Synthesis, molecular modeling and QSAR studies in chiral 2,3-disubstituted-1,2,3,4-tetrahydro-9H-pyrido(3,4-b)indoles as potential modulators of opioid antinociception

Article abstract of DOI:10.1016/S0968-0896(01)00042-6

In view of coexistence of opioid and cholecystokinin (CCK) in the brain areas concerned with pain processing, some semirigid racemic and chiral anlogues of a potent CCK receptor antagonist (benzotript) have been synthesized and tested for their modulatory

Full text of DOI:10.1016/S0968-0896(01)00042-6

Bioorganic & Medicinal Chemistry 9 (2001) 1559–1570
Synthesis, Molecular Modeling and QSAR Studies in Chiral
2,3-disubstituted-1,2,3,4-tetrahydro-9H-pyrido(3,4-b)indoles
as Potential Modulators of Opioid Antinociceptiony
Anil K. Saxena,a,* Suresh K. Pandey,a Ravish C. Tripathia and Ram Raghubirb
aDivision of Medicinal Chemistry, Central Drug Research Institute, Lucknow 226001, India
bDivision of Pharmacology, Central Drug Research Institute, Lucknow 226001, India
Received 16 October 2000; accepted 31 January 2001
Abstract—In view of coexistence of opioid and cholecystokinin (CCK) in the brain areas concerned with pain processing, some
semirigid racemic and chiral analogues of a potent CCK receptor antagonist (benzotript) have been synthesized and tested for their
modulatory role on opioid antinociception, which may be mediated by CCK-B receptor. Some of these compounds, 3e, 3g, 3h, 4a,
4b and 4h, exhibited antinociceptive potentiation comparable to benzotript and proglumide. In order to identify the essential
chemical structural features important for this potentiation, molecular modeling and quantitative structure activity relationship
(QSAR) studies have been carried out in the S and R enantiomers of some of these semi-rigid compounds. The 3D-biophore
models, common to all molecules of the training set have been derived. These models with superimposition (match value >0.25)
depicted three biophoric sites one each for, p/hydrophobic interactions, hydrogen bonding and ionic interactions among the phenyl/
pyrrole ring, indole nitrogen, amidic oxygen, pyridyl nitrogen and lone pair of amidic oxygen. The total hydrophobicity and S
absolute stereochemistry are found to positively contribute to potentiation of antinociception induced by morphine and the
resulting quantitative pharmacophoric model with good correlation is found to well describe the observed activity. # 2001 Elsevier
Science Ltd. All rights reserved.
Introduction
development to morphine in rodents;3À6 however, their
clinical use in the pain management as adjuvants is lim-
ited due to their own adverse side effects.7
Despite several analgesics, the opioids still remain the
main drug of choice for the relief of the pain in humans
and morphine is still a drug of choice in such situations.
There is a high patient-to-patient variability in response
to opioids1 and complete pain control is only achieved
by increasing the dose, which causes side effects like
nausea, respiratory depression and mood disturbance.
This is further complicated by the other major draw-
back of tolerance development2 to the analgesic effect
(of opioid/morphine). Therefore, the efforts have been
to develop adjuvant drugs, which can potentiate the
analgesic effect of morphine/opioids and at the same
time counteract their abuse potential.
The 5-HT receptor antagonists have been shown to
inhibit the expression of tolerance,8À10 the inter-connec-
tion of opioid and serotonergic systems in the circuitry
underlying nociceptive neurotransmission in the
CNS.11À13 However, the precise role of the serotonergic
system in the tolerance process is still not well
explained.8
In view of the coexistence of opioid and cholecystokinin
(CCK) in the brain areas concerned with pain, proces-
sing indicates neuromodulatory role of CCK in opioid
antinociception. The observations that exogenous cho-
lecystokinin selectively antagonizes opiate analgesia and
endogenous cholecystokinin system opposes the action
of opiates, the CCK receptors have been implicated in
the attenuation of morphine-induced antinociception.14,15
CCK is a polypeptide consisting of 33-amino acid and is
distributed in various molecular forms in the central ner-
vous system (CNS) as well as in the peripheral tissues.
The biological actions of CCK are manifested by two
receptor subtypes, CCK-A (alimentary) and CCK-B
Several strategies to block the opioid tolerance include
NMDA antagonists and nitric oxide synthase inhibi-
tors, which have been shown to attenuate the tolerance
yCDRI Commun. No. 6054. Part of the work has been presented in
the Vth IUPAC International Symposium on Bio-organic Chemistry
(ISBOC 5), 30 January—4 February 2000 at NCL, Pune.
*Corresponding author. Tel.: +91-522-212411-18, extn. 4268; fax:
+91-522-223405/223938/229504; e-mail: anilsak@hotmail.com
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00042-6

1560
A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
(brain), both of which have been cloned and their
expression is now known.16À18 The CCK-A receptor is
predominantly found in peripheral tissues such as the
pancreas, colon, and gallbladder, although low levels of
this subtype are also found in discrete regions of the
CNS.19 The CCK-B receptor is primarily located in the
CNS and is indistinguishable from peripheral gastrin
receptors on the basis of both their binding properties20
and comparision of the amino acid sequences deduced
from cloned receptor cDNA of brain and gastric
mucosa.21 The coexistence and interaction of CCK-B
with other neurotransmitters, for example, dopamine,
GABA22,23 and neuropeptides, for example, neurotensin
and substance P,24 support involvement of CCK-B in
variety of neurological disorders such as anxiety, pain
and panic disorder.25,26
hydrogen bonds (presence), charge–transfer complexes
(HOMO, LUMO), hydrophobic interactions, van der
Waals (London) dispersion forces (p electron density on
atoms), as well as atomic contributions to hydro-
phobicity and molar refractivity. This program com-
pares the descriptors and their distances with respect to
active and inactive analogues and stores the results as
rules in a knowledge base, which can be used to predict
the activities of novel compounds.
So the 3D-QSAR models based on common biophores
have been derived on total set of 13 compounds (11
chiral title esters, benzotript and proglumide) and the
models have been used to compare the observed activity
of the partially racemized acids with the predicted
activity of the corresponding chiral acids and results
produced are reported in this paper.
CCK receptor antagonists could be broadly classified
into the four classes: (1) derivatives of cyclic nucleo-
tides, (2) partial sequences of the C-terminal region of
CCK, (3) derivatives of amino acids, for example, pro-
glumide, benzotript27 and its analogues, (4) benzodiaze-
pine derivatives. As benzotript is a flexible molecule, it
may have many possible conformations. Out of them,
only certain discrete conformations may be responsible
for the specific biological activity. In view of this, some
semirigid racemic analogues of benzotript were synthe-
sized which showed potent anti-ulcer activity compar-
able with benzotript in our earlier work.28
Chemistry
The key intermediates S and R methyl 1,2,3,4-tetra-
hydro-9H-pyrido(3,4-b)indole-3-carboxylate (2a and
2b) were synthesized essentially according to literature
method36 starting from S and R tryptophan methyl
ester hydrochloride (1a and 1b) by Pictet Spengler
cyclyzation of the corresponding imines obtained by
condensation with formaldehyde, followed by the basi-
fication of the ester hydrochloride. These esters (2a and
2b) on condensation with different acid chlorides yielded
the corresponding amides 3a–3l (Scheme 1, Table 1).
Selective hydrolysis of 3a–3k with NaOH afforded the
partially racemic title acids 4a–4k (Scheme 1, Table 1).
These amidic esters 3a–3l were chirally pure and their
enantiomeric excess (ee) was >95% as was determined
on chiral stationary phase column, chiraDex1 by ana-
lytical HPLC, LaChrom of Merck. The compounds 3a–
3l had reasonable value of optical rotation almost of
similar magnitude for both C-3 (S) and C-3 (R) enan-
tiomers. However, in general, 1H NMR spectrum of
these amidic esters 3a–3l showed two sets of signal for
each proton in variable ratios depending on the steric
bulk of the amidic group possibly due to the restricted
rotation around an amide N–CObond resulting in the
existence of two conformers in solution state at room
temperature. In order to confirm this, the 1H and
13C NMR studies were carried out on relatively
simple compound S-(À)-methyl 2-acetyl-1,2,3,4-tetra-
hydro-9H-pyrido(3,4-b)indole-3-carboxylate (3l), which
was prepared by the acetylation of the C-3 (S) ester 2a
with acetic anhydride.
Some of the analogues, which did not show significant
anti-ulcer activity, showed potentiation of morphine
antinociception, which may be mediated by CCK-B
receptor. Thus, there was a need to identify the essential
chemical structural features important for this activity
for which the S and R enantiomers of some of these
semi-rigid compounds were synthesized, evaluated and
3D-QSAR study undertaken. Such QSAR’s are a useful
tool in the medicinal chemistry in defining the forces
governing the pharmacological activity of a particular
class of compounds. Particularly in cases where the
crystallographic structure of the pharmacological target
is unknown, it is possible to use molecular modeling
techniques to construct models of the receptor sites or a
graphical pharmacophore29 which may be used to
improve description and prediction of activity. Among
such approaches, hydrophobic interactions between
molecules (HINT),30 distance geometry,30,31 receptor
modeling based on the three-dimensional structure and
physicochemical properties of the ligand molecules
(REMOTEDISC),32,33 hypothetical active site lattice
(HASL),30,34 comparative molecular field analysis
(CoMFA)30 and the Apex-3D expert system based on
the logicostructural35 program are used. Apex-3D iden-
tifies biophores, which represents a certain structural
and electronic pattern in a bioactive molecule, which is
responsible for the manifestation of activity possible due
to receptor interaction. Descriptor centers of biophores
can be atoms, pseudo atoms, ring centers, hydrophobic
regions and hydrogen bonding sites that can participate
in ligand–receptor interactions. These descriptor centers
are based on certain physical properties such as electro-
static interaction, charges, electron acceptor or donor,

A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
1561
Scheme 1. (i) (a) HCHO, MeOH, H2O, Pictet–Spengler (PS) cyclization reaction; (b) NaHCO3 saturated solution; (ii) RCOCl, Et3N, THF, 30 ꢀC–
reflux; (iii) (a) 2N aq NaOH; (b) HCl.
Table 1. Characterization data of the title compoundsa
Compound no.
R
Configuration
at C-3
mp (ꢀC)
Yield (%)
[a]2D0
Concentration
(g/mL)b
Molecular
Formula
3a
3b
3c
3d
3e
3f
3g
3h
3i
4-Cl–C6H4
2-Quinolyl
Cyclohex-1-yl
2-COCH3–C6H4
3-F-C6H4
S
S
S
S
S
S
S
S
S
R
R
S
S
S
S
S
S
S
245
125
149
158
120
170
102
159
164
134
235
167
222
215
218
250
230
155
233
182
228
227
231
222
254
233
246
242
152
52
60
73
35
71
67
63
43
58
42
36
85
54
50
77
40
51
53
54
62
78
63
76
80
52
80
51
54
41
À8.28
À21.74
+84.39
À38.74
+6.72
+43.75
+62.90
+11.32
À10.52
+20.16
+7.54
+143.68
À1.91
À2.22
+3.22
À2.56
À0.90
À2.23
À2.94
À1.49
À5.93
+2.96
+2.96
+6.48
NAc
0.22
0.22
0.19
0.28
0.20
0.19
0.23
0.10
0.11
0.12
0.11
0.21
0.22
0.23
0.21
0.21
0.22
0.27
0.24
0.20
0.12
0.12
0.11
0.20
NA
NA
NA
NA
NA
C20H17O3N2Cl
C23H19O3N3
C20H24O3N2
C22H20O4N2
C20H17O3N2F
C21H19O3N2F
C21H20O4N2
C25H24O4N2
C20H17O3N2Cl
C23H19O3N3
C20H17O3N2Cl
C15H16O3N2
C19H15O3N2Cl
C22H17O3N2
C19H22O3N
C21H18O4N2
C19H15O3N2F
C20H17O3N2F
C20H18O4N2
C24H22O4N2
C19H15O3N2Cl
C22H17O3N3
C19H15O3N2Cl
C14H14N2O3
C19H15O3N2Cl
C22H17O3N3
C19H15O3N2
C21H20O3N2
C21H20O5N2
4-F-CH2-C6H4
OCH2–C6H5
2-OCH3-(CH¼CH)2–C6H4
3-Cl–C6H4
3j
3k
3l
2-Quinolyl
4-Cl–C6H4
CH3
4-Cl–C6H4
2-Quinolyl
Cyclohex-1-yl
2-COCH3–C6H4
3-F-C6H4
4a
4b
4c
4d
4e
4f
4g
4h
4i
4j
4k
4l
5a
5b
5c
5d
5e
4-F-CH2–C6H4
OCH2–C6H5
2-OCH3–(CH¼CH)2–C6H4
3-Cl–C6H4
S
S
S
2-Quinolyl
4-Cl–C6H4
CH3
4-Cl-C6H4
2-Quinolyl
3-F-C6H4
R
R
S
RS
RS
RS
RS
RS
NA
NA
NA
NA
3,5-(CH3)2C6H3
3,5-(OCH3)2C6H3
aSynthesis of racemic acids 5a–5e was done as per ref 28.
bOptical rotations of the esters 3a–3l were taken in MeOH while of the acids 4a–4l were taken in DMF.
cNA, not applicable.
The compound S-(À)-methyl 2-acetyl-1,2,3,4-tetra-
hydro-9H-pyrido(3,4-b)indole-3-carboxylate (3l) existed
as two conformers, one major and the other minor. H
spectrum of 3l showed two sets of signal for each proton
(Table 2), namely 8.63 (ex, brs, 1H, NH, min) and 8.47
(ex, brs, 1H, NH, maj), 7.51 (t, J=4.8 Hz, 1H, Ar-5-H),
7.26–7.32 (m, 1H, Ar-8-H), 7.09–7.20 (m, 2H, Ar-6 and
7-H) 5.95 (d, J=6.2 Hz, 1H, CHCO2CH3, maj) and 4.94
(d, J=5.7 Hz, 1H, CHCO2CH3, min), 4.39 and 5.21 (d,
J=16.8 and 17.1 Hz, respectively, 1H, CH2N, min), 4.73
and 4.84 (d, J=15.3 Hz for both, 1H, CH2N, maj), 3.62
(s, 3/2 H, OCH3, min), 3.59 (s, 3/2H, OCH3, maj), 3.06
and 3.47 (dd and d, J=6.3, 15.9 and 15.6 Hz, respec-
tively, 1H, CH2CHCO2CH3, maj),3.15 and 3.54 (dd and
1
d,
J=5.76
and
15.6 Hz,
respectively,
CH2CHCO2CH3, min), 2.27 (s, 3/2 H, COCH3, maj)
1H,

1562
A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
Table 2. 1H NMR spectrum of S-(À)-methyl 2-acetyl-1,2,3,4-tetrahydro-9H-pyrido(3,4-b)indole-3-carboxylate 3l (exists in two conformations,
major and minor, depending on the bulk of R group)
Comformer
Major
Minor
1H
3H
4H
5H
6 and 7H
8H
9H
11H
13H
4.84 and 4.73 (d, J=15.3 for both, 1H)
5.95 (d, J=6.2, 1H)
3.47 and 3.06 (d and dd, J=15.6 and 6.3, 15.9, respectively, 1H) 3.54 and 3.15 (d and dd, J=15.6 and 5.76, 15.6, respectively, 1H)
5.21 and 4.39 (d, J=17.1 and 16.8, respectively, 1H)
4.94 (d, J=5.7, 1H)
7.51 (t, J=4.8, 1H, indistinguishable)
7.09–7.20 (m, 2H, indistinguishable)
7.26–7.32 (m, 1H, indistinguishable)
8.47 (ex, brs, 1H)
2.27 (s, 3H)
3.59 (s, 3H)
8.63 (ex, brs, 1H)
2.26 (s, 3H)
3.62 (s, 3H)
Figure 1. 1H spectrum of 3l in CDCl3 at different temperatures (298, 308, 318, 328 and 338 K). At 338 K, the two set of signals of COCH3 and OCH3
merged together due to free rotation of these groups.

A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
1563
and 2.26 (s, 3/2 H, COCH3, min) and positions were
ascertained by 1H correlation spectroscopy (COSY)
spectrum. The presence of two conformers was further
reactivity (Don_01)) of site B corresponding to the
amidic oxygen in the compounds (3a–k), benzotript and
proglumide are 0.986ꢁ0.027, À0.362ꢁ0.029 and
8.370ꢁ0.115 A, respectively. The spatial disposition
with respect to each in terms of distances between three
biophoric centers A, B and C for the compounds is
described as the mean values between A and B, B and C
and A and C which are 7.583ꢁ0.217, 3.000ꢁ0.001 and
9.697ꢁ0.465 A, respectively. The similarity of these
results with earlier reported molecular modeling results
in case of CCK-B antagonist37 points that anti-noci-
ceptive action of title molecule is mainly due to their
CCK-B antagonism and the model selected conform to
the earlier suggested model of binding CCK-B antago-
nist to the receptor where hydrogen bond is formed
between the carbonyl oxygen of the ligand and properly
oriented receptor donor function and possibly hydro-
phobic pocket exist in the receptor cavity which offers
additional sites for anchoring. The best quantitative
structure–activity relationship derived for this model
(model no. 1; Table 5) is shown by the two-parameter
equation (eq. 1):
1
confirmed by recording the H spectra at different tem-
peratures (at 298, 308, 318, 328 and 338 K) (Fig. 1). At
normal temperature, the energy difference between two
conformers is too large; therefore, in the NMR time
scale both sets were observed separately. At higher
temperature, the two signals of each COCH3 and OCH3
merged together due to free motion, but the protons of
the ring were observed separately due to restricted
rotation of the ring between two forms. The sets of
proton of two conformers were also differentiated with
the help of rotatory Overhauser effect spectroscopy
(ROESY) spectrum. In ROESY spectra, COCH3 of one
conformer (minor) showed a cross-peak with H-3 pro-
ton, while in the other conformer (major), H-1 showed a
cross-peak with H-5 phenyl proton. Exactly similar
observations were there in 13C NMR spectra and com-
1
plete assignment of chemical shift values for H NMR
of the representative title acid and ester compounds are
given in the Experimental.
Log ðC Â 100Þ ¼ 0:255 ðtotal hyd:Þ À 0:628 ðIÞ þ 0:452
n ¼ 13; R ¼ 0:880; RMSA ¼ 0:140; F2;10 ¼ 17:224;
Results and Discussion
In initial studies, it was observed that the racemic semi-
rigid cyclic analogue (5a) of benzotript was almost
equiactive to benzotript in the modulation of anti-
nociceptive potentiation similar to the case of CCK
mediated anti-ulcer activity.28 Among its two partially
racemized (S and R) enantiomers, the enantiomer with
predominant S chiral center (4a) was found to be more
active and the enantiomer with predominant R chiral
center (4k) was found to be less active than the racemic
one (5a). Hence, the order of activity among the three
semi-rigid analogues is partially racemized S (4a)>RS
(5a)>partially racemized R (4k). Similar results were
also observed with its racemic quinolylcarboxamide (5b)
and its partially racemic S (4b) and R (4k) analogues
where both 4b and 5b were almost equiactive to benzo-
tript. In view of it, only compounds corresponding to S
enantiomer were synthesized and studied for the mod-
ulation of anti-nociception. Among these, the com-
pounds 3e, 3g, 3h and 4h exhibited anti-nociceptive
potentiation comparable to the benzotript and proglu-
mide in addition to the above discussed compound and
the overall order of antinociceptive modulation was 3h
> 4h > 3e=4b > 3g=4a.
F2;10ꢀ2 :0:002 ¼ 14:9
ð1Þ
The equation has reasonably good correlation coeffi-
cient value (R=0.880) of high statistical significance
>99.8% (F2,10=17.224; F2,10a2:0.002=14.9). The equa-
tion is also of good predictive value (R2 cross-vali-
dated=0.730). The positive regression coefficient with
total hydrophobicity points towards the positive con-
tribution and importance of this parameter for the
potentiation of antinociception (C) in these compounds.
The negative contribution of the indicator variable (I)
which has value of 0 and 1 for S and R stereochemistry
respectively, at C-3 center simply quantifies the positive
contribution of S stereochemistry at this center for
activity. This model (eq 1), describes well the experi-
mentally observed activity data and the calculated as
well as predicted (leave one out) activity data (Table 6).
A careful comparison of this data indicates that pre-
dicted and calculated error values 0.32 and 0.25 were
the highest for proglumide in this series of compounds.
This may possibly be due to structural dissimilarity
between proglumide and rest of the molecules. The same
may also reason for the poor superimposition, that is,
low match value (0.26). Hence, attention was focused to
the biophore models with 12 compounds (proglumide
excluded). Among several of these models, five models
(model nos. 2–6, Fig. 3) with match value >0.35, which
had three biophoric sites, were considered. As expected
this set of 12 compounds showed significant improve-
ment in the superimposition as indicated by higher
match values than that of model no. 1 because of the
exclusion of most dissimilar structure proglumide. The
model no. 2 had two biophoric sites amidic oxygen and
its lone pair similar to model no. 1, however it took
indole nitrogen atom as third biophoric site as com-
pared to the phenyl ring center of indole in model no.1.
In model no. 3, all the three biophore sites, namely
The preliminary molecular modeling studies for all the
13 compounds (3a–3k, benzotript (S configuration) and
proglumide (S configuration)) resulted in the develop-
ment of biophore models with poor superimposition
and only one model (model no. 1) showed match value
>0.25 with three biophore sites common to all mol-
ecules (Table 4, Fig. 2); site A is the phenyl ring of
indole part in terms of its 6p electrons presence which
possibly involved in p–p interactions, site B being the
amidic oxygen in terms of p population charge and the
electron donating index for electrostatic as well as for
ionic binding and its lone pair as third site C for
hydrogen bonding. The mean value of the three
properties (p-population, charge and electron donor

1564
A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
phenyl ring center, pyrrole ring center and nitrogen of
the pyrrole ring exists in indole nucleus near to each
other. The two-biophore sites in model nos. 4 and 5
were the same in all respect, that is, in terms of ring
center of phenyl ring and nitrogen of the pyridine ring
except in model no. 5 which had third biophore site as
pyrrole ring while in model no. 4 it was nitrogen atom
of the indole ring. However, model no. 4 has been better
in terms of match value as compared to model no. 5.
Model no. 6 was not found appropriate because benzo-
tript, which is a flexible molecule, so in its alignment
with other molecules it used its benzene ring center at a
particular biophoric site while the rest of the molecules
used their amidic oxygen; however, the rest of the bio-
phoric sites in the model were the same for the total set
of molecules (Fig. 2). These biophore models (nos. 2–6)
were also used to derive quantitative structure activity
relationships similar to the one corresponding biophore
model 1 (eq (1)) and surprisingly the best results corre-
sponded to eq (1) and were similar for all these biophore
models (eq. (2)) because it did not contain any 3D or
secondary site parameters.
the phenyl ring of indole part in terms of its 6p electrons
presence which are possibly involved in p–p interac-
tions, the second site B being the amidic oxygen in terms
of p population, charge and the electron of its lone pair
for electrostatic as well as ionic density index and the
third site C for the hydrogen bonding. All these sites are
important and are involved in the potentiation of anti-
nociceptive activity possibly through CCK–B receptor
as similar sites, namely a hydrogen bonding site between
the carbonyl oxygen of the ligand and properly oriented
receptor donor function and in p-rich hydrophobic site
between an aromatic p region (indole/phenyl ring) of
the ligand in the receptor cavity for hydrophobic inter-
actions have also been suggested for the binding of
CCK-B antagonist with the CCK-B receptor. Further,
the derived QSAR models not only explain the data for
the chiral ester molecules (3a–3k; training set com-
pounds, Table 6) used in deriving these models but also
explains the observed antinociceptive activity of the
partially racemic acids (4a–4k; test set compounds,
Table 7) and thus can advantageously be explored in the
designing of new leads.
Log ðC Â 100Þ ¼ 0:333 ðtotal hyd:Þ À 0:617 ðIÞ þ 0:202
n ¼ 12; R ¼ 0:927; RMSA ¼ 0:113; F2;9 ¼ 27:636;
Experimental
All chemicals used were of reagent grade unless stated
otherwise. Solvents and substituted benzoyl chlorides
were distilled prior to use according to standard meth-
ods. In high performance liquid chromatography
(HPLC) run solvents were used directly from HPLC
grade OmniSolv1 sure seal bottles. Melting points were
determined in open capillaries on an electrically heated
melting point apparatus and are otherwise uncorrected.
Thin layer chromatography (TLC) was run on self-
made plates of silica gel G (Acme Synthetic Chemicals,
India) or 0.25 mm ready made plates of silica gel 60F254
(Kieselgel 60F254, E. Merck, Darmstadt, Germany) and
detection was done by iodine vapors/spraying with
Dragon droff spray reagent or by UV radiation. Col-
umn chromatography was performed with silica gel
(Acme Synthetic Chemicals, India; 60–120 mesh). Pro-
ton nuclear magnetic resonance spectra (1H NMR) were
recorded on Bruker AVANCE DPX 200 MHZ or DRX
300 MHZ or 400 MHz FT NMR spectrometer at room
temperature using CDCl3 or DMSO-d6 as the solvent.
Proton chemical shifts (d) are reported in parts per mil-
lion (ppm) relative to tetra methyl silane (TMS,
0.00 ppm). Coupling constants (J) are reported in hertz
(Hz), and s, d, t, q, m, br, and ex refer to singlet, doub-
let, triplet, quartet, multiplet, broad, and exchangeable,
respectively. Maj and min refers to major and minor
conformations, respectively. Infra-red spectra (IR) were
recorded on Perkin-Elmer infrared models 557, 881, or
Shimadzu FTIR model PC spectrophotometer (nmax in
cmÀ1). Electron impact mass spectra (EIMS) were
recorded on Jeol-JMSD-300 spectrometer. Micro-
analysis (C, H, N) were performed on a Carlo erba
instruments CHNS-OEA1108-Elemental Analyzer and
results were within 0.4% of the theoretical values unless
stated otherwise. Optical rotations were determined on
Autopol III Automatic polarimeter using the sodium D-
line (c in g/100 mL). NMR, IR, EIMS, microanalyses
F2;9 a2:0:001 ¼ 19:9
ð2Þ
The comparison of the regression coefficient of both the
parameter total hydrophobicity and indicator variable
showed similarity both on terms of magnitude and sign.
This stability of regression coefficients and the
improvement in correlation coefficient value (R=0.927)
and statistical significance >99.9% (F2,9=27.636;
F2,9a:0.001=19.9) and R2 cross-validated=0.829 indi-
cated it to be a robust model. In order to check the
validity and robustness of these models, they were used
to predict the activity of partially racemic acids (Table
7), which were not used in deriving these models
because of the non-availability of the chiral acids for
evaluation. There has been a good correlation between
the observed activity of the partially racemized (S and
R) and racemic acids and the predicted activity for
chiral S and R acids (Table 7) and in general the order
of activity was chiral S > partially racemized S >
racemic > partially racemized R > chiral R. Such an
agreement between the observed and predicted activities
of the partially racemized (S and R) and racemic acids
clearly points towards the predictive comparability of
these models, which may be useful for designing mod-
ulators which can potentiate anti-nociceptive activity.
In summary, the semi-rigid chiral analogues of benzo-
tript have shown potent potentiation of anti-nociception
comparable to benzotript suggesting that the conforma-
tions corresponding to the semi-rigid analogue have
major contribution for the activity. The biophore mod-
els with three sites corresponding to ring centers of the
phenyl and pyrrole ring, nitrogen atom of the pyrrole
and pyridine, oxygen atom of the amide or ester and
their hydrogen bonding sites (p-electrons) indicate that
the three sites A, B and C; the first is A represented by

A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
1565
and optical rotations were performed at the Regional
Sophisticated Instrument Center (RSIC) and chiral
HPLC was performed on an analytical Lachrom Merck,
Hitachi system at the Medicinal Chemistry Division,
Central Drug Research Institute, Lucknow, India. The
column used was LiChroCART1 250-4 ChiraDex1
(244ꢂ4 mm) of Merck, with an L-7455 Diode Array
Detector set for 200–400 nm range, using MeOH/H2O
(30/70) solvent system with flow rate set at 0.8 mL/min
having a pressure of 174 psi at 20 ꢀC.
and 3.47 (dd and d, J=6.3, 15.6 and 15.9, 1H,
CH2CHCO2CH3), 3.15 and 3.54 (dd and d, J=5.76 and
15.6, 1H, CH2CHCO2CH3), 3.59 and 3.62 (s, 3H,
OCH3), 4.73 and 4.84 (d, J=15.3 for both, 1H, CH2N),
4.39 and 5.21 (d, J=16.8 and 17.1, 1H, CH2N), 4.94
and 5.95 (d, J=5.7 and 6.2, 1H, CHCO2CH3), 7.09–
7.20 (m, 2H, Ar-6 and 7-H), 7.26–7.32 (m, 1H, Ar-8-H),
7.51 and 8.63 (t, J=4.8, 1H, Ar-5-H), 8.47 (ex, brs, 1H,
NH); 13C NMR (CDCl3, 300 MHz) d 22.0 and 22.1 (C-
11), 22.9 and 23.8 (C-4), 39.1 and 42.2 (C-1), 50.5 and
55.9 (C-3), 52.4 and 52.7 (C-13), 104.8 and 106.6 (C-4a),
110.9 and 111.0 (C-8), 117.8 and 118.1 (C-5), 119.4 and
119.6 (C-6), 121.8 and 122.0 (C-7), 126.4 and 126.5 (C-
4b), 128.4 and 129.4 (C-9a),136.4 and 136.5 (C-8a),
170.9 and 171.3 (C-10), 171.5 and 171.6 (C-12); FTIR
(KBr, cmÀ1): 748, 1020, 1188, 1234, 1327, 1435, 1645,
1738, 3217, 3393; MS (EI): m/z 272 (M+). Anal. calcd
for C15H16N2O3: C 73.75; H 6.60; N 11.47%. Found: C
73.60; H 6.50; N 11.35%.
Synthesis
pyrido(3,4-b)indole-3-carboxylate (2a)
of
S-(À)-methyl-1,2,3,4-tetrahydro-9H-
Formaldehyde (38 wt% solution in water, 11 mL) was
added to a solution of l-tryptophan methyl ester
hydrochloride 1a (22 g, 0.86 mol) in aq methanol
(154 mL; ratio 10:1) during 30 min. The reaction
mixture was stirred for 4 h at room temperature, con-
centrated and cooled to give S-(À)-methyl-1,2,3,4-tetra-
hydro-9H-pyrido(3,4-b)indole-3-carboxylate hydrochlo-
ride which was basified with sodium bicarbonate to give
2a. Yield 11.4 g (57%): M.P. 165 ꢀC; [a]D20 À71.00ꢀ (c
Synthesis of S-(À)-Methyl-2-(4-chlorobenzoyl)-1,2,3,4-
tetrahydro-9H-pyrido(3,4-b)indole-3-carboxylate
R=C6H4-4-Cl)
(3a,
1
0.6, MeOH); H NMR (200 MHz, CDCl3): d 2.83–2.96
(m, 1H, CH2), 3.16 (dd, J=4.64 and 15.28, 1H, CH2),
3.80 (s, 4H, OCH3 and CH), 4.13 (bs, 2H, NCH2), 7.06–
7.19 (m, 2H, ArH), 7.30 (d, J=7.10, 1H, ArH), 7.47 (d,
J=8.18, 1H, ArH), 7.83 (brs, 1H, indole NH); FTIR
(KBr, cmÀ1): 738, 820, 866, 1004, 1056, 1128, 1190,
1234, 1302, 1346, 1442, 1504, 1590, 1626, 1740, 1916,
2312, 2752, 2938, 3058, 3150, 3324; MS (EI): m/z 230
(M+). Anal. calcd for C13H14O2O2: C 67.81; H 6.13; N
12.17%. Found: C 67.85; H 6.14; N 12.26%.
A mixture of S-(À)-methyl 1,2,3,4-tetrahydro-9H-pyr-
ido(3,4-b)indole-3-carboxylate 2a (2.30 g, 0.01 mol),
4-chlorobenzoyl chloride (1.75 g, 0.01 mol) and triethyl-
amine (1.60 mL) in dry THF (50 mL) was refluxed at
90 ꢀC for 8 h. It was concentrated and diluted with H2O
(30 mL). The solid separated was filtered and dried. It
was purified by column chromatography on silica gel
column using 2% methanol in chloroform as an eluant.
Yield 1.9 g (52%): mp 245 ꢀC; [a]D20 À8.25ꢀ (c 0.21,
1
MeOH); H NMR (400 MHz, CDCl3): d 3.10–3.25 (m,
Compound R-(+)-methyl 1,2,3,4-tetrahydro-9H-pyrido
(3,4-b) indole-3-carboxylate 2b was synthesized by the
similar method starting from d-tryptophan methyl ester
hydrochloride 1b. Yield 58%: mp 150 ꢀC; [a]D20 +62.84ꢀ
(c 0.5, MeOH); 1H NMR (200 MHz, CDCl3): d 1.75 (bs,
1H, pyrido NH), 2.83–2.96 (m, 1H, CH2), 3.15 (dd,
J=4.58 and 15.38, 1H, CH2), 3.80 (brs, 4H, OCH3 and
CH), 4.12 (brs, 2H, NCH2), 7.06–7.19 (m, 2H, ArH),
7.30 (d, J=7.04, 1H, ArH), 7.48 (d, J=7.80, 1H, ArH),
7.85 (brs, 1H, indole NH); FTIR (KBr, cmÀ1): 724, 802,
1006, 1176, 1256, 1336, 1442, 1738, 1880, 2340, 2848,
2935, 3198, 3726, 3826; MS (EI): m/z 230 (M+). Anal.
calcd for C13H14N2O2: C 67.81; H 6.13; N 12.17%.
Found: C 67.86; H 6.15; N 12.19%.
1H, CH2), 3.40–3.55 (m, 1H, CH2), 3.65 and 3.70 (s, 3H,
OCH3), 4.55 (d, J=16.67, 1H, NCH2), 4.90 and 5.35
(m, 1H, NCH2), 4.85 and 6.00 (d, J=6.67, 1H, CH),
7.00–7.15 (m, 2H, ArH), 7.25–7.55 (m, 6H, ArH), 9.82
and 10.12 (brs, 1H, indole NH); FTIR (KBr, cmÀ1):
820, 980, 1080, 1160, 1200, 1280, 1400, 1600, 1720,
2900, 2920, 3060, 3160; MS (EI): m/z 368 (M+). Anal.
calcd for C20H17O3N2Cl: C 65.13; H 4.65; N 7.60%.
Found: C 65.20; H 4.78; N 7.72%.
Compound R-(+)-methyl 2-(4-chlorobenzoyl)-1,2,3,4-
tetrahydro-9H-pyrido(3,4-b)indole-3-carboxylate
3k,
R=C6H4-4-Cl was synthesized by the similar method
starting from 2b. Yield 36%: mp 265oC; [a]D20 +9.54ꢀ (c
0.22, MeOH); 1H NMR (400 MHz, CDCl3): d 3.10–3.25
(m, 1H, CH2), 3.40–3.55 (m, 1H, CH2), 3.60 and 3.70 (s,
3H, OCH3), 4.55 (d, J=16.67, 1H, NCH2), 4.80 and
5.35 (d, J=16.67, 1H, NCH2), 5.95 and 4.85 (d,
J=6.67, 1H, CH), 7.00–7.15 (m, 2H, ArH), 7.25–7.55
(m, 6H, ArH), 10.05 and 10.35 (brs, 1H, indole NH);
FTIR (KBr, cmÀ1): 633, 723, 756, 843, 1015, 1092, 1142,
1173, 1198, 1232, 1421, 1523, 1597, 1632, 1740, 2513,
2750, 2866, 2914, 2955, 3111, 3196; MS (EI): m/z 368
(M+). Anal. calcd for C20H17O3N2Cl: C 65.13; H 4.65;
N 7.60%. Found: C 65.22; H 4.79; N 7.75%.
Synthesis of S-(À)-methyl 2-acetyl-1,2,3,4-tetrahydro-
9H-pyrido(3,4-b) indole-3-carboxylate (3l, R=CH3)
Acetic anhydride (1.10 mL, 0.011 mol), dry triethyl-
amine (1.60 mL, 0.011 mol) were added to a solution of
S-(À)-methyl-1,2,3,4-tetrahydro-9H-pyrido(3,4-b)indol-
e-3-carboxylate 2a (2.30 g, 0.01 mol) in dry tetra-
hydrofuran (THF) (50 mL) and the resulting solution
was stirred at room temperature for 4 h. It was then
concentrated under reduced pressure and triturated with
water to yield crude product (3l), which was recrys-
tallized with methanol. Yield=2.3 g (85%): mp 167–
168 ꢀC; [a]2D0 +143.64ꢀ (c 0.22, MeOH); 1H NMR
(300 MHz, CDCl3) d 2.26 and 2.27 (s, 3H, COCH3) 3.06
Other compounds (3b–3j) were similarly synthesized
and their physicochemical data are given in Table
1.

1566
A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
Synthesis of S-(À)-2-(4-chlorobenzoyl)-1,2,3,4-tetrahydro-
9H-pyrido(3,4-b) indole-3-carboxylic acid (4a, R=C6H4-
4-Cl)
NMR (400 MHz, DMSO-d6): d 2.25–2.35 (m, 1H, CH2),
2.50–2.65 (m, 1H, CH2), 3.70 (t, J=13.33, 1H, NCH2),
3.90 and 4.35 (d, J=13.33, 1H, NCH2), 3.95 and 5.00
(d, J=6.67, 1H, CH), 6.20–6.35 (m, 2H, ArH), 6.45–
6.90 (m, 6H, ArH), 9.90 and 10.20 (brs, 1H, CO2H);
FTIR (KBr, cmÀ1): 615, 679, 719, 752, 837, 889, 968,
1007, 1092, 1140, 1173, 1209, 1242, 1319, 1408, 1448,
1522, 1597, 1628, 1713, 2561, 2862, 3329; MS (EI): m/z
355(M+). Anal. calcd for C19H15O3N2Cl: C 64.32; H
4.26; N 7.98%. Found: C 64.28; H 4.22; N 7.86%.
2 N NaOH (0.6 mL) was added to a stirred solution of
S-(À)-methyl 2-(4-chlorobenzoyl)-1,2,3,4-tetrahydro-9H-
pyrido(3,4-b)indole-3-carboxylate 3a (0.368 g, 0.001 mol)
in methanol or dioxane (10 mL) and the stirring was
continued at 25 ꢀC overnight. Reaction mixture was
acidified with acetic acid and solution was concentrated
to half of its volume and then it was diluted with water
(50 mL). The separated solid was filtered, washed with
water and dried, purified by passing through silica gel
column using 6% methanol in chloroform as eluant.
Yield 0.19 g (54%): mp 222 ꢀC; [a]D20 À0.91ꢀ (c 0.2,
DMF); 1H NMR (400 MHz, DMSO-d6): d 2.25–2.35
(m, 1H, CH2), 2.50–2.65 (m, 1H, CH2), 3.70 (t,
J=13.33, 1H, NCH2), 3.90 and 4.35 (d, J=13.33, 1H,
NCH2), 3.95 and 5.00 (d, J=6.67, 1H, CH), 6.20–6.35
(m, 2H, ArH), 6.45–6.90 (m, 6H, ArH), 9.90 and 10.20
(brs, 1H, CO2H); FTIR (KBr, cmÀ1): 642, 682, 722, 754,
838, 888, 908, 968, 1008, 1092, 1136, 1174, 1212, 1246,
1325, 1404, 1448, 1596, 1628, 1714, 2342, 2866, 2928,
3062, 3336; MS (EI): m/z 354 (M+). Anal. calcd for
C19H15O3N2Cl: C 64.32; H 4.26; N 7.90. Found C
64.22; H 4.20; N 7.96.
Other compounds (4l, 4b–4j) were synthesized by
adopting above procedure and their characterization
data is given in Table 1.
Assessment of antinociceptive activity
The antinociceptive activity of morphine plus test com-
pound and morphine alone was determined by hot plate
test.38 Mice of either sex (18–23 g) bred at the National
Laboratory Animal Center of the Institute were used in
the group of 10 mice each. The nociceptive response
(reaction time, RT) of each animal was monitored on
electrically heated and thermostatically controlled hot
plate, maintained at 56ꢁ0.5 ꢀC. Two pre-drug (base
line) reaction times were observed 15 min apart, and the
control reaction times varied from 2.5 to 3.5 s. First,
graded doses of morphine were administered and per-
cent analgesia was determined at each dose level and
ED50 calculated by probit analysis.39 Subsequently var-
ious doses of test compound and morphine (ED50 dose)
Compound
R-(+)-2-(4-chlorobenzoyl)-1,2,3,4-tetra-
hydro-9H-pyrido(3,4-b)indole-3-carboxylic acid (4k,
R=C6H4-4-Cl) was prepared by the similar method.
Yield 76%: mp 231 ꢀC; [a]2D0 +2.96ꢀ (c 0.18, DMF); H
1
Table 3. Potentiation of morphine antinociception
Compound
Index of antinociception
Morphine+compound=(B)
Change (C=BÀA)
Log (Cꢂ100)
Morphine (A)
3a
3b
3c
3d
3e
3f
3g
3h
3i
3j
3k
4a
4b
4c
4d
4e
4f
4g
4h
4i
4j
4k
5a
5b
5c
5d
0.325ꢁ0.028
0.283ꢁ0.053
0.305ꢁ0.062
0.259ꢁ0.047
0.287ꢁ0.027
0.285ꢁ0.018
0.279ꢁ0.038
0.305ꢁ0.026
0.306ꢁ0.053
0.259ꢁ0.013
0.243ꢁ0.013
0.389ꢁ0.075
0.382ꢁ0.075
0.302ꢁ0.044
0.259ꢁ0.045
0.343ꢁ0.060
0.308ꢁ0.043
0.271ꢁ0.034
0.349ꢁ0.056
0.306ꢁ0.036
0.243ꢁ0.013
0.237ꢁ0.013
0.348ꢁ0.080
0.541ꢁ0.103
0.305ꢁ0.044
0.394ꢁ0.045
0.359ꢁ0.049
0.338ꢁ0.055
0.369ꢁ0.062
0.491ꢁ0.068
0.400ꢁ0.104
0.395ꢁ0.073
0.338ꢁ0.082
0.459ꢁ0.060
0.376ꢁ0.047
0.449ꢁ0.055
0.600ꢁ0.064
0.468ꢁ0.066
0.305ꢁ0.036
0.279ꢁ0.045
0.559ꢁ0.060
0.554ꢁ0.088
0.415ꢁ0.077
0.290ꢁ0.027
0.365ꢁ0.076
0.407ꢁ0.064
0.340ꢁ0.073
0.609ꢁ0.121
0.380ꢁ0.056
0.328ꢁ0.056
0.288ꢁ0.037
0.514ꢁ0.089
0.721ꢁ0.066
0.463ꢁ0.066
0.452ꢁ0.067
0.624ꢁ0.079
0.619ꢁ0.086
0.571ꢁ0.076
0.166
0.117
0.090
0.079
0.172
0.091
0.170
0.295
0.162
0.046
0.036
0.170
0.172
0.113
0.031
0.022
0.099
0.069
0.260
0.074
0.085
0.051
0.166
0.180
0.158
0.058
0.265
0.231
0.202
1.220
1.068
0.982
0.898
1.235
0.959
1.230
1.469
1.209
0.663
0.556
1.230
1.235
1.053
0.491
0.342
0.996
0.839
1.415
0.869
0.929
0.700
1.220
1.255
1.198
0.763
1.423
1.363
1.305
5e
Benzotript
Proglumide

A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
1567
were co-administered and anti-nociceptive response was
measured at 15 min intervals till it reached the basal
level. The control group of mice received an equivalent
amount of vehicle. Morphine served as control and
naltrexone, an opioid antagonist, was used to check the
specificity of opioid antinociceptive activity. Proglumide
and benzotript were used as standard.
APEX-3D module of the molecular simulations soft-
ware29 on the set of total 13 compounds (Table 5). The
overall match quality calculation42 is based on all
pairwise molecular similarities according to:
n
n
X X
2
Match ¼
n ðn À 1Þ
i¼1 j¼1þ1
Mol sim ði; jÞ
The data of each group of animals were averaged and
converted to an ‘index of antinociception’40 according
to the formula:
ð2Þ
Mol sim ði; iÞ þ Mol sim ðj; jÞ À Mol sim ði; jÞ
where n is the number of compounds and the Mol
sim(i,j) function is calculated according to
Index of Antinociception ðIAÞ
Test reaction time ðTRTÞ À Control RT
nn
n
¼
X X
At sim ðai; ajÞ
10 À Control RT
Mol sim ði; jÞ ¼
W3D ðai; ajÞ Â W2D ðai; ajÞ
ai¼1 aj ¼ 1
The control RT is the mean of the two basal readings
taken at 15 min intervals and a cut-off time of 10 s was used
in these experiments to avoid any damage to paw. IA
values of 1.0 indicate 100 percent antinociception and IA
values of 0 indicate no change in basal RT. The significant
increase in IA in compound-treated animals as compared
to morphine alone was taken as potentiation of morphine
anti-nociception by the test compound (Table 3).
where n is the number of atoms in molecule I, m is the
number of atoms in molecule j, At sim (ai, aj) is a func-
tion that calculates the similarity of atoms ai and aj,
Building up of the molecular model
The molecular structure of the compounds (3a–3k, 4a–
4k, benzotript and proglumide) were constructed on Sili-
con graphics INDY R-4000 workstation using the
sketch program in the Builder Module of INSIGHT II
software from the Molecular Simulation Incorporation
(MSI).29 The compounds were optimized for their geo-
metry (net charge 0.00 kcal/mol) by setting the forcefield
potential action, partial charge action and formal
charge action as fixed. The energy minimization method
for each compound employed the steepest descent, con-
jugate gradients and Newton–Raphson’s algorithms in
sequence followed by Quasi-Newton optimized pro-
cedure with a maximum number of iteration set at 1000
and using energy tolerance value of 0.001 kcal/mol. The
molecular structures were stored in MDL format and
used for the calculation of MOPAC 6.0 (MNDO
Hamiltonian41)-based indices, including atomic charges,
p population, and electron donor and acceptor indices.
The biophores (pharmacophores) were built with the
Figure 2. Superimposition of the compounds (3a–3k, benzotript and
proglumide) in model no. 1; biophoric sites represented by white/yel-
low solid spheres and cone: a biophoric pattern for anti-nociceptive
activity.
Table 4. 3D-QSAR Models (1–6) describing correlation and statistical reliability for antinociceptive activity
Model
RMSAa
RMSAb
R2
Chance
Size
Match
Var.c
Comp.
1
2
3
4
5
6
0.14
0.11
0.11
0.11
0.11
0.11
0.15
0.14
0.14
0.14
0.14
0.14
0.78
0.86
0.86
0.86
0.86
0.86
0.00
0.01
0.01
0.01
0.00
0.02
3
3
3
3
3
3
0.26
0.39
0.51
0.49
0.45
0.44
2
2
2
2
2
2
13
12
12
12
12
12
aRMSA, root mean square error of approximation.
bRMSP, root mean square error of leave-one-out prediction.
cVar., variable.

1568
A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
Table 5. Apex-3D Statistical Parameters for the selected six best models
Model
R2
RMSA
RMSP
F
n
Total hydrophobicitya
Ib
1
2–6
0.775
0.860
0.140
0.113
0.152
0.144
17.224
27.636
13
12
0.255
0.333
À0.628
À0.617
aTotal hydrophobicity, calculated as logP (sum of atomic increaments).42
Table 6. Experimental, calculated and predicted activity data and parameter values for the chiral title esters from Apex-3D model no. 1
Compound
Total hydrophobicitya
Experimental
Calculated
Calculated error
Predicted
Predicted error
3a
3b
3c
3d
3e
3f
3g
3h
3i
3j
3k
3.10
3.05
2.60
1.85
2.70
2.75
3.05
3.35
3.10
3.05
3.10
3.30
2.35
1.22
1.07
0.98
0.90
1.24
0.96
1.23
1.47
1.21
0.66
0.56
1.36
1.30
1.24
1.23
1.12
0.92
1.14
1.15
1.23
1.31
1.24
0.60
0.62
1.29
1.05
À0.02
À0.16
À0.13
À0.03
0.09
1.25
1.25
1.13
0.96
1.13
1.17
1.23
1.26
1.25
0.54
0.68
1.28
0.99
À0.03
À0.18
À0.15
À0.06
0.10
À0.20
À0.22
0.00
0.16
0.00
0.21
À0.03
À0.04
0.06
0.12
À0.06
0.07
0.25
À0.12
0.09
0.32
Benzotript
Proglumide
aThe other parameter I, Indicator has value of 1.00 for 3j and 3k (R config.) and 0.00 for rest compounds (3a–3i) (S config.) of the series.
Figure 3. Biophoric sites (A, B and C) represented by dotted circles
and ovals
on the chiral title esters (3a–3k) of model nos 1–6.

A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
1569
Table 7. Comparative data of experimentally observed and calcu-
lated biological activity of partially racemized title acids from Apex-
3D models
5 mm multinuclear inverse probehead with Z-shielded
gradient at room temperature using CDCl3 as a solvent.
300 MHz H, H COSY. relaxation delay d1=2 s; 90ꢀ
pulse, 6.88 ms for H; 1 k points in t2; spectral width,
1
1
Compound
Experimental
Model
no. 1
Model
no. 2
Model
nos. 3–6
1
8 ppm in both dimension; 256 experiments in t1; linear
prediction to 512 points; zero filling up to 1 k and apo-
dization with sine bell in both dimension prior to double
Fourier transform.
4a
4b
4c
4d
4e
4f
4g
4h
4i
1.23
1.23
1.22
1.10
0.92
1.13
1.14
1.22
1.31
1.23
0.59
0.60
—
1.22
1.20
1.05
0.82
1.09
1.10
1.20
1.32
1.22
0.56
0.60
1.235
1.053
0.491
0.342
0.996
0.839
1.415
0.869
0.920
0.70
1.20
1.05
0.82
1.09
1.10
1.20
1.32
1.22
0.58
0.60
300 MHz H, H ROESY. relaxation delay d1=2 s; 90ꢀ
1
1
1
pulse, 6.88 ms for H; spin lock=250 ms; 1 k points in t2;
spectral width 8 ppm in both dimension; 256 experi-
ments in t1; linear prediction to 512 points; zero filling
up to 1 k and apodization with 90ꢀ shifted sine bell in
both dimension prior to double Fourier transform.
4j
4k
W3D (ai, aj) is weighing function for matching atoms
based on the Cartesian distances between them, and
W2D(ai, aj) is a weighting function based on the topo-
logical distances of atoms from the biophore.
Acknowledgements
The authors are thankful to Dr. Raja Roy and Mr. B.S.
Joshi for the NMR studies, Mr. A.S. Kushwaha for the
technical assistance in molecular modeling and QSAR
and SKP is grateful to CSIR, New Delhi for a senior
research fellowship.
The 3D-QSAR equations were derived with the site
radius set at 0.60, the occupancy at 8, the sensitivity at
0.80, and the randomization at 100. The total hydro-
phobicity, total refractivity and indicator were selected
as global properties. The biophoric centers and second-
ary sites combined to global properties (total hydro-
phobicity, total refractivity, and indicator) were used to
obtain an equation to predict the percent potentiation
of morphine antinociception. The biophoric centers
were set to charges, p population, HOMO, LUMO,
hydrogen acceptor, hydrogen donor, refractivity and
hydrophobicity. The secondary sites were set to hydro-
gen acceptor, presence; hydrogen donor, presence; het-
eroatom, presence; hydrophobic, hydrophobicity; steric,
refractivity; ring, presence.
References
1. Portnoy, R. K. In Proceedings of the 7th World Congress on
Pain, Progress in Pain Research and Management; Crebhart,
G. F., Hammond, D.L., Jensen, T. S. Eds. IASP, Seattle,
WA,1994; Vol. 2, pp 595–619.
2. Foley, K. M. Clinical Tolerance to Opioids: In Towards a
New Pharmacotherapy of Pain; Basbaum, A. I.; Besson, J. M.
Eds. John Wiley & Sons Ltd. New York, 1991; pp 181–203.
3. Herman, B. H.; Vocci, F.; Bridge, P. Neuropsychopharma-
cology 1995, 13, 269.
4. Kolesnikov, Y. A.; Pick, C. G.; Pasternak, G. W. Eur. J.
Pharmacol. 1992, 221, 399.
Determination of enantiomeric excess (ee)
5. Trujillo, K. A.; Akil, H. Science 1991, 251, 85.
6. Trujillo, K. A.; Akil, H. Brain Res. 1994, 633, 178.
7. Trujillo, K. A.; Akil, H. Drug Alcohol Depend. 1995, 38, 139.
8. Arends, R. H.; Hayashi, T. G.; Luger, T. J.; Shen, D. D.
JPET 1998, 286, 585.
9. Hui, S. G.; Sevilla, E. L.; Ogle, C. W. Br. J. Pharmacol.
1996, 118, 1044.
10. Zarrindast, M. R.; Sajedian, M.; Rezayat, M.; Ghazi-
Khansari, M. Eur. J. Pharmacol. 1995, 273, 203.
11. Basbaum, A. I.; Fields, H. L. Ann. Rev. Neurosci. 1984, 7,
309.
12. Fields, H. L.; Heinrichu, M. M.; Mason, P. Ann. Rev.
Neurosci. 1991, 14, 219.
13. Le Bars, D. In Neuronal Serotonin Osborne, N. N.,
Hamon, M. Eds.; John Wiley & Sons Ltd. New York, 1988;
pp 171–228.
14. Watkins, L. R.; Kinscheck, I. B.; Mayer, D. J. Science
1984, 224, 395.
15. Fairs, P. L.; Komisaruk, B. R.; Watkins, L. R.; Mayer,
D. J. Science 1983, 219, 310.
Enantiomeric excess (ee) was determined for each title
compound and was found to be greater than 95% for
each chiral title esters (3a–3l) and was assessed in the
following manner. Using chiral HPLC analysis (Chir-
adex1 column) eluting with isocratic mixture of MeOH/
H2O(30/70), all the chiral esters appeared as a single
distinct peak. However, for the partially racemic title
acids (4a–4l), baseline separation could not be obtained
due to tailing. Several solutions of each synthesized (R/
S enantiomer) chiral title esters were prepared (2.0 mg/
5.0 mL of methyl alcohol) and mixed in the following
proportions (95:5, 97.5:2.5 and 98.75:1.25). These solu-
tions were analyzed, and it was found that each enan-
tiomer could easily be detected as a shoulder in the
97.5:2.5 mixture but not in the 98.75:1.25 mixture,
which implied that the ee of each of these chiral title
esters was at a minimum >95%.
16. Wank, S. A.; Pisegna, J. R.; DeWeerth, A. Proc. Natl.
Acad. Sci. U.S.A. 1992, 89, 8691.
17. Lee, Y.-M.; Beinborn, M.; McBride, E. W.; Lu, M.;
Kolakowski, L. F., Jr.; Kopin, A. S. J. Biol. Chem. 1993, 268,
8164.
NMR experiment
All experiments were carried out on Bruker AVANCE
DRX 300 MHz FT NMR spectrometer equipped with a

1570
A. K. Saxena et al. / Bioorg. Med. Chem. 9 (2001) 1559–1570
18. DeWeerth, A.; Pisegna, J. R.; Huppi, K.; Wank, S. A.
Biochem. Biophys. Res. Commun. 1993, 194, 811.
19. Wank, S. A.; Pisegna, J. R.; DeWeerth, A. Ann. N. Y.
Acad. Sci. 1994, 713, 49.
20. Beinfeld, M. C. Neuropeptides 1983, 3, 411.
21. Kopin, A. S.; Lee, Y. M.; McBride, E. W.; Miller, L. J.;
Lu, M.; Lin, H. Y.; Kolakowski, L. F., Jr.; Beinborn, M. Proc.
Natl. Acad. Sci. USA. 1992, 89, 3605.
31. Crippen, G. M. J. Comput. Phys. 1977, 24, 96.
32. Viswanadhan, V. N.; Ghose, A. K.; Weinstein, J. N. Bio-
chem. Biophys. Acta 1990, 1039, 356.
33. Viswanadhan, V. N.; Ghose, A. K.; Hanna, N. B.; Mat-
sumoto, S. S.; Avery, T. L.; Revankar, G. R.; Robins, R. K. J.
Med. Chem. 1991, 34, 526.
34. Saxena, A. K.; Saxena, M.; Chi, H.; Weise, M. Med.
Chem. Res. 1993, 3, 201.
22. Pisegna, J. R.; DeWeerth, A.; Huppi, K.; Wank, S. A.
Biochem. Biophys. Res. Commun. 1992, 189, 296.
23. Lindefors, N.; Linden, A.; Brene, S.; Sedvall, G.; Persson,
H. Prog. Neurobiol. 1993, 40, 671.
24. Altar, C. A.; Boyar, W. C. Brain Res. 1989, 483, 321.
25. Singh, L.; Lewis, A. S.; Field, M. J.; Hughes, J.; Woo-
druff, G. N. Proc. Natl. Acad. Sci. USA. 1991, 88, 1130.
26. Bradwejn, J. In Cholecystokinin Anxiety: Neuron Behavior;
Bradwejn, J., Vasar, E., Eds.; Landes: Austin, TX, 1995; pp
73–86.
27. Hahne, W. F.; Jensen, R. T.; Lemp, G. F.; Gardner, J. D.
Proc. Natl. Acad. Sci. USA. 1981, 78, 6304.
28. Tripathi, R. C.; Patnaik, G. K.; Saxena, A. K. Indian J.
Chem. 1989, 28B, 333.
29. InsightII, version 2.3.0; Biosym Technologies: San Diego,
1993.
30. Kubinyi, H., Ed. 3D QSAR in Drug Design. Theory,
Methods, and Applications; ESCOM Science: Leiden, 1993.
35. Golender, V. E.; Rozenblit, A. B. In Logical and Comina-
torial Algorithms for Drug Design; Research Studies: UK,
1983.
36. Saxena, A. K.; Jain, P. C.; Anand, N.; Dua, P. R. J. Med.
Chem. 1973, 16, 560.
37. Greco, G.; Novellino, E.; Silipo, C.; Vittoria, A. Proceed-
ings of the 9th European symposium on Structure Activity
Relationships: QSAR and Molecular Modelling; Wermuth, C.
G., Ed.; ESCOM: Leiden, 1993; pp 466–470.
38. Eddy, N. B.; Leimbach, D. J. Pharmacol. Exp. Ther. 1953,
107, 385.
39. Finney, D. J. In Probit Analysis. Cambridge University
Press, Cambridge, UK, 1971.
40. Azami, J.; Llewelyn, M. B.; Roberts, M. H. T. Pain 1982,
12, 229.
41. Dewar, M. J. S.; Thiel, W. J. Amer. Chem. Soc. 1977, 99,
4899, 4907.
42. Apex-3D User Guide; Biosym/MSI: San Diego, 1993.