’s
rule of five and other drug-like filters such as number of rings,
number of halogens and undesirable reactive groups. A subset of
about 600,000 compounds were selected and used to generate a
database of conformers. The conformational analysis search was
carried out via molecular dynamic (MD) as implemented in the
MOE software package. For each compound, an MD run of 1 ns
was performed using the NVT ensemble with T = 300 K, time step
of 0.002 ps using MMFF94 force field and the Born solvation model.
A total of 1000 conformations were collected for each MD run and
they were energy-minimized requiring a RMSD of the gradient less
than 0.05 kcal/mol.
4. Conclusion
In conclusion, a pharmacophore based virtual screening of our
in-house database and further chemical optimization led us to de-
sign novel gelatinase inhibitors. The compounds were synthesized
developing and applying a solid phase approach. Docking simula-
tions were performed to study the possible motifs of selectivity be-
tween MMP-2 and MMP-9. Notably, a very interesting selectivity
was observed for some compounds with good in silico drug-like-
ness profile. In particular the 4-thiazolydinyl-N-hydroxycarboxya-
mide derivatives 9a and 10a showed a MMP-2 inhibitory activity
in the low nanomolar range and a good selectivity over the MMPs
analysed in this study. The achieved results represent a good start-
ing point for future research activities, directed toward further
medicinal chemistry optimizations.
5.1.3. Pharmacophore design
MOE Pharmacophore Applications module has been used to de-
sign pharmacophore hypotheses and to perform database screen-
ing. The PCH pharmacophore scheme included the following
atomic tags: H-Bond Acceptor, H-Bond Donor, Cation, Anion, Aro-
matic ring center, Hydrophobic Region, Metal Ligator, Volume Con-
straint. In detail, by using SC-74020 ligand coordinates as reference
template, the first pharmacophoric point was fixed on hydroxamic
group setting a tolerance radius of 1.8 Å to include one or both oxy-
gen atoms of alternative other ZBGs (e.g. carboxylic acids, hydroxa-
mic acids, phosphates). The second pharmacophoric point was
defined on the P01 phenyl group, with a tolerance radius compara-
ble with the size of an aromatic ring (1.8 Å). Finally, the overall
shape of the protein around the catalytic site and the S01 pocket
was modeled considering the excluded volume defined by the
protein residues with at least one atom within 4.5 Å from any atom
5. Experimental section
5.1. Molecular modeling
5.1.1. MoDa-2D database
MoDa is a web-based application developed using the Perl pro-
gramming language, containing about 2 millions distinct mole-
cules from about 9 millions commercially available compounds.
MoDa uses MySQL as Relational Database management System
and Apache as web server.

2330
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
X
N
OH
+
HO
N
Fmoc
O
Fmoc
O
a
O
20
NovaSyn TG-OH
37 (X=CH2)
38 (X=S)
b
S
N
OH
+
HO
Fmoc
O
c
20
NovaSyn TG-OH
X
S(O)n
d
O
O
N
S
O
N
S
O
O
O
O
R
R
O
39 a (X=CH2)
40 a, h-l (X=S)
41 a (n=1)
42 a (n=2)
e
f
f
X
HO
O
N
S
O
N
H
O
O
R
7a
(X=CH2)
8a, h-l (X=S)
9a
[X=S(O)]
10a,k [X=S(O)2]
Scheme 6. Reagents and conditions: (a) HATU, DIPEA, dry DCM/DMF (1:1), rt, 12 h; (b) DMAP, dry DMF, rt, 12 h; (c) (1) 20% piperidine in DMF, 20 min, rt; (2) DMAP, R-SO2–Cl,
THF/DCM (2:1) rt, 12 h; (d) MCPBA, dry DCM, rt, 27 h; (e) MCPBA, dry DCM, rt, 27 h; (f) NH2OH 50% wt in H2O, THF, rt, 5 h.
Table 2
Gelatinase A and B inhibitory activity of 1–8a compounds
were aligned, the water molecules, cofactors and (MMP-9) co-
crystallized ligand were removed; by using Protonate 3D tool,
the hydrogen atoms were added, fixing the correct protonation of
protein residues. The inhibitors described in this work were built
from the MOE fragment library and minimized by setting MMFF94
force field.
H
N
X
OH
Y
N
O
Z
Docking simulations were performed using the GOLD program
(Cambridge Crystallographic Data Centre, Cambridge, UK) by set-
ting the GoldScore (GS) fitness functions to quantify the interaction
energy of ligand–protein complexes. The fitness function parame-
ters were taken from the GOLD parameter file. The binding cavities
contained the protein residues in a sphere of a 12 Å radius centred
on the SC-74020 aromatic carbon C18. Zinc trigonal/bipyramidal
coordination geometry was imposed. For each compounds, 10
poses were generated into the binding site of MMP-2 and MMP-
9, respectively. The binding mode inspections were carried out
by Hermes Viewer implemented in GOLD software.
IC50
MMP2wha
(l
M)
MMP9cdb
Compd
X
Y
Z
1a
2a
3a
4a
5a
6a
7a
8a
CH2
CH2
S
CH(OH)
S
CH2
CH2
S
C(O)
CH2
CH2
CH2
C(O)
C(O)
CH2
CH2
CH2
CH2
CH2
CH2
CH2
SO2
SO2
SO2
9
1
144 10
ND
ND
>500
330 15
>500
ND
6.80
1
47
1.9 0.1
76
0.23 0.05
8
1.7 0.5
4
1
4
0.20 0.01
5.1.5. Additional softwares
a
Visual inspections of docking and pharmacophore analysis were
performed with PyMOL.26 In addition MOE, Babel, CDK and JOELib
were used as chemoinformatic tools for SMARTS search, conver-
sion, and other routine tasks.
MMP2wh: MMP2 full length.
MMP9: full length; MMP9: catalytic domain.
b
of SC-74020. Around 200 atom points were used to describe the
excluded volume in such a manner that also larger inhibitors
should satisfy the same query (Fig. 3b).
5.2. Chemistry
Reagents and solvents were obtained from commercial suppli-
ers and used without further purification. Thin layer chromatogra-
phy (TLC) analytical separations were conducted with E. Merck
Silica Gel F-254 plates of 0.25 mm thickness and were visualized
with UV light (254 nm) or I2. Flash chromatography was performed
5.1.4. Docking simulations
In the present study, the 3D coordinates of MMP-2 and MMP-9
(pdb codes: 1qib and 2ow1, respectively) were retrieved from the
PDB. Protein preparation was performed in MOE: the two proteins

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2331
Table 3
Elementary analyses were performed on a Perkin-Elmer 2400
Series II Elemental Analyser and a microbalance AD-4 autobalance
Perkin-Elmer.
MMP-2 e MMP-9 screening results of 8a analogues
(O)
n
S
Optical rotations were determined at 25 °C on a Perkin-Elmer
341 polarimeter (concentrations are reported in units of grams
per 100 mL).
H
N
OH
N
O
SO2
5.2.1. General methods for solid phase synthesis
R
Solvents were distilled for resin reactions. Unless otherwise sta-
ted reactions were performed in 8 mL polypropylene cartridges
with 70 mm PE frits. Cartridges and stopcocks were obtained from
Applied Separations. All Fmoc amino acids, coupling reagents, and
resins were purchased from Novabiochem, and other chemicals
were from Aldrich.
IC50 (nM)
MMP2wh
Compd
n
R
MMP9cd
8a
0
200 10
230 50
5.2.2. General procedures for coupling of Fmoc-(D/L)-Glu-(OAll)-
OH to NovaSynTGOH
8h
8i
0
0
31,000 500
710 40
>50,000
560 30
Example: Synthesis of (R)-12.
N,N0-Diisopropylcarbodiimide (DIC) (300
l
l, 1.9 mmol) was added
N
to a solution of Fmoc-
D-Glu-(OAll)-OH (R)-11 (1.55 g, 3.8 mmol) in
10 ml dry DCM at 0 °C. After stirring at room temperature for 1 h,
DCM was evaporated, 3 ml of DMF were added and the solution
was dropped into the cartridge in which hydroxyl group function-
alized resin (NovaSynTGOH) (1.26 g, 0.38 mmol) was pre-swollen
in DCM. 4-Dimethylaminopyridine (DMAP) (5 mg, 0.04 mmol) in
0.3 ml dry DMF was introduced and the cartridge was gently
rocked for 2 h at room temperature.
After filtration the resin was rinsed with DMF (3 ꢀ 5 ml), DCM
(3 ꢀ 5 ml) and TBME (3 ꢀ 5 ml) and finally dried in vacuo.
The efficiency of loading (loading yield: 81%) was determined
on 5.9 mg of resin by Fmoc removal analysis (treatment of the resin
with piperidine 20% in DMF) and absorbance measurement at
301 nm (UV–vis Spectrophotometer Shimadzu).
H3CO
Cl
8j
0
0
0
9
13
10
1
5
2
2
0.3
9
O
8k
8l
16
5
O
F
0.5
O
9a
1
9
3
600 100
5.2.3. General procedures for Fmoc deprotection and reductive
amination of polymer bound O-All-(D/L) glutamate
10a
10k
2
2
53 16
480 80
Example: Synthesis of (R)-13a.
After Fmoc deprotection with 20% piperidine in DMF for
20 min (ꢀ2), (R)-12 (0.08 mmol) was washed (2 ꢀ DMF, 2 ꢀ DCM)
and swollen in DCM, then biphenyl-4-carboxaldehyde (120 mg,
0.66 mmol) in 0.8 ml trimethyl orthoformate (TMOF) was added
Cl
0.8 0.1
0.5 0.07
O
dropwise. Subsequently methanol (MeOH) (200
(38 l, 0.66 mmol) and NaCNBH3 (124 mg, 1.97 mmol) in MeOH
(400 l) were respectively introduced into the reaction mixture.
ll), acetic acid
l
l
using E. Merck silica gel. 1H nuclear magnetic resonance (NMR)
spectra were recorded in the specified deuterated solvents on a
Bruker Avance 300 spectrometer operating at 300 MHz. Chemical
shifts are reported in parts per million (d) from the tetramethylsil-
ane resonance in the indicated solvent (TMS: 0.0 ppm). Com-
pounds purities and mass spectra were determined by an LC/MS
platform (Gilson/ThermoFinnigan) using the positive electrospray
ionization technique (+ES) with a mobile phase of acetonitrile/
water with 0.1% TFA.
After shaking for 1 h at room temperature, the resin was filtered
and washed with MeOH (ꢀ3), DMF (ꢀ3), DCM (ꢀ3) and TBME
(ꢀ3).
5.2.4. General procedures for allyl deprotection and cyclization
Example: Synthesis of (R)-14a.
PhSiH3 (500 ll, 4.1 mmol) and Pd[P(Ph)3]4 (95 mg, 0.08 mmol)
in 2 ml dry DCM were added to 0.08 mmol of (R)-13a pre-swollen
in DCM.
Table 4
Biological data collected for compounds 8a analogues
IC50 (nM)
MMP3cd
Compd
MMP2
MMP9
MMP1cd
MMP8cd
MMP12cd
MMP14cd
8j
8k
8l
9a
10a
10k
9
13
10
9
53 16
0.8 0.1
1
5
2
3
2
16
5
0.3
9
0.5
ND
ND
ND
ND
ND
400 100
100 40
ND
ND
ND
2500 500
500 10
ND
ND
ND
160 20
3800 500
1000 50
>50,000
4000 500
2500 100
ND
160 30
700 100
2000 300
4000 100
ND
600 100
480 80
0.5 0.07
4
1
19
6
5
0.5

2332
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
Figure 7. Docking results of compound 10a into the binding site of MMP-2 (a) and MMP-9 (b). The proteins are rendered as green and cyan surfaces, respectively. The
relevant residues are depicted as line figures. The most important interactions between ligand and protein residues and metal coordinations are rendered as blue dashed
lines. Zinc ion is rendered as gray sphere.
After repeating the treatment twice (2 ꢀ 5 min) the resin was
filtered, rinsed first with DCM, and later with a solution of DIEA
(0.5%) and NaEt2NCS2 (0.5%) in DMF, and again with DMF, DCM
and DMF.
CH3OH). Anal. Calcd for C18H18N2O3: C, 69.66; H, 5.85; N, 9.03; O,
15.47. Found: C, 69.93; H, 5.86; N, 9.04; O, 15.42. Overall yield 15%.
5.2.5.2. (R)-1b: (R)-1-(2-Chlorobenzyl)-N-hydroxy-5-oxopyrrol-
DIEA (17
ll, 0.01 mmol) and a solution of HATU (2-(1H-9-
idine-2-carboxamide.
Amorphous solid—1H NMR (300 MHz,
azabenzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophos-
phate) (38 mg, 0.01 mmol) previously dissolved in 0.7 ml of dry
DMF were introduced into the cartridge; the cyclization proceeded
for 3 h at room temperature. The resin was filtered and rinsed with
DMF (ꢀ3), MeOH (ꢀ3), DCM (ꢀ3) and TBME (ꢀ3).
DMSO-d6): d 1.84–1.91 (m, 1H), 2.18–2.27 (m, 1H), 2.29–2.45 (m,
2H), 3.83–3.87 (m, 2H), 4.79 (d, 1H, J = 15.6 Hz), 7.21–7.23 (m,
1H), 7.32–7.34 (m, 2H), 7.45–7.47 (m, 1H), 9.07 (br s, 1H), 10.74
(br s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 269.1 (MH+,
100), C12H13ClN2O3: 268.7. Specific optical rotation ½a D25
ꢂ68
ꢁ
(c 0.09, CH3OH). Anal. Calcd for C12H13ClN2O3: C, 53.64; H, 4.88;
Cl, 13.19; N, 10.43; O, 17.86. Found: C, 53.44; H, 4.87; Cl, 13.15;
N, 10.45; O, 17.90.
5.2.5. General procedures for the cleavage of hydroxamate
derivatives 1
Example: (R)-1-(4-phenylbenzyl)-N-hydroxy-5-oxopyrrolidine-
2-carboxamide [(R)-1a].
Functionalized resin derivative (R)-14a (0.04 mmol) was swol-
len in THF (1.8 ml) and cleaved from the solid support by treat-
ment with 50% NH2OH in H2O (300 ll, 5.6 mmol). After rocking
for 5 h at room temperature, the resin was filtered and the filtrate
5.2.5.3. (S )-1b: (S)-1-(2-Chlorobenzyl)-N-hydroxy-5-oxopyrrol-
idine-2-carboxamide.
Amorphous solid—1H NMR (300 MHz,
DMSO-d6): d 1.84–1.92 (m, 1H), 2.16–2.26 (m, 1H), 2.29–2.45
(m, 2H), 3.82–3.84 (m, 1H), 3.85 (d, 1H, J = 15.6 Hz), 4.79 (d, 1H,
J = 15.6 Hz), 7.21–7.23 (m, 1H), 7.31–7.34 (m, 2H), 7.45–7.47 (m,
1H), 9.06 (br s, 1H), 10.75 (br s, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 269.1 (MH+, 100), C12H13ClN2O3: 268.7. Specific
was evaporated to dryness.
The crude product was further purified by semi-preparative RP-
HPLC (C18, Nucleosil 300 mm ꢀ 25 mm, 3 ml/min, detection at
230 nm, eluent A: Water/TFA: 99.92:0.08, eluent B: acetonitrile/
TFA: 99.92:0.08; gradient B from 20% to 80% in 14 min, from 80%
to 98% in 3 min) to afford (R)-1a.
optical rotation
½
a 2D5
ꢁ
+66 (c 0.09, CH3OH). Anal. Calcd for
C12H13ClN2O3: C, 53.64; H, 4.88; Cl, 13.19; N, 10.43; O, 17.86.
Found: C, 53.50; H, 4.87; Cl, 13.17; N, 10.45; O, 17.88. Overall yield
11%.
Amorphous solid—1H NMR (300 MHz, DMSO-d6): d 1.82–1.90
(m, 1H), 2.1–2.2 (m, 1H), 2.29–2.45 (m, 2H), 3.65 (d, 1H,
J = 14.7 Hz), 3.82 (dd, 1H, J = 8.8, 3.9 Hz), 4.90 (d, 1H, J = 14.7 Hz),
7.27 (d, 2H, J = 7.8 Hz), 7.36 (t, 1H, J = 14.7, 7.8 Hz), 7.46 (dd as t,
2H, J = 15.6, 7.8 Hz), 7.63 (d, 2H, J = 7.8 Hz), 7.65 (d, 2H,
J = 5.9 Hz), 9.08 (br s, 1H), 10.79 (br s, 1H). LC–MS (ESI) (H2O/
CH3CN, 1%TFA) m/z (%): 311.3 (MH+, 100), C18H18N2O3: 310.3. Spe-
5.2.5.4. (R )-1c: (R)-1-(4-Chlorobenzyl)-N-hydroxy-5-oxopyrrol-
idine-2-carboxamide.
Amorphous solid—1H NMR (300 MHz,
DMSO-d6): d 1.81–1.88 (m, 1H), 2.09–2.19 (m, 1H), 2.26–2.43 (m,
2H), 3.65 (d, 1H, J = 15.6 Hz), 3.76 (dd, 1H, J = 7.8, 3.91 Hz), 4.79
(d, 1H, J = 15.6 Hz), 7.21 (d, 2H, J = 8.8 Hz), 7.39 (d, 2H, J = 8.8 Hz),
9.08 (br s, 1H), 10.74 (br s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA)
m/z (%): 269.1 (MH+, 100), C12H13ClN2O3: 268.7. Specific optical
cific optical rotation ½a D25
ꢂ51 (c 0.02, CH3OH). Anal. Calcd for
ꢁ
C18H18N2O3: C, 69.66; H, 5.85; N, 9.03; O, 15.47. Found: C, 69.99;
H, 5.83; N, 9.05; O, 15.23. Overall yield 12%).
rotation ½a 2D5
ꢂ55 (c 0.04, CH3OH). Anal. Calcd for C12H13ClN2O3:
ꢁ
C, 53.64; H, 4.88; Cl, 13.19; N, 10.43; O, 17.86. Found: C, 53.58;
H, 4.87; Cl, 13.16; N, 10.41; O, 17.82. Overall yield 13%.
5.2.5.1. (S )-1a: (S)-1-(4-Phenylbenzyl)-N-hydroxy-5-oxopyrrol-
idine-2-carboxamide.
Amorphous solid—1H NMR (300 MHz
5.2.5.5. (S )-1c: (S)-1-(4-Chlorobenzyl)-N-hydroxy-5-oxopyrrol-
idine-2-carboxamide.
DMSO-d6): d 1.83–1.88 (m, 1H), 2.11–2.21 (m, 1H), 2.29–2.46
(m, 2H), 3.65 (d, 1H, J = 15.6 Hz), 3.82 (dd, 1H, J = 8.8, 3.9 Hz),
4.91 (d, 1H, J = 15.6 Hz), 7.27 (d, 2H, J = 7.8 Hz), 7.36 (dd as t, 1H,
J = 14.7, 6.86 Hz), 7.46 (dd as t, 2H, J = 14.7, 6.86 Hz), 7.63 (d, 2H,
J = 7.8 Hz), 7.65 (d, 2H, J = 5.9 Hz), 9.08 (br s, 1H), 10.79 (br s,
1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 311.2 (MH+, 100),
Amorphous solid—1H NMR (300 MHz,
DMSO-d6): d 1.81–1.89 (m, 1H), 2.07–2.17 (m, 1H), 2.25–2.41
(m, 2H), 3.65 (d, 1H, J = 15.6 Hz), 3.75 (dd, 1H, J = 7.8, 3.9 Hz), 4.78
(d, 1H, J = 15.6 Hz), 7.21 (d, 2H, J = 8.8 Hz), 7.39 (d, 2H, J = 8.8 Hz),
9.11 (br s, 1H), 10.82 (br s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA)
m/z (%): 269.1 (MH+, 100), C12H13ClN2O3: 268.7. Specific optical
C
18H18N2O3: 310.3. Specific optical rotation ½a D25
ꢁ
+53 (c 0.02,
rotation ½a 2D5
ꢁ
+56 (c 0.04, CH3OH). Anal. Calcd for C12H13ClN2O3: C,

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2333
53.64; H, 4.88; Cl, 13.19; N, 10.43; O, 17.86. Found: C, 53.61; H, 4.87;
Cl, 13.18; N, 10.41; O, 17.84. Overall yield 13%.
5.2.6.5. (S )-10c: (S )-1-(4-Chlorobenzyl)-5-oxopyrrolidine-2-car-
boxylic acid.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d
1.91–1.98 (m, 1H), 2.21–2.37 (m, 3H), 3.91 (d, 1H, J = 9.0 Hz),
3.95 (d, 1H, J = 15.3 Hz), 4.79 (d, 1H, J = 15.3 Hz), 7.23 (d, 2H,
J = 8.2 Hz), 7.39 (d, 2H, J = 8.2 Hz). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 254.1 (MH+, 75), C12H12ClNO3: 253.7. Specific opti-
5.2.6. General procedures for the cleavage of carboxylate
derivatives 10
Example: (R)-1-(4-phenylbenzyl)-5-oxopyrrolidine-2-carboxylic
acid [(R)-10a].
cal rotation ½a 2D5
ꢁ
+79 (c 0.02, CH3OH). Anal. Calcd for C12H12ClNO3:
C, 56.81; H, 4.77; Cl, 13.98; N, 5.52; O, 18.92. Found: C, 56.72; H,
4.76; Cl, 13.96; N, 5.50; O, 18.89. Overall yield 14%.
Functionalized resin derivative (R)-14a (0.04 mmol) was swol-
len in THF (1.8 ml) and cleaved by NaOH 0.5% solution (2.5 ml,
0.37 mmol). After rocking for 1.5 h at room temperature, the resin
was filtered and the filtrate was evaporated to dryness.
Deprotected (R)-10a was purified by semi-preparative RP-HPLC
(C18, Nucleosil 300 mm ꢀ 25 mm, 3 ml/min, detection at 230 nm,
eluent A: Water/TFA: 99.92:0.08, eluent B: acetonitrile/TFA:
99.92:0.08; gradient B from 25% to 60% in 13 min, from 60% to
98 in 1 min).
5.2.7. General procedures for coupling of Fmoc-
OH to Rink amide resin
D/L-Glu-(OAll)-
Example: Synthesis of (R)-15.
0.12 mmol of Rink amide resin (Novabiochem) previously
Fmoc-deprotected (double treatment of the resin with piperidine
20% in DMF) were swollen in DCM.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d 1.93–2.0 (m, 1H),
2.26–2.38 (m, 3H), 3.93–3.98 (m, 2H), 4.91 (d, 1H, J = 14.9 Hz), 7.29
(d, 2H, J = 7.8 Hz), 7.36 (t, 1H, J = 7.4 Hz), 7.46 (t, 2H, J = 7.4 Hz),
7.63 (d, 2H, J = 8.2 Hz), 7.65 (d, 2H, J = 7.4 Hz). LC–MS (ESI) (H2O/
CH3CN, 1%TFA) m/z (%): 296.2 (MH+, 100), C18H17NO3: 295.3. Spe-
A solution of Fmoc-D-Glu-(OAll)-OH (R)-11 (0.24 g, 0.59 mmol),
2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexa-
fluorophosphate (HBTU) (0.22 g, 0.59 mmol), N-hydroxybenzotria-
zole (HOBt) (80 mg, 0.59 mmol), and N-methylmorpholine (NMM)
(130 ll, 1.18 mmol) in 1 ml dry DMF was introduced into the
cific optical rotation ½a D25
ꢁ
ꢂ50 (c 0.01, CH3OH). Anal. Calcd for
cartridge. After stirring at room temperature for 2 h, the resin
was filtered, rinsed with DMF (3 ꢀ 5 ml), DCM (3 ꢀ 5 ml) and
TBME (3 ꢀ 5 ml) and finally dried in vacuo.
C18H17NO3: C, 73.20; H, 5.80; N, 4.74; O, 16.25. Found: C, 73.17;
H, 5.79; N, 4.75; O, 16.26. Overall yield 15%.
A positive Kaiser test27 resulted, indicating no free amine
groups on the resin.
5.2.6.1. (S)-10a: (S)-1-(4-Phenylbenzyl)-N-hydroxy-5-oxopyrrol-
idine-2-carboxamide.
Waxy solid—1H NMR (300 MHz,
DMSO-d6): d 1.94–2.0 (m, 1H), 2.26–2.38 (m, 3H), 3.93–3.97
(m, 2H), 4.9 (d, 1H, J = 15.3 Hz), 7.29 (d, 2H, J = 8.2 Hz), 7.36
(t, 1H, J = 7.4 Hz), 7.46 (dd as t, 2H, J = 7.4 Hz), 7.63 (d, 2H,
J = 8.2 Hz), 7.65 (d, 2H, J = 7.8 Hz). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 296.2 (MH+, 100), C18H17NO3: 295.3. Specific opti-
5.2.8. General procedures for Fmoc deprotection and reductive
amination of polymer bound O-all-glutamate
The same procedure described for NovaSynTGOH bound deriv-
atives (Synthesis of (R)-13) was followed.
cal rotation ½a 2D5
ꢁ
+56 (c 0.05, CH3OH). Anal. Calcd for C18H17NO3: C,
73.20; H, 5.80; N, 4.74; O, 16.25. Found: C, 73.15; H, 5.81; N, 4.75;
O, 16.28. Overall yield 17%.
5.2.9. General procedures for synthesis of polymer bound
pyroglutamate ring
The same procedure described for NovaSynTGOH bound deriv-
atives (synthesis of (R)-14a) was followed.
5.2.6.2. (R)-10b: (R)-1-(2-Chlorobenzyl)-5-oxopyrrolidine-2-car-
boxylic acid.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d
5.2.10. General procedures for the cleavage of carboxamide
derivatives 100
1.98 (m, 1H), 2.29–2.37 (m, 3H), 3.96 (d, 1H, J = 6.3 Hz), 4.11 (d,
1H, J = 15.6 Hz), 4.85 (d, 1H, J = 15.3 Hz), 7.24–7.26 (m, 1H), 7.31–
7.35 (m, 2H), 7.44–7.48 (m, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 254.1 (MH+, 98), C12H12ClNO3: 253.7. Specific opti-
Example: (R)-1-(4-chlorobenzyl)-5-oxopyrrolidine-2-carboxy-
amide [(R)-100c].
(R)-17c was subject to cleavage by treatment with TFA/TIS
cal rotation ½a 2D5
ꢁ
ꢂ18 (c 0.07, CH3OH). Anal. Calcd for C12H12ClNO3:
(95:5) (120 l
l, 2 mmol) in order to obtain (R)-100c carboxamide.
C, 56.81; H, 4.77; Cl, 13.98; N, 5.52; O, 18.92. Found: C, 56.94; H,
4.78; Cl, 13.99; N, 5.53; O, 18.96. Overall yield 18%.
After rocking for 1 h at room temperature, the resin was filtered
and the filtrate was evaporated to dryness.
The crude compound (R)-100c did not need further purification
as shown by HPLC–MS analysis (97% of pure compound at
220 nm; 88% at 254 nm).
5.2.6.3. (S)-10b: (S)- 1-(2-Chlorobenzyl)-5-oxopyrrolidine-2-car-
boxylic acid.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d
1.98 (m, 1H), 2.33 (m, 3H), 3.97 (d, 1H, J = 6.3 Hz), 4.11 (d, 1H,
J = 15.6 Hz), 4.85 (d, 1H, J = 15.6 Hz), 7.25–7.28 (m, 1H), 7.31–
7.35 (m, 2H), 7.44–7.48 (m, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 254.1 (MH+, 100), C12H12ClNO3: 253.7. Specific
Waxy solid—1H NMR (300 MHz, DMSO-d6): d 1.79–1.89 (m, 1H),
2.05–2.37 (m, 3H), 3.73 (d, 1H, J = 15.2 Hz), 3.85 (dd, 1H, J = 8.8,
3.42 Hz), 4.81 (d, 1H, J = 15.2 Hz), 7.16 (br s, 1H), 7.21 (d, 2H,
J = 8.3 Hz), 7.39 (d, 2H, J = 8.3 Hz), 7.55 (br s, 1H). LC–MS (ESI)
(H2O/CH3CN, 1%TFA) m/z (%): 253.2 (MH+, 100), C12H13ClN2O2:
optical rotation
½
a 2D5
+23 (c 0.1, CH3OH). Anal. Calcd for
ꢁ
C12H12ClNO3: C, 56.81; H, 4.77; Cl, 13.98; N, 5.52; O, 18.92. Found:
252.1. Specific optical rotation ½a D25
ꢂ56 (c 0.04, CH3OH). Anal.
ꢁ
C, 56.88; H, 4.78; Cl, 13.97; N, 5.53; O, 18.94. Overall yield 15%.
Calcd for C12H13ClN2O2: C, 57.04; H, 5.19; Cl, 14.03; N, 11.09; O,
12.66. Found: C, 57.14; H, 5.20; Cl, 14.05; N, 11.11; O, 12.69. Over-
all yield 32%.
5.2.6.4. (R)-10c: (R)-1-(4-Chlorobenzyl)-5-oxopyrrolidine-2-car-
boxylic acid.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d
1.92–1.98 (m, 1H), 2.22–2.37 (m, 3H), 3.93–3.97 (m, 2H), 4.79 (d,
1H, J = 15.3 Hz), 7.23 (d, 2H, J = 8.6 Hz), 7.39 (d, 2H, J = 8.6 Hz).
LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 254.1 (MH+, 75),
5.2.10.1. (R )-100a: (R)-1-(4-Phenylbenzyl)-5-oxopyrrolidine-2-
carboxyamide.
Waxy solid—1H NMR (300 MHz, DMSO-d6):
d 1.81–1.89 (m, 1H), 2.07–2.42 (m, 3H), 3.74 (d, 1H, J = 15.3 Hz),
3.91 (dd, 1H, J = 9.0, 3.09 Hz), 4.93 (d, 1H, J = 15.3 Hz), 7.19 (s,
1H), 7.28 (d, 2H, J = 8.2 Hz), 7.36 (t, 1H, J = 7.4 Hz), 7.46 (t, 2H,
J = 7.8 Hz), 7.58 (s, 1H), 7.63 (d, 2H, J = 7.8 Hz), 7.65 (d, 2H,
J = 6.6 Hz). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 295.2 (MH+,
C
12H12ClNO3: 253.7. Specific optical rotation ½a D25
ꢂ78 (c 0.03,
ꢁ
CH3OH). Anal. Calcd for C12H12ClNO3: C, 56.81; H, 4.77; Cl, 13.98;
N, 5.52; O, 18.92. Found: C, 56.65; H, 4.76; Cl, 13.94; N, 5.50; O,
18.87. Overall yield 15%.

2334
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
100), C18H18N2O2: 294.3. Specific optical rotation
½
a 2D5
ꢁ
ꢂ66
A solution of Fmoc-
1 mmol), and DIPEA (348
L
-proline (0.34 g, 1 mmol), HATU (0.38 g,
(c 0.03, CH3OH). Anal. Calcd for C18H18N2O2: C, 73.45; H, 6.16; N,
9.52; O, 10.87. Found: C, 73.20; H, 6.17; N, 9.54; O, 10.83. Overall
yield 34%.
l
l, 2 mmol) in 2 ml of dry DMF/DCM
(1:1) was added dropwise on the functionalized resin 19
(0.2 mmol) after Fmoc deprotection via classical procedure (20%
Piperidine in DMF for 20 min (ꢀ2)). The cartridge was shaked for
12 h at room temperature.
After filtration the resin was rinsed with DMF (3 ꢀ 5 ml), DCM
(3 ꢀ 5 ml) and TBME (3 ꢀ 5 ml) and finally dried in vacuo.
The efficiency of loading (loading yield: 53%) was determined
on 3 mg of resin by Fmoc removal analysis (treatment of the resin
with piperidine 20% in DMF) and absorbance measurement at
301 nm (UV–vis Spectrophotometer Shimadzu).
5.2.10.2. (S)-100a: (S)-1-(4-Phenylbenzyl)-5-oxopyrrolidine-2-car-
boxyamide.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d
1.80–1.89 (m, 1H), 2.06–2.41 (m, 3H), 3.73 (d, 1H, J = 15.3 Hz),
3.91 (dd, 1H, J = 9.0, 3.91 Hz), 4.91 (d, 1H, J = 15.3 Hz), 7.19
(s, 1H), 7.27 (d, 2H, J = 7.8 Hz), 7.35 (t, 1H, J = 7.4 Hz), 7.46 (t, 2H,
J = 7.4 Hz), 7.60 (s, 1H), 7.63 (t, 4H, J = 6.6 Hz). LC–MS (ESI) (H2O/
CH3CN, 1%TFA) m/z (%): 295.2 (MH+, 100), C18H18N2O2: 294.3. Spe-
cific optical rotation ½a D25
ꢁ
+69 (c 0.04, CH3OH). Anal. Calcd for
C18H18N2O2: C, 73.45; H, 6.16; N, 9.52; O, 10.87. Found: C, 73.30;
H, 6.17; N, 9.53; O, 10.85. Overall yield 37%.
5.2.13. General procedures for Fmoc deprotection and reductive
amination on polymer bound proline/thioproline/
hydroxyproline rings
The same procedure described for the synthesis of (R)-13 was
followed.
5.2.10.3. (R )-100b: (R)-1-(2-Chlorobenzyl)-5-oxopyrrolidine-2-
carboxyamide.
Waxy solid—1H NMR (300 MHz, DMSO-d6):
d 1.82–1.91 (m, 1H), 2.07–2.37 (m, 3H), 3.89–3.95 (m, 2H), 4.84
(d, 1H, J = 16.1 Hz), 7.16 (br s, 1H), 7.24 (t, 1H, J = 4.4 Hz), 7.33
(t, 2H, J = 4.4 Hz), 7.46 (t, 1H, J = 4.4 Hz), 7.54 (br s, 1H). LC–MS
(ESI) (H2O/CH3CN, 1%TFA) m/z (%): 253.2 (MH+, 100), C12H13ClN2O2:
5.2.14. General procedures for cleavage of hydroxamates
derivatives 2a, 3a, 4a
Example: Synthesis 2a.
252.1. Specific optical rotation ½a D25
ꢂ89 (c 0.03, CH3OH). Anal.
ꢁ
After swelling functionalized resin 26, the crude product 2a
(0.1 mmol) was cleaved by treatment with 2 ml of AcOH/TFE/
DCM (1:1:8) for 1 h at room temperature. The resin was filtered
and the filtrate was evaporated under N2.
Calcd for C12H13ClN2O2: C, 57.04; H, 5.19; Cl, 14.03; N, 11.09; O,
12.66. Found: C, 56.95; H, 5.18; Cl, 13.98; N, 11.06; O, 12.62. Over-
all yield 33%.
The crude product 2a was further purified by semi-preparative
RP-HPLC (C18, Nucleosil 300 mm ꢀ 25 mm, 3 ml/min, detection at
230 nm, eluent A: Water/TFA: 99.92:0.08, eluent B: acetonitrile/
TFA: 99.92:0.08; gradient B from 20% to 80% in 14 min, from 80%
to 98% in 3 min).
5.2.10.4. (S)-100b: (S)-1-(2-Chlorobenzyl)-5-oxopyrrolidine-2-car-
boxyamide.
Waxy solid—1H NMR (300 MHz, DMSO-d6): d
1.82–1.90 (m, 1H), 2.07–2.37 (m, 3H), 3.89–3.92 (m, 1H), 3.93
(d, 1H, J = 15.6 Hz), 4.84 (d, 1H, J = 15.6 Hz), 7.17 (br s, 1H), 7.24
(t, 1H, J = 4.9 Hz), 7.33 (t, 2H, J = 4.4 Hz), 7.46 (t, 1H, J = 4.9 Hz),
7.54 (br s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 253.1
1H NMR (300 MHz, DMSO-d6): waxy solid—1H NMR (300 MHz,
DMSO-d6): d 1.77–1.94 (m, 2H), 2.00–2.13 (m, 1H), 3.14–3.35 (m,
1H), 3.93–4.04 (m, 1H), 4.10–4.20 (m, 1H), 4.25–4.45 (m, 3H),
7.38–7.77 (m, 9H), 9.31 (br s, 1H), 11.12 (br s, 1H). LC–MS (ESI)
(H2O/CH3CN, 1%TFA) m/z (%): 297.3 (MH+, 100), C18H20N2O2:
(MH+, 100), C12H13ClN2O2: 252.1. Specific optical rotation ½a 2D5
ꢁ
+91 (c 0.05, CH3OH). Anal. Calcd for C12H13ClN2O2: C, 57.04; H,
5.19; Cl, 14.03; N, 11.09; O, 12.66. Found: C, 57.01; H, 5.18; Cl,
13.99; N, 11.07; O, 12.64. Overall yield 34%.
296.1. Specific optical rotation ½a D25
ꢂ17 (c 0.02, CH3OH). Anal.
ꢁ
Calcd for C18H20N2O2: C, 72.95; H, 6.80; N, 9.45; O, 10.80. Found:
C, 73.98; H, 6.79; N, 9.46; O, 10.77. Overall yield 22%.
5.2.10.5. (S)-100c: (S)-1-(4-Chlorobenzyl)-5-oxopyrrolidine-2-car-
boxyamide.
Waxy solid—1H NMR (300 MHz, DMSO-
5.2.14.1. Compound 3a.
Waxy solid—1H NMR (300 MHz,
d6 + D2O): d 1.77–1.85 (m, 1H), 2.14–2.41 (m, 3H), 3.74 (d, 1H,
J = 15.2 Hz), 3.89 (dd, 1H, J = 8.8, 3.91 Hz), 4.75 (d, 1H,
J = 15.2 Hz), 7.17 (d, 2H, J = 8.3 Hz), 7.35 (d, 2H, J = 8.3 Hz). LC–MS
(ESI) (H2O/CH3CN, 1%TFA) m/z (%): 253.1 (MH+, 100),
DMSO-d6): d 3.10 (s, 3H), 3.89–3.91 (m, 1H), 3.95–4.00 (m, 1H),
4.06–4.09 (m, 1H), 4.23–4.28 (m, 1H), 4.44–4.47 (m, 1H), 7.37–
7.70 (m, 9H), 9.88–9.99 (m, 1H), 11.26 (br s, 1H). LC–MS (ESI)
(H2O/CH3CN, 1%TFA) m/z (%): 315.2 (MH+, 100), C17H18N2O2S:
314.1.
C
12H13ClN2O2: 252.1. Specific optical rotation ½a D25
ꢁ
+55 (c 0.03,
CH3OH). Anal. Calcd for C12H13ClN2O2: C, 57.04; H, 5.19; Cl,
14.03; N, 11.09; O, 12.66. Found: C, 57.10; H, 5.20; Cl, 14.04; N,
11.11; O, 12.67. Overall yield 31%.
Specific optical rotation ½a D25
ꢂ8 (c 0.06, CH3OH). Anal. Calcd for
ꢁ
C
17H18N2O2S: C, 64.94; H, 5.77; N, 8.91; O, 10.18; S, 10.20. Found:
C, 64.71; H, 5.76; N, 8.93; O, 10.22; S, 10.25. Overall yield 20%.
5.2.11. General procedures for loading Fmoc-NHOH on
chlorotritylresin
5.2.14.2. 4a: (2S,4R)-N,4-Dihydroxy-1-benzyl(40-phenyl)-pyrroli-
Waxy solid—1H NMR (300 MHz, DMSO-
Example: Synthesis of 19.
A solution of Fmoc-NHOH (51 mg, 0.2 mmol) and DIPEA (70 ll,
dine-2-carboxamide.
d6): d1.99–2.10 (m, 1H), 2.15–2.27 (m, 1H), 3.03–3.17 (m, 1H), 3.55–
3.70 (m, 2H), 4.10–4.20 (m, 1H), 4.33–4.49 (m, 3H), 7.38–7.76
(m, 9H), 9.29 (br s, 1H), 11.06 (br s, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 313.2 (MH+, 100), C18H20N2O3: 312.5.
0.4 mmol) in 2 ml DCM was dropped on chlorotrityl resin (Nova-
biochem) (0.2 mmol) pre-swollen in DCM.28
After shaking for 60 h at room temperature, the resin was fil-
tered and rinsed with DCM (3 ꢀ 5 ml) DMF (3 ꢀ 5 ml), DCM
(3 ꢀ 5 ml) and TBME (3 ꢀ 5 ml) and finally dried in vacuo.
The efficiency of loading (loading yield: 53%) was determined
on 3 mg of resin by Fmoc removal analysis (treatment of the resin
with piperidine 20% in DMF) and absorbance measurement at
301 nm (UV–vis Spectrophotometer Shimadzu).
Specific optical rotation ½a D25
ꢂ2 (c 0.1, CH3OH). Anal Calcd for
ꢁ
C
18H20N2O3: C, 69.21; H, 6.45; N, 8.97; O, 15.37. Found: C, 69.10;
H, 6.43; N, 8.79; O, 15.40. Overall yield 27%.
5.2.15. General procedures for coupling of Fmoc-
to NovaSynTGOH resin
D-Cys-(trt)-OH
A solution of Fmoc-
(0.65 g, 1.7 mmol), and DIEA (600
DMF was dropped on the resin NovaSynTGOH (1.12 g, 0.34 mmol).
D
-Cys-(trt)-OH (1.02 g, 1.7 mmol), HATU
5.2.12. General procedures for coupling of proline/thioproline/
hydroxyproline rings to -NHOH functionalized resin
Example: Synthesis of 23.
ll, 3.4 mmol) in 3 ml of dry

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2335
DMAP (8.5 mg, 0.2 mmol) in 0.5 ml of dry DMF was added, and the
5.2.20. General procedures for Fmoc deprotection and N-
sulphonylation
cartridge was shaked at room temperature for 6 h.
After filtration of the solution the resin was rinsed with DMF
(3 ꢀ 5 ml), DCM (3 ꢀ 5 ml) and TBME (3 ꢀ 5 ml) and finally dried
in vacuo.
The efficiency of loading (loading yield: 80%) was determined
on 8.3 mg of resin by Fmoc removal analysis (treatment of the resin
with piperidine 20% in DMF) and absorbance measurement at
301 nm (UV–vis Spectrophotometer Shimadzu).
After Fmoc deprotection of (R)-34 with 20% piperidine in DMF
for 20 min (ꢀ2), the resin (70 mg, 0.07 mmol) was washed
(2 ꢀ DMF, 2 ꢀ DCM), and DIEA (0.41 mmol, 73
ll) and a solution
of bisphenylsulfonyl chloride (0.27 mmol, 68 mg) in 1.5 ml of dry
DCM (11 mg, 0.09 mmol) were added dropwise. After shaking
overnight at room temperature, the resin was filtered and rinsed
with DCM (ꢀ3), DMF (ꢀ3), DCM (ꢀ3) and TBME (ꢀ3).
5.2.21. General procedures for allyl deprotection and cyclization
5.2.16. General procedures for Fmoc deprotection and reductive
amination
The same procedure described for (R)-14 was followed.
The same procedure described for the synthesis of (R)-13 was
followed.
5.2.22. General procedures for cleavage of hydroxamates 6a
The crude product 6a was cleaved by treatment of 36a with
2.5 ml of TFA (5%) in DCM for 2 h at room temperature. The resin
was then filtered and filtrate was evaporated under N2.
5.2.17. General procedures for trityl deprotection and
cyclization
6a was further purified by semi-preparative RP-HPLC (C18,
Nucleosil 300 mm ꢀ 25 mm, 3 ml/min, detection at 230 nm, eluent
A: Water/TFA: 99.92:0.08, eluent B: acetonitrile/TFA: 99.92:0.08;
gradient B from 20% to 80% in 14 min, from 80% to 98% in 3 min).
Amorphous solid—1H NMR (300 MHz, DMSO-d6): d 1.84–1.97
(m, 1H), 2.23–2.49 (m, 3H), 4.67 (d, 1H, J = 8.1 Hz), 7.42–7.59 (m,
3H), 7.7 (d, 2H, J = 7.6 Hz), 7.91 (d, 2H J = 8.7 Hz), 8.04 (d, 2H
J = 8.7 Hz), 9.18 (s, 1H), 11 (s, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 361.2 (MH+, 75), C17H16N2O5S: 360.4. Specific opti-
Trityl deprotection of 31 (0.04 mmol pre-swollen in DCM) was
performed with TFA/triisopropyl silane (TIS)/DCM 10:5:85 (4 ꢀ
30 min). Then the resin was filtered, rinsed with DCM and a solu-
tion of 10% Et3N in DMF was added.
After filtering and washing (DMF, DCM), a solution of CDI
(60 mg, 0.37 mmol) dissolved in 1 ml dry DCM was introduced into
the cartridge; the cyclization proceeded for 6 h at room tempera-
ture. Functionalized resin 32 was then rinsed with DCM (ꢀ3),
DMF (ꢀ3), DCM (ꢀ3), and TBME (ꢀ3).
cal rotation ½a 2D5
ꢁ
+11 (c 0.02, CH3OH). Anal. Calcd for C17H16N2O5S:
C, 56.66; H, 4.47; N, 7.77; O, 22.20; S, 8.90. Found: C, 56.55; H, 4.49;
N, 7.80; O, 22.22; S, 8.92. Overall yield 12%.
5.2.18. General procedures for the cleavage of hydroxamate
derivatives 5a
The same procedure described for (R)-1a was followed.
Compound 5a: Amorphous solid—1H NMR (300 MHz, DMSO-d6):
d 3.19–3.24 (m, 1H), 3.60 (dd, 1H, J = 8.6, 10.9 Hz), 3.83 (d, 1H,
J = 15.1 Hz), 4.09–4.19 (m, 1H), 4.89 (d, 1H, J = 15.7 Hz), 7.28–
7.41 (m, 3H), 7.42–7.51 (m, 2H), 7.66 (d, 4H J = 8.1 Hz), 9.21 (s,
1H), 10.89 (s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%):
329.2 (MH+, 10), C17H18N2O6S2: 328.4. Specific optical rotation
5.2.23. General procedure for synthesis of series 7, 8, 9, 10
The synthesis of 7, 8 series was realized according to a common
parallel strategy except for the coupling to NovaSynTGOH resin:
N-Fmoc proline and N-Fmoc thioproline were loaded respectively
via HATU and via fluoride. The subsequent synthetic steps were
performed with the same experimental procedures.
½
a 2D5
ꢁ
ꢂ37 (c 0.03, CH3OH). Anal. Calcd for C17H16N2O3S: C, 62.18;
5.2.24. General procedures for coupling of N-Fmoc proline to
NovaSynTGOH
H, 4.91; N, 8.53; O, 14.62; S, 9.76. Found: C, 62.16; H, 4.87; N,
8.58; O, 14.64; S, 9.78. Overall yield 35%.
Example: Synthesis of 37.
A solution of Fmoc-
1 mmol), and DIPEA (348
(1:1) was dropped on the resin NovaSynTGOH (0.2 mmol).
L
-proline (0.34 g, 1 mmol), HATU (0.38 g,
l, 2 mmol) in 2 ml of dry DMF/DCM
5.2.19. General procedure for synthesis of Fmoc-
NHOH and loading on chlorotrityl resin
D-Glu(OAll)-
l
A solution of Fmoc-D-Glu-(OAll) chloride (obtained by the reac-
The cartridge was rocked 12 h at room temperature.
tion between Fmoc- -Glu-(OAll)-OH (0.32 g, 0.77 mmol) and SOCl2
D
After filtration the resin was rinsed with DMF (3 ꢀ 5 ml), DCM
(3 ꢀ 5 ml) and TBME (3 ꢀ 5 ml) and finally dried in vacuo.
The efficiency of loading (loading yield: 80%) was determined
on 8.3 mg of resin by Fmoc removal analysis (treatment of the resin
with piperidine 20% in DMF) and absorbance measurement at
301 nm (UV–vis Spectrophotometer Shimadzu).
(1.5 ml, 21 mmol) in dioxane for 2.5 h at 50 °C) in 1 ml of dioxane
was added to a freshly prepared solution of hydroxylamine, ob-
tained by reaction between hydroxylamine hydrochloride (0.27 g,
3.85 mmol) and DIEA (690 ll, 3.85 mmol) in dioxane/water. After
being stirred at room temperature overnight, the mixture was
evaporated under reduced pressure and DCM and HCl 0.1 M solu-
tion were added. The organic layer was separated, and the aqueous
one was extracted with DCM (2 ꢀ 5 mL). The combined organic
solutions were dried (MgSO4), and evaporated to dryness.
A solution of the pure compound 33 (95.7% 220 nm, 91.2%
254 nm) (0.15 mg, 0.35 mmol) and DIEA (165 ll, 0.92 mmol) in
5.2.25. General procedures for coupling of N-Fmoc thiazolidine
to NovaSynTGOH
Example: Synthesis of 38.
A solution of N-Fmoc thiazolidine (0.34 g, 1 mmol), HATU
(0.38 g, 1 mmol), and DIPEA (348 ll, 2 mmol) in 2 ml of dry DMF/
3 ml DMF/DCM (1:10) was added dropwise on chlorotrityl resin
(Novabiochem) (160 mg, 0.23 mmol) pre-swollen in DCM. After
shaking for 60 h at room temperature, the resin was filtered and
rinsed first with DCM/MeOH/DIEA (17:2:1) (3 ꢀ 5 min) and later
with DCM (3 ꢀ 5 ml), DMF (3 ꢀ 5 ml), DCM (3 ꢀ 5 ml) and TBME
(3 ꢀ 5 ml) and finally dried in vacuo. The efficiency of loading
(loading yield: 60%) was determined on 2.8 mg of resin by Fmoc re-
moval analysis (treatment of the resin with piperidine 20% in DMF)
and absorbance measurement at 301 nm (UV–vis Spectrophotom-
eter Shimadzu).
DCM (1:1) was added dropwise on was added to NovaSynTGOH
(0.2 mmol) after Fmoc deprotection via classical procedure (20%
Piperidine in DMF for 20 min (ꢀ2)). The cartridge was shaked for
12 h at room temperature.
After filtration the resin was rinsed with DMF (3 ꢀ 5 ml), DCM
(3 ꢀ 5 ml) and TBME (3 ꢀ 5 ml) and finally dried in vacuo.
The efficiency of loading (loading yield: 53%) was determined
on 3 mg of resin by Fmoc removal analysis (treatment of the resin
with piperidine 20% in DMF) and absorbance measurement at
301 nm (UV–vis Spectrophotometer Shimadzu).

2336
A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
5.2.26. General procedures for N-sulphonylation of polymer
bound proline/thioproline
23.86; S, 13.66. Found: C, 38.41; H, 3.01; F, 12.16; N, 8.98; O,
23.90; S, 13.63. Overall yield 25%.
Example: Synthesis of 40a.
After Fmoc deprotection with 20% piperidine in DMF for 20 min
(ꢀ2), 39 (0.03 mmol) was washed (2 ꢀ DMF, 2 ꢀ DCM) and pre-
swollen in 2 ml of THF in presence of DMAP (3.4 mg, 0.03 mmol).
A solution of bisphenylsulphonylchloride (0.14 mmol) in 1 ml of
dry DCM was added and the cartridge was rocked for 12 h at room
temperature.
5.2.27.5. Compound 8j.
Amorphous solid—1H NMR
(300 MHz, DMSO-d6): d 2.81–2.88 (m, 1H), 3.0 (dd, 1H, J = 5.4,
11.90 Hz), 4.39–4.49 (m, 2H), 4.75 (d, 1H, J = 10.8 Hz), 7–7.08 (m,
4H), 7.13 (d, 2H, J = 9.2 Hz), 7.87 (d, 2H, J = 8.7 Hz), 9.07 (br s,
2H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 411.1 (MH+, 45),
C
17H18N2O6S2: 410.5. Specific optical rotation ½a D25
ꢂ22 (c 0.1,
ꢁ
Then the resin was filtered and washed with MeOH (ꢀ2), DMF
CH3OH). Anal Calcd for C17H18N2O7S2: C, 47.88; H, 4.25; N, 6.57;
O, 26.26; S, 15.04. Found: C, 47.81; H, 4.26; N, 6.51; O, 26.23; S,
15.05. Overall yield 29%.
(ꢀ2), DCM (ꢀ2) and TBME (ꢀ2).
5.2.27. General procedures for cleavage of hydroxamate
derivatives 7, 8
5.2.27.6. Compound 8k.
Amorphous solid—1H NMR
Example: Synthesis of 8a.
Functionalized resin derivative 40a (0.03 mmol) was swollen in
THF (1.8 ml) and cleaved by solid support by treatment with 50%
NH2OH in H2O (300 ll, 5.6 mmol). After rocking for 5 h at room
temperature, the resin was filtered and the filtrate was evaporated
(300 MHz, DMSO-d6): d 2.70–2.80 (m, 1H), 2.98–3.05 (m, 1H),
4.40–4.52 (m, 2H), 4.72 (d, 1H, J = 10.3 Hz), 7.09–7.24 (m, 4H),
7.53 (d, 2H, J = 8.7 Hz), 7.91 (d, 2H, J = 8.7 Hz), 8.61–8.93 (br s,
2H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 415.1 (MH+, 90),
C16H15ClN2O5S2: 414.9. Specific optical rotation ½a D25
ꢂ21 (c 0.1,
ꢁ
to dryness.
CH3OH). Anal. Calcd for C16H15ClN2O6S2: C, 44.60; H, 3.51; Cl,
8.23; N, 6.50; O, 22.28; S, 14.88. Found: C, 45.58; H, 3.52; Cl,
8.21; N, 6.51; O, 22.26; S, 14.86. Overall yield 35%.
The crude product 8a was further purified by semi-preparative
RP-HPLC (C18, Nucleosil 300 mm ꢀ 25 mm, 3 ml/min, detection at
230 nm, eluent A: Water/TFA: 99.92:0.08, eluent B: acetonitrile/
TFA: 99.92:0.08; gradient B from 20% to 80% in 14 min, from 80%
to 98% in 3 min).
5.2.27.7. Compound 8l.
Amorphous solid—1H NMR
(300 MHz, DMSO-d6): d 2.70 (dd, 1H, J = 7.6, 10.8 Hz), 3.10 (dd,
1H, J = 3.2, 10.8 Hz), 4.36 (d, 1H, J = 10.3 Hz), 4.66–4.75 (m, 2H),
7.05–7.14 (m, 2H), 7.18–7.36 (m, 4H), 7.86–7.93 (m, 2H), 8.75 (s,
1H), 8.87 (s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%): 399.1
5.2.27.1. Compound 8a.
Amorphous solid—1H NMR
(300 MHz, DMSO-d6): d 2.83–2.93 (m, 1H), 2.97–3.04 (m, 1H),
4.46 (d, 1H, J = 10.8 Hz), 4.52 (dd, 1H, J = 5.9, 7.57 Hz), 4.85
(d, 1H, J = 10.8 Hz), 7.44–7.57 (m, 3H), 7.79 (d, 2H, J = 7.6 Hz),
7.91–8.03 (m, 4H), 9.05 (s, 1H), 10.87 (s, 1H). LC–MS (ESI) (H2O/
CH3CN, 1%TFA) m/z (%): 365.2 (MH+, 100), C16H16N2O4S2: 364.4.
(MH+, 100), C16H15FN2O5S2: 398.4. Specific optical rotation ½a 2D5
ꢁ
ꢂ18 (c 0.07, CH3OH). Anal Calcd for C16H15FN2O6S2: C, 46.37; H,
3.65; F, 4.58; N, 6.76; O, 23.16; S, 15.47. Found: C16H15FN2O6S2:
C, 46.42; H, 3.65; F, 4.57; N, 6.78; O, 23.20; S, 15.45. Overall yield
30%.
Specific optical rotation ½a D25
ꢂ125 (c 0.03, CH3OH). Anal. Calcd
ꢁ
for C16H16N2O5S: C, 50.51; H, 4.24; N, 7.36; O, 21.03; S, 16.86.
Found: C, 50.60; H, 4.25; N, 7.41; O, 21.06; S, 16.90. Overall yield
34%.
5.2.28. General procedures for synthesis of oxidized derivatives
41
Example: Synthesis of 41a.
5.2.27.2. Compound 7a.
Amorphous solid—1H NMR
MCPBA (31 mg, 0.17 mmol) in 4 ml of dry DCM was added to a
DCM pre-swollen 40a (0.13 mmol) at 0 °C. After rocking for 27 h at
room temperature, the resin was filtered and washed with DCM
(ꢀ2), DMF (ꢀ2), DCM (ꢀ2) and TBME (ꢀ2) and finally dried in
vacuo.
(300 MHz, DMSO-d6): d 1.45–1.58 (m, 1H), 1.67–1.93 (m, 3H),
3.14–3.26 (m, 1H), 3.38–3.49 (m, 1H), 3.98 (dd, 1H, J = 4.4,
8.12 Hz), 7.44–7.57 (m, 3H), 7.77 (d, 2H, J = 7.6 Hz), 7.93 (s, 4H),
8.95 (s, 1H), 10.70 (s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z
(%): 347.2 (MH+, 45), C17H18N2O4S: 346.4. Specific optical rotation
½
a 2D5
ꢁ
ꢂ110 (c 0.04, CH3OH). Anal Calcd for: C17H18N2O4S: C, 58.94;
5.2.29. General procedures for synthesis of oxidized derivatives
42
H, 5.24; N, 8.09; O, 18.48; S, 9.26. Found: C, 58.88; H, 5.21; N, 8.12;
O, 18.50; S, 9.21. Overall yield 38%.
Example: Synthesis of 42a.
MCPBA (37 mg, 0.213 mmol) in 4 ml of dry DCM was added to a
DCM pre-swollen 41a (0.04 mmol) at 0 °C. After rocking for 27 h at
room temperature, the resin was filtered and washed with DCM
(ꢀ2), DMF (ꢀ2), DCM (ꢀ2) and TBME (ꢀ2) and finally dried in
vacuo.
5.2.27.3. Compound 8h.
Amorphous solid—1H NMR
(300 MHz, DMSO-d6): d 2.75–2.83 (m, 1H), 2.98 (dd, 1H, J = 5.4,
11.36 Hz), 4.47 (d, 1H, J = 10.8), 4.57 (dd, 1H, J = 5.4, 7.03 Hz), 4.9
(d, 3H), 7.67–7.79 (m, 2H), 7.90–7.96 (m, 1H), 8.05–8.23 (m, 3H),
8.60 (s, 1H), 8.71 (s, 1H), 9.11 (s, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 339.2 (MH+, 100), C14H14N2O4S2: 338.4. Specific
5.2.30. General procedures for cleavage of hydroxamate
derivatives 9 and 10
optical rotation
14H14N2O5S2: C, 47.45; H, 3.98; N, 7.90; O, 22.57; S, 18.10. Found:
C, 47.42; H, 4.00; N, 7.99; O, 22.49; S, 18.12. Overall yield 35%.
½
a 2D5
ꢁ
ꢂ80 (c 0.06, CH3OH). Anal Calcd for
C
We follow the same procedure described for 8a.
5.2.30.1. Compound 9a.
Amorphous solid—1H NMR
5.2.27.4. 8i-TFA salt.
Amorphous solid—1H NMR (300 MHz,
(300 MHz, DMSO-d6): d 2.96 (dd, 1H, J = 3.8, 13.5 Hz), 3.17 (dd,
1H, J = 8.7, 13.5 Hz), 4.17 (d, 1H, J = 11.9 Hz), 4.72 (dd,1H, J = 3.8,
8.7 Hz), 5.30 (d, 1H, J = 11.9 Hz), 7.42–7.58 (m, 3H), 7.79 (d, 2H,
J = 7.0 Hz), 7.97 (m, 4H), 9.12 (s, 1H), 10.89 (s, 1H). LC–MS (ESI)
(H2O/CH3CN, 1%TFA) m/z (%): 381.1 (MH+, 100), C16H16N2O5S2:
DMSO-d6): d 2.80–2.89 (m, 1H), 2.99 (dd, 1H, J = 3.8, 11.36 Hz),
4.44 (d, 1H, J = 9.7), 4.95 (d, 1H, J = 9.7 Hz), 5.27 (dd, 1H, J = 4.3,
7.6 Hz), 7.72–7.84 (m, 2H), 8.37 (d, 1H, J = 8.1 Hz), 8.46 (d, 1H,
J = 7.0 Hz), 8.58 (d, 1H, J = 8.1 Hz), 9.09 (d, 1H, J = 4.3 Hz), 10.75
(s, 1H), 11.05 (s, 1H), 12.19 (s, 1H). LC–MS (ESI) (H2O/CH3CN,
1%TFA) m/z (%): 340.2 (MH+ꢂCF3CO2ꢂ, 100), C15H14F3N3O6S2:
380.4. Specific optical rotation ½a D25
ꢂ159 (c 0.07, CH3OH). Anal.
ꢁ
Calcd for C16H16N2O6S2: C, 48.47; H, 4.07; N, 7.07; O, 24.21; S,
16.18. Found: C, 48.35; H, 4.06; N, 7.05; O, 24.16; S, 16.14. Overall
yield 17%.
453.4. Specific optical rotation ½a D25
ꢂ66 (c 0.05, CH3OH). Anal
ꢁ
Calcd for C15H14F3N3O7S2: C, 38.38; H, 3.01; F, 12.14; N, 8.95; O,

A. Topai et al. / Bioorg. Med. Chem. 20 (2012) 2323–2337
2337
5.2.30.2. Compound 10a.
Amorphous solid—1H NMR
measurements T0 and T3h, using a range of test compound con-
centrations (at least six concentrations, each one in duplicate).
From the resulting dose–response curve IC50 values (the
concentration of inhibitor required to inhibit 50% of the enzyme
activity) were calculated from a three-parameter logistic fit of
the data (GraphPad Prism version 4.00 for Windows, GraphPad
Software, San Diego California USA).
(300 MHz, DMSO-d6): d 3.51–3.64 (m, 2H), 4.30 (d, 1H, J =
12.4 Hz), 4.91 (d, 1H, J = 12.4 Hz), 5.10 (dd, 1H, J = 4.9, 8.7 Hz),
7.42–7.58 (m, 3H), 7.78 (d, 2H, J = 7.0 Hz), 7.96 (m, 4H), 9.12
(s, 1H), 10.89 (s, 1H). LC–MS (ESI) (H2O/CH3CN, 1%TFA) m/z (%):
397.2 (MH+, 100), C16H16N2O6S2: 396.4 Specific optical rotation
½
a 2D5
ꢁ
ꢂ46 (c 0.05, CH3OH). Anal. Calcd for C16H16N2O7S2: C, 46.59;
H, 3.91; N, 6.79; O, 27.15; S, 15.55. Found: C, 46.65; H, 3.90; N,
6.80; O, 27.21; S, 15.59. Overall yield 12%.
Acknowledgments
5.2.30.3. Compound 10k.
Amorphous solid—1H NMR
The authors thank Drs. Tatiana Guzzo and Teresa Fabiola
Miscioscia for their essential support. This work was financed by
MIUR (Ministero dell’Istruzione, dell’Università e della Ricerca).
(300 MHz, DMSO-d6): d 3.24–3.33 (m, 1H), 3.51–6.66 (m, 1H),
4.41 (d, 1H, J = 12.5 Hz), 4.64 (dd, 1H, J = 6.5, 8.12 Hz), 4.95 (d,
1H, J = 13.0 Hz), 7.13–7.24 (m, 4H), 7.52 (d, 2H, J = 9.2 Hz), 7.94
(d, 4H, J = 9.2 Hz), 9.27 (s, 1H), 11.03 (s, 1H). LC–MS (ESI) (H2O/
CH3CN, 1%TFA) m/z (%): 447.1 (MH+, 60), C16H15ClN2O7S2: 446.8.
References and notes
Specific optical rotation ½a D25
ꢂ60 (c 0.05, CH3OH). Anal Calcd for
ꢁ
1. Coussens, L. M.; Werb, Z. Chem. Biol. 1996, 3, 895.
2. Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing, A. J. H. Chem. Rev. 1999, 99,
2735.
C
16H15ClN2O8S2: C, 41.52; H, 3.27; Cl, 7.66; N, 6.05; O, 27.65; S,
13.85. Found: C, 41.48; H, 3.26; Cl, 7.63; N, 6.03; O, 27.59; S,
13.81. Overall yield 9%.
3. Baker, A. H.; Edwards, D. R.; Murphy, G. J. Cell. Sci. 2002, 115, 3719.
4. Li, N. G.; Shi, Z. H.; Tang, Y. P.; Duan, J. A. Curr. Med. Chem. 2009, 16, 3805.
5. Aureli, L.; Gioia, M.; Cerbara, I.; Monaco, F.; Fasciglione, G. F.; Marini, S.;
Ascenzi, P.; Topai, A.; Coletta, M. Curr. Med. Chem. 2008, 15, 2192.
6. Oshitari, T.; Okuyama, Y.; Miyata, Y.; Kosano, H.; Takahashi, H.; Natsugari, H.
Bioorg. Med. Chem. 2011, 19, 7085.
5.3. Enzymatic assay
7. Wang, J.; Medina, C.; Radomski, M. W.; Gilmer, J. F. Bioorg. Med. Chem. 2011, 19,
4985.
8. Dorman, G.; Kocsis-Szommer, K.; Spadoni, C.; Ferdinandy, P. Recent Pat.
Cardiovasc. Drug Discov. 2007, 2, 186.
9. Renkiewicz, R.; Qiu, L.; Lesch, C.; Sun, X.; Devalaraja, R.; Cody, T.; Kaldjian, E.;
Welgus, H.; Bagagi, V. Arthritis Rheum. 2003, 48, 1742.
The enzymatic assays for MMP1, 2, 3, 8, 9, 12, and 14 were
based upon the protocol described by Knight et al.29 Active cata-
lytic domains of MMP1, 3, 8, 9, 12, and 14 were purchased from
Biomol (Biomol International, Plymouth Meeting) while proMMP2
(from R&D Systems, Inc.) was activated by incubation for 1 h at
37 °C with aminophenylmercuric acetate (APMA) 1 mM final con-
centration solution.
All enzymatic solutions were then diluted in Tris–HCl assay buf-
fer (Tris/HCl 50 mM, NaCl 0.1 M, CaCl2 10 mM, 0.05% Brij 35, pH
7.5) in order to obtain final concentrations of the enzymes used
in the assay included between 0.5 and 3 nM.
10. Pavlaki, M.; Zucker, S. Cancer Met. Rev. 2003, 22, 177.
11. Coussens, L. M.; Fingleton, B.; Matrisian, L. M. Science 2002, 295, 2387.
12. Overall, C. M.; Kleifeld, O. Nat. Rev. Cancer 2006, 6, 227.
13. Nuti, E.; Tuccinardi, T.; Rossello, A. Curr. Pharm. Des. 2007, 13, 2087.
14. Haas, T. L. Can. J. Physiol. Pharmacol. 2005, 83, 1.
15. Sluijter, J. P.; de Kleijn, D. P.; Pasterkamp, G. Cardiovasc. Res. 2006, 69, 595.
16. Deryugina, E. I.; Quigley, J. P. Cancer Metastasis Rev. 2006, 25, 9.
17. Overall, C. M.; Kleifeld, O. Br. J. Cancer 2006, 94, 941.
18. Higashi, S.; Miyazaki, K. Biol. Chem. 2008, 283, 10068.
19. Berman, H. M.; Westbrook, J.; Feng, Z.; Gilliland, G.; Bhat, T. N.; Weissig, H.;
Shindyalov, I. N.; Bourne, P. E. Nucleic Acids Res. 2000, 28, 235.
20. Shindyalov, I. N.; Bourne, P. E. Protein Eng. 1998, 11, 739.
21. MOE: Chemical Computing Group Inc. (2003): 1010 Sherbrooke Street West,
Suite 910, Montreal, H3A 2R7, Canada.
22. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J. Adv. Drug Deliv. Rev.
2001, 46, 3.
23. Xu, J.; Stevenson, J. J. Chem. Inf. Comput. Sci. 2000, 40, 1177.
24. Oprea, T. I. J. Comput. Aided Mol. Des. 2000, 14, 251.
25. Kiyama, R.; Tamura, Y.; Watanabe, F.; Tsuzuki, H.; Ohtani, M.; Yodo, M. J. Med.
Chem. 1999, 42, 1723.
The fluorogenic substrate was Mca-Pro-Leu-Gly-Leu-Dpa-Ala-
30
Arg-NH2
(Biomol Research Laboratories, Inc.) for all tested
MMP. The final substrate concentrations performed in assays con-
ditions were in the range of 2–7 M.
l
Test compounds were dissolved in dimethyl sulfoxide (DMSO)
and diluted with test buffer solution so that no more than 1%
DMSO was present. Test compound and enzymes were added to
each well of a 96-well plate and allowed to equilibrate for 1 h at
37 °C prior to the addition of substrate. Determinations of fluores-
cence was performed using a Varian Cary Eclipse spectrofluorime-
ter (kex 328 nm, kem 393 nm), respectively after addiction of
fluorogenic substrate (T0) and after 3 h incubation at 37 °C (T3h).
Inhibitors potency was evaluated from the amount of substrate
cleavage calculated from fluorescence delta between the two
26. DeLano, W.L. The PyMOL molecular graphics system. 2002 (http://
www.pymol.org).
27. Kaiser, E.; Colescott, R. L.; Bossinger, C. D.; Cook, P. I. Anal. Biochem. 1970, 34, 595.
28. Mellor, S. L.; McGuire, C.; Chan, W. C. Tetrahedron Lett. 1997, 38, 3311.
29. Knight, C. G.; Willenbrock, F.; Murphy, G. FEBS Lett. 1992, 296, 263.
30. Will, H.; Atkinson, S. J.; Butler, G. S.; Smith, B.; Murphy, G. J. Biol. Chem. 1996,
271, 17119.