Design and discovery of 1,3-benzodiazepines as novel dopamine antagonists

Article abstract of DOI:10.1016/j.bmcl.2009.07.012

A series of novel 1,3-benzodiazapine based D1 antagonists was designed according to the understanding of pharmacophore models derived from SCH 23390 (1b), a potent and selective D1 antagonist. The new design features an achiral cyclic-amidine that maintai

Full text of DOI:10.1016/j.bmcl.2009.07.012

Bioorganic & Medicinal Chemistry Letters 19 (2009) 5218–5221
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Design and discovery of 1,3-benzodiazepines as novel dopamine antagonists
a
a
a
a
Zhaoning Zhu a, , Zhong-Yue Sun , Yuanzan Ye , Brian McKittrick , William Greenlee ,
*
Michael Czarniecki a, Ahmad Fawzi b, Hongtao Zhang b, Jean E. Lachowicz b
a Department of Medicinal Chemistry, Schering Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
b Department of CNS biology, Schering Plough Research Institute, 2015 Galloping Hill Road, Kenilworth, NJ 07033, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of novel 1,3-benzodiazapine based D1 antagonists was designed according to the understanding
of pharmacophore models derived from SCH 23390 (1b), a potent and selective D1 antagonist. The new
design features an achiral cyclic-amidine that maintains desired basicity. Solid phase synthesis was
developed for SAR development of the novel dopamine antagonists.
Received 28 May 2009
Revised 23 June 2009
Accepted 2 July 2009
Available online 9 July 2009
Ó 2009 Elsevier Ltd. All rights reserved.
Keywords:
Dopamine
D1 receptor antagonist
Benzodiazepine
The neurotransmitter dopamine plays important roles in neuro-
nal functions involving reward processes, approach behavior, eco-
nomic decision making, adaptive behavior, motion and cognition.1
Dopamine receptors fall into two subclasses, with D1 and D5
receptors sharing homology and coupling to Gs, and D2, D3, and
D4 receptors coupling to Gi. Selective D1 receptor antagonists have
been studied as potential therapeutics for Parkinson’s disease, psy-
chotic behavior, substance abuse and obesity2 in animal models
and in human clinical trials.3
A representative benzazepine class of D1 antagonists 1a and 1b
(SCH 23390)4 are conformationally constrained catecholamine
analogs. According to a D1 receptor mapping study,5 a phenol OH
group, an H-bond donor with a well-defined directionality,6 and
a basic nitrogen placed 7 Å apart are the key features of the phar-
macophore represented by 1b. The N3-methyl group is not re-
quired for D1 binding affinity.7 The ‘accessory phenyl’ group at
the C5 position is preferred to be somewhat coplanar with the phe-
nol ring and can be replaced with a large range of other hydropho-
bic groups (Fig. 1).
Cl
Cl
1
5
3
N
3
N
2
1
N
Cl
HO
HO
N
HO
1a
D1 Ki 50nM
1
1b (SCH23390)
D1 Ki 1nM
D1 Antagonist Design
Figure 1. Novel D1 antagonist design based on 1b (SCH 23390).
NH
NH2
Bu
N
H
N
N
pKa 10.8
pKa 8.9
Figure 2. pKa values of N-phenyl amidine and N-phenylguanidines.
In an effort to identify novel D1 antagonists, we decided to
examine the 1,3-benzodiazepine core structure 1 as a potential
replacement for the benzazepine ring in 1a and 1b. N-aryl amidine
and N-guanidine model systems have pKa values around 8–9 and
10–11, respectively (Fig. 2),8 within a range matching the basicity
of the tert-azepine nitrogen center in 1a and 1b.9
The geometry of the 1,3-benzodiazepine (1) is close to 1b as
demonstrated by superimposing the MMFF94 minimized confor-
mations of the two core structures (Fig. 3).10 Despite the fact that
Figure 3. The top (a) and side (b) views of superimposition of minimized
conformation of 1b (purple) and 1 (grey) with hydrogens removed for clarity. Cl—
green, O—orange, N—blue.
* Corresponding author. Tel: +1 908 740 5651.
E-mail address: zhaoning.zhu@spcorp.com (Z. Zhu).
0960-894X/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2009.07.012

Z. Zhu et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5218–5221
5219
the C5 sp3 center with the attached phenyl group in 1b is replaced
with a sp2 nitrogen (N1) in 1, both scaffolds can place the substit-
uents in approximately the same region based on the minimized
conformations. The new achiral scaffold offers ease of synthesis
and an attractive way to explore SAR at the C2 and N1 positions.
In addition, absence of the metabolically labile N3-methyl group
in the new core may address this major metabolic pathway ob-
served in SCH23390 and improve bioavailability.11
The preparation of a prototype 1,3-benzodiazepine 8 was
accomplished using the synthetic route shown in Scheme 1.
NO2
NO2
NO2
N
NC
MeO
NC
a
+
S
S
OMe
OMe
N
Cl
Cl
Cl
4b
b
2
3
4a
NH2
Cl
NO2
H
N
HN
H2N
H
N
e
c, d
O
OMe
OMe
O
OMe
Cl
Cl
7
6
5
f, g
Cl
N
HO
N
8
Scheme 1. Reagents and conditions: (a) DMSO/NaOH 60% yield, 4a/4b ratio 1:2; (b) BH3/THF; (c) acetic anhydride/TEA/DCM; (d) SnCl2Á2H2O; (e) PhB(OH)2/Cu(OAc)2/TEA/4 Å
MS; (f) POCl3; (g) BBr.
Cl
Cl
Cl
OMe
OH
O
b,c
a
H2N
H2N
TeocHN
NO2
5
NO2
9
NO2
10
d
Cl
Cl
O
O
O
e, f
N
H2N
H
NH2
NO2
12
11
g
Cl
O
O
Cl
HO
N
N
H
h, i, j
N
NH
13
14
Scheme 2. Reagents and conditions: (a) BBr3; (b) TMS-ethyl p-nitrophenyl carbonate/K2CO3/THF/H2O; (c) brominated Wang resin/Bu4N+IÀ/DIEA/DMF,60 °C/6 h; (d) TBAF/
THF; (e) acetic anhydride/DIEA/DCM; (f) SnCl2Á2H2O/DIEA; (g) pre-formation of imine overnight, NaBH4/DCE/MeOH; (h) 50%TFA/DCM; (i) HCI salt conversion; (j) POCl3/DCE,
85 °C/4.5 h, then HPLC purification.

5220
Z. Zhu et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5218–5221
generated compound 7.13 Cyclization of compound 7 under reflux
Table 1
1,3-Benzodiazepinesa
with POCl3 generated diazepine 8 after demethylation using BBr3
in DCM in high yield.14
N
Cl
HO
Compound 8 binds to human D1 dopamine receptors with an
affinity of 200 nM.15 Even though its affinity is two orders of
magnitude weaker than SCH 23390, diazepine 8 is still an attrac-
tive starting point for further optimization. The C2 SAR develop-
ment would offer an opportunity for further improvement of
binding affinity. Enhancement of the basicity of the amidine
group through replacement of the N-1 phenyl with an alkyl or
benzyl group may also lead to better D1 receptor binding affin-
ity. In order to efficiently expand the SAR of this scaffold, we
decided to develop a solid phase synthesis route (Scheme 2).
Thus phenol 9, obtained by demethylation of 5, was reacted with
Teoc–OSu to give corresponding Teoc carbamate, which was sub-
sequently loaded onto the brominated Wang resin. The loading
level was determined to be 0.8 mmol/g after TFA cleavage of
10 to give compound 9. Treatment of 10 with TBAF led to the
resin bound intermediate 11 which was subsequently acetylated
with acetic anhydride to give 12. Amide 12 was reduced using
SnCl2Á2H2O in the presence of DIEA to prevent premature cleav-
age of the intermediate from the resin. Subsequent reductive
alkylation of 12 following an imine pre-formation and fast
NaBH4 reduction sequence led to the resin bound 13. After TFA
cleavage, 13 was cyclized with POCl3 followed by HPLC purifica-
tion to generate the final diazepine 14.
R2
N
R1
Compound
R1
R2
D1 Ki (nM)
D2/D1
36
8
Me
200
14
15
16
17
Me
Me
Me
Me
930
87
3
32
55
31
S
181
321
18
19
154
370
23
8
Similarly, solid phase N-arylation chemistry using boronic acid
and Cu(OAc)2 were used to generate N1-arylbenzodiazapines.16
The affinities for human D1 and D2 receptors were determined
for a number of benzodiazepines generated in this fashion (Table
1). The highest affinity was observed with R1 as 3-thienyl (15).
Methyl substitutions on phenyl (16 and 17) had little impact on
binding affinity, while replacing phenyl with benzyl (14) led to
an approximately fourfold increase in the Ki value. Other C2-sub-
stitutions were shown to lack major impact on binding affinity.
For example, replacement of C2-Me with 2-(p-dimethylamino-
phenyl)vinyl did not change the D1 affinity or D1/D2 selectivity
(8, 200 nM vs 18, 154 nM). N1-substituents are not required to
maintain comparable level of D1 receptor affinity; for example,
23. This mirrors similar findings with the benzazepine class of
D1 antagonist represented by 1a.4a However, the fact that diaze-
pine 26, having the smallest R2 groups, showed the weakest D1
affinity indicates that either R1 or R2 must be a relatively large
hydrophobic group (Scheme 3).
One major difference between the diazepine ring in 1 and aze-
pine in 1b is basicity, with the 1,3-benzodiazepines being at least
an order of magnitude less basic. It was of interest to determine
whether increased basicity would lead to improved D1 receptor
binding affinity. One way to boost the basicity of the benzodiaze-
pines is to replace the cyclic N-aryl amidine with a cyclic N-aryl
guanidine.8b Synthetically, this was easily done with a slight mod-
ification of the solid phase chemistry. Therefore, resin bound com-
pound 11 was reduced with SnCl2Á2H2O in the presence of an
excess triethylamine. Thiophosgene treatment of 27 under basic
conditions led to cyclic thiourea 28, which was further activated
though treatment with MeI to give 29. Subsequent treatment with
ammonia in THF led to diazepine 30 after TFA cleavage and HPLC
purification.
N
N
20
21
22
H
H
H
198
442
475
4
17
42
O
23
H
H
284
206
7
3
24
25
O
O
H
H
311
3
1
26
1500
a
Each Ki value is an average of three determinations, and the standard errors for
all Ki determinations are less than 10%.
In contrast to 1a, the un-substituted 2-amino-1,3-benzodiaze-
pine (30) lacks D1 receptor affinity (Table 2). Substitution on the
exocyclic amino group led to improved D1 binding affinity with
the most hydrophobic group (31) having the lowest Ki value. This
agrees with the SAR trend in the cyclic-amidine series in Table 1
(20–26). It is clear that the cyclic N-aryl guanidine did not improve
the binding affinity compared to the cyclic N-aryl amidine
counterpart.
Compound 3, which was generated with N,N-dimethyl thiocarba-
mate and chloroacetonitrile, reacted with commercially available
m-MeO-p-Cl-nitrobenzene to produce a 1:2 mixture of 4a and 4b
with a combined yield of 60%.12 The nitrile 4b was reduced using
BH3–THF to give amine 5, which was subsequently acetylated
and reduced using SnCl2Á2H2O to give amide 6. Aniline N-arylation
chemistry using phenyl boronic acid with anhydrous Cu(OAc)2

Z. Zhu et al. / Bioorg. Med. Chem. Lett. 19 (2009) 5218–5221
5221
Cl
Cl
Cl
NH
S
O
O
a
b
N
H
O
H2N
H2N
NO2
11
NH2
27
28
c
Cl
Cl
N
N
d
N
H
N
H
S
NH2
O
HO
30
29
Scheme 3. Reagents and conditions: (a) SnCl2Á2H2O/TEA (b) thiophosgene/TEA; (c) MeI; (d) (i) NH3/THF/sealed tube/Heat, (ii) TFA, (iii) HPLC purification.
Acknowledgments
Table 2
SAR 2-aminobenzodiazepinesa
We would like to thank Dr. Johannes Voigt for reproducing
Fig. 3a and b, Dr. Duane Burnett for proof reading the manuscript,
Ms. Emily Luk for LC–MS analysis and Dr. T. M. Chan and Ms. Re-
becca Osterman for helping with NMR structural determinations.
N
Cl
R2
N
H
X
HO
Compound
X
R2
H
D1 Ki (nM)
D2/D1
1
30
NH
10,200
References and notes
1. Schultz, W. Annu. Rev. Neurosci. 2007, 30, 259.
31
NH
129
12
2. (a) Olver, J. S.; O’Keefe, G.; Jones, G. R.; Burrows, G. D.; Tochon-Danguy, H. J.;
Ackermann, U.; Scott, A.; Norman, T. R. J. Affect. Disord. 2009, 114, 321; (b)
Tupala, E.; Tiihonen, J. Psychiat. Res.: Neuroimaging 2008, 162, 1; (c) Noda, Y.;
Nabeshima, T. Ann. N.Y. Acad. Sci. 2004, 1025, 62; (d) Bourne, J. A. CNS Drug Rev.
2001, 7, 399.
3. (a) Haney, M.; Ward, A. S.; Foltin, R. W.; Fischman, M. W. Psychopharmacology
2001, 155, 330; (b) Nann-Vernptica, E.; Donny, E. C.; Bigelow, G. E.; Walsh, S. L.
Psychopharmacology 2001, 155, 338; (c) Astrup, A.; Greenway, F. L.; Ling, W.;
Pedicone, L.; Lachowicz, J.; Strader, C. D.; Kwan, R. Obesity (Silver Spring) 2007,
15, 1717.
32
33
NH
NH
391
465
12
1
4. 1a: (a) Chang, W. K.; Peters, M.; Fevig, V. P.; Kozlowski, J. A.; Zhou, G.; Lowe, D.
B.; Guzik, H.; McQuade, R. D.; Duffy, R.; Coffin, V.; Berger, J. G. Bioorg. Med.
Chem. Lett. 1992, 2, 399; 1b: (b) Iorio, L. C.; Barnett, A.; Leitz, F. H.; Houser, V. P.;
Korduba, C. A. J. Pharm. Exp. Ther. 1983, 226, 462; c see also Ref. 2d.
5. Mottola, D. M.; Laiter, S.; Watts, V. J.; Tropsha, A.; Wyrick, S. D.; Nichols, D. E.;
Mailman, R. B. J. Med. Chem. 1996, 39, 285.
34
35
S
445
575
2
6
NH
6. Wu, W.-L.; Burnett, D. A.; Spring, R.; Greenlee, W. J.; Smith, M.; Favreau, L.;
Fawzi, A.; Zhang, H.; Lachowicz, J. E. J. Med. Chem. 2005, 48, 680.
7. Baindur, N.; Tran, M.; Niznik, H. B.; Guan, N. H. C.; Seeman, P.; Neumeyer, J. L. J.
Med. Chem. 1992, 35, 67.
36
37
NH
S
Me
Me
783
900
13
5
8. (a) Oszczapowicz, J.; Kuminska, M. J. Chem. Soc., Perkin Trans. 2 1994, 103; (b)
Hirt, R. C.; Halverson, F.; King, F. T. Spectrochim. Acta 1959, 15, 962.
9. N-methyl azepine was reported to have pKa of 10.4 in aqueous solution: Hall, H.
K. J. Org. Chem. 1964, 29, 3135.
a
Each Ki value is an average of three determinations, and the standard errors for
all Ki determinations are less than 10%.
10. Macromodel, Monte Carlo Conformational Search, BatchMin V9.1, MMFF94s,
GBSA water solvation model, Schrödinger LLC, New York, NY.
11. (a) Tedford, C. E.; Coffin, V. L.; Ruperto, V.; Cohen, M.; McQuade, R. D.; Johnson,
R.; Kim, H.-K.; Lin, C.-C. Psychophamacology 1993, 113, 199; (b) Tedford, C. E.;
Ruperto, V. B.; Coffin, V. L.; Cohen, M.; Libonati, M.; Barnett, A. Drug Dev. Res.
1992, 26, 389; (c) Tedford, C. E.; Ruperto, V. B.; Barnett, A. Drug Metab. Dispos.
1990, 19, 1152.
Another fundamental difference between the benzazepine and
benzodiazepine scaffolds is the geometry of the basic nitrogen.
Even though the basicity of benzodiazepines based on phenyl ami-
dine or phenyl guanidine scaffold may approach the same range of
a tert-amine, the amidine nitrogen (sp2) geometry differs from that
of the tert-amine (sp3) in 1b (Fig. 3). If the directionality of these
interactions involving the nitrogen electron lone pair, or its proton-
ated form, is important, the triangular coplanar geometry of the
amidine would result in some intrinsic differences which will alter
the binding affinities. Further investigation in this direction may
offer more insight to this question and would certainly facilitate
exploration of similar designs applicable to other dopamine and
serotonin receptor ligands.
12. Makosza, M.; Danikiewicz, W.; Wojciechowski, K. Liebigs Ann. Chem. 1988, 203.
13. Chan, D. M. T.; Monaco, K. L.; Wang, R.-P.; Winters, M. P. Tetrahedron Lett. 1998,
39, 2933.
14. (a) All new compounds gave satisfactory 1H NMR and MS in accord with their
assigned structure. All final targets showed LC–MS purities >95%. (b) Analytical
data for 8: LC–MS, M+ = 287. 1H NMR (CD3OD, ppm), d 7.58–7.63, m, 3H; 7.49–
751, m, 2H;7.27, s, 1H;6.49, s, 1H;3.76–3.77, m, 2H;3.32–3.33, m, 2H;2.20, s, 3H.
15. Ltk- cells stably expressing D1 and D2 receptors were processed for D1 and D2
receptor binding with [3H] SCH 23390 as the ratio ligand. For detailed assay
6
information, see Ref.
.
16. Combs, A. P.;Saubern, S.;Rafalski, M.;Lam, P. Y. S. Tetrahedron Lett. 1999, 40, 1623.