Pteridine-sulfonamide conjugates as dual inhibitors of carbonic anhydrases and dihydrofolate reductase with potential antitumor activity

Article abstract of DOI:10.1016/j.bmc.2010.05.072

Recent evidences suggest that cancer treatment based on combination of cytostatic and conventional chemostatic therapeutics, which are usually cytotoxic, can provide an improved curative option. On the sequence of our previous work on methotrexate (MTX) derivatives, we have developed and evaluated novel MTX analogues, containing a pteridine moiety conjugated with benzenesulfonamide derivatives, thus endowed with the potential capacity for dual inhibition of dihydrofolate reductase (DHFR) and carbonic anhydrases (CA). These enzymes are often overexpressed in tumors and are involved in two unrelated cellular pathways, important for tumor survival and progression. Their simultaneous inhibition may turn beneficial in terms of enhanced antitumor activity. Herein we report the design and synthesis of several diaminopteridine-benzenesulfonamide and -benzenesulfonate conjugates, differing in the nature and size of the spacer group between the two key moieties. The inhibition studies performed on a set of CAs and DHFR, revealed the activities in the low nanomolar and low micromolar ranges of concentration, respectively. Some inhibitors showed selectivity for the tumor-related CA (isozyme IX). Cell proliferation assays using two tumor cell lines (the non-small cell lung carcinoma, A549, and prostate carcinoma, PC-3) showed activities only in the millimolar range. Nevertheless, this fact points out the need of improving the cell intake properties of these new compounds, since the general inhibitory profiles revealed their potential as anticancer agents.

Full text of DOI:10.1016/j.bmc.2010.05.072

Bioorganic & Medicinal Chemistry 18 (2010) 5081–5089
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Pteridine–sulfonamide conjugates as dual inhibitors of carbonic anhydrases
and dihydrofolate reductase with potential antitumor activity
Sérgio M. Marques a, Éva A. Enyedy a, Claudiu T. Supuran b, Natalia I. Krupenko c, Sergey A. Krupenko c,
M. Amélia Santos a,
*
a Instituto Superior Técnico, Centro de Química Estrutural, Av. Rovisco Pais 1, 1049-001 Lisboa, Portugal
b Università degli Studi di Firenze, Laboratorio di Chimica Bioinorganica, Rm. 188, Via della Lastruccia 3, I-50019 Sesto Fiorentino (Firenze), Italy
c Medical University of South Carolina, Department of Biochemistry and Molecular Biology, Charleston, SC 29425, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Recent evidences suggest that cancer treatment based on combination of cytostatic and conventional
chemostatic therapeutics, which are usually cytotoxic, can provide an improved curative option. On
the sequence of our previous work on methotrexate (MTX) derivatives, we have developed and evaluated
novel MTX analogues, containing a pteridine moiety conjugated with benzenesulfonamide derivatives,
thus endowed with the potential capacity for dual inhibition of dihydrofolate reductase (DHFR) and car-
bonic anhydrases (CA). These enzymes are often overexpressed in tumors and are involved in two unre-
lated cellular pathways, important for tumor survival and progression. Their simultaneous inhibition may
turn beneficial in terms of enhanced antitumor activity.
Herein we report the design and synthesis of several diaminopteridine–benzenesulfonamide and -ben-
zenesulfonate conjugates, differing in the nature and size of the spacer group between the two key moi-
eties. The inhibition studies performed on a set of CAs and DHFR, revealed the activities in the low
nanomolar and low micromolar ranges of concentration, respectively. Some inhibitors showed selectivity
for the tumor-related CA (isozyme IX). Cell proliferation assays using two tumor cell lines (the non-small
cell lung carcinoma, A549, and prostate carcinoma, PC-3) showed activities only in the millimolar range.
Nevertheless, this fact points out the need of improving the cell intake properties of these new com-
pounds, since the general inhibitory profiles revealed their potential as anticancer agents.
Ó 2010 Elsevier Ltd. All rights reserved.
Received 8 February 2010
Revised 24 May 2010
Accepted 26 May 2010
Available online 2 June 2010
Keywords:
Multi-target drugs
Dihydrofolate reductase inhibitors
Carbonic anhydrase inhibitors
Cancer therapy
1. Introduction
isozymes IX, XII, and CA-related protein VIII are highly abundant in
tumors and are involved in tumorigenesis and tumor progression,7,8
Since most cancers proved to often evade the conventional che-
motherapeutic treatments, the combination of several anticancer
treatments has been one of the recent medical research strategies,
alongwiththesearchfornewtargets. Thereisworldwideinterestto-
wards the development of multifunctional and multitargeting
drugs, which is an alternative to drug combination protocols for
the treatment of diseases with complex etiologies (such as can-
cer).1,2 This trend has also challenged us to embark on the develop-
ment and study of new potential drugs with dual activity against
cancer pathways.3,4 The antifolate methotrexate (MTX, see Fig. 1)
is the oldest and one of the most widely used anticancer drugs. Its
activity is based on blockage of folate pathways, important for can-
cer cellsurvival, byinhibitingdihydrofolate reductase(DHFR).5,6 An-
other pathway, which might be relevant to malignant proliferation,
involves carbonic anhydrases (CAs). Recent studies revealed that CA
and that has led to their validation as new therapeutic targets for
cancer chemotherapy intervention.9,10 Thus, we have pursued a re-
search strategy focused on the novel hybrid compounds with poten-
tial dual inhibitory activity against DHFR and CAs. For this purposed,
two key chemical moieties (a pteridine and a benzenesulfonamide)
are combined in the same molecular entity for blocking the function
of the enzymes from two unrelated metabolic pathways involved in
tumor progression, with expected beneficial in terms of enhancing
the antitumor activity.
We present herein the design and synthetic route of several
arylsulfonamide- and arylsulfonate–diaminopteridine conjugates,
which differ in the length and the nature of spacer (linker) between
the two molecular moieties (pteridine and benzenesulfonyl groups,
see Chart 1). The inhibitory activities of these compounds against
three carbonic anhydrase isoforms (CA I, II and IX) and DHFR, as
well as their antiproliferative activities in the A549 (non-small cell
lung carcinoma) and PC-3 (prostate carcinoma) cell lines, are also
presented and discussed.
* Corresponding author. Tel.: +351 218419273; fax: +351 218464455.
E-mail address: masantos@ist.utl.pt (M. Amélia Santos).
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2010.05.072

5082
S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
O
N
N
CH3
N
H
SO2NH2
S
O
COOH
AAZ
γ
OH
NH2
N
N
H
α
N
N
O
N
N
SO2NH2
O
H2N
MTX
N
H
NH2SO2
SPESB
Figure 1. Structures of methotrexate (MTX,
a DHFR inhibitor and current antitumor drug), and of reference CA inhibitors: acetazolamide (AAZ) and N-(4-
sulfamoylphenylethyl)-4-sulfamoylbenzamide (SPESB).
NH2
N
pounds. This synthetic route is shown in Scheme 1 for compounds
1 and 2 (a and b analogues).
O
S
N
Spacer
X
In the case of the p-aminobenzoyl (pABA)- or N-methyl-p-amin-
obenzoyl (NMepABA)-containing inhibitors (compounds 3a–4a,
and 3b), before the coupling to the pteridine moiety, the benzene-
sulfonyl-containing intermediates were previously synthesized
through different steps (see Scheme 2). The pABA amine group
was first protected with the tert-butoxycarbonyl (Boc) group, yield-
ing compound 5. Boc-pABA (5) and NMepABA were then conju-
gated with the sulfonamide- (or sulfonate-) containing
fragments, via the TBTU-activation of the corresponding carboxylic
groups, generating compounds 6 (a, b) and 8. After the Boc-depro-
tection of 6 with TFA, to afford the respective amines, 7 and 8 were
coupled with the halogenated pteridine, PtBr, as described for com-
pounds 1 and 2, in order to obtain the target compounds.
O
H2N
N
N
3. Results and discussion
3.1. Enzyme inhibition
Chart 1.
The new compounds were tested in terms of their inhibitory
activities towards a set of physiologically relevant human carbonic
anhydrases (the ubiquitous CA I and II, and the tumor-associated
isoform CA IX) and dihydrofolate reductase (DHFR). Results of
these bioassays are presented in Table 1. Our data demonstrated
that most of the new compounds inhibited the tested enzymes,
but the activities vary from the low nanomolar to the micromolar
range.
2. Chemistry
For the preparation of the new inhibitors, the benzenesulfonyl-
containing moieties were coupled to the 2,4-diamino-6-(hydroxy-
methyl)pteridine (PtOH). This was carried out in two main steps:
firstly, PtOH was treated with thionyl bromide, to give the corre-
sponding halogenated analogue (PtBr). Afterwards, an amine-con-
taining derivative having the benzenesulfonyl moiety attached was
coupled with the pteridine moiety via nucleophilic substitution at
PtBr under strictly anhydrous conditions, yielding the final com-
Regarding the CAs inhibition profile, all compounds displayed
moderate activity with CA I (KI values ranging from 1.2 to
36
lM), as compared with reference inhibitors (KI values of 250
NH2
NH2
N
i
N
N
ii
N
OH
Pt-Br
N
NHR1
NH2R1
H2N
N
H2N
N
N
Pt-OH
-PhSO2X (1)
R1
=
(a)
X = NH2
(b)
OH
-(CH2)2PhSO2X (2)
Scheme 1. Synthesis of compounds 1 and 2 (a and b analogues). Reagents and conditions: (i) SOBr2, rt; (ii) NaH, CaCO3, DMF, rt.

S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
5083
i
ii
Boc-NH-PhCOOH
Boc-NH-PhCONHR 2
YHN
COOH
NH2R2
5
6a-b
ABA
p
(Y = H)
NMepABA
(Y = CH3)
iii
NH2R2
ii
CONHR2
NH2
N
iv
iv
NH2-PhCONHR2
N
CH3NH-PhCONHR2
N
N
Y
Pt-Br
Pt-Br
7a-b
8
H2N
N
(a)
(b)
X = NH2
OH
R2
=
-(CH2)2PhSO2X, Y = H (3)
-(CH2)2PhSO2X, Y = CH3 (4)
Scheme 2. Synthesis of compounds 3a, 3b, and 4a. Reagents and conditions: (i) Boc2O, Na2CO3, H2O/dioxane, rt; (ii) TBTU, NMM, DMF, 0 °C; (iii) 50% TFA, CH2Cl2, rt; (iv) NaH,
CaCO3, DMF, rt.
Table 1
Activities of the reference inhibitors AAZ, SPESB, MTX, and the synthesized compounds toward human CA I, II, IX, and DHFR, compared to their
respective antiproliferative effects in A549 and PC-3 cancer cells lines
Compounda
Spacer
KI (nM)
Select. CA IX/II
IC50 (nM)
A549
X
CA I
CA II
CA IX
DHFR
PC-3
1a
1b
2a
2b
3a
3b
4a
AAZ
SPESB
MTX
NH
NH
NHCH2CH2
NHCH2CH2
pABA–NH–CH2–CH2
pABA–NH–CH2–CH2
NMepABA-NH–CH2–CH2
NH2
OH
NH2
OH
NH2
OH
NH2
1150
2170
27,400
8170
3560
36,400
1290
250
4.5
2.1
>1000
4.7
>1000
2.5
>1000
3.7
2.2
<0.35
9.6
<2.6
369
<3.7
70
0.48
0.28
2500
2000
n.a.b
5 ꢀ 106
6 ꢀ 106
3 ꢀ 106
1 ꢀ 106
2 ꢀ 106
5 ꢀ 106
6 ꢀ 106
0.55 ꢀ 106
0.5 ꢀ 106
0.3 ꢀ 106
0.46 ꢀ 106
0.4 ꢀ 106
0.5 ꢀ 106
0.5 ꢀ 106
350
45.0
26,400
920
37,100
260
12
n.a.b
20,000
1300
1800
25
18
40
5
5
5
1
a
With the formula as in Chart 1.
n.a. means no inhibition at 0.5 mM.
b
and 40 nM, for AAZ and SPESB, respectively, see Fig. 1). This is a
good feature of these compounds, as CA I is an off target enzyme.
Concerning the inhibition of isozyme CA II, the activities presented
a more pronounced variations, the highest activities being ob-
served for compounds 1a and 4a, with KI of 4.5 and 260 nM,
more active than AAZ and SPESB (KI of 4.5, 12 and 5 nM, respec-
tively), suggesting that the 2,4-diamino-6-methylpteridine moiety
interacts favourably within the CA II active centre, but this interac-
tion is worsened when the distance between that moiety and the
ZBG of the compound is increased.
respectively, and the lowest for 3b, with 37
l
M. For this isoform,
Regarding CA IX, this isozyme had the highest inhibitory activ-
ities observed, namely with compounds 1a, 3a and 4a, displaying
KI values of 2.1, 2.5 and 3.7 nM, respectively. In fact, all the sulfon-
amide-based compounds (a-type series) proved to be more active
against CA IX than the reference inhibitors AAZ and SPESB, which
have KI values of 25 and 18 nM, respectively. As it was found for
CA II, the sulfonic derivatives (b-type series) are much less potent
inhibitors of the isoform IX than their corresponding sulfonamide
it was clear the activity dependency on the nature of the benzene-
sulfonyl moiety present in the inhibitor, the benzenesulfonamide
analogues (–PhSO2NH2) being more active than their sulfonic acid
homologues (PhSO3H) (i.e., compounds 1a vs 1b, 2a vs 2b and 3a vs
3b). This behaviour was expected and is according to the litera-
ture;11,12 in fact, it is well known that the aromatic sulfonamides
are able to establish stronger interactions with the active site of
the CAs than most of other zinc-binding groups (ZBG). This feature
is due to the coordination of the catalytic zinc(II) ion by the N-atom
of the sulfonamide moiety, and the formation of two H-bonds be-
tween the threonine residue in the vicinity of that metal ion with
the sulfonamide SO2NH2 group.13 The activity of the inhibitors to-
wards CA II also decreased with increasing of the spacer’s length
(i.e., following this order in the series 1a, 2a, and 3a). However,
compound 4a revealed higher affinity for this enzyme than 3a, thus
indicating that the N-methyl group on the pABA spacer should pro-
vide some extra lipophilic binding interactions. Comparison with
the reference inhibitors shows that only compound 1a is slightly
analogues, all of them presenting KI values larger than 1 lM,
mainly because the sulfonate is a much less efficient ZBG than
the sulfonamide.14 In this case, the order of inhibitory activities
of the sulfonamide-containing compounds is different from that
of CA II, since the decreasing order for CA IX is 1a > 3a > 4a > 2a,
while for CA II it is 1a > 2a > 4a > 3a. This fact must be intrinsically
related to the differences in the structures of the two isozymes, and
it may be determinant for a rational design of selective inhibitors.
However, the difference in their activities towards CA IX is not as
significant as for CA II; in fact the KI values are of the same order
of magnitude, and thus not so dependent on the nature and length

5084
S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
of the spacer. The sulfonamide-based compounds, unlike the refer-
ence inhibitors AAZ and SPESB, also revealed more selective for CA
IX with respect to the other tested CAs. With CA IX/CA II inhibition
ratios ranging from 2.2 to 369, compound 3a proved to be the most
selective, whereas 1a, in spite of being the most active inhibitor,
was the less selective.
With regards to DHFR inhibition, the results showed greater
variability in the potency of the synthesized antimetabolites to-
wards this enzyme. Thus, compounds 2a and 2b did not affect
the enzyme activity at concentrations up to 0.5 mM, while the
other compounds demonstrated moderate inhibitory effects with
the IC50 values in the low micromolar range (from 1.3 to 20 lM).
It has to be noted that methotrexate (MTX), a widely used antican-
cer drug, inhibits DHFR at low nanomolar concentrations (IC50 is
about 5 nM). Correspondingly, compounds 3b and 4a, which are
structurally more similar to MTX, demonstrated the highest inhib-
itory activities (IC50 values of 1.3 and 1.8
other hand, the compounds with the simplest structures, 1a and
1b, have similar activities (IC50 values of 2.5 and 2.0 M, respec-
lM, respectively). On the
l
tively). Interestingly, in these two compounds, the aromatic ring
of the benzenesulfonyl moiety seems to occupy the same position
as the aromatic ring of pABA moiety of MTX, while the sulfonyl
moiety is most probably positioned in the same region as the
a-
carboxyl group of the glutamate residue in MTX (see Fig. 1). The
fact that the sulfonic analogue (compound 1b) demonstrated
slightly stronger inhibitory activity than the corresponding sulfon-
amide (compound 1a) supports this suggestion. A similar trend
was observed in case of compounds 3a and 3b, for which the differ-
ence in the IC50 values is about 15-fold. In this case, however, the
sulfonyl moiety should bind in other area, since the distance from
this group to the pteridine moiety is significantly higher.
Summarizing, the presence of a benzene ring in the proximity of
the 2,4-diamino-6-methylpteridine moiety is crucial for the bind-
ing and inhibiting DHFR, since even a slightly more remote posi-
tion (difference in two methylene groups) yields inactive
compounds (2a and 2b).
3.2. Molecular modelling
3.2.1. Docking of inhibitors to carbonic anhydrases (CAs)
In order to rationalize the inhibition results towards the CAs
and DHFR with the new inhibitors, molecular modelling studies
were carried out.
The compounds were docked into the crystal structures of CA
II15 and IX16 (respectively, entries 1G54 and 3IAI of the RCSB Pro-
tein Data Bank),17 using the GOLD software,18 following a previously
validated method.13
Figure 2. Docking of compounds into the CA IX active site: 1a (magenta)
superimposed with 1b (light green) (a); 1a (magenta) superimposed with 3a
(orange) (b). Red surfaces represent the hydrophobic and blue the hydrophilic
regions of the protein; black dashed lines represent the ligand coordination to the
metal, and the full lines represent the H-bonds formed with the protein.
The docking result for the most active compound (1a) with CA
IX is depicted in Figure 2. Analysis of this figure shows a major
binding feature, common to all ligands, involving the benzenesul-
fonyl moiety, which is well accommodated in the active site of
the CA, and coordinates the catalytic zinc ion by the anionic N-
atom of the sulfonamide, while for the sulfonate the metal coordi-
nation involves the O-atom. In the case of the sulfonamides, two H-
bonds are formed, one between its NH group and the hydroxyl
group of Thr332 (numeration according to the hCA IX structure
of UniProtKB database, entry Q16790);19 another one between
the SO group and the backbone NH of the same residue. In the case
of the sulfonic analogues (such as 1b, Fig. 2a), only one H-bond is
formed, between the SO group oxygen and the backbone NH of
Thr332. These differences explain the much lower inhibitory activ-
ities normally observed for the sulfonic analogues, as compared
with the sulfonamides. In Figure 2a, it is also possible to verify that
1a and 1b bind in somewhat different manners, besides the similar
interactions involving the ZBGs. The pteridine moiety of 1a is well
adjusted to the hydrophobic wall of the active cavity of CA IX and
establishes close contacts with the several valine and leucine resi-
dues forming that wall (namely Leu331, Leu266, Val262), which
results in a good ligand–enzyme affinity. On the other hand, 1b
does not form so many hydrophobic interactions, but one of the
NH2 group forms an H-bond with the side-chain carbonyl group
of Gln203. Furthermore, considering the proximity of the pteridine
ring of compound 1a towards the hydrophilic residue Gln224
(Fig. 2a), a slight flip of these groups may be expect, thus allowing
a H-bonding between the side-chain NH2 group and the aromatic
nitrogen atoms of the pteridine, with a concomitant considerable
increase of the adduct stability.
As regards compound 2a, it binds with a conformation in be-
tween 1a and 1b, and therefore still preserves hydrophobic inter-
action with the protein, while establishes H-bonding with the
NH2 group of Gln224 (Fig. S1 of Supplementary data). Considering
compounds 3a and 4a, their binding conformations are quite sim-

S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
5085
ilar to each other, but compared with 1a, the longer spacers in-
creased the distance between the pteridine ring and the benzene-
sulfonamide moiety, leading to a decrease of the hydrophobic
interactions with the enzyme. On the other hand, the pteridine ring
of 3a may form an H-bond with the backbone carbonyl group of
Leu223 (Fig. 2b). Moreover, it is noticeable the proximity of the
carboxylic acid of Asp263 and the NH group of the pABA moiety,
for 3a. Therefore, some H-bond interaction between these two
groups may be also expected. This interplay of interactions may
lead to complexes with similar stabilities, thus explaining the very
close inhibitory activities of the different ligands (1a–4a) toward
CA IX.
actions with the enzyme were reduced (see Fig. 3a). Compounds
2a, 3a and 4a tend, in general, to be leaning over the hydrophobic
part of the cavity, in a similar way to what happened with CA IX
(see Fig. 3b). However, the interactions established are much
weaker with CA II than with CA IX. The main reason for this differ-
ence is the type of hydrophobic residues forming this hydrophobic
wall, which, in CA IX, is mostly based on leucine and valine resi-
dues. Thus, the substitution of Leu266 in CA IX for the smaller
Val134 in CA II, and, respectively, of Val262 for Phe130, decreases
the number of hydrophobic interactions established with CA II. On
the other hand, while on CA IX the Asp263 residue may, in certain
cases, form hydrophilic interactions with our inhibitors (such as for
Analyzing the docking results with CA II, it was possible to con-
firm that the structural differences on the two proteins also lead to
different binding modes of some ligands. In fact, compounds 1a
and 1b were found to bind preferably the hydrophilic part of the
cavity, forming a H-bond between the NH2 group connected to
the pteridine and the side-chain carbonyl of Asn67 (when referring
to CA II, the residue numeration is according to the hCA II structure
of UniProtKB, entry P00918), and therefore the hydrophobic inter-
Figure 3. Docking of compounds into the CA II active site: 1a (magenta)
superimposed with 3a (orange) (a); 1a (magenta) superimposed with 3a (orange)
and 4a (green) (b). Red surfaces represent the hydrophobic and blue the hydrophilic
regions of the protein; black dashed lines represent the ligand coordination to the
metal, and the full lines represent the H-bonds formed with the protein.
Figure 4. Docking of compounds to DHFR: MTX (pink) superimposed with 4a
(green) (a); 4a (green) superimposed with 1a (magenta) (b). Red surfaces represent
the hydrophobic and blue the hydrophilic regions of the protein; the full black lines
represent the H-bonds formed between the ligand and the protein.

5086
S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
3a), on CA II it is substituted by Gly131, and so this type of inter-
actions is no longer possible (see Supplementary data, Fig. S2).
Altogether, these features seem to explain the activity and
selectivity profiles displayed by these compounds towards CA II
and IX, which proved to be in general more selective for inhibiting
CA IX than the isozyme II. Furthermore, the suggested differences
between the structures of these two enzymes may be used in the
future for the rational design of more active and CA IX-selective
inhibitors of this type.
compounds were not transported across the cell membrane. While
CA IX was shown to be a membrane isoform with the active site
extracellularly exposed,8 that could be affected by the inhibitors
in the medium, DHFR, as well as CA I and II, are strictly intracellu-
lar-located proteins, and for their inhibition the compounds should
be transported inside the cells. Apparent absence of the effect of CA
IX inhibition on cell grows in culture can be explained by the fact
that both CA IX and XII grant the survival advantage to hypoxic tu-
mor cells by regulating and maintaining pH.10 In our cell culture
models, cells grew in monolayer and have never become hypoxic.
So, inhibition of the mechanism that helps survive hypoxic condi-
tions had no effect in cell culture. Tumor xenograft model would be
more useful to evaluate the effect of this CA inhibition on tumor
cell survival.
3.2.2. Docking of inhibitors to DHFR
The target compounds were also docked into a crystal structure
of L. casei DHFR, originally complexed with MTX20 (PDB entry
3DFR). The results showed that for all the compounds, the pteri-
dine moiety is able to have a good accommodation into the cavity
of the DHFR active centre, and is positioned in a very similar way as
MTX. The interactions between that group within the hydrophobic
pocket of the enzyme may be established in the same way, namely
4. Conclusions
New bifunctional compounds based on the conjugation of the
aminopteridine and arylsulfonamide pharmacophores have been
developed for the blockage of two unrelated pathways involved
in cancer development and proliferation, which are dependent on
the enzymatic activity of DHFR and CA IX. In particular, after the
design and preparation of the dual targeting compounds, their
inhibitory activities were evaluated towards a set of CAs (CA I, II
and IX), DHFR, and their antiproliferative effects were tested in
two cancer cell lines (A549 and PC-3). As for the CA inhibition,
some compounds presented low nanomolar activity, with com-
pound 1a displaying the lowest IC50 value (2.1 nM) over CA IX,
and selectivity for the cancer-related CA IX over the ubiquitous
CA II rising up to 369 for compound 3a. In terms of DHFR inhibi-
tion, most of the compounds presented activities in the low micro-
the p–p contact with the residue Phe31 (numeration according to
entry P00381 of UniProtKP database) and H-bonds between the
NH2 groups and the carboxylate of Asp27, the hydroxyl of
Thr117, and the backbone carbonyls of Leu5 and Ala98 (see
Fig. 4a). Considering the most active inhibitors (compounds of
types 1, 3, and 4), they all contain an aromatic group directly
bound to the pteridine moiety, as the pABA group in the case of
MTX. Moreover, the docking studies demonstrated that this group
is placed in a similar way as that of MTX, as was suggested before
by our SAR analysis. This feature results in the same strong hydro-
phobic interactions, namely with the residues Phe31, Leu28,
Phe50, and Leu55. In fact, the lack of this pABA moiety in com-
pounds 2a and 2b (which have a CH2CH2 group in its place, see
Fig. S3 in Supplementary data) results in the very poor inhibitory
activity displayed by these compounds, thus revealing how impor-
tant this feature is for a good interaction with DHFR. Concerning
compounds 3a, 3b and 4a, they all bind in a similar manner,
namely the benzenesulfonyl moiety lying over the same hydropho-
bic valley as MTX (see Fig. 4b and Fig. S4 of Supplementary data).
However, while this inhibitor forms strong H-bonds between the
molar range, namely with 2.5 lM for 1a and 1.3 lM for 3b. The
antiproliferative properties of the new compounds against the
two cell lines were more disappointing, with activities in the mil-
limolar concentration range, probably due to deficiency of the
inhibitor transport inside the cells. In general, the inhibitory pro-
files towards the CAs and DHFR revealed good antitumor potential
of these new bifunctional compounds, but apparently better cell
permeability properties would be required to improve their effi-
cacy. Thus, further development of this type of compounds, based
on a rational molecular design including physicochemical and
pharmacokinetic considerations, could be of great interest in find-
ing new effective multi-target anticancer agents.
a-carboxylate and the guanidinium group of Arg58, and the c-car-
boxylate with the side-chain NH of His29, our inhibitors form only
weak hydrophobic interactions with the residues forming this val-
ley (namely Leu28 and Pro25), and this fact accounts for their ca.
1000-fold weaker inhibitory activity. The docking calculations also
demonstrated that the pteridine and pABA groups form the same
interactions in these three families of compounds (1, 3 and 4).
Since compounds 1a and 1b presented IC50 values close to inhibi-
tors of type 3 and 4, the possibility of the sulfonyl group in family 1
to form a H-bond with Arg58 (as MTX) probably overcomes the lar-
ger number of hydrophobic interactions established in families 3
and 4, thus resulting in similar inhibitory activities (Fig. 4b).
5. Experimental part
5.1. General methods
All the commercially available reagents were of the highest pur-
ity and were used without further purifications. Folic acid, metho-
trexate, O-benzotriazol-1-yl-N,N,N0,N0-tetramethyluronium-tetra-
fluoroborate (TBTU), N-methylmorpholine (redistilled), NH2OH,
thionylbromide, di-tert-butyl dicarbonate, sulfanilic acid, 4-(2-
aminoethyl)-benzenosulfonamide were purchased from Sigma–Al-
drich, and 4-aminobenzoic acid, 4-(methylamino)benzoic acid,
2,4-diamino-6-(hydroxylmethyl)pteridine hydrochloride, sulfanil-
amide from Acros Organics. The solvents were purchased from
Acros Organics or Merck and, whenever necessary, they were puri-
fied and dried according to standards methods.22 All moisture-sen-
sitive reactions were performed under nitrogen atmosphere. The
chemical reactions were followed by TLC using silica gel plates
(G-60 F254, Merck). A Bio-Rad Merlin, FTS 3000 MX spectrometer
was used to record solid state IR spectra (KBr pellets). 1H and 13C
NMR spectra were recorded on a Varian Unity 300 FT NMR spec-
trometer at 25 °C. When necessary, assignment of the signals of
the 13C NMR signals were confirmed by DEPT. Chemical shifts are
3.3. Antiproliferative effect
Antiproliferative effects of the synthesized compounds were
evaluated on two cancer cell lines: non-small cell lung carcinoma
(A549) and prostate carcinoma (PC3). Effects of the compounds
were compared to antiproliferative effects of MTX as a control. Cell
viability was assessed using MTT cell proliferation assay, as previ-
ously described.21 These assays demonstrated (see Table 1) that
PC3 cells, in general, are more sensitive to the inhibition by both
MTX (IC50 is fivefold lower) and by the synthesized compounds
(IC50 are 2–10-fold lower). However, despite the fact that most of
the synthesised compounds inhibited both DHFR and CAs at low
micromolar and low nanomolar concentrations, respectively, ef-
fects on cell proliferation were noted only at low millimolar con-
centrations for both cell lines. This suggests that, most likely, the

S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
5087
reported in ppm (d) with tetramethylsilane (TMS) as internal refer-
ence, in organic solvents and sodium 3-(trimethylsilyl)-[2,2,3,3-
D4]-propionate (DSS) in D2O solutions. The following abbreviations
are used: s = singlet; d = doublet; t = triplet; m = multiplet;
br = broad. Mass spectra (FAB) were performed in a VG TRIO-
2000 GC/MS instrument, and the ESI mass spectra on a 500 MS
LC Ion Trap (Varian Inc., Palo Alto, CA, USA) mass spectrometer
equipped with an ESI ion source, operated either in positive or neg-
ative ion modes. Elemental analyses were performed on a Fisons
EA 1108 CHNF/O instrument.
5.2.5. 4-(2-Aminoethyl)benzenesulfonic acid
To a solution of AEBS (0.50 g, 2.5 mmol) in a mixture of THF
(5 mL) and 2 M HCl solution (10 mL) was added NaNO2 (0.30 g,
4.25 mmol), and the mixture was stirred at 40 °C for 24 h. After
evaporation of the organic solvent, the resulting solution was taken
into CH3CN (10 mL), and the white precipitate was filtered and
dried. Recrystallization with dry methanol allowed removal of
the inorganic material, affording the pure hydrochloric salt of the
title compound as white crystals (0.416 g, 70% yield); mp higher
than 320 °C; 1H NMR (300 MHz, D2O) d: 7.76 (2H, d, CH@C–S),
7.42 (2H, d, CH2–C@CH), 3.29 (2H, t, NH–CH2), 3.04 (2H, t, CH2–
5.2. Synthesis of the compounds
C@CH); IR (KBr, cmꢁ1): 3134, 3078 (
mSO3H); MS (FAB) m/z: 202
(M+H)+.
5.2.1. 6-(Bromomethyl)pteridine-2,4-diamine (PtBr)
2,4-Diamino-6-(hydroxymethyl)-pteridine
(PtOH,
0.5 g,
5.2.6. 4-(2-((2,4-Diaminopteridin-6-yl)methylamino)ethyl)
benzenesulfonic acid (2b)
2.60 mmol) was dissolved in thionyl bromide (10 mL) and the reac-
tion mixture was stirred for 24 h at rt. Thionyl bromide was evap-
orated in vacuum and the solid residue was washed with large
amount of dry toluene (ca. 100 mL), affording the product as brown
crystals (0.623 g, 94% yield); mp 168–170 °C. 1H NMR (300 MHz,
DMSO-d6) d: 9.04 (1H, s, CH(Pt)), 8.60 and 7.97 (2H each, s,
NH2(Pt)), 4.89 (2H, s, CH2(Pt)); 13C NMR (DEPT) (300 MHz,
DMSO-d6) d: 149 (CH(Pt)), 30.9 (CH2(Pt)); IR (KBr, cmꢁ1): 578,
From PtBr and 4-(2-aminoethyl)benzenesulfonic acid, as in the
case of 1a, but using 3 equiv of NaH. Yellow crystals (40%), mp
higher than 300 °C; 1H NMR (300 MHz, DMSO-d6) d: 8.63 (1H, s,
CH(Pt)), 7.58 (2H, d, CH@C–S), 7.22 (2H, d, CH2–C@CH), 6.99 (1H,
m, NH), 5.01 (1H, s, SO3H), 3.94 (2H, s, CH2(Pt)), 2.75 (2H, t, NH–
CH2), 2.58 (2H, t, CH2–C@CH); MS (ESI) m/z: 398.1 (M+Na)+.
523, 721, 611 (
peaks).
m
C-Br); MS (FAB) m/z: 255, 257 (M+1, M+3) (twin
5.2.7. 4-(tert-Butoxycarbonylamino)benzoic acid (5)
A solution of 4-aminobenzoic acid (pABA, 1.03 g, 7.5 mmol) in
water (8 mL) and 1,4-dioxane (8 mL) was cooled in an ice bath at
0 °C, followed by the parallel addition of Na2CO3 (1.53 g, 15 mmol)
in water (10 mL) and di-tert-butyl-dicarbonate (Boc2O, 1.80 g,
8.25 mmol) in 1,4-dioxane (10 mL). The reaction mixture was stir-
red for 4 h at 0 °C and then it was left at rt overnight. The pH of the
solution was lowered up to 2 with 2 M HCl and the product precip-
itated as a white solid, which was filtered and washed with large
amount of HCl solution to give the pure product as white crystals
(1.69 g, 95% yield); mp 205–206 °C; 1H NMR (300 MHz, acetone-
d6) d: 7.85 (2H, d, CH@C–CO), 7.39 (2H, d, NH–C@CH), 1.51 (9H,
5.2.2. 4-((2,4-Diaminopteridin-6-yl)methylamino)
benzenesulfonamide (1a)
General procedure for preparation of the target compounds
(for 1a–4a and 1b–3b), from PtBr: a mixture of sulfanilamide
(0.200 g, 1.16 mmol) and NaH (80% in paraffin oil, 0.035 g,
1.16 mmol) in dry DMF (15 mL) was stirred for 40 min. Then
PtBr (0.295 g, 1.16 mmol) and CaCO3 (0.058 g, 0.58 mmol) were
added, and the mixture was stirred under nitrogen, in the dark
at rt for a week. The solvent was then evaporated under high
vacuum. The solid residue was taken into water and a yellow so-
lid precipitated, which was then filtered and washed with etha-
nol and diethyl ether. Recrystallization from DMF afforded the
pure compound as yellow crystals (0.370 g, 92% yield). Mp
265–266 °C; 1H NMR (300 MHz, DMSO-d6) d: 9.09 and 9.12
(2H each, s, NH2(Pt)), 8.82 (1H, s, CH(Pt)), 7.66 (2H, d, CH@C–
S), 7.06 (1H, t, NH), 6.98 (2H, s, SO2NH2), 6.79 (2H, d, NH–C@CH),
s, C(CH3)3); IR (KBr, cmꢁ1): 3390, 3423, 3484, 3566 (
(ESI) m/z: 236.4 (MꢁH)ꢁ.
mNH); MS
5.2.8. tert-Butyl 4-(4-sulfamoylphenethylcarbamoyl)
phenylcarbamate (6a)
To an ice-cooled solution of N-methylmorpholine (0.23 mL,
2.10 mmol) and 5 (0.25 g, 1.05 mmol) in dry DMF (10 mL) under
nitrogen was added TBTU (0.34 g, 1.05 mmol), and the mixture
was stirred for 50 min. A solution of AEBS (0.21 g, 1.05 mmol) in
dry DMF (5 mL) was then added and resulting solution was left
stirring for 5 h at 0 °C, and, then, it was left to heat up to rt and
was stirred overnight. The solvent was evaporated under high vac-
uum and the solid residue was taken into large amount of water
(50 mL), and then a pale yellow solid was precipitated, which
was filtered, washed with methanol (40 mL), and dried under vac-
uum, affording the pure product as white crystals (0.198 g, 45%
yield); mp 255–258 °C; 1H NMR (300 MHz, MeOD) d: 7.72 (2H, d,
CH@C–S), 7.58 (2H, d, CH@C–CO), 7.37 (2H, d, NH–C@CH), 7.31
(2H, d, CH2–C@CH), 3.50 (2H, t, NH–CH2), 2.88 (2H, t, CH2–C@CH),
1.41 (9H, s, C(CH3)3); MS (ESI) m/z: 442.1 (M+Na)+.
4.60 (2H, s, CH2(Pt)); IR (KBr, cmꢁ1): 3356, 3154 (
1315 ( SO2NH2), 569, 545 ( SN), 831 (mNH aromatic); MS (ESI) m/z:
347.1 (M+H)+.
mNH), 1151,
m
m
,
5.2.3. 4-((2,4-Diaminopteridin-6-yl)methylamino)
benzenesulfonic acid (1b)
From PtBr and sulfanilic acid, as in the case of 1a. Yellow
crystals (34%), mp higher than 300 °C. 1H NMR (300 MHz,
DMSO-d6) d: 9.08 and 8.90 (2H each, s, NH2(Pt)), 8.82 (1H, s,
CH(Pt)), 7.56 (2H, d, CH@C–S), 6.98 (1H, s, SO3H), 6.81 (2H, d,
NH–C@CH), 6.76 (1H, t, NH), 4.61 (2H, s, CH2(Pt)); MS (ESI) m/
z: 348.1 (M+H)+.
5.2.4. 4-(2-((2,4-Diaminopteridin-6-yl)methylamino)ethyl)
benzenesulfonamide (2a)
5.2.9. 4-Amino-N-(4-sulfamoylphenethyl)benzamide (7a)
Removal of the Boc protecting group was performed by stirring
a solution of 6a (0.15 g, 0.36 mmol) in trifluoroacetic acid (TFA,
1 mL) and CH2Cl2 (1 mL) at rt for 5 h. After evaporation of the sol-
vent, recrystallization from dry methanol/diethyl ether afforded
the pure product as white crystals (0.062 g, 54% yield); mp 243–
245 °C; 1H NMR (300 MHz, MeOD) d: 7.73 (2H, d, CH@C–S), 7.45
(2H, d, CH@C–CO), 7.35 (2H, d, CH2–C@CH), 6.65 (2H, d, NH2–
C@CH), 3.50 (2H, t, NH–CH2), 2.90 (2H, t, CH2–C@CH); 13C NMR
(DEPT) (300 MHz, DMSO-d6) d:134 (CH@C–CO), 132 (CH2–C@CH),
From PtBr and (4-(2-aminoethyl)-benzenesulfonamide (AEBS),
as in the case of 1a. Yellow crystals (14%); mp higher than
300 °C; 1H NMR (300 MHz, DMSO-d6) d: 8.63 (1H, s, CH(Pt)), 7.73
and 7.48 (2H each, s, NH2(Pt)), 7.63 (2H, d, CH@C–S), 7.32 (2H, d,
CH2–C@CH), 6.57 (1H, m, NH), 4.13 (2H, s, SO2NH2), 3.92 (2H, s,
CH2(Pt)), 2.89 (2H, t, NH–CH2), 2.73 (2H, t, CH2–C@CH); IR (KBr,
cmꢁ1): 3356 (
MS (ESI) m/z: 375.2 (M+H)+, 397.1 (M+Na)+.
mNH), 1159 (mSO2NH2), 707 (mSN), 875 (mNH,aliphatic);

5088
S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
130 (CH@C–S), 43.8 (NH–CH2), 120 (NH2–C@CH), 40.1 (CH2–
C@CH); IR (KBr, cmꢁ1): 3328, 3210 NH), 1190, 1342 cmꢁ1
SO2NH2), 593, 545 ( SN), 1726 (
CO); MS (FAB) m/z: 320 (M+H)+.
3220 (
mNH), 1330, 1317 (mSO2NH2), 598, 548, 650 (mSN); MS (ESI) m/
(
m
z: 508.2 (M+H)+.
(m
m
m
5.3. CA inhibition assays
5.2.10. 4-((2,4-Diaminopteridin-6-yl)methylamino)-N-(4-
sulfamoylphenethyl)benzamide (3a)
Recombinant human CA isoforms I, II and IX have been pre-
pared as reported earlier by our group,23,24 and their activity as-
sayed by a stopped flow CO2 hydration assay.25 An Applied
Photophysics (Oxford, UK) stopped-flow instrument has been used
for assaying the CA-catalyzed CO2 hydration activity. A phenol red
solution (0.2 mM) was used as indicator, working at the absor-
bance maximum of 557 nm, with 10 mM Hepes (pH 7.5) as buffer,
0.1 M Na2SO4 (for maintaining constant the ionic strength), follow-
ing the CA-catalyzed CO2 hydration reaction for a period of 10–
100 s. The CO2 concentrations ranged from 1.7 to 17 mM for the
determination of the kinetic parameters and inhibition constants.
For each inhibitor at least six traces of the initial 5–10% of the reac-
tion have been used for determining the initial velocity. The uncat-
alyzed rates were determined in the same manner and subtracted
from the total observed rates. Stock solutions of inhibitor (1 mM)
were prepared in distilled–deionized water with 10–20% (v/v)
DMSO (which is not inhibitory at these concentrations) and dilu-
tions up to 0.1 nM were done thereafter with distilled–deionized
water. Inhibitor and enzyme solutions were preincubated together
for 15 min at room temperature prior to assay, in order to allow for
the formation of the E–I complex. The inhibition constants were
obtained by non-linear least-squares methods using PRISM 3, from
Lineweaver–Burk plots, as reported earlier, and represent the aver-
age from at least three different determinations.23,24
From PtBr and 7a, as in the case of 1a. Yellow crystals (83%);
mp 280–281 °C; 1H NMR (300 MHz, DMSO-d6) d: 8.71 (1H, s,
CH(Pt)), 8.20 (1H, t, CONH), 7.97 and 7.85 (2H each, s, NH2(Pt)),
7.75 (2H, d, CH@C–S), 7.65 (2H, d, CH@C–CO), 7.42 (2H, d, CH2–
C@CH), 6.74 (2H, d, NH–C@CH), 6.63 (1H, m, NH), 4.50 (2H, s,
CH2(Pt)), 4.13 (2H, s, SO2NH2), 3.19 (2H, t, NH–CH2), 2.91 (2H,
t, CH2–C@CH); IR (KBr, cmꢁ1): 3351, 3220 (
SO2NH2), 590, 550 ( SN), 875 ( NH,aliphatic); MS (ESI) m/z: 494.2
(M+H)+.
mNH), 1191, 1327
(m
m
m
5.2.11. 4-(2-(4-(tert-Butoxycarbonylamino)benzamido) ethyl)
benzenesulfonic acid (6b)
From 5 and 4-(2-aminoethyl)benzenesulfonic acid, as in the
case of 6a. White crystals (32%); mp 200 °C (decomposition); 1H
NMR (300 MHz, D2O) d: 7.70 (2H, d, CH@C–S), 7.54 (2H, d,
CH@C–CO), 7.37 (4H, m, NH–C@CH, CH2–C@CH), 3.61 (2H, t, NH–
CH2), 2.95 (2H, t, CH2–C@CH), 1.47 (9H, s, C(CH3)3); MS (ESI) m/z:
419.1 (MꢁH)ꢁ.
5.2.12. 4-(2-(4-Aminobenzamido)ethyl)benzenesulfonic acid
(7b)
From 6b, by removal of the protecting group, as in the case of
7a. White crystals (70%); mp 280–290 °C (decomposition); 1H
NMR (300 MHz, D2O) d: 7.70 (2H, d, CH@C–S), 7.57 (2H, d,
CH@C–CO), 7.38 (2H, d, CH2–C@CH), 7.17 (2H, d, NH–C@CH),
3.65 (2H, t, NH–CH2), 2.98 (2H, t, CH2–C@CH); IR (KBr, cmꢁ1):
5.4. DHFR inhibition assays
3325, 3073 (
m
SO3H); MS (ESI) m/z: 321.1 (M+H)+.
DHFR activity (the enzyme from L. Casei was used), in the pres-
ence and in the absence of inhibitors, has been assayed spectro-
photometrically by monitoring the decrease in absorbance at
340 nm due to oxidation of NADPH. All assays were performed at
30 °C in a Shimadzu 2401PC double-beam spectrophotometer.
5.2.13. 4-(2-(4-((2,4-Diaminopteridin-6-yl)methylamino)
benzamido)ethyl) benzenesulfonic acid (3b)
From PtBr and 7b, as in the case of 1a. Yellow crystals (95%); mp
higher than 300 °C; 1H NMR (300 MHz, DMSO-d6) d: 8.76 (1H, s,
CH(Pt)), 8.15 (1H, t, CONH), 7.62 (2H, d, CH@C–S), 7.49 (2H, d,
CH@C–CO), 7.14 (2H, d, CH2–C@CH), 6.80 (1H, m, NH), 6.70 (2H,
d, NH–C@CH), 5.07 (1H, s, SO3H), 4.54 (2H, s, CH2(Pt)), 2.78 (4H,
The reaction mixture contained 100
150 M of NADPH and 2.5 g/mL purified DHFR. The reaction
was started by the addition of the enzyme (1–2.5 g) in a final vol-
lM of dihydro-folic acid,
l
l
l
ume of 1.0 mL and read against a blank cuvette containing all com-
ponents except the enzyme. Synthesized compounds were added
m, NH–CH2, CH2–C@CH); IR (KBr, cmꢁ1): 3325, 3073 (
mSO3H),
1647 (
m
CO); MS (ESI) m/z: 493.1 (MꢁH)ꢁ.
in the reaction mixture to final concentrations 0.1–1000
lM. For
a reference inhibition curve, MTX was added into reaction mixture
5.2.14. 4-(Methylamino)-N-(4-sulfamoylphenethyl)benzamide
(8)
to final concentrations of 0.001–0.1 lM.
From 4-(methylamino)benzoic acid (NMepABA) and AEBS, as in
the case of 6a. White crystals (82%); mp 264–267 °C; 1H NMR
(300 MHz, DMSO-d6) d: 8.24 (1H, t, CONH), 7.80 (2H, d, CH@C–S),
7.68 (2H, d, CH@C–CO), 7.47 (2H, d, CH2–C@CH), 7.34 (1H, m,
NHCH3), 6.70 (2H, d, NH–C@CH), 6.25 (2H, s, SO2NH2), 3.49 (2H,
t, NH–CH2), 2.85 (2H, t, CH2–C@CH), 2.76 (3H, s, NCH3); 13C NMR
(DEPT) (300 MHz, DMSO-d6) d: 134 (CH@C–CO, CH2–C@CH), 131
(CH@C–S), 115 (NH–C@CH), 44.4 (NH–CH2), 40.2 (CH2–C@CH),
5.5. Cell culture assays
The lung carcinoma cell line A549 was obtained from American
Type Culture Collection. The prostate carcinoma cell line PC3 was a
kind gift from Dr. James Norris, Medical University of South Caro-
lina. Cells were maintained in RPMI-1640 supplemented with 10%
heat-inactivated fetal bovine serum, 2 mM glutamine and 1 mM
sodium pyruvate (complete medium). All cells were grown at
37 °C under humidified air containing 5% CO2. Cells were plated
in 96-well plates at a density of about 5000 cell/well. Treatment
with different concentrations of a corresponding inhibitor was per-
formed constantly for 72 h. MTT cell proliferation assay was per-
formed using CellTiter 96 kit (Promega) according to
manufacturer’s directions.
34.5 (NCH3); IR (KBr, cmꢁ1): 3411, 3351, 3315 (
1164 ( SO2NH2), 740, 690, 543, 590 ( SN), 1524 (
z: 334 (M+H)+.
m
NH), 1334, 1311,
m
m
mCO); MS (FAB) m/
5.2.15. 4-((2,4-Diaminopteridin-6-yl)methyl)(methyl)amino)-N-
(4-sulfamoylphenethyl)-benzamide (4a)
From PtBr and 8, as in the case of 1a. Yellow crystals (75%); mp
245–250 °C; 1H NMR (300 MHz, DMSO-d6) d: 8.63 (1H, s, CH(Pt)),
8.23 (1H, t, CONH), 7.72 (2H, d, CH@C–S), 7.65 (2H, d, CH@C–CO),
7.40 (2H, d, CH2–C@CH), 7.28 (2H, s, NH2), 6.79 (2H, d, NH–C@CH),
4.81 (2H, s, CH2(Pt)), 4.38 (2H, s, SO2NH2), 3.44 (2H, t, NH–CH2),
3.24 (3H, s, NCH3), 2.89 (2H, t, CH2–C@CH); IR (KBr, cmꢁ1): 3355,
5.6. Molecular modelling
5.6.1. Docking of ligands with CA II and CA IX
The X-ray structures of hCA II and IX complexed with inhibitors
were taken from the RCSB Protein Data Bank (entries 1G54 and

S. M. Marques et al. / Bioorg. Med. Chem. 18 (2010) 5081–5089
5089
3IAI, respectively).17 With the purpose of further result compari-
son, these structures were aligned, using UCSF Chimera software.26
The crystal structures were then treated with Maestro 7.5,27 and all
counterions, co-crystallization ligands and solvent molecules were
removed. The hydrogen atoms were added using the all-atom
model, the ligands were extracted from the complex structure
and then saved in different files to be further used for defining
the binding site in the docking calculations. The structures of our
inhibitors were built using Maestro 7.5, and were minimized by
means of Macromodel.28 The conjugated gradient method was ap-
plied, until a convergence value of 0.05 kJ/Å mol was reached,
using the MMFFs force field and a water environment model (gen-
2002 (E. Enyedy) and SFRH/BPD/29874/2006 (S. Marques). We also
thank the Portuguese NMR and MS networks (IST-UTL Center), cre-
ated by the Portuguese Foundation for Science and Technology
(FCT), for providing access to their facilities. This research was also
financed in part by an EU grant of the 7th Fp (Metoxia projecy) to
C.T. Supuran.
All molecular modeling figures and alignments of the enzyme
structures were produced using the UCSF Chimera package, from
the Resource for Biocomputing, Visualization, and Informatics at
the University of California, San Francisco (supported by NIH Grant
P41 RR-01081).
eralized-Born/surface-area model), with
a distance-dependent
Supplementary data
dielectric constant of 1.0. The minimized ligands were then sub-
jected to a conformational search (CS) of 100 steps, in which an
algorithm based on the Monte Carlo method was used, with the
same force field and parameters as in the minimization.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmc.2010.05.072.
The minimized ligands were docked into the two CA structures
with the GOLD program,18 version 4.0, following a previously vali-
dated procedure for docking and virtual screening of ligands with
CAs.13 The region of interest used by Gold was defined in order to
contain the residues within 15 Å from the position of the original li-
gands in the X-ray structures. The ‘allow early termination’ option
was deactivated, while the possibility for the ligand to flip ring cor-
ners was activated. The zinc ion was set with a tetrahedral coordina-
tion, and the three water residues were allowed for spinning during
the docking, in order to find better hydrogen orientation. The
remaining Gold default parameters were used, and the ligands were
submitted to 200 genetic algorithm (GA) runs. The ChemScore fit-
ness function was used, and two protein H-bond constraints were
imposed, one between the hydroxyl O-atom and another with the
NH H-atom of Thr198 residue (hCAII numeration), and the ligands.
References and notes
1. Morphy, R.; Rankovic, Z. J. Med. Chem. 2005, 48, 6523.
2. Zimmermann, G. R.; Lehár, J.; Keith, C. T. Drug Discovery Today 2007, 12, 34.
3. Santos, M. A.; Enyedy, É. A.; Rossello, A.; Carelli, P.; Krupenko, N. I.; Krupenko, S.
A. Bioorg. Med. Chem. 2007, 15, 1266.
4. Marques, S. M.; Nuti, E.; Rossello, A.; Supuran, C. T.; Tuccinardi, T.; Martinelli,
A.; Santos, M. A. J. Med. Chem. 2008, 51, 7968.
5. Curtin, N. J.; Hughes, A. N. Lancet Oncol. 2001, 2, 298.
6. Gangjee, A.; Jain, H. D.; Phan, J.; Lin, X.; Song, X.; McGuire, J. J.; Kisliuk, R. L. J.
Med. Chem. 2006, 49, 1055.
7. Pastorek, J.; Pastoreková, S.; Callebaut, I.; Mornon, J. P.; Zelník, V.; Opavsky´, R.;
Zat’ovicová, M.; Liao, S.; Portetelle, D.; Stanbridge, E. J.; Zavada, J.; Burny, A.;
Kettmann, R. Oncogene 1994, 9, 2877.
8. Thiry, A.; Dogné, J.-M.; Masereel, B.; Supuran, C. T. Trends Pharmacol. Sci. 2006,
27, 566.
´
ˇ
9. Svastova, E.; Hulıkováa, A.; Rafajová, M.; Zat’ovicová, M.; Gibadulinová, A.;
Casini, A.; Cecchi, A.; Scozzafava, A.; Supuran, C. T.; Pastorek, J.; Pastoreková, S.
FEBS Lett. 2004, 577, 439.
10. Chiche, J.; Ilc, K.; Laferrière, J.; Trottier, E.; Dayan, F.; Mazure, N. M.; Brahimi-
Horn, M. C.; Pouysségur, J. Cancer Res. 2009, 69, 358.
11. Supuran, C. T.; Scozzafava, A.; Casini, A. Med. Res. Rev. 2003, 23, 146.
12. Supuran, C. T. Nat. Rev. Drug Disc. 2008, 7, 168.
5.6.2. Docking of ligands with DHFR
To validate the best docking procedure using Gold software, se-
ven crystal structures of DHFR complexes were taken from the
RCSB Protein Data Bank (entries 3DFR, 1U72, 2W3M, 2W3B,
3GHW, 2W3A, and 1BOZ), and these structures were treated with
Maestro 7.5, as described above for the CAs, removing all solvent,
ions and NADPH molecules when they were present. The respec-
tive ligands in each complex structure were energy-minimized
by means of Macromodel, as described before for the ligands. The
resulting structures were then docked into the respective protein
structures with Gold. In each case, the binding site was defined
as the atoms within 15 Å from the original position of the respec-
tive ligand. The ‘allow early termination’ option was deactivated,
while the possibility for the ligand to flip ring corners was acti-
vated. All the remaining Gold default parameters were used with
no further constraints, and the ligands were submitted to 100 GA
steps runs, using the three scoring functions supplied by Gold:
ASP, ChemScore and GoldScore. The best ranked solution in each
case was compared with the original conformation in the respec-
tive X-ray structure, and the rmsd (root mean square deviation)
was calculated for the heavy atoms. From the three fitness func-
tions tested, ASP gave the best docking results (average rmsd for
the seven docking results of 0.83 Å), and it was further used to dock
our ligands, following the same procedure. For this purpose, the
DHFR structure from the PDB complex 3DFR was used.
13. Tuccinardi, T.; Nuti, E.; Ortore, G.; Supuran, C. T.; Rossello, A.; Martinelli, A. J.
Chem. Inf. Model. 2007, 47, 515.
14. Briganti, F.; Pierattelli, R.; Scozzafava, A.; Supuran, C. T. Eur. J. Med. Chem. 1996,
31, 1001.
15. Kim, C.-Y.; Chang, J. S.; Doyon, J. B.; Baird, T. T., Jr.; Fierke, C. A.; Jain, A.;
Christianson, D. W. J. Am. Chem. Soc. 2000, 122, 12125.
16. Alterio, V.; Hilvo, M.; Di Fiore, A.; Supuran, C. T.; Pan, P.; Parkkila, S.; Scaloni, A.;
Pastorek, J.; Pastorekova, S.; Pedone, C.; Scozzafava, A.; Monti, S. M.; De
Simone, G. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 16233.
17. 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.
18. Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R. J. Mol. Biol. 1997, 267,
727.
19. The UniProt Consortium, The Universal Protein Resource (UniProt) in 2010,
Nucleic Acids.
20. Bolin, J. T.; Filman, D. J.; Matthews, D. A.; Hamlin, R. C.; Kraut, J. J. Biol. Chem.
1982, 257, 13650.
21. Studzinski, G. P., Ed.Cell Growth, Differentiation and Senescence: a Practical
Approach; Oxford University Press: New York, 1999.
22. Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory Chemicals, 4th ed.;
Butterworth-Heinemann: Oxford, 1999.
23. (a) Cecchi, A.; Hulikova, A.; Pastorek, J.; Pastorekova, S.; Scozzafava, A.; Winum,
J.-Y.; Montero, J.-L.; Supuran, C. T. J. Med. Chem. 2005, 48, 4834; (b) Casey, J. R.;
Morgan, P. E.; Vullo, D.; Scozzafava, A.; Mastrolorenzo, A.; Supuran, C. T. J. Med.
Chem. 2004, 47, 2337.
24. (a) Pastorekova, S.; Casini, A.; Scozzafava, A.; Vullo, D.; Pastorek, J.; Supuran, C.
T. Bioorg. Med. Chem. Lett. 2004, 14, 869; (b) Vullo, D.; Innocenti, A.; Nishimori,
I.; Pastorek, J.; Scozzafava, A.; Pastorekova, S.; Supuran, C. T. Bioorg. Med. Chem.
Lett. 2005, 15, 963.
25. Gupta, S. P.; Maheswaran, V.; Pande, V.; Kumar, D. J. Enzyme Inhib. Med. Chem.
2003, 18, 7.
26. Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.;
Meng, E. C.; Ferrin, T. E. J. Comput. Chem. 2004, 25, 1605.
27. Maestro, version 7.5. Schrödinger Inc.: Portland, OR, 2005.
28. Macromodel, version 8.5. Schrödinger Inc.: Portland, OR, 1999.
Acknowledgements
The authors would like to thank the portuguese Fundação para
a Ciência e Tecnologia for the post-doc Grants SFRH/BPD/11653/