6076
M. Zagami et al. / Tetrahedron Letters 53 (2012) 6075–6077
Table 1
O
O
O
O
H
H
H
H
Screened conditions to racemize pregabalin nitrile
S
R
NaDCC
toluene/water
TEA
NBn
NBn
NBn
Entry
Base
Solvent
Temperature
N
N
N
H
H
O
O
Cl
1
2
3
4
5
6
7
8
MeONa
Et3N
Et3N
Et3N
Et3N
tBuOK
tBuOK
tBuOK
MeOH
Acetone
Acetone
DMF
DMF
DMF
RT
RT
50 °C
RT
60 °C
RT
RT
H2, Pd/C
butyl acetate
O
H
NBn
N
H
H
O
DMSO
iPrOH
cis racemate
80 °C
Scheme 2. Racemization of moxifloxacin intermediate.
we have efficiently racemized the undesired enantiomer of the
N-benzylimide of cis-piperidine-2,3-dicarboxylic acid, key
intermediate of the antibiotic moxifloxacin, by (a) chlorination of
piperidine nitrogen, (b) dehydrochlorination to give C(2)@C(3)
double bond conjugated to imide carbonyls and (c) cis-hydrogena-
tion (Scheme 2).8
two main peaks, namely a 228.07 m/z peak, equivalent to proton-
ated N-dichloro pregabalin, with M+2 and M+4 peaks, of two-third
and one-ninth intense respectively, and a 160.15 m/z peak (base
peak), equivalent to protonated pregabalin. After verifying that
N-dichlorination was quantitative, pregabalin nitrile was synthe-
sized in one-pot from pregabalin without isolating the dichlorinat-
ed intermediate. Treatment of pregabalin in dichloromethane first
with TCCA and then, after filtration, with triethylamine afforded
(S)-3-cyano-5-methylhexanoic acid in 85% yield by simple concen-
tration of the reaction mixture previously washed with aqueous
HCl.12 The specific rotation was in agreement with the values re-
cently reported in the literature for pregabalin nitrile.13,14
To racemize pregabalin nitrile, we screened a series of bases
(triethylamine, sodium methoxide, potassium tert-butoxide) and
solvents (methanol, isopropanol, DMF, acetone) at different tem-
peratures (Table 1). We observed racemization only using potas-
sium tert-butoxide in DMSO or in isopropanol. In particular, we
found that this base induced racemization, which was complete
after a few minutes in DMSO at room temperature or after 4 h in
isopropanol at 80 °C. Under these latter conditions, optically inac-
tive 3-cyano-5-methylhexanoic acid was obtained from (S)-3-cya-
no-5-methylhexanoic acid in 87% yield.15 The overall yield of the
procedure, calculated on pregabalin, was 74%.
In conclusion, the conversion of pregabalin into pregabalin ni-
trile, which has never previously been considered, can easily be
achieved via N-dichlorination and double dehydrochlorination.
The subsequent quantitative racemization of the stereocentre un-
der basic conditions offers the chance to access rac-3-amino-
methyl-5-methylhexanoic acid by nitrile hydrogenation. We
believe that such a procedure, applied to the R enantiomer of pre-
gabalin, improves the preparation of pregabalin based on the reso-
lution of rac-3-aminomethyl-5-methylhexanoic acid allowing the
undesired enantiomer to be recycled. The recently reported
comparison16 of the classical resolution of 3-aminomethyl-
5-methylhexanoic acid with the enzymatic resolution of 2-car-
boxyethyl-3-cyano-5-methylhexanoic acid ethyl ester17 has to be
reformulated on account of the present racemization method,
which might be further improved by replacing DCM with a more
environmentally friendly solvent.
a
These observations prompted us to consider the potential of N-
chlorination to racemize the c-aminoacid pregabalin. In this case,
it was unlikely that a C@C double bond involving the stereogenic
C(3) of pregabalin would result from N-dehydrochlorination so as
to allow racemization by successive hydrogenation. However, we
considered that the primary amine function of pregabalin might
be dichlorinated; subsequent double dehydrochlorination would
convert the dichloroaminomethyl group into nitrile. This would
make the hydrogen of the stereogenic methyne, now in a-position
to CN, acidic and allow the nitrile to be easily racemized by treat-
ment with a base. Subsequent hydrogenation of racemic 3-cyano-
5-methylhexanoic acid should be facile, since it is reported as the
penultimate step in the pregabalin synthesis preceding the resolu-
tion and providing racemic 3-aminomethyl-5-methylhexanoic acid
in high yield.6
The racemization experiments were carried out on the commer-
cially available enantiomer (S)-(+)-3-aminomethyl-5-methylhexa-
noic acid (pregabalin) (Scheme 3). Trichloroisocyanuric acid
(TCCA) was chosen as a chlorinating agent. According to De Luca
and Giacomelli,10 the reaction between primary amines and equi-
molar TCCA under mild conditions gives the corresponding dichlo-
roamines in very high or quantitative yields and not the nitriles as
previously reported.11 Our experiments confirmed the N-dichlori-
nation. Treatment of pregabalin with TCCA in dichloromethane at
room temperature afforded the N-dichloro-c-aminoacid, which
was isolated as an oil in near quantitative yield by simple filtration
and concentration of the reaction mixture. The TLC analysis
showed the disappearance of the starting substrate and the forma-
tion of a unique UV detectable product, eluted by relatively apolar
mobile phases, such as cyclohexane/ethyl acetate mixtures, unable
to carry up pregabalin. 1H NMR and elemental analyses were
consistent with the formation of N-dichlorinated pregabalin as a
single product and the mass spectrum showed the presence of
Acknowledgment
We thank the Italian Ministry of University and Research for
financial support.
OK
H2N
Cl2N
TCCA
DCM
TEA
NC
NC
i-PrOH
DCM
HOOC
HOOC
HOOC
HOOC
References and notes
(S)-(+)-3-aminomethyl
-5-methylhexanoic
acid
(S)-(-)-3-cyano-5-
methylhexanoic
acid
rac-3-cyano-5-methyl
hexanoic acid
1. Yuen, P.; Kanter, G. D.; Taylor, C. P.; Vartanian, M. G. Bioorg. Med. Chem. Lett.
1994, 6, 823.
2. Lawson, K. Curr. Opin. Invest. Drugs 2006, 7, 631.
3. Marino, J.; Taylor, K. CA 2530904 A1, 2007.
4. Silverman, R.B.; Andruszkiewicz, R.; Yuen, P.W.; Sobieray, D.M.; Franklin, L.C.;
Schwindt, M.A. WO 9323383 A1, 1993.
(pregabalin)
(pregabalin nitrile)
rac-3-aminomethyl-
5-methylhexanoic
acid
5. Grote, T.M.; Huckabee, B.K.; Mulhern, T.; Sobieray, D.M.; Titus, R.D. WO
9640617 A1, 1996.
Scheme 3. Racemization of pregabalin.