Design, synthesis and QSAR studies on N-aryl heteroarylisopropanolamines, a new class of non-peptidic HIV-1 protease inhibitors

Article abstract of DOI:10.1016/S0968-0896(02)00119-0

A series of N-aryl heteroarylisopropanolamines in which an indole or a 3-arylpyrrole moiety was linked to an aryl group through an isopropanolamine linker, were designed and synthesized as potential anti-HIV-1-PR agents. Series was tested for their ability in blocking PR activity. As a rule, indole derivatives of class 1 exhibited more potency than pyrrole analogues of class 2 while tert-butylamide substituents increased anti-PR potency. In fact, bis tert-butylamide 1e showed the highest activity with IC₅₀ = 25 μM. Even if not very potent, a simple class of anti-PR agents, with a facile synthetic pathway was discovered. QSAR studies on isopropanolamines 1 and 2 were performed in comparison with diarylbutanols, a new class of non peptidic anti-PR agents, recently discovered by Agouron Pharmaceuticals. QSAR and CoMFA models based on 30 diarylbutanols used as a training set were developed. The obtained models were used to investigate the binding mode of the newly synthesized derivatives 1 and 2. The results of this study suggest that N-aryl heteroarylisopropanolamines bind to the PR active site similarly to the diarylbutanols of Agouron.

Full text of DOI:10.1016/S0968-0896(02)00119-0

Bioorganic & Medicinal Chemistry 10 (2002) 2511–2526
Design, Synthesis and QSAR Studies on N-Aryl
Heteroarylisopropanolamines, a New Class of
Non-Peptidic HIV-1 Protease Inhibitors
Roberto Di Santo,a Roberta Costi,a Marino Artico,a,* Silvio Massa,b Rino Ragno,c
Garland R. Marshalld,* and Paolo La Collae
a
`
Istituto Pasteur-Fondazione Cenci Bolognetti, Dipartimento di Studi Farmaceutici, Universita degli Studi di Roma ‘La Sapienza’,
P.le A. Moro 5, I-00185 Roma, Italy
bb
`
Dipartimento Farmaco Chimico Tecnologico, Universita degli Studi di Siena, Via A. Moro 5, San Miniato, I-53100 Siena, Italy
Dipartimento di Studi di Chimica e Tecnologia delle Sostanze Biologicamente Attive, Universita degli Studi di Roma ‘La Sapienza’,
c
`
P.le Aldo Moro 5, I-00185 Roma, Italy
dCenter for Molecular Design, Washington University, St. Louis, MO 63110, USA
e
`
Dipartimento di Biologia Sperimentale, Sezione di Microbiologia, Universita degli Studi di Cagliari,
Cittadella Universitaria, I-09042 Monserrato, Cagliari, Italy
Received 23 October 2001; accepted 2 April 2002
Abstract—A series of N-aryl heteroarylisopropanolamines in which an indole or a 3-arylpyrrole moiety was linked to an aryl group
through an isopropanolamine linker, were designed and synthesized as potential anti-HIV-1-PR agents. Series was tested for their
ability in blocking PR activity. As a rule, indole derivatives of class 1 exhibited more potency than pyrrole analogues of class 2 while
tert-butylamide substituents increased anti-PR potency. In fact, bis tert-butylamide 1e showed the highest activity with IC50=25
mM. Even if not very potent, a simple class of anti-PR agents, with a facile synthetic pathway was discovered. QSAR studies on
isopropanolamines 1 and 2 were performed in comparison with diarylbutanols, a new class of non peptidic anti-PR agents, recently
discovered by Agouron Pharmaceuticals. QSAR and CoMFA models based on 30 diarylbutanols used as a training set were
developed. The obtained models were used to investigate the binding mode of the newly synthesized derivatives 1 and 2. The results
of this study suggest that N-aryl heteroarylisopropanolamines bind to the PR active site similarly to the diarylbutanols of Agouron.
# 2002 Elsevier Science Ltd. All rights reserved.
Introduction
2-fold rotational C2 symmetry with each monomer
containing one of the catalytic aspartic acid residues.3
PR is essential for cleavage of the gag/pol polyprotein
into viral structural proteins and enzymes during var-
ious stages of viral maturation. The ability of HIV pro-
tease to cleave phenylalanine(tyrosine)-proline peptide
bonds makes PR unique in its substrate specificity as
most proteolytic enzymes do not cleave on the N-term-
inal side of a prolyl residue.
The first approach to AIDS therapy utilized inhibitors
of immunodeficiency virus type-1reverse transcriptase
enzyme (RT).1 The rapid emergence of drug-resistant
strains during AIDS clinical trials with HIV-1RT inhi-
bitors, due to the high levels of mutations and high
replication levels of virus in vivo, led to targeting dif-
ferent viral enzymes such as HIV protease (PR) that
plays a vital role in viral replicating cycle.2
Studies on peptidomimetic protease inhibitors led to the
discovery of saquinavir, the first HIV protease inhibitor
to undergo clinical evaluation.4 After saquinavir, a
number of peptidomimetic derivatives were found to be
very potent PR inhibitors (ritonavir,5 indinavir,6
amprenavir,7 nelfinavir8) and are currently used in clin-
ical practice. Although PR inhibitors represented a
major advance in the management of HIV disease,
adverse side effects and viral resistance have encouraged
PR is classified as an aspartic proteinase containing the
typical Asp-Thr-Gly active-site sequence. The crystal
structure of HIV-1PR shows a dimer exhibiting exact
*Corresponding authors. Tel.: +39-06-446-2731; fax: +39-06-446-2731;
e-mail: marino.artico@uniroma1.it (M. Artico). Tel.: +1-314-362-1567;
fax: +1-314-362-0234 ; e-mail: garland@ibc.wustl.edu (G. Marshall)
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00119-0

2512
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
research to develop PR inhibitors with different profiles
of resistance to HIV strains. A further limitation of
current PR inhibitors is their complexity and the con-
comitant arduous synthetic pathway with high costs of
production. In order to overcome the above difficulties,
it is desirable to identify new non-peptidic HIV PR
inhibitors (PRI) of simpler structure. Up till now, only a
few non-peptidic drugs, such as nelfinavir, DuPont
Merck cyclic ureas9 or Parke-Davis dihydropyranones,10
reached trials, signifying the urgent need for new non-
peptidic inhibitors.
aromatic rings, being typical moieties shown to improve
anti-PR activity (i.e., saquinavir, indinavir, nelfinavir
and diarylbutanols).
As the three-dimensional (3-D) structure of HIV-1-PR
bound to some diarylbutanols13 was available, an
important goal of our study was to determine whether
or not the new PRIs share a common binding mode
with Agouron’s reference compounds. To reach this
objective, we have undertaken classical and three-
dimensional 3-D quantitative structure–activity rela-
tionships (QSARs) to derive statistical models using a
training set of a series of diarylbutanols.12ꢀ14 The mod-
els obtained were then used as tools to evaluate binding
modes of the new compounds (see Appendixes).
A structural analysis of peptidic drugs such as saquinavir,
nelfinavir, amprenavir,7,11 caused us to question whether
the central isopropanolamine unit of the above substrate
analogues was a binding determinant. Moreover, some
diarylbutanols12,13 were recently reported by Agouron
Pharmaceuticals to possess good anti-PR activity con-
sistent with the hypothesis that two aryl moieties, through
a four-atoms linker with carbinol unit mimicking the tran-
sition state, is a useful approach to design PR inhibitors.
Chemistry
Isopropanolamines 1 and 2 were obtained as reported in
Schemes 1–3. Indole derivatives 1 were synthesized by
alkylation of ethyl indole-3-carboxylate or ethyl 5-
chloroindole-3-carboxylate with epichloridrin in alka-
line medium (K2CO3) and the oxiranyl derivatives 5 and
6 which formed, were reacted with methyl anthranilate
at 90 ꢁC to obtain isopropanolamines 1a and 1d.
Finally, treatment of the latter compounds with tert-
BuNH2 and Al(CH3)3 gave a mixture of mono tert-
butylamides 1b and 1e and bis tert-butylamides 1c and
1f, respectively (Scheme 1).
Therefore, we designed and synthesized a series of N-
aryl heteroarylisopropanolamines 1 and 2 reported in
Figure 1in which an indole ring, mimicking the ami-
noindane group of indinavir, was linked to an aryl
moiety through an isopropanolamine group (structures
1). The indole was also replaced with its open form
represented by 3-arylpyrrole (structures 2) and a num-
ber of substituents on indole, arylpyrrole and aryl
groups were tested to better define the SAR in this ser-
ies. In particular, tert-butylcarboxamide groups were
linked in the 2 or 3-position of the aromatic or hetero-
A similar synthetic route led to pyrrole derivatives 2a–q.
In fact, the appropriate pyrrole 7–10 was alkylated with
Figure 1. Newly synthesized anti-PR agents compared with known PR inhibitors.

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2513
Scheme 1. (a) Epichlorohydrin, K2CO3; (b) methylanthranilate, 90 ꢁC; (c) tert-butylamine, Al(Me)3.
Scheme 2. (a) Epichlorohydrin, K2CO3; (b) substituted aniline or thiophenol, 90 ꢁC.
Scheme 3. (a) tert-Butylamine, Al(Me)3.
epichloridrin using potassium carbonate as a catalyst
and the oxiranylmethyl pyrroles 11–14 which formed,
underwent ring opening by reaction with methyl
anthranilate at 90 ꢁC without solvent to afford 2a–h,
j–m. Similarly, arylthiopropanols 2n–o were obtained by
reaction of oxirane 11 with the appropriate thiophenol
in refluxing ethanol (Scheme 2).
Finally, diester 2o was treated with Al(CH3)3/tert-
BuNH2 complex to give a mixture of mono tert-butyla-

2514
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
mide 2p and bis tert-butylamide 2q, while treatment of
pyrrole derivative 2h with the same reagent gave only
bis tert-butylamide 2i (Scheme 3).
The QSAR and 3-D QSAR were derived using as
training set a series of diarylbutanols taken from the
literature.12ꢀ14 The QSAR method used 3-D calculated
parameters compiled by mean of the VALIDATE16,19
and VALIDATE II20 procedures while the second
method was a comparative molecular field analysis
(CoMFA) as implemented in SYBYL.21
Indole and pyrrole derivatives 3, 4 and 7–10 used as
starting materials, were synthesized either by annulation
of tosylmethylisocyanide (TosMIC) with the appro-
priate Michael acceptor as already reported15,16 or by
esterification of commercially available pyrrole-2-car-
boxylic acid and indole-2-carboxylic acid. The anti-
HIV-1PR activity and the cytotoxicity of the deriva-
tives of classes 1 and 2 are reported in Table 1.
VALIDATE and VALIDATE II are hybrid approaches
(QSAR and scoring functions) that make the maximal
use of the three-dimensional information’s from the 3-D
coordinates of known ligand-receptor complexes. In this
approach the contributions to the overall free energy
variation upon complex formation were estimated by
twenty-nine chemical and physical-chemical calculated
parameters/scores, that were then correlated with the
binding affinities (pKis or pIC50). Due to the redun-
dancy of some parameters, fractional factorial design
(FFD) was used for variable selection as implemented in
the program GOLPE22 to select the more informative
parameters to define the best PLS model. Six para-
meters out of 29 were selected to be the more important
(Sfit; DipMom; HOMO; MW; Volume and
G_CDS_aq). The optimal number of principle compo-
nents (PCs) was then chosen on the basis of the highest
cross-validated q2 value and the smallest standard
deviation error of prediction (SDEP). Final PLS with
Molecular Modeling and QSAR Methods
All the computational work was performed on Silicon
Graphics computers (O2 R10000 225 MHz). The mea-
surements of biological activity used to develop the
QSAR and 3-D QSAR were expressed as (pIC50)=
ꢀlog(IC50). To the derivatives that were found to be
‘inactive’ (IC50>200 mM), an activity value equal to
50% of that of the compound reported as the less active
was arbitrarily assigned;17 for enantiomeric modeled
compound of which no information of the actual IC50
value, half of the corresponding racemate inhibitory
concentration was used.
Table 1. PR enzyme assay and cytotoxicity of indole and pyrrole derivatives 1–2
a
b
Compd
X
R1
R2
R3
R4
CC50
IC50
1a
1b
1c
1d
1e
1f
2a
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
—
—
—
—
—
—
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
NH
S
H
H
H
Cl
Cl
Cl
COOEt
CONH-tert-Bu
CONH-tert-Bu
COOEt
CONH-tert-Bu
CONH-tert-Bu
COOEt
COOMe
COOMe
CONH-tert-Bu
COOMe
COOMe
CONH-tert-Bu
—
—
—
—
—
—
2-Me
3-Me
4-Me
2-Cl
3-Cl
2-NH2
4-NH2
39
39
27
39
112
66
33
22.2
>200
47.3
42.7
0.8
4.3
37
101
61.3
45
>200
>200
75.3
50
nd
114
57
72
25
nd
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
H
H
H
H
H
H
H
H
32% (200 mM)
29% (200 mM)
27% (200 mM)
39% (200 mM)
36% (200 mM)
26% (200 mM)
22% (200 mM)
174
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
2-COOMe
2-CONH-tert-Bu
4-COOMe
H
CONH-tert-Bu
COOEt
COOEt
H
H
H
161
118
27% (200 mM)
21% (200 mM)
17% (200 mM)
24% (200 mM)
200
COOEt
H
COOEt
H
COOMe
H
H
H
2m
2n
2o
2p
2q
H
H
H
H
S
S
S
COOEt
COOEt
CONH-tert-Bu
2-COOEt
2-CONH-tert-Bu
2-CONH-tert-Bu
94
101
29% (200 mM)
nd
nd, not determined.
aCompound dose (mM) required to reduce the viability of mock-infected cells by 50% as determined by the MTT method.
bCompound dose (mM) required to reduce by 50% the activity of PR enzyme.

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2515
no cross-validation was then carried out using the opti-
mal number of PCs.
interactions at sterically bad contacts. For both
CoMFA fields, the atomic point charges were those
included in the AMBER all-atom force field imple-
mented in Macromodel 6.5. CoMFA regression ana-
lyses utilized the SYBYL implementation of the PLS
algorithm, initially with leave-one-out (LOO) cross-
validation to reduce the possibility of obtaining chance
correlations and two principal components (PCs).
CoMFA-like studies (SYBYL-CoMFA,23 GRID/
GOLPE22) are widely used tools for the study of
QSARs at the 3-D level (3-D QSAR). Unlike the pre-
vious QSAR approach, that relies on calculated para-
meters, such a 3-D QSAR relate the biological activity
of a series of molecules with molecular fields sampled at
grid points of a large 3-D lattice box where the pre-
aligned molecules are oriented. By means of suitable
molecular probes positioned at each grid point, the 3-D
QSAR columns are then computed using standard
molecular mechanics fields, that is steric from Lennard-
Jones interactions, electrostatic from Coulombic inter-
actions. Partial least square (PLS) is usually the regres-
sion method to develop the relationship between the
molecular fields and biological activity, where the opti-
mum number of components (latent variables) is deter-
mined by the maximum q2 value (cross-validated r2). An
important feature of this approach is the graphical
representation indicating the regions where the varia-
tion of the molecular fields of different molecules in a
data set is correlated with the variation of biological
activity.
The optimal number of PCs was then chosen on the
basis of the highest cross-validated q2 value, the smallest
standard error of prediction (SEP), and the minimum
number of components. To improve the signal-to-noise
ratio, the minimum sigma value was set to 2.0 kcal/mol.
The steric and electrostatic field columns were weighted
according to the CoMFA-STD default scaling option,
where a field is considered as a whole and every CoMFA
variable is affected by the overall field mean and stan-
dard deviation. Final PLS with no cross-validation was
then carried out using the optimal number of PCs.
Results and Discussion
Anti-protease activity
It is a well-established assumption to consider that a
similar class of molecules (congeneric series) binds to
the putative receptor site adopting a similar geometry
and orientation. All test and training-set molecules were
built using the MODIFY/ATOM module implemented
in the program SYBYL21 using as template molecules
the diarylbutanols AG1132, AG1157 and AG1284 to
model respectively the S and R absolute configurations
of the carbinol atoms.24 A training set of 30 known
diarylbutanol were compiled (Table 2) along with two
test-sets of 20 molecules represented from the S and R
configurations of the newly synthesized molecules of
classes 1 and 2.
Cytotoxicity and anti-protease activity of iso-
propanolamines 1–2 are reported in Table 1. As a rule
indole derivatives 1 were more potent than their pyrrole
counterparts (Table 1). Among the indole derivatives
(1a–f), compounds 1b–1e showed interesting PR inhibi-
tion with IC50 ranging from 25 to 114 mM. The intro-
duction of a chlorine atom at position 5 of the indole
ring increased the activity (compare 1e and 1b). In the
pyrrole series (2a–q), only compounds 2h–j and 2o
showed some slight inhibitory activity in the PR enzyme
assay.
Among tested compounds, tert-butylderivatives (1b, c, e
and 2i) showed the highest anti-PR activity with IC50
ranging from 25 to 161 mM, with the sole exception of
the thio derivative 2p. These data demonstrate the cru-
cial role of tert-butylamide group to obtain derivatives
with any anti-PR activity. Interesting potency was
shown by esters 1d and 2h (IC50=72 and 174 mM
respectively) or mixed amidoester derivatives 1e and 1b
(IC50=25 and 114 mM). The slight activity of the ortho-
ester 2h (IC50=174 mM) was retained when the ester
group was shifted to the para position (2j, IC50=118
mM), while replacement of carbomethoxy group with
methyl, chlorine, methoxy or amino group led to inac-
tive anti-PR compounds. The anti-PR activities were
evaluated by Prof. Chi-Huey Wong, The Scripps
Research Institute, La Jolla, CA, USA.26
Because of the high molecular similarity and employ-
ment of both ligand’s absolute configurations in the
training set and test set, no efforts were made to inves-
tigate different binding modes from those modeled from
the crystal structures.
For the QSAR parameters and the CoMFA molecular
alignment, all the modeled inhibitors were minimized
inside HIV-PR by replacing the co-crystallized inhibi-
tors in the complexes. Minimizations were performed
using MACROMODEL 6.525 allowing an 8 A core of
atoms to relax during the minimization. An external
fixed shell of 10 A was also included for long-range
interactions.
CoMFA calculations were performed with the QSAR
module of SYBYL and the following parameters. The
grid in which the aligned molecules were embedded was
regularly spaced (2 A) with dimensions of 28ꢂ24ꢂ20 A.
Steric and electrostatic interaction energies were calcu-
lated using a carbon sp3 probe with a +1charge, a dis-
tance-dependent dielectric constant (1/r), and an
energetic cutoff of 30 kcal/mol with no electrostatic
Finally, only a few compounds (viz., 2c, 2l, and 2m)
were not cytotoxic for MT-4 cells at doses as high as 200
mM, whereas the majority of propanolamines deriva-
tives showed CC50 values at concentration ranging
between 0.8 and 113.5 mM.
The most active derivatives in anti-protease assays,
namely indoles 1, were tested for their capability to

2516
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
Table 2. Training set of diarylbutanols
Compd
pIC50
Conf.
R1
X–Y
R2
Name
Refs
14
AG 1
14.92
R
CH2–CH2
ag1132R
PNAS 5R
PNAS 5R
AG 2
AG 3
AG 4
AG 5
2
4.92
6.25
6.25
8.31
S
R
S
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
ag1132R
ag1157R
ag1157S
ag1284R
14
14
14
14
3
4
5
PNAS 9S
PNAS 9S
PNAS 12R
R
AG 6
6
7
5.85
4.76
S
CH2–CH2
ag1284S
PNAS 12S
14
13
AG 7
R
CH2–CH2
JMC1996
15R
AG 8
AG 9
8
9
4.76
5.96
S
CH2–CH2
CH2–CH2
JMC1996
JMC1996
15S
16R
13
13
R
(continued on next page)

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2517
Table 2 (continued)
Compd
pIC50
5.96
Conf.
R1
X–Y
R2
Name
Refs
13
AG 10
AG 11
AG 12
AG 13
AG 14
AG 15
AG 16
10
11
12
13
14
15
16
S
CH2–CH2
JMC1996
16S
20R
20S
21S
21R
5R
6.22
6.22
5.46
5.46
4.90
4.90
R
S
S
R
R
S
CH2–CH2
CH2–CH2
CH2–CH2
CH2–CH2
CH2–S
JMC1996
JMC1996
JMC1996
JMC1996
JMC1996
JMC1996
13
13
13
13
13
13
CH2–S
5S
AG 17
AG 18
17
18
4.22
4.22
R
S
CH2–NH
CH2–NH
JMC1996
JMC1996
7R
7S
13
13
AG 19
AG 20
19
20
7.35
7.35
R
S
CH2–CH2
CH2–CH2
PNAS
PNAS
10R
10R
14
14
AG 21
217.96
R
CH2–CH2
PNAS
11R
14
(continued on next page)

2518
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
Table 2 (continued)
Compd
pIC50
7.96
Conf.
R1
X–Y
R2
Name
PNAS
Refs
14
AG 22
AG 23
AG 24
22
23
24
S
CH2–CH2
11S
13R
13S
9.00
9.00
R
S
CH2–CH2
PNAS
PNAS
14
14
CH2–CH2
AG 25
AG 26
AG 27
25
26
27
4.37
4.37
4.60
R
S
CH2–S
CH2–S
PNAS
PNAS
PNAS
4R
4S
7R
14
14
14
R
CH2–CH2
AG 28
AG 29
AG 30
28
29
30
6.32
4.76
4.76
S
R
S
CH2–CH2
PNAS
PNAS
PNAS
7S
8R
8S
14
14
14
CH2–S
CH2–S
selectively inhibit the HIV-1-induced cytopathic effect in
de novo infected MT-4 cells. Unfortunately, none of the
compounds under investigation inhibited HIV-1multi-
plication at concentrations below those cytotoxic for
MT-4 cells (data not shown), probably due to a low cell
penetration.
with an associated SEP value of 0.743 using only 2 PCs.
The non-cross-validated PLS resulted in a conventional
correlation coefficient (r2) value of 0.812 (Fig. 2) and a
standard error of estimation (SEE) of 0.646 log unit of
IC50 (F test=58.252). The resulting QSAR regression
equation achieved a strong correlation between the cal-
culated parameters and the biological activity. In gen-
eral the more important parameters were the steric
descriptors (MW and Volume) since their global con-
tribution to the regression equation represent more than
66% of the total (Table 3). This agrees with the fact that
the bigger are the structures, the better they can fit
Molecular Modeling and QSAR
QSAR. The calculated parameters when correlated with
the reported IC50 of the training-set led to an internal
self-consistent PLS model showing q2 value of 0.751

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2519
inside HIV-PR active site with more interaction with the
enzyme subsites.27,28
mode for some of Agouron’s compounds were available
from crystalline complexes, molecular alignment was
achieved by minimization of newly modeled HIV-PR/
inhibitor complexes upon replacement of the co-crys-
tallized compound with the analogous compounds.
pIC50 ¼ À4:284 À ð0:659ÞÃSfit þ ð0:010ÞÃDip-mom
À ð0:340ÞÃHOMO þ ð0:008ÞÃMW
Cross-validation led to a model with an optimal number
of PC of 2, a q2 value of 0.607 and a SEP of 0.933. The
corresponding non-cross-validated analysis (2 PCs)
resulted in a fitted r2 value of 0.869 (Fig. 2), with a SEE
of 0.540 (F-test=89.169).
þ ð0:008ÞÃVolume À ð0:053ÞÃG CDS aq
The QSAR model, when applied to the newly synthe-
sized compound 1b–e and 2a–p (used as external test-
set) calculated their pIC50 values (Table 4) with minimal
errors of prediction (Average Absolute Error of Predic-
tion for R configurations set, AAEPR=0.36 pIC50,
Average Absolute Error of Prediction for S configur-
ations set, AAEPS=0.48 pIC50). Moreover, the QSAR
model successfully predicted as inactive, compounds
2a–g, 2k–n, with the sole exception of 2p that was pre-
dicted an order of magnitude higher than the experi-
mental value. Application of different arbitrary activity
values to the inactive derivatives, such as 40 and 60% of
the least active compound, led to minimal difference in
the AAEPs values. (AAEPR-40%=0.33 pIC50, AAEPS-
40%=0.46 pIC50, AAEPR-60%=0.41pIC 50, AAEPS-
60%=0.51pIC 50).
The CoMFA model is in a good agreement with the
previous QSAR model showing a major contribution of
the steric field (Table 5) over the electrostatic field.
Moreover, the 3-D QSAR model was able to success-
fully predict the anti-PR active pIC50 (Table 4,
AAEPR=0.63 pIC50, AAEPS=0.74 pIC50).
Beside the models’ agreement, the 3-D QSAR correctly
showed (Table 4) its inability to discriminate anti-PR
active compounds (1b–e, 2h–j and 2o) from inactive
compounds (2a–g and 2k–n) since they were docked in
the enzyme active site although they do not compete
with any significant affinity. In the previous QSAR
model, the explicative variable were mostly chemical
and physical descriptors of ligands alone, and thus the
final regression equation is totally independent from
molecular alignment; while in the CoMFA model, the
molecules are all embedded in a three-dimensional grid
CoMFA. In developing a CoMFA model, the alignment
of the molecular training-set is a crucial part of the 3-D
QSAR. Since structural information on the binding
Table 3. Relative contributions of the calculated parameters to the regression equation
Parameter
Description
Norm. Coeff.
Fraction
Sfit
Steric fit between ligand and receptor18
0.087
0.058
0.126
0.3910.334
0.3810.326
0.127
0.074
0.050
0.108
Dip-Mom
HOMO
MW
Volume
G_CDS_aq
Dipole moment (AMSOL29
)
HOMO ligand energy (AMSOL29
Molecular weight (AMSOL29
Molecular volume (AMSOL29
Cavity-dispersion-solvent free energy (AMSOL29
)
)
)
)
0.109
Table 4. Predictions of compounds 1b–e and 2a–p
Compd
Actual pIC50
pIC50 QSAR predictions
pIC50 CoMFA predictions
R Configuration
S Configuration
R Configuration
S Configuration
1b
1c
1d
1e
2a
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
4.24
4.54
4.44
4.90
3.69
3.69
3.69
3.69
3.69
3.69
3.69
4.06
4.09
4.23
3.69
3.69
3.69
3.69
4.00
3.69
4.169
4.865
4.203
4.498
3.579
3.34
3.329
3.654
3.536
3.273
3.536
4.005
5.329
4.316
3.650
3.183
3.248
3.145
4.583
4.817
3.960
4.756
4.2614.599
4.478
3.480
3.385
3.349
3.913
3.7515.204 25
3.508
3.442
4.240
5.488
4.309
3.6815.1 4.722
2.180
1.952
3.047
4.223
4.794
4.365
4.593
4.755
4.459
5.058
5.099
5.325
5.277
5.1
5.082
5.015
4.908
5.267
5.149
4.776
5.192
4.486
5.072
5.132
4.936
4.466
4.879
4.831
4.834
5.292
4.971
07
5.029
4.421
4.964
5.358
5.316
4.833
4.383
5.160
5.342
5.570
2m
2n
2o
2p

2520
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
mimicking the receptor and assuming they all to bind to
active site. The CoMFA analysis is thus totally depen-
dent upon the molecular alignment.
Complementary to the previous QSAR model, the
CoMFA model permits the graphical display of the
results. The CoMFA steric and electrostatic field con-
tour maps gave information about those regions where
a modification of the chemical structure can enhance, or
decrease, the biological activity. The surfaces are the
representation of those lattice points, where differences
in field values are correlated with differences in activity.
In Fig. 3 are shown the maps for the CoMFA model
(StDEVꢂCOEFF), where the orange contours repre-
sent regions of decreased tolerance for positive charge
(20% contribution), the light blue contours represent
regions of decreased tolerance for negative charge (80%
contribution), the yellow contours represent regions of
low steric tolerance (20% contribution) and the green
contours represent regions of high steric tolerance (80%
contribution).
Figure 2. Training set experimental versus recalculated pIC50
.
The interpretation of these maps is highly subjective; the
absence of lattice points does not mean that a given
pharmacophoric element of the ligand has no influence
on the biological activity, but only that almost all the
compounds included in the training set exert similar
steric and/or electrostatic influence in that particular
area, or, alternatively, that the lack of polyhedra in
some regions delineates a less explored space.
Table 5. Relative contributions of CoMFA fields
Parameter
Norm. Coeff.
Fraction
COMFA (2145 vars) (Steric)
COMFA (2145 vars) (Electrostatic)
1.547
1.182
0.567
0.433
Table 6. Chemical and physical data of derivatives 1, 2, 5, 6 and 11–14
Compd
Formula
Mp (ꢁC)
Recrystna Solvent
Yield
Chromatb System
1a
lb
1c
ld
le
lf
2a
2b
2c
2d
2e
2f
2g
2h
2i
2j
2k
2l
2m
2n
2o
2p
2q
6
C22H24N2O5
C24H29N3O4
C27H36N4O3
C22H23ClN2O5
C24H25C1N3O4
C27H35C1N4O3
C23H26N2O3
C23H26N2O3
C23H26N2O3
C22H23ClN2O3
C22H23ClN2O3
C23H25N3O3
C22H25N3O3
C24H26N2O5
C29H38N4O3
C24H26N4O5
C22H23ClN2O3
C16H20N2O3
Cl6H20N2O3
C22H23NO3S
C25H27NO5S
C27H32N2O4S
C29H37N3O3S
C14H15NO3
97–99
oil
a
—
b
a
c
b
a
—
c
—
c
d
d
——
a
c
b
c
a
—
—
—
c
a
a
—
a
—
—
47
48
34
30
62
21
68
65
76
30
60
32
51
98
44
88
58
66
93
64
100
68
34
58
30
42
31D
96
55
A
B
B
138–140
108–110
124–126
132–134
84–86
oil
130–132
oil
76–78
137–138
118–120
oil
73–74
121–123
91–93
107–109
78–80
oil
oil
oil
96–98
53–54
83–84
oil
A
A
B
C
C
C
D
D
E
D
G
H
D
I
I
D
D
L
H
B
D
D
6
C14H14ClNO3
C16H17NO3
C16H16ClNO3
C10H13NO3
C10H13NO3
11
12
13
14
53–55
oil
oil
M
D
aRecrystallization solvents. a: Cyclohexane; b: benzene; c: benzene/cyclohexane; d: ethanol.
bChromatography eluent. A: n-hexane/ethyl eter 2:l; B: n-hexane/ethyl eter 1:1; C: ethyl acetate/chloroform 1:3; D: chloroform; E: ethyl acetate/
chloroform 1:1; F: ethyl acetate; G: ethyl acetate/chloroform 1:5; H: ethyl acetate/chloroform 1:4; I: n-hexane/acetone 5:2; L: ethyl acetate/chloro-
form 1:2; M: ethyl acetate/chloroform 1:20.

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2521
Figure 3. The contours of the electrostatic map are shown in orange and light blue (A), and those of the steric map are shown in yellow and green
(B). Greater values of ‘Bio-Activity Measurement’ are correlated with: more bulk near green; less bulk near yellow; more positive charge near light
blue, and more negative charge near orange. In C are showed the CoMFA contours overlapped to the HIV-Pr site color-coded by atom types. In A
and B the two enantiomers of compound 2e (S in white and R in purple) are also shown to help interpretation.
As can be seen in Fig. 3A, big light blue polyhedra are
situated near the region where the more active com-
pounds show a positive potential (i.e., an amidic or
hydroxylic hydrogen) by which an hydrogen bond can
take place, in fact, in ref 12 is reported on the experimental
hydrogen-binding scheme of some Agouron’s compound.
The introduction of hydrogen-bonding groups, where
lacking in this part of the inhibitors, could enhance the
complex formation. On the other hand, little orange
polyhedra are shown in those region where hydroxyethyl
oxygen is present in the more potent compounds. Four
yellow polyhedra (Fig. 3B) are situated that correspond
to enzyme sub sites S1, S10, S2 and S20 meaning that it is
not possible to enhance the bulkiness of the molecules in
these regions, but instead for some molecules, the
decrease of steric hindrance in this spatial region could
relieve the ligand–receptor steric interaction leading to
more active compounds. Two green polyhedra are
instead shown to correspond to the S3 and S30 sub sites
that in some cases are partially filled by the ethyl portion
of the hydroxyethyl group of the highly active com-
pounds. In Figure 3C is shown the superposition of the
CoMFA contour maps with the HIV-PR internal surface
(atom type color coded) by which is possible to observe
the good agreement between the red surface parts (amide
oxygen of Gly27/127 and Asp30/130) with the CoMFA
blue spots, and the white parts (enzyme backbone and
hydrophobic side chains) with the CoMFA yellow poly-
hedra. The latter observation fully supports the use of a
CoMFA model to describe the ligand–receptor interac-
tions, of course, when molecular alignment is based on
crystallographic information.
Conclusions
A series of N-aryl-pyrrolyl and N-aryl-indolylisopropa-
nolamine derivatives were designed and synthesized as

2522
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
inhibitors of HIV-1PR (transition-state analogues). The
cited compounds are strictly related to diaryl-
butanols12ꢀ14 synthesized by Agouron Pharmaceuticals
and were tested against HIV-1PR showing interesting
activity for such simple molecules. In fact, a number of
indolyl derivatives showed anti-PR activity with IC50
ranging from 25 to 114 mM. Unfortunately, few pyrrolyl
derivatives showed higher IC50s (118–200 mM) as most
of them were unable to inhibit the PR activity at con-
centrations lower than 200 mM. The ability of the most
potent PR inhibitors (viz., compounds 1) to block HIV-
1virus replication, was tested in cell-based assays, but
no indolyl derivative showed effective concentrations
(EC50) for viral replication lower than the cytotoxic
concentration.
Ethyl 1-oxiranylmethyl-4-phenyl-1H-pyrrole-3-carboxy-
late (11). A well stirred suspension of ethyl 4-phenyl-
1H-pyrrole-3-carboxylate (10.0 g, 46.5 mmol), anhy-
drous potassium carbonate (12.8 g, 93 mmol) and epi-
chlorohydrin (5.0 g, 54 mmol) in dry N,N-
dimethylformamide (50 mL), was heated at 90 ꢁC for 3
h. After cooling at room temperature, the mixture was
treated with water (200 mL) and extracted with ethyl
acetate (3ꢂ100 mL). The collected organic extracts were
washed with brine (3ꢂ300 mL), dried and the solvent
was evaporated to give crude derivative which was
chromatographed on silica gel column (chloroform as
eluent) to afford pure 11 (5.3 g, 42% yield): IR: n 1700
1
(CO) cmꢀ1; H NMR: d 1.28 (t, 3H, J1=7.2 Hz, CH3),
2.57 (dd, 1H, Jgem=4.6 Hz, Jvic1=2.5 Hz, oxyrane
CHH), 2.90 (dd, 1H, Jgem=4.6 Hz, Jvic2=4.0 Hz, oxy-
rane CHH), 3.30 (m, 1H, CH), 3.95 (m, 1H, CHH–pyr-
role), 4.22 (m, 3H, CHH–pyrrole and CH2CH3), 6.75 (d,
1H, J2=2.5 Hz, pyrrole C5-H), 7.30–7.50 (m, 6H, pyr-
role C2-H and benzene H); anal. C16H17NO3 (271.32)
C, H, N.
QSAR and CoMFA models were developed using as the
training set 30 of Agouron’s HIV-1PR inhibitors
belonging to the diarylbutanols series. The models
obtained were then used as tools to investigate on the
binding mode of newly tested classes 1 and 2. The
agreement between the predicted and the experimental
pIC50 values suggests that the new anti-PR active classes
1 and 2 likely bind into the enzyme active site similarly
to the diarylbutanol series that were used as template
for the molecular modeling and CoMFA alignment of
the tested derivatives.
This procedure was used for the synthesis of oxiranes 5,
6 and 12–14 and their spectroscopic data are reported
below
5. IR: n 1695 (CO) cmꢀ1 1H NMR: d 1.42 (t, 3H,
;
J=7.2 Hz, CH3), 2.52 (dd, 1H, Jgem=4.4 Hz, Jvic1=2.5
Hz, oxyrane CHH), 2.77 (dd, 1H, Jgem=4.4 Hz,
Jvic2=4.4 Hz, oxyrane CHH), 3.37 (m, 1H, CH), 4.36
(q, 2H, J=7.2 Hz, CH2CH3), 4.55 (dd, 1H, Jgem=15.0
Hz, Jvic3=5.1Hz, C HH-indole), 5.03 (dd, 1H,
Jgem=15.0 Hz, Jvic3=3.1Hz, CH H-indole), 7.12–7.69
(m, 5H, indole H); anal. C14H15NO3 (245.28) C, H, N.
Experimental
Chemistry
Melting points were determined on a Buchi 530 melting
point apparatus and are uncorrected. Infrared (IR) spec-
tra (Nujol mulls) were recorded on a Perkin–Elmer 297
instrument. 1H NMR spectra were recorded at 200 MHz
on a Bruker AC 200 spectrometer. Chemical shifts are
reported in d (ppm) units relative to the tetramethylsilane
(TMS). All compounds were routinely checked by TLC
and 1H NMR. TLC was performed by using Stratocrom
SIF Fluka (silica gel precoated plates with fluorescent
indicator) or Stratocrom ALF Fluka (aluminum oxide
precoated plates with fluorescent indicator). Developed
plates were visualized by UV light. Chromatographic
purifications were performed on Merck aluminum oxide
(70–230 mesh) and Merck silica gel (70–230 mesh). Sol-
vents were reagent grade and, when necessary, were pur-
ified and dried by standard methods. Concentration of
solutions after reactions and extractions involved the use
of a rotary evaporator operating at a reduced pressure of
approximately 20 torr. Organic solutions were dried over
anhydrous sodium sulfate. Analytical results agreed to
withinꢃ0.40% of the theoretical values. Microanalyses
were performed by Laboratories of Dipartimento di
Scienze Farmaceutiche, Universita di Padova, Italy.
Chemical and physical data of newly synthesized com-
pounds 1–2, 5–6 and 11–14 are reported in Table 6.
6. IR: n 1700 (CO) cmꢀ1 1H NMR: d 1.42 (t, 3H,
;
J=7.2 Hz, CH3), 2.48 (dd, 1H, Jgem=4.4 Hz, Jvic1=2.5
Hz, oxyrane CHH), 2.79 (dd, 1H, Jgem=4.4 Hz,
Jvic2=4.4 Hz, oxyrane CHH), 3.36 (m, 1H, CH), 4.33–
4.52 (m, 3H, CHH–indole and CH2CH3), 5.07 (dd, 1H,
Jgem=15.2 Hz, Jvic3=2.8 Hz, CHH–indole), 7.25–7.64
(m, 4H, indole H); anal. C14H14ClNO3 (280.73) C, H,
Cl, N.
1
12. IR: n 1690 (CO) cmꢀ1; H NMR: d 1.25 (t, 3H,
J1=7.2 Hz, CH3), 2.51(dd, 1H, Jgem=4.5 Hz, Jvic1=2.6
Hz, oxyrane CHH), 2.86 (dd, 1H, Jgem=4.5 Hz,
Jvic2=4.5 Hz, oxyrane CHH), 3.27 (m, 1H, CH), 3.87
(dd, 1H, Jgem=14.8 Hz, Jvic3=6.0 Hz, CHH–pyrrole),
4.15–4.26 (m, 3H, CHH–pyrrole and CH2CH3), 6.70 (d,
1H, J2=2.6 Hz, pyrrole C5-H), 7.26–7.44 (m, 5H, pyr-
role C2-H and benzene H); anal. C16H16ClNO3 (305.76)
C, H, Cl, N.
1
13. IR: n 1700 (CO) cmꢀ1; H NMR: d 1.35 (t, 3H,
J=7.0 Hz, CH3), 2.48 (dd, 1H, Jgem=4.7 Hz, Jvic1=2.5
Hz, oxyrane CHH), 2.85 (dd, 1H, Jgem=4.7 Hz,
Jvic2=4.5 Hz, oxyrane CHH), 3.26 (m, 1H, CH), 3.93
(dd, 1H, Jgem=14.8 Hz, Jvic3=5.7 Hz, CHH–pyrrole),
4.18–4.34 (m, 3H, CHH–pyrrole and CH2CH3), 6.60–
6.68 (m, 2H, pyrrole C4-H and C5-H), 7.35 (dd, 1H,
J2,4=1.5 Hz, J2,5=1.7 Hz, pyrrole C2-H); anal.
C10H13NO3 (195.22) C, H, N.
Syntheses
Specific examples presented below illustrate general
synthetic procedures.

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2523
1
14. IR: n 1680 (CO) cmꢀ1; H NMR: d 1.39 (t, 3H,
J=7.0 Hz, CH3), 2.47 (dd, 1H, Jgem=4.7 Hz, Jvic1=2.5
Hz, oxyrane CHH), 2.83 (dd, 1H, Jgem=4.7 Hz,
Jvic2=4.5 Hz, oxyrane CHH), 3.40 (m, 1H, CH), 4.25–
4.37 (m, 3H, CHH–pyrrole and CH2CH3), 4.83 (dd, 1H,
Jgem=14.7 Hz, Jvic3=5.6 Hz, CHH–pyrrole), 6.19 (dd,
1H, J3,4=4.0 Hz, J4,5=2.6 Hz, pyrrole C4-H), 6.92–7.03
(m, 2H, pyrrole C3-H and C5-H); anal. C10H13NO3
(195.22) C, H, N.
2d. IR: n 3400 (NH and OH), 1680 (CO) cmꢀ1
;
1H
NMR: d 1.23 (t, 3H, J=7.1Hz, CH ), 2.65 (bs, 1H,
3
OH), 3.08–3.35 (m, 2H, CH2NH), 3.90–4.24 (m, 5H,
CH2–pyrrole, CH2CH3, and CH), 4.60 (bs, 1H, NH),
6.62–7.49 (m, 11H, pyrrole and benzene H); anal.
C22H22ClN2O3 (398.89) C, H, Cl, N.
2e. IR: n 3380 (NH and OH), 1670 (CO) cmꢀ1 1H
;
NMR: d 1.23 (t, 3H, J=7.1Hz, CH ), 2.66 (bs, 1H,
3
OH), 3.05 (dd, 1H, Jgem=12.8 Hz, Jvic1=7.1Hz,
CHHNH), 3.23 (dd, 1H, Jgem=12.8 Hz, Jvic2=3.7 Hz,
CHHNH), 3.88–4.24 (m, 6H, CH2–pyrrole, CH2CH3,
CH and NH), 6.45–7.48 (m, 11H, pyrrole and benzene
H); anal. C22H22ClN2O3 (398.89) C, H, Cl, N.
Ethyl 1-[3-(2-methyl)phenylamino-2-hydroxypropyl]-4-
phenyl-1H-pyrrole-3-carboxylate (2a). A mixture of 11
(500 mg, 1.9 mmol) and 2-methylaniline (200 mg, 1.9
mmol) were heated at 90 ꢁC for 5 h. The cooled mixture
was dissolved in chloroform and chromatographed on
silica gel column (chloroform as eluent) to obtain pure
2a (490 mg, 68% yield): IR: n 3400 (NH and OH), 1680
2f. IR: n 3400, 3370, 3220, 3100 (NH2, NH and OH),
1
1670 (CO) cmꢀ1; H NMR: d 1.23 (t, 3H, J=7.1Hz,
(CO) cmꢀ1
;
1H NMR: d 1.26 (t, 3H, J=7.0 Hz,
CH3), 3.06 (dd, 1H, Jgem=12.8 Hz, Jvic1=7.6 Hz,
CHHNH), 3.26 (dd, 1H, Jgem=12.8 Hz, Jvic2=3.7 Hz,
CHHNH), 3.35 (bs, 4H, NH2, NH and OH), 3.90–4.25
(m, 5H, CH2–pyrrole, CH2CH3, and CH), 6.61–6.86 (m,
5H, pyrrole C5-H and aniline H), 7.26–7.50 (m, 6H,
pyrrole C2-H and benzene H); anal. C22H25N3O3
(379.46) C, H, N.
CH2CH3), 2.16 (s, 3H, CH3), 2.55 (bs, 1H, OH), 3.16
(dd, 1H, Jgem=13.0 Hz, Jvic1=7.3 Hz, CHHNH), 3.33
(dd, 1H, Jgem=13.0 Hz, Jvic2=3.9 Hz, CHHNH), 3.80–
4.25 (m, 6H, CH2–pyrrole, CH2CH3, CH and NH),
6.58–7.50 (m, 11H, pyrrole and benzene H); anal.
C23H26N2O3 (378.47) C, H, N.
The cited reaction was used to obtain propanolamines
1a, 1d, 2b–h, 2j–m, while arylthiopropanoles 2n–o were
obtained with the same procedure using ethanol as a
solvent and refluxing the mixture for 6 h. Spectroscopic
data of these derivatives are reported below:
2g. IR: n 3400, 3340, 3100 (NH2, NH and OH), 1670
1
(CO) cmꢀ1; H NMR: d 1.23 (t, 3H, J=7.1Hz, CH ),
3
2.98 (dd, 1H, Jgem=12.6 Hz, Jvic1=7.2 Hz, CHHNH),
3.19 (dd, 1H, Jgem=12.6 Hz, Jvic2=3.5 Hz, CHHNH),
3.20 (bs, 4H, NH2, NH and OH), 3.85–4.25 (m, 5H,
CH2–pyrrole, CH2CH3, and CH), 6.55 (m, 4H, aniline
H), 6.69 (d, 1H, J2,5=2.3 Hz, pyrrole C5-H), 7.21–7.50
(m, 6H, pyrrole C2-H and benzene H); anal.
C22H25N3O3 (379.46) C, H, N.
1a. IR: n 3480, 3340 (NH and OH), 1680 (CO) cmꢀ1; 1H
NMR: d 1.40 (t, 3H, J1=7.2 Hz, CH2CH3), 3.07 (d, 1H,
J2=4.2 Hz, OH), 3.37 (m, 2H, CH2NH), 3.85 (s, 3H,
OCH3), 4.31–4.41 (m, 3H, CH2CH3 and CH), 4.67 (m,
2H, CH2–indole), 6.56–7.93 (m, 12H, indole and ben-
zene H), 8.03 (bs, 1H, NH); anal. C22H24N2O5 (396.44)
C, H, N.
2h. IR: n 3420, 3350 (NH and OH), 1670 (CO) cmꢀ1; 1H
NMR: d 1.23 (t, 3H, J=7.1Hz, CH CH3), 1.60 and
2
2.39 (2bs, 2H, NH and OH), 3.31(m, 2H, C H2NH),
3.86 (s, 3H, OCH3), 4.02 (m, 2H, CH2–pyrrole), 4.14–
4.24 (m, 3H, CH2CH3 and CH), 6.62–6.72 (m, 3H, pyr-
role C5-H and anthranilate C3-H and C5-H), 7.21–7.50
(m, 7H, pyrrole C2-H, anthranilate C4-H and benzene
H), 7.93 (m, 1H, anthranilate C6-H); anal. C24H26N2O5
(422.48) C, H, N.
1d. IR: n 3490, 3340 (NH and OH), 1680 (CO) cmꢀ1; 1H
NMR: d 1.40 (t, 3H, J=7.0 Hz, CH2CH3), 2.95 (bs, 1H,
OH), 3.36 (m, 2H, CH2NH), 3.85 (s, 3H, OCH3), 4.32–
4.42 (m, 3H, CH2CH3 and CH), 4.70 (m, 2H, CH2-
indole), 6.59–6.71(m, 2H, anthranilate C3-H and C5-
H), 7.21–7.93 (m, 6H, indole and benzene H), 8.03 (bs,
1H, NH); anal. C22H23ClN2O3 (430.89) C, H, Cl, N.
2j. IR: n 3380 (NH and OH), 1685 (CO) cmꢀ1 1H
;
NMR: d 1.22 (t, 3H, J1=7.2 Hz, CH2CH3), 2.79 (d,
1H, J2=4.2 Hz, OH), 3.14–3.31 (m, 2H, CH2NH),
3.85 (s, 3H, OCH3), 3.90–4.10 (m, 3H, CH2–pyrrole
and CH), 4.18 (q, 2H, J1=7.2 Hz, CH2CH3), 4.50
(bs, 1H, NH), 6.56 (m, 2H, aniline C2-H and C6-H),
6.69 (d, 1H, J2,5=2.5 Hz, pyrrole C5-H), 7.25–7.48
(m, 6H, pyrrole C2-H and benzene H), 7.85 (m, 2H,
aniline C3-H and C5-H); anal. C24H26N2O5 (422.48) C,
H, N.
2b. IR: n 3380 (NH and OH), 1680 (CO) cmꢀ1 1H
;
NMR: d 1.22 (t, 3H, J=7.0 Hz, CH2CH3), 2.27 (s, 3H,
CH3), 2.80 (bs, 1H, OH), 3.05 (dd, 1H, Jgem=13.0 Hz,
Jvic1=7.2 Hz, CHHNH), 3.24 (dd, 1H, Jgem=13.0 Hz,
Jvic2=3.8 Hz, CHHNH), 3.83–4.23 (m, 6H, CH2–pyr-
role, CH2CH3, CH and NH), 6.41–7.49 (m, 11H, pyr-
role and benzene H); anal. C23H26N2O3 (378.47) C, H,
N.
2c. IR: n 3440, 3380 (NH and OH), 1670 (CO) cmꢀ1; 1H
NMR: d 1.22 (t, 3H, J=7.2 Hz, CH2CH3), 2.24 (s, 3H,
CH3), 2.60 (bs, 1H, OH), 3.05 (dd, 1H, Jgem=13.0 Hz,
Jvic1=7.4 Hz, CHHNH), 3.25 (dd, 1H, Jgem=13.0 Hz,
Jvic2=3.7 Hz, CHHNH), 3.70–4.24 (m, 6H, CH2–pyrrole,
CH2CH3, CH and NH), 6.49–7.77 (m, 11H, pyrrole and
benzene H); anal. C23H26N2O3 (378.47) C, H, N.
2k. IR: n 3280 (NH and OH), 1665 (CO) cmꢀ1 1H
;
NMR: d 1.29 (t, 3H, J=7.0 Hz, CH3), 2.50 (bs, 1H,
OH), 3.20 (dd, 1H, Jgem=13.2 Hz, Jvic1=7.2 Hz,
CHHNH), 3.31(dd, 1H, Jgem=13.2 Hz, Jvic2=4.0 Hz,
CHHNH), 3.95–4.29 (m, 6H, CH2–pyrrole, CH2CH3,
CH and NH), 6.66–7.47 (m, 11H, pyrrole and benzene
H); anal. C22H23ClN2O3 (398.89) C, H, Cl, N.

2524
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2l. IR: n 3480, 3440, 3340 (NH and OH), 1670 (CO)
With the same procedure were synthesized the amides 1b, c,
1e, f, 2p, q and their spectroscopic data are reported below.
1
cmꢀ1; H NMR: d 1.33 (t, 3H, J1=7.1Hz, CH ), 2.50
3
(d, 1H, J2=3.6 Hz, OH), 3.06 (dd, 1H, Jgem=13.1 Hz,
Jvic1=7.2 Hz, CHHNH), 3.27 (dd, 1H, Jgem=13.1 Hz,
Jvic2=3.4 Hz, CHHNH), 3.87–4.31(m, 6H, C H2–pyr-
role, CH2CH3, CH and NH), 6.59–7.35 (m, 8H, pyrrole
H and benzene H); anal. C16H20N2O3 (288.34) C, H, N.
1b. IR: n 3320, 3240 (NH and OH), 1680 (ester CO),
1630 (amide CO) cmꢀ1; 1H NMR: d 1.48 (s, 9H, CCH3),
3.41(m, 2H, C H2NH), 3.92 (s, 3H, OCH3), 4.32 (m, 1H,
CH), 4.61(m, 2H, C H2–indole), 5.69 (bs, 1H, OH), 6.16
(bs, 1H, CONH), 6.58–7.40 (m, 7H, indole and benzene
H), 8.10 (bs, 1H, NH aniline); anal. C24H29N3O4
(423.51) C, H, N.
1
2m. IR: n 3380 (NH and OH), 1685 (CO) cmꢀ1; H
NMR: d 1.35 (t, 3H, J=7.2 Hz, CH3), 2.96 (bs, 1H,
OH), 3.08 (dd, 1H, Jgem1=12.7 Hz, Jvic1=6.8 Hz,
CHHNH), 3.29 (dd, 1H, Jgem1=12.7 Hz, Jvic2=4.2 Hz,
CHHNH), 4.10–4.34 (m, 5H, CHH-pyrrole, CH2CH3,
CH and NH), 4.60 (dd, 1H, Jgem2=13.6 Hz, Jvic3=3.5 Hz,
CHH-pyrrole), 6.15 (dd, 1H, J3,4=4.0 Hz, J4,5=2.7 Hz,
pyrrole C4-H), 6.63–7.25 (m, 7H, pyrrole C3-H and C5-H
and benzene H); anal. C16H20N2O3 (288.34) C, H, N.
1c. IR: n 3360, 3260, 3160 (NH and OH), 1640 (amide CO)
cmꢀ1; 1H NMR: d 1.47 (s, 18H, CCH3), 3.30–3.49 (m, 2H,
CH2NH), 4.27 (m, 1H, CH), 4.59 (m, 2H, CH2–indole),
5.67 (d, 1H, J=4.4 Hz, OH), 5.94 and 6.18 (2bs, 2H,
CONH), 6.58–7.67 (m, 9H, indole and benzene H), 7.70
(bs, 1H, NH aniline); anal. C27H36N4O3 (464.61 ) C, H, N.
2n. IR: n 3420 (OH), 1680 (CO) cmꢀ1; 1H NMR: d 1.23
(t, 3H, J1=7.2 Hz, CH3), 2.71(d, 1H, J 2=3.2 Hz, OH),
2.86 (dd, 1H, Jgem=13.7 Hz, Jvic1=6.9 Hz, CHHS),
3.03 (dd, 1H, Jgem=13.7 Hz, Jvic2=4.5 Hz, CHHS),
3.91–4.24 (m, 5H, CH2–pyrrole, CH2CH3, and CH),
6.66 (d, 1H, J2,5=2.5 Hz, pyrrole C5-H), 7.23–7.48 (m,
11H, pyrrole C2-H and benzene H); anal. C22H23NO3S
(381.49) C, H, N, S.
1e. IR: n 3320, 3240, 3060 (NH and OH), 1680 (ester
1
CO), 1625 (amide CO) cmꢀ1; H NMR: d 1.47 (s, 9H,
CCH3), 3.39 (m, 2H, CH2NH), 3.87 (s, 3H, OCH3), 4.25
(m, 1H, CH), 4.55 (m, 2H, CH2–indole), 5.53 (d, 1H,
J=4.4 Hz, OH), 6.19 (bs, 1H, CONH), 6.59–7.95 (m,
8H, indole and benzene H), 8.05 (bs, 1H, NH aniline);
anal. C24H28ClN3O4 (457.96) C, H, Cl, N.
1f. IR: n 3380, 3270, 3150 (NH and OH), 1640 (amide CO)
cmꢀ1; 1H NMR: d 1.47 (s, 18H, CCH3), 3.22–3.53 (m, 2H,
CH2NH), 4.24 (m, 1H, CH), 4.58 (m, 2H, CH2–indole),
5.52 (bs, 1H, OH), 5.94 and 6.20 (2bs, 2H, CONH), 6.55–
7.66 (m, 8H, indole and benzene H), 7.65 (bs, 1H, NH
aniline); anal. C27H35ClN4O3 (499.05) C, H, Cl, N.
2o. IR: n 3440 (OH), 1700 (CO) cmꢀ1; 1H NMR: d 1.23
(t, 3H, J1=7.1Hz, CH CH3 pyrrole ester), 1.40 (t, 3H,
J2=7.1Hz, CH CH3 benzene ester), 2.86–3.13 (m, 2H,
2
2
CH2S), 3.27 (bs, 1H, OH), 4.04 (m, 3H, CH2–pyrrole
and CH), 4.20 (q, 2H, J1=7.1Hz, C H2CH3 pyrrole
ester), 4.37 (q, 2H, J2=7.1Hz, C H2CH3 benzene ester),
6.70 (d, 1H, J2,5=2.4 Hz, pyrrole C5-H), 7.21–7.48 (m,
9H, pyrrole C2-H and benzene H), 7.88 (dd, 1H,
Jo=7.7 Hz, Jm=1.2 Hz, phenylthio C3-H); anal.
C25H27NO5S (453.55) C, H, N, S.
2p. IR: n 3300 (NH and OH), 1700 (ester CO), 1630
1
(amide CO) cmꢀ1; H NMR: d 1.23 (t, 3H, J1=7.1Hz,
CH2CH3), 1.487 (s, 9H, CCH3), 2.69 (dd, 1H,
Jgem=14.0 Hz, Jvic1=8.2 Hz, CH2S), 3.01(dd, H1 ,
Jgem=14.0 Hz, Jvic2=3.0 Hz, CH2S), 3.94–3.96 (m, 3H,
CH2–pyrrole and CH), 4.17 (q, 2H, J1=7.1Hz,
CH2CH3), 5.23 (d, 1H, J2=4.9 Hz, OH), 5.78 (bs, 1H,
CONH), 6.65 (d, 1H, J2,5=2.3 Hz, pyrrole C5-H), 7.23–
7.52 (m, 10H, pyrrole C2-H and benzene H); anal.
C27H32N2O4S (480.62) C, H, N, S.
N-tert-Butyl 1-{2-hydroxy-3-[2-(tert-butylaminocarbo-
nyl)phenylamino]propyl}-4-phenyl-1H-pyrrole-3-carboxa-
mide (2i). tert-Butylamine (670 mg, 9.2 mmol) was
added into a cooled solution (T=ꢀ12 ꢁC) of trimethy-
laluminum (4.7 mL 2.0 M in hexane, 9.6 mmol) in tolu-
ene (45 mL) and the mixture was stirred at room
temperature for 45 min, then cooled at 0 ꢁC and finally
treated with a solution of 2h (880 mg, 2.1mmol) in tol-
uene (30 mL). The resulting solution was refluxed for 12
h, then cooled at room temperature and treated with
1N hydrochloric acid (100 mL) and finally extracted
with ethyl acetate (3ꢂ50 mL). The organic extracts were
collected, washed with brine (20 mL) and dried.
Removal of the solvent gave a residue which was chro-
matographed on silica gel column (chloroform/ethyl
acetate 2:1as eluent then ethyl acetate and finally ethyl
acetate/ethanol 1:1) to obtain 450 mg of 2i (44% yield):
2q. IR: n 3400, 3260 (NH and OH), 1625 (CO) cmꢀ1; 1H
NMR: d 1.18 (s, 9H, CCH3 pyrrole), 1.48 (s, 9H, CCH3
benzene), 2.77 (dd, 1H, Jgem=13.6 Hz, Jvic1=7.9 Hz,
CH2S), 2.98 (dd, 1H, Jgem=13.6 Hz, Jvic2=2.8 Hz,
CH2S), 3.87–3.98 (m, 3H, CH2–pyrrole, and CH), 5.21
(d, 1H, J=4.5 Hz, OH), 5.34 (bs, 1H, CONH pyrrole),
5.88 (bs, 1H, CONH phenylthio), 6.60 (d, 1H, J2,5=2.3
Hz, pyrrole C5-H), 7.22–7.38 (m, 10H, pyrrole C2-H and
benzene H); anal. C29H37N3O3S (507.69) C, H, N, S.
Microbiology
1
IR: n 3400, 3280 (NH and OH), 1620 (CO) cmꢀ1; H
NMR: d 1.18 (s, 9H, CCH3 pyrrole), 1.44 (s, 9H, CCH3
benzene), 3.32 (m, 2H, CH2NH), 3.99–4.03 (m, 3H,
CH2–pyrrole, and CH), 5.37 and 5.93 (2bs, 4H, CONH
and OH), 6.61–6.71 (m, 3H, pyrrole C5-H and anthra-
nilamide C3-H and C5-H), 7.24–7.38 (m, 8H, pyrrole
C2-H, anthranilamide C4-H and C6-H and benzene H);
anal. C29H38N4O3 (490.64) C, H, N.
Compounds. Test compounds were dissolved in DMSO
at an initial concentration of 200 mM and then were
serially diluted in culture medium.
Cells. MT-4 cells [grown in RPMI 1640 containing 10%
foetal calf serum (FCS), 100 UI/mL penicillin G and
100 mg/mL streptomycin] were used for cytotoxicity and

R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
2525
Appendix B. Analytical results of newly synthesized
derivatives 2f–d
anti-HIV-1assays. Cell cultures were checked periodi-
cally for the absence of mycoplasma contamination
with a MycoTect Kit (Gibco).
Compd
Elemental analyses calculated/found
Cytotoxicity and anti-HIVassays. Activity against the
HIV-1(HIV-,1 III
C
H
N
S
Cl
strain, obtaineded from super-
B
natants of persistently infected H9/IIIB cells.) multi-
plication in acutely infected cells was based on
inhibition of virus-induced cytopathogenicity in MT-4
cells. Briefly, 50 mL of RPMI 10% FCS containing
1ꢂ104 cells were added to each well of flat-bottomed
microtiter trays containing 50 mL of medium and serial
dilutions of test compounds. 20 mL of an HIV-1sus-
pension containing 100 CCID50 were then added. After
a 4 day incubation at 37 ꢁC, the number of viable cells
was determined by the 3-(4,5-dimethylthiazol-2-yl)-2,5-
diphenyl-tetrazolium bromide (MTT) method. Cyto-
toxicity of compounds, based on the viability of mock-
infected cells as monitored by the MTT method, was
evaluated in parallel with their antiviral activity.
2f
69.64
69.60
69.64
69.60
68.23
68.18
70.99 .42 71.811 —
71.03
68.23
6.64
6.67
6.64
6.58
6.20
6.31
11.07
11.03
11.07
11.05
6.63
—
—
—
—
—
—
—
—
—
—
—
—
2g
2h
2i
6.43
—
7.73
6.20
11.64
6.63
—
8.65
7.12
9.72
—
—
—
—
—
2j
68.2516.1 6.58
2k
2l
66.24
66.14
66.65
66.717.00
66.65
66.63
5.817.02
5.87
6.99
8.89
8.71
7.99
—
9.76
9.84
8.40
—
8.91
—
9.518.02
9.72
9.83
2m
2n
2o
2p
2q
6.99
6.88
6.08
—
—
—
Acknowledgements
69.27
3.67
69.216.07
66.216.00
66.01
67.47
67.23
3.69
3.09
8.41
7.07
Thanks are due to Ministero della Sanita, Istituto
Superiore di Sanita, Progetto AIDS 1999 (grant no.
40C.8) and to Italian MURST for partial support.
Thanks are due to Professor Chi-Huey Wong, The
Scripps Research Institute, La Jolla, CA, USA, for the
PR enzyme assays.
—
6.18
6.715.83
6.73
3.10
7.05
—
6.67
—
5.916.71
6.31
6.41
—
—
—
68.617.35
68.517.09
8.28
8.48
Appendix A. Analytical results of newly synthesized
derivatives 1a–f, 2a–e
Appendix C. Analytical results of newly synthesized
derivatives 5, 6 and 11–14
Compd
Elemental analyses calculated/found
Compd
Elemental analyses calculated/found
C
H
N
S
Cl
C
H
N
S
Cl
1a
1b
1c
ld
66.65 6.10
66.516.217 7.1
7.07
—
—
—
—
5
68.56
68.3516.1 5.83
6.16
5.71
—
4.99
—
—
—
68.07
68.18
6.90
6.85
9.92
9.87
—
—
—
—
6
59.90
59.97
70.83
70.816.40
62.85
62.83
61.53
61.50
61.53
61.50
5.39
—
12.63
—
5.314.87
6.32
5.25
5.27
5.33
6.71
6.68
6.71
6.81
—2.511
—
—
69.80
69.85
72.80161
7.95
—
12.18
—
—
11
12
13
14
5.16
—
—
61.32
61.28
5.38
5.29
6.50
6.66
—
—
8.23
8.35
4.58
4.72
7.17
7.15
7.17
7.20
—
—
—
—
—
—
11.60
11.51
—
1e
1f
62.95
62.816.25
6.16
9.18
—
—
7.85
7.74
9.31
—
—
—
64.98
64.85
7.07
7.21
11.23
1l.35
—
—
7.10
7.15
2a
2b
2c
2d
2e
72.99
73.03
6.92
6.90
7.40
7.37
—
—
—
—
72.99
72.87
6.92
6.95
7.45
7.43
—
—
References and Notes
1. Artico, M. Farmaco 1996, 51, 305.
2. Martin, J. A.; Redshaw, S.; Thomas, G. J. Prog. Med.
Chem. 1995, 32, 239.
3. Appelt, K. Perspect. Drug Discov. Des. 1993, 1, 23.
4. Roberts, N.; Martin, J.; Kinchington, D.; Broadhurst, A.;
Craig, J.; Duncan, I.; Galpin, S.; Handa, B.; Kay, J.; Krohn,
A.; Lambert, R.; Merrett, J.; Mills, J.; Parkes, K.; Redshaw,
72.99
72.85
6.92
6.917.40
7.40
8.65
—
8.71 —
66.24
66.35
5.817.02
5.80
7.99
8.89
8.02
6.99
9.76
7.00
8.93
8.91
66.24
66.34
5.817.02
5.90
8.89
9.84

2526
R. Di Santo et al. / Bioorg. Med. Chem. 10 (2002) 2511–2526
S.; Ritchie, A.; Taylor, D.; Thomas, G.; Machin, P. Science
1990, 248, 358.
Trippe, A. J.; Appelt II, K.; Davies, J. F.; Wu, B. W.; Musick,
L. J. Med. Chem. 1996, 39, 2781.
5. Kempf, D. J.; Marsh, K. C.; Denissen, J. F.; McDonald,
E.; Vasavanonda, S.; Flentge, C. A.; Green, B. E.; Fino, L.;
Park, C. H.; Kong, X. P.; Wideburg, N. E.; Saldivar, A.;
Ruitz, L.; Kati, W. M.; Sham, H. L.; Robins, T.; Stewat,
K. D.; Hsu, A.; Plattner, J. J.; Leonard, J. M.; Norbeck, D. W.
Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 2484.
6. Dorsey, B. D.; Levin, R. B.; McDaniel, S. L.; Vacca, J. P.;
Guare, J. P.; Darke, P. L.; Zugay, J. A.; Emini, E. A.; Schleif,
W. A.; Quintero, J. C.; Lin, J. H.; Chen, I.-W.; Holloway,
M. K.; Fitzgerald, P. M. D.; Axel, M. G.; Ostovic, D.;
Anderson, P. S.; Huff, J. R. J. Med. Chem. 1994, 37,
3443.
14. Reich, S. H.; Melnick, M.; II Davies, J. F.; Appelt, K.;
Lewis, K. K.; Fuhry, M. A.; Pino, M.; Trippe, A. J.; Nguyen,
D.; Dawson, H.; Wu, B.; Musick, L.; Kosa, M.; Kahil, D.;
Webber, S.; Gehlhaar, D. K.; Andrada, D.; Shetty, B. Proc.
Natl. Acad. Sci. U.S.A. 1995, 92, 3298.
15. Massa, S.; Di Santo, R.; Mai, A.; Botta, M.; Artico, M.;
Panico, S.; Simonetti, G. Il Farmaco 1990, 45, 833.
16. Van Leusen, A. M.; Siderius, H.; Hoogenboom, B. E.;
Van Leusen, D. Tetrahedron Lett. 1972, 52, 5337.
17. Ragno, R.; Marshall, G. R.; Di Santo, R.; Costi, R.;
Massa, S.; Rompei, R.; Artico, M. Bioorg. Med. Chem. 2000,
8, 1423.
7. Painter, G. R.; Ching, S.; Reynolds, D.; Clair, M.St.;
Sadler, B. M.; Elkins, M.; Blum, R.; Dornsife, R.; Livingston,
D. J.; Partaledis, J. A.; Pazhanisamy, S.; Tung, R. D.; Tisdale,
M. Drugs Future 1996, 21, 347.
1 8. Head, R. D.; Smythe, M. L.; Oprea, T. I.; Waller, C. L.;
Green, S. M.; Marshall, G. R. J. Am. Chem. Soc. 1996, 118, 3959.
19. Marshall, G. R.; Head, R. D.; Ragno R. Thermodinamics
in Biology. In Press.
8. Kaldor, S. W.; Kalish, V. J.; Davies, J. F.; Shetty, B. V.;
Fritz, J. E.; Appelt, K.; Burgess, J. A.; Campanale, K. M.;
Chirgadze, N. Y.; Clawson, D. K.; Dressman, B. A.; Hatch,
S. D.; Khalil, D. A.; Kosa, M. B.; Lubbehusen, P. P.; Mues-
ing, M. A.; Patick, A. K.; Reich, S. H.; Su, K. S.; Tatlock,
J. H. J. Med. Chem. 1997, 40, 3979.
9. Nugiel, D. A.; Jacobs, K.; Cornelius, L.; Chang, C. H.;
Jadhav, P. K.; Holler, E. R.; Klabe, R. M.; Bacheler, L. T.;
Cordova, B.; Garber, S.; Reid, C.; Logue, K. A.; Gorey-Feret,
L. J.; Lam, G. N.; Erickson-Viitanen, S.; Seitz, S. P. J. Med.
Chem. 1997, 40, 1465.
10. Boyer, F. E.; Vara Prasad, J. V.; Domagala, J. M.; Ells-
worth, E. L.; Gajda, C.; Hagen, S. E.; Markoski, L. J.; Tait,
B. D.; Lunney, E. A.; Palovsky, A.; Ferguson, D.; Graham,
N.; Holler, T.; Hupe, D.; Nouhan, C.; Tummino, P. J.; Uru-
mov, A.; Zeikus, E.; Zeikus, G.; Gracheck, S. J.; Sanders,
J. M.; VanderRoest, S.; Brodfuehrer, J.; Iyer, K.; Sinz, M.;
Gulnik, S. V. J. Med. Chem. 2000, 43, 843.
20. Ragno, R.; Head, R. D.; Marshall, G. R. 216th ACS
National Meeting 1998, COMP 032.
21. SYBYLTM (Molecular Modeling System); Tripos Associ-
ates: 1699 S Hanley Rd, St. Louis, MO 63144, USA.
22. Baroni, M.; Costantino, G.; Cruciani, G.; Riganelli, D.;
Valigi, R.; Clementi, S. Quant. Struct.-Act. Relat. 1993, 12, 9.
23. Cramer, R. D. I.; Patterson, D. E.; Bunce, J. D. J. Am.
Chem. Soc. 1988, 110, 5959.
24. After the co-crystal structure of AG1132, AG1157 and
1284 with HIV-PR were solved, an inversion of binding mode
was observed. Turning from secondary to tertiary amide the
carbinol configuration of Agouron’s compounds change from S
to R when bound to HIV-PR. For this reason we decided to
model all the compounds in both the configuration (see ref. 14).
25. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Lis-
kamp, R.; Lipton, M.; Caufield, C.; Chang, G.; Hendrickson,
T.; Still, W. C. J. Comput. Chem. 1990, 11, 440.
26. Lee, T.; Le, V.-D.; Lim, D.; Lin, Y.-C.; Morris, G. M.;
Wong, A. L.; Olson, A. J.; Elder, J. H.; Wong, C.-H. J. Am.
Chem. Soc. 1999, 121, 1145.
11. Navia, M. A.; Sato, V. L.; Tung, R. D. Int. Antivir. News
1995, 9, 143.
12. Melnick, M.; Reich, S. H.; Lewis, K. K.; Mitchell, L. J.Jr.;
Nguyen, D.; Trippe, A. J.; Dawson, H.; II Davies, J. F.;
Appelt, K.; Wu, B. W.; Musick, L.; Gehlhaar, D. K.; Webber,
S.; Shetty, B.; Kosa, M.; Kahil, D.; Andrada, D. J. Med.
Chem. 1996, 39, 2795.
27. Marcorin, G. L.; Da Ros, T.; Castellano, S.; Stefancich,
G.; Bonin, I.; Miertus, S.; Prato, M. Org. Lett. 2000, 2, 3955.
28. Friedman, S. H.; Ganapathi, P. S.; Rubin, Y.; Kenyon,
G. L. J. Med. Chem. 1998, 41, 2424.
29. Cramer, C. J.; Truhlar, D. G. J. Comput. Aided Mol. Des.
1992, 6, 629.
13. Reich, S. H.; Melnick, M.; Pino, M. J.; Fuhry, M. A.;