Antifungal agents. 11. N-substituted derivatives of 1-[(aryl)(4-aryl-1H- pyrrol-3-yl)methyl]-1H-imidazole: Synthesis, anti-Candida activity, and QSAR studies

Article abstract of DOI:10.1021/jm048997u

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazoles were recently reported by our group as potent anti-Candida agents belonging to the antifungal azole class. In the present paper the synthesis, anti-Candida activities, and QSAR studies on a novel serie

Full text of DOI:10.1021/jm048997u

5140
J. Med. Chem. 2005, 48, 5140-5153
Antifungal Agents. 11. N-Substituted Derivatives of
1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole: Synthesis, Anti-Candida
Activity, and QSAR Studies
Roberto Di Santo,*,†,‡ Andrea Tafi,‡,§ Roberta Costi,† Maurizo Botta,*,§ Marino Artico,† Federico Corelli,§
Michela Forte,† Fabiana Caporuscio,| Letizia Angiolella, and Anna Teresa Palamara
“Istituto Pasteur-Fondazione Cenci Bolognetti”, Dipartimento di Studi Farmaceutici, Universita` di Roma “La Sapienza”,
P.le Aldo Moro 5, I-00185 Roma, Italy; Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena, Via Aldo Moro,
S. Miniato, I-53100 Siena, Italy; and Dipartimento di Studi di Chimica e Tecnologia delle Sostanze Biologicamente Attive,
and Istituto di Microbiologia, Universita` di Roma “La Sapienza”, P.le Aldo Moro 5, I-00185 Roma, Italy
Received December 10, 2004
1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazoles were recently reported by our group as
potent anti-Candida agents belonging to the antifungal azole class. In the present paper the
synthesis, anti-Candida activities, and QSAR studies on a novel series of N-substituted 1-[(aryl)-
(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole derivatives are reported. The newly synthesized
azoles were tested against 12 strains of Candida albicans together with bifonazole, miconazole,
itraconazole, fluconazole, and compounds 1a, 1b, 3a, 3b, and 3c used as reference drugs. In
general, tested derivatives showed good antifungal activities, and the most potent compound
was 1d (MIC90 ) 0.032 µg/mL), which was from 4- to 250-fold more potent than reference drugs.
Catalyst software was applied to develop a quantitative pharmacophore model to be used for
the rational design of new antifungal azoles. Some key interactions, as well as excluded volumes,
further to the coordination bond of azole antifungals with the demethylase enzyme, are
highlighted.
Introduction
accumulation of 14R-methylated sterols alter membrane
fluidity, with concomitant reduction in activity of mem-
brane-associated enzymes and increased permeability.
The net effect is to inhibit fungal growth and replica-
tion.9
During the past 2 decades life-threatening fungal
infection frequency and type have been increasing in
immunocompromised patients, such as people affected
with AIDS (over 90%), bone marrow and organ trans-
plant patients, and cancer patients.1,2 Candida albicans
(CA) has been identified as the major opportunistic
pathogen in the etiology of fungal infections; however,
the frequency of other Candida species is increasing.3
The current standard of therapies are the fungicidal (but
toxic) polyene, amphotericin B, and the safe (but fungi-
static) azoles. In particular, the latter drugs are impor-
tant antifungal agents widely used for AIDS-related
mycotic pathologies.4
Azole antifungal agents prevent the synthesis of
ergosterol, a major component of fungal membranes, by
inhibiting the cytochrome P-450-dependent enzyme 14R-
lanosterol demethylase (also referred as CYP51).5,6 This
enzyme contains an iron protoporphyrin unit located in
its active site, which catalyzes the oxidative removal of
the 14R-methyl group of lanosterol, by typical monooxy-
genase activity.7 Azole antifungal agents bind to the iron
of the porphyrin8 and cause the blockade of the fungal
ergosterol biosynthesis pathway, by preventing the
access of the natural substrate lanosterol to the active
site of the enzyme. The depletion of ergosterol and
In the last 3 decades a number of antifungal azoles
were discovered and introduced in clinical practice.10-16
Despite the important advances in this field, there is a
continuing increase in the incidence of fungal infections,
together with a gradual rise in azole resistance.17 This
scenario highlights the urgent need for new and effective
antifungal agents. While the crystal structure of Myco-
bacterium tuberculosis CYP51 cytochrome (MT-CYP51)
is known,18 unfortunately, no three-dimensional struc-
tural information on Candida albicans (CA-CYP51) or
other fungal CYP51 enzymes is available. Due to this
fact, homology modeling19 and pharmacophore model-
ing20 techniques are currently used to attempt the
rational design of new antifungal leads.
Being engaged in the last 10 years in searches on new
antimycotic agents,21-30 we reported the design and
synthesis of 1-[(aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-
imidazoles,31 which were found to be potent anti-
Candida agents either in vitro or in vivo assays.32 Our
SAR studies on that class of compounds, with different
substituents either on the arylpyrrole moiety or on the
1-benzylimidazole portion, led us to obtain an interest-
ing group of azole antimicrobials. Pyrrole derivatives
1a-7a, 1b, 3b, and 3c were in fact synthesized (Chart
1) that showed anti-Candida activities comparable or
superior to that of miconazole, bifonazole, and keto-
conazole, used as reference drugs.33-35 It is worthy of
note that 1b, 3b, and 3c, the sole N-substituted syn-
* To whom correspondence should be addressed. R.D.S.: phone and
fax, +39-6-49913150; e-mail: roberto.disanto@uniroma1.it. M.B.: phone,
+39-577-234306; fax, +39-577-234333; e-mail, botta@unisi.it.
† Dipartimento di Studi Farmaceutici, Universita` di Roma “La
Sapienza”.
‡ These authors contributed equally to this work.
§ Dipartimento Farmaco Chimico Tecnologico, Universita` di Siena.
| Dipartimento SCTSBA, Universita` di Roma.
Istituto di Microbiologia, Universita` di Roma.
10.1021/jm048997u CCC: $30.25 © 2005 American Chemical Society
Published on Web 07/07/2005

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5141
Chart 1. Anti-Candida Agents Considered as Lead Compounds 1a-7a, 1b, and 3b,c and Newly Synthesized
Derivatives 1c-j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b, and 8a-c
thesized derivatives, were more active than the parent
compounds 1a and 3a, respectively.35 Moreover, even
if we were already involved in computational studies
aimed at enlightening the molecular bases of the anti-
Candida activity of azole compounds,34,35 at that time
no systematic investigations were performed on the
biological activities of the nitrogen-substituted deriva-
tives.
Chemistry
Phenylpyrrolylmethanones 13a-g31,32 (Table 2) were
used as starting materials to obtain the new imidazole
derivatives 1c-e,g,h,j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, and
7b (Scheme 1). Alkylation of pyrroles 13a-g with the
appropriate alkyl halide in alkaline medium (K2CO3)
gave N-alkylpyrrolylmethanones 14a-c,e,h-t. This
procedure was not successful for dimethylallyl deriva-
tive 14f, which was achieved in high yields by reaction
of dimethylallyl bromide with 13a, in the presence of
NaH as a catalyst. Arylation of 13a to afford [4-(4-chloro-
phenyl)-1-phenyl-1H-pyrrol-3-yl]phenylmethanone 14g
was performed in Suzuki reaction conditions using
phenylboronic acid, Cu(II) acetate, and pyridine. The
best yields were obtained in the presence of N-meth-
ylpyrrolidone (NMP) by microwave-assisted heating (60
W, 120 °C; 3 × 50 s). Reduction of ketones 14a-c,e-t
with LiAlH4 gave methanols 15a-c,e-t, which were
then treated with 1,1′-carbonyldiimidazole (CDI) to
afford the required imidazoles 1c-e,g,h,j, 2b,c, 3d-f,
4b, 5b-e, 6b,c, and 7b (Scheme 1).
A similar synthetic pathway gave derivatives 8a-c
(Scheme 2). The key intermediate (2,4-dichlorophenyl)-
[4-(naphthalen-1-yl)-1H-pyrrol-3-yl]methanone 13h was
obtained by condensation the naphthalene-1-carboxal-
dehyde with 2′,4′-dichloroacetophenone in aqueous so-
dium hydroxide to afford propenone 16, followed by
TosMIC cycloaddition in the presence of sodium hydride.
The synthesis of vinyl derivative 1f is depicted in
Scheme 3. Interestingly, the reaction of 13a with 1,2-
dichloroethane under phase transfer catalysis conditions
(tetrabutylammonium hydrogen sulfate (Bu4NHSO4),
aqueous sodium hydroxide and dichloromethane) gave
the expected methanone 14d, with a concomitant for-
mation of dimer methanone 17. Reduction of 14d
In this paper, we report the synthesis and anti-
Candida activity of a series of 1-[(aryl)(4-aryl-1H-pyrrol-
3-yl)methyl]-1H-imidazoles (1c-j, 2b,c, 3d-f, 4b, 5b-
e, 6b,c, 7b, and 8a-c) bearing different substituents
on the nitrogen of pyrrole ring (Chart 1, Table 1). In
particular, small alkyl (methyl, ethyl, n-propyl, c-
propylmethyl, dimethoxyethyl), alkenyl (vinyl, allyl,
dimethylallyl, methyl 3-acryloylate) or phenyl groups
were chosen. Introduction of bulky substituents was not
considered on the basis of our previous experience with
the N-2,4-dichlorobenzyl derivative of 1a and congeners
that were found to be inactive.31 Although the newly
synthesized derivatives were characterized by a stereo-
genic center, separation of the enantiomers was not
performed. In fact, we recently demonstrated that levo
and dextro forms of congeners of this series, which were
separated by HPLC on polysaccharide-based chiral
stationary phases, showed the same antifungal activities
in in vitro tests.25,27 In addition, taking into account that
the new derivatives showed antifungal activities higher
than any other compound previously synthesized by us,
we applied the program Catalyst to develop a new
quantitative pharmacophore model, endowed with pre-
dictive power to be used for the rational design of new
antifungal azoles.

5142 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Di Santo et al.
Table 1. Chemical, Physical, and Analytical Data of Derivatives 1c-j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b, 8a-c
mp
recryst
solva
reaction chromat
cmpd
R1
R2 R3
R4
R5
H
X
(°C)
yield (%) time (h) systemb
analyses
1c
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
Cl
H
-
-
-
H
H
H
H
H
H
H
H
Cl
Cl
H
H
H
Cl
Cl
Cl
Cl
Cl
H
H
H
-
H
C2H5
C3H7
oil
oil
oil
oil
oil
-
-
-
-
-
a
73
76
79
70
69
61
37
79
60
59
64
62
99
28
40
63
77
80
69
71
36
80
58
63
15
A
B
B
B
A
B
A
B
A
A
B
A
-
A
A
A
B
B
A
A
A
A
A
A
C,H,N,Cl
(C22H20ClN3)
C,H,N,Cl
(C23H22ClN3)
C,H,N,Cl
(C24H22ClN3)
C,H,N,Cl
(C22H18ClN3)
C,H,N,Cl
(C23H20ClN3)
C,H,N,Cl
(C25H24ClN3)
C,H,N,Cl
(C24H20ClN3O2)
C,H,N,Cl
(C26H20ClN3)
C,H,N,Cl
(C21H17Cl2N3)
C,H,N,Cl
(C22H19Cl2N3)
C,H,N,Cl
(C23H20Cl3N3)
C,H,N,Cl
(C23H18Cl3N3)
C,H,N,Cl
(C24H22Cl3N3O2)
C,H,N,Cl
(C21H16Cl3N3)
C,H,N,Cl
(C21H15Cl4N3)
C,H,N,Cl
(C22H17Cl4N3)
C,H,N,Cl
(C23H19Cl4N3)
C,H,N,Cl
(C23H17Cl4N3)
C,H,N,Cl
(C22H19Cl2N3)
C,H,N,Cl
(C23H21Cl2N3)
C,H,N,Cl
(C25H20Cl2N4)
C,H,N,Cl
(C24H17Cl2N3)
C,H,N,Cl
1d
1e
1f
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
2
CH2-c-C3H5
CHdCH2
15
1
1g
1h
1i
CH2CHdCH2
1
CH2CHdC(CH3)2 48-50
CHdCHCOOCH3 oil
48
6
-
-
a
1j
Ph
oil
2
2b
2c
3d
3e
3f
CH3
C2H5
148-150
149-151
123-125
oil
15
15
15
1
a
Cl Cl C3H7
b
Cl Cl CH2CHdCH2
Cl Cl CH2CH(OCH3)2
Cl Cl CH3
-
b
136-139
106-108
oil
1
4b
5b
5c
5d
5e
6b
6c
7b
8a
8b
8c
a
72
15
15
15
15
Cl
Cl
Cl
Cl
Cl
Cl
Cl Cl CH3
-
-
-
-
-
b
Cl Cl C2H5
oil
Cl Cl C3H7
oil
Cl Cl CH2CHdCH2
oil
CH3
CH3
H
H
CH3
oil
2.5
C2H5
105-106
oil
72
15
1
1-pyrrolyl
Cl Cl CH3
-
-
a
-
-
-
-
-
-
-
-
-
H
oil
-
CH3
C2H5
182-185
oil
1
(C25H19Cl2N3)
C,H,N,Cl
(C26H21Cl2N3)
-
-
72
a Recrystallization solvents: (a) benzene/cyclohexane, (b) cyclohexane. b Chromatographic system: (A) aluminum oxide/ethyl acetate,
(B) aluminum oxide/chloroform.
furnished the corresponding alcohol 15d, which was
condensed with CDI to afford imidazole derivative 1f.
At last, derivative 1i was obtained by direct reaction
of 1a with methyl propiolate in the presence of tetrabu-
tylammonium fluoride (Bu4NF) (Scheme 4).
the newly synthesized imidazole derivatives showed
good antifungal activities when compared with reference
drugs (MIC range from 0.032 to 64 µg/mL). Compounds
1d-i, 2b,c, 3e,f, 4b, 5c-e, 6b,c, 7b, and 8a were from
2 to 250 times more potent than bifonazole (MIC90 ) 8
µg/mL), with 1d (MIC90 ) 0.032 µg/mL), the most active
among test derivatives, being 4-, 15-, 30- and 250-fold
more potent than itraconazole, fluconazole, miconazole,
and bifonazole, respectively.
As a rule, the monochloro derivatives 1b-j were more
potent than N-unsubstituted compound 1a. The best
results were obtained with compounds 1b, 1d, and 1e
bearing methyl, propyl, or cyclopropylmethyl groups
(MIC90 range 0.032-0.125 µg/mL). Alkenyl derivatives
1f-h were just a little less potent than the above alkyl
counterparts, with MIC90 values ranging from 0.250 to
0.5 µg/mL. Introduction of a carboxymethyl function on
the vinyl group of 1f led to acrylate 1i, which was 16-
fold less potent than the parent compound. A similar
Results and Discussion
Microbiology. In Vitro Anti-Candida Activities.
The in vitro antifungal activities of newly synthesized
azole derivatives 1c-j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b,
and 8a-c against 12 strains of Candida albicans are
reported in Table 3 together with those of bifonazole,
miconazole, itraconazole, fluconazole, and compounds
1a, 1b, 3a, 3b, and 3c used as reference drugs.36 Five
out of 12 Candida albicans strains used in this study
were resistant to fluconazole. With few exceptions,
MIC50 and MIC values paralleled the related MIC90
data, which were chosen for a preliminary comparison
between the new azoles and reference drugs. In general,

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5143
Table 2. Chemical and Physical Data of Derivatives 13a-h, 14a-v, 15a-w, 16, and 17a
mp
(°C)
recryst
solvb
yield
(%)
reaction
time (h)
chromat
systemc
compd
R1
R2
R3
R4
R5
X
13a
13b
13c
13d
13e
13f
13g
13h
14a
14b
14c
14d
14e
14f
14 g
14h
14i
14j
14k
14l
14m
14n
14o
14p
14q
14r
14s
14t
14u
14v
15a
15b
15c
15d
15e
15f
Cl
Cl
Cl
H
Cl
Cl
H
H
H
H
H
Cl
H
-
H
H
H
H
H
-
-
H
H
H
H
H
H
H
H
H
H
Cl
Cl
H
-
-
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
Cl
Cl
H
-
-
-
-
-
H
Cl
H
Cl
Cl
H
H
-
H
H
H
Cl
Cl
Cl
H
Cl
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
a
-
-
-
-
-
-
-
-
H
-
Cl
Cl
Cl
CH3
Cl
-
-
-
-
-
-
-
-
-
-
-
-
-
-
A
B
B
A
A
A
A
-
A
B
A
A
-
-
A
B
A
-
A
-
-
A
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
A
-
-
-
1-pyrrolyl
-
-
-
-
-
210-212
82
0.5
60
7
Cl
H
H
H
H
H
-
H
H
H
H
H
H
-
C2H5
C3H7
oil
oil
-
-
b
37
Cl
H
78
Cl
H
CH2-c-C3H5
103-104
49
6
Cl
H
CHdCH2
oil
-
b
77
4.5
25
1
Cl
H
CH2CHdCH2
93-94
32
-
-
-
-
84-85
b
91
-
-
-
-
155-156
c
19
13
2
38
1
20
24
15
4
26
2.5
4
2
80
3
1
2
1.5
1
5.5
3.5
0.5
0.5
0.5
1
0.25
1
2
Cl
Cl
Cl
H
H
H
Cl
Cl
Cl
Cl
Cl
H
H
H
-
H
H
H
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
H
Cl
-
CH3
112-115
b
100
27
Cl
H
C2H5
129-130
b
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
CH3
CH3
Cl
-
C3H7
100-102
b
66
84
100
91
100
48
57
67
Cl
CH2CHdCH2
CH2CH(OCH3)2
CH3
oil
-
b
Cl
81-82
H
oil
oil
-
-
b
b
d
-
-
a
c
b
-
-
-
-
-
b
-
-
-
-
-
-
b
Cl
CH3
Cl
C2H5
123-125
Cl
C3H7
80-82
Cl
CH2CHdCH2
CH3
81-82
Cl
oil
oil
100
56
Cl
C2H5
1-pyrrolyl
CH3
190-193
93
-
-
CH3
131-133
93
-
-
-
C2H5
122-123
67
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
H
H
H
H
H
H
H
Cl
Cl
H
H
H
Cl
Cl
Cl
Cl
Cl
H
H
H
-
H
H
H
H
H
H
H
H
H
H
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
H
H
Cl
-
C2H5
-
75
H
C3H7
-
100
74
H
CH2-c-C3H5
CHdCH2
CH2CHdCH2
CH2CHdC(CH3)2
Ph
-
H
-
100
100
92
H
-
H
-
15g
15h
15i
15j
15k
15l
H
-
100
100
100
100
100
98
H
CH3
-
H
C2H5
-
Cl
Cl
Cl
Cl
Cl
Cl
Cl
Cl
CH3
CH3
Cl
-
C3H7
-
CH2CHdCH2
CH2CH(OCH3)2
CH3
-
-
1.5
2
15m
15n
15o
15p
15q
15r
15s
15t
15u
15v
15w
16
-
98
Cl
Cl
Cl
Cl
Cl
Cl
1-pyrrolyl
-
CH3
-
-
-
-
-
b
100
100
100
100
96
2
C2H5
-
2
C3H7
-
3
CH2CHdCH2
CH3
-
3
-
1
C2H5
-
e
86
96
100
98
100
76
6
1
CH3
-
-
-
-
-
f
2
H
-
15
1
-
-
-
-
CH3
-
-
-
-
-
C2H5
-
1.5
0.5
4.5
-
-
-
-
-
-
114-116
17
-
-
-
-
-
-
a Data of known derivatives 13a-g are reported in ref 32. b Recrystallization solvents: (a) DMF/H2O, (b) cyclohexane, (c) benzene/
cyclohexane, (d) n-hexane, (e) benzene, (f) ethanol. c Chromatographic system: (A) aluminum oxide/chloroform, (B) silica gel/chloroform.
decrease of activity was noted when a phenyl moiety
replaced the alkenyl group.
At last, it is worthy of note that derivatives 1a-c,i,j,
2b,c, 3a,b,d-f, 4b, 5b,c, 6b,c, 7b, and 8a-c showed a
percentage resistant strains (MIC > 64 µg/ mL) com-
parable to that of reference drugs, while compounds
1d-h and 5e (R% ) 0) were found to be potent
inhibitors of the fungal growth even on fungi strains
resistant to the reference drugs.
Molecular Modeling Studies. For reasons of con-
venience, the azoles used in the QSAR studies [1c-j,
2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b, 8a-c (newly synthe-
sized) and 9a-k, 10a, 11a-k, 12a-l, (previously syn-
thesized)] are classified in Chart 2 as belonging to five
different structural classes, named A, B, C, D, and E.
Generally, the introduction of one, two, or three
chlorine atoms on compounds 1b-d or 1g led to alkyl
derivatives 2b,c, 3d-f, 4b, and 5b-e, endowed with
lower potency (MIC90 > 1 µg/mL). Similarly, introduc-
tion of a methyl or a pyrrolyl groups on the phenyl rings
or replacement of the phenyl with a naphthalene ring
led to derivatives 6b,c, 7b, and 8a-c which showed
MIC90 > 2 µg/mL. As an assumption, this drop in the
activities could be ascribed to the highly hydrophobic
groups introduced in the molecules, leading to an altered
hydrophobic/hydrophilic balance.

5144 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Di Santo et al.
Scheme 1a
a Reagents: (a) alkyl iodide or bromide, K2CO3, DMF; (b) 1-bromo-3-methyl-2-butene, NaH, THF; (c) PhB(OH)2, Cu(OAC)2, pyridine,
NMP, microwave 60 W, 120 °C, 3 × 50 s; (d) LiAlH4, THF; (e) 1,1′-carbonyldiimidazole, MeCN.
Scheme 2a
a Reagents: (a) 2,4-dichloroacetophenone, NaOH, EtOH; (b) toluene-4-sulfonylmethyl isocyanide (TosMIC), NaH, DMSO, Et2O; (c) LiAlH4,
THF; (d) alkyl iodide, K2CO3, DMF; (e) 1,1′-carbonyldiimidazole, MeCN.
We have been involved since 1996 in computational
studies on antifungal azoles,34,35 and our efforts have
produced remarkable results, whose soundness has been
confirmed by other authors.37 Despite these attain-
ments, however, our work has not been able to fully
satisfy our expectations for a long time. The expected

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5145
Scheme 3a
Table 3. In Vitro Antifungal Activities of Imidazole
Derivatives 1a-j, 2b,c, 3a-f, 4b, 5b-e, 6b,c, 7b, and 8a-c
against Candida albicans
MIC ( SD
(µg/mL)
MIC50
MIC90
range
(µg/mL)
compd
%Ra
(µg/mL) (µg/mL)
1a
1b
1c
1d
1e
1f
33.33 2.5 ( 1.45
41.6 0.097 ( 0.031
33.33 5.59 ( 10.97
0.125
4
0.065-4
0.062
0.125 0.062-0.125
1
8
0.25-32
0
0
0
0
0
0.096 ( 0.188
0.111 ( 0.184
0.125 ( 0.114
0.082 ( 0.150
0.194 ( 0.196
0.016
0.032 0.016-0.5
0.125 0.016-0.5
0.25 0.016-0.5
0.25 0.016-0.5
0.016
0.032
1 g
0.016
1h
0.065
0.5
4
16
0.062-0.5
0.5-4
2-16
0.25-1
0.25-4
0.25-0.5
0.125-2
0.5-2
1i
16.6 1.62 ( 1.212
50 7 ( 6.985
41.6 0.46 ( 0.267
41.66 1.46 ( 1.326
33.33 4.2 ( 0.168
2
4
0.5
1
0.25
1
2
1
1j
2b
1
2
0.5
2
2
2c
3a
3b
50
25
66
1.31 ( 1.25
1.5 ( 0.61
3c
3d
8.56 ( 15.629
32
0.25-2
0.125-2
2-4
3e
33.33 0.73 ( 0.599
0.5
2
2
4
2
3f
50
36
16
2.66 ( 1.095
1.71 ( 0.626
14.6 ( 12.997
4b
2
0.5-2
5b
8
32
2
2-32
a Reagents: (a) 1,2-dichloroethane, Bu4NHSO4, NaOH, CH2Cl2;
(b) LiAlH4, THF; (c) 1,1′-carbonyldiimidazole, MeCN.
5c
41.66 1.21 ( 0.566
1
0.5-2
5d
100
0
>64
0.65 ( 0.709
>64
>64
1
>64
5e
0.25
0.5
2
4
0.5
4
0.062-2
0.25-2
0.125-2
1-4
Scheme 4a
6b
33.33 0.87 ( 0.731
33.33 2.18 ( 2.419
33.33 3.12 ( 1.246
2
6c
2
7b
4
8a
50
50
66
10
1 ( 0.547
4.58 ( 5.78
5.5 ( 3
2
0.5-2
0.5-16
2-8
0.5-8
0.062-2
8b
16
8
8c
4
2
bifonazole
miconazole
3.5 ( 2.74
8
1
33.33 0.65 ( 0.714
0.062
0.062
0.25
itraconazole 25
fluconazole 45.45 0.24 ( 0.150
0.096 ( 0.063
0.125 0.062-0.25
0.5
0.125-0.5
a Percentage of resistant strains (strains with MIC > 64 µg/
mL); see Experimental Section (Microbiology) for details.
a Reagents: (a) methyl propiolate, Bu4NF, THF.
demonstrated, in fact, that the binding of azoles to
CYP51 enzymes diverges to some extent from any fixed
distance and angle.40-42 Notably, similar arguments
have been recently taken into account to develop a
modified version of the docking program DOCK suitable
for CYP enzymes.43 Second, Catalyst’s HypoRefine
module,44 which allows one to generate quantitative
hypotheses with excluded volumes, thus accounting for
steric hindrance problems, was used in the new com-
putational protocol instead of HypoGen (Catalyst’s
original hypothesis generator). HypoGen algorithm, in
fact, has been reported to possibly have difficulties in
generating hypotheses that correlate well, if the steric
properties of the data set make a large contribution to
the activities.44 Compounds being inactive due to in-
compatible steric clashes with the target might likely
have their activities overpredicted by HypoGen, if they
would contain the same pharmacophore as active mol-
ecules. HypoRefine, conversely, by the strategic place-
ment of excluded volume spheres in the hypothesis, can
approximate steric repulsive interactions. Finally, mini-
mum inhibitory concentration mean values (MIC),
instead of the previously used MIC90 values,34,35 were
preferred in this study to express the anti-Candida
activities of the studied compounds. Although this choice
restricted the comprehensive set of compounds used in
the 3D-QSAR, as only formerly synthesized compounds
whose MIC was known could be included, it was in
agreement with the majority of the recent computa-
tional papers dealing with antifungal agents.45 The MIC
values of all the compounds (63 compounds plus flu-
predictive ability of a CoMFA model we proposed in the
past,34 in fact, was contradicted afterward by the
synthesis of other analogues,35 and a similar event
recurred whensin advance of the modeling studies
reported heresour subsequent pharmacophore model
(HYPO1)35 was applied to predict the activity of new
derivatives 1c-j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b, and
8a-c (r2
) 0.19). As a matter of fact, HYPO1’s
pred
predictive fault was not surprising if one relates its
framework38 to chemical structure and antifungal activ-
ity of 1c-j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b, and 8a-c
(Chart 1 and Table 3). MIC values of these molecules,
in fact, were revealed to be strongly dependent on the
presence of a substituent (possibly aliphatic) on the
nitrogen of the pyrrole ring, while HYPO1 could not
recognize this obviousness, as it could not match such
a substituent by any of its features.
Such being the case, further pharmacophore modeling
research was performed on newly and formerly synthe-
sized azoles (see Chart 2), using the software Catalyst,39
with the purpose to bring up to date our previous
model.38 On the occasion of this new study, the most
tricky points of our former computational protocol were
carefully revised, attempting to finally hit the “predic-
tive power” target, essential for the future rational
design of possibly active compounds.
First of all, the definition of the coordination feature
(HYPO1’s AWLP)38 was updated and renamed UNA
(unsubstituted nitrogen aromatic). Differently than
AWLP, UNA was defined as a sphere; X-ray data have

5146 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Chart 2. Azoles Used in the 3D-QSAR Studies
Di Santo et al.
conazole, miconazole, and bifonazole as the reference
Because no experimental data on the biologically
relevant conformations of the selected compounds were
available (for example, atomic coordinates derived from
X-ray crystallographic studies of their complexes with
the putative receptor), we resorted to a molecular
mechanics approach to build the conformational models
to be used for pharmacophore generation. By means of
Catalyst 2D-3D sketcher, alternative stereoisomers of
all compounds were automatically generated since, in
compounds), which covered 4 orders of magnitude, were
34,35
normalized and formulated as MICcompd/MICbifonazole
.
Each set of compounds considered during the hypoth-
eses generation was cut into a training set and a
complementary test set, and the MIC prediction of the
latter was used as the rule to select the hypothesis
representing “the pharmacophore model” among the
highest scoring ones generated by Catalyst.

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5147
Figure 1. Compound 1g (in yellow), the most active of the
whole set, mapped onto the pharmacophore hypothesis MOD1.
Pharmacophore features are color coded: blue for the unsub-
stituted aromatic nitrogen (UNA), red for aromatic ring (RA),
green for hydrophobic (HY1 and HY2).
agreement with the synthetic pathways, the chirality
of the asymmetric centers was not specified. A local
minimum geometry of each compound was built and
submitted to a conformational search protocol with the
aim of collecting a representative set of conformers,
chosen within a range of energetically reasonable
conformations. The following chemical features were
taken into account to build the pharmacophoric hypoth-
eses: UNA, hydrogen-bond acceptors (HBA), hydrogen-
bond donors (HBD), aromatic rings (RA), and hydro-
phobic groups (HY).46 Due to both the molecules’
flexibility and functional complexity, the hypothesis
generator was constrained to report hypotheses with at
least four features and to include UNA in each phar-
macophore, to satisfy the key interaction between azole
inhibitors and the enzyme.
Figure 2. Regression of experimental versus estimated/
predicted anti-Candida activities (MIC values) based on MOD1
(logarithmic scale), for each member of the training and test
sets.
was the highest scoring one and consisted of four
chemical features: the coordination interaction (UNA)
plus one aromatic ring (RA) and two hydrophobics (HY1
and HY2). Two findings more than any other led us to
credit MOD1: first of all, an interaction pattern very
similar to the one proposed by this hypothesis had been
already determined in a study on a series of obtusifoliol
14R-methyl demethylase azole inhibitors;48 secondarily,
and as expected, the aliphatic substituent on the pyrrole
nitrogen of all the most active compounds (1b, 1d, 1e,
1f, 1g, 1h) was found to match one of the two hydro-
phobic features (HY1).
A training set conceived to maximize the structural
and microbiological information content of the whole set
of studied compounds (66) was selected, accordingly to
the Catalyst guidelines, to generate a three-dimensional
pharmacophoric model. In particular, the following two
rules were applied: (i) inclusion of the most active
compound 1g in the set, due to the fact that Catalyst
pays particular attention to this molecule when it
generates the chemical feature space,47 and (ii) maxi-
mization of the kinds and relative positions (substitution
pattern) of the chemical features shared by the mol-
ecules, because the program recognizes the molecules
as collections of chemical features, not as assemblies of
atoms or bonds. Accordingly, 24 molecules were selected
out of the whole set of azole derivatives (namely,
compounds 1a, 1b, 1d, 1e, 1g, 1h, 1i, 1j, 3f, 4b, 6b, 8a,
9a, 9f, 9j, 10a, 11b, 11e, 11h, 12i, 12k, bifonazole,
fluconazole, miconazole), whose activities represented
all the orders of magnitude covered by each of the
structural groups A, B, C, D, and E. The training set
was characterized by MIC values spanning about 4
orders of magnitude, the minimum value 0.019 being
associated with compound 1g and the maximum value
97 associated with compound 9f.
The regression of experimental versus estimated/
predicted MIC values based on MOD1 (shown in Figure
2) exhibited a correlation coefficient r2 ) 0.84 for the
training set [root-mean-square deviation (rmsd) ) 1.23].
Comparison between estimated and experimentally
measured MIC values of the compounds showed, in the
worst case, a 24-fold difference, while, in most cases,
the comparison was inferior to a 2-fold difference. These
findings indicated a reliable ability of MOD1 to estimate
affinities within the training set. The execution of the
same analysis on the test set revealed, differently, a less
convincing scenario (see Figure 2), as a quite high
correlation coefficient (r2 ) 0.73) did not match a
homogeneous distribution of the corresponding marks
around the bisector. The MIC values of several com-
pounds were in fact predicted with very high accuracy,
while the inhibitory potency of some derivatives was
much overestimated (up to 2 orders of magnitude) and
only one compound was underestimated (9b, 6.9 times).
Experimental and calculated (estimated or predicted by
Catalyst) MIC values of all the compounds used in the
computational studies are shown in Table 4.
The prediction of the test set by MOD1 was regarded
as unsatisfactory and presumed to be the after effect of
possibly missing a quantitative correlation between MIC
activity data of some inhibitors and their binding
interaction with the enzyme active site (MIC has been
found to depend on permeability and metabolic factors
as well).49 With the aim to verify the substantiality of
such a guess, a reduced subset of 35 compounds out of
66 was selected for a second hypothesis generation, in
Further to this approach 10 hypotheses were gener-
ated by Catalyst. After a brief assessment of the
statistical parameters of all 10 hypotheses, it was
preferred to privilege the one with the best predictive
correlation coefficient on the complementary 42 com-
pound test set for further investigation. The selected
hypothesis (named MOD1 hereafter) is displayed in
Figure 1 superposed on the most active compound. It

5148 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Di Santo et al.
Table 4. Experimental and Estimated Anti-Candida Activity
Values for All the Compounds Used in the 3D-QSAR Study
MICcompd (µmol/mL)/MICbifonazole (µmol/mL)]
MICcompd/MICbifonazole
error
estimated estimated
compd
experimental (MOD1)
(MOD3) MOD1 MOD3
1aa
0.66
0.025
0.11
0.023
0.025
0.031
0.019
0.043
0.34
1.5
0.11
0.33
0.92
0.28
0.31
1.70
0.15
0.48
0.36
2.9
0.23
0.12
0.19
0.47
0.62
0.21
0.94
1.1
1.1
0.86
0.13
0.016
0.0064
0.052
0.26
0.0076
0.063
0.26
1.0
0.13
0.33
0.58
0.60
0.063
0.13
0.045
0.54
0.28
0.28
0.33
0.23
0.13
0.19
3.3
0.74
0.46
0.39
0.62
26
4.3
23
6.6
23
11
30
32
22
22
3.9
21
1.0
1.7
2.1
-5.8
-1.7
3.5
1.3
1.0
1.3
5.1
-6.7
-3.6
2.1
8.3
-2.5
1.5
-1.3
-1.5
1.2
1.0
-1.6
2.1
-4.9
-13
-3.3
1.1
-1.3
-10
1.5
1.9
1ba
1c
0.053
0.018
0.014
0.087
0.041
0.020
0.031
0.11
0.56
0.046
0.014
1.0
1da
1ea
1f
1ga
1ha
1ia
-1.4
-3.1
-2.7
-2.3
1ja
2b
2c
-24
3a
1.1
Figure 3. Compound 1g (in yellow) mapped onto the phar-
macophore hypothesis MOD2. Pharmacophore features are
color coded: blue for the unsubstituted aromatic nitrogen
(UNA), red for aromatic ring (RA1, RA2 and RA3). In black
are shown the excluded volumes (EV1 and EV2).
3b
0.034
0.016
0.014
0.011
0.80
0.80
0.039
0.0090
0.013
0.054
0.0098
0.74
1.4
-8.3
3c
-19
-120
-14
3d
3e
3fa
1.7
2.2
-73
-25
-9.3
-3.6
-48
4ba
5b
steadily overestimated by MOD1, and consequently,
only the six most active molecules from this class (1
order of magnitude) were considered in this generation,
while the remaining 21 were arbitrarily discarded. A
second set of pharmacophore hypotheses was thus built,
considering a training set and a test set of, respectively,
20 and 15 compounds, from which a pharmacophore was
selected (named MOD2 hereafter) according to the
criteria outlined above. MOD2 showed a relevant in-
crease of the correlation for both training set (r2 ) 0.94,
rmsd ) 0.84) and test set (r2 ) 0.90); however, as was
to be expected owing to the rule with which the training
set had been selected, the correlation coefficient de-
creased dramatically when all the class E compounds
were included into an exhaustive (46 compounds) test
set (r2 ) 0.52). MOD2 is displayed in Figure 3 super-
posed on 1g. This time, the coordination interaction plus
three aromatic rings (RA1, RA2, and RA3) and two
excluded volumes (EV1 and EV2) were found by Cata-
lyst to have pharmacophoric relevance. Unexpectedly,
and probably due to the too close proximity that might
have appeared between RA1 and a further feature on
the pyrrole nitrogen substituent, HY1 was ruled out
from MOD2.
5c
5e
6ba
6c
-1.5
-2.5
5.4
7b
1.2
6.5
8aa
8b
3.5
0.79
0.65
1.3
-1.2
-1.7
1.3
-2.0
-2.8
-1.6
19
8c
9aa
9b
1.0
1.4
45
9.8
6.9
9c
1.2
-38
-22
-8.1
-24
-7.6
-2.0
-23
-10
9d
63
51
97
2.9
-2.8
-7.7
-4.1
1.4
9e
6.3
9fa
4.1
9g
7.8
1.0
9h
17
44
49
23
8.4
1.8
9i
1.9
-1.4
-2.2
-1.0
-1.7
9.2
9ja
7.7
-6.3
-2.8
-1.2
-1.7
1.2
-6.4 -12
-2.4
1.6
9k
8.0
10aa
11a
11ba
11c
11d
11ea
11f
11g
11i
11ha
11j
6.8
5.8
2.3
1.2
7.3
2.8
0.70
3.1
2.7
1.8
4.1
1.1
1.4
1.7
3.8
1.4
1.4
-1.2
1.1
0.62
0.63
0.61
0.61
0.83
1.2
-4.4
1.1
-1.1
-5.1
-3.3
12
1.5
-2.1
-1.6
-1.3
-2.7
1.1
1.7
1.4
22
1.5
3.0
0.68
3.7
6.6
2.5
-1.4
-1.7
2.7
1.2
11k
12a
12b
12c
12d
12e
12f
12g
12h
12ia
12j
1.6
1.2
The inadequate predictivity of the 46-compound ex-
haustive test set by MOD2 clearly demonstrated its
overall unreliability. Therefore, a critical comparison
was performed between this hypothesis and MOD1 to
rationalize the not entirely satisfactory results obtained
up to then in our studies. Two structural aspects of
MOD2 appeared in fact to be really interesting and
deserved some more consideration. The first unfavorable
peculiarity was the presence of aromatic ring features
only (RA1, RA2, and RA3) to express the attractive
interactions of azoles with the amino acids of the active
site of CYP51. This result was clearly a computational
oversight largely induced by the exclusion from the
training set of all the less active E compounds. As a
consequence of that choice, in fact, Catalyst could
recognize the relevance of different dispositions in three-
dimensional space of aromatic rings (including the
pyrrole, mapped by RA1) of the selected A, B, C, D, and
E derivatives more effectively than the presence/absence
of a pyrrole substituent (HY1 in MOD1), peculiar of
1.6
-1.1
-2.2
3.8
1.8
-1.6
-4.0
1.7
63
1.1
15
11
3.0
0.52
13
19
17
0.18
21
0.92
4.0
-58
6.0
4.3
2.2
4.3
4.5
1.1
-5.5
-1.9
-2.6
-1.2
-1.3
-4.4
-1.6
-4.2
-3.8
-4.7
21
2.2
-1.4
-4.3
3.0
0.86
3.7
3.4
6.5
0.82
6.1
0.74
4.7
1.5
0.60
4.2
28
1.3
26
2.8
1.0
0.069
0.14
-1.7
-7.1
-1.2
-3.1
4.0
12ka
12l
bifonazolea
fluconazolea
miconazolea
0.59
0.27
8.6
2.0
4.4
a Training set compounds.
which only derivatives 8a-c, 9a-k, 10a, 11a-k, 1b,
1d, 1e, 1f, 1g, 1h, bifonazole, miconazole, and flucona-
zole were included. The majority of the derivatives
belonging to the structural class E (Chart 2)scovering
by itself four MIC orders of magnitudeshad been in fact

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5149
Figure 4. Compound 1g (in yellow) mapped onto the phar-
macophore hypothesis MOD3 (the final pharmacophore).
Pharmacophore features are color coded: blue for the unsub-
stituted aromatic nitrogen (UNA), red for aromatic ring (RA),
green for hydrophobic (HY1 and HY2). In black are shown the
excluded volumes (EV1 and EV2).
Figure 5. Regression of experimental versus estimated/
predicted anti-Candida activities (MIC values) based on MOD3
(logarithmic scale), for each member of the training and test
sets.
compounds E. The finding of two excluded volumes in
MOD2 (EV1 and EV2) was interpreted as the second
remarkable (favorably) indication provided by that
hypothesis. Excluded volumes had been in fact chased
since the beginning of this studies, and their appearance
enforced a primal guess that the average overprediction
of the 42-compound test set by MOD1 (see Figure 2) was
mainly due to missing the excluded volumes in that
pharmacophore. Actually, neither MOD1 nor MOD2
displayed polar features besides UNA; moreover, the
possibility that steric interactions might play a relevant
role in the binding of azoles into the active site of CYP51
had been already assessed both in our previous re-
search,35 based on the lipophilicity of our compounds
and, more generally, by other authors.19,45
The analysis just described clarified at a molecular
level the reasons for the minor reliability of MOD2 with
respect to MOD1 and suggested a possible adjustment
to be made to correct our computational protocol.
Catalyst’s default spacing valuesthat is, the minimum
distance between actual features locationswas hypoth-
esized to be the critical control parameter that had not
worked properly for our set of compounds. A further and
final set of three-dimensional hypotheses was conse-
quently generated from the first general training set,
after having shortened the value of the Catalyst spacing
parameter from 3.0 (default value) to 1.0 Å. The phar-
macophore with the best predictivity on the test set (the
highest scoring hypothesis) was then selected for further
investigation and named MOD3. MOD3 (shown in
Figure 4 with the most active compound 1g superposed)
was a four-features plus two excluded volumes hypoth-
esis, which showed a remarkable increase of the cor-
relation for the training set (24 compounds, r2 ) 0.93,
rmsd ) 0.80) with respect to MOD1. Comparison
between estimated and experimental MIC values gave,
in the worst case (fluconazole), a 8.5-fold difference and
in most cases was inferior to a 2-fold difference. Inter-
estingly, the regression based on MOD3 (shown in
Figure 5) exhibited a correlation coefficient for the whole
42-compound test set equal to that given by MOD1 (r2
) 0.73), whereas the distribution around the bisector
was definitely more homogeneous than that arising from
MOD1 (Figure 2), indicating the better capacity of
MOD3 to predict the activities of the test set. The MIC
value of all the test set compounds was in fact predicted
within its measured order of magnitude, with the
exception of compounds 3d, 9b, and 11i, whose anti-
fungal activity was underestimated by a factor of 13,
19, and 12, respectively.
MOD3 appears in Figure 4 as constituted by the
coordination feature UNA, two hydrophobics (HY1 and
HY2), one ring aromatic (RA), and two excluded volumes
(EV1 and EV2). The antifungal activity of 1g was
correctly estimated by the model, due to the fact that a
complete mapping of the molecule onto the pharma-
cophore was possible. In detail, the unsubstituted imi-
dazole nitrogen matched UNA, the pyrrole ring RA, and
the substituent on the pyrrole nitrogen HY1, while the
p-Cl substituent was mapped by HY2. The very low MIC
value of this molecule suggested that it possesses many
of the molecular features required for antifungal activ-
ity. On the basis of the experimentally determined
MIC1g/MICbifonazole (0.019, Table 4) and taking into
account that the pharmacophore model correctly esti-
mated the MIC of this compound (0.076, Table 4), one
might conclude that all the pharmacophore features
possibly accounted for major interactions between an-
tifungal agents and CYP51 active site.
MOD3 accounted for the trend of the biological data
associated with the compounds of both training and test
sets. While the most active compounds were found to
match all pharmacophoric features, the markedly de-
creased inhibitory activity of almost all the compounds
of classes A, B, C, and D and of the less active
compounds of class E (see Table 4) was ascribed as being
due to the unmatching of pharmacophoric features. It
is questionable whether the excluded volumes properly
represented the surrounding atoms in the binding
pocket of CA-CYP51 or more simply regions randomly
selected by the HypoRefine algorithm in the aligned
inactive molecules far away from the active ones that
could not contain any topology.44 Nevertheless, volume
spheres helped to improve the predictivity of MOD3, as
they specified spherical spaces in proximity to the
pharmacophore that could not contain any atoms or

5150 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Di Santo et al.
compound 1d, which was more potent than the refer-
ence drugs. Interestingly, the cited derivatives and some
congeners were found to be highly active against Can-
dida strains resistant to itraconazole, bifonazole, mi-
conazole, and fluconazole used as reference drugs.
As a summary of the 3D-QSAR studies reported in
this paper, the new pharmacophore model developed
and named MOD3 seems to be more definite, accurate,
flexible, and realistic with respect to the model we
proposed in our previous studies, as aromatic π-π
stacking interactions appear no more to be the sole
interactions able to modulate the activities of different
antifungal agents. Some key interactions, as well as
excluded volumes, further to the coordination bond of
azole antifungals with the demethylase enzyme are now
highlighted. The pharmacophore will be validated by
database searching and by the prediction of the anti-
fungal activities of new compounds. As a matter of fact,
however, a preliminary successful test of MOD3 has
been allowed51 by the recent publication of some novel
anti-Candida agents analogues of fluoxetine.52 This
approach, hopefully, will allow us to endow our com-
pounds with substituents able to form H-bonds within
the enzyme’s active site. Finally, even if the chosen
approach improved considerably our previous pharma-
cophore, nevertheless it highlighted the inherent limita-
tions in the use of MIC values for the development of
quantitative interaction models of antifungal azoles.
Soon, 3D models of CA-CYP51 will be probably available
to improve this ligand-based approach and to combine
it with a structure based one, to design new potent
derivatives with higher probability of success. The
possibility to discover a more accurate method to
describe the relevant coordination interaction between
the azole ring and the cytochrome heme moiety is still
under study.
Finally, the hypothesis that the antifungal activity
of fluconazole might partially be due to a hydrogen-
bonding interaction within the active site of CA-CYP51
is quite reasonable. It strengthens and adjusts at the
same time one of the conclusions of our previous
pharmacophore study on azole anti-Candida agents.35
In that paper we concluded that aromatic interactions
with amino acids localized in proximity of heme could
be responsible for the different activities of diverse
antifungal agents, while hydrogen bonds or similar
hydrophilic interactions (not pointed out in that study)
could fortuitously increase the affinity of an inhibitor
with respect to another one. The results of the studies
reported here corroborate that work if we amend the
above-mentioned conclusions assuming more hydropho-
bic than aromatic interactions, within the catalytic site
of CA-CYP51.
Figure 6. Fluconazole (in yellow) mapped onto the pharma-
cophore hypothesis MOD3 (the final pharmacophore). Phar-
macophore features are color coded: blue for the unsubstituted
aromatic nitrogen (UNA), red for aromatic ring (RA), green
for hydrophobic (HY1 and HY2). In black are shown the
excluded volumes (EV1 and EV2).
bonds, and this was a constraint preventing an advan-
tageous matching of conformations of the less active
compounds onto the pharmacophore.
The case of fluconazole could not be traced back to
the general paradigm; however, it is worth of note. In
Figure 6 is shown the matching of that compound onto
MOD3. Due to the peculiarity of its structure, flucona-
zole could match only three pharmacophore features out
of four: one of the two triazole rings was mapped by
UNA, the second triazole by RA, and the substituted
phenyl by one of the two hydrophobic features (HY1),
while the fourth feature of MOD3, HY2, was too far from
UNA to be possibly matched. On the other hand,
Catalyst did not recognize the OH of fluconazole as a
pharmacophoric feature, even though this group is
actually potentially able to act both as hydrogen-bond
donor and acceptor. The mismatching between MOD3
and fluconazole and the fact that the activity of flu-
conazole was underestimated 8.5 times by MOD3,
considered together, could be an indication that the high
antifungal activity of this compound might partially due
to a hydrogen-bonding interaction within the active site
of CA-CYP51, not recognized by our model (that conse-
quently underestimated it). Notably, a structure-based
de novo design study of inhibitors of CA-CYP51 has been
recently published, based on the three-dimensional
structure of a model of the enzyme built by homology
modeling.45c In that paper is reported that a catalytically
essential amino acid residue (Thr311), located on the
wall of the S1 subsite of the active site (already
described as one of the few hydrophilic areas in the
interior of the cytochromes P450s),50 “can act as an
important candidate to form hydrogen bond with
ligands”. This finding indirectly strengthens our specu-
lation.
Experimental Section
Conclusions
1. Chemistry. Melting points were determined on a Bu¨chi
530 melting point apparatus and are uncorrected. Infrared (IR)
spectra (Nujol mulls) were recorded on a Perkin-Elmer Spec-
trum-one spectrophotometer. 1H NMR spectra of imidazole
derivatives were recorded at 400 MHz on a Bruker AC 400
Ultrashield spectrophotometer. Column chromatographies
were performed on alumina (Merck; 70-230 mesh) or silica
gel (Merck; 70-230 mesh) column. All compounds were
routinely checked by TLC by using aluminum-baked silica gel
plates (Fluka DC-Alufolien Kieselgel 60 F254). Developed
plates were visualized by UV light. Solvents were reagent
The high activities showed by N-methyl- and N-
ethylpyrrolyl derivatives 1b, 3b, and 3c reported in
previous works31,35 led us to design and synthesize a
number of derivatives of imidazoles 1-7a, bearing
substituents on 1-position of the pyrrole ring. The newly
synthesized 1-[(aryl)(4-aryl-1-substituted-1H-pyrrol-3-
yl)methyl]-1H-imidazoles (1c-j, 2b,c, 3d-f, 4b, 5b-e,
6b,c, 7b, and 8a-c) showed high potency against
Candida albicans, and the most active derivative was

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5151
grade and, when necessary, were purified and dried by
standard methods. Concentration of solutions after reactions
and extractions involved the use of a rotary evaporator (Bu¨chi)
operating at a reduced pressure (ca. 20 Torr). Organic solutions
were dried over anhydrous sodium sulfate. Analytical results
agreed to within (0.40% of the theoretical values. The newly
synthesized azole derivatives were analyzed for C, H, N, and
Cl.
2. Syntheses. Specific examples presented below illustrate
the general synthetic procedures. The chemical and physical
properties of the newly synthesized derivatives and their
intermediates are reported in Tables 1 and 2.
1-(2,4-Dichlorophenyl)-3-(naphthalen-1-yl)prope-
none (16). A mixture of naphthalene-1-carboxaldehyde (21.9
g, 140 mmol) and 2′,4′-dichloroacetophenone (26.5 g, 140 mmol)
in ethanol (290 mL) was treated with a solution of sodium
hydroxide (13.6 g, 340 mmol) in water (290 mL). The mixture
was stirred at room temperature for 20 min and then treated
with water (600 mL) and the precipitate that formed was
filtered off. The filtrate was washed with water and recrystal-
lized from ethanol to afford 34.7 g (76%) of pure 16.
(2,4-Dichlorophenyl)[4-(naphthalen-1-yl)-1H-pyrrol-3-
yl]methanone (13h). A solution of 16 (16 g, 48.9 mmol) and
toluene-4-sulfonylmethylisocyanide (9.6 g, 48.9 mmol) in a
mixture of anhydrous dimethyl sulfoxide/ethyl ether (150:300
mL) was added dropwise into a well-stirred suspension of
sodium hydride (60% in white oil; 8.9 g, 220 mmol) in dry ethyl
ether (260 mL) under argon stream. After the addition, the
mixture was stirred at room temperature for 15 min and then
treated with water (900 mL). The formed precipitate was
filtered off, and the filtrate was washed with ethanol (3 × 50
mL) to afford pure 13h (14.77 g, 82%).
[4-(4-Chlorophenyl)-1-propyl-1H-pyrrol-3-yl]phenyl-
methanone (14b). A well-stirred suspension of 13a (1.5 g, 5.3
mmol), 1-bromopropane (4.7 g, 38 mmol), and anhydrous
potassium carbonate (1.5 g, 11 mmol) in dry N,N-dimethyl-
formamide (8.5 mL) was heated at 90 °C for 7 h. After cooling,
the mixture was treated with water (50 mL) and extracted
with ethyl acetate (3 × 100 mL). The organic extracts were
collected, washed with brine (3 × 50 mL), and dried, and the
solvent was evaporated under reduced pressure. The residue
was chromatographed on silica gel (chloroform as eluent) to
furnish pure 14b (1.3 g, 78%).
[4-(4-Chlorophenyl)-1-(3-methyl-2-butenyl)-1H-pyrrol-
3-yl]phenylmethanone (14f). A solution of 13a (1.0 g, 3.5
mmol) in dry tetrahydrofuran (40 mL) was added dropwise
into a well-stirred suspension of sodium hydride (60% in white
oil; 0.14 g, 3.5 mmol) in dry tetrahydrofuran (32 mL). After
the addition, the mixture was stirred at room temperature for
10 min and then a solution of 1-bromo-3-methyl-2-butene (0.5
g, 3.5 mmol) in dry tetrahydrofuran (12 mL) was added. The
mixture was stirred at room temperature for 1 h and then was
treated with water (100 mL). After the removal of the solvent,
the residue was extracted with ethyl acetate (3 × 100 mL).
The organic extracts were collected, washed with brine (3 ×
50 mL), and dried, and the solvent was evaporated. The
residue was chromatographed on aluminum oxide (chloroform
as eluent) to furnish 1.1 g (92%) of pure 14f.
[4-(4-Chlorophenyl)-1-ethenyl-1H-pyrrol-3-yl]phenyl-
methanone (14d). Into a well-stirred solution of 13a (2.0 g,
7.1 mmol) and tetrabutylammonium hydrogen sulfate (0.24
g, 0.71 mmol) in 1,2-dichloroethane (10 mL) cooled at 0 °C was
added NaOH 50% (13.5 mL). The resulting mixture was
refluxed for 4.5 h, allowed to cool, diluted with water (100 mL),
and extracted with ethyl acetate (3 × 150 mL). The collected
organic extracts were washed with brine (3 × 100 mL) and
dried, and the solvent was removed to give a yellow residue
that was chromatographed on aluminum oxide (chloroform as
eluent) to afford 14d (1.68 g, 77% yield) and 17 (0.25 g, 6%).
[4-(4-Chlorophenyl)-1-phenyl-1H-pyrrol-3-yl]phenyl-
methanone (14g). In a 5-mL glass tube were placed 13a (0.10
g, 0.54 mmol), cupric acetate (0.32 g, 1.8 mmol), phenylboronic
acid (0.13 g, 1.1 mmol), NMP-pyridine mixture (0.5:0.5 mL),
and a magnetic stir bar. The vessel was sealed with a septum
and placed into the microwave cavity. Microwave irradiation
of 60 W was used, the temperature being ramped from room
temperature to 120 °C. Once 120 °C was reached, the reaction
mixture was held at this temperature for 3 × 50 s (after each
cycle, the reaction vessel was cooled and cupric acetate (1.8
mmol), phenylboronic acid (1.1 mmol), and NMP-pyridine (0.5:
0.5 mL) were added). The reaction vessel was opened, the
mixture was diluted with tetrahydrofuran (10 mL) and filtered
off, and the solvent was evaporated. The residue was dissolved
in ethyl acetate (50 mL), washed with 1 N HCl (2 × 20 mL)
and then with brine (3 × 20 mL), dried, and evaporated. The
crude product was chromatographed on aluminum oxide
(chloroform as eluent) to furnish 0.05 g (41%) of pure 14g.
4-(4-Chlorophenyl)-1-ethyl-r-phenyl-1H-pyrrole-3-
methanol (15a). To a well-stirred suspension of lithium
aluminum hydride (0.08 g, 2.0 mmol) in dry tetrahydrofuran
(10 mL), at 0 °C, was added dropwise a solution of 14a (0.43
g, 1.4 mmol) in dry tetrahydrofuran (10 mL). After the
addition, the mixture was stirred at room temperature for 1.5
h and then carefully treated with crushed ice. The inorganic
precipitate was removed, and the solution was concentrated
and extracted with ethyl acetate (3 × 50 mL). The combined
organic extracts were washed with brine (3 × 50 mL), dried,
and evaporated to give 15a (0.33 g, 75%), which was used in
the following reaction without further purification.
1-[[4-(4-Chlorophenyl)-1-propyl-1H-pyrrol-3-yl]phenyl-
methyl]-1H-imidazole (1d). A solution of 15b (1.1 g, 3.4
mmol) and 1,1′-carbonyldiimidazole (2.4 g, 15 mmol) in anhy-
drous acetonitrile (60 mL) was stirred at room temperature
for 2 h. The solvent was removed, and the residue was
dissolved in ethyl acetate (100 mL). The organic solution was
washed with brine (3 × 50 mL) and dried, and the solvent was
evaporated. The crude product was chromatographed on
aluminum oxide (chloroform as eluent) to furnish 0.97 g (76%)
of pure 1d.
3-[3-(4-Chlorophenyl)-4-(1H-imidazol-1-ylphenylmeth-
yl)-1H-pyrrol-1-yl]-2-propenoic Acid Methyl Ester (1i). A
solution of 1 M tetrabutylammonium fluoride in tetrahydro-
furan (5.0 mL, 5.0 mmol) was added onto a well-stirred
solution of 1a (0.83 g, 2.8 mmol) in anhydrous tetrahydrofuran
(9 mL). A solution of methyl propiolate (0.35 g, 4.2 mmol) in
dry tetrahydrofuran (9 mL) was added dropwise. After the
addition the solution was stirred at room temperature for 6 h
and then treated with water (100 mL) and extracted with ethyl
acetate (3 × 150 mL). The organic solution was washed with
brine (3 × 100 mL) and dried, and the solvent was evaporated.
The crude product was chromatographed on aluminum oxide
(ethyl acetate as eluent) to give pure 1i (0.16 g, 37%).
3. Microbiology. Anti-Candida in Vitro Assays. The in
vitro antifungal activities of the derivatives 1c-j, 2b,c, 3d-
f, 4b, 5b-e, 6b,c, 7b, and 8a-c were tested against Candida
albicans. The antifungal potency was evaluated by means of
the minimal inhibitory concentration (MIC), using the serial
test in the broth microdilution modification method published
by the National Committee for Clinical Laboratory Standard
(NCCLS) method M 27-A2.36 MIC was defined as the lowest
concentration of test substances at which there was no visible
growth, as compared with a blank experiment, after the
present incubation time. For the preparation of the dilution
series, all derivatives were dissolved in dimethyl sulfoxide
(DMSO) at a concentration of 6.4 mg/mL and were stored
frozen at -20 °C. The solution was added to the medium at
the right concentration. Further progressive double dilutions
with a test medium furnished the required concentrations in
the range from 0.016 to 64 µg/mL. Blanks were prepared in
the test medium, without adding test substances.
Bifonazole, miconazole, itraconazole, and fluconazole were
used as reference antifungal drugs. Bifonazole, miconazole,
and itraconazole were dissolved in DMSO, while fluconazole
was dissolved in distilled water at a concentration of 6.4 mg/
mL. All microorganisms were preliminarily incubated at 28
°C for 24 h on RPMI 1640 (Sigma) medium buffered pH 7.0.
Antimicrobial tests were performed with RPMI 1640 medium
using inocula of approximately 104 cells/ mL. Readings of MICs

5152 Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16
Di Santo et al.
were taken at 24 h of growth at 28 °C. MIC values were
calculated by the expression MIC ) (ΣMICi)/st, where MICi is
the minimal inhibitory concentration values of all strains at
the used concentration Ci and st is the total number of strains.
MIC50 and MIC90 refer to MIC for 50 and 90% of strains,
respectively. %R refers to percentage of resistant strains at
64 µg/mL.
Twelve strains of C. albicans, four strains originally isolated
from AIDS patients with oropharyngeal candidiasis, four
isolates from HIV-infected subjects with recurrent vulvovagini-
tis, two reference strains (ATCC 24433 and 20891), and two
laboratory strains (3153 and CA2) were employed in these
studies. All the Candida strains used in this study has been
tested for their sensitivity to bifonazole, miconazole, itracona-
zole, and fluconazole. Among those being fluconazole-resistant,
three were derived from HIV infected patients (AIDS 68 and
AIDS 126), one was an ATCC resistant strain (20891), and
one was a laboratory-selected strain obtained through several
passages in fluconazole-enriched medium (CO23rflu). All
strains were identified by conventional diagnostic procedures.
tion run), none was found with a total cost lower than the
chosen hypothesis, so there was at least a 95% probability that
the three hypotheses did not represent chance correlation in
the data.
Test set molecules were fitted to the generated hypothesis
using the Catalyst fast fit feature. Fast fit refers to the method
of finding the optimum fit of the substrate to the hypothesis
among all the conformers of the molecule without performing
an energy minimization on the conformers of the molecule.
This method was also used to fit the training set molecules to
the pharmacophore.
Acknowledgment. Thanks are due to Ministero
della Sanita`, Istituto Superiore di Sanita`, “Programma
Nazionale di ricerca sull’AIDS” (grant no. 30F.19), and
to Italian MIUR (PRIN 2004) for partial support.
Elemental analyses were performed by Dott. M. Zan-
cato, University of Padova, Italy.
Supporting Information Available: Spectroscopic and
elemental analysis data of the newly synthesized derivatives
1c-j, 2b,c, 3d-f, 4b, 5b-e, 6b,c, 7b, and 8a-c. This material
is available free of charge via the Internet at http://pubs.
acs.org.
4. Molecular Modeling. All the biological activity MIC
mean values were expressed, after having converted all the
µg/mL values to µmol/mL units, as MICcompd/MICbifonazole, and
they span 4 orders of magnitude.
Because no experimental data on the biologically relevant
conformations of the studied compounds were available, the
conformational populations to be used for the pharmacophore
generation were built by the software Catalyst procedure.
Training and test set compounds were submitted to both fast
and best conformational searches, using the poling algorithm
and the CHARMm force field as implemented in Catalyst, with
the aim of collecting a representative set of conformers chosen
within a range of 20 kcal/mol with respect to the global
minimum. Moreover, alternative stereoisomers were automati-
cally considered during the conformational search, since the
chirality of the asymmetric center was not specified, in
agreement with the synthetic pathways.
Ten hypotheses were generated for each training set after
the selection of the following features: hydrophobic (HY), ring
aromatic (RA), hydrogen-bond acceptor (HBA), hydrogen-bond
donor (HBD), and the new defined feature “unsubstituted
nitrogen aromatic” (UNA). UNA was implemented by modify-
ing the definition of the “hydrogen-bond-acceptor lipid” feature,
already present into the internal software database, so as to
include only aromatic nitrogens bearing a lone pair. The N-2
atom of the triazole ring was excluded, as azole compounds
do not inhibit the enzyme by coordination of that atom. The
hypothesis generator was constrained to report hypotheses
with at least four features and to include UNA in each
pharmacophore. During the hypotheses generation, the Hy-
poRefine algorithm was used. All the control parameters
values were set to their default values except for the Spacing
value, which was set to a value of 100 picometers instead of
297, and for the Variable Tolerance value, which was set to a
value of 1 instead of 0, during the generation of MOD3.
A brief assessment of the statistical parameters of all 10
hypotheses was done after every generation through the cost
analyses reported here. MOD1: cost ) 114.8 bits. Total fixed
cost (ideal hypothesis) ) 95.7 bits. Cost of the null hypothesis
) 140.8 bits. Cost range over the 10 best generated hypotheses
) 7.1 bits. MOD2: cost ) 89.4 bits. Total fixed cost (ideal
hypothesis) ) 82.2 bits. Cost of the null hypothesis ) 131.6
bits. Cost range over the 10 best generated hypotheses ) 7.6
bits. MOD3: cost ) 117.3 bits. Total fixed cost (ideal hypoth-
esis) ) 109.1 bits. Cost of the null hypothesis ) 140.8 bits.
Cost range over the 10 best generated hypotheses ) 17.5 bits.
Special care was taken to test the models for chance
correlation. The three pharmacophores were evaluated for
statistical significance using a randomization trial procedure
derived from the Fischer method. Experimental activities of
the training set were scrambled 19 times to obtain 19
spreadsheets with randomized activity data. A total of 19
hypotheses were generated using the scrambled training sets,
and among the 190 resulting hypotheses (10 for each computa-
References
(1) Richardson, M. D. Opportunistic and Pathogenic Fungi. J.
Antimicrob. Chemother. 1991, 28 (Suppl. A), 1-11.
(2) Ablordeppey, S. Y.; Fan, P.; Ablordeppey J. H.; Mardenborough,
L. Systemic Antifungal Agents Against AIDS-Related Op-
portunistic Infections: Current Status and Emerging Drugs in
Development. Curr. Med. Chem. 1999, 6, 1151-1196.
(3) Beck-Sague, C. M.; Jarvis, W. R. Secular Trends in the Epide-
miology of Nosocomial Fungal Infections in the United States,
1980-1990. J. Inf. Dis. 1993, 167, 1247-1251.
(4) Koltin, Y. Target for Antifungal Drug Discovery. Annu. Rep.
Med. Chem. 1990, 25, 141-148.
(5) Ghannoum, M. A. Future of Antimycotic Therapy. Dermatol.
Ther. 1997, 3, 104-111.
(6) Asai, K.; Tsuchimori, N.; Okonogi, K.; Perfect, J. R.; Gotoh, O.;
Yoshida, Y. Formation of Azole-Resistant Candida albicans by
Mutation of Sterol 14-Demethylase P450. Antimicrob Agents
Chemother. 1999, 43, 1163-1169.
(7) Dawson, J. H.; Sono, M. Cytochrome P-450 and Chloroperoxi-
dase: Thiolate-Ligated Heme Enzymes. Spectroscopic Determi-
nation of Their Site Structures and Mechanistic Implications of
Thiolate ligation. Chem. Rev. 1987, 87, 1255-1276.
(8) Odds, F.; Brown, A. J. P.; Gow, N. A. R. Antifungal Agents:
Mechanism of Action. Trends Microbiol. 2003, 11, 272-
279.
(9) Adams, J. L.; Metcalf, B. W. Therapeutic Consequences of the
Inhibition of Sterol Metabolism. In Comprehensive Medicinal
Chemistry; Hansch, C., Sammes, P. G., Taylor, J. B., Eds.;
Pergamon Press: Oxford, England, 1990; Vol. 2, pp 333-
364.
(10) Buchel, K. H.; Draber, W.; Regel, E.; Plempel, M. Synthesen und
Eigenschften von Clotrimazol und Weiteren Antimykotischen
1-Triphenylmethylimidazolen. (Syntheses and Properties of Clo-
trimazole and Further Antimycotic 1-(Triphenylmethyl)imida-
zoles.) Arzneim.-Forsch. 1972, 22, 1260-1272.
(11) Bannatyne, R. M.; Cheung, R. Susceptibility of Candida albicans
to Miconazole. Antimicrobial Agents Chemother. 1978, 13, 1040-
1041.
(12) Heel, R. C.; Brodgen, R. N.; Carmine, A.; Morley, P. A.; Speight,
T. M.; Avery, G. S. Ketoconazole: A Review of Its Therapeutic
Efficacy in Superficial and Systemic Fungal Infections. Drugs
1982, 23, 1-36.
(13) Plempel, M.; Regel, E.; Buchel, K. H. Antimycotic Efficacy of
Bifonazole in Vitro and in Vivo. Arzneim.-Forsch. 1983, 33, 517-
524.
(14) Richardson, K.; Brammer, K. W.; Marriott, M. S.; Troke, P. F.
Activity of UK-49,858, a Bis-Triazole Derivative, Against Ex-
perimental Infections with Candida albicans and Trichophyton
mentagrophytes. Antimicrobial Agents Chemother. 1985, 27,
832-835.
(15) George, D.; Miniter, P.; Androile, V. T. Efficacy of UK-109496,
a New Azole Antifungal Agent, in an Experimental Model of
Invasive Aspergillosis. Antimicrobial Agents Chemother. 1996,
40, 86-91.
(16) Espinel-Ingroff, A.; Shadomy, S.; Gebhart, R. J. In Vitro Studies
with R 51,211 (Itraconazole). Antimicrobial Agents Chemother.
1984, 26, 5-9.

1-[(Aryl)(4-aryl-1H-pyrrol-3-yl)methyl]-1H-imidazole
Journal of Medicinal Chemistry, 2005, Vol. 48, No. 16 5153
(17) Sanglard D.; Odds, F. C. Resistance of Candida Species to
Antifungal Agents: Molecular Mechanisms and Clinical Con-
sequences. Lancet Infect. Dis. 2002, 2, 73-85.
(35) Tafi, A.; Costi, R.; Botta, M.; Di Santo, R.; Corelli, F.; Massa, S.;
Ciacci, A.; Manetti, F.; Artico, M. New Derivatives of 1-[(Aryl)-
[4-aryl-1H-pyrrol-3-yl]methyl]-1H-imidazole, Synthesis, Anti-
Candida Activity, and Quantitative Structure-Activity Rela-
tionship Studies. J. Med. Chem. 2002, 45, 2720-2732.
(36) National Committee for Clinical Laboratory Standards. 2002.
Reference methods for broth dilution antifungal susceptibility
testing of yeast, approved standard. NCCLS document M27-A2.
National Committee for Clinical Laboratory Standards, Waine,
PA.
(37) Liu, J.; Pan, D.; Tseng, Y.; Hopfinger, A. J. 4D-QSAR Analysis
of a Series of Antifungal P450 Inhibitors and 3D-Pharmacophore
Comparisons as a Function of Alignment. J. Chem. Inf. Comput.
Sci. 2003, 43, 2170-2179.
(38) HYPO1 was composed by one ad hoc defined aromatic-nitrogen-
with-lone-pair vectorial feature (AWLP) to simulate the coordi-
nation interaction of azole inhibitors with the iron atom of the
enzyme protoporphyrin system and three aromatic rings.35
(39) Catalyst 4.8, Accelrys, Inc.: 9685 Scranton Road, San Diego,
CA.
(40) Poulos, T. L.; Howard, A. J. Crystal Structures of Methyrapone-
andPhenylimidazole-InhibitedComplexesofCytochromeP-450cam.
Biochemistry 1987, 26, 8165-8174.
(41) Yano, J. K.; Koo, L. S.; Shuller, D. J.; Li, H.; Ortiz De Montellano,
P. R.; Poulos, T. L. Crystal Structure of a Thermophilic Cyto-
chrome P450 from the Archaeon Sulfolobus solfataricus. J. Biol.
Chem. 2000, 275, 31086-31092.
(42) Podust, L. M.; Poulos, T. L.; Waterman, M. R. Crystal Structure
of Cytochrome P450 14alpha-Sterol Demethylase (Cyp51) from
Mycobacterium tubercolosis in Complex with Azole Inhibitors.
Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 3086-3073.
(43) Verras, A.; Kuntz, I. D.; Ortiz De Montellano, P. R. Computer-
Assisted Design of Selective Imidazole Inhibitors for Cytochrome
P450 Enzymes. J. Med. Chem. 2004, 47, 3572-3579.
(44) http://www.accelrys.com/catalyst/hyporefine.html.
(45) (a) Sharma, P.; Rane, N.; Gurram, V. K. Synthesis and QSAR
Studies of Pyrimido[4,5-d]pyrimidine-2,5-dione derivatives as
Potential Antimicrobial Agents. Bioorg. Med. Chem. 2004, 14,
4185-4190. (b) Gollapudy, R.; Ajmani, S.; Kulkarni, A. Modling
and Interaction of Aspergillus fumigatus Lanosterol 14-R Dem-
ethylase with Azole Antifungals. Bioorg. Med. Chem. 2004, 12,
2937-2950. (c) Ji H.; Zhang W.; Zhang M.; Kudo M.; Aoyama
Y.; Yoshida Y.; Sheng C.; Song Y.; Yang S.; Zhou Y.; Lu¨ J.; Zhu
J.; Structure-Based de Novo Design, Synthesis, and Biological
Evaluation of Non-Azole Inhibitors Specific for Lanosterol 14R-
Demethylase of Fungi. J. Med. Chem. 2003, 46, 474-85. (d)
Wiktorowicz, W.; Markuszewski, M.; Krysinski, J.; Kaliszan, R.
Quantitative Structure-Activity Relationships Study of a Series
of Imidazole derivatives as Potential New Antifungal Drugs.
Acta Pol. Pharm. 2002, 59, 295-306.
(18) Podust, L. M.; Poulos, T. L.; Waterman, M. R. Crystal Structure
of Cytochrome P450 14R-Sterol Demethylase (CYP51) from
Mycobacterium tuberculosis in Complex with Azole Inhibitors.
Proc. Natl. Acad. Sci. U.S.A. 2001, 98(6), 3068-3073.
(19) Xiao, L.; Madison, V.; Chau, A. S.; Loebenberg, D.; Palermo, R.
E.; McNicholas, P. M. Three-Dimensional Models of Wild-Type
and Mutated Forms of Cytochrome P450 14R-Sterol Demethyl-
ase from Aspergillus fumigatus and Candida albicans Provide
Insights into Posocanazole Binding. Antimicrobial Agents
Chemother. 2004, 48, 568-574.
(20) De Groot, M. J.; Ekins, S. Pharmacophore Modeling of Cyto-
chromes P450. Adv. Drug Del. Rev. 2002, 54, 367-383.
(21) Di Santo, R.; Massa, S.; Artico, M. Synthesis of Biologically
Active Azoles via TosMIC. Farmaco 1993, 48, 209-229.
(22) Di Santo, R.; Massa, S.; Costi, R.; Simonetti, G.; Retico, A.;
Apuzzo, G.; Troccoli, F. Antifungal Agents. VIII. Synthesis and
Antifungal Activities of Bipyrryl Analogues of Bifonazole. Far-
maco 1994, 49, 229-236.
(23) Di Santo, R.; Costi, R.; Artico, M.; Massa, S.; Musiu, C.; Scintu,
F.; Putzolu, M.; La Colla, P. Antifungal estrogen-like imidazoles.
Synthesis and antifungal activities of thienyl and 1H-pyrrolyl
derivatives of 1-aryl-2-(1H-imidazol-1-yl)ethane. Eur. J. Med.
Chem. 1997, 32, 143-149.
(24) Di Santo, R.; Costi, R.; Artico, M.; Massa, S.; Musiu, C.; Milia,
C.; Putzolu, M.; La Colla, P. N-(1-Naphthylmethyl)-N-(1-alkyl-
4-aryl-1H-pyrrol-3-ylmethyl)methylami nes Related To Naftifine.
Synthesis And Antifungal Activity. Med. Chem. Res. 1997, 7,
98-108.
(25) Cirilli, R.; Costi, R.; Di Santo, R.; Ferretti, R.; La Torre, F.;
Angiolella, L.; Micocci, M. Analytical and Semipreparative
Enantiomeric Separation of Azole Antifungal Agents by High-
Performance Liquid Cromatography on Polysaccharide-Based
Chiral Stationary Phases. Application to In Vitro Biological
Studies. J. Chromatogr. A 2002, 942, 107-114.
(26) Artico, M.; Massa, S.; Di Santo, R.; Costi, R.; Retico, A.; Apuzzo,
G.; Simonetti, N. Antifungal Agents. 5. Chloro and Amino
Derivatives of 1,2-Diaryl-1-(1H-imidazol-1-yl)ethane with Potent
Antifungal Activities. Eur. J. Med. Chem. 1993, 28, 715-720.
(27) Botta, M.; Corelli, F.; Gasparrini, F.; Messina, F.; Mugnaini, C.
Chiral Azole Derivatives. 4.1. Enantiomers of Bifonazole and
Related Antifungal Agents: Synthesis, Configuration Assign-
ment, and Biological Evaluation. J. Org. Chem. 2000, 65(15),
4736-4739.
(28) Massa, S.; Di Santo, R.; Retico, A.; Artico, M.; Simonetti, N.;
Panico, S.; Fabrizi, G.; Lamba, D. Antifungal Agents. 1. Syn-
thesis and Antifungal Activities of Estrogen-Like Imidazole and
Triazole Derivatives. Eur. J. Med. Chem. 1992, 27, 495-502.
(29) Massa, S.; Di Santo, R.; Artico, M.; Costi, R.; Apuzzo, G.
Simonetti, G.; Artico, M. Novel In Vitro Highly Active Antifungal
Agents with Pyrrole and Imidazole Moieties. Med. Chem. Res.
1992, 2, 148-153.
(30) Massa, S.; Di Santo, R.; Costi, R.; Mai, A.; Artico, M.; Retico, A.;
Apuzzo, G.; Artico, M.; Simonetti, G. Antifungal Agents. 4.
Synthesis of neo-Isopyrrolnitrin and Other 3-Aryl-4-nitro-1H-
pyrroles via TosMIC. Med. Chem. Res. 1993, 3, 192-199.
(31) Massa, S.; Di Santo, R.; Artico, M.; Costi, R.; Di Filippo, C.;
Simonetti, G.; Retico, A.; Artico, M. Researches on Antibacterial
and Antifungal Agents. XV. 3-Aryl-4-[R-(1H-imidazol-1-yl)ben-
zyl]pyrroles with Potent Antifungal Activity. Eur. Bull. Drug
Res. 1992, 1, 12-17.
(32) Artico, M.; Di Santo, R.; Costi, R.; Massa, S.; Retico, A.; Artico,
M.; Apuzzo, G.; Simonetti, G.; Strippoli, V. Antifungal Agents.
9. 3-Aryl-4-[R-(1H-imidazol-1-yl)arylmethyl]pyrroles: A New
Class of Potent Anti-Candida Agents. J. Med. Chem. 1995, 38,
4223-4233.
(46) Greene, J.; Kahn, S. D.; Savoj, H.; Sprague, P.; Teig, S. L.
Chemical Function Queries for 3D Database Search. J. Chem.
Inf. Comput. Sci. 1994, 34, 1297-1308.
(47) Catalyst 4.5 Tutorials, August 1999, Molecular Simulation, Inc.,
San Diego, CA 92121.
(48) Bargar, T. M.; Secor, J.; Markley, L. D.; Shaw, B. A.; Erickson,
J. A. A Comparative Molecular Field Analysis of Obtusifoliol
14R-Methyl Demethylase Inhibitors. Pestic. Sci. 1999, 55, 1059-
1069.
(49) Upadhayaya, R. S.; Sinha, N.; Jain, S., Kishore, N., Chandra,
R.; Arora, S. K. Optically Active Antifungal Azoles: Synthesis
and Antifungal Activity of (2R,3S)-2-(2,4-Difluorophenyl)-3-(5-
[2-[4-aryl-piperazin-1-yl]-ethyl]-tetrazol-2-yl/1-yl)-1-[1,2,4]-tria-
zol-1-yl-butan-2-ol. Bioorg. Med. Chem. 2004, 12, 2225-38.
(50) Hasemann, C. A.; Kurumbail, R. G.; Boddupalli, S. S.; Peterson,
J. A.; Deisenhofer, J. Structure and Function of Cytochromes
P450: A Comparative Analysis of Three Crystal Structures.
Structure 1995, 2, 41-62.
(33) Massa, S.; Di Santo, R.; Costi, R.; Simonetti, G.; Retico, A.;
Apuzzo, G.; Artico, M. Antifungal Agents. III. Naphthyl and
Thienyl Derivatives of 1H-Imidazol-1-yl-4-phenyl-1H-pyrrol-3-
ylmethane. Farmaco 1993, 48, 725-736.
(34) Tafi, A.; Anastassopoulou, J.; Theophanides, T.; Botta, M.;
Corelli, F.; Massa, S.; Artico, M.; Costi, R.; Di Santo, R.; Ragno,
R. Molecular Modeling of Azole Antifungal Agents Active Against
Candida albicans. 1. A Comparative Molecular Field Analysis
Study. J. Med. Chem. 1996, 39, 1227-1235.
(51) Unpublished results.
(52) Silvestri, R.; Artico, M.; La Regina, G.; Di Pasquali, A.; De
Martino, G.; D’Auria, F. D.; Nencioni, L.; Palamara, A. T.
Imidazole Analogues of Fluoxetine, a Novel Classs of Anti-
Candida Agents. J. Med. Chem. 2004, 47, 3924-3926.
JM048997U