Scaffold hopping, synthesis and structure-activity relationships of 5,6-diaryl-pyrazine-2-amide derivatives: A novel series of CB1 receptor antagonists

Article abstract of DOI:10.1016/j.bmc.2007.03.075

A scaffold hopping approach has been exploited to design a novel class of cannabinoid (CB1) receptor antagonists for the treatment of obesity. On the basis of shape-complementarity and synthetic feasibility the central fragment, a methylpyrazole, in Rimonabant was replaced by a pyrazine. The synthesis and CB1 antagonistic activities of a new series of 5,6-diaryl-pyrazine-2-amide derivatives are described. Several compounds showed antagonist potency below 10 nM for the CB1 receptor.

Full text of DOI:10.1016/j.bmc.2007.03.075

Bioorganic & Medicinal Chemistry 15 (2007) 4077–4084
Scaffold hopping, synthesis and structure–activity relationships
of 5,6-diaryl-pyrazine-2-amide derivatives: A novel series
of CB1 receptor antagonists
Jonas Bostrom,a,* Kristina Berggren,a Thomas Elebring,b
¨
Peter J. Greasleya and Michael Wilstermanna
aLead Generation Department, AstraZeneca R&D Mo¨lndal, S-431 83 Mo¨lndal, Sweden
bMedicinal Chemistry Department, AstraZeneca R&D Mo¨lndal, S-431 83 Mo¨lndal, Sweden
Received 7 September 2006; revised 23 February 2007; accepted 27 March 2007
Available online 30 March 2007
Abstract—A scaffold hopping approach has been exploited to design a novel class of cannabinoid (CB1) receptor antagonists for the
treatment of obesity. On the basis of shape-complementarity and synthetic feasibility the central fragment, a methylpyrazole, in
Rimonabant was replaced by a pyrazine. The synthesis and CB1 antagonistic activities of a new series of 5,6-diaryl-pyrazine-2-amide
derivatives are described. Several compounds showed antagonist potency below 10 nM for the CB1 receptor.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
Sanofi-Aventis is currently launching their cannabinoid
CB1 receptor antagonist Rimonabant (AcompliaTM), to
treat obesity. However, there is a need for novel cannab-
inoid CB1 modulators with improved physicochemical
properties, DMPK and/or pharmacodynamic proper-
ties. Consequently, a scaffold hopping approach aiming
at replacing the central fragment, the methylpyrazole, in
Rimonabant was initiated. Several different compound
classes, such as thiazoles, pyrroles and pyrazines, were
scrutinized for their chemical feasibility and their
shape-complementarity to the putative bioactive confor-
mation of Rimonabant.4 Mini-libraries of around 20
compounds per substance class were synthesized to eval-
uate the different classes. Among the most interesting of
the classes was the pyrazine series.5 In the current study,
we present the synthesis and the CB1 antagonistic activ-
ities of a new series of 5,6-diaryl-pyrazine-2-amide
derivatives.
Cannabis is well known for stimulating appetite and the
short bursts of extreme hunger are commonly known as
the munchies. This characteristic effect of cannabis is be-
lieved to occur because D9-THC activates the cannabi-
noid CB1 receptor in the brain. The crux of the CB1
hypothesis is that antagonism of CB1 receptors would
have the opposite effect and it has been postulated that
cannabinoid CB1 modulators are useful in the treatment
of obesity.1–3 The cannabinoid CB1 receptor belongs to
the large superfamily of G-protein coupled receptors,
which have a seven transmembrane helix structure. As
is the case for most transmembrane proteins, cannabi-
noid receptors do not readily crystallize and no receptor
structures have yet been experimentally obtained.
Hence, in the quest for designing high-affinity CB1
receptor antagonists with a superior clinical profile for
the treatment of obesity a ligand-based strategy has to
be employed.
Cl
Cl
N
O
O
N
N
HN
N
N
H
N
N
Keywords: Scaffold hopping; Shape-matching; Cannabinoid;
Antagonist.
Cl
Cl
Cl
Cl
*
Corresponding author. Tel.: +46 31 706 52 51; fax: +46 31 776 37
10; e-mail: jonas.bostrom@astrazeneca.com
1: Rimonabant
2a
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.03.075

4078
J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
2. Results and discussion
energetically unrealistic conformations. The calculated
conformational energy penalties of the proposed bioac-
tive conformations for compounds 2a–b, 5a–f, 6a–e,
7a–e were obtained as described in the Section 4.3. The
calculated conformational energies are all low
(<3.0 kcal/mol) for the proposed bioactive conforma-
tions of 2a–b, 5a–f, 6a–e, 7a–e, see Figure 2. Such low
values are compatible with ligands showing high poten-
cies for the CB1 receptor.
2.1. Molecular design
A putative bioactive conformation of Rimonabant was
deduced on the basis of SAR, small molecule X-ray data
and conformational analysis. It largely agrees with the
later published proposal from Shim and co-workers.6
Taking into account that bioactive conformation, we
embarked on a scaffold hopping route steered by shape
similarity. With the aim of retaining the global shape of
Rimonabant, significant molecular modifications were
performed to obtain the designed compounds. Using
this strategy in combination with ease of synthesis, sev-
eral new scaffolds were ranked and selected. ROCS was
used to calculate the shape-overlays.7,8 Evidently, aro-
matic five- and six-membered rings are structurally dif-
ferent and do not, for example, share the same
optimal bond-angles. This was reflected in a slightly
lower shape overlay for the pyrazine containing com-
pound 2a, as compared to the other selected five-mem-
bered counterparts, the pyrrole (shape Tanimoto: 0.94)
and thiazole (shape Tanimoto: 0.92) derivatives. Never-
theless, the shape Tanimoto for 2a was high (0.90),
meaning that it is virtually identical to Rimonabant with
respect to shape. In addition, the shape-based alignment
shows that not only is the global shape retained, but it
also imposes the ring nitrogens to occupy the same re-
gion in space. The molecular alignment of 1 and the cor-
responding pyrazine compound 2a is shown in Figure 1.
Mini-libraries of the two other highly ranked scaffolds
(pyrrole and thiazole) were also synthesized. These series
have been reported upon elsewhere.9,10
2.2. Chemistry
Six commercially available symmetrical diketones of
varying size and electronic properties were selected to
gain knowledge on how the substitution pattern on the
phenyl rings affected affinity as compared to the substi-
tution pattern seen in the corresponding Rimonabant
derivatives. The symmetrical diketones were condensed
with monohydrochloride of 2,3-diaminopropionic acid
(3), Scheme 1, under basic conditions in refluxing meth-
anol.14 The reaction proved to be dependent upon the
ketone, and the reaction times varied between 20 min
and 16 h to yield the dihydrointermediates. Subsequent
to the condensation, air was bubbled through the reac-
tion mixture until the desired pyrazines, 4a–f, were
formed. We decided to make three different amides
and therefore compounds 4a–f were further coupled
with 1-aminopiperidine, cyclohexylamine and aniline,
respectively, by using EDC and DMAP in DCM at
room temperature overnight to give the final products
5a–f, 6a–e and 7a–e, respectively.
The synthesis of 2a and 2b is described in Scheme 2.
Friedel-Crafts acylation of dichlorobenzene 9 with 4-
chlorophenylacetyl chloride 8 yielded 10,15 which
was then oxidized by PCC in 1,2-dichloroethane to
give the 1,2-diketone 11. Condensation of 11 with
2,3-diaminopropionic acid (3) furnished the two
regioisomers 12a and 12b, which were transformed
into the corresponding acid chlorides by treatment
with thionyl chloride in toluene at reflux temperature
for 3 h. The acid chlorides were then reacted with 1-
aminopiperidine in DCM at room temperature
overnight yielding a mixture of 2a and 2b. The final
compounds 2a and 2b were purified by preparative
chromatography.
Moreover, conformational energy penalties may signifi-
cantly influence the affinity of a ligand.11 Thus, a stan-
dard approach to conformational analysis is to apply
conformational energy cut-offs.11–13 This pragmatic ap-
proach reduces the number of conformations in a calcu-
lated ensemble, while at the same time removing
2.3. Receptor functional studies
A GTPcS CB1 functional assay was used to assess the
potency of the pyrazine derivatives described above.
Receptor inhibition potencies for the compounds modi-
fied at the R1–R4 position, with different N-substituted
carboxamides, are reported in Table 1 (piperidylamide),
2 (cyclohexylamide) and 3 (aniline). For all three series
of carboxamides the di-para-substituted analogues (5a–
e, 6a–d and 7b–d) showed higher potency for CB1 than
the corresponding ortho-substituted derivatives (5f, 6e
and 7e) (Tables 2 and 3).
Figure 1. Molecular superimposition of compounds
1 and 2a,
illustrating the scaffold hopping. The methylpyrazole moiety in
Rimonabant (1) is exchanged for a pyrazine moiety in our compound
2a. The hydrogens have been removed for clarity. The shape Tanimoto
value is 0.9 meaning that they are virtually identical with respect to
shape. The carbon atoms and the molecular shape are visualized grey
for 1 and green for 2a.
The SAR of different N-substituted carboxamides at the
2-position of the pyrazine ring was also investigated. For
example, the bis(para-chloro-phenyl) N-piperidylamide

J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
4079
3.0
2.5
2.0
1.5
1.0
0.5
0.0
1
2a 2b 5a 5b 5c 5d 5e 5f 6a 6b 6c 6d 6e 7a 7b 7c 7d 7e
Ligand Number
Figure 2. Calculated conformational energy penalties of the proposed bioactive conformations of ligands 1, 2a–b, 5a–f, 6a–e, 7a–e. All ligands
display values less than 3 kcal/mol. However, the conformational energy penalties are slightly higher for the N-phenyl substituted compounds 7a–e,
providing a possible rationale for their somewhat lower CB1 receptor affinities.
R3
R4
R3
R4
ClH
O
R3
H2N
H2N
O
N
N
N
N
i, ii
iii
O
+
O
O
O
R4
R1
R2
OH
HN
R1
R2
R1
R2
R5
3
4a R1=R2=R3=R4=H
5a R1=R2=R3=R4=H, R5=1-piperidine
5b R1=R3=Br, R2=R4=H, R5=1-piperidine
5c R1=R3=Me, R2=R4=H, R5=1-piperidine
5d R1=R3=MeO, R2=R4=H, R5=1-piperidine
5e R1=R3=Cl, R2=R4=H, R5=1-piperidine
5f R1=R3=H, R2=R4=Cl, R5=1-piperidine
6a R1=R3=Br, R2=R4=H, R5=cyclohexyl
6b R1=R3=Me, R2=R4=H, R5=cyclohexyl
6c R1=R3=MeO, R2=R4=H, R5=cyclohexyl
6d R1=R3=Cl, R2=R4=H, R5=cyclohexyl
6e R1=R3=H, R2=R4=Cl, R5=cyclohexyl
7a R1=R2=R3=R4=H, R5=phenyl
4b R1=R3=Br, R2=R4=H
4c R1=R3=Me, R2=R4=H
4d R1=R3=MeO, R2=R4=H
4e R1=R3=Cl, R2=R4=H
4f R1=R3=H, R2=R4=Cl
7b R1=R3=Me, R2=R4=H, R5=phenyl
7c R1=R3=MeO, R2=R4=H, R5=phenyl
7d R1=R3=Cl, R2=R4=H, R5=phenyl
7e R1=R3=H, R2=R4=Cl, R5=phenyl
Scheme 1. Reagents and condition: (i) NaOH, MeOH, reflux; (ii) air bubbling; (iii) EDC, DMAP, DCM, rt.
Cl
Cl
O
Cl
Cl
Cl
ii
i
+
O
Cl
O
Cl
O
Cl
Cl
Cl
8
9
10
11
iii, iv
Cl
Cl
Cl
Cl
Cl
R3
R1
R4
N
N
N
N
v, vi
+
O
O
O
N
N
HN
HN
OH
Cl
R2
N
N
12 R1=R2=R3=Cl, R4=H
13 R1=R3=R4=Cl, R2=H
2b
2a
Scheme 2. Reagents and conditions: (i) AlCl3, neat, 0 ꢁC, Prt; (ii) PCC, pyridine, 1,2-dichloroethane, reflux; (iii) 3, NaOH, MeOH, reflux; (iv) air
bubbling; (v) SOCl2, toluene, reflux; (vi) 1-aminopiperidine, TEA, DCM, rt.
analogue (5e) gave the optimal potency for CB1 (IC50 of
4 nM) followed by the N-cyclohexylamide derivative
(6d), whereas the corresponding aromatic N-carboxam-
ide (7d) showed somewhat lower potency. This general
trend of lower potencies for the aromatic N-carboxamide
compounds (7a–e), as compared to the N-piperidylamide
(5a–f) and N-cyclohexylamide analogues (6a–e), may be
rationalized by their somewhat higher conformational
energy penalties (Fig. 2).11–13
Compound 5b, a piperidino derivative incorporating
bromine atoms at both the para-substituted phenyl
rings, was found to be the most potent antagonist for
the cannabinoid CB1 receptor (IC50: 1 nM), whereas

4080
J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
Table 1. Structure and activity for compounds with N-piperidylamide
substituents
the unsubstituted aromatic carboxamide (7a) displayed
the poorest activity (IC50: 2790 nM).
R4
In our in vitro functional assays, the pyrazine derivatives
show common characteristics in SAR with that previ-
ously reported for the pyrazole series,4 providing support
for the shape-based molecular alignment shown in Figure
1. For example, the bis(para-phenyl) pyrazoles substi-
tuted analogues to ligand 1 also display significantly high-
er potency compared with the bis(ortho)phenyl
substituted pyrazoles. Furthermore, the aromatic carbox-
amide pyrazole analogues also showed decreased po-
tency.4 This suggests a common mode of binding for the
pyrazole and pyrazine series at cannabinoid receptors.
R3
N
O
N
HN
N
R2
R1
Compound
R1
R2 R3
Cl Cl
R4 GTPcS IC50 values (nM)
1 (Rimonbant)
3
2a
2b
5a
5b
5c
5d
5e
5f
Cl
Cl
H
Cl
H
H
H
H
H
114
10
766
1
H
H
H
H
H
H
Cl
Cl
H
Br
H
Br
Recently, several other CB1 receptor antagonists, with
variations in the central scaffold, have been published.
For example, the pyrazole ring in 1 has been exchanged
with the corresponding thiazole,9 pyrrole10 and pyri-
dine.16,17 All these classes include ligands with high poten-
cies for the CB1 receptor. Since these fragments are
electronically quite different we believe that the central
scaffold is mainly essential for geometrical reasons. That
is, the function of the central scaffold is not to make any di-
rect interactions with the cannabinoid receptor, but rather
to place the substituted phenyl rings and the N-carboxya-
mide fragments in anoptimal 3D orientation. This hypoth-
esis is in line with previous work in the CB1 field.18
Me
OMe
Cl
Me
OMe
Cl
6
21
4
H
H
Cl 174
The experimental potency data are given as IC50 values, for the CB1
antagonists.
Table 2. Structure and activity for compounds with N-cyclohexyla-
mide substituents
R4
R3
N
O
N
3. Conclusions
HN
R2
A scaffold hopping approach has been exploited for the
design of a novel class of CB1 antagonists for the treat-
ment of obesity. The synthesis and CB1 antagonistic
activities of a new series of 5,6-diaryl-pyrazine-2-amide
derivatives have been described. Several compounds
showed in vitro potency below 10 nM for the CB1 recep-
tor. Bis para-substitution improves potency as compared
with bis ortho-substituted pyrazines. We believe that the
scaffold hopping was successful because the global shape
was preserved and that the central scaffold seems mainly
to be essential for geometric reasons.
R1
Compound R1
R2 R3
R4 GTPcS IC50 values (nM)
6a
6b
6c
6d
6e
Br
H
H
H
H
Cl
Br
H
H
H
H
11
13
8
Me
OMe
Cl
Me
OMe
Cl
19
H
H
Cl 1578
The experimental potency data are given as IC50 values, for the CB1
antagonists.
Table 3. Structure and activity for compounds with N-phenyl substit-
uents
4. Experimental
R4
4.1. General synthetic methods
R3
N
Mass spectra were recorded on either a Micromass ZQ
single quadrupole or a Micromass LCZ single quadrupole
mass spectrometer both equipped with a pneumatically
assisted electrospray interface (LC-MS). 1H NMR mea-
surements were performed on either a Varian Mercury
O
N
HN
R2
R1
1
300 or a Varian Inova 500, operating at H frequencies
Compound R1
R2 R3
R4 GTPcS IC50 values (nM)
of 300 and 500 MHz, respectively. Chemical shifts are gi-
ven in ppm with CDCl3 as internal standard. Purification
was performed on a semipreparative HPLC with a mass
triggered fraction collector, Shimadzu QP 8000 single
quadrupole mass spectrometer equipped with
19 · 100 mm C8 column. The mobile phase used was, if
nothing else stated, acetonitrile and buffer (0.1 M
NH4Ac:acetonitrile 95/5).
7a
7b
7c
7d
7e
H
H
H
H
H
Cl
H
H
H
H
H
Cl
2790
326
119
80
Me
OMe
Cl
Me
OMe
Cl
H
H
763
The experimental potency data are given as IC50 values, for the CB1
antagonists.

J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
4081
For isolation of isomers
a
kromasil-CN E9344
MS m/z 335 (M+H)+.
(250 · 20 mm id) column was used. As mobile phase
heptane:ethyl acetate:DEA 95/5/0.1 was used. Fractions
were collected by UV.
4.1.5. 5,6-Bis-(4-chloro-phenyl)-pyrazine-2-carboxylic
acid (4e). The compound 4e was prepared as described
in Section 4.1.3 using 4,40-dichlorobenzil (993 mg,
3.56 mmol). Reflux for 1 h gave the oxidised pyrazine di-
rectly. The crude product (923 mg, 75%) was used in
steps described below without further purification.
4.1.1. 5,6-Diphenyl-pyrazine-2-carboxylic acid (4a).
Compound 3 (500 mg, 3.56 mmol) and benzil (890 mg,
4.23 mmol) were added to a solution of sodium hydrox-
ide (677 mg, 16.93 mmol) in methanol (10 mL). An extra
portion of methanol was added (5 mL) and the reaction
mixture was refluxed for 20 min. The mixture was
cooled to 25 ꢁC and air was bubbled through for
30 min. Hydrochloric acid (aq, 2 M) was added until
the reaction mixture reached pH 2. The solution was ex-
tracted with diethyl ether. The combined diethyl ether
phases were dried (MgSO4), filter and the solvent was
evaporated under reduced pressure to give the crude
product. The crude product was used in steps described
below without further purification.
MS m/z 343, 345, 347 (MꢀH)ꢀ.
4.1.6.
5,6-Bis-(2-chloro-phenyl)-pyrazine-2-carboxylic
acid (4f). The title compound was prepared as de-
scribed in Section 4.1.3 using 2,20-dichlorobenzil
(993 mg, 3.56 mmol). The crude product (895 mg,
73%) was used in steps described below without further
purification.
MS m/z 343, 345, 347 (MꢀH)ꢀ.
MS m/z 277 (M+H)+.
4.1.7. 5,6-Diphenyl-pyrazine-2-carboxylic acid piperidin-
1-ylamide (5a). Compound 4a (500 mg, 1.81 mmol) was
dissolved in DCM (4 mL) and DMF (150 lL). DMAP
(22 mg, 0.18 mmol) and 1-aminopiperidine (218 mg,
2.17 mmol) were added and the solution was cooled to
0 ꢁC. A slurry of EDC (1.99 mmol, in 2 mL DCM and
100 lL DMF) was added dropwise. The reaction mix-
ture was stirred at 25 ꢁC. After 17 h, additional 1-amin-
opiperidin (40 mg, 0.40 mmol) and EDC (76 mg,
0.40 mmol) were added and the mixture was stirred for
another 3 h. The crude was diluted with DCM (5 mL)
and washed with a saturated solution of NaHCO3.
The organic phase was dried (MgSO4), filtered and the
solvent was evaporated. Flash chromatography (SiO2,
ethyl acetate:hexane 2/1) gave the title compound
(160 mg, 25%) as a white solid.
4.1.2.
acid (4b). The title compound was prepared essentially
as described in Section 4.1.1, using (600 mg,
5,6-Bis-(4-bromo-phenyl)-pyrazine-2-carboxylic
3
4.26 mmol) and 4,40-dibromobenzil (1.745 g, 4.26 mmol,
90%) as starting materials. The reaction mixture was re-
fluxed for 2 h and air was bubbled through for 1 h.
Hydrochloric acid (aq, 2 M) was added until reaching
pH 2. The mixture was evaporated under reduced pres-
sure and the residue was dissolved in water. The solution
was extracted with diethyl ether, the combined diethyl
ether phases were dried (MgSO4), filter and the solvent
was evaporated under reduced pressure. The crude
product (500 mg, 27%) was used in steps described be-
low without further purification.
MS m/z 435, 437, 439 (M+H)+.
1H NMR (300 MHz) d 9.41 (s, 1H), 8.52 (s, 1H), 7.50–
7.29 (m, 10H), 2.94 (t, 4H), 1.81 (m, 4H), 1.50 (m,
2H). MS m/z 359 (M+H)+.
4.1.3. 5,6-Di-p-tolyl-pyrazine-2-carboxylic acid (4c). The
compound 4c was prepared as described in Section 4.1.1
using 4,40-dimethylbenzil (848 mg, 3.56 mmol). The
reaction mixture was however refluxed for 1 h and air
was bubbled through the reaction mixture for about
7 h. The mixture was evaporated and the residue was
dissolved water. Hydrochloric acid (aq, 2 M) was added
until reaching pH 2. The solution was extracted with
diethyl ether, the combined diethyl ether phases were
dried (MgSO4), filter and the solvent was evaporated un-
der reduced pressure. The crude product (918 mg, 85%)
was used in steps described below without further
purification.
4.1.8.
5,6-Bis-(4-bromo-phenyl)-pyrazine-2-carboxylic
acid piperidin-1-ylamide (5b). To compound 4b
(108 mg, 0.25 mmol), DMAP (0.025 mmol, in 500 lL
DCM), 1-aminopiperidine (0.25 mmol, in 1100 lL
DCM), and EDC (0.27 mmol, in 1100 lL DCM and
cooled to 8 ꢁC) were added. The reaction mixture was
stirred at 25 ꢁC for 20 h, then washed with saturated
NaHCO3 solution, dried (MgSO4), filtered and the sol-
vent was evaporated. Semipreparative HPLC (0.01%
TEA in the buffer phase) gave the title compound
(6.7 mg, 5.4%).
MS m/z 305 (M+H)+.
1H NMR (300 MHz) d 9.41 (s, 1H), 8.48 (s, 1H), 7.54 (d,
2H), 7.51 (d, 2H), 7.36 (d, 2H), 7.34 (d, 2H), 2.94 (t, 4H),
1.81 (m, 4H), 1.55–1.45 (m, 2H).
4.1.4. 5,6-Bis-(4-methoxy-phenyl)-pyrazine-2-carboxylic
acid (4d). The title compound was prepared as described
in Section 4.1.3 using 4,40-dimethoxybenzil (961 mg,
3.56 mmol) as starting material. The reaction mixture
was refluxed overnight and air was bubbled through
the mixture for 8 h. The crude product (435 mg, 36%)
was used in steps described below without further
purification.
MS m/z 515, 517, 518 (M+H)+.
4.1.9. 5,6-Di-p-tolyl-pyrazine-2-carboxylic acid piperidin-
1-ylamide (5c). Compound 4c (76 mg, 0.25 mmol) was
used as described in Section 4.1.8 to give the title com-
pound (27 mg, 28%).

4082
J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
1H NMR (300 MHz) d 9.35 (s, 1H), 8.57 (s, 1H), 7.38 (d,
4H), 7.18 (d, 2H), 7.13 (d, 2H), 2.92 (t, 4H), 2.40 (s, 3H),
2.37 (s, 3H), 1.86–1.75 (m, 4H), 1.54–1.44 (m, 2H).
MS m/z 386 (M+H)+.
4.1.15. 5,6-Bis-(4-methoxy-phenyl)-pyrazine-2-carboxylic
acid cyclohexylamide (6c). Compound 4d (76 mg,
0.25 mmol) was used essentially as described in Section
4.1.13 but the reaction mixture was first stirred over-
night, then more cyclohexylamine (25 mg, 0.25 mmol)
was added and the mixture was stirred for another
two days prior to workup. Semipreparative HPLC
(0.15% TFA in the buffer phase) gave the title com-
pound (12 mg, 11%).
MS m/z 387 (M+H)+.
4.1.10. 5,6-Bis-(4-methoxy-phenyl)-pyrazine-2-carboxylic
acid piperidin-1-ylamide (5d). Compound 4d (84 mg,
0.25 mmol) was used as described in Section 4.1.8 to give
the title compound (20 mg, 19%).
1H NMR (300 MHz) d 9.31 (s, 1H), 8.57 (s, 1H), 7.46 (d,
2H), 7.44 (d, 2H), 6.90 (d, 2H), 6.86 (d, 2H), 3.86 (s,
3H), 3.84 (s, 3H), 2.93 (t, 4H), 1.80 (m, 4H), 1.54–1.45
(m, 2H).
1H NMR (300 MHz) d 9.32 (s, 1H), 7.76 (d, 1H), 7.47
(d, 2H), 7.45 (d, 2H), 6.90 (d, 2H), 6.86 (d, 2H), 4.10–
3.96 (m, 1H), 3.86 (s, 3H), 3.84 (s, 3H), 2.09–1.17 (m,
10H).
MS m/z 419 (M+H)+.
MS m/z 418 (M+H)+.
4.1.11. 5,6-Bis-(4-chloro-phenyl)-pyrazine-2-carboxylic acid
piperidin-1-ylamide (5e). Compound 4e (86 mg, 0.25
mmol) was used as described in Section 4.1.8 to give
the title compound (16 mg, 15%).
4.1.16.
5,6-Bis-(4-chloro-phenyl)-pyrazine-2-carboxylic
acid cyclohexylamide (6d). Compound 4e (86 mg,
0.25 mmol) was used as described in Section 4.1.15 to give
the title compound (7 mg, 8%) after washing with Na2CO3.
1H NMR (300 MHz) d 9.40 (s, 1H), 8.49 (s, 1H), 7.45–7.31
(m, 8H), 2.94 (t, 4H), 1.80 (m, 4H), 1.54–1.45 (m, 2H).
1H NMR (300 MHz) d 9.41 (s, 1H), 7.69 (s, 1H), 7.47–
7.30 (m, 8H), 4.10–3.97 (m, 1H), 2.10–1.18 (m, 10H).
MS m/z 427, 429, 431 (M+H)+.
MS m/z 426, 428, 430 (M+H)+.
4.1.12. 5,6-Bis-(2-chloro-phenyl)-pyrazine-2-carboxylic
acid piperidin-1-ylamide (5f). Compound 4f (86 mg,
0.25 mmol) was used as described in Section 4.1.8 to give
the title compound (6 mg, 6%).
4.1.17. 5,6-Bis-(2-chloro-phenyl)-pyrazine-2-carboxylic
acid cyclohexylamide (6e). Compound 4f (86 mg,
0.25 mmol) was used as described in Section 4.1.15 to
give the title compound (14 mg, 13%).
1H NMR (300 MHz) d 9.52 (s, 1H), 8.52 (s, 1H), 7.44–
7.17 (d, 8H), 2.94–2.88 (t, 4H), 1.85–1.70 (m, 4H),
1.52–1.44 (m, 2H).
1H NMR (300 MHz) d 9.51 (s, 1H), 7.74 (s, 1H), 7.41–
7.18 (m, 8H), 4.10–3.97 (m, 1H), 2.07–1.14 (m, 10H).
MS m/z 427, 429, 431 (M+H)+.
MS m/z 426, 428, 430 (M+H)+.
4.1.13. 5,6-Bis-(4-bromo-phenyl)-pyrazine-2-carboxylic
acid cyclohexylamide (6a). Compound 4b (109 mg,
0.25 mmol) was reacted essentially as described in Section
4.1.8 but with cyclohexylamine (0.25 mmol in 1 mL
DCM) instead. DMF (100 lL) was also added. Semipre-
parative HPLC (0.15% TFA/water:acetonitrile 95/5 in-
stead of the buffer phase) gave the title compound
(7 mg, 8%) after washing with Na2CO3.
4.1.18. 5,6-Diphenyl-pyrazine-2-carboxylic acid phenyla-
mide (7a). Compound 4a (70 mg, 0.25 mmol) was used
as described in Section 4.1.13 but aniline (0.25 mmol
in 1 mL DCM) was used instead. The reaction mixture
was worked up as described in Section 4.1.8. Semipre-
parative HPLC (0.15% TFA/water:acetonitrile 95/5 in-
stead of the buffer phase) gave the title compound
(27 mg, 30%) after washing with Na2CO3.
1H NMR (300 MHz) d 9.41 (s, 1H), 7.68 (s, 1H), 7.54 (d,
2H), 7.50 (d, 2H), 7.36 (d, 2H), 7.34 (d, 2H), 4.11–3.96
(m, 1H), 2.12–1.20 (m, 10H).
1H NMR (300 MHz) d 9.75 (s, 1H), 9.52 (d, 1H), 7.80
(d, 2H), 7.55–7.32 (m, 12H), 7.20 (t, 1H).
MS m/z 352 (M+H)+.
MS m/z 514, 516, 518 (M+H)+.
4.1.19. 5,6-Di-p-tolyl-pyrazine-2-carboxylic acid phenyla-
mide (7b). Compound 4c (77 mg, 0.25 mmol) was used
as described in Section 4.1.18 to give the title compound
(28 mg, 29%).
4.1.14. 5,6-Di-p-tolyl-pyrazine-2-carboxylic acid cyclo-
hexylamide (6b). Compound 4c (76 mg, 0.25 mmol) was
used as described in Section 4.1.13. Semipreparative
HPLC (0.01% TEA in the buffer phase) gave the title com-
pound (4 mg, 4%).
1H NMR (500 MHz) d 9.78 (s, 1H), 9.49 (s, 1H), 7.81 (d,
2H), 7.47–7.43 (m, 6H), 7.25–7.17 (m, 5H), 2.45 (s, 3H),
2.41 (s, 3H).
1H NMR (300 MHz) d 9.36 (s, 1H), 7.77 (d, 1H), 7.39
(d, 4H), 7.18 (d, 2H), 7.13 (d, 2H), 4.10–3.96 (m, 1H),
2.40 (s, 3H), 2.37 (s, 3H), 2.09–1.20 (m, 10H).
MS m/z 380 (M+H)+.

J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
4083
4.1.20. 5,6-Bis-(4-methoxy-phenyl)-pyrazine-2-carboxylic
acid phenylamide (7c). Compound 4d (85 mg,
0.25 mmol) was used as described in Section 4.1.18 to
give the title compound (33 mg, 32%).
1H NMR (500 MHz) d 7.97 (d, 2H), 7.84 (d, 1H), 7.52
(d, 2H), 7.46 (s, 1H), 7.44 (d, 1H).
4.1.25. 5-(4-Chloro-phenyl)-6-(2,4-dichloro-phenyl)-pyra-
zine-2-carboxylic acid (12a) and 6-(4-chloro-phenyl)-5-
(2,4-dichloro-phenyl)-pyrazine-2-carboxylic acid (12b).
The title compounds were prepared as described in
(4a) using 11 (1.85 g, 5.90 mmol) and 3 (0.61 g,
5.90 mmol) as starting materials. The mixture was re-
fluxed for 30 min and then directly worked up. The
crude product was allowed to stand overnight to sponta-
neously aromatize. Flash chromatography (SiO2,
DCM:methanol 10/1, 1% acetic acid) gave the isomer
mixture (0.2 g, 10%). MS m/z 377, 379, 381 (MꢀH)ꢀ.
1H NMR (300 MHz) d 9.74 (s, 1H), 9.42 (s, 1H), 7.79 (d,
2H), 7.50 (d, 4H), 7.42 (t, 2H), 7.19 (t, 1H), 6.94 (d, 2H),
6.89 (d, 2H), 3.88 (s, 3H), 3.85 (s, 3H).
MS m/z 412 (M+H)+.
4.1.21. 5,6-Bis-(4-chloro-phenyl)-pyrazine-2-carboxylic
acid phenylamide (7d). Compound 4e (87 mg,
0.25 mmol) was used as described in Section 4.1.18 to
give the title compound (6 mg, 6%).
4.1.26.
5-(4-Chloro-phenyl)-6-(2,4-dichloro-phenyl)-
1H NMR (300 MHz) d 9.66 (s, 1H), 9.52 (s, 1H), 7.79 (d,
2H), 7.48–7.35 (m, 10H), 7.21 (t, 1H).
pyrazine-2-carboxylic acid piperidin-1-ylamide (2a)
and 6-(4-chloro-phenyl)-5-(2,4-dichloro-phenyl)-pyra-
zine-2-carboxylic acid piperidin-1-ylamide (2b). The mix-
ture of 12a and 12b (78 mg, 0.205 mmol) and thionyl
chloride (147 mg, 1.23 mmol) was refluxed in toluene
(2 mL) for 3 h. The solvent and the excess of reagent were
evaporated under reduced pressure. The intermediates
were dissolved in DCM (1 mL). TEA (42 mg, 0.41 mmol)
and 1-aminopiperidine (21 mg, 0.205 mmol) were both
dissolved in DCM (1 mL) and added. The reaction mix-
ture was stirred at 25 ꢁC overnight. The solvent was evap-
orated under reduced pressure and the crude product was
directly purified by flash chromatography (SiO2, hep-
tane:ethyl acetate 1/1) which gave the isomers (45 mg,
MS m/z 420, 422, 424 (M+H)+.
4.1.22. 5,6-Bis-(2-chloro-phenyl)-pyrazine-2-carboxylic
acid phenylamide (7e). Compound 4f (87 mg, 0.25 mmol)
was treated as described in Section 4.1.18 to give the title
compound (27 mg, 25%).
1H NMR (500 MHz) d 9.73 (s, 1H), 9.66 (s, 1H), 7.81(d,
2H), 7.46–7.22 (m, 11H).
MS m/z 420, 422, 424 (M+H)+.
1
47%). H NMR (300 MHz) d 9.46 (s, 1H), 8.39 (s, 1H),
4.1.23. 2-(4-Chloro-phenyl)-1-(2,4-dichloro-phenyl)-etha-
none (10). To a dried round-bottomed flask, AlCl3
(4.94 g, 37.0 mmol) and (11) (38.9 g, 264.5 mmol) were
added under nitrogen. The slurry was cooled to 0 ꢁC
by an ice bath. (10) (5.0 g, 26.5 mmol) was added drop-
wise during 10 min. The ice bath was removed and the
reaction mixture was left to warm to room temperature
while stirring during night. The excess of reagent was re-
moved by rotatory evaporation under reduced pressure.
To the remaining, ice (100 mL) in HCl (concd 10 mL)
was added and the aqueous phase was extracted with
DCM. The organic phase was washed with NaOH
(2 M, aq), brine, water, dried (MgSO4), filtered and
the solvent was evaporated. Flash chromatography
(SiO2, toluene:heptane 1/1) gave the title compound
(3.12 g, 39%).
7.47–7.28 (m, 7H), 3.02–2.84 (m, 4H), 1.89–1.73 (m,
4H), 1.57–1.41 (m, 2H) and 9.42 (s, 1H), 8.51 (s, 1H),
7.47–7.28 (m, 7H), 3.02–2.84 (m, 4H), 1.89–1.73 (m,
4H), 1.57–1.41 (m, 2H).
4.1.27. Compound 2a was isolated from its isomer by
preparative chromatography (9 mg). 1H NMR (300 MHz)
d 9.46 (s, 1H), 8.38 (s, 1H), 7.46–7.24 (m, 7H), 2.89 (t, 4H),
1.78 (p, 4H), 1.52–1.40 (m, 2H).
4.1.28. Compound 2b was isolated from its isomer by
preparative chromatography (11 mg). 1H NMR
(300 MHz) d 9.42 (s, 1H), 8.50 (s, 1H), 7.39–7.30 (m,
7H), 2.93 (t, 4H), 1.80 (p, 4H), 1.54–1.43 (m, 2H).
4.2. GTPcS assay
1H NMR (500 MHz) d 7.46 (d, 1H), 7.38 (d, 1H), 7.28–
7.34 (m, 3H), 7.17 (d, 2H), 4.23 (s, 2H).
[35S] GTPcS binding assays were performed at 30 ꢁC for
45 min in membrane buffer (100 mM NaCl, 5 mM,
1 mM EDTA, and 50 mM HEPES, pH 7.4) containing
0.025 lg/lL of membrane protein with 0.01% bovine
serum albumin (fatty acid free), 10 lM GDP, 100 lM
DTT and 0.53 nM [35S] GTPcS (Amersham Pharmacia
Biotech) in a final volume of 200 lL. Non specific bind-
ing was determined in the presence of 20 lM GTPcS.
For antagonist experiments the EC80 for CP55940 was
determined and used to activate the receptor. The reac-
tion was terminated by addition of ice cold wash buffer
(50 mM Tris–HCl, 5 mM MgCl2, and 50 mM NaCl, pH
7.4) followed by rapid filtration under vacuum through
Printed Filtermat A glass fibre filters (Wallac) (0.05%
PEI treated) using a Micro 96 Harvester (Skatron
MS m/z 297, 299, 301, 303 (M+H)ꢀ.
4.1.24. 1-(4-Chloro-phenyl)-2-(2,4-dichloro-phenyl)-eth-
ane-1,2-dione (11). Compound 10 (2.7 g, 9.01 mmol)
was dissolved in 1,2-dichloroethane (25 mL) and freshly
made PCC (3.89 g, 18.02 mmol), pyridine (1.43 g,
18.02 mmol) and molecular sieves were added. The reac-
tion mixture was refluxed under inert atmosphere over-
night. The solution was cooled to 25 ꢁC, filtered through
silica and the solvent was evaporated under reduced
pressure. The crude product (1.9 g, 66%) was used di-
rectly in the next step.

4084
J. Bostro¨m et al. / Bioorg. Med. Chem. 15 (2007) 4077–4084
Instruments). The filters were dried for 30 min at 50 ꢁC,
then a paraffin scintillate pad was melted onto the filters
and the bound radioactivity was determined using a
1450 Microbeta Trilux (Wallac) scintillation counter.
Supplementary data
Supplementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmc.2007.03.075.
4.3. Computational methods
Tanimoto values were calculated to assess the shape sim-
ilarity of compounds 2a–b, 5a–f, 6a–e, 7a–e as compared
to the proposed bioactive conformation of Rimonabant.
The shape complementary program ROCS8 was used.
Shape Tanimoto values above 0.8 correspond to struc-
tures of visually very similar shape. The shape Tanimoto
values were calculated as follows. A multi conformational
database was generated for compounds 2a–b, 5a–f, 6a–e,
7a–e according to standard procedure.19 That is, the
SMILES codes were converted to 3D by using Corina.20
The structures obtained from Corina were subjected to ra-
pid geometry optimization (25 iterations) using the
MMFF94s force field as implemented in Szybki.21 A con-
formational ensemble was generated using the OMEGA
program.22 The compounds were aligned onto the pro-
posed bioactive conformation of Rimonabant using
ROCS. The bioactive conformation for each compound
was deduced on the basis of the highest shape Tanimoto.
References and notes
1. Pertwee, R. G. Addict. Biol. 2000, 5, 37.
2. Gomez, R.; Navarro, M.; Ferrer, B.; Trigo, J. M.; Bilbao,
A.; Del Arco, I.; Cippitelli, A.; Nava, F.; Piomelli, D.; De
Fonseca, F. R. J. Neurosci. 2002, 22, 9612.
3. Reggio, P. H. In Biology of Marijuana—From Gene to
Behavior; Onaivi, E. S., Ed.; Taylor & Francis, 2002; pp
449–490.
4. Lan, R.; Liu, Q.; Fan, P.; Lin, S.; Fernando, S. R.;
McCallion, D.; Pertwee, R.; Makriyannis, A. J. Med.
Chem. 1999, 42, 769.
5. Berggren, A. I. K.; Bostro¨m, S. J.; Elebring, S. T.;
Greasley, P.; Terricabras, E.; Wilstermann, J. M. WO
2003051851, 2003.
6. Shim, J. Y.; Welsh, W. J.; Cartier, E.; Edwards, J. L.;
Howlett, A. C. J. Med. Chem. 2002, 45, 1447.
7. Grant, J. A.; Gallardo, M. A.; Pickup, B. T. J. Comput.
Chem. 1996, 17, 1653.
8. ROCS is available from OpenEye Science Software, 3600
Cerrillos Rd., Suite 1107 Santa Fe, NM 87507, USA.
9. Berggren, A. I. K.; Bostro¨m, S. J.; Elebring, T. S.;
Fa¨llefors, L.; Wilstermann, J. M.; Greasley, P. WO
2004058255, 2004.
The affinity of a ligand may be significantly influenced by
the conformational energy component.11 Thus, conforma-
tional energy penalties were calculated, as follows. First,
the global minimum for each ligand was obtained from
an exhaustive conformational analysis using the Mixed
Torsional/Low Mode search, implemented in MacroMod-
el v9.0.23 The MMFFs force field24,25 and the Generalized
Born/Surface Area (GB/SA)26 solvent model was used.
Second, for each ligand the structure obtained by shape-
matching was energy minimized with ‘flat bottomed’
Cartesian constraints. This procedure is to make bond
lengths and bond-angles consistent with the MMFFs force
field, ensuring that deviations in these terms do not make
spurious contributions to the conformational energy.11 Fi-
nally the conformational energy penalty was calculated
from the difference in the global energy minimum and the
conformational energy after constrained optimization.
10. Berggren, A. I. K.; Bostro¨m, S. J.; Cheng, L.; Elebring, S.
˚ ˚
T.; Greasley, P.; Nagard, M.; Wilstermann, J. M.;
Terricabras, E. WO 2004058249, 2004.
11. Bostro¨m, J.; Norrby, P. O.; Liljefors, T. J. Comput. Aided
Mol. Des. 1998, 12, 383.
12. Nicklaus, M. C.; Wang, S.; Driscoll, J. S.; Milne, G. W.
Bioorg. Med. Chem. 1995, 3, 411.
13. Perola, E.; Charifson, P. S. J. Med. Chem. 2004, 47, 2499.
`
14. Felder, E.; Pitre, D.; Boveri, S.; Grabitz, E. B. Chem. Ber.
1967, 100, 555.
15. Bonadies, F.; Bonini, C. Synth. Commun. 1988, 18, 1573.
16. Finke, P. E.; Meurer, L. C.; Debenham, J. S.; Toupence,
R. B.; Walsh, T. F. WO 2003082191, 2003.
17. Barth, F.; Martinez, S.; Rinaldi Carmona, M. WO
2003084930, 2003.
18. Lange, J. H. M.; Kruse, C. G. Curr. Opin. Drug Discov.
2004, 7, 498.
To facilitate new ideas and the chemist’s awareness of 3D
the above-described superposition procedure was auto-
mated and an interactive web interface was designed, using
dynamic HTML documents. For an input structure, con-
formation ensembles were generated by OMEGA, ROCS
was used to identify the most similar conformations
accordingtoshapeandthealignedstructuresandtheshape
Tanimoto similarity value were reported.
19. Bostro¨m, J.; Greenwood, J. R.; Gottfries, J. J. Mol.
Graph. Model. 2003, 21, 449.
20. CORINA Molecular Networks, GmbH Computerchemie
Langemarckplatz 1, Erlangen, Germany, 2000.
21. SZYBKI is available from OpenEye Science Software,
3600 Cerrillos Rd., Suite 1107 Santa Fe, NM 87507,
USA.
22. OMEGA is available from OpenEye Science Software,
3600 Cerrillos Rd., Suite 1107 Santa Fe, NM 87507, USA.
23. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Lipton, M.; Caufield, C.; Chang, G.;
Hendrikson, T.; Still, W. C. J. Comput. Chem. 1990, 11,
440 (http://www.schrodinger.com).
Acknowledgments
The authors would like to thank Maria Lindskog, for
purification work, Christina Lo¨wendahl and Hanna
Nelander for separating isomers 2a and 2b and Gunnar
Gro¨nberg for structrual elucidation of isomers 2a and 2b.
24. Halgren, T. A. J. Comput. Chem. 1996, 17, 490.
25. Halgren, T. A. J. Comput. Chem. 1999, 20, 720.
26. Still, W. C.; Tempczyk, A.; Hawley, R. C.; Hendrickson,
T. J. Am. Chem. Soc. 1990, 112, 6127.