Discovery of New Potential Anti-Infective Compounds Based on Carbonic Anhydrase Inhibitors by Rational Target-Focused Repurposing Approaches

Article abstract of DOI:10.1002/cmdc.201600180

In academia, compound recycling represents an alternative drug discovery strategy to identify new pharmaceutical targets from a library of chemical compounds available in house. Herein we report the application of a rational target-based drug-repurposing approach to find diverse applications for our in-house collection of compounds. The carbonic anhydrase (CA, EC 4.2.1.1) metalloenzyme superfamily was identified as a potential target of our compounds. The combination of a thoroughly validated docking screening protocol, together with in vitro assays against various CA families and isoforms, allowed us to identify two unprecedented chemotypes as CA inhibitors. The identified compounds have the capacity to preferentially bind pathogenic (bacterial/protozoan) CAs over human isoforms and represent excellent hits for further optimization in hit-to-lead campaigns.

Full text of DOI:10.1002/cmdc.201600180

DOI: 10.1002/cmdc.201600180
Full Papers
Discovery of New Potential Anti-Infective Compounds
Based on Carbonic Anhydrase Inhibitors by Rational
Target-Focused Repurposing Approaches
Giannamaria Annunziato+,[a] Andrea Angeli+,[b] Francesca D’Alba,[a] Agostino Bruno,*[a]
Marco Pieroni,[a] Daniela Vullo,[c] Viviana De Luca,[d] Clemente Capasso,[d]
Claudiu T. Supuran,*[b, c] and Gabriele Costantino[a]
In academia, compound recycling represents an alternative
drug discovery strategy to identify new pharmaceutical targets
from a library of chemical compounds available in house.
Herein we report the application of a rational target-based
drug-repurposing approach to find diverse applications for our
in-house collection of compounds. The carbonic anhydrase
(CA, EC 4.2.1.1) metalloenzyme superfamily was identified as
a potential target of our compounds. The combination of
a thoroughly validated docking screening protocol, together
with in vitro assays against various CA families and isoforms, al-
lowed us to identify two unprecedented chemotypes as CA in-
hibitors. The identified compounds have the capacity to prefer-
entially bind pathogenic (bacterial/protozoan) CAs over human
isoforms and represent excellent hits for further optimization
in hit-to-lead campaigns.
Introduction
Drug repurposing or repositioning denotes an ensemble of
tasks aimed at the identification of new drug indications for
existing drugs[1] and is an emerging alternative strategy in
drug discovery programs, in both pharma and academia fields.
In academia, drug repurposing can be also translated into
compound recycling, that is, the repurposing of compound li-
brary collections already available in house.[1–3] Indeed, already
synthesized small molecules that were inactive against a target
of interest can be tested on other targets, which can lead to
a new purpose for an old molecule. Potential new targets for
unemployed compounds can be identified through cross-anal-
ysis of the fragment composition of the whole compound li-
brary with the data available in the literature for a specific
class of protein (target-focused approach).[1,4–6] Indeed, in aca-
demia, compound recycling approaches have found great ap-
plication and have led to the development of several success-
ful repurposing strategies.[2,3]
In this scenario, we embarked on a project aimed at the re-
purposing of the compound libraries available in house to look
for new potential applications for our compounds. Therefore,
we proceeded with the analysis of the fragments and chemo-
types present in our libraries by applying the maximum
common substructure (MCS) decomposition approach.[7,8] The
analysis of the data available in the literature for similar classes
of chemical structures allowed us to identify the carbonic an-
hydrase (CA, EC 4.2.1.1) metalloenzyme family as a potential
target of some of our compound series. Prompted by these re-
sults, we set up and thoroughly validated a docking protocol
against all of the CA classes and isoforms crystallized so far.
Such a method allowed us to identify eleven compounds as
potential CA inhibitors (CAIs). The compounds were then
tested in vitro for their ability to inhibit different classes and
isoforms of the CA superfamily, which led to the discovery of
a series of CAIs that are active in the low mm range and are
characterized by 1) two unprecedented chemotypes for CAIs,
2) an unprecedented selectivity profile for this class of mole-
cules, with the ability to bind preferentially microbial CAs over
human ones, and 3) good ligand efficiencies and binding effi-
ciency indexes[9] with respect to that of marketed CAIs. Overall,
the newly reported CAIs represent interesting hits, which can
be further optimized as potential antimicrobial agents.
[a] Dr. G. Annunziato,+ Dr. F. D’Alba, Dr. A. Bruno, Dr. M. Pieroni,
Prof. G. Costantino
Universitꢀ degli Studi di Parma
Dipartimento di Farmacia, P4T group
Parco Area delle Scienze, Via G.P. Usberti 27A, 43121, Parma (Italy)
E-mail: agostino.bruno@unipr.it
[b] Dr. A. Angeli,+ Prof. C. T. Supuran
Universitꢀ degli Studi di Firenze
Neurofarba Dept., Section of Pharmaceutical and Nutriceutical Sciences
Via U. Schiff 6, 50019 Sesto Fiorentino, Florence (Italy)
E-mail: claudiu.supuran@unifi.it
[c] Dr. D. Vullo, Prof. C. T. Supuran
Universitꢀ degli Studi di Firenze
Polo Scientifico, Laboratorio di Chimica Bioinorganica, Rm. 188
Via della Lastruccia 3, 50019 Sesto Fiorentino, Florence (Italy)
[d] Dr. V. De Luca, Dr. C. Capasso
Istituto di Bioscienze e Biorisorse, CNR
Via Pietro Castellino 111, 80131 Napoli (Italy)
[+] These authors contributed equally to this work.
Supporting information for this article can be found under http://
dx.doi.org/10.1002/cmdc.201600180.
ChemMedChem 2016, 11, 1 – 12
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Docking studies were performed by means of Glide software,
and during the validation steps of the docking procedure, we
observed that several CA X-ray crystal structures have a glycerol
molecule in the CA active site establishing contacts with the
co-crystallized ligand (figure S1 in the Supporting Information).
As a result, the stripping of the glycerol molecule (during the
preparation phase of the docking protocol) led to a bad place-
ment of the co-crystallized inhibitors during the re-docking
phases (see the Supporting Information for details). Thus, all
CA X-ray crystal structures with co-crystallized molecules
bound within the active site were discharged from further
analysis. Once we had identified the best docking parameters
and the proper X-ray dataset, we screened the compound col-
lection. When possible, we used more than one X-ray crystal
structure per CA isoform, in order to include side chains and
protein flexibilities in our calculations. In Figure 2, the heat
map scores for both human isoforms (Figure 2A) and microbial
families (Figure 2B) are depicted. From the analysis of Fig-
ure 2A, it can be noticed that 1) several compounds have
good Glide scores (>7.5Æ0.1 kcalmolÀ1, dark blue) against all
but hCA III isoforms (hCA III inset in Figure 2), so hCa III is appa-
rently not druggable by our chemotypes, and 2) some com-
pounds display preferential interactions with specific confor-
mations of the same hCA isoform. For example, compound C1
Results and Discussion
Repurposing of the in-house 3D chemical library
The knowledge of the main chemotypes/fragments present in
a chemical library can direct the identification of new potential
targets for compounds in that library. Therefore, we extrapolat-
ed and analyzed the main fragments present in our in-house
chemical library, composed of 12 main fragments/chemotypes,
by applying the MCS approach.[7,8] The analysis revealed that,
among others, the most abundant fragments are 2-aminothia-
zole (1), 2-phenylcyclopropane-1-amino (2), 2-hydroxypyridini-
um (3), pyridine-N-oxide (4), cyclopropanecarboxylic acid(5),
and benzoic acid (6; Figure 1). Taking into account that most
of the compounds available in our lab were designed as anti-
microbial agents[10–14] or inhibitors of metalloenzymes,[15] we
decided to focus our attention on all of those targets that sat-
isfied the following criteria: 1) they are metalloenzymes ubiqui-
tously expressed in most living organisms, and 2) there are re-
ported inhibitors that are characterized by fragments similar to
those present in our library. An extensive knowledge-based lit-
erature search allowed us to identify the carbonic anhydrase
superfamily as an eligible protein class.[16–20] Indeed, sulfona-
mides are the most common CAIs[16,18] (similar to fragment 12,
in Figure 1), but recently other CAI chemotypes have been dis-
closed, such as 2-aminothiazole[16,18] (similar to 1), phenols[21]
(similar to 3), hydroxypyridin(thi)ones[22,23] (similar to 4), benzo-
ic acid[2] (similar to 5, 6, and 8), coumarins[24,25](similar to 7),
and polyamines[26] (not present in our database).
(inset in Figure 2) shows good scores (>7.5Æ0.1 kcalmolÀ1
,
dark blue) for only a decreased number of hCA II conforma-
tions, which supports the observation that small conformation-
al changes in the CA active site can have important effects on
the docking results.
Encouraged by these findings, we set up a computational
protocol aimed at the identification of potential CAIs among
our compound collection. To this aim, we collected all the CA
X-ray crystal structures available in the Protein Data Bank from
different organisms, families, and isoforms (about 750 struc-
tures). A docking protocol was designed and optimized in
a stepwise pipeline, with the aim of 1) identifying the best
docking parameters (by means of the re-docking procedure;
see the Supporting Information), 2) identifying the CA X-ray
crystal structures most suitable for our docking purposes, and
finally 3) reliably screening the in-house collection library.
If docked on the microbial CAs (Figure 2B), our compounds
generally displayed low Glide scores. This can be explained by
the fact that most of these CAs belong to the b family, which
is characterized by a narrower and more solvent-exposed
active site (figure S2 in the Supporting Information). Exceptions
are represented by those microbial CAs characterized by the
a-family fold type (that is, the Sulfurihydrogenibium bacterial
CAs, highlighted by the black bar and arrow in Figure 2B),
which show higher Glide scores than the others.
Figure 1. Structures of the main fragments identified in our compound collection library. The asterisks highlight the main points of substitution for the report-
ed scaffolds.
&
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Figure 2. A) Heat map scores for the human (h) CAs. The red lines indicate separation between the different hCAs in the following order: hCA I, II, III, IV, VII,
XII, and XIII. Highlighted in the insets are the scores for the hCA III isoform and for two different compounds (C1 and C2). B) Heat map scores for microbial
CAs. Scores are reported as Glide scores in kcalmolÀ1 (indicated by the colored bar beside the heat map).
Figure 3. Structures of compounds selected to be tested in vitro.
In vitro assays against different CA isoforms and families
yet available and so was not present in our initial structure da-
On the basis of the preliminary computational observations,
we selected a small set of ligands (Figure 3; see the Experi-
mental Section for the compound selection strategy) to be
tested in vitro against three hCA isoforms (I, II, and III) and
against the Sulfurihydrogenibium yellowstonensis (Ssp) CA.
Table 1 lists the obtained inhibition constant (Ki) values for the
selected compounds (Figure 3), along with the corresponding
Glide scores produced by the docking calculations. The experi-
mental results are in line with our computational predictions,
and compounds 13, 15, 17, and 18 are endowed with mm af-
finity against to hCA I and II. As predicted by our calculation,
none of the compounds was active against hCA III (>100 mm).
Again in line with the docking prediction, the compounds re-
ported in Figure 3 showed Ki values in the low mm range for
the bacterial a-family SspCA (Table 1).
taset. Nevertheless, PfCA represents an already identified
target for treating malaria,[16,27,28] and it belongs to the h
family, which is the most similar to the a family with respect to
the other CA genetic families.[29,30] In Table 2, the Ki values for
PfCA are reported and, as expected, the compounds show in-
hibitory activities similar to those observed for SspCA. Even
more interestingly, they all have a significant selectivity over
the human CAs. The affinity of our compounds, either pyri-
dine-N-oxides or phenylcyclopropane carboxylates, is much
lower than that of reference compounds such as AAZ. Never-
theless, it must be stressed that our compounds come from
a recycling approach and that they were optimized against
very different targets.[10–15] We are confident that some of them
can still be highly optimized against microbial CAs. For in-
stance, two widely used metrics, namely the ligand efficiency
(LE) and the binding efficiency index (BEI), reported in Table 2,
indicate that most of our compounds have potential for large
improvement in further hit-to-lead optimization cycles.
Intrigued by the possibility to identify new CAI chemotypes
with the ability to preferentially bind microbial CAs over
human CAs, we decided to challenge our compounds against
another CA family that is relevant for the treatment of human
infections. Plasmodium falciparum, the causative agent of ma-
laria, expresses a CA (PfCA). The crystal structure of PfCA is not
Among the compounds reported in Figure 3, compound 13
was available in our lab as a racemic mixture, whereas com-
pound 14 was a single enantiomer (1R,2S). Therefore, we pro-
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
the data indicates that hCA isoforms show an appre-
ciable stereospecificity, because compound 24 is
active against hCA II, whereas compound 25 is active
against hCA I, which thus highlights the possibility to
obtain isoform selectivity by modulating the stereo-
chemistry of the phenylcyclopropane carboxylate
scaffold. The introduction of the ethyl moiety on the
aforementioned scaffold has a detrimental effect on
the activity against the human isoforms, because
both compounds 14 and 26 are completely inactive
against the human CAs. On the other hand, such
a modification of the phenylcyclopropane carboxyl-
ate scaffold led to a slight improvement of the activi-
ty for SspCA, because 14 is more active than 24, 25,
and 26 (14>26>25>24; see Table 2 and Table 3).
In contrast, PfCA seems to be unaffected by the
stereochemistry and the modification of the main
core (phenylcyclopropane carboxylate), because com-
pounds 13, 14, 24, 25, and 26 show similar activity
against PfCA (Table 2 and Table 3).
Table 1. Inhibition constant values for hCA I, II, and III and SspCA.[a]
Compd
Ki [mm]
hCA I
hCA II
hCA III
SspCA
13
70.6Æ6.3
(À7.72Æ0.46)
100Æ8.2
70.3Æ5.9
>100
2.80Æ0.21
(À7.37Æ0.42)
(À9.01)
14
>100
(À6.41Æ0.45)
>100
(À5.29Æ0.52)
>100
(À4.55Æ0.38)
>100
(À7.35Æ0.42)
>100
(À4.81Æ0.54)
>100
(À7.53Æ0.70)
>100
(À7.12Æ0.59)
>100
(À7.58Æ0.67)
>100
(À7.18Æ0.45)
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
>100
1.71Æ0.11
(À7.73)
3.45Æ0.22
(À6.32Æ0.37)
71Æ5.8
15
(À5.42Æ0.46)
100Æ6.3
(À8.91)
16
36.30Æ3.5
(7.71)
(À4.84Æ0.52)
90.3Æ8.0
17
5.96Æ0.42
(À8.02Æ0.43)
91.7Æ8.3
(7.43)
18
25.70Æ2.4
(À8.94)
(À7.55Æ0.65)
72Æ5.9
19
8.20Æ0.63
(À8.00)
(À7.53Æ0.34)
92.7Æ7.5
20
1.89Æ0.13
(À9.60)
(À7.25Æ0.24)
21
>100
3.47Æ0.20
(À7.82)
(À7.28Æ0.30)
>100
(À7.49Æ0.37)
>100
(À5.57Æ0.34)
0.25Æ0.008
22
2.59Æ0.16
(À7.21)
Finally, compound 27 shows similar activity against
all CAs considered in this study relative to 15. This
observation suggests that the chlorine atom and the
amino group in position 2 of the pyridine-N-oxide
ring are well tolerated, whereas substitutions at the
other positions of the main core play a crucial role in
terms of selectivity and activity.
23
35.20Æ1.8
(À8.62)
(À5.62Æ0.42)
0.012Æ0.001
AAZ[b]
0.17Æ0.006
0.005Æ0.0001
[a] Values are the mean Æ standard deviation (SD) of three different assays, n=
3. Glide scores (kcalmolÀ1) obtained for each ligand are given in parentheses.
For hCA I and II, the average Glide scores and standard deviations among differ-
ent conformations are reported, whereas only one X-ray structure is available for
SspCA. Glide scores >7.5Æ0.1 kcalmolÀ1 are listed in bold. [b] Acetazolamide.
Binding mode and selectivity against the different
CA isoforms and families
Table 2. Inhibition constant values for SspCA and PfCA, with the corresponding fold
selectivity, LE, and BEI values.
Among the tested compounds, 18 is quite a promis-
ing one, because it shows the best combination of
selectivity over hCAs, LE and BEI parameters, and, in-
terestingly, was devoid of any activity toward the
target for which was originally designed,[15] which
thus highly decreases the risk of cross-interaction
with other receptors/enzymes.
Compd
Ki [mm][a]
Selectivity[b]
SspCA PfCA SspCA PfCA SspCA PfCA
LE[c]
BEI[d]
SspCA
PfCA
13
14
15
16
17
18
19
20
21
22
23
AAZ
2.80Æ0.21
1.71Æ0.11
3.45Æ0.22
36.3Æ3.5
5.82Æ0.23
6.73Æ0.31
39.00Æ2.0
5.71Æ0.44
9.07Æ0.72
6.38Æ0.30
6.09Æ0.51
37.60Æ1.84
39.80Æ2.16
41.30Æ4.01
8.07Æ0.62
25.11 12.08 0.648 0.611 0.048 0.045
58.48 14.86 0.577 0.517 0.042 0.038
20.58
2.76 17.51 0.414 0.489 0.030 0.036
15.15 9.96 0.522 0.504 0.037 0.036
1.82 0.956 0.772 0.059 0.048
5.96Æ0.42
25.7Æ2.4
In Figure 4, the proposed binding modes of 18
into the hCA I–III and SspCA structures are depicted.
Compound 18 is nicely accommodated in the SspCA
active site, with the N-oxide moiety chelating the
Zn2+ ion. The 2-amino and 3-methyl ester groups are
accommodated in the hydrophilic half of the active
site, defined by the Asn62, Thr65, and Gln67 residues,
whereas the 5-ethyl moiety protrudes into the hydro-
phobic half, which is defined by Val110, Val120,
Val173, and Val182. Finally, the pyridine ring engages
a p–H-bond interaction with Thr175 (Figure 4A).
Through a comparison of the binding modes of the
3.57 14.37 0.459 0.520 0.033 0.037
8.78 11.82 0.548 0.562 0.039 0.040
8.20Æ0.63
1.89Æ0.13
3.47Æ0.20
2.59Æ0.16
35.0Æ1.8
49.04
28.82
38.61
2.46 0.616 0.477 0.040 0.031
2.51 0.637 0.513 0.041 0.033
2.42 0.711 0.558 0.045 0.035
2.84 12.39 0.520 0.594 0.033 0.038
2.40 0.07 0.894 0.729 0.052 0.043
0.005Æ0.0001 0.17Æ0.003
[a] Values are the meanÆSD of three different assays, n=3. [b] Calculated as hCA Ki
(with the lowest Ki value determined)/(Ssp or Pf)CA Ki. [c] Ligand efficiency (LE)=
RTpKi/N, in which N is the number of non-hydrogen atoms. [d] Binding efficiency
index (BEI)=RTpKi/Mr, in which Mr is the molecular weight in kDa.
ceeded with the resolution of the racemic mixture of 13
[24(1S,1S) and 25(1R,1R)] and with the synthesis of the 1S,2R
enantiomer of 14 (26). Moreover, to expand the exploration
around the pyridine-N-oxide scaffold, 2-amino-pyridine-N-oxide
was also synthesized (27). The compounds were tested against
all of the CA families and isoforms previously described, and
the data are reported in Table 3. Interestingly, the analysis of
same compound in hCA I (Figure 4B) and hCA III (Figure 4c), it
is clear that amino acid differences in both the hydrophilic and
hydrophobic halves (Figure 4d) of the active site explain:
1) the reduced activity of 18 for hCA I and 2) the lack of activi-
ty toward hCA III. Indeed, 18 in hCA I is no longer able to es-
tablish the p–H-bond interaction observed in SspCA. Moreover
Asn62 and Val120 are substituted with Val62 and Ala121 in the
&
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
file toward the different CAs considered (Figure 4). The en-
gagement of specific interactions with strategic residues inside
the CA active site could lead to the design of selective CAIs
that preferentially bind microbial CAs over human CAs.
Table 3. Inhibition constant values for hCA I, II, III, and SspCA.
Compd
Ki [mm][a]
hCA III
hCA I
>100
hCA II
SspCA
PfCA
24
25
26
27
41.2Æ2.0 >100 8.44Æ0.42 6.87Æ0.37
66.3Æ4.8 >100
>100 6.19Æ0.25 7.48Æ0.39
>100 3.66Æ0.12 8.24Æ0.50
>100 6.36Æ0.33 49.8Æ3.9
Chemistry
>100
>100
83.5Æ7.1 >100
Compounds 17, 18, and 19 were obtained by starting from
the commercially available 2-chloro-6-methylnicotinonitrile,
which reacted with methyl iodide in the presence of NaH as
a base to yield compound 29.[15] The chlorine atom in posi-
tion 2 was substituted by an amino group through the reac-
tion of compound 29 with benzylamine, conducted in a micro-
wave oven, which led to compound 30. This compound was
debenzylated with 98% H2SO4. Subsequently, a basic hydrolysis
in the presence of 10% KOH at reflux temperature[15] gave the
acid 32, which was methylated by an esterification reaction in
the presence of trimethylsilyldiazomethane (TMS-diazome-
thane).[15] The desired compound 18 was obtained by oxida-
tion of the pyridine nitrogen atom with methyltrioxorhenium(-
VII) (MTO) in a catalytic amount;[15] subsequent basic hydrolysis
of the ester gave desired compound 19. Finally, compound 18
through a reaction with hydroxylamine and sodium methoxide
as a base[31] gave compound 17 with a high yield (Scheme 1).
For the synthesis of compound 16, a protocol already re-
ported was followed,[15] in which the last step was conducted
in a sealed tube in the presence of sodium azide (Scheme 2).
Compounds 20, 21, 22, and 23 were obtained through a syn-
thesis already reported (Scheme 3).[15]
[a] Values are the meanÆSD of three different assays, n=3.
Figure 4. A) Binding mode of compound 18 into the SspCA active site
(white sticks and cartoon), B) into the hCA I active site (cyan sticks and car-
toon), and C) into the hCA III active site (yellow sticks and cartoon). D) Hy-
drophilic (blue surface) and hydrophobic (red surface) halves of hCA I (gray
surface).
The oxidation of compounds 40 and 41, which are commer-
cially available, was performed with m-CPBA in dichlorome-
thane and gave the desired compounds 15 and 27
(Scheme 4).[22]
Compounds 14 and 26 were prepared through a straightfor-
ward protocol that has already been reported (Scheme 5).[14]
Compound 13 was commercially available. The enantiopure
compounds 24 and 25 were obtained from compound 13 as
a racemic mixture that was treated with (R)-(À)-2-phenylglyci-
nol[32] to obtain the two diastereomers (+)-46 and (À)-46.
Compounds (+)-46 and (À)-46 were separated through flash
chromatography and hydrolyzed in acidic conditions[32] to give
compounds 24 and 25 as enantiopure compounds (Scheme 6).
human isoform, which leads to a less favorable interaction of
18 with the hCA I active site. On the other hand, 18 cannot
adopt a similar binding mode in hCA III as a result of two im-
portant amino acid differences: 1) His64 (in SspCA) is substitut-
ed with Lys64 (in hCA III), which clashes with the 3-methyl
ester group and 2) the hCA III active site is further narrowed by
the presence of Phe198; these differences hamper the proper
accommodation of the 6-ethyl moiety as observed in SspCA or
hCA I.
Compound 18 can be considered as a prototype structure
for a new class of zinc-chelating agents as CAIs. Most of the
CAIs so far reported have a sulfonamide group playing the
principal role in the binding to the CA active sites. As a result
of the prominent role played by the sulfonamide in the inter-
action with CA active sites, modifications at the core structure
have only decreased effects in terms of affinity/selectivity. We
can hypothesize that, because the N-oxide is a weaker zinc-
chelating group, compounds bearing this functionality will be
more amenable to modification of the organic core with
a much higher potential for better control of the affinity and
selectivity profile, with respect to the sulfonamide structures.
As a matter of fact, the elucidation of the binding mode of 18
shed light on the structural basis underlying its selectivity pro-
Conclusions
In the context of the economic downturn and the decreased
success rate in drug development programs that is hampering
classical drug discovery activities, medicinal chemists have to
identify alternative strategies to overcome productivity prob-
lems.[33,34] In academia, one possibility can be represented by
the ability to disclose new purposes for compound libraries al-
ready available in house. By applying a rational target-focused
repurposing approach, we were able to identify a new purpose
for some compounds in our whole collection. First of all, we
proceeded with the analysis of the fragments and chemotypes
present in our libraries by applying the MCS approach.[7,8] The
analysis of the data available in the literature for similar classes
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Scheme 1. Reagents and conditions: a) NaH, CH3I, DMF, 15 min at 08C, 4 h at RT, 68%; b) benzylamine, 150 W microwaves, 2008C, 15 min, 86%; c) 98% H2SO4,
RT, 18 h, 98%; d) 10% KOH, 1008C, 3 h, 98%; e) TMS-diazomethane, toluene/methanol, 08C, 30 min, 95%; f) MTO, 35% aqueous H2O2, EtOH, RT, 3 h, 78%;
g) NH2OH·HCl, MeONa, MeOH, RT, 18 h, 76%.
Scheme 2. Reagents and conditions: a) NaH, CH3I, DMF, 15 min at 08C, 4 h at RT, 68%; b) EtOH/NH3, sealed tube, 24 h, 2008C, 45%; c) MTO, 35% aqueous
H2O2, EtOH, RT, 3 h, 78%; d) NaN3, NH4Cl, DMF, 1308C, 3 h, 68%.
Scheme 3. Reagents and conditions: a) TMS-diazomethane, toluene/methanol, 08C, 30 min, 95%; b) meta-chloroperoxybenzoic acid (m-CPBA), CH2Cl2, 4 h, RT,
67%; c) 10% KOH, 1008C, 3 h, 98%.
are characterized by a low mm affinity for microbial CAs. The
obtained modest Ki values are reasonable because the mole-
cules used were not designed to target primarily CAs. Even if
the activity profile of the compounds needs to be improved,
the identified molecules may represent excellent hits to be fur-
ther optimized in hit-to-lead campaigns. Indeed, compound 18
seems to be the most promising one, and the analysis of its
binding mode in the different CAs allowed us to pave the way
for the design of potential selective CAIs that preferentially in-
hibit microbial CAs. Indeed, the combination of the herein-
identified zinc-chelating group with organic cores that are
properly decorated may further increase activity and selectivity
Scheme 4. Reagents and conditions: a) m-CPBA, CH2Cl2, 4 h, RT, 67%.
of chemical structures allowed us to identify the carbonic an-
hydrase (EC 4.2.1.1) metalloenzyme family as a potential target
of our compound libraries. Modeling studies, together with in
vitro assays, allowed us to identify new CAI chemotypes, which
&
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Scheme 5. Reagents and conditions: a) NaH, CH3CH2I, 1,2-dimethoxyethane (DME), 2 h at RT, 2 h at 608C, 65%; b) nBuLi, DME, 30 min, RT, 18 h, 908C, 73%;
c) LiOH, THF, MeOH, H2O, microwaves, 10 min’, 1008C, 85%.
Scheme 6. Reagents and conditions: a) (R)-(À)-2-phenylglycinol, O-(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium tetrafluoroborate (TBTU), 3-(3-dimethylami-
nopropyl)-1-ethylcarbodiimide (EDC)·HCl, triethylamine, CH2Cl2, 1 h at 08C, 4 h at RT, 75%; b) 3n H2SO4, dioxane, 18 h, 1008C, 98%.
General procedure for the esterification of nicotinic acids: (Tri-
methylsilyl)diazomethane (2.0m in diethyl ether; 2.00 equiv) was
added dropwise, over a period of 15 min, to a stirred, cooled (08C)
suspension of substrate (1.00 equiv) in dry toluene/methanol (3:2
v/v; 10 mLmmolÀ1) under a nitrogen atmosphere. The reaction
mixture was stirred at this temperature for 30 min. At the end of
this time, the reaction mixture became a clear, yellow solution.
TLC, by elution with chloroform/methanol (9:1), showed complete
consumption of the starting material. The reaction was stopped.
The solvents were evaporated under reduced pressure. The crude
residue, dissolved in chloroform, was washed with saturated aque-
ous NaHCO3 and brine. The organic phase was dried over anhy-
drous Na2SO4, filtered, and concentrated under reduced pressure.
The desired product was purified by silica flash chromatography,
by elution with chloroform, then chloroform/methanol (95:5). Com-
pounds 33, 38, and 39 were obtained by this method (75–95%
yield). Analytical data for these compounds matched the data al-
ready published.[15]
against CAs of human pathogens characterized by the a- and
h-family fold types.
Experimental Section
Chemistry
Materials and methods: All reagents were purchased from Sigma–
Aldrich, Alfa-Aesar, and Enamine at reagent purity and, unless oth-
erwise noted, were used without any further purification. Dry sol-
vents used in the reactions were obtained by distillation of techni-
cal-grade materials over appropriate dehydrating agents. Multi-
component reactions were performed by using a CEM Microwave
Synthesizer, Discover model. Reactions were monitored by thin
layer chromatography on silica-gel-coated aluminum foils (silica gel
on Al foils, SUPELCO Analytical, Sigma–Aldrich) at both 254 and
365 nm wavelengths. Where indicated, intermediates and final
products were purified through Merck silica gel 60 flash chroma-
tography (silica gel, 0.040–0.063 mm) by using appropriate solvent
mixtures.
General procedure for the oxidation of the pyridine nitrogen
atom: Method A: A 1:2 (v/v) solution of 35% (w/w) aqueous H2O2
in absolute ethanol was stirred over anhydrous Na2SO4 (5 g/30 mL)
at room temperature for 3 h. After filtration, MTO (0.10 equiv) was
added to this oxidant solution (4 mLmmolÀ1), followed by sub-
strate (1.00 equiv). The reaction mixture was stirred at room tem-
perature for 4 h, after which TLC, by elution with chloroform/meth-
anol (9:1), showed complete consumption of starting material. The
reaction was stopped. Brine was added to the reaction mixture,
and it was extracted with chloroform. The organic phase was dried
over anhydrous Na2SO4, filtered, and concentrated under reduced
pressure. The desired compounds were purified by silica gel flash
chromatography, by elution with chloroform, then chloroform/
methanol (from 99:1 to 95:5). Compounds 18 and 35 were ob-
tained by this method (68–88% yield). Analytical data for these
compounds matched the data already published.[15]
1H NMR and 13C NMR spectra were recorded on a BRUKER AVANCE
spectrometer at 300 and 75.5 MHz, respectively. 1H NMR spectra
are reported in this order: multiplicity and number of protons.
Standard abbreviations indicating the multiplicity were used as fol-
lows: s: singlet; d: doublet; dd: doublet of doublets; t: triplet; q:
quadruplet; m: multiplet; br: broad signal. HPLC–MS experiments
were performed with an Agilent 1100 series HPLC instrument,
equipped with a Waters Symmetry C18, 3.5 mm, 4.6ꢁ75 mm
column, and an Applied Biosystem/MDS SCIEX mass spectrometer,
equipped with an API 150EX ion source, or, alternatively, HPLC–MS
experiments were performed on an Acquity UPLC apparatus,
equipped with a diode array and a Micromass SQD single quadru-
ple (Waters). HRMS experiments were performed with an LTQ ORBI-
TRAP XL THERMO instrument. All compounds were tested as 95%
purity samples or higher (by HPLC–MS).
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
General procedure for the oxidation of the pyridine nitrogen
atom: Method B: The appropriate substrate (1 equiv) was com-
bined with m-CPBA (2 equiv) in dichloromethane (4 mLmmolÀ1),
and the mixture was stirred at room temperature for 4 h. After this
time, TLC, by elution with dichloromethane/methanol (95:5),
showed complete consumption of starting material. The solution
was washed with saturated aqueous NaHCO3, and the solvent was
evaporated under reduced pressure. The product was purified by
silica gel flash chromatography, by elution with dichloromethane/
methanol (95:5). Compounds 15, 20, 23, and 27 were obtained by
this method (56–67% yield). Analytical data for these compounds
matched the data already published.[15]
petroleum ether/ethyl acetate (9:1), showed almost complete con-
sumption of the starting material. The reaction was stopped. The
reaction mixture was cooled at 08C, then 1m NaOH aqueous solu-
tion was carefully added. The mixture was extracted with chloro-
form. The organic phase was washed with brine, dried over
Na2SO4, filtered, and concentrated under reduced pressure. The
crude residue was used for the next step without further purifica-
tion.
Synthesis of 2-amino-6-ethyl-3-(hydroxycarbamoyl)pyridine 1-
oxide (17): Compound 18 (1.00 equiv) was added carefully, drop-
wise through dropping funnel, to a suspension of NH2OH·HCl
(1.5 equiv) and MeONa (1.5 equiv) in MeOH at 08C. The reaction
was stirred at room temperature for 18 h. TLC after this time, by
elution with CH2Cl2/MeOH (85:15), showed complete consumption
of the starting material. The reaction was stopped. The reaction
mixture was cooled to 08C and 1n HCl aqueous solution was care-
fully added. The mixture was extracted with chloroform. The or-
ganic phase was washed with brine, dried over Na2SO4, filtered,
and concentrated under reduced pressure. The crude residue was
purified by silica gel flash chromatography, by elution with chloro-
form/methanol/formic acid (8.7:1:0.3), to afford the target product
(76%) as a white solid.
General procedure for the hydrolysis of methyl esters and 2-
amino-6-ethylnicotinamide (33) to the corresponding carboxylic
acids: A fine suspension of the substrate in 10% (w/v) aqueous
KOH (1 mLmmolÀ1) was heated at reflux and stirred at this temper-
ature for 3 h. TLC after this time showed complete consumption of
the starting material. The reaction was stopped. After cooling at
room temperature, the clear, colorless solution so obtained was
cooled to 08C and carefully acidified to pH 4–5, with 3n aqueous
HCl. A white precipitate immediately formed. The desired product
was collected by filtration in vacuo in quantitative yield. Com-
pounds 32, 19, 21, and 22 were obtained by this method. Analyti-
cal data for these compounds matched the data already pub-
lished.[15]
Synthesis of 2-amino-6-ethylnicotinonitrile (34): A solution of the
substrate in ammonia-saturated ethanol (1 mLmmolÀ1) was placed
in a sealed tube, heated to 2008C, and maintained at this tempera-
ture for 24 h. After the mixture had cooled to room temperature,
the solvent was removed by evaporation in vacuo, and the desired
product was purified by silica gel flash chromatography, by elution
with petroleum ether/ethyl acetate (from 9:1 to 6:4), to afford the
target product (45%). Analytical data for this compound matched
the data already published.[15]
Synthesis of 2-chloro-6-ethylnicotinonitrile (29): Sodium hydride
(60% dispersion in mineral oil; 1.50 equiv) was carefully added,
under a nitrogen atmosphere, to a stirred, cooled (08C) solution of
2-chloro-6-methylnicotinonitrile (28; 1.00 equiv) in dry N,N’-dime-
thylformamide (2.5 mLmmolÀ1). After the reaction mixture had
stirred at this temperature at this temperature for 15 min, iodome-
thane (4.00 equiv) was added. The reaction mixture was allowed to
warm to room temperature and stirred for 4 h under nitrogen. TLC
after this time, by elution with petroleum ether/ethyl acetate (7:3),
showed almost complete consumption of the starting material.
The reaction was stopped. The reaction mixture was cooled to
08C, then water was carefully added. The mixture was extracted
with diethyl ether. The organic phase was washed with brine, dried
over anhydrous Na2SO4, filtered, and concentrated under reduced
pressure. The crude residue was purified by silica gel flash chroma-
tography, by elution with petroleum ether/ethyl acetate (from 9:1
to 8:2) to afford the target product (68%) as a colorless oil. Analyti-
cal data for this compound matched the data already published.[15]
Synthesis of 6-ethyl-3-(1H-tetrazol-5-yl)pyridin-2-amine (16): A
sealed tube was charged with 2-amino-3-cyano-6-ethylpyridine 1-
oxide (35; 1.00 equiv), NaN3 (3.00 equiv), NH4Cl (3.00 equiv), and
dry DMF (1 mLmmolÀ1). The reaction was stirred at 1308C for 3 h.
TLC after this time, by elution with CH2Cl2/MeOH (85:15), showed
complete consumption of the starting material. The reaction was
stopped. The reaction mixture was cooled at 08C, and water was
added. The mixture was extracted with chloroform, and the organ-
ic phase was washed with brine, dried over Na2SO4, filtered, and
concentrated under reduced pressure. The crude residue was puri-
fied by silica gel flash chromatography, by elution with chloroform/
methanol/formic acid (8.7:1:0.3), to afford the target product (68%)
as a white solid.
Synthesis of 2-(benzylamino)-6-ethylnicotinonitrile (30): A 10 mL
microwave tube was charged with 2-chloro-6-ethylnicotinonitrile
(29, 1.00 equiv) and benzylamine (2.00 equiv). The tube was placed
in a microwave and irradiated at 2008C for 15 min (maximum
power input, 150 W; maximum pressure, 160 PSI; power max, ON;
stirring, ON). TLC after this time, by elution with petroleum ether/
ethyl acetate (9:1), showed almost complete consumption of the
starting material. The reaction was stopped. The reaction mixture
was cooled to 08C, then 1n HCl aqueous solution was carefully
added. The mixture was extracted with chloroform. The organic
phase was washed with brine, dried over anhydrous Na2SO4, fil-
tered, and concentrated under reduced pressure. The crude resi-
due was purified by silica gel flash chromatography, by elution
with petroleum ether/ethyl acetate (from 98:2 to 9:1), to afford the
target product (86%) as a colorless oil.
Synthesis of ethyl 2-(diethoxyphosphoryl)butanoate (43): Ethyl
2-(diethoxyphosphoryl)acetate (42; 1.00 equiv) was added drop-
wise to a cooled suspension of NaH (1.1 equiv) in dry DME
(2 mLmmolÀ1). After the reaction mixture had stirred at 258C for
2 h, ethyl iodide (1.1 equiv) was added, and the mixture was
heated at 608C for 2 h. The reaction mixture was poured into ice
water and extracted with ethyl acetate, and the combined organic
layers were washed with H2O and brine, dried over Na2SO4, and
evaporated under reduced pressure. The crude material was puri-
fied through flash chromatography, by elution with dichlorome-
thane/diethyl ether (95:5), to give the desired compound as a color-
less oil (65%). Analytical data for this compound matched the data
already published.[14]
Synthesis of (1R,2S)-ethyl 1-ethyl-2-phenylcyclopropanecarboxy-
late (44a): n-Butyllithium (2.5m in hexane, 2 equiv) was added
dropwise over 5 min to a solution of compound 38 (2 equiv) in dry
DME (5 mLmmolÀ1) at 258C. After the reaction mixture had stirred
Synthesis of 2-amino-6-ethylnicotinamide (31): H2SO4 (95%) was
carefully added to compound 30, and the dark mixture was stirred
at room temperature for 18 h. TLC after this time, by elution with
&
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
at the same temperature for 30 min, (S)-(À)-styrene oxide (1 equiv)
was added portionwise, and the mixture was heated at 908C for
18 h. After the mixture had cooled, a saturated aqueous solution
of NH4Cl was added, and the organic layers were extracted with
ethyl acetate, washed with brine, dried over anhydrous Na2SO4,
and evaporated under reduced pressure. The oil obtained was pu-
rified through flash column chromatography, by elution with petro-
leum ether/ethyl acetate (95:5), to give the desired product as
a yellowish oil (85%). Analytical data for this compound matched
the data already published.[14]
with dichloromethane/methanol (from 99:1 to 95:5) to afford the
two diastereomers (+)-46 and (À)-46 as white solids in 43 and
31% yields, respectively.
General procedure for the hydrolysis of N-(2-hydroxy-1-phenyl-
ethyl)-2-phenylcyclopropanecarboxamides: A solution of sub-
strate in 3n H2SO4 and dioxane (1:1; 1 mLmmolÀ1) was stirred at
1008C for 18 h. TLC after this time, by elution with dichlorome-
thane/methanol (9:1), showed almost complete consumption of
the starting material. The reaction was stopped. Water was careful-
ly added to the reaction mixture, which was then extracted with
chloroform. The organic phase was washed with brine, dried over
anhydrous Na2SO4, filtered, and concentrated under reduced pres-
sure. The crude residue was purified by silica gel flash chromatog-
raphy, by elution with dichloromethane/methanol (from 99:1 to
95:5), to afford the target compounds as white solids (98%). Com-
pounds 24 and 25 were obtained by this method.
Synthesis of (1S,2R)-ethyl 1-ethyl-2-phenylcyclopropanecarboxy-
late (44b): n-Butyllithium (2.5m in hexane, 2 equiv) was added
dropwise over 5 min to a solution of compound 38 (2 equiv) in dry
DME (5 mLmmolÀ1) at 258C. After the reaction mixture had stirred
at the same temperature for 30 min, (R)-(+)-styrene oxide (1 equiv)
was added portionwise, and the mixture was heated at 908C for
18 h. After the mixture had cooled, a saturated aqueous solution
of NH4Cl was added, and the organic layers were extracted with
ethyl acetate, washed with brine, dried over anhydrous Na2SO4,
and evaporated under reduced pressure. The oil obtained was pu-
rified through flash column chromatography, by elution with petro-
leum ether/ethyl acetate (95:5), to give the desired product as
a yellowish oil (85%). Analytical data for this compound matched
the data already published.[14]
2-(Benzylamino)-6-ethylnicotinonitrile (30): 1H NMR (300 MHz,
CDCl3): d=7.58 (d, J=7.8 Hz, 1H), 7.41–7.30 (m, 5H), 6.51 (d, J=
7.8 Hz, 1H), 5.45 (brs, 1H), 4.75 (d, J=5.6 Hz, 2H), 2.76–2.68 (q,
2H), 1.30–1.25 ppm (t, 3H); HPLC–MS: m/z found [M+H]+: 238.30.
1
2-Amino-6-ethylnicotinamide (31): H NMR (300 MHz, DMSO): d=
7.86 (d, J=7.7 Hz, 1H), 7.80 (brs, 2H), 7.16 (brs, 2H), 6.43 (d, J=
7.7 Hz, 1H), 2.55–2.50 (q, 2H), 1.19–1.14 ppm (t, 3H); HPLC–MS: m/
z found [M+H]+: 166.19.
Synthesis of (1R,2S)-1-ethyl-2-phenylcyclopropanecarboxylic
acid (14): Compound 44 (1 equiv) and LiOH·H2O (4 equiv) were dis-
solved in a solution of THF/MeOH/H2O (3:1:1; 1 mLmmolÀ1) and
heated under stirring in a microwave oven at 1008C for 7 min. The
reaction mixture was then evaporated in vacuo, and the crude resi-
due was taken up with H2O, acidified with HCl 1n, and extracted
with ethyl acetate. The organic layer was, in turn, washed with
brine and dried over anhydrous Na2SO4. After the evaporation of
the solvent, the crude material was purified through flash column
chromatography, by elution with dichloromethane/methanol
(95:5), to give the desired product as a white solid (85%). Analyti-
cal data for this compound matched the data already published.[14]
1
2-Amino-6-ethylnicotinic acid (32): H NMR (300 MHz, DMSO): d=
13.48 (brs, 1H), 7.94 (d, J=7.7 Hz, 1H), 7.18 (brs, 2H), 6.46 (d, J=
7.7 Hz, 1H), 2.57–2.50 (q, 2H), 1.19–1.14 ppm (t, 3H); HPLC–MS: m/
z found [M+H]+: 167.18.
1
Methyl 2-amino-6-ethylnicotinate (33): H NMR (300 MHz, DMSO):
d=7.96 (d, J=8.0 Hz, 1H), 7.11 (brs, 2H), 6.51 (d, J=8.0 Hz, 2H),
3.79 (s, 3H), 2.61–2.54 (q, 2H), 1.20–1.15 ppm (t, 3H); HPLC–MS: m/
z found [M+H]+: 181.20.
1
Methyl 2-amino-6-ethylnicotinate 1-oxide (18): H NMR (300 MHz,
DMSO): d=7.73 (brs, 2H), 7.68 (d, J=8.4 Hz, 1H), 6.71 (d, J=
8.4 Hz, 1H), 3.86 (s, 3H), 2.88–2.80 (q, 2H), 1.24–1.89 ppm (t, 3H);
13C NMR (75.5 MHz, DMSO): d=168.3, 154.7, 152.2, 128.2, 109.6,
106.7, 52.8, 24.7, 10.8 ppm; HPLC–MS: m/z found [M+H]+: 197.20.
Synthesis of (1S,2R)-1-ethyl-2-phenylcyclopropanecarboxylic
acid (26): Compound 44b (1 equiv) and LiOH·H2O (4 equiv) were
dissolved in a solution of THF/MeOH/H2O (3:1:1; 1 mLmmolÀ1) and
heated under stirring in a microwave oven at 1008C for 7 min. The
reaction mixture was then evaporated in vacuo, and the crude resi-
due was taken up with H2O, acidified with HCl 1n, and extracted
with ethyl acetate. The organic layer was, in turn, washed with
brine and dried over anhydrous Na2SO4. After the evaporation of
the solvent, the crude material was purified through flash column
chromatography, by elution with dichloromethane/methanol
(95:5), to give the desired product as a white solid (85%). Analyti-
cal data for this compound matched the data already published.[14]
2-Amino-6-ethyl-N-hydroxynicotinamide 1-oxide (17): 1H NMR
(300 MHz, DMSO): d=11.20 (brs, 1H), 9.34 (brs, 1H), 7.73 (brs, 2H),
7.62 (d, J=8.4 Hz, 1H), 6.47 (d, J=8.4 Hz, 1H), 2.79–2.73 (q, 2H),
1.21–1.17 ppm (t, 3H); 13C NMR (75.5 MHz, DMSO): d=161.2, 155.7,
152.0, 127.7, 110.0, 106.7, 24.7, 10.8 ppm; HPLC–MS: m/z found
[M+H]+: 198.19.
6-Ethyl-3-(1H-tetrazol-5-yl)pyridin-2-amine 1-oxide (16): 1H NMR
(300 MHz, DMSO): d=11.72 (brs, 1H), 7.88 (brs, 2H), 7.81 (d, J=
8.2 Hz, 1H), 6.87 (d, J=8.2 Hz, 1H), 2.91–2.85 (q, 2H), 1.27–
1.23 ppm (t, 3H); 13C NMR (75.5 MHz, DMSO): d=160.2, 157.3,
151.7, 127.0, 110.5, 106.7, 24.7, 10.8 ppm; HPLC–MS: m/z found
[M+H]+: 206.4.
Synthesis of N-(2-hydroxy-1-phenylethyl)-2-phenylcyclopropane-
carboxamide ((+/À)-46): (R)-(À)-2-Phenylglycinol (1.5 equiv), TBTU
(1 equiv), EDC·HCl (1.5 equiv), and triethylamine (1 equiv) were
added to a solution of compound 40 (1 equiv) in dichloromethane
(1 mLmmolÀ1). The reaction mixture was stirred for 1 h at 08C and
then for 4 h at room temperature. TLC after this time, by elution
with dichloromethane/methanol (9:1), showed almost complete
consumption of the starting material. The reaction was stopped.
An aqueous solution of NH4Cl was carefully added to the reaction
mixture, which was then extracted with chloroform. The organic
phase was washed with brine, dried over anhydrous Na2SO4, fil-
tered, and concentrated under reduced pressure. The crude resi-
due was purified by silica gel flash chromatography, by elution
1
2-Chloropyridine 1-oxide (15): H NMR (300 MHz, DMSO): d=8.44
(d, 1H), 7.81–7.72 (t, 1H), 7.57–7.53 (t, 1H), 7.40 ppm (d, 1H);
13C NMR (75.5 MHz, DMSO): d=125.3, 111.3, 105.8 ppm; HPLC–MS:
m/z found [M+H]+: 114.15.
1
Pyridin-2-amine 1-oxide (27): H NMR (300 MHz, DMSO): d=8.38
(d, 1H), 7.97 (brs, 1H), 7.67–7.62 (t, 1H), 7.51–7.48 (t, 1H),
7.35 ppm (d, 1H); 13C NMR (75.5 MHz, DMSO): d=124.3, 110.3,
104.2 ppm; HPLC–MS: m/z found [M+H]+: 95.11.
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
(1R,2R)-N-(2-Hydroxy-1-phenylethyl)-2-phenylcyclopropanecar-
boxamide ((+)-46): 1H NMR (300 MHz, DMSO): d=8.57 (d, 1H),
7.33–7.12 (m, 10H), 4.91–4.87 (m, 1H), 3.56–3.52 (t, 2H), 2.26–22.21
(m, 1H), 2.06–2.02 (m, 1H), 1.32–1.28 (m, 1H), 1.20–1.18 ppm (m,
1H); HPLC–MS: m/z found [M+H]+: 282.35.
tered on the catalytic Zn2+ metal atom, with a box dimension of
15ꢁ15ꢁ15 ꢂ, 3) ligand preparation by using LigPrep[40] in combi-
nation with Epik,[40] 4) each docking run was carried out with the
standard precision method, and the van der Waals scaling factor of
nonpolar atoms was set to 0.8, 5) final refinement of the obtained
docking pose by applying the extra-precision method, and 6) com-
putation of the re-docked ligand RMSD with respect to the crystal
structure. In the second step, the docking protocol identified in
the first step was challenged against a greater number of CA X-ray
structures (258hCA X-ray crystal structures). This final step allowed
the performance of the selected protocol to be robustly assessed
and allowed issues related to the X-ray crystal structures (such as
the presence of the co-crystallization factor in the CA active site)
to be identified. Details about the docking procedure, the per-
formance, and the ten CA X-ray crystal structures used are dis-
cussed in Supporting Information.
(1S,2S)-N-(2-Hydroxy-1-phenylethyl)-2-phenylcyclopropanecar-
boxamide ((À)-46): 1H NMR (300 MHz, DMSO): d=8.57 (d, 1H),
7.33–7.12 (m, 10H), 4.91–4.87 (m, 1H), 3.56–3.52 (t, 2H), 2.26–22.21
(m, 1H), 2.06–2.02 (m, 1H), 1.32–1.28 (m, 1H), 1.20–1.18 ppm (m,
1H); HPLC–MS: m/z found [M+H]+: 282.74.
(1R,2R)-2-Phenylcyclopropanecarboxylic acid ((+)-24): 1H NMR
(300 MHz, DMSO): d=11.06 (brs, 1H), 7.35–7.22 (m, 5H), 2.67–2.60
(m, 1H), 1.96–1.91 (m, 1H), 1.72–1.68 (m, 1H), 1.47–1.40 ppm (m,
1H); 13C NMR (100.6 MHz, CDCl3): d=129.3, 128.2, 126.9, 33.3, 31.2,
21.5 ppm; HPLC–MS: m/z found [MÀH]À: 161.19; [a]D = +47.62.
Docking: Once the validation step of the docking protocol had
been concluded, the identified method was applied to screen the
in-house library against the CA X-ray crystal structure dataset gen-
erated. In this case, the Protein Data Bank dataset was defined by
1) the X-ray crystal structures of the CA a family (for different or-
ganisms, such as human (I, II, III, IV, VII, XII, and XIII isoforms), bacte-
rial, protozoan, etc.), excluding those X-ray structures with a co-
crystallization factor inside the CA active site and 2) the X-ray crys-
tal structures of the CA b family (for different organisms), which
gave a total of 530 out of 753CA X-ray crystal structures. One pose
per ligand was retained and the Glide score was evaluated by
using the heat map representation.
In vitro assays
An SX.18MV-R Applied Photophysics (Oxford, UK) stopped-flow in-
strument was used to assay the catalysis/inhibition of various CA
isozymes as reported by Khalifah.[35] The CA-catalyzed CO2 hydra-
tion reaction was followed for a period of 5–10 s. Phenol red (at
a concentration of 0.2 mm) was used as the indicator, working at
the absorbance maximum of 557 nm, with 10 mm 2-[4-(2-hydrox-
yethyl)-1-piperazinyl]ethanesulfonic acid (pH 7.4) as buffer and
0.1m Na2SO4 or NaClO4 for constant maintenance of the ionic
strength (these anions are not inhibitory in the concentration
used). Saturated CO2 solutions in water at 258C were used as sub-
strates. Stock solutions of inhibitors were prepared at a concentra-
tion of 10 mm (in the assay buffer) and dilutions up to 0.1 nm were
done with the same assay buffer. Inhibitor and enzyme solutions
were pre-incubated together for 15 min at room temperature prior
to assay, to allow formation of the enzyme–inhibitor complex. The
inhibition constants were obtained by nonlinear least-squares
methods by using PRISM 3 and the Cheng–Prusoff equation, as re-
ported earlier,[36,37] and represent the mean from at least three dif-
ferent determinations. hCA II was purchased from Sigma–Aldrich
and used without further purification, whereas all the other CA iso-
forms were recombinant enzymes obtained in house as reported
earlier.[30,37,38] The concentrations of the enzymes in the assay sys-
tems were 9.2 nm for hCA I, 7.1 nm for hCA II, 13.6 nm for hCA III,
8.4 nm for SspCA, and 12.0 nm for PfCA.
Compounds selection for in vitro assays: Among the different com-
pounds screened during the docking studies, only those which
1) showed a glide score >7.5 kcalmolÀ1 (if more than one X-ray
crystal structure was available for the same enzyme, the average
and its standard deviation were considered) and 2) were physically
available in the lab were selected to be tested in vitro.
Keywords: carbonic anhydrase · computational chemistry ·
drug design · enzymes · inhibitors
[1] J. Langedijk, A. K. Mantel-Teeuwisse, D. S. Slijkerman, M.-H. D. B. Schut-
jens, Drug Discovery Today 2015, 20, 1027–1034.
[2] M. Mori, Y. Cau, G. Vignaroli, I. Laurenzana, A. Caivano, D. Vullo, C. T. Su-
puran, M. Botta, ACS Chem. Biol. 2015, 10, 1964–1969.
[3] M. Kaiser, L. Maes, L. P. Tadoori, T. Spangenberg, J.-R. Ioset, J. Biomol.
Screening 2015, 20, 634–645.
Computational methods
[4] R. Sawada, H. Iwata, S. Mizutani, Y. Yamanishi, J. Chem. Inf. Model. 2015,
55, 2717–2730.
[5] S. N. Deftereos, C. Andronis, E. J. Friedla, A. Persidis, A. Persidis, Wiley In-
terdiscip. Rev. Syst. Biol. Med. 2011, 3, 323–334.
Maximum common scaffold decomposition: The maximum common
scaffold decomposition was performed by using the Library MCS
tool of ChemAxon[39] and the default parameters.
[6] Z. Liu, H. Fang, K. Reagan, X. Xu, D. L. Mendrick, W. Slikker, W. Tong,
Drug Discovery Today 2013, 18, 110–115.
Identification and validation of the docking protocol: The best dock-
ing parameters and the X-ray structures to be used in the study
were identified by applying a two-step pipeline protocol. In the
first step, a re-docking procedure was performed on a small set
(ten) of CA X-ray crystal structures, by using Glide.[40] The default
docking parameters were modified until the identification of the
combination that turned out the best performance in terms of the
ligand re-docking root mean square deviation (RMSD). In this case,
the ten CA X-ray crystal structures were chosen so as to account
for the different nature of the CAIs crystallized so far. The final
docking procedure is defined by the following steps: 1) protein
preparation by using the metal-binding-site function and removal
of co-crystallization factors (that is, glycerol), 2) the grid was cen-
[7] J. W. Raymond, P. Willett, J. Comput.-Aided Mol. Des. 2002, 16, 521–533.
[8] T. Kawabata, J. Chem. Inf. Model. 2011, 51, 1775–1787.
[9] C. Abad-Zapatero, J. T. Metz, Drug Discovery Today 2005, 10, 464–469.
[10] L. Amori, S. Katkevica, A. Bruno, B. Campanini, P. Felici, A. Mozzarelli, G.
Costantino, MedChemComm 2012, 3, 1111.
[11] M. Pieroni, B. Wan, S. Cho, S. G. Franzblau, G. Costantino, Eur. J. Med.
Chem. 2014, 72, 26–34.
[12] P. Vincetti, F. Caporuscio, S. Kaptein, A. Gioiello, V. Mancino, Y. Suzuki, N.
Yamamoto, E. Crespan, A. Lossani, G. Maga, G. Rastelli, D. Castagnolo, J.
Neyts, P. Leyssen, G. Costantino, M. Radi, J. Med. Chem. 2015, 58, 4964–
4975.
[13] M. Pieroni, D. Machado, E. Azzali, S. Santos Costa, I. Couto, G. Costanti-
no, M. Viveiros, J. Med. Chem. 2015, 58, 5842–5853.
&
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
[14] M. Pieroni, G. Annunziato, C. Beato, R. Wouters, R. Benoni, B. Campanini,
T. A. Pertinhez, S. Bettati, A. Mozzarelli, G. Costantino, J. Med. Chem.
2016, 59, 2567–2578.
[15] G. P. Vallerini, L. Amori, C. Beato, M. Tararina, X.-D. Wang, R. Schwarcz, G.
Costantino, J. Med. Chem. 2013, 56, 9482–9495.
[16] C. T. Supuran, Nat. Rev. Drug Discovery 2008, 7, 168–181.
[17] C. T. Supuran, Front. Pharmacol. 2011, 2, 34.
[18] V. Alterio, A. Di Fiore, K. D’Ambrosio, C. T. Supuran, G. De Simone, Chem.
Rev. 2012, 112, 4421–4468.
[29] S. Del Prete, D. Vullo, G. M. Fisher, K. T. Andrews, S.-A. Poulsen, C. Capa-
sso, C. T. Supuran, Bioorg. Med. Chem. Lett. 2014, 24, 4389–4396.
[30] G. De Simone, A. Di Fiore, C. Capasso, C. T. Supuran, Bioorg. Med. Chem.
Lett. 2015, 25, 1385–1389.
[31] D. Chaiyaveij, A. S. Batsanov, M. A. Fox, T. B. Marder, A. Whiting, J. Org.
Chem. 2015, 80, 9518–9534.
[32] S. J. Cho, N. H. Jensen, T. Kurome, S. Kadari, M. L. Manzano, J. E. Malberg,
B. Caldarone, B. L. Roth, A. P. Kozikowski, J. Med. Chem. 2009, 52, 1885–
1902.
[19] C. Capasso, C. T. Supuran, Expert Opin. Ther. Pat. 2013, 23, 693–704.
[20] C. T. Supuran, J. Enzyme Inhib. Med. Chem. 2016, 31, 345–360.
[21] S. K. Nair, P. A. Ludwig, D. W. Christianson, J. Am. Chem. Soc. 1994, 116,
3659–3660.
[22] D. P. Martin, P. G. Blachly, J. A. McCammon, S. M. Cohen, J. Med. Chem.
2014, 57, 7126–7135.
[23] J. Schulze Wischeler, A. Innocenti, D. Vullo, A. Agrawal, S. M. Cohen, A.
Heine, C. T. Supuran, G. Klebe, ChemMedChem 2010, 5, 1609–1615.
[24] H. Vu, N. B. Pham, R. J. Quinn, J. Biomol. Screening 2008, 13, 265–275.
[25] Y. Lou, P. C. McDonald, A. Oloumi, S. Chia, C. Ostlund, A. Ahmadi, A.
Kyle, U. Auf dem Keller, S. Leung, D. Huntsman, B. Clarke, B. W. Suther-
land, D. Waterhouse, M. Bally, C. Roskelley, C. M. Overall, A. Minchinton,
F. Pacchiano, F. Carta, A. Scozzafava, N. Touisni, J.-Y. Winum, C. T. Supur-
an, S. Dedhar, Cancer Res. 2011, 71, 3364–3376.
[33] T. T. Ashburn, K. B. Thor, Nat. Rev. Drug Discovery 2004, 3, 673–683.
[34] T. I. Oprea, J. E. Bauman, C. G. Bologa, T. Buranda, A. Chigaev, B. S. Ed-
wards, J. W. Jarvik, H. D. Gresham, M. K. Haynes, B. Hjelle, R. Hromas, L.
Hudson, D. A. Mackenzie, C. Y. Muller, J. C. Reed, P. C. Simons, Y. Smag-
ley, J. Strouse, Z. Surviladze, T. Thompson, O. Ursu, A. Waller, A. Wan-
dinger-Ness, S. S. Winter, Y. Wu, S. M. Young, R. S. Larson, C. Willman,
L. A. Sklar, Drug Discov. Today Ther. Strateg. 2011, 8, 61–69.
[35] R. G. Khalifah, J. Biol. Chem. 1971, 246, 2561–2573.
[36] A. Maresca, D. Vullo, A. Scozzafava, G. Manole, C. T. Supuran, J. Enzyme
Inhib. Med. Chem. 2013, 28, 392–396.
[37] A. Maresca, A. Scozzafava, D. Vullo, C. T. Supuran, J. Enzyme Inhib. Med.
Chem. 2013, 28, 384–387.
[38] D. Vullo, V. D. Luca, A. Scozzafava, V. Carginale, M. Rossi, C. T. Supuran,
C. Capasso, Bioorg. Med. Chem. 2013, 21, 1534–1538.
[26] F. Carta, C. Temperini, A. Innocenti, A. Scozzafava, K. Kaila, C. T. Supuran,
J. Med. Chem. 2010, 53, 5511–5522.
[39] https://www.chemaxon.com.
[40] http://www.schrodinger.com.
[27] J. Krungkrai, S. R. Krungkrai, C. T. Supuran, Bioorg. Med. Chem. Lett.
2008, 18, 5466–5471.
Received: March 31, 2016
Revised: May 16, 2016
[28] J. Krungkrai, A. Scozzafava, S. Reungprapavut, S. R. Krungkrai, R. Ratta-
najak, S. Kamchonwongpaisan, C. T. Supuran, Bioorg. Med. Chem. 2005,
13, 483–489.
Published online on && &&, 0000
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
11
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

FULL PAPERS
G. Annunziato, A. Angeli, F. D’Alba,
New uses for old chemicals: The appli-
cation of a rational target-focused drug-
repurposing approach led to the iden-
tification of new carbonic anhydrase
(CA) inhibitors that are able to selective-
ly inhibit microbial over human en-
zymes. Compound 18 is a promising hit
because of an interesting CA selectivity
profile and a lack of activity against the
target for which it was originally de-
signed. Compound 18 is a good starting
point for a new class of CA inhibitors.
A. Bruno,* M. Pieroni, D. Vullo, V. De Luca,
C. Capasso, C. T. Supuran,* G. Costantino
&& – &&
Discovery of New Potential Anti-
Infective Compounds Based on
Carbonic Anhydrase Inhibitors by
Rational Target-Focused Repurposing
Approaches
&
ChemMedChem 2016, 11, 1 – 12
www.chemmedchem.org
12
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!