Conformational considerations in the design of dual antagonists of platelet-activating factor (PAF) and histamine

Article abstract of DOI:10.1016/S0968-0896(99)00075-9

Following the discovery of the first dual antagonist of platelet-activating factor (PAF) and histamine, 1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]cyclohepta[1,2-b]pyridin-11-ylidene)piperidine, Sch 37370, 1, a related series of structures, exemplified by (±)-1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]-cyclohepta[1,2-b]pyridin-11-yl)piperazine, Sch 40338, 2, were prepared. Interestingly, the compounds exhibited a parallel structure-antiallergy activity relationship, suggesting that the two series may adopt a common conformation at the PAF receptor. Conformational analysis led to a proposal for this bioactive conformation accessible to both series. The synthesis of novel conformationally constrained analogues that might mimic the proposed bioactive conformation of these compounds, and the evaluation of their in vitro antiallergy activity form the subject matter of this report. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0968-0896(99)00075-9

Bioorganic & Medicinal Chemistry 7 (1999) 1413±1423
Conformational Considerations in the Design of Dual Antagonists
of Platelet-Activating Factor (PAF) and Histamine
James J. Kaminski, a,* Nicholas I. Carruthers, a Shing-Chun Wong, a Tze-Ming Chan, a
M. Motassim Billah, a S. Tozzi a and Andrew T. McPhail b
aSchering-Plough Research Institute, 2015 Galloping Hill Road, K15-1, 1800, Kenilworth, NJ 07033, USA
bDepartment of Chemistry, Duke University, Durham, NC 27708, USA
Received 5 October 1998; accepted 8 February 1999
AbstractÐFollowing the discovery of the ®rst dual antagonist of platelet-activating factor (PAF) and histamine, 1-acetyl-4-(8-
chloro-5,6-dihydro-11H-benzo[5,6]cyclohepta[1,2-b]pyridin-11-ylidene)piperidine, Sch 37370, 1, a related series of structures, exem-
pli®ed by (‹)-1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]-cyclohepta[1,2-b]pyridin-11-yl)piperazine, Sch 40338, 2, were pre-
pared. Interestingly, the compounds exhibited a parallel structure±antiallergy activity relationship, suggesting that the two series
may adopt a common conformation at the PAF receptor. Conformational analysis led to a proposal for this bioactive conforma-
tion accessible to both series. The synthesis of novel conformationally constrained analogues that might mimic the proposed
bioactive conformation of these compounds, and the evaluation of their in vitro antiallergy activity form the subject matter of this
report. # 1999 Elsevier Science Ltd. All rights reserved.
Introduction
antagonist of PAF and histamine evaluated in clinical
trials.5,6
The pathophysiology of allergic and in¯ammatory dis-
eases such as asthma is complex and probably involves
the biosynthesis and/or release of multiple mediators.
These mediators may be preformed, such as histamine,
which acts through speci®c histamine H1-receptors.
Alternatively, they can be synthesized in response to
antigenic stimulation, such as the leukotrienes, (LTB4,
LTC4, and LTD4), the prostaglandins, thromboxane A2
(TxA2), and platelet-activating factor (PAF). Regardless
of their genesis, the presence of these mediators elicit the
physiological e€ects that characterize the disease, i.e.
smooth muscle contraction, vasopermeability, in¯am-
matory cell in¯ux, edema, and mucous production. The
importance of the contribution that any one of these
mediators makes to clinical asthma is largely unknown.
However, with the expectation that an inhibitor of more
than one mediator would be clinically more e€ective than
a selective mediator inhibitor, we reported the results of
the structure±antiallergy activity relationship studies1±4
that culminated with the identi®cation of 1-acetyl-4-(8-
chloro-5,6-dihydro-11H-benzo[5,6]cyclohepta-[1,2-b]pyr-
idin-11-ylidene)piperidine, Sch 37370 1, as the ®rst dual
Following the disclosure of (1), we described a related, yet
structurally distinct, series of compounds exempli®ed by
(‹)-1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo-[5,6]
cyclohepta[1,2-b]pyridin-11-yl)piperazine, Sch 40338 (2)
which also exhibited in vitro dual PAF and histamine
H1-receptor antagonist activity (Table 1).7,8 This series
of compounds contains a piperazine ring attached to the
11-position of the tricyclic system in place of the piper-
idinylidene ring.
Replacement of the Á11-carbon-to-carbon double bond
with a carbon-to-nitrogen single bond, in addition to
introducing a chiral center at C-11 in the molecule, also
signi®cantly increases the conformational space acces-
sible to analogues in the piperazine series relative to
those in the piperidinylidene class.
Key words: Platelet activating factor (PAF); H1-receptor antagonist;
conformational analysis; rigid analogues; molecular modeling.
* Corresponding author. Tel.: +1-908-740-3521; fax: +1-908-740-7545;
e-mail: james.kaminski@spcorp.com
0968-0896/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(99)00075-9

1414
J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
More recently, we have extended our investigations to
include (‹)-1-acetyl-4-(8-chloro-5,6-dihydro-11H-ben-
zo[5,6]cyclohepta[1,2-b]pyridin-11-yl)piperidine (3) and
(‹)-1-acetyl-4-(8-chloro-5,6-dihydro-11-methyl-11H-
benzo[5,6]cyclohepta[1,2-b]-pyridin-11-yl)piperidine 4
Chemistry
Synthesis of the desired rigid analogues proceeded from
the tricycle 16 which was obtained from ketone 149,10
(Scheme 1). Thus, 14 was reduced to alcohol 15, and the
hydroxyl group was removed via conversion to the cor-
responding chloride followed by treatment with zinc
and acetic acid. Tricycle 16 was condensed with methyl
1,2,5,6-tetrahydro-1-methylnicotinate 17 to a€ord the
Michael addition product 18. Condensation of 16 with
17 proceeded in satisfactory yield only when the anion
of 16 was generated with sodium amide in ammoniacal
dimethoxyethane, with the ammonia being removed by
distillation prior to the addition of 17. Hydrolysis of 18
to its corresponding carboxylic acid was followed by
.
In these two classes of compounds, a piperidine ring is
attached to the 11-position of the tricyclic system in
place of either the piperazine or piperidinylidene rings.
Furthermore, in 4, the C-11 hydrogen atom in 2 has
been replaced by a methyl group, the presence of which
introduces additional gauche interactions with the
piperidine ring, thereby, a€ecting even further the con-
formational space accessible to analogues of this series.
Interestingly, while 3 and 4 exhibit di€ering levels of
PAF antagonist activity relative to either 1 or 2, both
compounds are essentially devoid of histamine H1-
receptor antagonist potency (Table 1). These limited
observations led us to propose the existence of an
interrelationship between the conformational require-
ments of these compounds and their antiallergy activity.
Conformational analysis of 1±4, examined using a vari-
ety of theoretical and experimental methods, suggested
a bioactive conformation for these series. Synthesis of
rigid analogues that might mimic the proposed bioactive
conformation of these compounds, and evaluation of
their in vitro antiallergy activity form the subject matter
of this report.
Scheme 1. Synthesis of (‹)-3-methyl-7-chloro-2,3,4-b,4a,9,10,14b-
a,14c-b octahydropyrido(300,200:60.70)cyclohepta(10,20,30:4,5)naphtho (2,
3-c)pyridin-5(1H)-one, 9, and (‹)-3-methyl-7-chloro-2,3,4,4a-a,9,10,
14b-a,14c-b octahydropyrido(300,200:60.70)cyclohepta(10,20,30:4,5) naph-
tho(2,3-c)pyridin-5(1H)-one (19).
Table 1. In vitro antiallergy activity and Macromodel MM2 conformational energies of 1 and related analogues
c
Compd
PAF antagonista
IC50 (mM)
H1-bindingb
Ki (mM)
pA2
Eequatorial
(kcal/mol)
Eaxial
(kcal/mol)
ÁEaxial
>equatorial
(kcal/mol)
1
2
3
4
5
7
8
10
11
12
13
0.61‹0.05d
0.59‹0.05d
2.5g
0.32‹0.09e
5.4‹0.5e
Ðh
+29.9
+31.2
+40.6
+26.3
+26.2
+35.5
3.6f
5.0
5.1
>50g
>50g
Ð1
0.05g
7.30
3000g
4.8j
300g
3.7g
14.3g
+33.3
+32.3
1.0
>50g
0.00017g
0.001g
9.76
8.85
a
Values are a measure of the concentration of the drug required to cause a 50% inhibition of PAF-induced platelet aggregation of human pla-
telet-rich plasma when challenged with 25 nM PAF.1 The values represent the mean of two independent experiments unless otherwise indicated.
b
Values were determined using a receptor binding assay using rat brain membranes and the experimentally determined value of 2.7 nM for the
KD of [3H] pyrilamine.1
c
pA2= log [B], where [B] is the dose of antagonist that necessitates doubling the dose of agonist to compensate for the action of the antagonist.
Value is the mean‹the standard error of the mean for eleven independent experiments.
Value is the mean‹the standard error of the mean for eight independent experiments.
d
e
f
Protonated form of 2: Eequatorial=+31.3 kcal/mol and Eaxial=+29.9 kcal/mol, ÁEaxial
Approximate value of IC50 or Ki, correlation coecient (r2) for the analysis >0.9.
Binding to histamine-H1 receptor: 32% inhibition at 2.8 mM.
= 1.4 kcal/mol.
>equatorial
g
h
i
Binding to histamine-H1 receptor: 0% inhibition at 2.8 mM.
Global minimum energy conformation, E=+27.8 kcal/mol. Local minimum conformation, E=+32.6 kcal/mol.
j

J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
1415
Friedel±Crafts acylation to a€ord the pentacyclic ketones
9 and 19. Although 9 and 19 were separable, it was more
convenient to complete the synthesis using the mixture
since the di€ering chiral center was removed and regener-
ated in the subsequent steps (Scheme 2). The N-methyl
piperidines 9 and 19 were transformed to their N-acet-
ylpiperidines 21 via their corresponding carbamates 20.
Direct removal of the ketone carbonyl group in 21 was
unsuccessful using a variety of experimental conditions.
Furthermore, when the ketone carbonyl group in 21 was
reduced to produce alcohol 22, direct removal of the hy-
droxyl group in 22 was also unsuccessful using a variety of
experimental conditions. Therefore, 22 was dehydrated
using tri¯uoromethanesulfonic acid to produce ole®n 23.
Catalytic hydrogenation of 23 produced a mixture of the
desired pentacyclic targets 7 and 8.
25. Conversion of 25 to 10 was accomplished using the
procedures that were used to convert 9 and 19 to 21.
Analogues 3 and 11 were prepared from nitrile 26
(Scheme 4). Treatment of 26 with N-methylpiperidine-4-
magnesium chloride gave intermediate ketone 27.
Sodium borohydride reduction of 27 produced the
alcohol 28. Cyclization of 28 using polyphosphoric acid
produced the required ring systems, albeit in low yield.
The 8-chloro and 10-chloro isomers, 30 and 29, respec-
tively, were separated and transformed into 3 and 11 via
their corresponding carbamates as described previously.
Biological test methods
Platelet aggregation assay. Inhibition of PAF-induced
aggregation of human platelets was determined using the
experimental protocol described by Billah et al.1
To investigate the e€ect of steric bulk at the C-10 posi-
tion of the analogue on the biological activity
required the synthesis of 10 and 11. Synthesis of 10
proceeded from ketone 24 (Scheme 3). Treatment of
24 with N-methylpiperidine-4-magnesium chloride fol-
lowed by dehydration of the addition product produced
Histamine-H1 binding assay. The assay was performed
using the experimental protocol described by Billah et
al.1
Computational methods
Calculations were performed with SYBYL Version 5.3
and MacroModel V2.7 using a Silicon Graphics Chal-
lenge XL operating under IRIX.
Results and Discussion
Conformational analysis
A conformational search of 1 using a combination of
SYBYL Version 5.3 and MacroModel V2.7 suggested four
conformational states for the ring system (Fig. 1). In
addition, three distinct conformations of 1 within ca.
5 kcal/mol of its global minimum energy conformation
were also identi®ed (Fig. 2). The di€erence in energies
(ÁE) observed between the conformations that result from
interconversion of the central seven-membered ring of the
Scheme 2. Synthesis of (‹)-3-acetyl-7-chloro-1,2,3,4,4a-b,5,9,10,14b-
a,14c-b decahydropyrido(300,200:60,70)-cyclohepta(10,20,30:4,5)naphtho-
(2,3-c)pyridine (7) and (‹)-3-acetyl-7-chloro-1,2,3,4,4a-a,5,9,10,14b-a,
14c-b decahydropyrido(300,200:60,70)cyclohepta(10,20,30:4,5)naphtho(2,3-
c)pyridine (8).
Scheme 4. Synthesis of (‹)-1-acetyl-4-(8-chloro-5,6-dihydro-11H-
benzo[5,6]cyclohepta[1,2-b]pyridin-11-yl)piperidine (3) and (‹)-1-
acetyl-4-(10-chloro-5,6-dihydro-11H-benzo[5,6]cyclohepta[1,2-b]pyr-
idin-11-yl)piperidine (11).
Scheme 3. Synthesis of 1-acetyl-4-(10-chloro-5,6-dihydro-11H-ben-
zo[5,6]cyclohepta[1,2-b]pyridin-11-ylidene)piperidine (10).

1416
J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
Figure 1. Conformational analysis of 1-acetyl-4-(8-chloro-5,6-dihydro-
11H-benzo[5,6]cyclohepta[1,2-b]pyridin-11-ylidene)piperidine (1).
Figure 3. Comparison between the global minimum energy conforma-
tion of 1 and the solid-state conformation of loratadine, (5).
mand in SYBYL 5.3. Interestingly, the global minimum
energy conformation of 1 determined in the gas phase
di€ers from the solid-state conformation of 5 only in the
conformation of the central seven-membered ring of the
tricyclic system, i.e. the energy di€erence between twist
orientation conformers of 1 di€ers by 0.6 kcal/mol (Fig.
1). The weighted root mean square deviation for the ®t
is RMS=+0.4006 A.
Conformational searches and determination of the global
minimum energy conformations for 2, 3, and 4 were also
investigated; the results are summarized in Table 1. These
analogues contain either a piperazine or piperidine ring
attached to the 11-position of the tricyclic system in place
of the piperidinylidene ring. Replacement of the Á11-car-
bon-to-carbon double bond with a carbon-to-nitrogen
single bond in these analogues allows the heterocycle
attached to C-11 to occupy either an axial, or an equator-
ial position with respect to the conformation of the central
seven-membered ring of the tricyclic system. Comparison
between the MacroModel MM2 energies of the axial and
equatorial conformers of 2, 3, and 4 suggests that in each
case, the heterocyclic substituent at the 11-position prefers
to be axially oriented in its global minimum energy con-
formation (Table 1).
Figure 2. Conformations accessible to 1-acetyl-4-(8-chloro-5,6-dihydro-
11H-benzo[5,6]cyclohepta[1,2-b]pyridin-11-ylidene)piperidine (1) within
5 kcal/mol of its global minimum energy conformation.
tricyclic system are equal and opposite to the di€erences in
energies (ÁE) observed between conformations that result
from di€erent orientations of the piperidinylidene ring
with respect to the central seven-membered ring of the tri-
cyclic system, i.e. ÁE=‹0.6 kcal/mol. As a result, the
four conformational states accessible to the ring system are
in dynamic equilibrium and atropisomers are not observed
at ambient temperature.
Recently, the single-crystal X-ray structure of the ( )-
isomer of 2, as its N-acetyl d-leucine salt, has been
determined and its absolute con®guration has been
established as the S-stereoisomer, 2.7,8
However, while conformations of 1 that di€er only in
the orientation of the amide carbonyl are isoenergetic,
the rotational barrier of the amide was suciently high
1
relative to the H NMR time scale such that distinct
resonances corresponding to each rotameric form of the
amide were observed. Another higher energy conformer
of 1, but within 5 kcal/mol of the global minimum, has the
piperidine ring in a twist-chair conformation (Fig. 2).
The global minimum energy conformation of 1 was
compared to the solid-state conformation of loratadine,
5, determined by single-crystal X-ray analysis (Fig. 3.
The conformations were compared by ®tting the atoms
of the tricyclic ring system by use of the ``FIT'' com-
In the solid-state, the piperazine ring of 20 prefers to
adopt an axial orientation. In solution, an axial orien-
tation for the piperidine ring in 3 has also been estab-
lished with the observation of a positive nuclear

J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
1417
Overhauser enhancement (NOE) between H11 and the
1
of the C-11 substituent of each analogue described in the
Table 1 is consistent with this prediction. This observation
prompted the design and synthesis of conformationally
constrained analogues that might mimic the proposed
bioactive conformation of 1.
aromatic proton attached to C-10 in its H NMR spec-
trum. These observations are consistent with the pre-
ferred axial orientation of the piperazine and piperidine
rings predicted from the global minimum energy con-
formations of 2, 3, and 4, determined in the gas phase.
Furthermore, they are consistent with the results of
conformational studies described for similar systems.12
Synthesis of rigid analogues that might mimic the
proposed bioactive conformation of 1. The pentacyclic
ring system, exempli®ed by 6, was designed as a poten-
tial class of compounds that might mimic the proposed
bioactive conformation of 1. Conformational rigidity
imposed by the introduction of an additional six-mem-
bered ring formed by joining C-10 of the tricylic system
to the b-carbon atom of the piperidine ring through an
intermediate carbon atom could satisfy this require-
ment. However, the pentacyclic ring system proposed
contains three asymmetric centers and can potentially
exist as four possible diastereomeric pairs.
The in vitro potencies for PAF antagonism of 1 and 2
are identical (Table 1). In addition, the structure±anti-
allergy activity relationship studies for the two series
parallel each other closely.7,8 These observations suggest
that both compounds may bind to the PAF receptor in
a similar manner, and the two compounds may share a
common bioactive conformation. While conformational
analyses of 1 and 2 suggest that their global minimum
energy conformations are quite di€erent (Fig. 4a), a
common conformation is accessible in which the con-
formation of 2 is only slightly higher in energy, 3.6 kcal/
mol (Table 1). It should be noted that the pKa of the
basic piperazine nitrogen atom in 2 is such that at phy-
siologic pH 7.4 it will predominantly exist in its proto-
nated form. Under these circumstances, the axial
orientation of the piperazine ring is still preferred rela-
tive to its equatorial orientation, however, the di€erence
in energy (ÁE) between these two conformations is
lower, 1.4 kcal/mol (Table 1). Regardless of the state of
protonation of the piperazine nitrogen atom, the con-
formation in which the piperazine ring is equatorial
with respect to the central seven-membered ring of the
tricyclic system may represent the bioactive conforma-
tion at the PAF receptor (Fig. 4b). In order for 3 and 4
to mimic the proposed bioactive conformation of 1,
their axially oriented C-11 substituent must also occupy
an equatorially oriented position. Thus, we proposed that
for each analogue, the magnitude of the di€erence in
energy (ÁE) associated with orienting the C-11 sub-
stituent equatorially from its preferred axial global mini-
mum energy orientation, should be re¯ected in the in vitro
potency observed for PAF antagonism of the corre-
Synthesis of this novel ring system has been accom-
plished,13 and resulted in the preparation of two dia-
stereoisomers in racemic form, (‹)-3-acetyl-7-chloro-
1,2,3,4, 4a-b,5,9,10,14b-a,14c-b-decahydropyrido (300,200:
600,700)cyclohepta(10,20,30:4,5)-naphtho(2,3 -c)pyridine, 7
and (‹)-3-acetyl-7-chloro-1,2,3,4,4a-a,5,9,10,14b-a,14c-
b - decahydropyrido(300,200:600,700) - cyclohepta(10,20,30:4,5)
naphtho(2,3-c)pyridine, 8.
sponding analogue, i.e., the greater the ÁEaxial->equatorial
,
the lower the in vitro potency for PAF antagonism of the
analogue. The trend observed between the in vitro
potency for PAF antagonism and the energetic cost asso-
ciated with the axial-to-equatorial conformational change
Figure 4. (a) Global minimum energy conformations of 1 and 2. (b) A higher energy conformation of 2 that mimics the proposed bioactive conformation of 1.

1418
J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
Assignment of the relative stereochemistry at the ring
junctions in the two racemic diastereomeric products
was established by 2-D NOESY experiments, and the
1H NMR stereochemical assignment in (‹)-3-methyl-7-
chloro-2,3,4-b,4a,9,10,14b-b,14c-b-octahydropyrido(300,
200:60.70)cyclohepta(10,20,30:4,5) - naphtho(2,3 - c)pyridin -
5(1H)-one, 9, an intermediate precursor to the ®nal
products, was corroborated unequivocally by single-
crystal X-ray analysis (Fig. 5).
bioactive conformation of 1 is 4.8 kcal/mol. While the
magnitude of this energy di€erence (ÁE) might be con-
sidered sucient to account for the inactivity of 7 as an
antiPAF agent, it is inconsistent with the in vitro anti-
PAF potency observed for 3, IC50=2.5 mM, since 3
must also expend approximately 5 kcal/mol to adopt the
bioactive conformation of 1 (Table 1).
In the absence of further structural information regarding
the binding of 7 and 8 to the PAF receptor, these obser-
vations suggest that while attainment of the proposed
bioactive conformation of 1 may be a necessary condition,
alone it is not sucient to impart antiPAF activity to the
analogue. On this basis, an examination of the e€ect of
substitution at the 10-position in 1 and related analogues
on their antiPAF activity was initiated.
E€ect of a 10-substituent in 1 and related analogues on
in vitro PAF antagonism. The magnitude of the energy
di€erence between axial and equatorial orientations of a
C-11 substituent could be markedly a€ected by the
absence, or presence of a C-10 substituent. Comparison
of the in vitro potency for PAF antagonism of 1 relative
to its 10-chloro isomer, 10, suggests that the increased
steric interaction of the 10-chloro substituent decreases
the antiPAF potency of the analogue by approximately
sixfold, i.e. for 1, IC50=0.61‹0.05 mM relative to 10,
IC50=3.7 mM. Interestingly, comparison of the in vitro
potency for PAF antagonism of 3 relative to its 10-chloro
isomer, 11, suggests that the increased steric interaction of
the 10-chloro substituent in this analogue also decreases
the potency of the analogue by approximately sixfold, i.e.
for 3, IC50=2.5mM relative to 11, IC50=14.3 mM. How-
ever, in the latter case, the increased steric interaction of
the 10-chloro substituent in 11 concomitantly reduces the
magnitude of the di€erence in energy (ÁE) associated
with orienting the C-11 substituent equatorially from its
preferred axial global minimum energy orientation rela-
tive to 3, i.e. for 3, ÁEaxial->equatorial= 5 kcal/mol relative
to 11, ÁEaxial->equatorial= 1 kcal/mol.
The gas-phase global minimum energy conformations
of 7 and 8 were identi®ed using a combination of sys-
tematic search and molecular dynamics techniques. The
global minimum energy conformation of 7 determined
in the gas phase is consistent with the solid-state and
solution conformation of 9 determined by single-crystal
X-ray analysis and 1H NMR spectroscopy, respectively.
Comparison between the global minimum energy con-
formations of 7 and 8 to the proposed bioactive con-
formation of 1, suggests that 8 may better mimic the
bioactive conformation relative to 7. However, a local
minimum energy conformation is accessible to 7 which
is only slightly higher in energy, 4.8 kcal/mol, which
better mimics the bioactive conformation of 1, (Fig. 6a,
b and c).
In either case, determination of the in vitro potencies for
PAF antagonism of 7 and 8 suggested that both dia-
stereoisomers are inactive antiPAF agents. The di€er-
ence in energy associated with 7 adopting a local
minimum energy conformation rather than its global
minimum energy conformation to better mimic the
While the global minimum energy conformation of 8
approximates the bioactive conformation of 1, and the
energetic cost associated with 7 attaining the proposed
bioactive conformation of 1 is approximately the same
as that required for 3, the in vitro potency for PAF
antagonism observed for 8 and 7 are 120- and 1200-fold
less, respectively, relative to the in vitro potency for
PAF antagonism observed for 3. These observations
support the proposal that although the conformational
constraint of the six-membered pentacyclic system in 8
and 7 may o€er an entropic advantage contributing to
antiPAF potency, the conformational constraint
Fig. 5. Solid-state conformation of one enatiomer, (‹)-3-methyl-7-
chloro-2,3,4,4a-b,9,10,14b-a,14c-b octahydropyrido(300,200:60.70)-cyclo-
hepta(10,20,30:4,5)naphtho(2,3-c)pyridin-5(1H)-one (9), as determined
by single-crystal X-ray analysis.

J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
1419
Figure 6. (a) Global minimum energy conformation of (‹)-3-acetyl-7-chloro-1,2,3,4,4a-a,5,9,10,14b-a,14c-b decahydropyrido(300,200:60,70)-cyclo-
hepta(10,20,30:4,5)naphtho(2,3-c)pyridine (8) ®t to the global minimum energy conformation of 1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]cy-
clohepta[1,2-b]pyridin-11-ylidene)-piperidine (1).
(b) Global minimum energy conformation of (‹)-3-acetyl-7-chloro-1,2,3,4,4a-b,5,9,10,14b-a,14c-b decahydropyrido(300,200:60,70)-cyclo-
hepta(10,20,30:4,5)naphtho(2,3-c)pyridine (7) (red) and a local minimum energy conformation of 7 (green) which resides 4.8 kcal/mol above its global
minimum energy conformation ®t to the global minimum energy conformation of 1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]cyclohepta[1,2-
b]pyridin-11-ylidene)-piperidine (1).
(c) Global minimum energy conformation of (‹)-3-acetyl-7-chloro-1,2,3,4,4a-a,5,9,10,14b-a,14c-b decahydropyrido(300,200:60,70)cyclohepta
(10,20,30:4,5)naphtho(2,3-c)pyridine (8) and a local minimum energy conformation of (‹)-3-acetyl-7-chloro-1,2,3,4,4a-b,5,9,10,14b-a,14c-b decahy-
dropyrido(300,200:60,70)cyclohepta(10,20,30:4,5)naphtho(2,3-c)pyridine (7) which resides 4.8 kcal/mol above its global minimum energy conformation ®t
to the global minimum energy conformation of 1-acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]cyclohepta[1,2-b]pyridin-11-ylidene)piperidine (1).

1420
J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
invoked is at the expense of a signi®cant steric interac-
tion in an area of the PAF receptor which apparently
does not tolerate steric bulk. Obviously, this conclusion
is drawn in the absence of additional structural infor-
mation regarding the binding of these ligands to the
PAF receptor. Knowledge of this information could
identify the speci®c interaction of the ligand with the
receptor that is responsible for decreasing the antiPAF
potency of the ligand.
Experimental
(‹)-3-Methyl-7-chloro-2,3,4-b,4a,9,10,14b-ꢀ,14c-ꢁ-octa-
hydropyrido(300,200:60.70)cyclohepta(10,20,30:4,5) naphtho(2,
3-c)pyridin-5(1H)-one (9) and (‹)-3-Methyl-7-chloro-2,
3,4, 4a-ꢀ,9,10,14b-ꢀ,14c-ꢁ-octahydropyrido(300,200:60.70)
cyclohepta(10,20,30:4,5)-naphtho(2,3-c)pyridin-5(1H)-one
(19). To a solution of ketone 14 (46.5 g, 190 mmol) in
EtOH (400mL) was added NaBH4 (6.0g, 160 mmol) and
the mixture stirred for a further 90min. The reaction
mixture was treated with ice/H2O (300mL) and extracted
with CH2Cl2 (3Â200 mL). The organic portions were
separated, concentrated to a slurry and redissolved in
EtOAc (400mL). This solution was washed with H2O
(2Â100 mL), dried over MgSO4, ®ltered and concentrated
to a slurry. Trituration with diisopropylether a€orded
alcohol 15 (44.1 g). The alcohol was dissolved in benzene
(500mL) at 0 ꢀC and treated with thionyl chloride (23.8 g,
200 mmol). The mixture was stirred at 0 ꢀC for 15min and
at room temperature for 30min. The solution was cooled
to 0 ꢀC and treated with 25% NaOH (250mL), stirred for
5 min and treated with EtOAc (200mL). The organic layer
was separated, washed with H2O (2Â250 mL), dried over
MgSO4, ®ltered and evaporated to give a brown oil
(ꢁ50 g). This material was dissolved in AcOH (100 mL),
treated with Zn dust (12 g, 184 mmol) and heated at re¯ux
temperature for 60min. The reaction mixture was ®ltered
and the residue washed with hot H2O (3Â50mL). The
combined ®ltrates were cooled (ice bath) and treated with
excess concd NH4OH solution. This solution was extrac-
ted with CH2Cl2 (3Â100 mL) and the combined extracts
dried over MgSO4, ®ltered and evaporated to give a
brown oil (ꢁ55g). Trituration with diethylether a€orded
16 (18.8 g) and evaporation of the supernatant followed
by ¯ash chromatography eluting with 1±5% MeOH/
CH2Cl2 a€orded additional 16 (22.6 g): Overall yield
(41.4 g, 95%); mp 92±94 ꢀC.
The in vitro histamine H1-receptor antagonist potency
of (‹)-7-chloro-3-methyl-1,2,3,4,4a-a,5,9,10,14b-a,14c-
b - decahydropyrido(300,200:60,70) - cyclohepta(10,20,30:4,5)
naphtho(2,3-c)pyridine, 13, was determined relative to
the classical histamine H1-receptor antagonist, azata-
dine, 12 and the nonsedating histamine H1-receptor
antagonist, loratadine, 5 (Table 1).
Comparison between either the Ki or pA2 values sug-
gests that 13 is a more potent H1-receptor antagonist
relative to the nonsedating antihistamine, loratadine, 5,
but is a weaker H1-receptor antagonist relative to the
classical antihistamine, azatadine, 12. In vivo, oral
administration of 13, at a screening dose of 0.1 mg/kg,
completely inhibited a histamine-induced bronchospasm
in the guinea pig. However, no separation between H1-
receptor antagonist activity and sedation was observed
for 13 in rats and mice.
A solution of NaNH2, prepared from Na (1.2 g, 52 mmol),
.
NH3 (150mL) and Fe(NO3)3 9H2O (0.05g), was treated
with a solution of 16 (11.5 g, 50mmol) in DME (125mL)
over 15min. A second portion of DME (125mL) was then
added and the excess NH3 removed by warming the reac-
tion mixture to 40ꢀC. The solution was then cooled to
78ꢀC and treated with a solution of 17 (7.75 g, 50mmol)
in DME (100mL) over 60min. Once the addition was
complete, the cooling bath was removed and NH4Cl (25 g)
added, followed by saturated NH4Cl solution (125mL).
This mixture was stirred for 5 min, then treated with die-
thylether (300mL). The organic layer was separated and
washed with H2O (150mL). The aqueous fractions were
combined and extracted with diethylether (150mL). The
organic fractions were combined, dried over MgSO4, ®l-
tered and evaporated to a€ord an oil (21 g). Silica gel
chromatography eluting with 1.5±5.0% NEt3/EtOAc gave
18 (12.0 g, 62%).
Summary and Conclusions
In this paper, we have undertaken a conformational ana-
lysis of four series of related PAF antagonists, 1, 2, 3, and
4. These analyses suggested that the compounds might
adopt a common bioactive conformation and conse-
quently, a rigid pentacyclic ring system, 6, was designed to
mimic the bioactive conformation proposed. Two dia-
stereoisomers, 7 and 8 were prepared and 8 was found to
better approximate the bioactive conformation. However,
both 7 and 8 were determined to be inactive as PAF
antagonists. This result led to an evaluation of the con-
sequences of introducing a 10-substituent, necessary for
the construction of the rigid system, and thus 10 and 11
were prepared. Comparison of the in vitro potency for
PAF antagonism of the 10-chloro isomers 10 and 11 with
their 8-chloro analogues, 1 and 3, respectively, suggests
that the increased steric interaction of the 10-substituent is
detrimental to biological activity. These results suggest
that whilst the conformational constraints of the six-
membered ring present in 7 and 8 may o€er an entropic
advantage, this is obtained only at the expense of intro-
ducing unfavorable steric bulk.
Ester 18 (1.3 g, 3.38 mmol) was treated with con-
centrated HCl (15 mL) and stirred at re¯ux temperature
for 2 h. The reaction mixture was evaporated to dryness
and the residue treated with tri¯uoromethanesulfonic acid
(30 mL) and heated at 90±95 ꢀC for 3 h. The solution was
poured onto ice, treated with excess NH4OH solution and

J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
1421
extracted with CH2Cl2 (3Â40mL). The organic extracts
were combined, dried over MgSO4, ®ltered and evapo-
rated to give an oil. Silica gel chromatography eluting with
1±5% NEt3/EtOAc gave 9 (0.35 g, 29%); mp 180±181 ꢀC
and 19 (0.27 g, 22%); mp 220±221 ꢀC.
with Zn dust (4g, 61mmol) and heated at re¯ux tempera-
ture for 45min. The reaction mixture was ®ltered and the
residue washed with H2O (3Â20mL) and CH2Cl2
(2Â20mL). The ®ltrate was treated with excess con-
centrated NH4OH solution and the organic layer sepa-
rated. The aqueous portion was re-extracted with CH2Cl2
(2Â50mL) and the combined organic fractions dried over
MgSO4, ®ltered and evaporated to approximately one-
third volume. This solution was treated with Ac2O (3mL,
32mmol) and heated at re¯ux temperature for 60min.
The reaction mixture was poured onto ice, treated with
excess concentrated NH4OH and the organic layer sepa-
rated. The aqueous layer was re-extracted with CH2Cl2
(3Â50mL) and the combined organic fractions dried over
MgSO4, ®ltered and evaporated. Silica gel chromato-
graphy gave the corresponding acetamide (3.3 g, 85%).
The acetamide was dissolved in EtOH (30 mL), treated
with KOH (2.2 g, 39mmol) and stirred at ambient tem-
perature for 30h. The solution was poured onto ice and
adjusted to pH 5.0 by addition of concentrated HCl. This
solution was extracted with CH2Cl2 (3Â50mL) and the
extracts combined, dried over MgSO4, ®ltered and eva-
porated to give a gummy solid (3.5 g). This material was
dissolved in tri¯uoromethanesulfonic acid (30 mL) and
stirred at 70±80 ꢀC for 2 h. The solution was poured onto
ice, treated with excess concentrated NH4OH and the
precipitate obtained isolated by ®ltration. The precipitate
was dissolved in CH2Cl2 (150mL), washed with H2O
(50 mL), dried over MgSO4, ®ltered and evaporated to
give a gummy solid. Trituration with diethylether gave 21
(2.8 g, 55%) mp 170±195 ꢀC.
(‹)-3-Acetyl-7-chloro-1,2,3,4,4a-ꢁ,5,9,10,14b-ꢀ,14c-ꢁ-
decahydropyrido - (300,200:60,70) cyclohepta (10, 20, 30: 4, 5)
naphtho(2,3-c)pyridine (7) and (‹)-3-acetyl-7-chloro-1,2,
3,4,4a-ꢀ,5,9,10,14b-ꢀ,14c-ꢁ-decahydropyrido(300,200:600,
700)cyclohepta (10,20,30:4,5)naphtho(2,3-c)pyridine (8). A
solution of 9/19 (0.27 g, 0.76 mmol) in benzene (20 mL)
was treated with NEt3 (0.3 g, 3 mmol) and tri-
chloroethylchloroformate (0.5 g, 2.4 mmol). The mix-
ture was heated at re¯ux temperature for 60 min;
whereupon,
a
second portion of trichloroethyl-
chloroformate (0.2 g, 0.95 mmol) was added and heating
continued for an additional 30 min. The hot solution
was poured onto ice and treated with excess con-
centrated NH4OH solution and extracted with EtOAc
(50 mL). The organic extracts were washed with H2O
(2Â20 mL), dried over MgSO4, ®ltered and evaporated
to give 20 (0.36 g, 92%) mp 249±250 ꢀC.
Carbamate 20 (0.36 g, 0.7 mmol) in AcOH (15 mL) was
warmed to 60 ꢀC to obtain a homogeneous solution,
treated with Zn dust (1.2 g, 18.4 mmol) and heated at
80±90 ꢀC for 45 min. The reaction mixture was ®ltered
and the residue washed with H2O (3Â15 mL) and
CH2Cl2 (2Â10 mL). The ®ltrate was treated with excess
concentrated NH4OH solution and the organic layer
separated. The aqueous portion was re-extracted with
CH2Cl2 (2Â15 mL) and the combined organic extracts
dried over MgSO4, ®ltered and evaporated to a€ord a
waxy solid. This material was dissolved in CH2Cl2
(20 mL), treated with Ac2O (1 mL, 10.6 mmol) and
heated at re¯ux temperature for 45 min. The reaction
mixture was poured onto ice, treated with excess con-
centrated NH4OH solution and the organic layer sepa-
rated. The aqueous portion was re-extracted with
CH2Cl2 (3Â20 mL) and the combined organic extracts
dried over MgSO4, ®ltered and evaporated to give a
crude solid. Silica gel chromatography eluting with 1±
8% NEt3/EtOAc followed by recrystallization from
diethylether gave 21 (0.068 g, 23%) mp 188±190 ꢀC.
Ketone 21 (1.95 g, 5.1 mmol) in DME (35 mL) and
EtOH (15 mL) was treated with NaBH4 (0.8 g, 21 mmol)
and stirred at ambient temperature for 24 h. The reac-
tion mixture was treated with ice and extracted with
CH2Cl2 (3Â50 mL). The combined organic extracts
were dried over MgSO4, ®ltered and evaporated to give
22 (1.82 g, 92%) mp 200±210 ꢀC.
Alcohol 22 (0.82 g, 2.14 mmol) was treated with tri-
¯uoromethanesulfonic acid (20 mL) and heated, steam
bath, for 40 min. The reaction mixture was poured onto
ice, treated with excess concentrated NH4OH and the
precipitate obtained isolated by ®ltration. The solid was
redissolved in EtOAc (40 mL), washed with H2O
(10 mL), dried over MgSO4, ®ltered and evaporated to
give 23 (0.75 g, 96%).
A more convenient synthesis of 21 from 18 is as follows.
Ester 18 (5.1 g, 13.25 mmol) in benzene (250 mL) was
treated with NEt3 (2.5 g, 25 mmol) followed by tri-
chloroethylchloroformate (4.2 g, 20 mmol). The mixture
was heated at re¯ux temperature for 60 min and then
treated with additional NEt3 (1.5 g, 15 mmol) followed
by trichloroethylchloroformate (2Â1 g) at 30 min inter-
vals. After a total period of 3 h at re¯ux temperature,
the reaction mixture was poured onto ice and treated
with excess concentrated NH4OH. The solution was
extracted with diethylether (3Â100 mL) and the organic
extracts combined, dried over MgSO4, ®ltered and
evaporated to give a brown oil (ꢁ12g). Silica gel chro-
matography eluting with 5±20% EtOAc/CH2Cl2 a€orded
the corresponding trichloroethylcarbamate (5.3 g, 74%).
This material was dissolved in AcOH (25 mL), treated
Alkene 23 (0.5 g, 1.37 mmol in EtOH (20 mL) was trea-
ted with PtO2 (0.03 g) and hydrogenated at 60 psi for
8 h. The reaction mixture was ®ltered and the ®ltrate
evaporated to a€ord a foam. Silica gel chromatography
eluting with 1±2% MeOH/CH2Cl2 gave 7 (0.05 g, 10%)
and 8 (0.09 g, 20%).
1-Acetyl-4-(10-chloro-5,6-dihydro-11H-benzo[5,6]cyclo-
hepta[1,2-b]pyridin-11-ylidene)piperidine (10). To a sol-
ution of ketone 24 (1.5 g, 6.2 mmol) in THF (50 mL)
was added a THF solution (1.4 M) of N-methyl-
piperidine-4-magnesium chloride (6.0 mL, 8.4 mmol)
and the reaction mixture stirred at ambient tempera-
ture overnight. The reaction mixture was treated with

1422
J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
saturated NH4Cl solution (50 mL) and extracted with
Et2O (2Â50 mL). The combined extracts were dried
over MgSO4, ®ltered and evaporated to a€ord a viscous
yellow oil (2.2 g). Silica gel chromatography eluting with
2% NEt3/EtOAc gave intermediate alcohol (1 g, 47%).
The alcohol intermediate (0.9 g, 2.62 mmol) was dis-
solved in methanesulfonic acid (10 mL) and stirred at
100±105 ꢀC for 24 h. The reaction mixture was poured
onto ice, treated with excess concentrated NH4OH
solution and extracted with CH2Cl2 (3Â35 mL). The
organic extracts were combined, dried over MgSO4, ®l-
tered and evaporated. Silica gel chromatography eluting
with 1±2% NEt3/EtOAc gave 25 (0.55 g, 65%). N-
Methylpiperidine 25 (0.55 g, 1.7 mmol) in benzene
(30 mL) was treated with NEt3 (0.3 g, 3 mmol) and tri-
chloroethylchloroformate (0.5 g, 2.4 mmol). The mix-
ture was heated at re¯ux temperature for 2.5 h and the
solution cooled by addition of ice. The solution was
treated with excess concentrated NH4OH solution and
extracted with Et2O (2Â30 mL). The combined organic
extracts were dried over MgSO4, ®ltered and evaporated
to a€ord a yellow oil (1.6 g). The crude material was
puri®ed by silica gel chromatography eluting with 2±5%
NEt3/EtOAc to a€ord the intermediate carbamate. The
carbamate (0.44 g, 0.91 mmol) in AcOH (10 mL) was
treated with Zn dust (0.4 g, 6.1 mmol) and heated at 80±
90 ꢀC for 1 h. The reaction mixture was ®ltered and the
residue washed with hot H2O (2Â10 mL) and CH2Cl2
(2Â20 mL). The ®ltrate was treated with excess con-
centrated NH4OH solution and the organic layer sepa-
rated. The aqueous layer was re-extracted with CH2Cl2
(2Â20 mL) and the combined organic extracts dried
over MgSO4, ®ltered and evaporated to give a yellow oil
(0.28 g). This material was dissolved in CH2Cl2 (20 mL),
treated with Ac2O (1 mL, 10.6 mmol) and heated at
re¯ux temperature for 1 h. The reaction mixture was
poured onto ice, treated with excess concentrated
NH4OH solution and the organic layer separated. The
aqueous was re-extracted with CH2Cl2 (3Â20 mL) and
the combined organic extracts dried over MgSO4, ®l-
tered and evaporated to give an oil (0.32 g). Silica gel
chromatography eluting with 1±20% MeOH/CH2Cl2
gave 24 (0.25 g, 45%).
(3Â300 mL). The organic extracts were combined, dried
over MgSO4, ®ltered and evaporated to give a crude
viscous oil (8 g). Silica gel chromatography a€orded 29
(1.65 g, 5%); mp 132±134 ꢀC and 30 (1.05 g, 3%); mp
132±139 ꢀC.
N-Methylpiperidine 29 (1.12 g, 3.43 mmol) in benzene
(20 mL) was treated with NEt3 (0.35 g, 3.46 mmol) and
trichloroethylchloroformate (1.2 g, 5.8 mmol). The mix-
ture was heated at re¯ux temperature for 1.5 h and
poured onto ice (50 mL). The mixture was treated with
excess concentrated NH4OH solution and extracted
with EtOAc (2Â20 mL). The combined organic extracts
were dried over MgSO4, ®ltered and evaporated to give
an oil (2.2 g). The oil was puri®ed by silica gel chroma-
tography to give the intermediate carbamate (1.45 g).
The carbamate in AcOH (10 mL) was treated with Zn
dust (1.5 g, 23 mmol) and heated at 80±100 ꢀC for 1 h.
The reaction mixture was ®ltered through a glass wool
plug and the residue washed with hot H2O (2Â20 mL).
The combined organic extracts were dried over MgSO4,
®ltered and evaporated to give an oil (1.2 g). This mate-
rial was dissolved in CH2Cl2 (25 mL), treated with Ac2O
(1.5 mL, 15.9 mmol) and heated at re¯ux temperature
for 1 h. The reaction mixture was poured onto ice, trea-
ted with excess concentrated NH4OH solution and the
organic layer separated. The aqueous was re-extracted
with CH2Cl2 (2Â25 mL) and the combined organic
portions dried over MgSO4, ®ltered and evaporated.
The crude oily product was puri®ed by silica gel chro-
matography eluting with 2±5% MeOH/CH2Cl2 to
a€ord 3 (0.98 g, 81%). Isomeric N-methylpiperidine 30
(2.9 g, 8.9 mmol) was converted to 11 (1.32 g, 72%) in
an analogous manner.
X-ray crystal structure analyses of compounds 20, 5, and
9. Crystal data for 20: [C18H21ClN+3 ][C8H14NO3 ],
M=487.05, monoclinic, a=17.740(2) A, b=5.783(1) A,
c=12.886(1) A, b=98.05(1)ꢀ, V=1309.0(5) A3, Z=2,
Dcalcd=1.236g cm 3, m(Cu-Ka radiation, l=1.5418 A)
=15.6 cm 1; space group P21(C22) from the Laue sym-
metry and systematic absences (0k0 when kˆ2n); crystal
dimensions: 0.06Â0.06Â0.60mm. Crystal data for 5:
C22H23ClN2O2, M=382.89, monoclinic, a=28.299(3) A,
b=4.993(1) A, c=29.137(3) A, b=109.19(1)ꢀ, V=3888(2)
A3, Z=8, Dcalcd=1.308gcm 3, m(Cu-Ka radiation)
=19.0 cm 1; space group Cc(Cs4) or C2/c(C26h) from the
Laue symmetry and systematic absences (hkl when
h+kˆ2n; h0l when lˆ2n), shown to be C2/c by structure
solution and re®nement; crystal dimensions: 0.03Â0.16
Â0.60mm. Crystal data for 9: C21H20ClN2O, M=351.86,
triclinic, a=10.057(1) A, b=10.578(1) A, c=9.367(1) A,
(‹)-1-Acetyl-4-(8-chloro-5,6-dihydro-11H-benzo[5,6]
cyclohepta[1,2-b]pyridin-11-yl)piperidine, (3) and (‹)-
1-acetyl-4-(10-chloro-5,6-dihydro-11H-benzo[5,6]cyclo-
hepta[1,2-b]pyridin-11-yl)piperidine, (11). Ketone 27 was
prepared from nitrile 26 via addition of N-methylpiper-
idine-4-magnesium chloride according to the procedure
of Schumacher et al.11 Ketone 27 (34.3 g, 100 mmol) in
EtOH (400 mL) was treated, portionwise, with NaBH4
(5.8 g, 153 mmol) and stirred at ambient temperature for
3 h. The solvent was evaporated under reduced pressure
and the residue partitioned between CH2Cl2 (200 mL)
and H2O (200 mL). The organic layer was separated and
the aqueous re-extracted with CH2Cl2 (3Â75 mL). The
combined organic portions were dried over MgSO4, ®l-
tered and evaporated to give alcohol 28 (34 g, 99%).
The alcohol was treated with polyphosphoric acid
(300 g), heated at 170±175 ꢀC for 3 h then poured onto
ice (2 L). This solution was treated with excess con-
centrated NH4OH solution and extracted with EtOAc
a=111.10(1)ꢀ, b=93.90(1)ꢀ, g=69.95(1)ꢀ V=871.6(3) A3,
3
Z=2, Dcalcd=1.341g cm
,
m(Cu-Ka radiation)=
20.3cm 1; space group P1(C11) or P1(C1i ) from the Laue
symmetry, shown to be P1 by structure solution and
re®nement; crystal dimensions: 0.40Â0.40Â0.40 mm.
Oscillation and Weissenberg photographs yielded pre-
liminary unit-cell parameters and space group informa-
tion for each crystal. Intensity data (+h, k,‹l, 2972
nonequivalent re¯ections for 20; +h,+k,‹l, 4062
re¯ections for 5, +h,‹k,‹l, 3791 re¯ections for 9) were

J. J. Kaminski et al. / Bioorg. Med. Chem. 7 (1999) 1413±1423
1423
recorded on an Enraf±Nonius CAD-4 di€ractometer
[Cu-Ka radiation, graphite monochromator; w-2q
scans, qmax=75ꢀ, scanwidths (1.15+0.14 tanq)ꢀ for 20,
(1.00+0.14 tanq)ꢀ for 5 and 9]; the intensities of four
reference re¯ections, remeasured every 2 h, showed no
signi®cant variation (<1%) throughout. Re®ned unit-
cell parameters were calculated in each case from the
di€ractometer setting angles for 25 re¯ections
(36ꢀ<q<40ꢀ for 20 and 9; 32ꢀ<q<36ꢀ for 5) widely
separated in reciprocal space. Intensity data were cor-
rected for the usual Lorentz and polarization e€ects. An
Crystallographic calculations were performed on
PDP11/44 and MicroVAX computers by use of the
Enraf±Nonius Structure Determination Package (SDP).
For all structure-factor calculations, neutral atom scat-
tering factors and their anomalous dispersion correc-
tions were taken from ref. 14.
Supplementary information available
ORTEP diagrams showing the crystallographic atom
numbering schemes, tables of fractional atomic coordi-
nates and temperature factor parameters, bond lengths,
bond angles, and torsion angles for 20, 5 and 9 (26
pages) are available from the authors.
empirical
absorption
correction
[Tmax:Tmin(rela-
tive)=1.00:0.80], based on the f-dependency of the inten-
sities of several re¯ections with c ca. 90ꢀ, was also applied
to the data for 5. Equivalent re¯ections for 5 and 9 were
averaged [Rmerge (on I)=0.03 for 5, 0.01 for 9] to yield 3973
and 3589 nonequivalent values. The structure analyses and
parameter re®nements were based on 1902, 1475, and 3289
re¯ections with I>3.0s(I) for 20, 5 and 9, respectively. All
three crystal structures were solved by direct methods. For
both 5 and 9, the centrosymmetric space group choices (P1
and C2/c, respectively) were assumed at the outset and
shown to be correct by the structure solutions and re®ne-
ments. E-maps yielded initial coordinates for all non-
hydrogen atoms in 20 and 9. For 20, the enantiomer was
selected to yield the known absolute stereochemistry of the
N-acetyl-d-leucyl anion. Approximate coordinates for the
non-hydrogen atoms of 5, other than those of the ethyl
moiety, were obtained from an E-map; positions for the
carbon atoms of the ethyl group, which was found to be
disordered over two orientations, were derived from a dif-
ference Fourier synthesis phased by the other atoms.
Atomic positional and thermal parameters (®rst isotropic
and then anisotropic) were adjusted by means of full-
matrix least-squares calculations during which SwD2
[w=1/s2(jFoj),d=(jFoj jFcj)] was minimized. Hydrogen
atoms in 5 and 9 (other than those of the disordered ethyl
group in 5) were located in di€erence Fourier syntheses,
and their positional and isotropic thermal parameters
were re®ned in addition to the non-hydrogen atom para-
meters in the subsequent least-squares iterations. For 20
and the ethyl group of 5, hydrogen atoms were incorpo-
rated at their calculated positions. An extinction correc-
tion (g) was included as a variable during the later least-
squares cycles for 20 and 9. The parameter re®nements
converged at R(=SjjFoj jFcjj)/SjFoj)=0.049, {Rw=
[Sw(jFoj jFcj)2/Sw(jFoj2]1/2=0.066, GOF=[SwD2/(N
References
1. Billah, M. M.; Chapman, R. W.; Egan, R. W.; Gilchrest,
H.; Piwinski, J. J.; Sherwood, J.; Siegel, M. I.; West, R. E., Jr.;
Kreutner, W. J. Pharmacol. Exp. Ther. 1990, 252, 1090.
2. Billah, M. M.; Egan, R. W.; Ganguly, A. K.; Green, M. J.;
Kreutner, W.; Piwinski, J. J.; Siegel, M. I.; Villani, F. J.;
Wong, J. K. Lipids 1991, 26, 1172.
3. Piwinski, J. J.; Wong, J. K.; Green, M. J.; Ganguly, A. K.;
Billah, M. M.; West, R. E., Jr.; Kreutner, W. J. Med. Chem.
1991, 34, 457.
4. Wong, J. K.; Piwinski, J. J.; Green, M. J.; Ganguly, A. K.;
Anthes, J. C.; Billah, M. M. Bioorg. Med. Chem. Lett. 1993, 3,
1073.
5. Billah, M. M.; Gilchrest, H. G.; Eckel, S.; Granzow, E.;
Radwanski, E.; A€rime, M.; Siegel, M. I. J. Allergy Clin.
Immunol. 1991, 87, 310.
6. Berkowitz, R.; Zora, J.; Tinkelman, D.; Cuss, F.; Danzig,
M.; Fourre, J. J. Allergy Clin. Immunol. 1991, 87, 167.
7. Green, M. J.; Piwinski, J. J.; Wong, J. K.; Albanese, M. A.;
Colizzo, F.; Kaminski, J. J.; Ganguly, A. K.; Billah, M. M.;
West, R. E., Jr.; Kreutner, W. E. Abstracts of Papers, 202nd
National Meeting of the American Chemical Society; Amer-
ican Chemical Society: Washington, D.C. 1991; MEDI 161.
8. Piwinski, J. J.; Wong, J. K.; Green, M. J.; Kaminski, J. J.;
Colizzo, F.; Albanese, M. A.; Ganguly, A. K.; Billah, M. M.;
Anthes, J. C.; West, R. E., Jr. Bioorg. Med. Chem. Lett. 1998,
8, 3469.
9. Villani, F. J.; Daniels, P. J. L.; Ellis, C. A.; Mann, T. A.;
Wang, K. C. J. Med. Chem. 1971, 8, 73.
10. Villani, F. J. U.S. Patent 3,326,924, 1967.
11. Schumacher, D. P.; Murphy, B. L.; Clark, J. E.; Tahbaz,
P.; Mann, T. A. J. Org. Chem. 1989, 54, 2242.
12. Young, S. D.; Baldwin, J. J.; Cochran, D. W.; Kin, S. W.;
Remy, D. C.; Springer, J. J. Org. Chem. 1985, 50, 339.
13. Kaminski, J. J.; Wong, S.-C.; Carruthers, N. I. U.S. Patent
5,648,360, 1997.
14. Hamilton, W. C.; Ibers, J. A., Eds. International Tables for
X-Ray Crystallography; The Kynoch Press: Birmingham,
England, 1974; Vol. IV.
Nparameters)]1/2=1.28, g=3.2 (5)Â10 6} for
observations
20; R=0.044 (Rw=0.055), GOF=1.41 for 5, R=0.038
5
(Rw=0.064), GOF=2.82, g=1.08(6)Â10 for 9. Final
di€erence Fourier syntheses contained no unusual fea-
tures [Dr(eA 3) max, min: 0.45, 0.19 (2); 0.13, 0.12
(5), 0.30, 0.21 (9)].