Concise, enantiospecific synthesis of (3S,4R)-3-amino-4-ethylpiperidine as partner to a non-fluoroquinolone nucleus

Article abstract of DOI:10.1016/S0040-4039(03)00421-0

An enantiospecific, eight-step synthesis of (3S,4R)-3-amino-4-ethylpiperidine 3 starting from readily available (S)-(-)-α-methyl-4-pyridinemethanol 6 has been achieved utilizing an Overman rearrangement of a chiral allylic trichloroacetimidate 13 as the key step. A diastereoselective hydrogenation of the resulting chiral allylic amine 15 afforded predominantly the desired trans-substituted piperidine. The conformation of the piperidine along with the directing nature of the amino function are implicated in the selectivity observed.

Full text of DOI:10.1016/S0040-4039(03)00421-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 44 (2003) 2927–2930
Concise, enantiospecific synthesis of
(3S,4R)-3-amino-4-ethylpiperidine as partner to a
non-fluoroquinolone nucleus
Michael Reilly,* Donald R. Anthony and Corey Gallagher
Chemical Development, Procter & Gamble Pharmaceuticals, Woods Corners, PO Box 191, Norwich, NY 13815, USA
Received 27 January 2003; revised 10 February 2003; accepted 12 February 2003
Abstract—An enantiospecific, eight-step synthesis of (3S,4R)-3-amino-4-ethylpiperidine 3 starting from readily available (S)-(−)-a-
methyl-4-pyridinemethanol 6 has been achieved utilizing an Overman rearrangement of a chiral allylic trichloroacetimidate 13 as
the key step. A diastereoselective hydrogenation of the resulting chiral allylic amine 15 afforded predominantly the desired
trans-substituted piperidine. The conformation of the piperidine along with the directing nature of the amino function are
implicated in the selectivity observed. © 2003 Elsevier Science Ltd. All rights reserved.
As part of an ongoing investigation into the synthesis
and biological evaluation of 7-substituted, non-fluoro
quinolone derivatives within the context of our topoiso-
merase targeted anti-infectives program, a number of
derivatives containing amine functionality in the side
chain portion were progressed into development 1 (Fig.
1).1 Most notably, a variety of differentially substituted,
chiral, aminopiperidines 2 were found to exhibit good
class activity and were desired in multi-gram quantities
for further in vivo evaluation. In particular, a (3S,4R)-
Retrosynthetically (Fig. 2) it was envisioned that pipe-
ridine 3 could be derived from an externally situated
amino olefin 4 via a straightforward hydrogenation
3-amino-4-ethylpiperidine
quinolone nucleus was chosen for further exploration.
3 as a partner to the
The research synthesis used to produce this side chain
suffered from a variety of drawbacks and was not
suitable for the production of larger quantities of the
desired piperidine.2 The synthesis involved employed a
Figure 1. Non-fluoroquinolone derivatives (1) and chiral
piperidine side chains (2 and 3).
lengthy 11-step sequence starting from a
D-serine
derivative and tedious chromatography of
a
diastereomers at the end of the synthesis. Moreover, it
was retrospectively discovered that significant epimer-
ization of the C3% chiral center in 3 had occurred during
the synthesis resulting in near complete racemization. A
concise, enantiospecific synthesis of (3S,4R)-3-amino-4-
ethylpiperidine was therefore necessitated.
Keywords: Overman rearrangement; enantiospecific; pipiridine; 3-
amino-4-ethylpiperidine; quinolone; hydrogenation.
* Corresponding author. Tel.: +0-607-335-2918; fax: +0-607-335-2010;
e-mail: reilly.m@pg.com
Figure 2. Retrosynthetic dissection of (3S,4R)-3-amino-4-
ethylpiperidine.
0040-4039/03/$ - see front matter © 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0040-4039(03)00421-0

2928
M. Reilly et al. / Tetrahedron Letters 44 (2003) 2927–2930
protocol. It was believed that the desired, relative, trans
geometry could be achieved predominantly during
hydrogenation by exploiting the directing nature of the
primary amino function.3 The directing ability of hap-
tophilic groups, such as R–NH2, during heterogeneous
hydrogenation has been well documented and utilized
to dramatic effect in the past.4 The absolute stereo-
chemical control required during the synthesis of the
chiral allylic amine 4 could, in turn, be obtained from a
similarly chiral allylic alcohol acetimidate 5 via oxygen–
nitrogen transposition utilizing a highly selective [3,3]-
sigmatropic Overman rearrangement reaction.5 Owing
to the highly organized transition state for this reaction
and an analogous carbocyclic example of this rear-
rangement,6 it was believed that excellent facial transfer
of chirality and control of absolute stereochemistry
could be achieved. The chiral allylic alcohol necessary
for the present synthesis, though synthetically
approachable in a multitude of ways, is represented
below as a 1,2,5,6-tetrahydropyridine derivative 5 that
could be obtained sequentially from a suitable chiral
pyridine7 6 via quarternization and subsequent hydride
reduction.8
ate quaternary iodide salt as a solid in 98% yield.
Reduction of this salt with sodium borohydride led to
the desired 1,2,5,6-tetrahydropyridine derivative 8 in
97% yield. Exposure of the secondary allylic alcohol to
trichloroacetonitrile in the presence of catalytic NaH
led to the corresponding allylic trichloroacetimidate 9
derivative in 55% yield after silica gel chromatography.
Rearrangement under thermal conditions was evaluated
in a variety of solvents (Table 1). Refluxing THF,
which was amenable to
a
foreseeable one-pot,
trichloroacetimidate formation/rearrangement protocol,
failed to give the desired product and only starting
material was recovered. The higher boiling solvent tolu-
ene afforded only slight solubility of the substrate and,
after prolonged heating, decomposition was observed.
The use of chlorobenzene as solvent offered excellent
solubility and, upon heating at reflux for several hours,
afforded the desired rearranged allylic trichloro-
acetamide 10 in 50% isolated yield. An insoluble, black,
tar-like residue, presumed to be due to starting material
decomposition, was also observed and believed to be
having a deleterious effect on yield. It was supposed
that trace acids, either from the solvent or formed
during the high temperature procedure, were contribut-
ing to the decomposition of the trichloroacetimidate
starting material 9. As a result, modified Overman
rearrangement conditions previously reported to inhibit
the decomposition of allylic trichloroacetimidates by
scavenging trace acid were employed.11 The addition of
2 mg K2CO3/mL solvent to the reaction mixture fol-
lowed by heating to reflux led to the clean transforma-
tion of the allylic acetimidate into the allylic
trichloroacetamide 10 in 95% yield and near analytical
purity.
Fortunately, for the purposes of our synthesis, both
antipodes of chiral a-methyl-4-pyridinemethanol were
commercially available.9 Prior to the consumption of
relatively expensive chiral starting materials however, it
was decided to explore the conditions for the key
Overman rearrangement step. With this in mind, a
model study was initiated using the prochiral 4-
acetylpyridine 7 as starting material (Scheme 1).
Following literature protocol,10 treatment of 4-
acetylpyridine with methyl iodide led to the intermedi-
With the optimized Overman rearrangement conditions
in hand, attention was focused on the transformation of
chiral (S)-(−)-a-methyl-4-pyridinemethanol 6 into the
desired optically pure 3-amino-4-ethylpiperidine
(Scheme 2). Acetylation of the secondary alcohol with
acetic anhydride followed by quarternization with ben-
zyl bromide afforded the quarternary bromide salt 11
as a solid in 94% yield. Reduction of the pyridinium
salt with sodium borohydride to afford the 1,2,5,6-tet-
rahydropyridine intermediate followed by in situ depro-
tection of the acetate with 1.0N NaOH produced the
requisite secondary alcohol 12 in preparation for for-
mation of the chiral trichloroacetimidate ester. The
preparation of the trichloroacetimidate ester 13 was
improved by treatment of the allylic alcohol with
trichloroacetonitrile and DBU to provide a 91% yield
Scheme 1. Reagents and conditions: (a) MeI, MeOH, reflux,
98%; (b) NaBH4, MeOH, 5–20°C, 97%; (c) Cl3CCN, cat.
NaH (10 mol%), THF, −15°C, 55%; (d) see Table 1.
Table 1. Model study on the racemic trichloroacetimidate 9
Entry
Solvent
Temp. (°C)
Time (h)
Additive
Yield (%) 10
a
1
2
3
4
THF
PhMe
PhCl
PhCl
Reflux
110
135
24
20
15
15
None
None
None
K2CO3
–
–
b
50
135
95c
a No rearrangement product observed.
b Substrate only partially soluble, decomposition observed.
c 2.0 mg K2CO3 per 1.0 mL PhCl added to reaction mixture.

M. Reilly et al. / Tetrahedron Letters 44 (2003) 2927–2930
2929
19/18 as the di-HCl salt. The debenzylation reaction in
methanol worked similarly, albeit more quickly (entry 3
versus 4). Hydrogenation of the olefin over PtO2 (entry
2) afforded a 1:5 mixture of N-benzyl cis/trans isomers
17/16 which were then subjected to the debenzylation
protocol using Pearlman’s catalyst in methanol fol-
lowed by treatment with HCl (entry 5). No further
investigation of catalysts was performed; rather the 1:5
ratio of 19/18 was deemed acceptable for our purposes
and carried forward through the synthesis for determi-
nation of the absolute configuration. Multi-gram quan-
tities of the pure trans diastereomer, needed for the
production of drug substance 20, could be obtained in
58% yield, however, via careful silica gel chromatogra-
phy of the N-benzyl piperidine mixture 17/16 prior to
deprotection with Pearlman’s catalyst.
Scheme 2. Reagents and conditions: (a) Ac2O, pyridine,
DMAP, DCM, 12 h, 97%; (b) BnBr, 2-butanone, reflux, 12 h,
97%; (c) NaBH4, MeOH, 5–8°C, 3 h, 97%; (d) Cl3CCN,
DBU, DCM, 0°C, 15 h, 91%; (e) PhCl, K2CO3, 135°C, 5 h,
94%;15 (f) NaBH4, EtOH, 1.0N NaOH, 15 h, 60%; (g) H2,
catalyst (see Table 2); (h) H2, Pearlman’s catalyst (see Table
2); (i) 4.0N HCl/dioxane.16
The 1:5 cis/trans mixture from the PtO2 reduction/
Pd(OH)2 debenzylation (entry 5) protocol was carried
forward and coupled via aromatic nucleophilic substitu-
tion to the quinolone nucleus for absolute and relative
structure confirmation. Chiral CE analysis17 of the final
drug substance14 20 indicated the desired (3S,4R)
configuration of the amine side chain by comparison
with samples of known configuration. High enantiospe-
cificity was also achieved during the synthesis, with
<5% of the (3R,4S) enantiomer observed in the mix-
ture. The cis/trans ratio of the piperidine mixture was
also confirmed by this method to be 1:4.8.
after brief chromatography on silica gel. Application of
the optimized conditions, discovered earlier for the
thermal Overman rearrangement, resulted in the clean
transformation of the chiral allylic trichloroacetimidate
13 into the E-olefinic trichloroacetamide 14 in 94%
yield.12 Deprotection of the trichloroacetamide group
with sodium borohydride led to the requisite 3-amino-
4-ethylidene pyridine 15 in 60% yield.13 It was decided
at this point that the absolute stereochemistry of the
piperidine would be determined by conversion to the
finished quinolone drug substance14 20, whose absolute
stereochemistry was known.
It is believed that the diastereoselectivity achieved dur-
ing the hydrogenation of the olefin is due, at least in
part, to the interaction of the haptophilic amino group
with the catalyst surface. Since the allylic amine group
would be axially disposed due to A(1,3) strain, it could
easily interact with the catalyst surface providing deliv-
ery of hydrogen predominantly from the same side. The
more electrophilic nature of PtO2 (Adams’ catalyst) and
greater affinity toward the amine function could possi-
bly explain the greater facial selectivity achieved with
the use of this catalyst.
With this in mind, diastereoselective hydrogenation of
the exocyclic allylic amine 15 was briefly studied to
ascertain the effect of two different heterogeneous cata-
lysts on the cis/trans ratio (Table 2). Hydrogenation
over standard Pd/C (10 wt%) in ethanol at 1 atm H2 led
to the reduction of the olefin 15 with only partial
debenzylation observed to give a 1:3 cis/trans ratio
17/16 (entry 1). Subsequent de-benzylation with Pearl-
man’s catalyst (20% wt, entry 3) and treatment of the
crude reaction mixture with HCl/dioxane led to the
formation of the desired cis/trans side chain mixture
In conclusion, a concise, stereo- and enantiospecific
synthesis of (3S,4R)-3-amino-4-ethylpiperidine has been
achieved in eight steps utilizing a highly selective Over-
man rearrangement as the key transformation. Applica-
tion of this synthesis to the (3R,4S)-antipodal isomer
has also been achieved with similar results.
Table 2. Hydrogenation of 15 to 17/16; debenzylation to 19/18: conditions and yieldsa
Entry
Solvent
Catalyst
Time (h)
cis/trans
Yield (%)b
1
2
3
4
5
EtOH
EtOH
EtOH
MeOH
MeOH
Pd/C
PtO2
Pd(OH)2
Pd(OH)2
Pd(OH)2
12
12
20
12
12
1:3 (17/16)
1:5 (17/16)
1:3 (19/18)
1:3 (19/18)
1:5 (19/18)
91c
95
95d
95d
92d
a All hydrogenations performed under 1 atm H2 at 25°C.
b Combined yield.
c Slight debenzylation was observed.
d Isolated as the di-HCl salt (4.0N HCl/dioxane).

2930
M. Reilly et al. / Tetrahedron Letters 44 (2003) 2927–2930
13. Greene, T. W.; Wuts, P. G. M. Protective Groups in
Acknowledgements
Organic Synthesis; 3rd ed.; John Wiley & Sons, 1999;
p. 556.
14. Comparison of our sample to the quinolone structure
of known configuration shown below was used to
establish absolute stereochemistry.
The authors would like to thank Drs. Patricia Matson,
Tom Cupps, and Jared Randall for their help during
the preparation of this manuscript. We would also like
to thank Drs. Kylen Whitaker and Paul Zoutendam for
conducting chiral capillary electrophoresis (CE) analy-
sis of the final compounds. In addition, we would like
to extend our appreciation to David Berry and Jason
Winters for conducting preparative chromatography on
the diastereomeric mixtures.
References
1. Ledoussal, B.; Hu, X. E.; Almstead, J.; Gray, J. L.; Reed,
T. D. WO Patent 02/48138 A1, 2002, 73 pp.
2. Hu, X. E.; Kim, N. K.; Ledoussal, B. Org. Lett. 2002, 4,
4499.
3. Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993,
93, 1307–1370.
15. A typical procedure for the Overman rearrangement of
trichloroacetimidate 13 is as follows: A solution of the
trichloroacetimidate (1.1 g, 3.0 mmol) was dissolved in
150 mL dry chlorobenzene. To this solution was added
0.30 g of anhydrous granular K2CO3 (2 mg per mL
PhCl). The mixture was heated to 135°C for 5 h and
was then cooled to ambient temperature. The reaction
solution was filtered through a pad of Celite and con-
centrated in vacuo to a pale brown oil 14 (1.0 g, 94%).
1H NMR (300 MHz, CDCl3) was consistent with the
desired structure and shows the Overman rearranged
product 14 in excellent purity.
16. Selected spectroscopic and analytical data for
3(S),4(R)-trans-3-amino-4-ethylpiperidine dihydro-chlo-
ride 18: 1H NMR (300 MHz, CD3OD) l 3.64 (dd, 1H,
J=12.1 Hz, 3.9 Hz), 3.46 (bd, 1H, J=12.9 Hz), 3.39–
3.28 (m, 1H), 3.09–2.99 (m, 2H), 2.17 (bd, 1H, J=13.4
Hz), 1.83–1.72 (m, 2H), 1.60–1.46 (m, 1H), 1.37–1.25
(m, 1H), 0.99 (t, 3H, J=7.3 Hz); 13C NMR (75 MHz,
CD3OD) l 50.5, 46.1, 44.9, 40.3, 26.6, 24.5, 10.3; MS
(+ESI): m/z=129 (M+H).
4. Thompson, H. W.; Wong, J. K. J. Org. Chem. 1985, 50,
4270–4276.
5. Overman, L. E. J. Am. Chem. Soc. 1976, 98, 2901–2910.
6. Grigg, R.; Santhakumar, V.; Sridharan, V.; Thornton-Pett,
M.; Bridge, A. W. Tetrahedron 1993, 49, 5177–5188.
7. (a) Nakamura, K.; Takenaka, K.; Fujii, M.; Ida, Y.
Tetrahedron Lett. 2002, 43, 3629–3631; (b) Orrenius, C.;
Mattson, A.; Norin, T. Tetrahedron: Asymmetry 1994, 5,
1363–1366.
8. Waters, J. A.; Spivak, C. E.; Hermsmeier, M.; Yadav, J.
S.; Liang, R. F.; Gund, T. M. J. Med. Chem. 1988, 31,
545–554.
9. Commercially available from Aldrich Chemical Co.
10. (a) Spivak, C. E.; Gund, T. M.; Liang, R. F.; Waters, J.
A. Eur. J. Pharm. 1986, 120, 127–131; (b) Lyle, R. E.;
Perlowski, E. F.; Troscianiec, H. J.; Lyle, G. C. J. Org.
Chem. 1955, 20, 1761.
11. Nishikawa, T.; Asai, M.; Ohyabu, N.; Isobe, M. J. Org.
17. Chiral capillary electrophoresis (CE) conditions: Agilent
CE system, Voltage −20 kV, 200 mbar injection of a
0.1 mg/mL sample, u=298 nm. Running buffer: 5%
(w/v) highly sulfated-g-cyclodextrin (HSCD, Beckman)
in a 25 mM phosphate buffer pH 2.5. Capillary: Bared
fused silica total length 64.5 cm, effective length 56.0
cm, 50 mm i.d. with a 150 mm detection window.
Chem. 1998, 63, 188–192.
12. Due to the chair-like transition state adopted during the
Overman rearrangement and the equatorial disposition of
the methyl group resulting from the minimization of a 1,3
diaxial interaction with the bulky trichloromethyl group,
the E isomeric olefin is expected along with retention of
facial selectivity during the oxygen–nitrogen transposition.