Synthesis of 3,4-disubstituted piperidines by carbonyl ene and prins cyclizations: Switching between kinetic and thermodynamic control with Bronsted and Lewis acid catalysts

Article abstract of DOI:10.1021/jo052532+

A novel approach to cis and trans 3,4-disubstituted piperidines is described. Carbonyl ene cyclization of aldehydes 4a-e catalyzed by MeAlCl ₂ in refluxing chloroform afforded the trans piperidines 7a-e with diastereomeric ratios of up to 93:7,

Full text of DOI:10.1021/jo052532+

Synthesis of 3,4-Disubstituted Piperidines by Carbonyl Ene and
Prins Cyclizations: Switching between Kinetic and Thermodynamic
Control with Brønsted and Lewis Acid Catalysts
Jodi T. Williams,† Perdip S. Bahia,† Benson M. Kariuki,† Neil Spencer,†
Douglas Philp,*,‡ and John S. Snaith*,†
School of Chemistry, The UniVersity of Birmingham, Edgbaston, Birmingham, B15 2TT, United Kingdom.
and Centre for Biomolecular Sciences, School of Chemistry, UniVersity of St. Andrews, North Haugh, St.
Andrews, Fife, KY16 9ST, United Kingdom
j.s.snaith@bham.ac.uk
ReceiVed December 9, 2005
A novel approach to cis and trans 3,4-disubstituted piperidines is described. Carbonyl ene cyclization of
aldehydes 4a-e catalyzed by MeAlCl2 in refluxing chloroform afforded the trans piperidines 7a-e with
diastereomeric ratios of up to 93:7, while aldehyde 4f afforded solely the cis product 6f, which was
resistant to isomerization to the trans isomer. It was demonstrated for 4a that the cyclization catalyzed
by a variety of Lewis acids at low temperature proceeded under kinetic control to afford predominantly
the cis piperidine 6a, and this isomerized to the thermodynamically more stable trans piperidine 7a on
warming. In contrast, Prins cyclization of 4a-e catalyzed by concentrated hydrochloric acid in CH2Cl2
at low temperature afforded cis piperidines 6a-e with diastereomeric ratios of up to >98:2. The yield
and diastereoselectivity of these cyclizations could be improved by using HCl-saturated CH2Cl2 to form
the corresponding chloride, followed by elimination of HCl effected by ammonia. Aldehydes 4f and 4g
also cyclized in good yield under the latter conditions. Mechanistic studies supported by DFT calculations
(B3LYP/6-31G(d)) suggest that the cyclizations proceed via a mechanism with significant carbocationic
character, with the cis carbocation being more stable than the trans carbocation. DFT calculations (B3LYP/
6-31G(d)) of the transition state energies for concerted cyclization show that the cis piperidine is also the
favored product from cyclization through a more concerted mechanism.
Introduction
In particular, methods for functionalizing the 3-, 4-, and
5-positions of the ring are rather limited. We now report in full
the results of our study into the synthesis of 3,4-disubstituted
piperidines by carbonyl ene and Prins cyclizations.4
The Lewis acid-catalyzed Type I intramolecular carbonyl ene
reaction is a very attractive method of ring closure, forming a
carbon-carbon bond with the concomitant generation of two
contiguous stereocenters.5 Such reactions are generally limited
to the formation of five- and six-membered rings, with the
cyclization of citronellal 1 being the prototypical Type I carbonyl
ene reaction (Scheme 1). Citronellal cyclizes to give principally
the two diastereomeric products 2 and 3. Diastereoselectivity
Functionalized piperidines occur widely in natural products1
and synthetic pharmaceuticals. The biological importance of
piperidines has led to the development of numerous synthetic
approaches to the ring system,2 but the wide variety of
functionality and substitution patterns present in piperidine
targets continues to drive the search for new methodologies.3
† University of Birmingham.
‡ University of St. Andrews.
(1) (a) Findlay, J. A. In The Alkaloids; Brossi, A., Ed.; Academic Press:
London, UK, 1985; Vol. 26, pp 89-183. (b) Pinder, A. R. Nat. Prod. Rep.
1986, 3, 171-180. (c) Pinder, A. R. Nat. Prod. Rep. 1987, 4, 527-537.
(d) Pinder, A. R. Nat. Prod. Rep. 1989, 6, 67-78. (e) Pinder, A. R. Nat.
Prod. Rep. 1990, 7, 447-455. (f) Pinder, A. R. Nat. Prod. Rep. 1992, 9,
491-504. (g) Plunkett, A. O. Nat. Prod. Rep. 1994, 11, 581-590. (h)
O’Hagan, D. Nat. Prod. Rep. 1997, 14, 637-651. (i) O’Hagan, D. Nat.
Prod. Rep. 2000, 17, 435-446.
(2) For reviews see: (a) Laschat, S.; Dickner, T. Synthesis 2000, 1781-
1813. (b) Weintraub, P. M.; Sabol, J. S.; Kane, J. M.; Borcherding, D. R.
Tetrahedron 2003, 59, 2953-2989. (c) Buffat, M. G. P. Tetrahedron 2004,
60, 1701-1729.
10.1021/jo052532+ CCC: $33.50 © 2006 American Chemical Society
Published on Web 02/11/2006
2460
J. Org. Chem. 2006, 71, 2460-2471

Synthesis of 3,4-Disubstituted Piperidines
SCHEME 1. Cyclization of Citronellal
Our initial investigation focused on the carbonyl ene reaction
of substrate 4a to afford the diastereomeric piperidines 6a and
7a, catalyzed by MeAlCl2. This Lewis acid has previously been
shown to be effective at catalyzing carbonyl ene reactions to
form six-membered rings,11 and our investigation focused on
the effect of temperature, time, and amount of catalyst on
diastereoselectivity (Table 1).
is dependent on the particular conditions and Lewis acid used,
but generally the trans diastereomer isopulegol 2 is favored,
with a diastereomeric ratio of up to 95:5.6
Application of the carbonyl ene cyclization to piperidine
synthesis has been little explored,7 and we were interested to
see whether the method could be used to synthesize 3,4-
disubstituted piperidines.
Conversion of 4a to the diastereomeric piperidines 6a and
7a was rapid and high yielding, with isolated yields of the crude
piperidine mixtures typically in excess of 90%. The reaction
was facile at -78 °C, even with substoichiometric amounts of
the Lewis acid (entries 1-4). At this temperature, the major
product was cis diastereomer 6a, and the diastereomeric ratio
proved to be relatively insensitive to the amount of Lewis acid
used. Raising the temperature at which the reaction was
performed to 25 °C and quenching after 2 h again gave 6a as
the major product (entry 5). Increasing the amount of Lewis
acid and the reaction time at this temperature resulted in
preferential formation of trans diastereomer 7a (entries 6 and
7), suggesting that under these conditions the carbonyl ene
reaction is reversible and that 6a is the kinetic product,
equilibrating to 7a on warming. This was confirmed by raising
the temperature to 61 °C (entry 8); under these conditions, 7a
predominated, with a diastereomeric ratio of 92:8 7a:6a. This
product ratio was also reached by subjecting a sample of 6a to
the same equilibrating reaction conditions.12
Along with the two major products, trace amounts (<5%) of
the cis chloride 8a were also isolated (for stereochemical
assignment vide infra). Such γ-chloro alcohols have been
reported previously as byproducts in alkylaluminum chloride
Lewis acid-catalyzed carbonyl ene reactions.13,14 This side
product was difficult to separate from 6a chromatographically,
but simply stirring a THF solution of a mixture of 6a and 8a
with aqueous ammonia in THF induced an elimination to afford
essentially quantitative recovery of pure 6a (Scheme 3).
The diastereomers were readily separated on silica to afford
pure 6a and 7a as white crystalline solids. The 1H NMR
spectrum of 7a exhibited a trans diaxial coupling constant of
10.1 Hz between the C3-C4 ring protons, and the trans
relationship between the two substituents was confirmed by
X-ray analysis. Coupling constants could not be extracted from
the 1H NMR spectrum of 6a, but it also proved possible to grow
single crystals of this diastereoisomer, and X-ray analysis
confirmed the cis relationship between the two substituents.
Other Lewis acids were screened at low temperature in an
effort to favor formation of the kinetic product 6a (Table 2).
Aluminum trichloride, which is more Lewis acidic than
MeAlCl2, was found to be an effective catalyst and favored
formation of the kinetic product 6a (entries 1-3). In contrast,
FeCl3 afforded only trace amounts (<5%) of the products (entry
4), with the remainder of the starting material returned
unchanged. Tin tetrachloride was effective at catalyzing the
reaction but the diastereoselectivity was poor (entry 5). Titanium
tetrachloride, on the other hand, was found to be an extremely
effective catalyst (entry 6), favoring the kinetic product with a
high diastereoselectivity, but also leading to the formation of
Results and Discussion
A range of cyclization precursors were straightforwardly
synthesized (Scheme 2) from 3-aminopropanol. These were
designed to explore the effect of steric bulk and the nucleophi-
licity of the ene component in the cyclization reaction.
N-Tosylation8 (94%) followed by N-alkylation with the
corresponding allylic bromide (59-86%) afforded the alcohols
4a-g in excellent overall yields. Subsequent oxidation was
generally carried out with PCC, but in the case of alcohol 1c,
trace formation of piperidinone 5 via the competing tandem
oxidation-cyclization-oxidation reaction9 led us to use subs-
toichiometric amounts of TPAP with NMO as reoxidant.10 This
avoided the side reaction, and aldehydes 4a-g were prepared
in 57-83% yields after flash column chromatography. These
â-amino aldehydes were generally used immediately, but they
could be stored for several weeks at -20 °C without significant
decomposition.
(3) (a) Watson, P. S.; Jiang, B.; Scott, B. Org. Lett. 2000, 2, 3679-
3681. (b) Brooks, C. A.; Comins, D. L. Tetrahedron Lett. 2000, 41, 3551-
3553. (c) Johnson, T. A.; Curtis, M. D.; Beak, P. J. Am. Chem. Soc. 2001,
123, 1004-1005. (d) Shu, C.; Alcudia, A.; Yin, J.; Liebeskind, L. S. J.
Am. Chem. Soc. 2001, 123, 12477-12487. (e) Harris, J. M.; Padwa, A. J.
Org. Chem. 2003, 68, 4371-4381. (f) Legault, C.; Charette, A. B. J. Am.
Chem. Soc. 2003, 125, 6360-6361. (g) Angoli, M.; Barilli, A.; Lesma, G.;
Passarella, D.; Silvani, A.; Danieli, B. J. Org. Chem. 2003, 68, 9525-
9527. (h) Amat, M.; Escolano, C.; Lozano, O.; Llor, N.; Bosch, J. Org.
Lett. 2003, 5, 3139-3142. (i) Hedley, S. J.; Moran, W. J.; Price, D. A.;
Harrity, J. P. A. J. Org. Chem. 2003, 68, 4286-4292. (j) Poupon, E.;
Francois, D.; Kunesch, N.; Husson, H. P. J. Org. Chem. 2004, 69, 3836-
3841. (k) Toure, B. B.; Hall, D. G. Angew. Chem., Int. Ed. 2004, 43, 2001-
2004. (l) Kuethe, J. T.; Comins, D. L. J. Org. Chem. 2004, 69, 2863-
2866. (m) Davis, F. A.; Zhang, J.; Li, Y.; Xu, H.; DeBrosse, C. J. Org.
Chem. 2005, 70, 5413-5419. (n) Legault, C. Y.; Charette, A. B. J. Am.
Chem. Soc. 2005, 127, 8966-8967. (o) Wurz, R. P.; Fu, G. C. J. Am. Chem.
Soc. 2005, 127, 12234-12235.
(4) For a preliminary account of this work see: Williams, J. T.; Bahia,
P. S.; Snaith, J. S. Org. Lett. 2002, 4, 3727-3730.
(5) For a review see: Snider, B. B. In ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Eds.; Pergamon: Oxford, UK, 1991; Vol. 2, pp
527-561.
(6) (a) Nakatani, Y.; Kawashima, K. Synthesis 1978, 147-148. (b)
Aggarwal, V. K.; Vennall, G. P.; Davey, P. N.; Newman, C. Tetrahedron
Lett. 1998, 39, 1997-2000.
(7) For examples see: (a) Laschat, S.; Fox, T. Synthesis 1997, 475-
479. (b) Monsees, A.; Laschat, S.; Kotila, S.; Fox, T.; Wurthwein, E.-U.
Liebigs Ann. 1997, 533-540 and 1041. For a highly diastereoselective
approach to indolizidines and quinolizidines via carbonyl ene reaction see:
(c) Laschat, S.; Grehl, M. Chem. Ber. 1994, 127, 2023-2034. (d) Overman
has reported piperidine synthesis via type II ene reactions: Overman, L.
E.; Lesuisse, D. Tetrahedron Lett. 1985, 26, 4167-4170.
(11) Snider, B. B.; Karras, M.; Price, R. T.; Rodini, D. J. J. Org. Chem.
1982, 47, 4535-4545.
(8) Bailey, T. R.; Garigipati, R. S.; Morton, J. A.; Weinreb, S. M. J.
Am. Chem. Soc. 1984, 106, 3240-3245.
(9) Bahia, P. S.; Snaith, J. S. J. Org. Chem. 2004, 69, 3226-3229.
(10) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639-666.
(12) For a review of the retro-ene reaction see: Ripoll, J.-L.; Valle´e, Y.
Synthesis 1993, 659-677.
(13) Snider, B. B.; Rodini, D. J.; Karras, M.; Kirk, T. C.; Deutsch, E.
A.; Cordova, R.; Price, R. T. Tetrahedron 1981, 37, 3927-3934.
(14) Snider, B. B. Acc. Chem. Res. 1980, 13, 426-432.
J. Org. Chem, Vol. 71, No. 6, 2006 2461

Williams et al.
SCHEME 2. Synthesis of Cyclization Precursors
TABLE 1. Cyclizations of Aldehyde 4a with MeAlCl2
TABLE 3. Cyclization of Aldehydes 4a-g with MeAlCl2
entry
aldehydea
time (h)
6:7b
yield (%)c
1
2
3
4
5
6
7
4a
4b
4c
4d
4e
4f
16
35
27
27
27
27
27
8:92
22:78
30:70
7:93
25:75
>98:2
74 (6)
55 (15)
53 (22)
74 (4)
61 (20)
90
a
entry
equiv of MeAlCl2
temp (°C)
time (h)
6a:7ab
4g
no reactiond
1
2
3
4
5
6
7
8
0.15
0.3
0.5
1.0
0.08
0.3
0.5
1.0
-78
-78
-78
-78
25
8
8
8
7
2
16
20
16
67:33
70:30
67:33
73:27
67:33
33:67
17:83
8:92
a Reactions were performed with 1 equiv of MeAlCl2 in chloroform at
61 °C. b The ratio was determined by integration of 1H NMR of a crude
mixture of piperidines. c Isolated yields of major (minor) isomers following
chromatography. d Starting material was recovered.
25
25
61c
Cyclization of the remaining aldehydes 4b-g was studied
under the optimized MeAlCl2 conditions; the results are shown
in Table 3.
a All reactions were performed in dry CH2Cl2 unless otherwise stated.
b The ratio was determined by integration of crude 1H NMR spectra. c The
reaction was performed in dry chloroform.
Analysis of the products 6b and 7b from cyclization of 4b
was complicated by the presence of E and Z double bond
isomers, and so the diastereomeric ratio was verified after
hydrogenation to the saturated products (entry 2). Aldehydes
4c-e, in which the alkene is exocyclic to a five-, six-, and seven-
membered ring, respectively, likewise all favored the trans
diastereoisomer (entries 3-5), a preference that was particularly
marked in the case of the cyclohexyl system (entry 4).
Surprisingly, 4f cyclized to give exclusively the cis piperidine
6f, identified from the characteristic coupling pattern of the C-4
proton. Only one double bond isomer was present, and an X-ray
crystal structure revealed this to be the Z-isomer.
The adamantyl substrate 4g did not undergo the carbonyl ene
cyclization with MeAlCl2, returning only unreacted starting
material. Aldehyde 4g would not be expected to undergo a
concerted cyclization as this would lead to formation of a
bridgehead double bond, compounds 11 and 12, in violation of
Bredt’s rule.15 Cyclization could be possible via a more stepwise
pathway (Scheme 4), but none of the products that could come
from the intermediacy of cations 9 and 10 was observed, e.g.,
interception by a nucleophile to afford 13 and 14.
SCHEME 3. Elimination of HCl from Chloride 8a
TABLE 2. Cyclization of Aldehyde 4a with a Variety of Lewis
Acids
entry
Lewis acida
equiv
time (h)
6a:7ab
1
2
3
4
5
6
7
8
9
AlCl3
AlCl3
AlCl3
FeCl3
SnCl4
TiCl4
ZnBr2
Sc(OTf)3
Cu(OTf)2
Yb(OTf)3
BF3‚Et2O
0.1
0.3
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
0.5
7
7
6
7
7
17
7
7
7
7
83:17
83:17
75:25
trace
67:33
92:8
no reactionc
50:50
no reactionc
no reactionc
67:33
10
11
7
Brønsted Acid-Catalyzed Reactions. Closely related to the
carbonyl ene reaction is the Prins reaction, the addition of an
aldehyde to an alkene catalyzed by a Brønsted acid.16 Reports
of intramolecular Prins reactions to form six-membered rings
are less common than their carbonyl ene counterparts, but we
were intrigued by a report from Holker of a highly diastereo-
selective example catalyzed by HCl.17 A small number of
Brønsted acids were therefore screened for the cyclization of
4a (Table 4).
a Reactions were performed in CH2Cl2 at -78 °C. b The ratio was
determined by integration of 1H NMR of the crude mixture of piperidines.
c Starting material was recovered.
significant amounts of the cis chloride 8a; the amount formed
varied between runs, but was typically between 20% and 40%.
Exploration of zinc bromide and a range of metal triflates
met with limited success (entries 7-10), with only scandium
triflate catalyzing the reaction with a surprising lack of selectiv-
ity. Boron(III) fluoride etherate gave 6a and 7a with poor
diastereoselectivity, along with up to 75% of the cis fluoride,
the fluoro analogue of chloride 8a. The stereochemistry of the
fluoride was assigned by comparison of the NMR spectra of
the two compounds.
(15) For a review see: Shea, K. J. Tetrahedron 1980, 36, 1683-1715.
(16) For a review see: Adams, D. R.; Bhatnagar, S. P. Synthesis 1977,
661-672.
(17) Chexal, K. K.; Holker, J. S. E.; Simpson, T. J.; Young, K. J. Chem.
Soc., Perkin Trans. 1 1975, 543-548.
2462 J. Org. Chem., Vol. 71, No. 6, 2006

Synthesis of 3,4-Disubstituted Piperidines
SCHEME 4
TABLE 4. Cyclization of 4a with a Variety of Brønsted Acids
TABLE 5. Cyclization of Aldehydes 4a-g with Concentrated
Hydrochloric Acid
entry Brønsted acida equiv temp (°C) time (h)
6a:7ab
entry
aldehydea
time (h)
6a:7ab
yield (%)c
1
2
3
4
5
6
7
8
CF3SO3H
TsOH
HCld
0.5
0.5
3.0
1.0
3.0
3.0
3.0
3.0
-78
-78
-78
-78
61f
-78
-78
-78
8
7
78:22
no reactionc
95:5
93:7
86:14
92:8
1
2
3
4
5
6
7
4a
4b
4c
4d
4e
4f
16
16
16
16
16
16
16
95:5
>98:2
90:10
89:11
80:20
79 (4)
86
71 (7)
72 (9)
62 (14)
no reactiond
no reactiond
16
64e
16
18
20
16
HCld
HCld
HBrg
HIh
H2SO4
90:10
87:13
i
4g
a All reactions were performed in CH2Cl2 unless otherwise stated. b The
ratio was determined by integration of 1H NMR of a crude mixture of
piperidines. c Starting material was recovered. d Refers to concentrated
(37%) hydrochloric acid. e The reaction was only 80% complete after this
time. f The reaction was performed in chloroform. g Refers to 48% hydro-
bromic acid. h Refers to 57% hydriodic acid. i Refers to concentrated sulfuric
acid.
a Reactions were performed with 3 equiv of concentrated (37%)
hydrochloric acid in CH2Cl2 at -78 °C. b The ratio was determined by
integration of 1H NMR of a crude mixture of piperidines. c Isolated yields
of major (minor) isomers following chromatography. d Starting material was
recovered.
TABLE 6. Cyclization of Aldehydes 4a-e in CH2Cl2 Saturated
with HCl
Results were disappointing with trifluoromethanesulfonic acid
and p-toluenesulfonic acid (entries 1 and 2), but hydrochloric
acid proved to be an extremely effective catalyst for the Prins
cyclization of 4a. Three equivalents of concentrated HCl at -78
°C in CH2Cl2, conditions presumed to lead to a low concentra-
tion of HCl in CH2Cl2, effected cyclization to 6a and 7a with
a diastereomeric ratio of 95:5 in favor of the kinetic isomer 6a
(entry 3). A trace (<5%) of chloride 8a was produced under
these conditions, which could be easily eliminated as before to
afford pure 6a. Reducing the amount of acid led to an
unacceptably slow reaction without improvement in the dia-
stereomeric ratio (entry 4). Although raising the temperature
of the reaction lowered the diastereoselectivity, the cyclization
still favored cis product 6a (entry 5). Hydrobromic acid and
hydriodic acid were also effective catalysts, but the diastereo-
selectivity of these reactions was not better than that with
hydrochloric acid (entries 6 and 7), and there were also traces
(<5%) of a side product produced under these conditions,
presumably the γ-bromo and γ-iodo alcohols.
Somewhat surprisingly, the diastereoselectivity with concen-
trated sulfuric acid in CH2Cl2 (entry 8) was relatively modest
compared to the results with HCl (entry 3). Switching to dilute
sulfuric acid (0.05 M aqueous, no cosolvent), cyclization
proceeded very slowly at room temperature (176 h) to give a
mixture of 6a, 7a, and the diols 15 and 16 with a ratio of 15:
8:54:23, respectively, i.e., a 69:31 ratio of cis:trans products
(Scheme 5).
entry aldehydea time (h)
6:8:7b
6:7 after elimb yield (%)c
1
2
3
4
5
6
7
8
9
4a
4a
4b
4b
4c
4c
4d
4d
4e
4e
2
6
2
6
2
6
2
6
2
6
57:40:3
4:93:3
54:45:1
47:52:1
89:6:5
80:8:12
97:3:0
86:6:8
94:6:0
82:14:4
97:3
>98:2
95:5
75 (1)
80 (0)
83 (4)
70 (0)
87 (0)
>98:2
>98:2
10
a Reactions were performed in CH2Cl2 saturated with HCl at -78 °C.
b The ratio was determined by integration of H NMR of a crude mixture
1
of piperidines. c Isolated yields of 6 (7) following chromatography.
All four products were readily separated on silica to afford
the pure compounds as white crystalline solids. The H NMR
1
spectrum of 16 identified a coupling constant of 10.5 Hz
between the C3-C4 ring protons, which is consistent with trans
diaxial coupling, while X-ray analysis of single crystals of 15
confirmed the cis relationship between the two substituents.
The remaining aldehydes 4b-g were cyclized under the
optimized concentrated hydrochloric acid conditions; the results
are shown in Table 5.
Cyclization of 4b exhibited a remarkable diastereoselectivity
of at least 98:2, with no trans product detectable by NMR; as
before, a mixture of double bond isomers was produced and
the diastereomeric ratio was verified both before and after
hydrogenation. Paralleling our earlier findings, around 5% of
the chloride resulting from addition of HCl to the double bond
of 6b was also obtained; subjecting a THF solution of the
chloride to aqueous ammonia effected elimination to the alkene.
Extending our study to aldehydes 4c-e, the preference for
formation of the cis diastereoisomers 6c-e was again marked
(entries 3, 4, and 5), although not as high as the acyclic
examples. Aldehyde 4f failed to give any cyclization products,
Diol 15 was the major product and was isolated in 51% yield.
Increasing the reaction temperature to 50 °C improved the rate
of reaction (reaction complete in 19 h), and also led to increased
amounts of the alkene products (a 25:13:44:18 mixture of 6a:
7a:15:16), although the overall cis:trans ratio remained un-
changed at 69:31. It is likely that under the conditions of
elevated temperature, dehydration of the tertiary alcohols takes
place to give the alkenes.
J. Org. Chem, Vol. 71, No. 6, 2006 2463

Williams et al.
SCHEME 5. Cyclization of Aldehyde 4a in Dilute Sulfuric Acid
SCHEME 6. Cyclization and Rearrangement of Aldehyde 4g
with only starting material recovered. The addition of more acid
and prolonged reaction times simply resulted in the formation
of decomposition products.
Adamantyl compound 4g also failed to cyclize under these
conditions. Although the concerted cyclization pathway is not
open to 4g (vide supra), we had hoped that under Brønsted acid
conditions a Prins-type cyclization could occur via generation
and trapping of the cations 9 and 10.
The successful results obtained with concentrated hydrochlo-
ric acid led us to explore the cyclization of 4a in dichlo-
romethane saturated with hydrogen chloride gas. Initially the
reaction was performed by bubbling anhydrous HCl gas through
a -78 °C CH2Cl2 solution of 4a for 30 min, followed by stirring
for a further 2 h at this temperature. This led to a 57:3:40 mixture
of 6a:7a:8a (i.e. 97:3 cis:trans), while in a separate experiment,
extending the reaction time to 6 h gave a 4:3:93 mixture of
6a:7a:8a (i.e. 97:3 cis:trans). No trans chloride was detected in
the crude NMR spectrum or in any of the fractions isolated on
purification.
easily separated from 6a and 8a chromatographically, so the
mixture of 6a and 8a isolated after purification was subjected
to HCl-saturated CH2Cl2 for a further 2 h to give a 3:97 mixture
of 6a:8a in an overall yield of approximately 75% from 4a; the
structure of 8a was confirmed by X-ray crystallography.
Cyclization of the remaining aldehydes in HCl-saturated CH2-
Cl2 led smoothly to the expected piperidines with excellent
diastereoselectivities in favor of cis products (Table 6). The best
diastereoselectivities were obtained after 2 h at -78 °C;
extending the reaction time to 6 h led unsurprisingly to increased
amounts of the corresponding cis chloride, but also to a slight
decrease in the overall cis:trans ratio in most cases. Once again,
trans chloride products were not detected.
Unlike the previous example, it was not possible to drive the
cis alkenes 6b-e completely to the corresponding cis chlorides,
possibly as a result of the greater steric hindrance present in
these systems. Instead, the crude mixture of cis and trans alkenes
and cis chloride was subjected directly to aqueous ammonia-
THF to effect elimination of HCl from the chloride, leaving
only the separable alkene products. This proved to be a very
efficient way of synthesizing the cis alkenes 6a-e in good
overall yields and with much improved diastereoselectivity over
the concentrated hydrochloric acid conditions.
Although the latter experiment led to chloride 8a being the
major product, it was difficult to isolate a pure sample of 8a as
it readily eliminated HCl on silica to give 6a, and 8a could not
be crystallized away from 6a and 7a. However, 7a could be
2464 J. Org. Chem., Vol. 71, No. 6, 2006

Synthesis of 3,4-Disubstituted Piperidines
SCHEME 7. Removal of the p-Toluenesulfonyl Protecting
Group
SCHEME 8. Transition States for Concerted Cyclizations
of 24 and 27
Cyclization of 4f for 2 h at -78 °C in HCl-saturated CH2Cl2
afforded exclusively the cis piperidine 6f, again as the Z double
bond isomer. Prolonged acid treatment and warming to room
temperature did not yield any of the chloride, presumably
because elimination to form the conjugated alkene is very facile.
Pleasingly, adamantyl system 4g also underwent a facile
cyclization in HCl-saturated CH2Cl2. After 2 h at -78 °C all
of the starting material had been consumed, and three products
SCHEME 9. Synthesis of Aldehydes 24 and 27
1
were visible in the crude H NMR, cis and trans chlorides 18
and 19 in a 52:23 ratio, along with 25% of chloroalkene 20.
X-ray crystal structures confirmed the identity of 18 and 19.
Extending the reaction time to 6 h at -78 °C led to an erosion
in the cis:trans selectivity to 42:36, with the remaining 22% of
the material being 20. To confirm that the cis chloride 18 could
isomerize to the trans chloride 19 a pure sample of 18 was
subjected to HCl-saturated CH2Cl2 at -78 °C for 24 h to give
a 60:40 ratio of 18:19.
Repeating the cyclization for 6 h at room temperature gave
a similar amount of the chlorides 18 and 19, but with the trans
chloride predominating (34:44 ratio of 18:19), while the amount
of 20 decreased slightly to 18%, and 4% of a new product was
present, identified as the diene 21. Extending the reaction time
to 24 h at room temperature resulted in an increase in the trans:
cis ratio for the chlorides, along with increased amounts of the
diene 21. Both chloro alcohols were unstable toward decom-
position to the diene on storage, with traces of the latter evident
after a few days in the refrigerator.
It is assumed that cyclization proceeds via the cis and trans
cations 9 and 10, with trapping by chloride ion leading to 18
and 19 (Scheme 6). The alternative pathway involving loss of
a proton from the adamantyl ring would afford bridgehead
alkenes 11 and 12, in violation of Bredt’s rule. The product
resulting from loss of a proton from the piperidine ring, 17, is
not observed either, presumably due to the significant A1,3 strain
present in this molecule, although it cannot be ruled out as an
intermediate in the formation of 20 or chlorides 18 and 19. Acid-
catalyzed elimination of water from 18 and 19 affords 20, the
main side product in the low-temperature cyclization, and loss
of HCl gives the diene 21.
dynamically more stable trans isomer isopulegol 2 is favored
with a variety of Lewis acids.6 There is no evidence that
cyclization proceeds through a kinetic intermediate that isomer-
izes. Indeed, Nakatani demonstrated that under the conditions
of zinc bromide in benzene, the cyclization of citronellal to
afford chiefly isopulegol 2 is irreversible.6a More recently,
however, Kocˇovsky´ has shown that several Lewis acids,
including zinc chloride in dichloromethane, rapidly isomerize
neoisopulegol 3 to isopulegol 2.18
A reversal in the sense of diastereoselectivity of citronellal
cyclization to favor the cis-diastereoisomer neoisopulegol 3 (dr
75:25) has been achieved with Wilkinson’s catalyst, although
this observation has not been satisfactorily explained.19
Koe`ovsk×c6 has explored the cyclization with bulky molybde-
num Lewis acids, observing a cis:trans ratio of up to 80:20. In
this case it is suggested that the bulky Lewis acid forces
cyclization through a boatlike transition state, leading to the
switch in diastereoselectivity.18
Given this precedent for citronellal, our cyclization results
for aldehydes 4a-e using Lewis acids under the equilibrating
conditions of elevated temperature were unsurprising and readily
explained by the accepted concerted cyclization mechanism. The
strong preference for formation of the kinetic product under
Brønsted acid conditions, or with Lewis acids at low temper-
ature, was much more difficult to account for, and did not appear
to fit with the classical concerted mechanism. Likewise, it was
Removal of the p-Toluenesulfonyl Protecting Group.
Removal of the p-toluenesulfonyl protecting group was carried
out on a representative range of examples by stirring the
N-tosylpiperidines with 6 equiv of sodium naphthalenide in THF
at -78 °C (Scheme 7). Deprotection proceeded cleanly to afford
good yields (52-79%) of the piperidines 22 and 23, although
the loss of a small amount of material on aqueous workup was
unavoidable due to the polar nature of the deprotected products.
Mechanistic Considerations. Carbonyl ene cyclizations to
form six-membered rings typically favor the thermodynamically
more stable stereoisomer. Considering the cyclization of cit-
ronellal, one of the most well-studied examples, the thermo-
(18) Kocˇovsky´, P.; Ahmed, G.; Srogl, J.; Malkov, A. V.; Steele, J. J.
Org. Chem. 1999, 64, 2765-2775.
(19) Funakoshi, K.; Togo, N.; Koga, I.; Sakai, K. Chem. Pharm. Bull.
1989, 37, 1990-1994.
J. Org. Chem, Vol. 71, No. 6, 2006 2465

Williams et al.
SCHEME 10. Proposed Dimerization Leading to 33
SCHEME 11. Proposed Reaction Pathway
difficult to rationalize the preference of 4f to give exclusively
this cis product 6f with MeAlCl2 in refluxing chloroform.
It was hoped that the cis and trans crotyl substrates 24 and
27 would provide a probe for the cyclization mechanism. We
reasoned that if the cyclizations proceeded through a concerted
mechanism, then the E-crotyl aldehyde 24 would be expected
to afford the trans piperidine 26 via transition state 25, while
the Z-crotyl aldehyde 27 would afford the cis piperidine 29 via
transition state 28 (Scheme 8). The synthesis of both aldehydes
was straightforward (Scheme 9).
Alkylation of 30 with commercially available crotyl chloride
(approximately 5:1 E:Z) gave the corresponding crotyl alcohol
with an E:Z ratio of 3:1. The stereoisomers were inseparable,
and so the mixture was oxidized with PCC to give aldehyde
24, also as a 3:1 E:Z mixture. To prepare the Z-isomer, 30 was
alkylated with 1-bromo-2-butyne to give alkyne 31, which
underwent smooth reduction on treatment with hydrogen and
Pd-BaSO4 poisoned with quinoline to afford the Z-crotyl
alcohol (20:1 Z:E). PCC oxidation afforded the Z-crotyl aldehyde
27 with an unchanged Z:E ratio.
previously optimized conditions of 3 equiv of HCl at -78 °C
in CH2Cl2, even with extended reaction time. Increasing the
reaction temperature to 25 °C again failed to result in any
reaction after 5 days. Even subjecting the two substrates to the
much more vigorous conditions of HCl-saturated CH2Cl2 at
room temperature failed to result in any reaction, giving after
25 h recovery of aldehyde starting materials. Extending the
reaction time to 80 h under these conditions gave rise to
extensive decomposition.
It appeared that removal of a methyl group from 4a to give
24 and 27 had completely deactivated the system toward Prins
cyclization. Since the Prins reaction is generally believed to
proceed through a stepwise mechanism via a cationic intermedi-
ate,16 the lack of reactivity of the E and Z crotyl compounds
under these conditions could be accounted for by their reluctance
to react through a much less favorable secondary carbocation
(cf. tertiary carbocation for substrate 4a). We therefore turned
our attention to the Lewis acid-catalyzed reactions of 24 and
27.
Both 24 and 27 proved to be completely unreactive at -78
°C with the Lewis acids that had previously been shown to be
effective at catalyzing the cyclization of 4a, but reaction
occurred on raising the temperature to 25 °C. Treatment of 24
With both cyclization precursors in hand we explored their
cyclization under Lewis and Brønsted acid conditions. Neither
24 nor 27 would undergo a Prins-type reaction under our
2466 J. Org. Chem., Vol. 71, No. 6, 2006

Synthesis of 3,4-Disubstituted Piperidines
SCHEME 12. Cyclization of Citronellal with HCl
ature suggests that there is significant cationic character to the
transition state. When 4a is cyclized in aqueous sulfuric acid
the overall cis:trans ratio is a more modest 2:1, suggesting that
the energy difference between the cis and trans cations is
reduced as a result of solvation by the surrounding water
molecules.20
In the case of the crotyl aldehydes 24 and 27, the loss of a
methyl group makes the resulting secondary cations (or any
cationic character developed during cyclization through a more
concerted pathway) much less stable than 36 and 37. As a
consequence, the stabilizing interaction with the oxygen lone
pair becomes even more significant, and cyclization through
the trans cation is energetically inaccessible. In a similar way,
the two-electron withdrawing aryl substituents present in 4f serve
to destabilize the trans cation relative to the cis cation, leading
to exclusive formation of the cis piperidine 6f.
As a reference point we performed the cyclization of
citronellal under our optimal hydrochloric acid conditions to
give a 16:33:11:40 mixture of 2:3:38:39 (Scheme 12).
Interestingly, the cyclization favored formation of the cis
products 3 and 39, but in marked contrast to our own system,
the combined ratio of cis:trans products was only 73:27. Using
HCl-saturated CH2Cl2 at -78 °C afforded after 4 h an 80:20
mixture of 39:38 as the sole products.21 Repeating the reaction
at room temperature while keeping the reaction time at 4 h led
to a 67:33 mixture of 39:38 as the sole products.
These results for citronellal under Brønsted acid catalysis,
conditions likely to result in a much more stepwise pathway,
would also appear to be consistent with the cation stability
arguments laid out above. The much higher diastereoselectivity
for the cyclization of aldehydes of the type 4 would therefore
suggest that there is a larger energy difference between cis and
trans cations 36 and 37 compared with the cis and trans cations
generated during the cyclization of citronellal, possibly as a
result of the electron withdrawing inductive effect of the N-tosyl
substituent.
FIGURE 1. Overlap with the oxygen lone pair stabilizes the cis cation
36.
and 27 with 0.5 equiv of MeAlCl2 at 25 °C led in both cases to
the formation of two products that were readily separable on
silica. The minor product, formed in around 20% yield in each
case, was the desired piperidine, and comparison of the 1H NMR
spectrum with those of 6a and 7a clearly revealed it to be the
cis diastereomer 29; no trans diastereomer was detected in the
crude reaction mixture from either reaction.
The major product from the reaction had a very complex 1H
NMR spectrum, suggesting that it was a mixture of diastere-
oisomers. It was apparent that there were resonances from more
than one tosyl group and there were also resonances from more
than one alkene. The product had a molecular weight of 580,
and the mass spectrum showed an isotope pattern characteristic
of one chlorine atom. Following extensive 2-D NMR we
tentatively assigned the structure 33 to this product, and propose
that it is formed via the dimerization process shown in Scheme
10. Dimerization is shown starting from 27, although an identical
pathway is assumed to operate for 24.
Due to the less electron-rich disubstituted double bond, the
rate of intramolecular cyclization is reduced, allowing inter-
molecular reactions to compete. Dimerization to form a hemi-
acetal of the type 32 is followed by Lewis acid-catalyzed
oxonium ion formation, and cyclization to a pyran as shown.
There are 16 possible stereoisomers of 33, and we believe at
least four of these are present in the mixture, making it
impossible to unequivocally assign a structure to this compound.
The formation of pyrans has been observed previously as a side
reaction competing with intermolecular ene reactions.6b
With AlCl3, FeCl3, and SnCl4 33 was the sole product of the
reaction, and so we sought to optimize the formation of
piperidine 29 with MeAlCl2. Performing the reaction at higher
dilution (1 mM in 24 or 27 compared to 20 mM in the earlier
runs) increased the yield of 29, but at the expense of consider-
ably increased reaction times. Happily, the use of 3 equiv of
MeAlCl2 at a 10 mM concentration of either substrate led to an
85:15 mixture of 29:33, with a 68% isolated yield of 29.
Surprisingly, both substrates gave solely the cis piperidine; no
trans isomer was detectable.
In an attempt to quantify the differences in cation stability,
we calculated the structures and relative stabilities of the cis
and trans cations derived from 4a and those derived from
citronellal at the B3LYP/6-31G(d) level of theory. The structures
of these cations are shown in Figures 2 and 3.
In the case of the cis and trans cations derived from citronellal,
the energy difference between the geometry-optimized structures
is 0.19 kcal in favor of the trans cation. This difference is clearly
at odds with the observed diastereoselectivity. However, com-
parison with the corresponding cations derived from 4a is
instructive. The stability of the two cations derived from 4a
(Figure 2b) is reversed, with the cis cation being more stable
that the trans cation by 0.82 kcal. Although this energy
The fact that both substrates afforded the cis piperidine
suggested that the reaction was not concerted, but proceeded
with significant stepwise character (Scheme 11).
The rate-determining step is likely to be the C-C bond-
forming step involving attack of the alkene on the activated
aldehyde, and the strong preference for formation of the cis
product suggests a lower energy pathway for the formation of
the cis carbocation 34 compared with the trans carbocation 35.
Our results for aldehydes 4a-f suggest that under Brønsted
acid catalysis these too cyclize via a pathway with significant
stepwise character. In the cis cation 36, overlap between the
oxygen lone pair and the empty p-orbital at the cationic center
could provide a stabilizing interaction (Figure 1). Such overlap
is geometrically unfavorable in the case of the trans cation 37.
The cyclization catalyzed by MeAlCl2 is more concerted,
although the preference to form the cis product at low temper-
(20) The cyclization of citronellal in aqueous sulfuric acid and in micelles
has been studied. In both cases a preference for cis products is reported.
See: (a) Zimmerman, H. E.; English, J. J. Am. Chem. Soc. 1953, 75, 2367-
2370. (b) Clark, B. C.; Chamblee, T. S.; Iacobucci, G. A. J. Org. Chem.
1984, 49, 4557-4559.
(21) Similar levels of diastereoselectivity have been reported for the
cyclization of citronellal with Et3N-5HF to form the analogous fluorides.
See: Hayashi, E.; Hara, S.; Shirato, H.; Hatakeyama, T.; Fukuhara, T.;
Yoneda, N. Chem. Lett. 1995, 205-206.
J. Org. Chem, Vol. 71, No. 6, 2006 2467

Williams et al.
FIGURE 4. Calculated (B3LYP/6-31G(d)) structures of (a) the cis
transition state derived from citronellal and (b) the trans transition state
derived from citronellal. Carbon atoms are green, oxygen atoms are
red, and hydrogen atoms are white.
FIGURE 2. Calculated (B3LYP/6-31G(d)) structures of (a) the cis
cation derived from citronellal and (b) the trans cation derived from
citronellal. Carbon atoms are green, oxygen atoms are red, and hydrogen
atoms are white.
FIGURE 5. Calculated (B3LYP/6-31G(d)) structures of (a) the cis
transition state derived from compound 4a and (b) the trans transition
state derived from compound 4a. Carbon atoms are green, oxygen atoms
are red, nitrogen atoms are blue, sulfur atoms are yellow, and hydrogen
atoms are white.
FIGURE 3. Calculated (B3LYP/6-31G(d)) structures of (a) the cis
cation derived from compound 4a and (b) the trans cation derived from
compound 4a. Carbon atoms are green, oxygen atoms are red, nitrogen
atoms are blue, sulfur atoms are yellow, and hydrogen atoms are white.
The surprising reversal in cation stability between citronellal
and compound 4a is intriguing, but can be rationalized readily
by examining the structures (Figure 6) of the cations in detail.
In the case of both cations derived from citronellal (Figure 6a,b),
weak C-H‚‚‚O interactions exist between a hydrogen atom on
one of the two methyl groups and the oxygen atom of the
hydroxyl group. By contrast, in the case of the cis cation derived
from compound 4a, this interaction is weaker, as judged by a
longer C-H‚‚‚O and less lengthening of the C-H bond. In the
case of the trans cation, this interaction is completely absents
the hydrogen atom lies well outside the sum of the van der
Waals radii of the interacting atoms. It is not clear from our
calculations why this interaction should be disfavored in the
case of trans 4a. However, it presumably results from the
differing juxtapositions of the interacting groups with respect
to the highly polar sulfonamide.
difference is too small to explain the absolute diastereoselec-
tivities observed in this system, these results suggest that, if
cation stability is a factor, then the reaction involving 4a would
be expected to be markedly more cis selective than that
involving citronellal.
Recognizing that the reaction pathways may be more
concerted under Lewis acidic conditions we calculated the
transition states for the concerted processes involving citronellal
and compound 4a at the B3LYP/6-31G(d) level of theory. The
calculated transition states are shown in Figures 4 and 5.
Although the actual activation barriers in the absence of Lewis
acid are very high, the structures and relative energies of these
transition states are instructive. In the case of citronellal, the
energy difference between the cis and trans concerted transition
states is 0.78 kcal in favor of the trans transition state. In the
case of compound 4a, the cis transition state is more stable by
0.79 kcal. These results suggest that even if the reaction has
some concerted character, the preference for a much more cis
selective process in the case of compound 4a when compared
to citronellal would be maintained.
Others have invoked the importance of electrostatic interac-
tions in calculations on related systems. Houk has shown through
computational modeling that the presence of a heteroatom lone
pair on an enophile can have a large effect on the endo/exo
2468 J. Org. Chem., Vol. 71, No. 6, 2006

Synthesis of 3,4-Disubstituted Piperidines
Experimental Section
Computational Methods Ab initio electronic structure calcula-
tions were carried out with GAMESS24 running on a Linux cluster.
The version dated 22 Nov 2004 was used in all calculations. The
transition states for the cyclizations of citronellal and compound
4a were located by generation of an initial guess, using the linear
synchronous transit (LST) method and then refinement at the HF/
6-31G level of theory within GAMESS. This model transition state
was then used to construct an initial guess for the transition state
leading to the appropriate cis or trans product. This guess was
refined at the B3LYP/6-31G(d) level of theory to a transition state
structure possessing a single imaginary vibration that corresponded
to the reaction coordinate.
Typical Procedure for the Alkylation: Preparation of N-(3-
Hydroxypropyl)-4-methyl-N-(3-methylbut-2-enyl)benzenesulfona-
mide. Cesium carbonate (1.85 g, 5.7 mmol) was added to a solution
of N-(3-hydroxypropyl)-4-methylbenzenesulfonamide (1.0 g, 4.4
mmol) in DMF (25 mL). The reaction mixture was stirred at
ambient temperature for 30 min before 1-bromo-3-methylbut-2-
ene (0.5 mL, 4.4 mmol) was added. The resulting mixture was
stirred for a further 90 min and then poured into water. The aqueous
phase was then extracted with diethyl ether, and the combined
organic phases washed with brine, dried over MgSO4, and
evaporated in vacuo to leave a yellow crystalline solid that was
purified by flash column chromatography (silica; eluent 2:1 hexane:
ethyl acetate) to give N-(3-hydroxypropyl)-4-methyl-N-(3-methyl-
but-2-enyl)benzenesulfonamide (1.13 g, 86%) as a white crystalline
solid: mp 42-43 °C (from hexane/ethyl acetate); Rf 0.28; IR
(Nujol) 3531, 2929, 2875, 1669, 1598, 1337, 1305, 1157 cm-1; 1H
NMR (400 MHz, CDCl3) δ 1.61 (s, 3H), 1.63 (s, 3H), 1.67-1.74
(m, 2H) 2.33 (t, J ) 7.0 Hz, 1H), 2.42 (s, 3H), 3.21 (t, J ) 6.5 Hz,
2H), 3.71-3.76 (m, 2H), 3.80 (d, J ) 7.0 Hz, 2H), 4.93-4.99 (m,
1H), 7.29 (d, J ) 8.0 Hz, 2H), 7.68 (d, J ) 8.0 Hz, 2H); 13C NMR
(125 MHz, CDCl3) δ 17.7, 21.5, 25.7, 31.0, 43.7, 45.8, 58.8, 118.9,
127.2, 129.6, 136.95, 136.97, 143.2; MS (CI) m/z 315 ([M + NH4]+,
4%), 298 ([M + H]+, 100), 247 (25), 230 (26), 143 (45). Anal.
Calcd for C15H23NO3S: C, 60.58; H, 7.79; N, 4.71. Found: C,
60.35; H, 8.0; N, 4.8.
FIGURE 6. C-H‚‚‚O interactions in the calculated (B3LYP/6-31G-
(d)) structures of (a) the cis cation derived from citronellal, (b) the
trans cation derived from citronellal, (c) the cis cation derived from
compound 4a, and (d) the trans cation derived from compound 4a.
Some hydrogen atoms are omitted for clarity. Carbon atoms are green,
oxygen atoms are red, nitrogen atoms are blue, sulfur atoms are yellow,
and hydrogen atoms are white.
stereoselectivity of the ene reaction.22 By modeling the transition
structures of the ene reaction of propene with formaldehyde
imine, he was able to show that electrostatic interactions between
the nitrogen lone pair and the central carbon atom of propene
actually dictate the endo/exo outcome of the reaction. More
recently, Mikami has performed similar calculations on the
Lewis acid-catalyzed carbonyl ene reaction of E-but-2-ene with
glyoxylate, showing that there is a similar electrostatic interac-
tion between the oxygen lone pair and the relevant carbon atom
of the ene component.23
Typical PCC Oxidation Procedure: Preparation of 4-Methyl-
N-(3-methylbut-2-enyl)-N-(3-oxopropyl)benzenesulfonamide (4a)
Celite (2 g) was added to a suspension of PCC (1.09 g, 5.0 mmol)
in CH2Cl2 (10 mL). The resultant slurry was stirred vigorously for
5 min before being cooled to 0 °C. A solution of N-(3-hydroxy-
propyl)-4-methyl-N-(3-methylbut-2-enyl)benzenesulfonamide (1.0
g, 3.36 mmol) in CH2Cl2 (5 mL) was then added in one portion.
After the solution was warmed to room temperature and stirred for
2 h, NaHSO4 (3.0 g) and diethyl ether (25 mL) were added and
the mixture was stirred vigorously for 15 min before being filtered
through silica, washed with diethyl ether, and dried over MgSO4.
Concentration of the filtrate in vacuo followed by flash column
chromatography (silica; eluent 3:1 petroleum ether:ethyl acetate)
gave 4-methyl-N-(3-methylbut-2-enyl)-N-(3-oxopropyl)benzene-
sulfonamide (4a) (0.67 g, 67%) as a colorless oil: Rf 0.42; IR (thin
Conclusion
In summary, we have discovered a highly diastereoselective
synthesis of cis and trans 3,4-disubstituted piperidines from
simple acyclic precursors. Cyclization catalyzed by Brønsted
acids or by Lewis acids at low temperature affords predomi-
nantly the cis product, which is the kinetic product of the
reaction, equilibrating to the thermodynamically more stable
trans isomer under Lewis acid catalysis at higher temperatures.
DFT calculations at the B3LYP/6-31G(d) level of theory indicate
that the cis carbocation that would result from a stepwise
cyclization mechanism is more stable than the trans carbocation,
possibly as a result of stabilizing overlap between the oxygen
lone pair and the empty p-orbital at the carbocationic center.
Calculations (B3LYP/6-31G(d)) of the transition states for
cyclization via a concerted mechanism also indicate that the
cis product should be favored, with the cis transition state being
lower in energy than the corresponding trans transition state.
The method should find application in the synthesis of more
complex molecules, and the switch between kinetic and
thermodynamic control with Brønsted and Lewis acid catalysts
may be applicable to other carbonyl ene cyclizations.
1
film) 2924, 2731, 1723, 1674, 1598, 1338, 1305, 1158 cm-1; H
NMR (300 MHz, CDCl3) δ 1.61 (s, 3H), 1.66 (s, 3H), 2.43 (s,
3H), 2.79 (dt, J ) 1.1, 7.1 Hz, 2H), 3.36 (t, J ) 7.1 Hz, 2H), 3.76
(d, J ) 7.0 Hz, 2H), 4.92-5.00 (m, 1H), 7.30 (d, J ) 8.0 Hz, 2H),
7.70 (d, J ) 8.0 Hz, 2H), 9.75 (t, J ) 1.1 Hz, 1H); 13C NMR (75
MHz, CDCl3) δ 17.9, 21.6, 25.8, 41.0, 44.1, 46.5, 118.7, 127.2,
129.8, 136.3, 137.7, 143.5, 200.6; MS (CI) m/z 313 ([M + NH4]+,
100%), 296 ([M + H]+, 22), 275 (5), 245 (70), 227 (11), 189 (4),
142 (64), 112 (7), 98 (7), 80 (8); HRMS (ES) calcd for C15H21-
NO3NaS, 318.1140 [M + Na]+, found 318.1129 [M + Na]+.
(22) Thomas, B. E.; Houk, K. N. J. Am. Chem. Soc. 1993, 115, 790-
(24) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.;
Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.;
Su, S. J.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem.
1993, 14, 1347-1363.
792.
(23) Yamanaka, M.; Mikami, K. HelV. Chim. Acta 2002, 85, 4264-
4271.
J. Org. Chem, Vol. 71, No. 6, 2006 2469

Williams et al.
TPAP Oxidation Procedure: Preparation of 4-Methyl-N-(2-
ethylidenecyclopentane)-N-(3-oxopropyl)benzenesulfonamide (4c).
Solid TPAP (5 mol %) was added in one portion to a stirred mixture
of N-(3-hydroxypropyl)-4-methyl-N-(2-ethylidenecyclopentane)-
benzenesulfonamide (0.20 g, 0.60 mmol), N-methylmorpholine
N-oxide (0.109 g, 0.90 mmol), and powdered 4 Å molecular sieves
(0.50 g) in CH2Cl2 (10 mL). The reaction mixture was stirred for
75 min and then filtered through silica, eluting with CH2Cl2. The
residue was dried over MgSO4 and concentrated in vacuo. Flash
column chromatography (silica; eluent 2:3 ethyl acetate:petroleum
ether) afforded 4-methyl-N-(2-ethylidenecyclopentane)-N-(3-oxo-
propyl)benzenesulfonamide (4c) (0.17 g, 85%) as a colorless oil:
46.7, 63.0, 112.2, 127.6, 129.6, 133.4, 143.4, 143.8; MS (CI) m/z
296 ([M + H]+, 100%), 278 (4), 257 (5), 189 (8), 143 (13), 124
(15), 105 (4), 79 (11). Anal. Calcd for C15H21NO3S: C, 60.99; H,
7.17; N, 4.74. Found: C, 61.0; H, 7.1; N, 4.6. Further elution
afforded (3S*,4S*)-3-isopropenyl-1-(toluene-4-sulfonyl)piperidin-4-
ol (7a) (4 mg, 4%) as a white crystalline solid, data as above.
Typical Procedure for the Prins-Type Cyclization with HCl-
Saturated CH2Cl2: Preparation of (3R*,4S*)-3-(1-Benzyl-2-
phenylvinyl)-1-(toluene-4-sulfonyl)piperidin-4-ol (6f). Anhydrous
HCl gas was bubbled through a -78 °C solution of 4f (152 mg,
0.34 mmol) in CH2Cl2 for 30 min, followed by stirring at this
temperature for a further 6 h. Removal of the solvent in vacuo and
purification of the residue by flash column chromatography (silica;
eluent 2:1 petroleum ether:ethyl acetate) afforded (3R*,4S*)-3-(1-
benzyl-2-phenylvinyl)-1-(toluene-4-sulfonyl)piperidin-4-ol (6f) (108
mg, 71%) as a white crystalline solid: mp 129-130 °C; Rf 0.38;
1
Rf 0.51; IR (neat) 2947, 2361, 1720, 1450, 1335, 1157 cm-1; H
NMR (300 MHz, CDCl3) δ 1.54-1.66 (m, 4H), 2.13-2.15 (m,
2H), 2.16-2.20 (m, 2H), 2.42 (s, 3H), 2.80 (t, J ) 7.0 Hz, 2H),
3.37 (t, J ) 7.0 Hz, 2H), 3.75 (d, J ) 7.4 Hz, 2H), 5.08-5.15 (m,
1H), 7.30 (d, J ) 8.1 Hz, 2H), 7.67 (d, J ) 8.1 Hz, 2H), 9.75 (s,
1H); 13C NMR (75 MHz, CDCl3) δ 21.5, 25.9, 26.2, 28.8, 33.8,
41.1, 44.0, 48.0, 114.1, 127.3, 129.7, 136.5, 143.3, 149.2, 200.5;
MS (ES) m/z 344 (100%, [M + Na]+).
1
IR (CDCl3) 3438, 3026, 1644, 1493, 1338, 1161 cm-1; H NMR
(300 MHz, CDCl3) δ 1.62-1.72 (m, 1H), 1.74-1.86 (m, 1H), 2.48
(s, 3H), 2.55 (dt, J ) 2.9, 11.4 Hz, 1H), 2.67 (t, J ) 11.4 Hz, 1H),
3.19-3.24 (m, 1H), 3.29-3.38 (m, 1H,), 3.47-3.54 (m, 1H), 3.62
(s, 2H), 4.00 (s, 1H), 6.46 (s, 1H), 7.22 (d, J ) 7.0 Hz, 2H), 7.27-
7.32 (m, 6H), 7.33-7.44 (m, 4H), 7.45 (d, J ) 8.1 Hz, 2H); 13C
NMR (75 MHz, CDCl3) δ 21.5, 32.8, 40.2, 41.8, 42.5, 44.6, 67.7,
126.5, 126.8, 127.6, 128.4, 128.5, 128.7, 128.9, 129.6, 132.1, 133.0,
137.4 139.6, 140.4, 143.3; MS (ES) m/z 470 (100%, [M + Na]+).
Anal. Calcd for C27H29NO3S: C, 72.45; H, 6.53; N, 3.13. Found:
C, 72.5; H, 6.6; N, 2.9.
Typical Procedure for the Ene-Type Cyclization: Prepara-
tion of (3S*,4S*)-3-Isopropenyl-1-(toluene-4-sulfonyl)piperidin-
4-ol (7a). Methyl aluminum dichloride (1 M solution in hexanes,
0.34 mL, 0.34 mmol) was added to a solution of aldehyde 4a (100
mg, 0.34 mmol) in chloroform (10 mL). The resulting mixture was
heated under reflux for 16 h, after which it was quenched by
addition of water. The aqueous phase was then extracted with
diethyl ether and the combined organic phases washed with brine,
dried over MgSO4, and evaporated in vacuo to leave a colorless
oil that was purified by flash column chromatography (silica; eluent
2:1 petroleum ether:ethyl acetate) to give first (3R*,4S*)-3-isopro-
penyl-1-(toluene-4-sulfonyl)piperidin-4-ol (6a) (6 mg, 6%) as a
white crystalline solid, data as below. Further elution afforded
(3S*,4S*)-3-isopropenyl-1-(toluene-4-sulfonyl)piperidin-4-ol (7a) (74
mg, 74%) as a white crystalline solid: mp 149-151 °C (from
petroleum ether/ethyl acetate); Rf 0.24; IR (neat) 3391, 2920, 2904,
Typical Procedure for the Prins-Type Cyclization-Elimina-
tion with HCl-Saturated CH2Cl2 Followed by Ammonia: Prepa-
ration of (3R*,4S*)-3-Cyclohex-1-enyl-1-(toluene-4-sulfonyl)-
piperidin-4-ol (6d). Anhydrous HCl gas was bubbled through a
-78 °C solution of 4d (50 mg, 1.48 mmol) in CH2Cl2 for 30 min,
followed by stirring at this temperature for a further 2 h. The solvent
was removed in vacuo and the residue dissolved in THF (10 mL).
Concentrated aqueous ammonia was added (10 mL) and the mixture
was stirred for 18 h. Following removal of the THF in vacuo the
aqueous phase was extracted with CH2Cl2 and the combined organic
phases washed with brine, dried over MgSO4, and evaporated in
vacuo. Purification of the residue by flash column chromatography
(silica; eluent 2:1 petroleum ether:ethyl acetate) afforded (3R*,4S*)-
3-cyclohex-1-enyl-1-(toluene-4-sulfonyl)piperidin-4-ol (6d) (35 mg,
70%) as a white crystalline solid: mp 94-95 °C (from petroleum
ether/ethyl acetate); Rf 0.62; IR (CDCl3) 2910, 2360, 1670, 1590,
1
1647, 1597, 1458, 1334, 1304, 1157 cm-1; H NMR (500 MHz,
CDCl3) δ 1.58-1.68 (m, 1H), 1.71 (s, 3H), 1.85 (br s, 1H), 2.01-
2.06 (m, 1H), 2.17-2.27 (m, 2H), 2.37 (dt, J ) 2.8, 12.4 Hz, 1H),
2.43 (s, 3H), 3.44 (dt, J ) 4.5, 10.1 Hz, 1H), 3.75-3.77 (m, 1H),
3.81-3.86 (m, 1H), 4.89 (s, 1H), 5.01 (s, 1H), 7.32 (d, J ) 8.0
Hz, 2H), 7.64 (d, J ) 8.0 Hz, 2H); 13C NMR (125 MHz, CDCl3)
δ 20.6, 21.5, 32.5, 45.0, 48.8, 51.6, 69.4, 114.7, 127.6, 129.7, 133.6,
142.4, 143.6; MS (CI) m/z 296 ([M + H]+, 100%), 278 (4), 247
(5), 189 (5), 143 (10), 124 (8), 79 (4), 69 (10). Anal. Calcd for
C15H21NO3S: C, 60.99; H, 7.17; N, 4.74. Found: C, 61.0; H, 7.1;
N, 4.6.
1
1334, 1161 cm-1; H NMR (300 MHz, CDCl3) δ 1.48-1.65 (m,
4H), 1.82-2.06 (m, 5H), 2.25-2.30 (m, 1H), 2.43 (s, 3H), 2.51-
2.61 (m, 2H), 3.50-3.60 (m, 2H), 3.91-3.93 (m, 1H), 5.31 (s,
1H), 7.32 (d, J ) 8.5 Hz, 2H), 7.65 (d, J ) 8.5 Hz, 2H); 13C NMR
(75 MHz, CDCl3) δ 21.5, 22.3, 22.8, 25.2, 28.6, 31.0, 40.6, 43.4,
46.9, 63.2, 123.6, 127.6, 129.7, 133.5, 136.2, 143.4; MS (ES) m/z
358 ([M + Na]+, 100%). Anal. Calcd for C18H25NO3S: C, 64.45;
H, 7.51; N, 4.18. Found: C, 64.5; H, 7.8; N, 4.0.
Typical Procedure for the Prins-Type Cyclization with
Concentrated Hydrochloric Acid: Preparation of (3R*,4S*)-3-
Isopropenyl-1-(toluene-4-sulfonyl)piperidin-4-ol (6a). Concen-
trated hydrochloric acid (37%, 0.10 mL, 1.02 mmol) was added to
a solution of aldehyde 4a (100 mg, 0.34 mmol) in CH2Cl2 (10 mL)
at -78 °C. The resulting mixture was stirred at -78 °C for 16 h,
after which it was quenched by addition of water. The aqueous
phase was then extracted with diethyl ether and the combined
organic phases washed with brine, dried over MgSO4, and
evaporated in vacuo to leave a colorless oil that was purified by
flash column chromatography (silica; eluent 2:1 petroleum ether:
ethyl acetate) to give (3R*,4S*)-3-isopropenyl-1-(toluene-4-sulfonyl)-
piperidin-4-ol (6a) (79 mg, 79%) as a white crystalline solid: mp
89-91 °C (from petroleum ether/ethyl acetate); Rf 0.36; IR (neat)
Typical Procedure for Removal of the p-Toluenesulfonyl
Group: Preparation of (3R*,4S*)-3-Cyclohept-1-enylpiperidin-
4-ol (22e). To a solution of 6e (62 mg, 0.18 mmol) in THF (0.8
mL) at -78 °C was added sodium naphthalenide (0.8 mL of 1 M
solution in THF). After 5 min the reaction was quenched by the
addition of MeOH (0.3 mL) and then allowed to warm to ambient
temperature. The reaction mixture was then diluted with H2O and
acidified to pH 1 with 3 M HCl. The aqueous layer was washed
with Et2O, basified to pH 9 with 3 M NaOH, and extracted with
EtOAc. The combined organic extracts were washed with brine,
dried over MgSO4, and concentrated in vacuo to yield (3R*,4S*)-
3-cyclohept-1-enylpiperidin-4-ol (22e) (25 mg, 71%) as a colorless
1
3519, 2920, 2904, 1651, 1597, 1447, 1323, 1304, 1157 cm-1; H
1
oil: IR (thin film) 3020, 2925, 1522, 1216, 756 cm-1; H NMR
NMR (500 MHz, CDCl3) δ 1.50 (br s, 1H), 1.77 (s, 3H), 1.81-
1.95 (m, 2H), 2.37 (d, J ) 12.1 Hz, 1H), 2.42 (s, 3H), 2.55-2.62
(m, 2H), 3.55-3.59 (m, 2H), 3.96 (d, J ) 2.2 Hz, 1H), 4.58 (s,
1H), 4.96 (s, 1H), 7.31 (d, J ) 8.1 Hz, 2H), 7.64 (d, J ) 8.1 Hz,
2H); 13C NMR (125 MHz, CDCl3) δ 21.5, 22.8, 31.1, 40.5, 43.5,
(300 MHz, CDCl3) δ 1.29-1.43 (m, 2H), 1.44-1.60 (m, 2H),
1.62-1.80 (m, 2H), 1.86-1.95 (m, 1H), 2.03-2.29 (m, 5H), 2.68-
3.03 (m, 5H), 4.00 (br s, 1H), 5.50 (t, J ) 6.3 Hz, 1H); 13C NMR
(75 MHz, CDCl3) δ 29.3, 29.5, 30.6, 34.3, 34.9, 35.8, 42.5, 45.9,
2470 J. Org. Chem., Vol. 71, No. 6, 2006

Synthesis of 3,4-Disubstituted Piperidines
51.9, 65.9, 130.5, 146.5; MS (ES) m/z 196 (100%, [M + Na]+);
HRMS (ES) calcd for C12H22NONa, 196.1701 [M + Na]+, found
196.1694 [M + Na]+.
Supporting Information Available: Experimental procedures
for all compounds not detailed in Experimental Section and 1H and
13C NMR spectra for all compounds; CIF files and ORTEP plots
for all crystal structures reported; coordinate files for the calcula-
tions. This material is available free of charge via the Internet at
http://pubs.acs.org.
Acknowledgment. We thank the Engineering and Physical
Sciences Research Council (studentships to P.S.B and J.T.W)
and the University of Birmingham for financial support, and
Mr Graham Burns for HPLC separations.
JO052532+
J. Org. Chem, Vol. 71, No. 6, 2006 2471