A high yielding route to substituted piperidines via the aza-Diels-Alder reaction

Article abstract of DOI:10.1016/S0040-4039(01)02148-7

The imine Ph₂CHN=CHCO₂Et, generated from benzhydrylamine and ethyl glyoxylate, is an excellent dienophile in aza-Diels-Alder reactions, giving diastereomerically pure cycloadducts in high yield.

Full text of DOI:10.1016/S0040-4039(01)02148-7

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 1067–1070
A high yielding route to substituted piperidines via the
aza-Diels–Alder reaction
Patrick D. Bailey,a,* Peter D. Smith,a Frederick Pederson,a William Clegg,b,c Georgina M. Rosaira
and Simon J. Teatb
aDepartment of Chemistry, Heriot-Watt University, Edinburgh EH14 4AS, UK
bCLRC Daresbury Laboratory, Warrington WA4 4AD, UK
cDepartment of Chemistry, University of Newcastle, Newcastle upon Tyne NE1 7RU, UK
Received 4 September 2001; revised 2 November 2001; accepted 8 November 2001
Abstract—The imine Ph2CHNꢀCHCO2Et, generated from benzhydrylamine and ethyl glyoxylate, is an excellent dienophile in
aza-Diels–Alder reactions, giving diastereomerically pure cycloadducts in high yield. © 2002 Elsevier Science Ltd. All rights
reserved.
The ubiquitous nature of the piperidine ring system in
diverse natural products has ensured that general routes
to this sub-structure have remained an important syn-
thetic goal for many years.1 One particularly powerful
approach is the aza-Diels–Alder reaction,2 especially
when an imino–ester is used as the dienophile (Scheme
1).3–7
Excellent asymmetric induction can be achieved if a
second matched auxiliary is introduced into the ester
functionality (2),6 offering a reliable route to enantio-
pure piperidines; however, the imine requires three
steps for its preparation from (−)-pulegone, and only
one enantiomeric series is readily accessible. An
efficient, diastereospecific aza-Diels–Alder approach
would therefore be very attractive, even if it did not
provide enantio-control; optically active cyclo-adducts
should be readily accessible by resolution because the
basic nature of piperidines allows the formation of
diastereomeric salts with chiral acids.
We initially developed an achiral version of this reac-
tion using PhCH2NꢀCHCO2Et 1a, allowing us to estab-
lish general conditions that yield piperidines in 50–90%
yields in a single step.3 We have been able to extend this
chemistry to an asymmetric procedure by using the
1-phenylethyl auxiliary on nitrogen (1b), available in
either enantiomeric form;4,5 this gives moderate asym-
metric induction for 2-substituted dienes (typically 6:1
diastereo-control), but the minor isomer must be
removed chromatographically, and 1-substituted dienes
form cyclo-adducts with poor asymmetric induction.
Although the aza-Diels–Alder reaction using the achiral
benzyl imine 1a (PhCH2NꢀCHCO2Et) provides an
extremely short and effective route to a range of
pipecolic acid derivatives, we had encountered three
problems:
Me
CO2Et
O
CO2R"
CO2R"
R'
R
N
R
R
O
N
N
Ph
R'
Me
N
1 a, R = H
b, R = Me
c, R = Ph
Me
Me
Ph
Ph
2
Scheme 1. Aza-Diels–Alder reaction.
Keywords: alkaloids; Diels–Alder reaction; piperidines.
* Corresponding author. Present address: Department of Chemistry, UMIST, PO Box 88, Manchester M60 1QD, UK. Tel.: 0161 200 4448; fax:
0161 200 4559; e-mail: p.bailey@umist.ac.uk
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)02148-7

1068
P. D. Bailey et al. / Tetrahedron Letters 43 (2002) 1067–1070
CO2Et
(a) Variable yields, which we eventually attributed to
inconsistency in the purity of the imine
(b) Modest yields (ca. 50%) with 1-substituted dienes
(c) Incomplete diastereo-control with some dienes
(e.g. cis/trans ratio of 1:12 with penta-1,3-diene).
OHC
CO2Et
CO2Et
H+
+
Bn
Bn
N
H
N
H
NH2
Bn
N
Bn
OHC
CO2Et
Concerning the purity of the imine, it is essentially
homogeneous and fridge stable for several weeks pro-
vided (i) absolutely pure ethyl glyoxylate is used; (ii) the
glyoxylate:benzylamine ratio is exactly 1:1; (iii) the
resulting imine is stored under dry, solvent-free condi-
tions at 0°C. However, we were often able to identify
by-products due to attack of additional nucleophiles on
the imine (i.e. PhCH2NH–CH(X)–CO2Et). We also
observed the formation of a further by-product during
the aza-Diels–Alder reactions and, by stirring the imine
under the cycloaddition conditions in the absence of the
diene, we obtained sufficient material for its crystallisa-
tion and identification as 4 by X-ray diffraction (see
Fig. 1). This was presumably formed from the interme-
diate betaine 3, as shown in Scheme 2.
CO2Et
Bn
CO2Et
Bn
CO2Et
Bn
Bn
N
Bn
Bn
N
N
N
N
N
H
3a
O
CO2Et
O
EtOH
Bn
EtO CO2Et
Bn
CO2Et
Bn
Bn
N
N
N
N
3b
O
EtO2C
O
4
CHO
EtO2C
We therefore turned our attention to an alternative
achiral imine, and looked at the benzhydryl derivative
1c. This turned out to be readily obtainable as a white,
homogeneous, stable solid by reacting benzhydrylamine
with ethyl glyoxylate under dehydrating conditions.
Scheme 2. Formation of the imidazolone by-product.
i
CO2Et
OH
(HO)2HC CO2Et
OH
We found the Roberts’ conditions for generating ethyl
glyoxylate hydrate were very convenient,8 and the ease
with which 1c could be reprecipitated in homogeneous
form overcame all problems of purity. Presumably
because it is a solid (cf. 1a, which is an oil), the
benzhydryl imine 1c is indefinitely stable, being usable
after storage for 6–12 months at 0°C. To our delight, it
behaved as an excellent and reliable dienophile in aza-
Diels–Alder reactions (Scheme 3), and the results are
summarised in Table 1.
CO2Et
OHC CO2Et
ii
CO2Et
Ph
CO2Et
iii
R
N
Ph
N
5
1c
Ph
Ph
Scheme 3. Reagents and conditions: (i) NaIO4, CH2Cl2/H2O
5:1, reflux, 1 h then cool to 0°C and add MgSO4 (94%); (ii)
Notable is that we used a standard set of conditions, so
yields are not optimised. Nevertheless, a wide range of
,
Ph2CHNH2 (0.75 equiv.), CH2Cl2, 3 A MS, rt (94%); (iii)
diene (2 equiv.), TFA (1 equiv.), TFE, −40°C (see Table 1,
and Refs. 13–15 for experimental details and data).
methylated butadienes all gave the Diels–Alder adducts
in yields of 42–95%. In all cases, the adduct was
relatively easy to purify, as it was usually the highest Rf
component. We observed total control of both regio-
chemistry (o/p products with respect to nitrogen) and
diastereochemistry (the all-cis products resulting from
pseudo endo attack of the imine on the diene). The
1-substituted dienes have in the past given slightly
lower yields (and polar by-products that are probably
polymeric), and small amounts of the ene products 6
were isolated (Scheme 4) in two cases. However, the
key observation is the rapidity and ease with which the
piperidine ring can be constructed using this methodol-
ogy, and with complete regio- and diastereo-control.
In summary, we have shown that the benzhydryl imine
1c (Ph2CHNꢀCHCO2Et) is a readily prepared, storable
dienophile that reacts with acyclic dienes in high yield
(typically ca. 70%), and with complete regio- and
Figure 1. X-Ray crystal structure of 4.9

P. D. Bailey et al. / Tetrahedron Letters 43 (2002) 1067–1070
1069
Table 1. Cycloadducts 5 from the aza-Diels–Alder reactions, which were single regio- and diastereo-isomers, as shown
entry
diene
cycloadduct 5
yielda
Me
Me
CO2Et
1
95%
N
CHPh2
Me
CO2Et
2
3
4
5
87%
N
CHPh2
CO2Et
62% (5%)
42% (6%)
N
CHPh2
Me
CO2Et
N
Me
CHPh2
Me
Me
CO2Et
60%
N
CHPh2
Me
a
Isolated yield of 5; yield of acyclic by-product 6 shown in brackets, see Scheme 4.
diastereo-control. It therefore provides a particularly
short and attractive route to substituted piperidines
and, because of the double bond within the ring, is also
amenable to further functionalisation. The application
of this approach to natural product synthesis, and the
indirect stereochemical control conferred by the benz-
hydryl group, is discussed in the following paper.
NMR and mass spectra, and Quintiles (Scotland) for
financial support.
References
1. Bailey, P. D.; Millwood, P. A.; Smith, P. D. Chem.
Commun. 1998, 633.
2. (a) Boger, D. L.; Weinreb, S. M. Hetero-Diels–Alder
Methodology in Organic Synthesis; Academic Press:
Orlando, 1987; (b) Waldmann, H. Synthesis 1994, 535; (c)
Buonora, P.; Olsen, J.-C.; Oh, T. Tetrahedron 2001, 57,
6099.
Acknowledgements
We thank Dr. A. S. F. Boyd and Dr. R. Fergusson for
3. Bailey, P. D.; Wilson, R. D.; Brown, G. R. Tetrahedron
Me
Lett. 1989, 30, 6781.
4. Bailey, P. D.; Wilson, R. D.; Brown, G. R. J. Chem. Soc.,
Perkin Trans. 1 1991, 1337.
5. Bailey, P. D.; Brown, G. R.; Korber, F.; Reid, A.;
Wilson, R. D. Tetrahedron: Asymmetry 1991, 2, 1263.
6. Bailey, P. D.; Londesbrough, D. J.; Hancock, T. C.;
Heffernan, J. D.; Holmes, A. B. Chem. Commun. 1994,
2543.
7. Abraham, H.; Stella, L. Tetrahedron 1992, 48, 9707.
8. Hursthouse, M. B.; Malik, K. M. A.; Hibbs, D. E.;
Roberts, S. M.; Seago, A. J. H.; Sik, V.; Storer, R. J.
Chem. Soc., Perkin Trans. 1 1995, 2419.
9. X-Ray data for 4. A small single crystal of 4 was
mounted on a glass wool fibre in perfluoropolyether oil.
The intensity data were collected at 160 K on a specially
adapted Siemens SMART CCD area detector diffrac-
R''
CO2Et
H
N
Ph Ph
Me
R''
CO2Et
Ph
N
6
Ph
Scheme 4. Formation of ene by-product 6 from Table 1,
entries 3/4 (R%%=H/Me).

1070
P. D. Bailey et al. / Tetrahedron Letters 43 (2002) 1067–1070
tometer located at Station 9.8 on the Daresbury Labora-
t, J 7.1 Hz), 4.24 (2H, m), 4.60–5.40 (3H, m, broad); lC
(50 MHz, CDCl3), l: 13.76 (CH3), 62.32 (CH2), 88.32
(CH), 168.87 (CꢀO).
tory Synchrotron Radiation Source.10 Crystal data for 4,
C26H30N2O6, Mw=466.52, monoclinic, space group
,
P21lc, a=13.162(3), b=8.8710(10), c=21.643(3) A, i=
14. Ethyl N-benzhydryliminoethanoate. To a solution of ethyl
glyoxylate hydrate (12.3 g, 120 mM) in dry CH2Cl2 (150
3
−3
,
102.980(13)°, V=2462.4(7) A , Z=4, Dc=1.258 mg m
,
v=0.090 mm−1, T=160(2) K. Crystal size=0.020×0.015×
0.010 mm3, independent reflections 5247 [R(int)=0.0367],
R1=0.0528, wR2=0.1515 for [I>2|(I)]. Data collection
and reduction used XSCANS,11 and structure determina-
tion and refinement used SHELXTL.12
,
ml) was added activated 3 A molecular sieves (ꢀ4 g).
The solution was then stirred at 0°C under Ar atmo-
sphere for about 5 min before addition of benzhydryl-
amine (16.5 g, 90 mM, 0.75 equiv.) via syringe as a
solution in dry CH2Cl2 (10 ml). The cloudy solution was
left to stir for 16 h at 4°C. Filtration to remove molecular
sieves followed by removal of solvent under reduced
pressure (30°C) gave a yellow oil which started to crys-
tallise. This was triturated with 40/60 pet. ether to
provide, after filtration, white crystalline product (22.6 g,
94%), mp 55–56°C. Data lH (200 MHz, CDCl3), l: 1.34
(3H, t, J 7.1 Hz), 4.34 (2H, q, J 7.1 Hz), 5.68 (1H, s),
7.19–7.39 (10H, m), 7.77 (1H, s); lC (50 MHz, CDCl3), l:
14.07 (CH3), 61.75 (CH2), 77.39 (CH), 127.44 (CH),
127.75 (CH), 128.54 (CH), 141.46 (C), 153.68 (CH),
163.08 (CꢀO); w (Nujol, cm−1): 2940.2, 2925.5, 2854.6,
1740.9, 1493.6, 1454.8, 1211.4, 1015.8, 737.3, 697.0; m/z
(EI) 535.6 (0.19%, 2M+1), 268.3 (5.2%, M+1), 167.1
(100%); HRMS, calcd for C17H17NO2: 267.126, found:
267.121.
10. (a) Cernik, R. J.; Clegg, W.; Catlow, C. R. A.; Bushnell-
Wye, G.; Flaherty, J. V.; Greaves, G. N.; Burrows, I.;
Taylor, D. J.; Teat, S. J.; Hamichi, M. J. Synchrotron
Rad. 1997, 4, 279; (b) Clegg, W.; Elsegood, M. R. J.;
Teat, S. J.; Redshaw, C.; Gibson, V. C. J. Chem. Soc.,
Dalton Trans. 1998, 3037.
11. XSCANS Data collection and reduction program, 1994,
Version 2.2, Bruker AXS, Madison, Wisconsin, USA.
12. Sheldrick, G. M. SHELXTL, Structure determination and
refinement programs, 1999, Version 5.1 Bruker AXS,
Madison, Wisconsin, USA.
13. Ethyl glyoxylate hydrate.8 A solution of diethyl-
L-tartrate
(20.9 g, 101.5 mM) in CH2Cl2 (200 ml) was rapidly
stirred (motorised stirrer) at room temperature in a flask
equipped with a reflux condenser. To this was added solid
sodium periodate (43.4 g, 203 mM, 2 equiv.) then H2O
(40 ml, approx. 20% by volume). The biphasic mixture
was heated to reflux with rapid stirring for ꢀ1 h, where-
upon a thick white precipitate had formed. Over the
course of a further hour, the solution became ever more
cloudy, with the white precipitate had to be manually
broken up on several occasions. TLC analysis (ammo-
nium molybdate/D visualisation) indicated complete con-
sumption of the tartrate, at which point the solution was
cooled to 0°C. At this point MgSO4 (approximately 80 g)
was added in portions to the stirred mixture (CAUTION:
exotherm), and the, by now, very thick white reaction
mixture was stirred for 15 min before all solids were
removed by filtration. The solids were washed with a
further portion of CH2Cl2 (200 ml), before the filtrate
solution was evaporated under reduced pressure to reveal
the ethyl glyoxylate hydrate, in the form of an opaque oil
(22.8 g, 94%). Data: lH (200 MHz, CDCl3), l: 1.31 (3H,
15. General procedure for the aza-Diels–Alder reaction using
ethyl N-benzhydryl-iminoethanoate. A solution of ethyl
N-benzhydryliminoethanoate in trifluoroethanol (10 ml/
g) is cooled to −40°C under Ar. To this solution is added,
via syringe, trifluoroacetic acid (1 equiv.), immediately
followed by slower syringe addition of the required diene
(2 equiv. or more). The solution is then left to stir at
−40°C until reaction is deemed complete (as indicated by
TLC monitoring, usually ꢀ20 min), whereupon the reac-
tion is warmed to room temperature. After evaporation
of the solvent under reduced pressure, the reaction mix-
ture is taken up in Et2O. Successive washings with satd
NaHCO3 solution (3×), then H2O (1×), then brine (1×)
and drying over MgSO4 followed by removal of solvent
under reduced pressure gives the crude product. Further
purification by column chromatography (SiO2) may be
carried out if necessary.