Carboxylate bioisosteres of pregabalin

Article abstract of DOI:10.1016/j.bmcl.2006.03.083

Several β-amino tetrazole analogs of gabapentin 1 and pregabalin 2 were prepared by one of two convergent, highly efficient routes, and their affinity for the α₂-δ protein examined. Two select compounds with potent affinity for α₂-δ,

Full text of DOI:10.1016/j.bmcl.2006.03.083

Bioorganic & Medicinal Chemistry Letters 16 (2006) 3559–3563
Carboxylate bioisosteres of pregabalin
Jacob B. Schwarz,a,* Norman L. Colbry,a Zhijian Zhu,a Brian Nichelson,a Nancy S. Barta,a
Kristin Lin,a Raymond A. Hudack,a Sian E. Gibbons,a Paul Galatsis,a
Russell J. DeOrazio,b David D. Manning,b Mark G. Vartanian,a Jack J. Kinsora,a
Susan M. Lotarski,a Zheng Li,a Melvin R. Dickerson,a Ayman El-Kattan,a
Andrew J. Thorpe,a Sean D. Donevan,a Charles P. Taylora and David J. Wustrowa
aPfizer Global Research and Development, Michigan Laboratories, 2800 Plymouth Road, Ann Arbor, MI 48105, USA
bAlbany Molecular Research, Inc., PO Box 15098, Albany, NY 12212, USA
Received 10 February 2006; revised 23 March 2006; accepted 24 March 2006
Available online 18 April 2006
Abstract—Several b-amino tetrazole analogs of gabapentin 1 and pregabalin 2 were prepared by one of two convergent, highly
efficient routes, and their affinity for the a2–d protein examined. Two select compounds with potent affinity for a2–d, 8a and
16a, were subsequently tested in vivo in an audiogenic seizure model and found to elicit protective effects.
Ó 2006 Elsevier Ltd. All rights reserved.
N
CO2H
NH2
HN
Recently, we described the preparation and biological
activity of a series of analogs of gabapentin 1, in which
the carboxylate had been replaced with various bioisos-
teres.1 When the replacement employed for 1 was tetra-
zole, binding to the a2–d protein was conserved, as was
the in vivo anticonvulsant activity. Binding to a2–d
appears to attenuate calcium currents in synaptic
terminals,2 leading to a reduction in the release of neu-
rotransmitters, including norepinephrine, substance P,
and glutamate.3–5 Mindful that pregabalin 2 has shown
more potent and robust activity than gabapentin 1 in
various models of epilepsy, neuropathic pain, and anxi-
ety,6 we were interested in subjecting 2 to a similar strat-
egy. However, when the carboxylic acid of 2 was
replaced by tetrazole, no significant binding of 3 to the
a2–d protein was observed. To address this issue, we re-
called that a series of constrained b-amino acid analogs
of pregabalin 2 demonstrated good potency for the a2–d
protein.7 As a result, we were intrigued at the possibility
of replacing the carboxylate of b-amino acid analogs
of pregabalin 2 with a tetrazole8 and examining the
pharmacological consequences.
CO2H
NH2
N
N
NH2
1
2
3
In 1995, workers at Novartis showed that 5-lithiotet-
razoles added to a variety of electrophiles, including
aldehydes, ketones, enones, amides, iodine, and dieth-
yl chlorophosphate.9 Although they found that no
addition to benzyl bromide, N-methyl benzylidene-
amine, or propylene oxide took place, nitroolefins
were not examined as a possible electrophilic partner
for this useful reagent. We noted that if such an
addition were feasible, it might serve as a key step
toward a general preparation of b-amino tetrazoles,
potentially useful bioisosteres of b-amino acids
(Fig. 1, path a).
To initiate the sequence, 1-benzyltetrazole 4 was meta-
lated with n-butyllithium at low temperature, and with-
out allowing the mixture to warm, nitroolefin 5 in THF
was added (Scheme 1). It had been reported previously
that the lithiated tetrazole underwent decomposition
with extrusion of N2 at temperatures above ꢀ70 °C. In
fact, we observed that a portion of lithiated N-benzyl-
cyanamide was formed under the reaction conditions
which also added to the nitroolefin to give cyanamide
Keywords: Gabapentin; Pregabalin; a2–d Subunit; Tetrazole.
*
Corresponding author. Tel.: +1 7346222580; fax: +1 7346225165;
E-mail: jacob.schwarz@pfizer.com
0960-894X/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2006.03.083

3560
J. B. Schwarz et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3559–3563
NH2
N
N
a,b
c
N
NH2
N
NC
CN
N
Tr
N
a
CN
N
H
N
NH
N
9
10
11
N
g
b
N
d,e
a
b
f
NO2
N
N
N
N
N
N
CN
N
N
Tr
Tr
N
O
P
CN
R2
N
N
CN
R1
N
N
N
N
N
P
Tr
N
R1
H
N
CN
R2
N
N
12
14
R1
Figure 1. Convergent routes to access b-amino tetrazoles.
13
NH2
NH2
NO2
R1
R2
NO2
h
Bn
R1
a
Bn
12 or 14
or
N
N
NO2
NH
NH
N
+
+
R1
R1
Bn
N
N
R1
N
N
N
N
N
N
N
N
N
CN
N
15
16
5
7
4
6
Scheme 2. Reagents and conditions: (a) NaN3, NH4Cl, DMF, 80 °C;
(b) Ph3CCl, Et3N, DMF; (c) H2, PtO2; (d) NaH, R1X, THF; (e) NaH,
R2X, THF; (f) NaH, ketone, THF; (g) Bu4NBr, aldehyde, PhH, 10%
NaOH (aq); (h) H2, PtO2, MeOH, HCl.
b
NH2
R1
NH
N
N
N
alkyl halides, direct deprotonation (NaH) was found
to be required. If performing an SN2 displacement, it
was found that the sequence could be repeated to form
quaternary cyanotetrazoles 12. The unsaturated cyano-
tetrazoles 13 and 14 were somewhat sensitive but could
be adequately purified by expedient (plug) chromatogra-
phy on silica gel. Exhaustive reduction under acidic con-
ditions resulted in 1,2- and 1,4-reduction of the nitrile
and detritylation as for 11 to furnish the desired amino-
tetrazoles 15 or 16.
8
Scheme 1. Reagents and conditions: (a) n-BuLi, THF, ꢀ78 °C; (b) H2,
Pd/C, MeOH, HCl.
7. Fortunately, this byproduct could be easily removed
from the desired adduct 6 by chromatography. Hydro-
genation of 6 in the presence of acid afforded the target
aminotetrazole 8 in just two steps from the nitroolefin.
The added acid was essential to effect debenzylation of
the azole. Although this route proved very efficient, we
decided to examine an alternate approach owing to the
fact that nitroolefin preparation, although routine,
nevertheless required 2–3 steps.
As an extension of this method, we found that the
unsaturated cyanotetrazoles could also be cyclopropa-
nated using a sulfur ylide. Hence, treatment of crude
condensation products 13 with the sodium anion of
trimethylsulfoxonium iodide at ambient temperature
followed by hydrogenation afforded the cyclopropa-
nated aminotetrazole 18 (Scheme 3). Alternately, the
unsaturated cyanotetrazoles could be directly reduced
and deprotected in one pot to give 17. The alde-
hyde-derived unsaturated cyanotetrazoles of type 14
also underwent smooth cyclopropanation by this
method.
Guided by the Knoevenagel condensation approach to
b-amino acids,10 it was envisioned that a modification
using a cyanotetrazole in place of cyanoacetate would
afford the desired framework (Fig. 1, path b). An inspec-
tion of the literature found that this type of reagent had
indeed been previously prepared and its reactivity exam-
ined.11 Hence, it was reported that N-alkyl or N–aryl
cyanotetrazoles not only underwent condensation with
aldehydes, but also SN2 addition to alkyl halides and
Michael addition to ynones. With the ability to perform
alkylations as well as condensations, this versatile
reagent seemed highly attractive for our purposes.
Reaction of malononitrile 9 with sodium azide and
ammonium chloride at 80 °C, followed by regioselective
tritylation,12 afforded the appropriately protected cya-
nomethylated tetrazole 10 (Scheme 2). Direct nitrile
reduction of 10 under acidic conditions proceeded with
concomitant trityl group cleavage to furnish the tetraz-
olyl b-alanine derivative 11.13 Mild condensation of 10
with aldehydes could be carried out under phase-trans-
fer conditions as outlined previously.11 However, to
effect condensations with ketones or substitutions of
Unlike the SAR trends observed previously for modifi-
cations of the pregabalin backbone, subtle modifications
in the b-amino tetrazole series did not have a dramatic
N
N
N
N
N
N
N
N
Tr
N
N
a
b,a
CN
R2
N
H
N
H
NH2
NH2
R1
R2
R1
R1
R2
17
13 or 14
18
Scheme 3. Reagents: (a) H2, PtO2, MeOH, HCl; (b) trimethylsulfoxo-
nium iodide, NaH, DMF.

J. B. Schwarz et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3559–3563
3561
effect on a2–d binding using pig brain membranes7 in the
aliphatic series. Noteworthy is the fact that unsubstitut-
ed aminotetrazole 11 did not bind to a2–d. The direct
pregabalin analog 8a was found to be equipotent to pre-
gabalin 1, and the truncated structure 8b was fourfold
less potent (entries 3 and 4). However, when the methyls
on the isobutyl side chain of 8a were extended to a (2-
ethyl)-butyl group, the most potent compound was ob-
tained (cf. 16a, 11 nM). In the aromatic series, a more
predictable SAR was observed with respect to distance
of the phenyl ring from the pharmacophore. Until the
ring was separated from the aminotetrazole by at least
three methylene units, only micromolar affinity was ob-
tained (entries 12–17). A well seemed to be reached at
the four methylene unit distance. Finally, in the quater-
nary series the optimal combination of substituents was
found to be methyl and benzyl, optimally substituted on
the ring with an electron-withdrawing group as in 15d
(entries 18–22) (Table 1).
was required to obtain low nanomolar a2–d potency.
Not surprisingly, from the sulfur ylide the thermody-
namic products containing the alkyl group cis to the
nitrile (and subsequently the aminomethyl group) were
obtained.16 (Table 2)
Finally, the in vivo activity of this series of heterocyclic
derivatives was examined. It was found that, like pre-
gabalin 2 (but unlike c-amino tetrazole 3), protection
of DBA/2 mice against audiogenic seizures could be
achieved after oral dosing with 8a and 16a17 (Table 3).
In summary, a series of heterocyclic carboxylate replace-
ments of pregabalin 2 were prepared via two highly con-
vergent routes. The first route employed the addition of
Table 2. a2–d Binding affinity of cyclopropyl tetrazoles
NH2
R1
In the case of the cyclopropyl series, the SAR tracked
closely with the amino acid series reported previously.
For instance, it was found that the binding of spiro-
fused bicyclic tetrazole 18c to a2–d correlated with the
corresponding amino acid,7 however with somewhat re-
duced potency. In the case of pregabalin-like substrates
18a–b, only marginal potency was achieved. This obser-
vation was in accordance with the amino acid series,
where it was found that the opposite relative orientation
about the cyclopropyl ring (alkyl and carboxylate cis)
NH
R2
N
N
N
Entry
Compound
R1
R2 a2–d Binding
(Ki, lM)
1
2
3
18a
18b
18c
CH(CH3)2
CH(CH2CH3)2
H
H
0.19 0.019
4.0 0.15
0.033 0.007
–CH2CH2CH2CH2CH2–
Table 1. a2–d Binding affinity of b-amino tetrazolesa
NH2
NH
R1
R2
N
N
N
Entry
Compound
R1
R2
a2–d Binding (Ki, lM)b
1
2
2
—
H
—
H
0.019 0.003
>10
11
3
8a14
8b
CH2CH(CH3)2
CH(CH3)2
H
0.017 0.004
0.13 0.016
0.21 0.058
0.33 0.058
0.011 0.001
0.28 0.040
0.018 0.003
0.15 0.009
0.57 0.072
>10
4
H
5
8c
8d
CH2CH2CH(CH3)2
CH2(CH2)4CH3
CH2CH(CH2CH3)2
H
6
H
7
16a
16b15
17a
17b
16c
8e
H
8
CH2CH(CH2CH2CH3)2
c-C5H9
c-C6H11
H
9
H
10
11
12
13
14
15
16
17
18
19
20
21
22
H
CH2-c-C6H11
Ph
CH2Ph
H
H
16d
16e
16f
16g
16h
15a
15b
15c
15d
15e
H
7.6
3.1
CH2CH2Ph
CH2CH2CH2Ph
CH2(CH2)2CH2Ph
CH2(CH2)3CH2Ph
CH3
H
H
2.1 0.25
0.13 0.019
0.50 0.13
>10
H
H
CH3
CH3
CH3
CH3
CH2Ph
CH2Ph
0.90 0.18
0.52 0.089
0.16 0.020
>10
CH2–m,p-di-F–C6H3
CH2–m-CF3–C6H4
CH2Ph
aFor clarity, R1 and R2 have been modified to fit the generic structure above the table. The compound number only reflects method of preparation.
bData reported as means SEM (N = 4 experiments).

3562
J. B. Schwarz et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3559–3563
Table 3. Anticonvulsant activity of 2, 3, 8a, and 16a
13. McManus, J. M.; Herbst, R. M. J. Org. Chem. 1959, 24,
1643.
Compound
a2–d Binding
(Ki, lM)
DBA/2 anticonvulsant assay
(% protection)a
14. Representative procedure for addition of lithiated tetra-
zole to nitroolefin and subsequent hydrogenation: To a
solution of 1-benzyl-1H-tetrazole 4 (0.25 g, 1.56 mmol) in
7.0 mL THF at ꢀ78 °C was added n-butyllithium (1.6 M
solution in hexanes, 1.17 mL, 1.87 mmol) slowly dropwise
over 5 min until the yellow color persisted (ca. 1.0 equiv).
The mixture was stirred at ꢀ78 °C for 15 min at which
time 4-methyl-1-nitropent-1-ene 5a (0.20 g, 1.56 mmol) in
1 mL THF was added dropwise. The mixture was stirred
at ꢀ78 °C for 20 min, then warmed directly to ambient
temperature and quenched with satd NH4Cl (aq). The
mixture was extracted with EtOAc, and the organic phase
dried (Na2SO4) and concentrated. Flash chromatography
of the residue (15%!20%!25% EtOAc/hexanes)
provided 0.27 g (60%) of 1-benzyl-5-(3-methyl-1-nitrom-
ethyl-butyl)-1H-tetrazole 6a as well as 0.058 g (14%) of
benzyl-(3-methyl-1-nitromethyl-butyl)-cyanamide 7a as
colorless oils. Nitrotetrazole 6a: 1H NMR (CDCl3) d
7.37 (m, 3H), 7.22 (m, 2H), 5.72 (d, J = 15.6 Hz, 1H), 5.50
(d, J = 15.6 Hz, 1H), 4.69 (dd, J = 14.6, 9.8 Hz, 1H), 4.56
(dd, J = 14.6, 5.1 Hz, 1H), 3.72 (m, 1H), 1.47 (m, 2H), 1.19
(hept, J = 6.6 Hz, 1H), 0.65 (d, J = 6.6 Hz, 3H), 0.62 (d,
t = 1 h
t = 2 h
2
0.019
0.38
100
0
100
0
3
8a
16a
0.017
0.011
40
60
80
100
a % protection is the fraction of DBA/2 mice (N = 5 animals) protected
from audiogenically induced tonic seizures by a 30 mg/kg po dose of
the test compound.
a metalated tetrazole to a nitroolefin, followed by
reduction to yield b-amino tetrazoles in two steps. Alter-
nately, a Knoevenagel-type condensation of a malononit-
rile-derived cyanotetrazole with carbonyl compounds
afforded an intermediate that could be hydrogenated to
the desired target structures. This second method obviat-
ed the need for nitroolefin preparation, and could be
extended to give quaternary substituted aminotetrazoles
through double alkylation. A wide range of substitution
on the b-aminotetrazole template was tolerated by the
a2–d protein, and two representative substrates 8a and
16a were also found to have in vivo protective activity
against seizures.
1
J = 6.6 Hz, 3H). Cyanamide 7a: H NMR (CDCl3) d 7.37
(m, 3H), 7.31 (m, 2H), 4.51 (dd, J = 13.2, 8.8 Hz, 1H), 4.33
(dd, J = 13.2, 4.6 Hz, 1H), 4.27 (d, J = 13.7 Hz, 1H), 4.14
(d, J = 13.9 Hz, 1H), 3.72 (m, 1H), 1.60 (m, 2H), 1.28 (m,
1H), 0.86 (d, J = 6.4 Hz, 3H), 0.73 (d, J = 6.6 Hz, 3H);
LRMS: m/z 262.0 (M+1). To a solution of 1-benzyl-5-(3-
methyl-1-nitromethyl-butyl)-1H-tetrazole
6a
(1.41 g,
4.87 mmol) in MeOH (50 mL) were added 20% Pd/C
(0.20 g) and concentrated HCl (0.6 g). The mixture was
hydrogenated in a Parr shaker at 48 psi for 70 h, filtered,
and concentrated. The residue was dissolved in 5 mL H2O
and loaded onto a plug of DOWEX-50WX4-100 ion
exchange resin. The column was eluted with 100 mL H2O,
then 50 mL of 5% NH4OH (aq), and finally with 50 mL of
10% NH4OH (aq). The alkaline fractions were concen-
trated, and the solid dissolved in the minimum amount of
MeOH. Precipitation with EtOAc and filtration afforded
0.33 g (40%) of 4-methyl-2-(1H-tetrazol-5-yl)-pentylamine
8a as a colorless solid, mp > 240 °C (dec). 1H NMR (D2O)
d 3.31 (m, 1H), 3.10 (m, 2H), 1.53 (m, 1H), 1.35 (m, 1H),
1.03 (m, 1H), 0.66 (d, J = 6.6 Hz, 3H), 0.61 (d, J = 6.6 Hz,
3H). LRMS: m/z 168.0 (Mꢀ1). Anal. Calcd for C7H15N5:
C, 49.68; H, 8.93; N, 41.38. Found: C, 49.51; H, 8.90; N,
41.20.
References and notes
1. Burgos-Lepley, C. E.; Thompson, L. R.; Kneen, C. O.;
Osborne, S. A.; Bryans, J. S.; Capiris, T.; Suman-Chau-
han, N.; Dooley, D. J.; Donovan, C. M.; Field, M. J.;
Vartanian, M. G.; Kinsora, J. J.; Lotarski, S. M.;
El-Kattan, A.; Walters, K.; Cherukury, M.; Taylor, C. P.;
Wustrow, D. J.; Schwarz, J. B. Bioorg. Med. Chem. Lett.
2006, 16, 2333.
2. Fink, K.; Dooley, D. J.; Meder, W. P.; Suman-Chauhan,
N.; Duffy, S.; Clusmann, H.; Gothert, M. Neuropharma-
cology 2002, 42, 229.
3. Dooley, D. J.; Donovan, C. M.; Pugsley, T. A. J.
Pharmacol. Exp. Ther. 2000, 295, 1086.
4. Maneuf, Y. P.; Hughes, J.; McKnight, A. T. Pain 2001, 93,
191.
15. Representative procedure for condensation of 10 with an
aldehyde and subsequent hydrogenation: To a solution
of (2-trityl-2H-tetrazol-5-yl)-acetonitrile 10 (2.0 g,
5.69 mmol) in 25 mL benzene were added 2-n-propylval-
eraldehyde (0.87 g, 6.79 mmol), 25 mL of 10% NaOH
(aq), and tetrabutylammonium bromide (50 mg). The
mixture was stirred for 1 h and then poured into EtOAc.
The phases were separated, and the organic phase dried
(MgSO4) and concentrated. Flash chromatography of the
residue (10%!15% EtOAc/hexanes) furnished 1.66 g
(64%) of 4-propyl-2-(2-trityl-2H-tetrazol-5-yl)-hept-2-ene-
5. Dooley, D. J.; Mieske, C. A.; Borosky, S. A. Neurosci.
Lett. 2000, 280, 107.
6. Lauria-Horner, B. A.; Pohl, R. B. Expert Opin. Investig.
Drugs 2003, 12, 663.
7. Schwarz, J. B.; Gibbons, S. E.; Graham, S. R.; Colbry, N.
L.; Guzzo, P. R.; Le, V.-D.; Vartanian, M. G.; Kinsora, J.;
Lotarski, S. M.; Li, Z.; Dickerson, M. R.; Su, T.-Z.;
Weber, M. L.; El-Kattan, A.; Thorpe, A. J.; Donevan, S.
D.; Taylor, C. P.; Wustrow, D. J. J. Med. Chem. 2005, 48,
3026.
8. (a) Herr, R. J. Bioorg. Med. Chem. 2002, 10, 3379; (b)
Yuan, H.; Silverman, R. B. Bioorg. Med. Chem. 2006, 14,
1331.
1
nitrile 14b as a colorless solid. H NMR (CDCl3) d 7.34
(m, 10H), 7.09 (m, 6H), 2.86 (m, 1H), 1.30–1.59 (m, 8H),
0.90 (t, J = 7.1 Hz, 6H). To a solution of 4-propyl-2-(2-
trityl-2H-tetrazol-5-yl)-hept-2-enenitrile
9. Satoh, Y.; Marcopulos, N. Tetrahedron Lett. 1995, 36,
1759.
14b
(1.60 g,
3.47 mmol) in 50 mL 1:1 MeOH/THF were added PtO2
(0.1 g), concentrated HCl (1.09 g), and the mixture was
hydrogenated in a Parr shaker at 48 psi for 30 h. The
mixture was filtered and concentrated. Flash chromatog-
raphy of the residue on silica gel (0.25:1.25:3.5 concd
NH4OH (aq)/MeOH/CH2Cl2), followed by a second
10. Lee, J.; Gauthier, D.; Rivero, R. A. J. Org. Chem. 1999,
64, 3060.
11. Brekhov, Y. V.; Buzilova, S. R.; Vereshchagin, L. I. Zh.
Org. Khim. 1992, 28, 1921.
12. Huff, B. E.; LeTourneau, M. E.; Staszak, M. A.; Ward, J.
A. Tetrahedron Lett. 1996, 37, 3655.

J. B. Schwarz et al. / Bioorg. Med. Chem. Lett. 16 (2006) 3559–3563
3563
column using the same conditions, yielded 0.40 g (51%) of
4-propyl-2-(1H-tetrazol-5-yl)-heptylamine 16b as a color-
less solid. H NMR (D2O) d 3.29 (m, 1H), 3.09 (m, 2H),
1.56 (m, 1H), 1.37 (m, 1H), 0.80–1.07 (m, 9H), 0.58 (t,
J = 6.6 Hz, 3H), 0.47 (t, J = 6.6 Hz, 3H). Anal. Calcd for
C11H23N5: C, 58.63; H, 10.29; N, 31.08. Found: C, 57.87;
H, 10.04; N, 30.73.
16. As determined by 1D 1H NMR, and 2D NOESY and
COSY spectra obtained with compound 18a.
17. (a) Schlesinger, K.; Greik, B. J. In Contributions to
Behavior-Genetic Analysis: The Mouse as a Prototype;
Lindzey, G., Theissen, D. D., Eds.; Appleton-Century-
Crofts: New York, 1970; pp 219–257; (b) Seyford, T. N.
Fed. Proc. 1979, 38, 2399.
1