Design, synthesis and biological evaluation of 5-hydroxy, 5-substituted-pyrimidine-2,4,6-triones as potent inhibitors of gelatinases MMP-2 and MMP-9

Article abstract of DOI:10.1016/j.ejmech.2012.09.036

Matrix metalloproteinases (MMPs) are attractive biological targets that play a key role in many physiopathological processes such as degradation of extracellular matrix proteins, release and cleavage of cell-surface receptors, tumour progression, homeostatic regulation and innate immunity. A series of 5-hydroxy, 5-substituted pyrimidine-2,4,6-triones were rationally designed, prepared and tested as inhibitors of gelatinases MMP-2 and MMP-9 and collagenase MMP-8. On one side, the presence of the 5-hydroxyl group, that represents an typical feature of this class of compounds, ensured an attractive pharmacokinetic profile while on the other suitably substituted biaryl molecular fragments, attached to position 5 through a ketomethylene linker, guaranteed favourable interaction in the deep region of the S₁′ enzymatic subsite. This rational design led to the discovery of highly potent MMP inhibitors. In particular, biphenyl derivatives bearing at the para position COCH₃ and OCF₃ substituents permitted to inhibit gelatinases MMP-2 and MMP-9, with IC₅₀ values as low as 30 nM and 21 nM, respectively, whereas the introduction at the same position of the bulkier SO₂CH₃ group afforded a potent collagenase MMP-8 inhibitor with an IC₅₀ value equal to 66 nM. Molecular docking simulations allowed us to elucidate key interactions driving the binding of the top active compounds towards their preferred MMP target.

Full text of DOI:10.1016/j.ejmech.2012.09.036

European Journal of Medicinal Chemistry 58 (2012) 368e376
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Design, synthesis and biological evaluation of 5-hydroxy, 5-substituted-
pyrimidine-2,4,6-triones as potent inhibitors of gelatinases MMP-2 and MMP-9
**
Orazio Nicolotti , Marco Catto, Ilenia Giangreco, Maria Barletta, Francesco Leonetti, Angela Stefanachi,
*
Leonardo Pisani, Saverio Cellamare, Paolo Tortorella, Fulvio Loiodice, Angelo Carotti
Dipartimento di Farmacia, Università degli Studi di Bari “Aldo Moro”, via Orabona 4, I-70125 Bari, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 July 2012
Received in revised form
7 September 2012
Accepted 25 September 2012
Available online 5 October 2012
Matrix metalloproteinases (MMPs) are attractive biological targets that play a key role in many phys-
iopathological processes such as degradation of extracellular matrix proteins, release and cleavage of
cell-surface receptors, tumour progression, homeostatic regulation and innate immunity. A series of 5-
hydroxy, 5-substituted pyrimidine-2,4,6-triones were rationally designed, prepared and tested as
inhibitors of gelatinases MMP-2 and MMP-9 and collagenase MMP-8. On one side, the presence of the 5-
hydroxyl group, that represents an typical feature of this class of compounds, ensured an attractive
pharmacokinetic profile while on the other suitably substituted biaryl molecular fragments, attached to
position 5 through a ketomethylene linker, guaranteed favourable interaction in the deep region of the
Keywords:
MMP
Gelatinases
ZBG
QSAR
Molecular docking
0
S1 enzymatic subsite. This rational design led to the discovery of highly potent MMP inhibitors. In
particular, biphenyl derivatives bearing at the para position COCH3 and OCF3 substituents permitted to
inhibit gelatinases MMP-2 and MMP-9, with IC50 values as low as 30 nM and 21 nM, respectively,
whereas the introduction at the same position of the bulkier SO2CH3 group afforded a potent collagenase
MMP-8 inhibitor with an IC50 value equal to 66 nM. Molecular docking simulations allowed us to
elucidate key interactions driving the binding of the top active compounds towards their preferred MMP
target.
Ó 2012 Elsevier Masson SAS. All rights reserved.
1. Introduction
identified. Most of them contain the hydroxamate function, which
strongly interacts with the zinc ion in the MMP catalytic site, and
The great interest towards matrix metalloproteinases (MMPs),
a family of zinc- and calcium-dependent endopeptidases, stems
from their primary role in the degradation of the extracellular
matrix (ECM) [1]. Overexpression and/or misregulation of MMPs
may trigger various inflammatory, malignant, and degenerative
diseases [2e5]. Selective MMP inhibition is therefore an important
goal in medicinal chemistry research to tackle severe and wide-
spread diseases [6,7].
MMP-2 and MMP-9 have a relevant role in the pathogenesis of
cancer, heart disease and inflammation and indeed the major focus
of the present investigation was the inhibition of these two gelat-
inases characterized by fibronectin-like inserts enabling them to
bind and process gelatine. Nowadays, a high number of orally
available broad-spectrum MMP inhibitors (MMPIs) have been
thus are often non selective compounds. Moreover, concerns have
been raised about the true clinical efficacy and significant toxicity of
this class of MMP inhibitors [8]. As a result, increasing attention has
been paid to the identification of new zinc binding groups (ZBGs)
that could represent a valuable alternative to the hydroxamate
function [9e13].
In this context, our effort has been devoted to the design of
novel barbiturate derivatives to be further developed as potent
MMPIs. A large number of 5,5-disubstituted pyrimidine-2,4,6-
triones, exhibiting high affinities over several MMP subtypes [14e
17], has been already reported and more recently they have been
studied also as putative radiotracers to non-invasively locate acti-
vated MMPs by means of positron emission tomography (PET)
[18,19]. Moreover, the very attractive drug-likeness [20] of barbi-
turates compared to hydroxamates makes them good starting
structures for the design of more promising and safer MMPIs. In
this respect, the need of achieving a good balance between binding
potency and pharmacokinetic properties prompted us to insert at
the C5 carbon two quite different substituents, one belonging to
* Corresponding author. Tel.: þ39 080 5442782; fax: þ39 080 5442230.
** Corresponding author. Tel.: þ39 080 5442551; fax: þ39 080 5442230.
E-mail addresses: orazio.nicolotti@uniba.it (O. Nicolotti), angelo.carotti@uniba.it
(A. Carotti).
0223-5234/$ e see front matter Ó 2012 Elsevier Masson SAS. All rights reserved.
http://dx.doi.org/10.1016/j.ejmech.2012.09.036

O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
369
a series of variously decorated biphenyl arms to explore the binding
interactions at the S10 subsite while the other was a hydroxyl group
to modify hydrophobic properties and ensure a fair pharmacoki-
netic profile, quite diverse from the barbiturates explored so far
carrying generally two highly hydrophobic substituents at position
5.
dynamics (MD) simulations and free energy perturbation calcula-
tions [21]. The encouraging binding affinity observed for the para-
carboxamido analogue (16; IC50 ¼ 87 nM at MMP-2) prompted us
to further explore the para position with a series of substituents
spanning a wider chemical space in terms of Hammett
s and
Hansch substituent constants [25]. Satisfactorily, our approach
p
Our molecular design was also inspired by findings from an our
recent paper [21], where we have examined a number of ZBGs (e.g.,
barbiturates, hydroxypyrones and hydroxypyridinones) and their
potential in biasing selectively MMPs [22]. In particular, we
described how the chelation of the catalytic zinc ion by the
pyrimidine-2,4,6-trione ring could be further reinforced by the
specific occurrence of additional interactions, including hydrogen
bonds and van der Waals (vdW), that may determine a consistent
gain of molecular affinity and selectivity [21]. Originally, such
a study started from the analysis of our electronic database
archiving a wealth of physicochemical and biochemical information
of a molecular library containing >2000 chemicals prepared in our
laboratories [21,23] that led us to find a number of 5-hydroxy, 5-
substituted-pyrimidine-2,4,6-triones prepared as chemical inter-
mediates for the synthesis of pyrimido[4,5-c]pyridazine inhibitors
of monoamine oxidases [24]. Given the weak activity towards
MMPs measured for the fished hits (1e4 of Table 1), the phenyl
moiety of such compounds was augmented to biphenyl, a privi-
led to the discovery of additional interesting compounds with
inhibitory activity towards MMP-2 comparable to or higher than
reference compound 16 and remarkably increased towards MMP-9.
Besides the para position, other structural variations were made to
explicitly target other sensitive enzymatic regions. Quantitative
structureeactivity relationships and docking studies allowed the
identification of the main molecular determinants responsible for
the high affinity showed by our derivatives.
2. Results
2.1. Chemistry
5-Hydroxy, 5-substituted pyrimidine-2,4,6-triones were
prepared via aldol addition of suitable methyl ketones to pyrimi-
dine-2,4,6-trione (i.e., alloxan) as already described [24]. When not
commercially available, substituted biaryl methyl ketones I were
prepared via microwave assisted SuzukieMiyaura reaction of 4-
bromoacetophenone and arylboronic acids (Scheme 1a). Reaction
of alloxan hydrate with biaryl methyl ketones I yielded comps 7e
15, 17 and 18 (Scheme 1a) whereas a mixture of isomeric
compounds 19 and 20 was obtained in the reaction with biphenyl
acetone (Scheme 1b).
Biological assays. MMP-2, -8, -9 inhibitory activities were
determined by a fluorimetric assay, where a pro-fluorescing
peptide was used as substrate, and the fluorogenic activity of its
cleavage product was measured after coincubation with MMP of
the tested compounds.
0
leged fragment for targeting the S1 subsite of MMPs. In agreement
with previous findings, this knowledge-based approach has
permitted to obtain a number of compounds (i.e., 5, 6 and 16)
inhibiting the gelatinases MMP-2 and MMP-9 and to a lesser extent
the collagenase MMP-8. The binding interactions of inhibitors 5, 6
and 16 have been deeply investigated by means of molecular
Table 1
MMP-2, MMP-9 and MMP-8 inhibitory activities of compounds 1e20.
H
O
N
O
O
2.2. Molecular modelling studies
HN
The program GOLD 5.1 was used in all the docking simulations
[26]. In the present investigation, GOLD returned a confident reli-
ability as it well reproduces the binding conformation of different
molecules co-crystallized with MMP proteins. As a measure of fit,
the rmsd value was calculated over all the heavy atoms between
the docking pose and the X-ray conformation of ligands in complex
with the two gelatinases MMP-2 (pdbcode 1HOV) and MMP-9
(pdbcode 2OVX) and the collagenase MMP-8 (pdbcode 1I73) and
deviations as small as 1.04 A, 1.24 A and 1.70 A, were observed.
The program QikProp (QikProp, version 3.4, Schrödinger, LLC,
New York, NY, 2011) was used to predict a number of pharmaceu-
tically relevant properties (e.g., octanol/water partition coefficient,
log P; aqueous solubility, log S; and binding to human serum
albumin, log Khsa) for five reference compounds [14,17,18,20] in
comparison with inhibitors 5, 8e17 of Table 1. As a measure of
goodness, the interested user can assess the high concordance
between measured (5.6 ꢀ 10ꢁ4 g/mL) and predicted (4.3 ꢀ 10ꢁ4 g/
mL) aqueous solubility for a compound elsewhere described [20].
OH
O
R
Entry
R
Gelatinases
Collagenase
MMP-8a
MMP-2a
MMP-9a
1b
2b
3b
4b
5b
6b
7
H
CF3
OCH3
NO2
(0%)
44,200 ꢂ 1600
(11%)
13,000 ꢂ 400
42,900 ꢂ 700
8200 ꢂ 100
6800 ꢂ 900
63,000 ꢂ 1000
266 ꢂ 36
39,000 ꢂ 700
21,200 ꢂ 400
(0%)
ꢀ
ꢀ
ꢀ
(8%)
517 ꢂ 10
C6H5
184 ꢂ 24
C6H4-30-Cl
C6H4-30-NO2
C6H4-40-CH3
C6H4-40-Cl
60,000 ꢂ 4000 19,800 ꢂ 3100
4500 ꢂ 100
8100 ꢂ 900
301 ꢂ 13
278 ꢂ 26
148 ꢂ 40
121 ꢂ 47
143 ꢂ 3
120 ꢂ 8
58 ꢂ 2
27,500 ꢂ 14,800 2600 ꢂ 800
8
9
103 ꢂ 30
66 ꢂ 12
561 ꢂ 82
141 ꢂ 16
21 ꢂ 1
336 ꢂ 37
271 ꢂ 43
210 ꢂ 8
184 ꢂ 42
715 ꢂ 100
313 ꢂ 38
66 ꢂ 11
109 ꢂ 20
586 ꢂ 91
2020 ꢂ 270
153 ꢂ 23
26%d
10
11
12
13
14
15
16b
17
18
19
20
C6H4-40-CH2OH
C6H4-40-OCH3
C6H4-40-OCF3
C6H4-40-SCH3
C6H4-40-SO2CH3
C6H4-40-COCH3
C6H4-40-CONH2
86 ꢂ 27
650 ꢂ 120
170 ꢂ 30
1800 ꢂ 210
3010 ꢂ 650
159 ꢂ 24
21%d
30 ꢂ 1
3. Results and discussion
87 ꢂ 24
C6H4-40-CON(CH3)2 1340 ꢂ 60
Thien-3-yl
181 ꢂ 8
The need of discovering novel ZBG-containing compounds
prompted us to first interrogate our electronic database storing
a chemical library of >2000 rationally designed biological active
compounds synthesized since late eighties in the laboratories of our
group. As shown in Fig. 1, the likelihood of our library to exhibit
activity in any therapeutic area has been estimated by comparing its
bioactiveprofiles (i.e., distribution of physicalproperties relevant for
10%d
c
38%d
28%d
71%d
c
a
Inhibition data are expressed as IC50 (nM) ꢂ SEM or % inhibition at 10
mM (in
parentheses) from 2 or 3 different experiments.
b
Data reported in Ref. [21].
Structures in Scheme 1b.
c
d
% inhibition at 100 mM concentration.

370
O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
Scheme 1. Reagents and conditions: (i) Na2CO3, (PPh3)2PdCl2, water/dioxane, microwave heating; (ii) acetic acid, reflux.
druglikeness such as molecular weight, log P, number of hydrogen
bond donor atoms and number of hydrogen bond acceptor atoms)
with those derived from DrugBank [27], a freely available database
containing FDA-approved small molecule drugs, FDA-approved
biotech drugs, nutraceuticals and experimental drugs.
(1e4; Table 1) with MMP inhibition potency in the high micromolar
range. The inhibitory potency of the selected pyrimidine-2,4,6-
trione derivatives has been then improved through a knowledge-
based molecular design rooted on the elementary concept of
assembling the pyrimidine-based core with a biphenyl arm,
0
Interestingly, besides a large series of compounds bearing
canonical ZBGs (i.e., carboxylic acids and thiols), four original 5,5-
disubstituted pyrimidine-2,4,6-triones have been also fished as
potential hits. The rationale behind such a selection has been based
on previously reported studies available in the literature [14e19].
Our attention has been focused on the 5-OH group featuring these
compounds because of its potential role in determining a more
appropriate hydrophilic/lipophilic balance and in engaging HB
interactions. In this respect, the four 5,5-disubstituted pyrimidine-
2,4,6-triones retrieved from our collection have been submitted to
docking calculations, and the plausible binding poses found in the
MMP-2 enzyme increased our confidence in raising up these
compounds as MMP-2 inhibitors [21]. Interestingly, biological
experimental data awarded a number of these selected derivatives
a fragment widely used to target the S1 subsite of MMPs. Our
efforts resulted three barbiturate-based derivatives (5, 6 and 16;
Table 1 [21]) with two of them showing interesting binding
potential towards MMPs. In that investigation, our attention has
been mostly paid to interpret, via MD and free energy perturbation
techniques, the dramatic difference of the binding affinities towards
MMP-2 between compounds 6 and 16, IC50 60,000 vs. 87 nM,
respectively, despitea slightstructural change. Theinterested reader
is referred to elsewhere for a complete description [21].
In the present study, we prefer to better explore the para posi-
tion of the biphenyl ring with the purpose to improve potency
towards MMP-2 while minor emphasis has been placed at this
stage in studying the substituent changes at the meta position given
the disappointing results recently obtained by compound 6. In
Fig. 1. The physical property profiles of our academic collection (in black) are compared with property profiles found in DrugBank (in white). The y-axis represents the percentage of
compounds, the x-axis represents the molecular weights (MW), log P, hydrogen bond donor (HBD) and hydrogen bond acceptors (HBA) profiles.

O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
371
doing so, we prepared a panel of variously substituted barbiturate-
based derivatives (7e15 and 17) by decorating the distal phenyl
ring with selected functional groups having easy commercial and
pIC50;MMPꢁ2 ¼6:55ðꢂ0:18Þþ0:404ðꢂ0:28Þsp þ0:350ðꢂ0:22ÞMR
s ¼ 0:225 n ¼ 10 r2 ¼ 0:565
ð2Þ
chemical access as well as different Hammett
s and Hansch p
However, despite the poor statistics recommended great
caution, equation (2) seemed to suggest the occurrence of polar
interactions in a non-hydrophobic enzymatic region, accounted for
by MR, whose strength is favoured by the EWG character of the para
substituent constants. Finally, isosteric 3-thienyl derivative 18 was
also prepared along with two additional derivatives 19 and 20
carrying longer and branched biphenyl-containing molecular
fragments at position 5, respectively. For the sake of completeness,
data for derivatives 1e6 and 16 already described in our recent
work [21] are also listed in Table 1.
substituents, as assessed by
s (vide infra).
A clearer and more consistent trend was observed when
modelling the MMP-9 binding affinity that, more specifically,
benefited of the presence of EWG with more pronounced hydro-
phobic character. In this respect, the most active compounds were
As shown in the Craig plot in Fig. 2, the physicochemical land-
scape was fairly explored changing the
s and p coordinates by
selecting a pool of suitable substituents. In particular, a number of
easily accessible functional groups was chosen to assess their effect
in targeting MMPs. Interestingly, it was observed that MMP-2
inhibitory activity is benefited by the presence of electron with-
drawing groups (EWG) while the hydrophobic effects are less
those combining high values for both s and p substituent constants.
Striking examples are the para-OCF3 (12) and the para-Cl (9)
derivatives inhibiting MMP-9 with IC50 values equal to 21 and
66 nM, respectively. Interestingly, the multivariate analysis of
inhibition data from parent compound and para derivatives (5, 8e
17) returned the following informative equation:
relevant although low values of
p are to some extent welcome. The
shift from the para-CONH2 (16) to the para-COCH3 (15) EWG
resulted an increase of affinity towards MMP-2 from 87 to 30 nM
pIC50;MMPꢁ9 ¼ 6:73ðꢂ0:09Þ þ 0:504ðꢂ0:29Þsp þ 0:617ðꢂ0:10Þ
p
IC50 likely due to an increment of
s from 0.36 to 0.50. Similarly,
changing the para-CONH2 (16) to the para-SO2CH3 EWG (14)
improved affinity towards MMP-2 from 87 to 58 nM following
s ¼ 0:243 n ¼ 10 r2 ¼ 0:861
ð3Þ
a variation of the
s
value from 0.36 to 0.72. On the other hand, the
showing a high r2 and even a confident value of q2, that is the
square correlation coefficient for prediction from the leave-one-out
procedure [34], as high as 0.738. Importantly, equation (3) explicitly
indicated the occurrence of favourable lipophilic interactions of the
para-substituents in the MMP-9 binding site.
replacement of para-CONH2 (16) EWG with a para-OCH3 electron
donor groups (EDG) (11) had a detrimental impact on the affinity
towards MMP-2. Focussing on parent compound and para deriva-
tives (5, 8e17), some linear tendency emerged by relating binding
affinity towards MMP-2 with both
s
and
p
constants, as follows:
Interestingly, this binding hypothesis and the additional infor-
mation coming from the QSAR equations were further confirmed
and complemented by the analysis of the structural features of the
MMP-2 and MMP-9 binding sites, in particular the S1 subsite [35],
and, more importantly, by docking simulations of selected inhibi-
tors into their binding pockets.
The informative QSAR models obtained for gelatinases MMP-2
and MMP-9 prompted us to extend the investigation to the MMP-
8, which is instead a member of the collagenases. Unfortunately,
the examination of the same series of para derivatives did not
furnish a clear picture of the physicochemical forces driving the
inhibitors binding to MMP-8. Given the uneven distribution and
narrow spread of pIC50 values, very disappointingly, a poor corre-
lation (r2 < 0.250) was observed by using the same substituents
parameters applied for modelling MMP-2 and MMP-9 inhibitory
activities in equations (1)e(3).
pIC50;MMPꢁ2 ¼ 6:82ðꢂ0:09Þ þ 0:458ðꢂ0:28Þsp ꢁ 0:116ðꢂ0:09Þ
p
0
s ¼ 0:237 n ¼ 10 r2 ¼ 0:520
ð1Þ
In this and in the following equations, s, n and r2 represent the
standard deviation, the number of the examined compounds and
the squared correlation coefficient, respectively. Normally, a nega-
tive coefficient with
p constant would indicate unfavourable
interactions in a non-hydrophobic or polar local environment [28].
In these cases, as amply demonstrated in past QSAR studies [29e
33], the use of the MR, that is the substituent molar refractivity
[25], allows to deriving more solid and interpretable QSAR models.
Actually, the replacement of
p with MR in equation (1) yielded
equation (2) having slightly improved statistical parameters.
To further extend our structureeactivity relationship study,
three more compounds were designed, prepared and tested. As the
isosteric substitution of one phenyl with a 3-thienyl ring in
a biphenyl moiety was successfully applied to HDAC inhibitors
affording much more potent compounds [36,37], the same
replacement was made on lead compound 5 affording compound
18. Very satisfactorily, this compound exhibited an equal inhibitory
potency at MMP-2 and a significantly higher potency at both MMP-
9 and MMP-8 compared to the phenyl isoster 5.
To better explore the binding region around the 5 position of the
pyrimidine-2,4,6-trione scaffold, two new compounds, that is
compounds 19 and 20 in Table 1, were designed and contempo-
rarily prepared through the single reaction reported in Scheme 1b.
Compound 19 was conceived to increase the length and the
conformational mobility of the biaryl arm whereas compound 20 to
look for additional favourable interactions around the keto-
methylene group. Unfortunately, both compounds displayed very
weak inhibition towards the three tested MMPs (Table 1).
Docking studies were undertaken to identify at a 3D level the
driving forces governing the interaction [38,39] at the MMP binding
sites of our inhibitors. Initial coordinates for the two gelatinases,
Fig. 2. Choice of substituents at the para position of the biphenyl fragment by
multidimensional mapping based on the Hammett (s) and Hansch (p) constants.

372
O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
MMP-2 and MMP-9, were taken from structural models deposited in
the Protein Data Bank (PDB) with the entry code 1HOV and 2OVX,
respectively. The PDB entry 1I73 was instead selected for docking
studies of the collagenase MMP-8. A wealth of information was
gathered by inspecting other available PDB structures of MMPs
complexed with barbiturate inhibitors [40,41]. On this basis, a model
where the N1 atom of the barbiturate coordinates the metal ionwith
atetrahedralgeometry, andtheoxygenatom interactsviaanHBwith
the side chain of Glu121 (numbering as in 1HOV) has been accepted
as the most reasonable. As a result, the enolic form of the barbiturate
is supposed to be the favoured tautomer in protein context over the
keto form that is instead prevalent in solution [42]. An additional HB
is observed between carbonyl oxygen atom of C4 and Ala84
(numbering as in 1HOV) that is part of the conserved AHA motif S1 of
the MMP binding site [43,44]. As described in Fig. 3a, the docking
pose of the most active compound 15 (IC50 ¼ 30 nM) into the MMP-2
active site showed that the barbiturate core acted indeed as ZBG. In
particular, the N1 atom ensured the coordination to the zinc ion by
assuming, along with the three catalytic histidine residues, a tetra-
hedral geometry, while the charged Glu121 residue was engaged in
ꢀ
HB with the N3 atom located at a distance of 2.7 A. As expected, the
0
para-COCH3 biphenyl arm was buried into the S1 subsite where
determinant HBs were established between the oxygen atom of the
acetyl substituent and guanidine groups of and Arg149 residue. A
scoring value as high as ꢁ71.88 kJ/mol was estimated when docking
compound 15. Despite the E121Q variation (1HOV numbering),
a similar posing (Fig. 3b) was observed when docking the highest
active compound 12 (IC50 ¼ 21 nM) into the MMP-9 active site and
a comparable value of scoring equal to ꢁ69.32 kJ/mo0l was in fact
observed. However, the comparison between the S1 subsites of
MMP-2 and MMP-9 disclosed an important residue change (from
arginine to proline at position 149; 1HOV numbering) [45]. As
a result, the binding at MMP-9 could be mainly driven by
0
hydrophobic interactions at the S1 subsite. Interestingly, such
information confirmed what already emerged in the 2D QSAR
analysis (equation (3)) indicating the positive influence of hydro-
phobic groups (p > 0) at the para position on the MMP-9 affinity.
Finally, the docking pose of the highest active compound 14
(IC50 ¼ 66 nM) into the MMP-8 active site is reported in Fig. 3c. As
can be seen, no appreciable differences were discernable in the way
the barbiturate scaffold binds the zinc ion. Importantly, the
anchoring of the bulkier para-SO2CH3 substituent into the deeper
region of the S10 subsite was assured by the formation of an HB with
Arg222 of MMP-8 with scoring value equal to ꢁ81.89 kJ/mol.
Besides our efforts in the design of optimal biphenyl arm
0
engaging S1 subsite, we kept an open eye on the evaluation of
a number of pharmaceutically relevant properties (e.g., octanol/
water partition coefficient, log P; aqueous solubility, log S; and
binding to human serum albumin, log Khsa). To this aim, we used
QikPro to predict such properties for inhibitors 5, 8e17 in Table 1 as
well as for five structurally close barbiturate derivatives retrieved
by recent literature [14,17,18,20] and used as reference compounds.
As can be seen in Fig. 4aec, the presence of the hydroxyl group at
C5 that is a unique feature of compounds 5, 8e17 compared to
reference literature compounds might improve the lipophilice
hydrophilic-balance, increase the aqueous solubility and reduce
the binding to human serum albumin with a likely significant gain
in terms of bioavailability.
Fig. 3. Zoomed in view of main interactions of top active compounds 15, 12 and 14
(coloured in white) at the protein binding sites, rendered as cartoon models, of MMP-2
(a), MMP-9 (b) and MMP-8 (c), respectively. For the ease of interpretation, key residues
are represented as ball and stick while black and red dashed lines represent the
coordination to the catalytic zinc ion and hydrogen bonds, respectively. (For inter-
pretation of the references to colour in this figure legend, the reader is referred to the
web version of this article.)

O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
373
4. Conclusions
The basic idea behind our current investigation was the repur-
posing of properly modified pyrimidine-2,4,6-triones as MMP
inhibitors to avoid the unwanted properties of hydroxamate MMP
inhibitors while ensuring more attractive drug-likeness. In this
respect, we tried to achieve a better pharmacokinetic profile by
inserting as a substituent at the C5 a hydroxyl group, which ensured
a better log P, an improved aqueous solubility and a limited binding
to human serum albumin as well. On the other hand, the second
substituent of the C5 was chosen by using a pool of variously
decorated biphenyl moieties having diverse hydrophobic, electronic
0
and steric properties to explore the chance of interactions at the S1
subsite. The good inhibition data obtained suggested that our
compounds merit further exploration (especially as selective
gelatinase inhibitors) [46]. And in fact, the design of novel MMP-2
inhibitors is still an open challenge in medicinal chemistry research.
5. Experimental section
5.1. Chemistry
Syntheses of biaryl methyl ketones intermediates
I and
compounds 7e15 and 17e20 were performed using a Milestone
Microsynth oven, and a Büchi Syncore parallel reactor, respectively.
Commercial reagents and solvents were purchased from Sigmae
Aldrich Europe. Melting points (mp) were taken on a Gallenkamp
MFB 595010 M apparatus and are uncorrected. Elemental analyses
were performed on a EuroEA 3000 analyser for C, H, N; experi-
mental results agreed to within ꢂ0.40% of the theoretical values.
ESI mass spectra (not reported) were performed on a Agilent 1100
series LC-MSD trap system VL apparatus. 1H NMR spectra were
recorded in DMSO-d6, unless otherwise stated, on a Varian Mercury
300 spectrometer. Chemical shifts are expressed in d (ppm) and the
coupling constants J in Hz. The following abbreviations were used:
s, singlet; d, doublet; t, triplet; m, multiplet. Exchange with
deuterium oxide was used to identify OH and NH protons. Chro-
matographic separation of isomer compounds 19 and 20 was per-
formed on a Biotage SP1 purification system using flash cartridges
ꢀ
pre-packed with KPSilÔ 32e63
m
m, 60 A silica.
5.1.1. Synthesis of biaryl methyl ketones I
In a sealed microwave vessel equipped with magnetic stirrer
and optical sensor for temperature detection, 1.0 mmol of 4-
bromoacetophenone (0.20 g), 1.2 mmol of opportune phenyl-
boronic acid, 0.05 mmol of dichlorobis(triphenylphosphine)palla-
dium(II) (0.04 g) and 3.0 mmol of sodium carbonate (0.32 g) were
suspended in 3 mL of dioxane and 3 mL of water. Microwave
heating was set as follows: 800 W, ramping from room temperature
(rt) to 150 ꢃC in 3 min, heating at 150 ꢃC for 10 min, then cooling to
rt. The residue was filtered, washed with diethyl ether and dried as
white precipitate. All biaryl methyl ketones I were characterized by
their LCeMS and 1H NMR spectra and used without further puri-
fication; yields 62e81%.
5.1.2. Synthesis of compounds 7e15 and 17e20
0.5 mmol of alloxan monohydrate (0.08 g) and suitable methyl
ketone were suspended in 5 mL of glacial acetic acid and reacted in
a Syncore apparatus set at the temperature of 115 ꢃC, shaking at
120 rpm and reaction time 3 h. All the targeted compounds
precipitated after cooling and were recrystallized from ethanol.
Compounds 19 and 20 were obtained as a mixture in a 36:64
ratio (total yield 75%); chromatographic purification of the crude
(gradient eluent: methanol in dichloromethane 0e10%) afforded
the pure final compounds.
Fig. 4. Histograms describing the variation of some calculated pharmaceutically
relevant properties that are: (a) octanol/water partition coefficient, log P; (b) aqueous
solubility, log S; and (c) binding to human serum albumin, log Khsa. The first five bars
are referred to compounds retrieved by recent literature where the letter R indicates
the corresponding endnote [14,17,18,20] and the digit the numbering of the original
paper (see also Supplementary Materials). The remaining bars are related with
inhibitors 5, 8e17 in Table 1 as indicated by the reported labels.

374
O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
5.1.2.1. 5-[2-(30-Nitrobiphen-4-yl)-2-oxoethyl]-5-hydroxy-hexahy-
mp > 250 ꢃC 1H NMR
2H, Jo ¼ 8.3), 7.81 (d, 2H, Jo ¼ 8.3), 7.52 (d, 2H, Jo ¼ 8.3), 7.32 (s,1H, OH),
3.93 (s, 2H), 2.98 (s, 3H), 2.94 (s, 3H). Anal. % (C21H19N3O6) calculated:
C 61.61, H 4.68, N 10.25; found C 61.22, H 4.94, N 10.49.
d
11.45 (s, 2H, NH), 8.05 (d, 2H, Jo ¼ 8.3), 7.88 (d,
dropyrimidine-2,4,6-trione (7). 64% Yield, mp 246e8 ꢃC 1H NMR
d
11.46 (s, 2H, NH), 8.51 (s, 1H), 8.21e8.29 (m, 2H), 8.09 (d, 2H,
Jo ¼ 8.1), 7.96 (d, 2H, Jo ¼ 8.1), 7.79 (t, 1H, Jo ¼ 8.1), 7.33 (s, 1H, OH),
3.95 (s, 2H). Anal. % (C18H13N3O7) calculated: C 56.40, H 3.42, N
10.96; found C 56.28, H 3.37, N 10.68.
5.1.2.11. 5-[2-(40-(Thien-3-yl)phenyl)-2-oxoethyl]-5-hydroxy-hex-
ahydropyrimidine-2,4,6-trione (18). 49% Yield, mp
>
250 ꢃC
5.1.2.2. 5-[2-(40-Methylbiphen-4-yl)-2-oxoethyl]-5-hydroxy-hexahy-
(decomp. 230 ꢃC). 1H NMR
d
11.43 (s, 2H, NH), 8.12 (s, 1H), 7.98 (d,
dropyrimidine-2,4,6-trione (8). 69% Yield, mp > 250 ꢃC 1H NMR
2H, Jo ¼ 8.3), 7.89 (d, 2H, Jo ¼ 8.3), 7.66e7.70 (m, 2H), 7.27 (s, 1H,
OH), 3.90 (s, 2H). Anal. % (C16H12N2O5S) calculated: C 55.81, H 3.51,
N 8.14; found C 55.52, H 3.87, N 8.42.
d
11.46 (s, 2H, NH), 8.02 (d, 2H, Jo ¼ 8.7), 7.82 (d, 2H, Jo ¼ 8.7), 7.65 (d,
2H, Jo ¼ 8.4), 7.30 (d, 2H, Jo ¼ 8.4), 7.31 (s, 1H, OH), 3.91 (s, 2H), 2.34
(s, 3H). Anal. % (C19H16N2O5) calculated: C 64.77, H 4.58, N 7.95;
found C 64.56, H 4.60, N 7.82.
5.1.2.12. 5-[3-(Biphen-4-yl)-2-oxopropyl]-5-hydroxy-hexahydropyr-
imidine-2,4,6-trione (19). 27% Yield, mp 173e5 ꢃC (decomp. 150 ꢃC).
5.1.2.3. 5-[2-(40-Chlorobiphen-4-yl)-2-oxoethyl]-5-hydroxy-hexahy-
1H NMR (acetone-d6)
d 10.24 (s, 2H, NH), 7.60e7.67 (m, 2H), 7.43e
dropyrimidine-2,4,6-trione (9). 60% Yield, mp > 250 ꢃC 1H NMR
7.54 (m, 3H), 7.30e7.37 (m, 3H), 7.16e7.19 (m, 1H), 5.80 (s, 1H, OH),
3.30 (s, 2H), 2.14 (s, 2H). Anal. % (C19H16N2O5) calculated: C 64.77, H
4.58, N 7.95; found C 64.47, H 4.27, N 7.67.
d
10.60 (s, 2H, NH), 7.81 (d, 2H, Jo ¼ 8.3), 7.49 (d, 2H, Jo ¼ 8.3), 7.40 (d,
2H, Jo ¼ 8.4), 7.28 (d, 2H, Jo ¼ 8.4), 6.83 (s, 1H, OH), 3.89 (s, 2H). Anal.
% (C18H13ClN2O5) calculated: C 58.00, H 3.52, N 7.51; found C 57.96,
H 3.67, N 7.49.
5.1.2.13. 5-[1-(Biphen-4-yl)-2-oxopropyl]-5-hydroxy-hexahydropyr-
imidine-2,4,6-trione (20). 48% Yield, mp
(acetone-d6) 10.21 (s, 2H, NH), 7.62e7.71 (m, 4H), 7.44e7.54 (m,
>
250 ꢃC 1H NMR
5.1.2.4. 5-[2-(40-(Hydroxymethyl)biphen-4-yl)-2-oxoethyl]-5-
d
hydroxy-hexahydropyrimidine-2,4,6-trione
(10). 62%
Yield,
4H), 7.37 (t, 1H, Jo ¼ 7.5), 5.72 (s, 1H, OH), 4.88 (s, 1H), 2.05 (s, 3H).
Anal. % (C19H16N2O5) calculated: C 64.77, H 4.58, N 7.95; found C
64.53, H 4.50, N 7.97.
mp > 250 ꢃC 1H NMR
d
11.45 (s, 2H, NH), 8.02 (d, 2H, Jo ¼ 8.3), 7.83
(d, 2H, Jo ¼ 8.3), 7.71 (d, 2H, Jo ¼ 8.4), 7.43 (d, 2H, Jo ¼ 8.4), 7.31 (s,
1H, OH), 5.25 (t, 1H, J ¼ 6.0, OH), 4.54 (d, 2H, J ¼ 6.0), 3.92 (s, 2H).
Anal. % (C19H16N2O6) calculated: C 61.96, H 4.38, N 7.61; found C
61.84, H 4.49, N 7.42.
5.2. Biological assays
Inhibitory activities were determined by a fluorimetric assay,
where a pro-fluorescing peptide is used as substrate, and the flu-
orogenic activity of its cleavage product is measured after coincu-
bation of the selected MMP with test compounds.
Assay reagents (Calbiochem) used in the fluorimetric assay were
as follows:
5.1.2.5. 5-[2-(40-Methoxybiphen-4-yl)-2-oxoethyl]-5-hydroxy-hex-
ahydropyrimidine-2,4,6-trione (11). 82% Yield, mp > 250 ꢃC 1H NMR
d
10.78 (s, 2H, NH), 7.63 (d, 2H, Jo ¼ 8.4), 7.33 (d, 2H, Jo ¼ 8.4), 7.27 (d,
2H, Jo ¼ 8.8), 6.74 (s, 1H, OH), 6.68 (d, 2H, Jo ¼ 8.8), 3.69 (s, 2H), 3.53
(s, 3H). Anal. % (C19H16N2O6) calculated: C 61.96, H 4.38, N 7.61;
found C 61.64, H 4.45, N 7.53.
- Substrate: Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg 25 mM in assay
5.1.2.6. 5-[2-(40-(Trifluoromethoxy)biphen-4-yl)-2-oxoethyl]-5-
buffer (Calbiochem). This low-fluorescent peptide is cleaved by
MMPs to give Mca-Pro-Leu-Gly, which is intensely fluorescent
(exc. 340 nm; em. 405 nm).
- Enzymes: pro-MMP-2250 pM; MMP-8 (recombinant catalytic
sequence) 500 pM; MMP-9 (recombinant catalytic sequence)
500 pM in assay buffer.
hydroxy-hexahydropyrimidine-2,4,6-trione (12). 69% Yield, mp
>
250 ꢃC 1H NMR
d
11.46 (s, 2H, NH), 8.05 (d, 2H, Jo ¼ 8.3), 7.84e7.90 (m,
4H), 7.49 (d, 2H, Jo ¼ 8.3), 7.32 (s, 1H, OH), 3.93 (s, 2H). Anal. %
(C19H13F3N2O6) calculated: C 54.04, H 4.55, N 6.63; found C 53.91, H
4.44, N 6.72.
- Assay buffer: triseHCl 0.1 M, pH 7.5, NaCl 0.1 M, CaCl2 10 mM,
0.05% v/v BRIJ35, APMA (p-aminophenylmercury acetate)
1 mM.
5.1.2.7. 5-[2-(40-(Methylthio)biphen-4-yl)-2-oxoethyl]-5-hydroxy-
hexahydropyrimidine-2,4,6-trione (13). 80% Yield, mp > 250 ꢃC 1H
NMR
d
11.45 (s, 2H, NH), 8.01 (d, 2H, Jo ¼ 8.5), 7.83 (d, 2H, Jo ¼ 8.5),
7.71 (d, 2H, Jo ¼ 8.4), 7.36 (d, 2H, Jo ¼ 8.4), 7.31 (s, 1H, OH), 3.91 (s,
2H), 2.51 (s, 3H). Anal. % (C19H16N2O5S) calculated: C 59.37, H 4.20,
N 7.29; found C 59.58, H 4.21, N 6.89.
In a typical experiment in 96-well white plate, a solution of
inhibitor (50
(50
L) was incubated at 25 ꢃC for 30 min; then 10
solution were added and the plate was incubated 3 h at 37 ꢃC.
Incubation solutions were quenched with acetic acid (100 L of 3%
m
L) at the opportune concentration and enzyme
m
mL of substrate
5.1.2.8. 5-[2-(40-(Methylsulphonyl)biphen-4-yl)-2-oxoethyl]-5-
m
hydroxy-hexahydropyrimidine-2,4,6-trione
(14). 65%
Yield,
soln.) and their fluorescence read with a plate reader spectrometer
mp > 250 ꢃC 1H NMR
d
11.44 (s, 2H, NH), 7.91e8.10 (m, 8H), 7.31 (s,
(Perkin Elmer Victor3).
1H, OH), 3.94 (s, 2H), 3.26 (s, 3H). Anal. % (C19H16N2O7S) calculated:
C 54.80, H 3.87, N 6.73; found C 54.50, H 3.57, N 6.86.
Inhibitory activities were calculated against a control and IC50
curves, derived from 6 different inhibitor concentrations, extrapo-
lated by means of statistical analysis program Prism (v. 4.0). Each
inhibitor concentration was used in triplicate and the experiments
5.1.2.9. 5-[2-(40-Acetylbiphen-4-yl)-2-oxoethyl]-5-hydroxy-hexahy-
dropyrimidine-2,4,6-trione (15). 55% Yield, mp > 250 ꢃC (decomp.
run twice. Low active compounds were tested at a single 10
concentration (100 M for compounds 19 and 20).
mM
235 ꢃC). 1H NMR
d
11.45 (s, 2H, NH), 8.05e8.08 (m, 4H), 7.90e7.93
m
(m, 4H), 7.31 (s, 1H, OH), 3.94 (s, 2H), 2.61 (s, 3H). Anal. %
(C20H16N2O6) calculated: C 63.16, H 4.24, N 7.37; found C 63.30, H
4.43, N 7.44.
5.3. Computational methods
5.3.1. Docking simulations
GOLD (version 5.1), a genetic algorithm based software, was
used for the docking study choosing GOLDSCORE and CHEMPLP as
5.1.2.10. 5-[2-(40-(N,N-Dimethylaminocarbonyl)biphen-4-yl)-2-
oxoethyl]-5-hydroxy-hexahydropyrimidine-2,4,6-trione (17). 62% Yield,

O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
375
scoring and rescoring functions, respectively. GOLDSCORE is made
Acknowledgements
up of four components that account for proteineligand binding
energy: proteineligand hydrogen bond energy (external H-bond),
proteineligand van der Waals energy (external vdw), ligand
internal van der Waals energy (internal vdw), and ligand torsional
strain energy (internal torsion). Parameters used in the fitness
function (hydrogen bond energies, atom radii and polarizabilities,
torsion potentials, hydrogen bond directionalities, and so forth)
were taken from the GOLD parameter file. CHEMPLP is an empirical
fitness function optimised for pose prediction and to model the
steric complementarity between protein and ligand. It is based on
the Piecewise Linear Potential (PLP) that models the attraction as
well as repulsion of protein and ligand heavy-atoms and takes into
account the distance- and angle-dependent hydrogen and metal
bonding terms from ChemScore fitness function. Parameters used
in CHEMPLP fitness function were taken by default. Importantly,
the use of CHEMPLP permitted to gain a better ranking of docking
solutions and comparable results with respect to GOLDSCORE. In
the present study, the 3D coordinates of MMP-2 (pdbcode 1HOV),
MMP-8 (pdbcode 1I73), and MMP-9 (pdbcode 2OVX) were
retrieved from the Protein Data Bank. The target proteins were
prepared by adding hydrogen atoms, completing and optimising
missing residues, removing water and the cocrystallized molecules
by using the Protein Preparation Wizard (Schrödinger 9.2, LLC, New
York, NY, 2011). As well known, the histidine side chains cannot
normally be placed into the electron density map unambiguously.
As a result, the protonation state of such residues was adjusted
according to the formation of HB networks that was further
confirmed upon visual inspection. The basic amino functional
groups were protonated, aromatic amino functional groups were
left uncharged and carboxylic groups were considered to be
deprotonated. For every protein GOLD was set to generate 10
The authors gratefully acknowledge the University of Bari “Aldo
Moro” (Progetto IDEA e Giovani Ricercatori 2008), the European
Commission (“CancerGrid” STREP project, FP VI, Contract No. LSHC-
CT-2006-03755) and Regione Puglia (Strategic Project “Neuro-
Biotech”) for financial support.
Appendix A. Supplementary data
Supplementary data related to this article can be found at http://
dx.doi.org/10.1016/j.ejmech.2012.09.036.
References
[1] C.J. Malemud, Front. Biosci. 11 (2006) 1696e1701.
[2] C.M. Overall, O. Kleifeld, Nat. Rev. Cancer 6 (2006) 227e239.
[3] J.E. Rundhaug, J. Cell. Mol. Med. 9 (2005) 267e285.
[4] E.I. Deryugina, J.P. Quigley, Cancer Metastasis Rev. 25 (2006) 9e34.
[5] J.H. Hu, P.E. Van den Steen, Q.A. Sang, G. Opdenakker, Nat. Rev. Drug Discov. 6
(2007) 480e498.
[6] G. Dormán, S. Cseh, I. Hajdú, L. Barna, D. Kónya, K. Kupai, L. Kovács,
P. Ferdinandy, Drugs 70 (2010) 949e964.
[7] A.K. Nalla, B. Gorantla, C.S. Gondi, S.S. Lakka, J.S. Rao, Cancer Gene Ther. 17
(2010) 599e613.
[8] B. Govinda Rao, Curr. Pharm. Des. 11 (2005) 295e322.
[9] E. Breuer, J. Frant, R. Reich, Expert Opin. Ther. Patent 15 (2005) 253e269.
[10] M.T. Rubino, M. Agamennone, C. Campestre, P. Campiglia, V. Cremasco,
R. Faccio, A. Laghezza, F. Loiodice, D. Maggi, E. Panza, A. Rossello, P. Tortorella,
ChemMedChem 6 (2011) 1258e1268.
[11] M.T. Rubino, M. Agamennone, C. Campestre, G. Fracchiolla, A. Laghezza,
F. Loiodice, E. Nuti, A. Rossello, P. Tortorella, ChemMedChem 4 (2009) 352e362.
[12] A.R. Folgueras, A. Fueyo, O. Garcia-Suarez, J. Cox, A. Astudillo, P. Tortorella,
C. Campestre, A. Gutierrez-Fernandez, M. Fanjul-Fernandez, C.J. Pennington,
D.R. Edwards, C.M. Overall, C. Lopez-Otin, J. Biol. Chem. 283 (2008) 9465e9474.
[13] M.T. Rubino, D. Maggi, A. Laghezza, F. Loiodice, P. Tortorella, Arch. Pharm. 344
(2011) 557e563.
[14] J.A. Blagg, M.C. Noe, L.A. Wolf-Gouveia, L.A. Reiter, E.R. Laird, S.P. Chang,
D.E. Danley, J.T. Downs, N.C. Elliott, J.D. Eskra, R.J. Griffiths, J.R. Hardink,
A.I. Haugeto, C.S. Jones, J.L. Liras, L.L. Lopresti-Morrow, P.G. Mitchell, J. Pandit,
R.P. Robinson, C. Subramanyam, M.L. Vaughn-Bowser, S.A. Yocum, Bioorg.
Med. Chem. Lett. 15 (2005) 1807e1810.
[15] L.H. Foley, R. Palermo, P. Dunten, P. Wang, Bioorg. Med. Chem. Lett. 11 (2001)
969e972.
[16] J. Wang, C. Medina, M.W. Radomski, J.F. Gilmer, Bioorg. Med. Chem. 19 (2011)
4985e4999.
[17] L.A. Reiter, K.D. Freeman-Cook, C.S. Jones, G.J. Martinelli, A.S. Antipas,
M.A. Berliner, K. Datta, J.T. Downs, J.D. Eskra, M.D. Forman, E.M. Greer,
R. Guzman, J.R. Hardink, F. Janat, N.F. Keene, E.R. Laird, J.L. Liras, L.L. Lopresti-
Morrow, P.G. Mitchell, J. Pandit, D. Robertson, D. Sperger, M.L. Vaughn-
Bowser, D.M. Wallera, S.A. Yocum, Bioorg. Med. Chem. Lett. 16 (2006) 5822e
5826.
ꢀ
docking poses per molecule in a sphere of 13 A radius centred on
the centroid atom of cocrystallized ligand.
GOLD was flagged for the automated determination of the
coordination geometry around the zinc metal ion. More specifically,
virtual coordination points were added at locations where GOLD
was missing a coordination site and these coordination points were
used as fitting points that could bind to acceptors. In order to
determine the coordination geometry, GOLD performs a permuted
superimposition of coordination geometry templates onto the
coordinating atoms found in the protein. Coordination fitting
points were thus generated using the template that gave the best fit
in terms of RMSD from the target. The tetrahedral and trigonal
bipyramidal coordination geometries were manually specified to
allow the docking software to prioritize the Zn metal ion. A visual
inspection of solutions demonstrated that our molecules satisfied
the structural requisites for a stable coordination with the metal as
[18] D. Schrigten, H.J. Breyholz, S. Wagner, S. Hermann, O. Schober, M. Schäfers,
G. Haufe, K. Kopka, J. Med. Chem. 55 (2012) 223e232.
[19] H.J. Breyholz, M. Schäfers, S. Wagner, C. Höltke, A. Faust, H. Rabeneck,
B. Levkau, O. Schober, K. Kopka, J. Med. Chem. 48 (2005) 3400e3409.
[20] M. Piette, B. Evrard, F. Frankenne, P. Chiap, P. Bertholet, D. Castagne, J.-
M. Foidart, L. Delattre, G. Piel, Eur. J. Pharm. Sci. 28 (2006) 189e195.
[21] I. Giangreco, G. Lattanzi, O. Nicolotti, M. Catto, A. Laghezza, F. Leonetti,
A. Stefanachi, A. Carotti, PLoS One 6 (2011) e25597.
[22] Y. Yan, M.T. Miller, Y. Cao, S.M. Cohen, Bioorg. Med. Chem. Lett. 19 (2009)
1970e1976.
[23] B. Flachner, Z. Lörincz, A. Carotti, O. Nicolotti, P. Kuchipudi, N. Remez, F. Sanz,
J. Tóvári, M.J. Szabó, B. Bertók, S. Cseh, J. Mestres, G. Dormán, PLoS One 7
(2012) e35582.
0
well as a suitable fitting of the biphenyl fragment into the S1
subsite.
Following the clues from MD analyses carried out by us in a recent
published work [21], which indicate the fluctuation of Arg149 (1HOV
numbering) within the time scale of MD simulation, the full side-
chain flexibility of such a residue was explicitly considered giving
GOLD the option to read alternative allowed rotamers from a library
compiled by the most commonly observed side-chain conformations
for the naturally occurring amino acids [47].
[24] C. Altomare, S. Cellamare, L. Summo, M. Catto, A. Carotti, U. Thull, P.-
A. Carrupt, B. Testa, H. Stoeckli-Evans, J. Med. Chem. 41 (1998) 3812e3820.
[25] C. Hansch, A. Leo, Exploring QSAR
e Fundamentals and Applications in
Chemistry and Biology e ACS Professional Reference Book, American Chem-
ical Society, Washington, DC, 1995.
[26] L. Verdonk, J.C. Cole, M.J. Hartshorn, C.W. Murray, R.D. Taylor, Proteins 52
(2003) 609e623.
[27] C. Knox, V. Law, T. Jewison, P. Liu, S. Ly, A. Frolkis, A. Pon, K. Banco, C. Mak,
V. Neveu, Y. Djoumbou, R. Eisner, A.C. Guo, D.S. Wishart, Nucl. Acids Res. 39
(2011) 1035e1041.
[28] C. Hansch, T.E. Klein, Methods Enzymol. 202 (1991) 512e543.
[29] A. Carotti, C. Raguseo, T.E. Klein, R. Langridge, C. Hansch, Chem. Biol. Interact.
67 (1988) 171e184.
[30] A. Carotti, C. Raguseo, F. Campagna, R. Langridge, T.E. Klein, Quant. Struct.e
Act. Relat. 8 (1989) 1e10.
5.3.2. Pharmaceutically relevant properties predictions
QikProp is an easy-to-use program for a quick and accurate
prediction of absorption, distribution, metabolism, and excretion
properties (ADME properties). QikProp predicts physically signifi-
cant descriptors and pharmaceutically relevant properties of
organic molecules, either individually or in batches.(QikProp,
version 3.4, Schrödinger, LLC, New York, NY, 2011.)

376
O. Nicolotti et al. / European Journal of Medicinal Chemistry 58 (2012) 368e376
[31] A. Carotti, C. Raguseo, C. Hansch, Quant. Struct.eAct. Relat.
145e149.
4
(1985)
[39] O. Nicolotti, T.F. Miscioscia, A. Carotti, F. Leonetti, A. Carotti, J. Chem. Inf.
Model. 48 (2008) 1211e1226.
[32] A. Carotti, R.N. Smith, S. Wong, C. Hansch, J.M. Blaney, R. Langridge, Arch.
Biochem. Biophys. 229 (1984) 112e125.
[33] A. Carotti, M.M. Mueller, J.M. Blaney, R. Langridge, J. Med. Chem. 27 (1984)
1401e1405.
[40] A. Tochowicz, K. Maskos, R. Huber, R. Oltenfreiter, V.R. Dive, A. Yiotakis,
M. Zanda, W. Bode1, P. Goettig, J. Mol. Biol. 371 (2007) 989e1006.
[41] H. Brandstetter, F. Grams, D. Glitz, D. Lang, R. Huber, W. Bode, H.W. Krell,
R.A. Engh, J. Biol. Chem. 276 (2001) 17405e17412.
[34] O. Nicolotti, A. Carotti, J. Chem. Inf. Model. 46 (2006) 264e276.
[35] L. Devel, B. Czarny, F. Beau, D. Georgiadis, E. Stura, V. Dive, Biochimie 92
(2010) 1501e1508.
[36] J. Grimm, B.J. Hubs, L.T. Miller, K. Otte, M.P. Siliphaivanh, M. Stanton,
G.K. Wilson, D. Witter, H. Zhou, (Merck & Co.), Patent WO2007100657,
2007.
[42] P. Pospisil, P. Ballmer, L. Scapozza, G. Folkers, J. Recept. Signal. Transduct. 23
(2003) 361e371.
[43] O. Nicolotti, T.F. Miscioscia, F. Leonetti, G. Muncipinto, A. Carotti, J. Chem. Inf.
Model. 47 (2007) 2439e2448.
[44] I. Giangreco, O. Nicolotti, A. Carotti, F. De Carlo, G. Gargano, R. Bellotti, BMC
Bioinform. 500 (2010) 1e13.
[37] K.J. Wilson, D.J. Witter, J.B. Grimm, P. Siliphaivanh, K.M. Otte, A.M. Kral,
J.C. Fleming, A. Harsch, J.E. Hamill, J.C. Cruz, M. Chenard, A.A. Szewczak,
R.E. Middleton, B.L. Hughes, W.K. Dahlberg, J.P. Secrist, T.A. Miller, Bioorg.
Med. Chem. Lett. 18 (2008) 1859e1863.
[45] J.M. Chen, F.C. Nelson, J.I. Levin, D. Mobilio, F.J. Moy, R. Nilakantan, A. Zask,
R. Powers, J. Am. Chem. Soc. 122 (2000) 9648e9654.
[46] J. Wang, S. O’Sullivan, S. Harmon, R. Keaveny, M.W. Radomski, C. Medina,
J.F. Gilmer, J. Med. Chem. 55 (2012) 2154e6212.
[38] O. Nicolotti, I. Giangreco, T.F. Miscioscia, A. Carotti, J. Chem. Inf. Model. 49
(2009) 2290e2302.
[47] S.C. Lovell, J.M. Word, J.S. Richardson, D.C. Richardson, Proteins 40 (2000)
389e408. 14.