A flexible strategy for the synthesis of tri- and tetracyclic lupin alkaloids: Synthesis of (+)-cytisine, (±)-anagyrine, and (±)-thermopsine

Article abstract of DOI:10.1002/anie.200504015

(Chemical Equation Presented) Looping the lupin: A general synthetic strategy for the construction of the lupin alkaloids has been defined and involves sequential formation of the N1-C10 and C6-C7 bonds, which are common to compounds of this class of natu

Full text of DOI:10.1002/anie.200504015

Angewandte
Chemie
Natural Product Synthesis
DOI: 10.1002/anie.200504015
A Flexible Strategy for the Synthesis of Tri- and
Tetracyclic Lupin Alkaloids: Synthesis of
(+)-Cytisine, (ꢀ )-Anagyrine, and
(ꢀ )-Thermopsine**
Diane Gray and Timothy Gallagher*
The lupin alkaloids, which comprise a varied range of
primarily bi-, tri-, and tetracyclic molecules based on a
quinolizidine core, represent an increasingly important group
of natural products that exhibit a diverse array of properties.[1]
(ꢁ)-Cytisine (1),[2,3] which is a potent and a4b2 subtype-
selective partial agonist at nicotinic acetylcholine receptors,
was recently exploited successfully by Coe, OꢀNeill et al. as
the natural product lead for the discovery of varenicline.[4] In
a quite different realm, the potential (and also limitations) of
(ꢁ)-sparteine (2) as a reagent and ligand for asymmetric
synthesis are now well established.[5–7] Many key structural
relationships between important members of the lupin
alkaloids are known.[1] For example, reduction of (+)-
anagyrine (3) gives (ꢁ)-sparteine (2), while reduction of
(ꢁ)-thermopsine (4), the C11 epimer of anagyrine, leads to
(ꢁ)-a- and (+)-b-isosparteine. Given this background, the
development of general and therefore flexible synthetic
entries to key lupin alkaloid targets becomes an attractive
challenge; earlier contributions in this area by van Tamelen
and Baran[3a,d] are of particular note.
We previously reported a novel approach to (ꢀ )-cyti-
sine.[8a] Although highly convergent, this chemistry had
limitations in terms of its broader utility.[8b] Herein, we
[*] D. Gray, Prof. T. Gallagher
School of Chemistry
University of Bristol
Bristol BS81TS (UK)
Fax: (+44)117-929-8611
E-mail: t.gallagher@bristol.ac.uk
[**] We thank Dr. Ernest Boehm (Apin Chemicals Ltd.) for an authentic
sample of natural anagyrine and the EPSRCfor financial support.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 2419 –2423
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
2419

Communications
report a second-generation approach that provides both a
more efficient and also an asymmetric entry to cytisine.
Furthermore, the core strategy focuses on establishing both
two steps). In this way, non-natural (+)-cytisine ((+)-1)[13] was
obtained in 10 steps in 3.7% overall yield from commercially
available materials and also allowed us to assign the absolute
stereochemistry of both (5R)-5 and (5S)-6 as indicated in
Scheme 1.
The broader applicability of this synthetic strategy is
exemplified for two representative tetracyclic lupin alkaloids,
anagyrine (3)[14,15] and its C11 epimer thermopsine (4),[15,16] as
shown in Scheme 2 and Scheme 3. Anagyrine (3) was first
isolated in 1895 from Anagyris foetida, and previous synthe-
ses of anagyrine were reported by van Tamelen and Baran,[3d]
Goldberg and Lipkin,[14a] and most recently by Padwa et al.[14b]
Our approach to 3 is shown in Scheme 2, and again it targeted
ꢁ
ꢁ
N1 C10 and C6 C7 bond constructions, which are not only
suited to cytisine (1) but also encompass a broader range of
lupin alkaloids. The more general applicability of our
approach is demonstrated here by the synthesis of both
anagyrine (3) and thermopsine (4).
Our approach to cytisine (1) is shown in Scheme 1 and
involves the efficient enzymatic resolution of ester 5,[8,9] which
ꢁ
ꢁ
the N1 C10 and C6 C7 bonds as the key disconnections,
requiring the quinolizidine-based bromide 13[17,18] as a
substrate for alkylation of 2-pyridone. Bromide 13 was
Scheme 1. Synthesis of (+)-cytisine ((+)-1). a) See Ref. [9], 56%;
b) 10% Pd/C, H2 (1 atm), Na2CO3, EtOH, 95%; c) a-chymotrypsin,
0.1m phosphate buffer, acetone, pH 7.4, (5R)-5 (42% yield, >98% ee)
and (5S)-6 (48% yield, 64% ee after derivatization to its methyl ester
for HPLCanalysis); d) LiAlH 4, THF; e) PBr3, toluene, 57% overall yield
from (5R)-5; f) 2-pyridone, K2CO3, Bu4NBr, H2O, PhMe, reflux, 66%
(+15% of the corresponding 2-alkoxypyridine); g) LHMDS (2 equiv),
THF, 708C, sealed tube, 15 h, 94%; h) MnO2, C HCl2, 79%;
2
i) BH3·THF, 08C!RT; j) Pd(OAc)2, H2 (1 atm), MeOH, HCl, 60%
overall from 10. Bn=benzyl; LHMDS=lithium hexamethyldisilazide.
Scheme 2. Synthesis of (ꢀ)-anagyrine (3). a) LHMDS, EtOCH2Cl, THF,
ꢁ788C!RT, 87%; b) KOtBu, THF, ꢁ788C, 84%; c) MeCO2tBu, LDA,
ꢁ788C, THF, 68%; d) 10% Pd/C, H2 (1 atm), 100%; e) AcOH,
toluene, reflux, 73%; f) LiAlH4, THF, then PBr3, toluene, 70% overall
yield; g) 2-pyridone, K2CO3, Bu4NBr, H2O, toluene, reflux, 66% (+14%
of the corresponding 2-alkoxypyridine); h) LDA (2 equiv), THF, room
is readily available in a two-pot operation from simple and
commercially available starting materials. Resolution of 5 was
achieved using a-chymotrypsin to give (5R)-5 in 42% isolated
yield and with over 98% ee (as determined by chiral
HPLC).[10] The acid (5S)-6 was obtained in 48% yield in
64% ee, and the absolute configuration of (5R)-5 was
established by its conversion into (+)-cytisine[3j] (the non-
natural enantiomer). Reduction of ester (5R)-5 and bromi-
nation provided (5R)-7 (57% overall), and reaction of 7 with
2-pyridone under basic conditions gave the N-alkylated
temperature, 3 h, 44%; i) MnO2, C HCl2, 76%; j) BH3·THF, 08C!RT,
2
85%. LDA=lithium diisopropylamide; AcOH=acetic acid.
ꢁ
adduct 8 (66%), so establishing the N1 C10 linkage.
The second key feature of our strategy is formation of the
ꢁ
C6 C7 bond, which is a structural feature common to many
lupin alkaloids. Previously, this bond was set using a Pd0-
mediated intramolecular a-arylation of the lactam unit,[8] but
ꢁ
the same C6 C7 linkage can now be made more simply.
Exposure of lactam 8 to LHMDS (or LDA) led to smooth
cyclization to give the internal 1,6-addition adduct 9 as a
single diastereomer (and alkene regioisomer) in 94% yield,
with the stereochemistry at C6 of 9 confirmed by NOE
studies.[12] Intramolecular 1,6-additions to 2-pyridone have
not previously been reported, and indeed there are only a few
examples known of the corresponding intermolecular pro-
cess.[11] Mild oxidation of 9 gave lactam 10 in 79% yield, which
was then converted through two steps[8] into (+)-cytisine by
reduction of the lactam and debenzylation (60% yield over
=
Scheme 3. Synthesis of (ꢀ)-thermopsine (4). a) LHMDS, (EtO2C)2C
CH2, THF, ꢁ788C!RT, 79%; b) TFA, CH2Cl2, THF, 08C, 100%;
c) AcOH, toluene, reflux, 89%; d) NaCl, H2O, DMSO, 1308C , 72 h,
72%; e) LiAlH4, THF, then PBr3, toluene, 72% overall yield; f) 2-
pyridone, K2CO3, Bu4NBr, H2O, toluene, reflux, 71% (+12% of the
corresponding 2-alkoxypyridine); g) LDA (2 equiv), THF, room temper-
ature, 2.5 h, 26%; h) MnO2, C HCl2, 70%; i) BH3·THF, 08C!RT, 57%.
2
TFA=trifluoroacetic acid; DMSO=dimethyl sulfoxide.
2420 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2419 –2423

Angewandte
Chemie
prepared through diastereoselective 1,4-addition[19] of the
enolate derived from tert-butyl acetate to the a,b-unsaturated
ester 11, which was available in two steps from racemic ethyl
N-benzyl homopipecolate. N-Debenzylation of the 1,4-addi-
tion adduct and then acid-mediated cyclization led to 12, and
reduction of the ester and bromination provided the requisite
bromide 13. Again, this intermediate underwent clean N-
alkylation with 2-pyridone, and cyclization of 14 under basic
conditions proceeded to give 15 in 44% yield. This step serves
to complete the tetracyclic core of anagyrine as a single
diastereomer, with the relative stereochemistry at C6 that
corresponds to that required for sparteine, as again deter-
mined by NOE studies.[20] Oxidation of 15, followed by
selective reduction of lactam 16 then gave (ꢀ )-anagyrine (3),
which we were able to compare directly with an authentic
sample of the natural product.[21]
The synthesis of (ꢀ )-thermopsine (4) was achieved by
application of the same strategy. The only significant require-
ment here was for the epimeric quinolizidine derivative
20,[17,18] which was prepared efficiently as shown in Scheme 3.
The key stereocenters were established by addition of the
enolate derived from (ꢀ )-17 to freshly prepared diethyl
methylenemalonate to give adduct 18 as a single diastereo-
mer. The stereochemical outcome of this key step was
consistent with that reported for the corresponding allylation
reaction, as reported by Knight and co-workers.[22] Cleavage
of the Boc residue from 18 and thermally induced cyclization
were followed by selective ester cleavage and decarboxyla-
tion to give 19, the structure and relative stereochemistry of
which was confirmed by single-crystal X-ray analysis.[23]
Conversion of 19 into bromide 20 followed by N-alkylation
of 2-pyridone gave 21. This intermediate underwent base
(LDA)-induced cyclization to provide 22 in 26% yield.[24,25] In
the same fashion as described for anagyrine, oxidation of 22
and subsequent reduction of the lactam in 23 gave (ꢀ )-
thermopsine (4). Although an authentic sample of thermop-
sine was not available, we were able to compare spectroscopic
data for our synthetic material with that available in the
literature[15c,d] and our crystallographic assignment of ester 19
underpins this conclusion.
In summary, we have disclosed a synthetic strategy that is
broadly applicable to representative lupin alkaloids, with
syntheses of anagyrine (3) and thermopsine (4) also consti-
tuting formal routes to sparteine (2) and a- and b-isospar-
teines, respectively. More significantly, however, the chemis-
try that we have described is highly convergent, concentrating
as it does on two key and common bonds within the core of
these target structures. This approach also offers new
opportunities to develop novel and significantly different
variants of important targets, including both cytisine and
sparteine, which will complement currently available meth-
ods. Our primary objective has been to validate the viability
of the basic synthetic strategy, but asymmetric entries are also
important and are being pursued.[18] However, the corner-
stone of this new approach is the intramolecular 1,6-addition
of a mono- or bicyclic lactam enolate to an N-alkylated
pyridone as exemplified by 8!9 (94%), 14!15 (44%), and
21!22 (26%). Such 1,6-addition processes are without direct
precedent, and this step needs to be examined in more detail
as it is clearly sensitive to the steric environment of the
substrate.[25] This type of process has applications elsewhere,
and further studies to understand more fully, optimize, and
extend the scope of this process are underway.[26]
Received: November 11, 2005
Published online: March 7, 2006
Keywords: alkaloids · cyclization · natural products ·
.
nucleophilic addition · synthetic methods
[1]N. J. Leonard in The Alkaloids, Vol. 3 (Eds: R. H. F. Manske,
H. L. Holmes), Academic Press, New York, 1953, p. 120; H. C. S.
Wood, R. Wrigglesworth in Roddꢀs Chemistry of Carbon
Compounds, Vol. IV, part H (Ed: S. Coffey), Elsevier, Amster-
dam, 1978, chap. 38, p. 285; K. A. Aslanov, Y. K. Kushmuradov,
A. S. Sadykov in The Alkaloids, Vol. 31 (Ed: A. Brossi),
Academic Press, New York, 1997, p. 117.
[2]H. Ing, J. Chem. Soc. 1932, 2778; A. Partheil, Arch. Pharm. 1894,
232, 161. For a recent overview of nicotinic pharmacology
relevant to cytisine, see: A. A. Jensen, B. Frølund, T. Lijefors, P.
Krogsgaard-Larsen, J. Med. Chem. 2005, 48, 4705.
[3]For previous syntheses of cytisine, see ref. [8]and: a) E. E.
van Tamelen, J. S. Baran, J. Am. Chem. Soc. 1955, 77, 4944; b) F.
Bohlmann, A. English, N. Ottawa, H. Sander, W. Weise, Chem.
Ber. 1956, 89, 792; c) T. R. Govindachari, S. Rajadurai, M.
Subramanian, B. S. Thyagarajan, J. Chem. Soc. 1957, 3839;
d) E. E. van Tamelen, J. S. Baran, J. Am. Chem. Soc. 1958, 80,
4659; e) B. T. OꢀNeill, D. Yohannes, M. W. Bundesmann, E. P.
Arnold, Org. Lett. 2000, 2, 4201; f) J. W. Coe, Org. Lett. 2000, 2,
4205; g) P. Nshimyumukiza, D. Cahard, J. Rouden, M. C. Lasne,
J. C. Plaquevent, Tetrahedron Lett. 2001, 42, 7787; h) D. Stead, P.
OꢀBrien, A. J. Sanderson, Org. Lett. 2005, 7, 4459; for asym-
metric routes to cytisine, see: i) B. Danieli, G. Lesma, D.
Passarella, A. Sacchetti, A. Silvani, A. Virdis, Org. Lett. 2004,
6, 493; for the synthesis of (+)-cytisine, see: j) T. Honda, R.
Takahashi, H. Namiki, J. Org. Chem. 2005, 70, 499.
[4]a) J. W. Coe, M. G. Vetelino, C. G. Bashore, M. C. Wirtz, P. R.
Brooks, E. P. Arnold, L. A. Lebel, C. B. Fox, S. B. Sands, T. I.
Davis, D. W. Schulz, H. Rollema, F. D. Tingley, B. T. OꢀNeill,
Bioorg. Med. Chem. Lett. 2005, 15, 2974; b) J. W. Coe, P. R.
Brooks, M. G. Vetelino, M. C. Wirtz, E. P. Arnold, J. H. Huang,
S. B. Sands, T. I. Davis, L. A. Lebel, C. B. Fox, A. Shrikhande,
J. H. Heym, E. Schaeffer, H. Rollema, Y. Lu, R. S. Mansbach,
L. K. Chambers, C. C. Rovetti, D. W. Schulz, F. D. Tingley, B. T.
OꢀNeill, J. Med. Chem. 2005, 48, 3474. Varenicline (Champix)
has recently been granted a priority review by the US FDA and
is currently in late-stage clinical trials as a smoking cessation
agent.
[5]a) D. Hoppe, T. Hense, Angew. Chem. 1997, 109, 2376; Angew.
Chem. Int. Ed. Engl. 1997, 36, 2283; b) T. Schutz, Synlett 2003,
901.
[6]For the synthesis, evaluation, and computational aspects of
sparteine mimics, see: a) M. J. Dearden, C. R. Firkin, J. P. R.
Hermet, P. OꢀBrien, J. Am. Chem. Soc. 2002, 124, 11870; b) M. J.
Dearden, M. J. McGrath, P. OꢀBrien, J. Org. Chem. 2004, 69,
5789; c) P. OꢀBrien, K. B. Wiberg, W. F. Bailey, J. P. R. Hermet,
M. J. McGrath, J. Am. Chem. Soc. 2004, 126, 15480; d) P. W.
Phuan, J. C. Ianni, M. C. Kozlowski, J. Am. Chem. Soc. 2004, 126,
15473.
[7]For asymmetric syntheses of sparteine, see: a) B. T. Smith, J. A.
Wendt, J. Aube, Org. Lett. 2002, 4, 2577; b) J. P. R. Hermet, M. J.
McGrath, P. OꢀBrien, D. W. Porter, J. Gilday, Chem. Commun.
2004, 1830; a number of approaches to racemic sparteine have
also been reported; see: T. Buttler, I. Fleming, S. Gonsior, B. H.
Angew. Chem. Int. Ed. 2006, 45, 2419 –2423
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2421

Communications
Kim, A. Y. Sung, H. G. Woo, Org. Biomol. Chem. 2005, 3, 1557,
and references therein.
1998, 36, 779. Mikhova and Duddeck[15d] report a comprehensive
listing of 13C NMR data for lupin alkaloids, including both
anagyrine and thermopsine.
[8]a) C. Botuha, C. M. S. Galley, T. Gallagher, Org. Biomol. Chem.
2004, 2, 1825; b) We have already examined the feasibility of
extrapolating this first-generation synthetic approach to ther-
mopsine (4). However, problems were encountered in both the
N-alkylation of 6-bromo-2-pyridone using bromide 20 to estab-
[16]Earlier syntheses of thermopsine were reported by van Tamelen
and Baran[3d] and Bohlmann et al. (F. Bohlmann, E. Winterfeldt,
H. Overwien, H. Pagel, Chem. Ber. 1962, 95, 944).
[17]A number of different approaches to racemic quinolizidines
esters and halides 12/19 and 13/20, respectively, have already
been reported[17a–h], but attempts to produce asymmetric variants
of these have had limited success.[17h] New synthetic approaches
to both 12 and 19 have been outlined herein (Schemes 2 and 3)
with the longer term aim of using these routes to provide
enantiomerically pure substrates. These studies are currently
underway.[18] a) F. Bohlmann, E. Winterfeldt, H. Laurent, W.
Ude, Tetrahedron 1963, 19, 195; b) T. Kappe, Monatsh. Chem.
1967, 98, 1852; c) B. Lal, D. N. Bhedi, H. Dornauer, N. J.
De Souza, J. Heterocycl. Chem. 1980, 17, 1073; d) M. L. Brem-
merr, S. M. Weinreb, Tetrahedron Lett. 1983, 24, 261; e) M. Ihara,
T. Kirihara, K. Fukumoto, T. Kametani, Heterocycles 1985, 23,
1097; f) M. Ihara, T. Kirihara, A. Kawaguchi, M. Tsuruta, K.
Fukumoto, J. Chem. Soc. Perkin Trans. 1 1987, 1719; g) T.
Nagasaka, H. Yamamoto, H. Hayashi, H. Kato, M. Kawaida, K.
Yamaguchi, F. Hamaguchi, Heterocycles 1989, 29, 1209; h) M. J.
Wanner, G. J. Koomen, Tetrahedron 1991, 47, 8431; i) E. D.
Edstrom, Tetrahedron Lett. 1991, 32, 5709.
[18]Methyl homopipecolate, the starting material used in Schemes 2
and 3, is now available in enantiomerically pure form by either
an efficient homologation process[18a] or by enzymatic resolu-
tion.[18b] a) D. Gray, C. Concellón, T. Gallagher, J. Org. Chem.
2004, 69, 4849; b) C. Pousset, R. Callens, M. Haddad, M.
LarchevÞque, Tetrahedron: Asymmetry 2004, 15, 3407.
[19]OꢀBrien and co-workers reported a related diastereoselective
1,4-addition reaction in their synthesis of (+)-sparteine.[7b] This
reaction used an N-a-methylbenzyl moiety to induce a very
modest level of diastereoselectivity in a cyclization reaction
leading to the piperidine ring. For a related alkylative cyclization
strategy also based on N-a-methylbenzyl as a chiral-directing
group that generates a very similar intermediate to that used by
OꢀBrien and co-workers, see: D. N. A. Fox, D. Lathbury, M. F.
Mahon, K. C. Molloy, T. Gallagher, J. Am. Chem. Soc. 1991, 113,
2652.
ꢁ
lish the N1 C10 linkage: O-alkylation predominated. This result
provided the motivation to find more straightforward solutions
ꢁ
ꢁ
to the formation of the N1 C10 and C6 C7 bonds.
[9]G. R. Cook, L. G. Beholz, J. R. Stille, J. Org. Chem. 1994, 59,
3575.
[10]Felluga et al. reported the enzymatic resolution of a series of N-
substituted pyrrolidinone analogues of 5: F. Felluga, G. Pitacco,
M. Prodan, S. Pricl, M. Visintin, E. Valentin, Tetrahedron:
Asymmetry 2001, 12, 3241. The enantiomeric excess of ester
(5R)-5 was determined by chiral HPLC (Chiralcel OJ column)
using racemic 5 as a standard. The enantiomeric excess of acid
(5S)-6 was determined by its conversion into ester (5S)-5 using
HCl in MeOH. We have observed the same sense of asymmetric
differentiation, that is, selective hydrolysis of (5S)-5 to (5S)-6, as
seen by Felluga et al. in the five-membered lactam series.
[11]For rare examples of intermolecular 1,6-additions of alkyl
lithium derivatives to 2-pyridones that lack additional activating
(electron-withdrawing) substituents, see: a) P. Meghani, J. A.
Joule, J. Chem. Soc. Perkin Trans. 1 1988, 1; b) E. W. Thomas, J.
Org. Chem. 1986, 51, 2184; an interesting photochemical 1,6-
addition to 2-pyridone has been suggested to involve electron
transfer and radical coupling; see: N. Sakurai, S. Ohmiya, J.
Chem. Soc. Chem. Commun. 1993, 297.
[12]Characterization data for all compounds reported herein are
available in the Supporting Information. NOE experiments were
carried out to determine the stereochemistry of 9. Irradiation of
C6-H showed an enhancement of both C8-Hax (2.8%) and C10-
Hax (6.2%). Irradiation of C8-Hax showed an enhancement of
both C6-H (9.1%) and C10-Hax (4.8%).
[13]Natural ( ꢁ)-cytisine is commercially available, and we used a
commercial sample of it to compare with our synthetic material.
A lack of information in the literature as to the pharmacological
profile of (+)-cytisine[26] prompted our interest in this enantio-
mer as our initial target, and (+)-cytisine is currently under
evaluation as a nicotinic ligand. Optical rotation data for
synthetic (+)-cytisine were recorded: [a]2D0 = + 120 (c = 0.1,
EtOH); (literature values: for (ꢁ)-1 [a]D = ꢁ110 (c = 0.5,
[20]NOE experiments were carried out to determine the stereo-
chemistry of 15. Irradiation of C6-H showed an enhancement of
C10-Hax (3.9%). Enhancement of C8-Hax was also observed but
was complicated because this signal overlapped with C13-Hax
and C13-Heq. As a result, selective irradiation of C8-Hax was not
possible. Irradiation of C10-Hax showed an enhancement of both
C6-H (7.9%) and C8-Hax, although the latter enhancement
could not be quantified.
EtOH)[13a]
;
for (+)-1 [a]D = + 113.5 (c = 0.3, EtOH)[3j]).
a) Y.-H. Wang, J.-S. Li, H. Kubo, K. Higashiyama, H. Komiya,
S. Ohmiya, Chem. Pharm. Bull. 1999, 47, 1308.
[14]The first synthesis of anagyrine was reported by van Tamelen
and Baran,[3d] who exploited related chemistry to synthesise
cytisine[3a-d] and thermopsine[3d]. For later routes to anagyrine,
see: a) S. I. Goldberg, A. H. Lipkin, J. Org. Chem. 1972, 37, 1823;
b) A. Padwa, T. M. Heidelbaugh, J. T. Kuethe, J. Org. Chem.
2000, 65, 2368.
[15]The crystal structure reported supposedly for anagyrine (A. U.
Rahman, A. Pervin, M. I. Choudhary, N. Hasan, B. Sener, J. Nat.
Prod. 1991, 54, 929) corresponds, in fact, to thermopsine, and this
has been the subject of a query note placed on the file in the
Cambridge Crystallographic Data Centre. The authors of the
original paper, however, have not commented on this apparent
error. There has been some dispute over the stereochemical
assignments of anagyrine and thermopsine; for a discussion of
the issues involved, see: a) D. S. Rycroft, D. J. Robins, I. H.
Sadler, Magn. Reson. Chem. 1991, 29, 936; b) Z. M. Liu, L. Yang,
Z. J. Jia, J. H. Chen, Magn. Reson. Chem. 1992, 30, 511. This
problem has since been resolved and the original assignments
confirmed: c) D. J. Robins, D. S. Rycroft, Magn. Reson. Chem.
1992, 30, 1125; d) B. Mikhova, H. Duddeck, Magn. Reson. Chem.
[21]A sample of natural anagyrine was kindly provided by Dr.
Ernest Boehm (Apin Chemicals Ltd.) and used for comparison
purposes (1H and 13C NMR spectroscopic data and TLC). We
were able to correlate 1H and 13C NMR data for synthetic
anagyrine (and synthetic thermopsine) with that reported by
Robins et al.,[15a,c] and Mikhova and Duddeck.[15d] The region of
the 13C NMR spectra[15d] displaying the resonances for C7 C15
ꢁ
for these two isomers is quite distinct, and full details are
available in the Supporting Information.
[22]C. Morley, D. W. Knight, A. C. Share, J. Chem. Soc. Perkin Trans.
1 1994, 2903; for related diastereoselective alkylations involving
N-benzyl piperidines, see: S. Ledoux, J. P. CØlØrier, G. Lhommet,
Tetrahedron Lett. 1999, 40, 9019; S. Ledoux, E. Marchalant, J. P.
CØlØrier, G. Lhommet, Tetrahedron Lett. 2001, 42, 5397.
[23]Crystal data for lactam 19: C11H17NO3, Mr = 211.26, colorless
plate, 0.50 ” 0.50 ” 0.10 mm3, MoKa radiation (l = 0.71073 Š) was
used, and intensity data were collected as w scans (frames, 0.38
width, 2qmax = 508); a multiscan absorption correction (G. M.
2422 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 2419 –2423

Angewandte
Chemie
Sheldrick, SADABS v2.03, University of Göttingen, Germany,
2003) was applied, (m = 0.094 mmꢁ1, Tmax = 1.00, Tmin = 0.94), the
structure was solved and refined by standard techniques (G. M.
Sheldrick, SHELXS-97, University of Göttingen, Germany,
1990; G. M. Sheldrick, SHELXL-97, University of Göttingen,
Germany, 1997); triclinic crystal system, a = 9.3234(19), b =
10.175(2), c = 11.724(2) Š, a = 88.69(3), b = 75.71(3), g =
88.28(3)8, V= 1077.2(4) Š3, 1 = 1.303, T= 173(2) K, space
¯
group P1, Z = 4, Rint = 0.0456 (for 10271 measured reflections),
273 parameters were used in the refinement, hydrogen atoms
were constrained to ideal geometries and refined with displace-
ment parameters equal to 1.5 times (methyl H atoms) or
1.2 times (all other H atoms) Ueq of their parent atom; largest
difference electron density map features were + 0.192,
ꢁ0.191 eŠꢁ3, R1 = 0.0423 [for 2532 unique reflections with
> 2s(I)], wR2 = 0.0994 (for all 3787 unique reflections).
CCDC 289282 contains the supplementary crystallographic
data for this paper. These data can be obtained free of charge
from the Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data request/cif.
[24]NOE experiments were carried out to determine the stereo-
chemistry of 22. Irradiation of C6-H showed an enhancement of
both C8-Hax (2.2%) and C10-Hax (3.4%). Irradiation of C10-Hax
showed an enhancement of both C6-H (4.4%) and C8-Hax
(3.2%).
[25]For cyclization to occur, there is a requirement that the enolates
derived from bicyclic lactams 14 and 21 maintain the pyridone
substituent associated with C1 (quinolizidine numbering) in an
axial environment. Interestingly, the crystal structure of bicyclic
lactam 19 (see Supporting Information) shows the CO2Me
substituent at C1 to be orientated in a pseudoequatorial
environment, and, significantly, the cyclization of 21 (derived
from 19 and which leads to thermopsine) is the least efficient of
those we have examined to date. However, at this time, the
factors that moderate the efficiency and scope of this reaction
have not yet been defined, and the processes reported here
remain unoptimized.
[26]Note added in proof: The activity of ( +)-cytisine at the
a4b2* nicotine acetylcholine receptor labeled with [3H]-epiba-
tidine has been determined: Ki = 0.79 ꢀ 0.30 mm for (+)-cytisine
versus Ki = 3.0 ꢀ 0.65 nm for (ꢁ)-cytisine. This low level of
activity could also be attributed to a small (ca. 0.4%) level of
contamination by (ꢁ)-cytisine as a consequence of the kinetic
resolution step involved. We thank Professor Susan Wonnacott
and Neal Innocent (University of Bath) for this determination.
Angew. Chem. Int. Ed. 2006, 45, 2419 –2423
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2423