Total synthesis of the marine alkaloid halichlorine: Development and use of a general route to chiral piperidines

Article abstract of DOI:10.1021/jo901481n

(Chemical Equation Presented) The total synthesis of the marine alkaloid halichlorine is described, based on an approach that involves constructing the fully substituted asymmetric center at an early stage. The five-membered ring is formed by 5-exo-trig r

Full text of DOI:10.1021/jo901481n

pubs.acs.org/joc
Total Synthesis of the Marine Alkaloid Halichlorine: Development and
Use of a General Route to Chiral Piperidines
Dazhan Liu, Hukum P. Acharya, Maolin Yu, Jian Wang, Vince S. C. Yeh, Shunzhen Kang,
Chandramouli Chiruta, Santosh M. Jachak, and Derrick L. J. Clive*
Chemistry Department, University of Alberta, Edmonton, Alberta T6G 2G2, Canada
derrick.clive@ualberta.ca
Received July 10, 2009
The total synthesis of the marine alkaloid halichlorine is described, based on an approach that
involves constructing the fully substituted asymmetric center at an early stage. The five-membered
ring is formed by 5-exo-trig radical cyclization and the unsaturated six-membered ring by a process
that formally represents a sequential combination of conjugate addition and SN20 displacement;a
method that is general for making bicyclic compounds with nitrogen at a ring fusion position. A
formal synthesis of (+)-halichlorine is also reported, based on the development of a general method
for preparing optically pure piperidines. The key step of this method, which was used to make one of
our intermediates, is the Claisen rearrangement of a 4-vinyloxy-3,4-dihydro-2H-pyridine-1-car-
boxylic acid benzyl ester. Such O-vinyl compounds are easily generated in situ from the correspond-
ing alcohols, which are themselves readily assembled from serine and terminal acetylenes.
Introduction
cess of cancer cells.4 The pinnaic acids inhibit cytosolic
phospholipase A2 (cPLA2),2 an enzyme involved at an early
stage in the cascade of reactions that leads to the formation
of certain inflammatory mediators.5 The potential for bio-
chemical and therapeutic applications, together with the
synthetic challenges posed by the complicated structures,
has attracted appreciable attention from organic chemists,
and a large body of work has been reported.6,7 Pioneering
studies in Danishefsky’s laboratory resulted in the total
synthesis of both 18 and 23 in optically pure form, as well
Halichlorine (1)1 and the pinnaic acids, 22,3 and 3,2,3 are
structurally related natural products isolated from different
marine organisms. The compounds possess significant bio-
logical properties and may have potential for use as bio-
chemical tools or as leads for drug design. Halichlorine
inhibits the induction of vascular cell adhesion molecule 1
(VCAM-1),1 which is a property relevant to the study of
inflammatory diseases and, importantly, the metastatic pro-
(1) (a) Kuramoto, M.; Tong, C.; Yamada, K.; Chiba, T.; Hayashi, Y.;
Uemura, D. Tetrahedron Lett. 1996, 37, 3867–3870. (b) Arimoto, H.;
Hayakawa, I.; Kuramoto, M.; Uemura, D. Tetrahedron Lett. 1998, 39,
861–862.
(2) Chou, T.; Kuramoto, M.; Otani, Y.; Shikano, M.; Yazawa, K.;
Uemura, D. Tetrahedron Lett. 1996, 37, 3871–3874.
(5) Formation of certain inflammatory mediators: (a) Leslie, C. C. J.
Biol. Chem. 1997, 272, 16709–16712. (b) Dennis, E. A. J. Biol. Chem. 1994,
269, 13057–13060. (c) Smith, W.; DeWitt, D. Adv. Immunol. 1996, 62, 167–
215. (d) Holmeide, A. K.; Skatteboøl, L. J. Chem. Soc., Perkin Trans. 1 2000,
2271–2276.
(3) (a) Carson, M. W.; Kim, G.; Hentemann, M. F.; Trauner, D.;
Danishefsky, S. J. Angew. Chem., Int. Ed. 2001, 40, 4450–4452. (b) Carson,
M. W.; Kim, G.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2001, 40, 4453–
4456.
(6) Review: Clive, D. L. J.; Yu, M.; Wang, J.; Yeh, V. S. C.; Kang, S.
Chem. Rev. 2005, 105, 4483-4514.
(7) For references to model studies subsequent to, or not included in, ref 6,
see: (a) Roulland, E.; Chiaroni, A.; Husson, H.-P. Tetrahedron Lett. 2005,
46, 4065–4068. (b) Keck, G. E.; Heumann, S. A. Org. Lett. 2008, 10, 4783–
4786. (c) Hurley, P. B.; Dake, G. R. J. Org. Chem. 2008, 73, 4131–4138. (d)
Sinclair, A.; Arini, L. G.; Rejzek, M.; Szeto, P.; Stockman, R. A. Synlett
2006, 2321–2324. (e) Yang, S.-H.; Caprio, V. Synlett 2007, 1219–1222.
(8) Trauner, D.; Schwarz, J. B.; Danishefsky, S. J. Angew. Chem., Int. Ed.
1999, 38, 3542–3545.
(4) Biological properties of VCAM-1: (a) Foster, C. A. J. Allergy Clin.
ꢀ
ꢀ
Immunol. 1996, 98, S270–S277. (b) Postigo, A. A.; Teixido, J.; Sanchez-
Madrid, F. Res. Immunol. 1993, 144, 723–735. (c) Vidal-Vanaclocha, F.;
Fantuzzi, G.; Mendoza, L.; Fuentes, A. M.; Anasagasti, M. J.; Martin, J.;
Carrascal, T.; Walsh, P.; Reznikov, L. L.; Kim, S.-H.; Novick, D.; Rubin-
stein, M.; Dinarello, C. A. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 734–739.
DOI: 10.1021/jo901481n
r
Published on Web 09/09/2009
J. Org. Chem. 2009, 74, 7417–7428 7417
2009 American Chemical Society

Liu et al.
JOCArticle
as clarification of certain configurational assignments and
insights into the stereochemistry at nitrogen for the natural
products and some of their synthetic precursors.9 A synthesis,
in racemic form, of halichlorine and the two pinnaic acids;
each elegantly derived from a β-lactam;has been reported by
Heathcock and Christie,10 and a third total syntheses of pinnaic
acid 2 was subsequently accomplished by Xu, Arimoto, and
Uemura.11 The remainder of the many publications in this area
have dealt very largely with construction of the spirocyclic core
system6,7 and with formal syntheses12 reliant on the Danishefs-
ky routes. We report here an independent total synthesis of
halichlorine and, based on this route, a formal synthesis of the
optically pure compound. Our approach involved the devel-
opment of new general methods to make six-membered nitro-
gen heterocycles in optically pure form, including substances
with fully substituted asymmetric carbons.
SCHEME 1. General Synthetic Plan
displacement),15,16 which has now been explored with a wide
range of examples to generate cyclic amines15 and also carbo-
cycles.16 The conversion of 1.1 into 1.2 is one of a series of
asymmetric alkylations suggested17 to proceed with high en-
antiomeric excess (ee). However, in our hands, this particular
reaction17 gives only a modest ee (ca. 67%) when run on a
useful scale (>5 g of 1.1);a fact which we discovered when we
were much further advanced in our studies.19 At that time, we
decided to accept this level of enantioselectivity, but we later
developed a new and general approach to optically pure
piperidines having a fully substituted center adjacent to nitro-
gen (see Schemes 8 and 9). Such piperidines are expected to
have a variety of uses besides the original application described
here because of the widespread occurrence of the piperidine
substructure in the medicinal chemical literature20 and as a
feature of many natural products.21
Results and Discussion
Construction of the Fully Substituted Asymmetric Center;
First Route. The cis-diester 1.1 was prepared by hydrogena-
tion (Pd/C) of the hydrochloride salt of dimethyl 2,6-pyr-
idinedicarboxylate,22 followed by N-benzylation23 (BnBr,
K2CO3), according to published procedures (Scheme 2).
The critical allylation 1.1 f 1.2 was then carried out using
the chiral base 417,24 to give the desymmetrized product 1.2.
Synthetic Plan
After extensive exploratory studies,13 we decided to adopt a
route (Scheme 1)14 in which the asymmetric spiro center is
introduced very early in the synthesis, and to this end, webased
our approach on the allylation of the symmetrical diester 1.1 to
the unsymmetrical product 1.2, in which the eventual spiro
center is already set. Procedures were then developed for
elaborating 1.2, via intermediates 1.3 and 1.4, into halichlor-
ine. Formation of ring A (see 1.3 f 1.4) was achieved by a new
general method for cyclization (intramolecular conjugate
This is a key intermediate because it contains the eventual
spiro center of halichlorine. We had assumed that the optical
(9) Trauner, D.; Churchill, D. G.; Danishefsky, S. J. Helv. Chim. Acta
2000, 83, 2344–2351.
(10) Heathcock, C. H.; Christie, H. S. Proc. Natl. Acad. Sci. U.S.A. 2004,
101, 12079–12084.
(11) Xu, S.; Arimoto, H.; Uemura, D. Angew. Chem., Int. Ed. 2007, 46,
5746–5749.
(17) (a) Goldspink, N. J.; Simpkins, N. S.; Beckmann, M. Synlett 1999,
1292-1294, especially footnote 9. (b) Huxford, T.; Simpkins, N. S. Synlett 2004,
2295-2298, especially footnote 9. (c) In our hands, when we made 1.2 using 100 mg
of 1.1, the product had [R]25D þ40.1 (c 2.7, CHCl3), but on a larger scale (5 g of 1.1),
the product had [R]25D þ26.5 (c 2.7, CHCl3). Various modes of addition were tried,
including the use of special apparatus (see ref 18) for addition of cold solutions.
(18) Suzuki, M.; Yanagisawa, A.; Noyori, R. J. Am. Chem. Soc. 1988,
110, 4718–4726.
(12) Formal syntheses: (a) Hayakawa, I.; Arimoto, H.; Uemura, D.
Heterocycles 2003, 59, 441–444. (b) Matsumura, Y.; Aoyagi, S.; Kibayashi,
C. Org. Lett. 2004, 6, 965–968. (c) Zhang, H.-L.; Zhao, G.; Ding, Y.; Wu, B.
J. Org. Chem. 2005, 70, 4954–4961. (d) Andrade, R. B.; Martin, S. F. Org.
Lett. 2005, 7, 5733–5735. (e) Wu, H.; Zhang, H.; Zhao, G. Tetrahedron 2007,
63, 6454–6461.
(13) (a) Clive, D. L. J.; Yeh, V. S. C. Tetrahedron Lett. 1999, 40, 8503–
8507. (b) For other studies from this laboratory, but not described here, see: Clive,
D. L. J.; Wang, J.; Yu, M. Tetrahedron Lett. 2005, 46, 2853-2855.
(14) Preliminary communications: (a) Yu, M.; Clive, D. L. J.; Yeh, V. S.
C.; Kang, S.; Wang, J. Tetrahedron Lett. 2004, 45, 2879–2881. (b) Clive, D. L.
J.; Yu, M.; Li, Z. J. Chem. Soc., Chem. Commun. 2005, 906–908.
(15) Clive, D. L. J.; Li, Z.; Yu, M. J. Org. Chem. 2007, 72, 5608–5617.
(16) Wang, L.; Prabhudas, B.; Clive, D. L. J. J. Am. Chem. Soc. 2009, 131,
6003–6012.
(19) We were alerted to the poor enantioselectivity of the allylation when
X-ray analysis of a more advanced intermediate showed the material to be a
racemate. See Supporting Information for structure and CCDC registry number.
(20) For comments on the pharmaceutical significance of piperidines, see:
Watson, P. S.; Jiang, B.; Scott, B. Org. Lett. 2000, 2, 3679-3681.
(21) The piperidine substructure occurs in many natural products. See,
for example: Daly, J. W.; Spande, T. F.; Garraffo, H. M. J. Nat. Prod. 2005,
68, 1556-1575.
^
(22) Chenevert, R.; Dickman, M. J. Org. Chem. 1996, 61, 3332–3341.
(23) Chong, H.-S.; Garmestani, K.; Bryant, L. H., Jr.; Brechbiel, M. W. J.
Org. Chem. 2001, 66, 7745–7750.
(24) Bambridge, K.; Begley, M. J.; Simpkins, N. S. Tetrahedron Lett.
1994, 35, 3391–3394.
7418 J. Org. Chem. Vol. 74, No. 19, 2009

Liu et al.
JOCArticle
SCHEME 2. Synthesis of the Key Aldehyde 2.7
SCHEME 3. Initial Approach to the Construction of Ring C
purity would be high,17 but as indicated above, we later
found that not to be the case and we were unable to identify
the factors that gave us inferior results to those reported.17
Nonetheless, we continued our investigation using this ma-
terial, with the intention of later developing our own method
for setting the asymmetric spiro center.
Both ester groups of 1.2 were reduced, and the resulting
diol was monoprotected as its pivaloate (1.2 f 2.1 f 2.2).
Our plan had been to protect the less hindered hydroxyl, but
the reaction took a surprisingly different course, and we
obtained the product of acylation at the more hindered site.
In contrast, silylation with t-BuMe2SiCl (imidazole, DMAP,
CH2Cl2, 87%) occurred at the less hindered hydroxyl. The
peculiar outcome with pivaloyl chloride made no change to
our approach because the essential requirement was that the
two hydroxyls of 2.1 be differently protected. The remaining
hydroxyl of 2.2 was masked as its MOM ether, and the
pendant allyl group was subjected to hydroboration (2.2 f
2.3 f 2.4). The resulting terminal alcohol was then protected
by silylation (2.4 f 2.5). Next, the pivaloyl group was
removed by reaction with DIBALH (2.5 f 2.6), and Swern
oxidation then gave aldehyde 2.7. This monocyclic com-
pound, which represents ring B of halichlorine, contains all
of the stereochemical and structural features required for
elaboration of rings A and C and is also properly constituted
to allow stereocontrolled introduction of the C(17) methyl
group by the sequence described later.
a3.1a = less polar isomer; 3.1b = more polar isomer. bThe benzyl
group can also be removed by using Pd-C/H2; under these conditions,
when 3.1a,b are hydrogenolyzed as a mixture and the product is heated
in PhMe, all of the material is converted into a mixture of 3.3a and 3.3b
(83% overall). cWhen 3.2a and 3.2b are processed as a mixture, the
yield of the 3.3a,b mixture was 89%. dWhen 3.3a,b are processed as a
mixture, the yield of 3.4 was 93%.
with methyl propanoate (Scheme 3) so as to provide a
mixture of the diastereoisomeric alcohols 3.1a and 3.1b.
These were separable, but the stereochemistry at the two
newly created asymmetric centers was not established and is
actually of no consequence. At this point, in order to provide
some conformational rigidity to the system, and as a prelude
to generating further asymmetric centers, the benzyl group
was removed by hydrogenolysis and the rigid bicyclic lac-
tams 3.3a,b were then formed by heating the hydrogenolysis
product in PhMe; again, two isomers were obtained from the
corresponding isomeric amines 3.2a,b. The alcohols 3.3a,b
were individually dehydrated via their mesylates to enone
3.4, and finally, the pendant triisopropylsiloxy group was
replaced by bromide (3.4 f 3.5 f 3.6) and, later, also by a
phenylseleno group (3.4 f 3.5 f 3.7).
Formation of Ring C. Having reached aldehyde 2.7, the
next task was to generate ring C, and to prepare for that, the
compound was converted sequentially into 3.7 (Scheme 3)
and 4.3a,b (Scheme 4), which each contains an additional
ring that is subsequently dismantled after it has served its
purpose. Aldehyde 2.7 was subjected to aldol condensation
When the bromide was subjected to radical cyclization, the
required five-membered ring was produced (3.6 f 3.8) but
the stereochemicaloutcome at C(17) (halichlorine numbering)
J. Org. Chem. Vol. 74, No. 19, 2009 7419

Liu et al.
JOCArticle
SCHEME 4. Formation of Ring C and Introduction of the
Eventual C(17) Methyl Group
number of seemingly attractive synthetic routes that we
examined, and which involved attachment of functionalized
units to the nitrogen, were thwarted by steric factors.
The phenyl selenide 3.7 was subjected to ozonolysis and
in situ reduction at a low temperature (Scheme 4). By this
procedure, the initially formed selenoxide was reduced back
to the selenide under conditions in which the selenoxide did
not fragment to an olefin.26 Accordingly, the product of this
sequence was the tricarbonyl selenide 4.1, and on treatment
with DBU, it underwent consecutive intramolecular aldol
condensation and dehydration to afford 4.2. When the same
sequence was tried with the bromide 3.6, the aldol condensa-
tion did not work because of the base sensitivity of the
primary bromide.
Our plan had been to carry out a radical cyclization using
4.2, but the enone system of this compound was too easily
reduced by Bu3SnH and so its reactivity was lowered by
reduction under Luche conditions27 to the corresponding R-
hydroxy lactams (90%); these were then protected by acet-
ylation (98%), bringing the route to 4.3a,b. The radical
cyclization of 4.3a and 4.3b (which were processed
individually) took an interesting course. The major products
were the expected acetates 4.4a [AcO and C(17a)H syn] and
4.4b [AcO and C(17a)H anti] from 4.3a and 4.3b, respec-
tively, but the corresponding rearranged acetates 4.5a and
4.5b and the enone 4.6 were also isolated. Formation of all
these products is readily understandable: the acetates 4.4a,b
are the result of simple radical closure, while 4.5a,b presum-
ably arise by intervention of an acyloxy migration28 of the
radical produced after cyclization. The enamide 4.6 is simply
the result of spontaneous elimination, probably from the
rearranged acetates 4.5a,b. The structure of acetate 4.4a was
confirmed by X-ray analysis.
The formation of a mixture in the radical cyclization is of
little consequence because all of the material in the mixture
can be converted into 4.6. Thus 4.4a and 4.4b were hydro-
lyzed (usually as a mixture) to the corresponding alcohols
(MeONa, MeOH) in at least 95% yield, and these could be
converted, either by mesylation and base treatment, into 4.6
or, more efficiently, by treatment with o-(O2N)C6H4SeCN
and Bu3P,29 followed by oxidation with 30% H2O2. Simi-
larly, 4.5a,b were hydrolyzed and converted into 4.6. The
structure of alcohol 4.8a [i.e., hydroxyl and C(17a)H syn) was
also confirmed by X-ray analysis.
aLess polar isomer (4.3a) gave 4.4a in 66% yield, 4.5a (18%) and 4.6
(13%); more polar isomer (4.3b) gave 4.4b (67%), 4.5b (<14%), and 4.6
(15%). bStarting with 4.4a containing some 4.5a, the yield of the
mixture of 4.7a and 4.8a is 97%. cStarting with 4.4b that contains
some 4.5b, the yield of the mixture of 4.7b and 4.8b is 95%. dFrom 4.5a.
Compound 4.5b was hydrolyzed only as a minor component of a mixture
with 4.4b. eOver two steps for mixture of 4.7a and 4.8a. fOver two
steps for mixture of 4.7b and 4.8b.
At this point, we needed to introduce a methyl group at the
eventual C(17) position. Conjugate addition of cuprates to
unsaturated lactams is not well-known,30 but in the present
case, reaction with Me2CuLi worked in good yield (96%),
and the convex shape of the starting lactam 4.6 ensured that
the methyl group was delivered in the correct stereochemical
sense to provide saturated lactam 1.3.
was unfavorable, with the major isomer (4:1) having the
C(17) methyl group anti to the adjacent ring fusion hydro-
gen. Although the isomer ratio could be nearly reversed (1:3)
by base-catalyzed equilibration (t-BuOK, t-BuOH, reflux),
we were unable to separate the isomers on a useful scale, and
so we had to modify the radical cyclization and associated
steps. These changes required a base-stable homolyzable
substituent, and we turned to the phenylseleno group, which
proved to be an ideal choice.
Although the conversion of 3.6 into 3.8 (Scheme 3) and
the corresponding processes with 4.3a,b (Scheme 4) were
A noteworthy feature of the sequence in Scheme 3 is that
the nitrogen of 3.2a,b is subsequently protected by an
intramolecular process; that nitrogen is hindered and, at
least with related compounds, often resisted protection by
intermolecular reactions.25 The inaccessibility of the nitro-
gen in related intermediates posed serious restrictions, and a
(26) (a) Reich, H. J.; Shah, S. K.; Chow, F. J. Am. Chem. Soc. 1979, 101,
6648–6656. (b) Clive, D. L. J.; Postema, M. H. D. J. Chem. Soc., Chem.
Commun. 1994, 235–236.
(27) Gemal, A. L.; Luche, J. L. J. Am. Chem. Soc. 1981, 103, 5454–5459.
(28) Beckwith, A. L. J.; Crich, D.; Duggan, P. J.; Yao, Q. Chem. Rev.
1997, 97, 3273–3312.
(29) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem. 1976, 41,
1485–1486.
(30) (a) Amat, M.; Prez, M.; Llor, N.; Bosch, J.; Lago, E.; Molins, E. Org.
Lett. 2001, 3, 611–614. (b) Herdeis, C.; Kaschinski, C.; Karla, R.; Lotter, H.
Tetrahedron: Asymmetry 1996, 7, 867–884.
(25) See Supporting Information for structures of the compounds in-
volved.
7420 J. Org. Chem. Vol. 74, No. 19, 2009

Liu et al.
JOCArticle
efficient, the use of radical cyclization in our studies on the
synthesis of halichlorine uncovered some unexpected reac-
tions. In early explorations, we found that the unsaturated
sulfone 5 underwent 6-endo closure instead of following the
desired 5-exo pathway (eq 1).13a The related substrate 7
behaved as expected (eq 2),13a but the bromide 9 led only
to the simple reduction product 10 (eq 3). The contrasting
behavior of 5 and 7 might be attributable to the rigidity of the
bicyclic system 5, as this rigidity could prevent the derived
SCHEME 5. Construction of Ring A
€
radical from easily following the required Burgi-Dunitz
trajectory, while the more flexible system 7 is not restricted
in this respect. The factors responsible for the different
outcomes with 7 and 9 are not obvious; possibly, the addi-
tional bulk of the C(17a) substituent in 9 introduces unfa-
vorable nonbonded interactions in the conformation needed
for ring closure.
and gave aldehyde 5.2 in 84% yield, evidently with no
epimerization since the aldehyde could be reduced back to
the starting alcohol 5.1. Likewise, due to the vulnerability of
the C(5) configuration in 5.2 to the basic conditions of the
Wittig reaction, we also checked that formation of the enol
ethers 5.3 did not incur any epimerization at C(5): ozonolysis
and reduction gave back alcohol 5.1.
In order to build up ring A, aldehyde 5.4 was subjected to a
Baylis-Hillman reaction with acrylonitrile. That process
was slow but provided the desired alcohols 5.5a,b33 in high
yield, and these were immediately converted into the corre-
sponding acetates 5.6a,b, the overall yield from the homo-
logated aldehyde 5.4 being 89%. We used AcCl and not
Ac2O for the acetylation because, in model studies, acetate
was found to add to the terminal double bond of structures
related to 5.5a,b, but this side reaction was not observed with
AcCl. Finally, the method we had found earlier for opening
the lactam ring31 of 1.3 was now applied to the acetates 5.6a,b
so as to release the lactam carbonyl as a methyl ester and the
nitrogen as a free amine. The resulting amines (5.7a,b)
underwent spontaneous intramolecular conjugate displace-
ment15,34 to afford the unsaturated nitrile 1.4, with the A, B,
and C rings of halichlorine in place. In our preliminary
studies,14b we had used methyl acrylate as the Michael
acceptor in the Baylis-Hillman condensation to eventually
give a diester corresponding to 1.4 (CO2Me in place of CN),
but we later found that an essential chemoselective reduction
of the two ester groups;one to an aldehyde and the other to
an alcohol;could not be effected reproducibly, and so we
Formation of Ring A. We initially felt that the appropriate
next step from 1.3 would be hydrolysis of the lactam. This
transformation required a very extensive effort, but we
eventually found that the lactam ring could be opened by
treatment with the Meerwein salt Me3O+BF4-, followed by
aqueous Na2CO3.31 However, by the time we had achieved
this lactam opening, our plans for constructing ring A had
been changed, and the new approach, which also requires
opening of the lactam (but at a later stage), was now
implemented as follows (Scheme 5).
The MOM group of 1.3 was removed by the action of
Me3SiBr, and the resulting alcohol was homologated by
oxidation, Wittig olefination with Ph3PdCH(OMe), and
then acid hydrolysis (1.3 f 5.1 f 5.2 f 5.3 f 5.4). The
oxidation step was unreliable when done by the Swern
method, as this gave variable and usually low yields. How-
32
ever, the use of n-Pr4NRuO4 was consistently successful
(31) (a) Overman, L. E.; Robichaud, A. J. J. Am. Chem. Soc. 1989, 111,
300–308. (b) Deslongchamps, P.; Lebreux, C.; Taillefer, R. Can. J. Chem.
1973, 51, 1665–1669. (c) A very similar method has been used to open a five-
membered lactam in the halichlorine series: See ref 12b.
(33) A small amount of the starting aldehyde was recovered, indicating
that the aldehyde did not epimerize (by retro-Michael addition and Michael
addition).
(32) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639–666.
(34) For a mechanistically related method, see ref 10.
J. Org. Chem. Vol. 74, No. 19, 2009 7421

Liu et al.
SCHEME 6. Side Chain Extension and Preparation for
Macrocyclization
JOCArticle
had to retrace our steps and try acrylonitrile, which proved to
be a satisfactory choice. As mentioned earlier, this method
(cf. 5.4 f 1.4) of constructing bicyclic compounds with
nitrogen at a ring junction is general.15,34 An important
feature of the nitrile ester 1.4 is the fact that the stereogenic
center in the side chain is β to a carbonyl, as opposed to R,
and consequently, the danger of stereochemical scrambling
at C(17) is avoided.
Side Chain Extension and Macrocyclization. The route we
followed in order to reach halichlorine from 1.4 required
extension of the side chain on the five-membered ring,
followed by macrocyclization. To this end, the best of several
approaches that we examined began with reduction of the
nitrile ester 1.4 with DIBALH in two stages. These specifi-
cally involved treatment with DIBALH (2 mol per mol nitrile
ester) in CH2Cl2-THF at -78 °C, followed by workup and
re-exposure at -78 °C to DIBALH (4 mol per mol of original
nitrile ester) in CH2Cl2-PhMe. Under these conditions, the
nitrile was converted to an aldehyde (after aqueous workup)
and the ester to an alcohol (1.4 f 6.1). The aldehyde group
was protected as a cyclic ketal, and oxidation (n-Pr4NRuO4)
then generated aldehyde 6.3.
At this stage, we wished to form a carbanion at the
aldehyde carbon [C(15) of halichlorine] and decided to use
a selenium-stabilized carbanion, the aim being to later
employ the selenium unit to make the C(15)-C(16) double
bond. The usual method of generating selenium-stabilized
carbanions35 is from seleno ketals, themselves obtained by
acid-catalyzed reaction of aldehydes and PhSeH.36 How-
ever, our experience in handling compounds of type 6.3 gave
us the impression that they were sensitive to Lewis and protic
acids, and so we felt obliged to produce the selenium-
stabilized carbanion by a method that avoids the use of
acids. Accordingly, aldehyde 6.3 was treated with Bu3SnLi,
and the resulting mixture of stannyl alcohols was immedi-
ately converted into the corresponding selenides 6.4, which
were isolated in modest yield (61%). When this material was
treated with BuLi (to generate the required selenium-stabi-
lized carbanion37 by preferential C-Sn heterolysis), and
then with the known chloro aldehyde 6.5,8,38 a mixture of
β-hydroxyselenides 6.6 was obtained. On oxidation, the
selenium was removed to liberate the double bond (6.6 f
6.7). 1H NMR measurements showed that the double bond
had the required E geometry (J = 15.6 Hz). Finally, the
hydroxyl group of 6.7 was protected by silylation (Scheme 6,
6.7 f 6.8).
aContains slight impurities.
was selectively deprotected with NH4F,41 which had also
been used by Danishefsky et al.,8 and the resulting hydroxy
acids were cyclized (58%) by the Keck macrocyclization
protocol.42 Finally, removal of the remaining hydroxyl
protecting group [54 or 86% corrected for the proportion
of the correct C(17) isomer present in the macrolactone] gave
halichlorine 7.5 (ꢀ1), whose NMR spectra corresponded to
those reported for the natural substance1 and for racemic
synthetic10 material. The undesired C(14) isomer of 7.5 was
obtained in very small amount, and we did not to attempt to
recycle it.
Formation of an Optically Pure Intermediate;A New
Route to Piperidines. Having established a synthetic route
to halichlorine, we returned to the key problem of making
one of our intermediates as a single enantiomer so as to
develop a formal synthesis of the optically pure alkaloid
based on the approach summarized in the above schemes. To
this end, we needed to devise a method for preparing a
suitable 2,2,6-trisubstituted piperidine. The structure of the
early intermediate 2.1 suggested the use of serine as a starting
material, and the requirement that this amino acid be
converted into a piperidine with a fully substituted asym-
metric carbon led us to consider compounds of type 11
To prepare for the macrocyclization, the ketal unit was
hydrolyzed (Scheme 7) by using Me3SiOSO2CF3 in the
presence of 2,6-lutidine.39 This experiment released alde-
hydes 7.1, which were converted by Pinnick oxidation40 to
the corresponding acids 7.2. Next, the primary siloxy group
(35) Review on selenium-stabilized carbanions: Clive, D. L. J. Organic
Compounds of Sulphur, Selenium, and Tellurium, Specialist Periodical Re-
ports; Royal Society of Chemistry: London, 1981; Vol. 6, pp 112-126.
(36) Review on the chemistry of selenoacetals: Krief, A.; Hevesi, L.
Janssen Chimica Acta 1984, 2, 3-14.
(37) This method of generating selenium-stabilized carbanions is general:
Fernandopulle, S.; Clive, D. L. J.; Yu, M. J. Org. Chem. 2008, 73,
6018-6021.
(38) Keen, S. P.; Weinreb, S. M. J. Org. Chem. 1998, 63, 6739–6741.
(39) Fujioka, H.; Okitsu, T.; Sawama, Y.; Murata, N.; Li, R.; Kita, Y. J.
Am. Chem. Soc. 2006, 128, 5930–5938.
(41) Selective deprotection of primary t-BuMe2SiOR in the presence of
secondary t-BuMe2SiOR: (a) Nelson, T. D.; Crouch, R. D. Synthesis 1996,
1031–1069. (b) White, J. D.; Porter, W. J.; Tiller, T. Synlett 1993, 535–538.
(42) Boden, E. P.; Keck, G. E. J. Org. Chem. 1985, 50, 2394–2395.
(40) Bal, B. S.; Childers, W. E., Jr.; Pinnick, H. W. Tetrahedron 1981, 37,
2091–2096.
7422 J. Org. Chem. Vol. 74, No. 19, 2009

Liu et al.
JOCArticle
SCHEME 7. Modification of Functional Groups, Macrolacto-
nization, and Formation of Halichlorine
rearrangements in the oxygen series (glycals) are, of course,
part of the classical repertoire of Ireland-ester enolate re-
arrangements,46 but the Claisen rearrangement defined by
the arrows in 14 has not been explored before and, in the
form described below, represents a general synthetic method.
We initially set out to examine the rearrangement using
compounds with the stereochemistry shown in 14 but en-
countered several difficulties: carbonyl reduction of the
parent dihydropyridinone 13 is most efficiently done in a
manner (NaBH4/CeCl3•7H2O)44 that gives the undesired
stereochemistry (hydroxyl anti to PgOCH2), so that a Mit-
sunobu inversion is required. Second, the ease of the re-
arrangement depends on the stereochemistry at C(4): the
reaction is very slow for the stereochemistry shown in 14
(Pg = Cbz or t-BuMe2Si, Pg0 = Cbz) but occurs at a
satisfactory rate for the opposite C(4) stereochemistry. Once
these facts had emerged, we were able to plan the synthesis of
an intermediate that overlaps with our route to halichlorine
because we were then in position to make a proper choice of
the C(2) substituent for the starting dihydropyridinone of
type 13.
(Pg = protecting group), our intention being to use the
asymmetric center at C(6) to control the stereochemical
outcome when we converted C(2) to an sp3 center. Prelimin-
ary experiments were aimed at delivering a functionalized
carbon intramolecularly along the lines summarized in 12,
where C* is a stabilized carbanion. Our attempts to carry out
such a reaction, however, were unsuccessful, as were at-
tempts at conjugate addition to 13 (Pg = t-BuMe2Si, Pg0
= Bn), using vinylmagnesium bromide and CuBr•SMe2 in
the presence of Me3SiCl, with or without HMPA. A con-
sequence of this unrewarding effort was that we decided to
try the Claisen rearrangement of enol ethers of type 14, which
we expected to be available43,44 from dihydropyridinones 13.
In view of the very large amount of research published on the
construction of piperidines,45 and the importance of this
compound class,20,21 it is surprising that the procedure
summarized in 14 has not been reported before. Related
A convenient method for preparing dihydropyridinones
was already available,47 and the transformation shown in
eq 4 served as a model for our needs, although it was not
certain that the route we intended to follow would preserve
the stereochemical integrity of the starting material, as
several related cyclizations47 incurred some epimerization.
(43) NaBH4/CeCl3 reduction of piperidinones: Comins, D. L.; Foley, M.
A. Tetrahedron Lett. 1988, 29, 6711-6714.
(44) Comins, D. L.; Kuethe, J. T.; Miller, T. M.; Fevrier, F. C.; Brooks, C.
A. J. Org. Chem. 2005, 70, 5221–5234.
ꢀ
L-Serine methyl ester hydrochloride was converted in four
simple steps into the iodide 8.1, using a reported procedure,48
and treatment with lithium divinyl cuprate then generated
the known oxazolidinone49 8.2 (Scheme 8). Ozonolysis of the
double bond and reaction with the lithium acetylide derived
(45) Reviews: (a) Buffat, M. G. P. Tetrahedron 2004, 60, 1701–1729. (b)
Felpin, F.-X.; Lebreton, J. Eur. J. Org. Chem. 2003, 3693–3712. (c) Wein-
traub, P. M.; Sabol, J. S.; Kane, J. M.; Borcherding, D. R. Tetrahedron 2003,
59, 2953–2989. (d) Bates, R. W.; Sa-Ei, K. Tetrahedron 2002, 58, 5957–5978.
(e) Enders, D.; Thiebes, C. Pure Appl. Chem. 2001, 73, 573–578. (f) Husson,
H.-P.; Royer, J. Chem. Soc. Rev. 1999, 28, 383–394. (g) Laschat, S.; Dickner,
T. Synthesis 2000, 1781–1813. (h) Deiters, A.; Martin, S. F. Chem. Rev. 2004,
104, 2199–2238. (i) Rubiralta, M.; Giralt, E.; Diez, A. Piperidines, Structure,
Preparation, Reactivity, and Synthetic Applications of Piperidine and Its
Derivatives; Elsevier: Amsterdam, 1991. Routes to non-racemic piperidines: (j)
Coombs, T. C.; Zhang, Y.; Garnier-Amblard, E. C.; Liebeskind, L. S. J. Am.
Chem. Soc. 2009, 131, 876–877. (k) Abrunhosa-Thomas, I.; Roy, O.; Barra, M.;
Besset, T.; Chalard, P.; Troin, Y. Synlett 2007, 1613–1615. (l) Adriaenssens, L. V.;
Hartley, R. C. J. Org. Chem. 2007, 72, 10287–10290. (m) Gotchev, D. B.;
Comins, D. L. J. Org. Chem. 2006, 71, 9393–9402. (n) Back, T. G.; Nakajima, K.
(46) (a) Castro, A. M. M. Chem. Rev. 2004, 104, 2939–3002. (b) Pereira,
S.; Srebnik, M. Aldrichimica Acta 1993, 26, 17–29. (c) Enders, D.; Knopp,
M.; Schiffers, R. Tetrahedron: Asymmetry 1996, 7, 1847–1882. (d) Overman,
L. E. Angew. Chem., Int. Ed. Engl. 1984, 23, 579–586. (e) Lutz, R. P. Chem.
Rev. 1984, 84, 205–247. (f) Rhoads, S. J.; Raulins, N. R. Org. React. 1975, 22,
1–252.
(47) Turunen, B. J.; Georg, G. I. J. Am. Chem. Soc. 2006, 128, 8702–8703.
(48) Sibi, M. P.; Rutherford, D.; Renhowe, P. A.; Li, B. J. Am. Chem. Soc.
1999, 121, 7509–7516.
€
J. Org. Chem. 1998, 63, 6566–6571. (o) Waldmann, H.; Braun, M.; Drager, M.
Angew. Chem., Int. Ed. Engl. 1990, 29, 1468–1471. (p) Kunz, H.; Pfrengle, W.
Angew. Chem., Int. Ed. Engl. 1989, 28, 1067–1068.
(49) We found this method more convenient than that reported in the
literature (Kim, S.-G.; Ahn, K. H. Synth. Commun. 1998, 28, 1387-1397).
J. Org. Chem. Vol. 74, No. 19, 2009 7423

Liu et al.
JOCArticle
SCHEME 8. Synthesis of the Precursor to the Rearrangement
Substrate 8.10
experiments: N-acylation of 8.8 with (S)-(+)-R-methoxy-R-
trifluoromethylphenylacetyl chloride gave a product with a
vinyl signal in the 1H NMR spectrum at 5.93 ppm (400 MHz,
CDCl3). With the enantiomeric reagent, the corresponding
signal is at 5.88 ppm; this signal is absent in the spectrum of
the derivative formed with the S-reagent and so 8.8 must be a
single enantiomer.
The crucial rearrangement was done in situ by heating
alcohol 8.10 in butyl vinyl ether in the presence of Hg(OAc)2
and Et3N (Scheme 9). Under these conditions (110 °C, 36 h),
the expected vinyl ether (cf. 14) was formed and it rearranged
in the required way to give 9.1, but attempts to isolate the
intermediate vinyl ether (before the reaction was complete)
were not successful because the material is unstable to
chromatography, during which it appears to undergo elim-
ination (1H NMR). The required conditions are mild com-
pared with those that are usual46f for the Claisen rearra-
ngement; possibly, mercuric ion catalysis is involved.46d,e We
can find no examples of Claisen rearrangement in which the
distal terminus of the allylic double bond system carries
nitrogen;as is the case in the present work;and only one
report of a correspondingly substituted acyclic Ireland-ester
enolate rearrangement.52
Although alcohol 8.10 was made specifically for the reac-
tions described, it should also have provided an opportunity
for generating the fully substituted center by way of an SN20
displacement.53,54 However, the earlier studies that led us
to develop the route shown in Schemes 8 and 9 did in-
clude experiments along such lines, and we had found that
O-acyl derivatives of alcohol 1855 decomposed on chroma-
tography (silica gel).56 When the crude O-acyl derivatives
were treated with vinyl metals, under conditions that nor-
mally result53,54,57 in an SN20 pathway, they failed to react,
decomposed, or (in one case53) appeared to undergo both
SN20 and SN2 displacement.
from 8.450 gave the expected alcohols 8.5, which were oxidized
to ketone 8.6. On treatment with Cs2CO3 in hot MeOH, the
ketone was transformed into the dihydropyridinone 8.7 in
80% yield. The use of a p-methoxybenzyl-protected oxygen for
the propargylic component was necessary as t-BuMe2Si and
t-BuPh2Si groups were labile under the cyclization conditions.
The presence of additional functionality in our case (8.6),
compared with compound 15 (eq 2), necessitated the use of
different conditions for the ring closure from those reported,47
and weavoided the use of HCl47 or Me3SiI.47 Silylation51 of the
free hydroxyl of 8.7 and protection of nitrogen as a carbamate
took the route as far as 8.9. Finally, reduction under Luche
conditions27 gave alcohol 8.10, setting the stage for the in-
tended rearrangement. The choice of a Cbz group for nitrogen
protection was based on the finding that reduction of the
ketone carbonyl in analogous compounds was unsuccessful
when the nitrogen is protected with a benzyl group and the fact
that related reductions have been reported43 where the nitro-
gen carries a PhOCO group. Reduction was stereoselective
(95:5) in favor of the indicated isomer 8.10. Presumably, the
reduction involves conformation 17,43,44 which is adopted in
order to alleviate A1,3 strain, and complexation occurs anti to
the siloxymethyl substituent so that the hydride is delivered
axially (syn to the substituent).
Because of the unfavorable properties of the O-acyl deri-
vatives, access to the inverted alcohol 19 was gained by
DIBALH-BHT reduction44 of the parent ketone (63% yield),
and with both 18 and 19 in hand, the effect of stereochemistry
on the Claisen rearrangement could be examined, as men-
tioned above. Alcohol 18 behaves as required, but the isomer
(52) Ylioja, P. M.; Mosley, A. D.; Charlot, C. E.; Carbery, D. R.
Tetrahedron Lett. 2008, 49, 1111–1114.
(53) (a) Kiyotsuka, Y.; Acharya, H. P.; Katayama, Y.; Hyodo, T.;
Kobayashi, Y. Org. Lett. 2008, 10, 1719–1722. (b) Kiyotsuka, Y.; Katayama,
Y.; Acharya, H. P.; Hyodo, T.; Kobayashi, Y. J. Org. Chem. 2009, 74, 1939–
1951.
(54) (a) Calaza, M. I.; Hupe, E.; Knochel, P. Org. Lett. 2003, 5, 1059–
1061. (b) Harrington-Frost, N.; Leuser, H.; Calaza, M. I.; Kneisel, F. F.;
Knochel, P. Org. Lett. 2003, 5, 2111–2114.
(55) The alcohol was coupled (EDCI, DMAP) with picolinic acid or
MeOCH2CO2H, and we attempted unsuccessfully to phosphorylate it with
(EtO)2P(O)Cl.
The optical purity of the starting serine was preserved
during the above sequence, as indicated by the following
(50) Clive, D. L. J.; Yang, W.; MacDonald, A. C.; Wang, Z.; Cantin, M.
J. Org. Chem. 2001, 66, 1966–1983.
(51) The route would have been a little shorter if the hydroxyl had been
protected as a MOM ether; this was not tried, however, because attempts to
so protect a related compound (i-Pr3SiOCH2CH2CH2 instead of
PMBOCH2) were unsuccessful and gave complex mixtures.
(56) For a related pivaloate that is stable to chromatography, see ref 44.
For related isolable acetates, see: (a) Dransfield, P. J.; Gore, P. M.; Prokes, I.;
Shipman, M.; Slawin, A. M. Z. Org. Biomol. Chem. 2003, 1, 2723–2733. (b)
Hanson, G. J.; Russell, M. A. Tetrahedron Lett. 1989, 30, 5751–5754.
(57) Nakata, K.; Kiyotsuka, Y.; Kitazume, T.; Kobayashi, Y. Org. Lett.
2008, 10, 1345–1348.
7424 J. Org. Chem. Vol. 74, No. 19, 2009

Liu et al.
JOCArticle
SCHEME 9. Claisen Rearrangement and Synthesis of Optically
Pure Intermediate 9.13 [Formal Synthesis of (þ)-Halichlorine]
SCHEME 10. Mechanistic Pathway
intermediate in our route to halichlorine; formation of (þ)-
9.13 constitutes a formal synthesis of (þ)-halichlorine be-
cause none of the remaining steps in our sequence can cause
racemization.
Mechanistic Considerations. The cyclization of 8.6 to 8.7
(and related cyclizations) clearly represents the outcome of a
number of consecutive steps, and we wanted to identify the
point in the overall sequence at which the oxazolidinone is
hydrolyzed. When ketone 20a, which was made58 along lines
developed for 8.6, was treated with Cs2CO3 in MeOH for 3 h
at room temperature (i.e., under milder conditions than used
for 8.6), we isolated the vinylogous ester 20b as the major
product, together with the expected dihydropyridinone 20c
(20b/20c = ca. 2:1) (Scheme 10). The ee of 20c was 99%,
based on the 1H NMR method described above for 8.8; the
olefin geometry shown for 20b is arbitrary.59 If the reaction is
stopped after 20 min, the only products isolated were again
20b (major) and 20c (ca. 4:1). When 20b was subjected to the
normal conditions for cyclization (Cs2CO3, MeOH, room
temperature and then 65 °C), it was converted into 20c.
When this reaction was stopped during the heating period,
1
but before completion, H NMR examination of the crude
mixture showed the signals for 20b and 20c, together with
signals of what were obviously minor components, but which
we were unable to isolate.
It would appear that the final dihydropyridinone is always
produced by an addition-elimination mechanism (cf. 20a f
20b f 20c), similar to that proposed by Turunen and
Georg.47
We also converted 20c into the cyclic carbamate 21 and
found that it rapidly yields 20c on treatment with Cs2CO3 in
MeOH at room temperature; in contrast, the oxazolidinone
8.2, which we used as a reference model, is largely unchanged
under the same conditions (Cs2CO3, MeOH, 4 h at room
temperature, 4 h at 65 °C), suggesting that cyclization occurs
with the oxazolidinone subunit intact, and the resulting
vinylogous imide (e.g., 21) undergoes rapid methanolysis.60
19 did not provide a rearrangement product; the main com-
pound obtained appeared to be the intermediate vinyl ether
(1H NMR), which again could not be isolated. The reluctance
of the vinyl ether derived from 19 to rearrange must be due to
unfavorable steric interactions, as both the CbzOCH2 group
and the vinyloxy substituent would have to occupy a pseu-
doaxial conformation; in the vinyl ether derived from 18, only
the CbzOCH2 group is pseudoaxial in the reacting conforma-
tion.
Conclusion
The halichlorine structure represents a complicated target
for total synthesis, a view that is supported by the fact that an
At this point, aldehyde 9.1, in which two key stereogenic
centers have been set, was elaborated into (þ)-9.13, largely
by the methods used earlier, as summarized in Scheme 9. This
compound is the optically pure version of 3.4, which is an
(58) See Supporting Information.
(59) Dabrowski, J.; Tencer, M. Bull. Chem. Soc. Jpn. 1976, 49, 981-986.
(60) Flynn, D. L.; Zelle, R. E.; Grieco, P. A. J. Org. Chem. 1983, 48, 2424–
2426.
J. Org. Chem. Vol. 74, No. 19, 2009 7425

Liu et al.
JOCArticle
unusually large number of model studies have been under-
taken. The synthesis reported here had two unanticipated
consequences: the development of the general process of
intramolecular conjugate displacement (cf. 5.6 f 1.4,
Scheme 5)15 and the route to piperidines. The former was
devised specifically for constructing ring A but is broadly
general for the preparation of nitrogen-containing hetero-
cycles15 and also, in a suitably modified form, for the
preparation16 of carbocycles. The approach to optically
active piperidines, which is based on a Claisen rearrange-
ment of readily accessible vinyl ethers of defined stereochem-
istry, is also general61 and can lead, as applied here, to the
formation of compounds having a fully substituted carbon.
The stereochemistry of that center is controlled by the
stereochemistry of the starting alcohol and the choice of
substituent on the double bond of the parent dihydropyr-
idinone. This method for making piperidines is expected to
be useful in its own right because piperidines represent a
pharmaceutically privileged compound class20 and are a
common subunit in natural products.21 In addition, inter-
mediates of type 9.1 should be amenable to many modifica-
tions, so as to give access to a wide range of optically pure
compounds.
out using the same amounts of starting material and reagents, all
of the solutions had been combined in the 500 mL round-
bottomed flask. The cooling bath was left in place but not
recharged, and stirring was continued overnight. Evaporation
of the solvent and flash chromatography of the residue over silica
gel (3 ꢁ 30 cm), using 1:3 to 1:1 EtOAc/hexane, gave 4.1 as a
colorless oil. The combined material from the above seven
batches together withthe total product froma further five batches
was converted, as described below, into the unsaturated keto
amide 4.2 in an overall yield of 73% (4.89 g) for the two steps.
(4R,9aR)-rel-1,3,4,9a-Tetrahydro-4-[(methoxymethoxy)methyl]-
9a-[3-(phenylseleno)propyl]-2H-quinolizine-6,7-dione (4.2). First
Procedure. DBU (15.51 mL, 103.7 mmol) was added dropwise
over 20 min to a stirred and cooled (-10 °C) solution of 4.1
(4.72 g, 10.4 mmol) in THF (200 mL). Stirring at this tempera-
ture was continued for 1 h, the cold bath was removed, and after
1 h, the solution was stirred and refluxed overnight. The solution
was cooled to room temperature and evaporated. Flash chro-
matography of the residue over silica gel (3 ꢁ 25 cm), using 2:1
EtOAc/hexane, gave 4.2 (3.63 g, 80%) as a colorless oil: FTIR
1
(CH2Cl2 cast) 1747, 1666 cm-1; H NMR (CDCl3, 300 MHz)
δ 1.42-1.73 (m, 5 H), 1.83-2.07 (m, 4 H), 2.44-2.59 (m, 1 H),
2.80-3.00 (m, 2 H), 3.34 (s, 3 H), 3.44-3.58 (m, 1 H), 4.06-4.20
(m, 2 H), 4.62 (AB q, J = 6.4 Hz, 4νAB = 11.4 Hz, 2 H), 6.36 (d,
J = 10.2 Hz, 1 H), 6.83 (d, J = 10.2 Hz, 1 H), 7.23-7.28 (m,
3 H), 7.43-7.48 (m, 2 H); 13C NMR (CDCl3, 125 MHz) δ 19.7
(t), 23.5 (t), 26.7 (t), 27.4 (t), 34.1 (t), 37.3 (t), 55.4 (q), 56.3 (d),
64.4 (s), 69.4 (t), 96.8 (t), 126.4 (d), 127.2 (d), 129.2 (d), 129.5 (s),
132.8 (d), 156.1 (d), 158.5 (s), 179.1 (s); exact mass (electrospray)
m/z calcd for C21H27NNaO480Se (M þ Na) 460.0998, found
460.0994.
Second Procedure. Seven combined batches of freshly ob-
tained ketoaldehyde 4.1 (see above) were immediately treated
with base as follows: DBU (7.4 mL, 50 mmol) was added
dropwise over 5 min to a stirred and cooled (0 °C) solution of
the ketoaldehyde in THF (170 mL). The ice bath was left in place
but not recharged, and stirring was continued overnight. Eva-
poration of the solution and flash chromatography of the
residue over silica gel (3 ꢁ 25 cm), using 3:1 EtOAc/hexane,
gave 4.2. A further five batches of ketoaldehyde were processed
in the same way to obtain 4.2 (4.89 g in all, 73% over two steps)
as a colorless oil.
Experimental Section
(2R,6R)-rel-1-(1,2-Dioxopropyl)-6-[(methoxymethoxy)methyl]-
2-[3-(phenylseleno)propyl]-2-piperidinecarboxaldehyde
(4.1).
First Method. A solution of 3.7 (500 mg, 1.18 mmol) in CH2Cl2
(40 mL) was cooled to -78 °C. A steady stream of O3 in O2
(dried by passing through a trap at -78 °C) was bubbled
through the solution for 35 min. The solution was then flushed
with O2 for 20 min to remove the excess of O3, and Ph3P (1.2 g,
4.7 mmol) was added in one portion. The cold bath was left in
place but not recharged, and stirring was continued overnight.
Evaporation of the solvent and flash chromatography of the
residue over silica gel (2 ꢁ 20 cm), using 1:2 EtOAc.hexane, gave
4.1 (451 mg, 84%) as a colorless oil: FTIR (CH2Cl2 cast) 1768,
1731, 1714, 1624 cm-1 1H NMR (CDCl3, 400 MHz) δ
;
1.49-2.07 (m, 9 H), 2.20-2.26 (m, 1 H), 2.32 (s, 3 H),
2.87-2.90 (m, 2 H), 3.32 (s, 3 H), 3.45 (dd, J = 4.1, 9.6 Hz,
1 H), 3.60 (t, J = 10.3 Hz, 1 H), 4.43-4.46 (m, 1 H), 4.58-4.61
(m, 2 H), 7.20-7.25 (m, 3 H), 7.45-7.47 (m, 2 H), 9.39 (s, 1 H);
13C NMR (CDCl3, 125 MHz) δ 13.0 (t), 21.7 (t), 24.6 (t), 25.7 (t),
27.1 (q), 28.2 (t), 33.4 (t), 51.0 (q), 55.6 (d), 65.2 (s), 70.5 (t), 96.2
(t), 126.9 (d), 129.0 (d), 130.0 (s), 132.7 (d), 168.4 (s), 196.8 (s),
197.0 (d); exact mass m/z calcd for C21H29NO580Se 455.1211,
found 455.1200.
Second Method (Batch Process). A solution of 3.7 (0.539 g,
1.27 mmol) in CH2Cl2 (40 mL) was cooled to -78 °C, and a
steady stream of O3 in O2 (dried by passing through a trap at
-78 °C) was bubbled through the solution for 30 min. Then the
mixture was flushed with O2 for 20 min to remove the excess of
O3, and Ph3P (1.2 g, 4.7 mmol) was added in one portion. The
mixture was stirred for 1 min and was then transferred to a
cooled (-78 °C) round-bottomed flask (500 mL containing a
magnetic stirring bar) using CH2Cl2 (ca. 3 mL) as a rinse. The
mixture was stirred at -78 °C.
(4R,7R,8aR,11aS)-rel-7-(Acetyloxy)decahydro-4-[(methoxy-
methoxy)methyl]-6H-cyclopenta[i]quinolizin-6-one (4.4a), (4R,
8S,8aS,11aS)-rel-8-(Acetyloxy)decahydro-4-[(methoxymethoxy)-
methyl]-6H-cyclopenta[i]quinolizin-6-one (4.5a), and (4R,8aR,
11aS)-rel-1,2,3,4,8a,9,10,11-Octahydro-4-[(methoxymethoxy)-
methyl]-6H-cyclopenta[i]quinolizin-6-one (4.6). Bu3SnH (0.53 mL,
2.0 mmol) and AIBN (65 mg, 0.40 mmol) in PhH (25 mL) were
added over 10 h (syringe pump) to a refluxing solution of 4.3a
(0.64 g, 1.3 mmol) in PhH (260 mL). After the addition, the
solution was refluxed for 4 h, cooled to room temperature, and
evaporated. Flash chromatography of the residue over silica gel
(2ꢁ10 cm), using 1:1 EtOAc/hexane, gave acetate 4.4a (combined
yield of two batches run on the same scale, 0.566 g, 66%), 4.5a
(combined yield of the two batches, 0.151 g, 18%), and 4.6
(combined yield of the two batches, 89 mg, 13%) as colorless oils.
Acetate 4.4a: FTIR (CH2Cl2 cast) 1694, 1667 cm-1; 1H NMR
(CDCl3, 500 MHz) δ 1.47-2.05 (m, 15 H), 2.13 (s, 3 H), 3.33 (s, 3
H), 3.71 (dd, J = 9.6, 3.8 Hz, 1 H), 3.77 (t, J = 9.3 Hz, 1 H),
3.85-3.89 (m, 1 H), 4.60 (AB q, J = 6.34 Hz, 4νAB = 19.25 Hz,
2 H), 5.24 (dd, J = 10.9, 5.2 Hz, 1 H); 13C NMR (CDCl3, 125
MHz) δ 18.2 (t), 21.0 (q), 23.0 (t), 23.5 (t), 30.1 (t), 30.5 (t), 35.6
(t), 38.3 (t), 42.6 (d), 53.7 (q), 55.2 (d), 67.2 (s), 67.3 (d), 69.0 (t),
96.6 (t), 168.5 (s), 170.3 (s); exact mass (electrospray) m/z calcd
for C17H27NNaO5 (M þ Na) 348.1781, found 348.1782.
Another solution of 3.7 (0.539 g, 1.27 mmol) in CH2Cl2
(40 mL) was subjected to ozonolysis following the above con-
ditions. After the addition of Ph3P (1.2 g, 4.7 mmol), the mixture
was transferred to the same 500 mL round-bottomed flask using
CH2Cl2 (ca. 3 mL) as a rinse. After five more experiments carried
Acetate 4.5a: FTIR (CH2Cl2 cast) 2948, 2879, 1737, 1653 cm-1
1H NMR (CDCl3, 500 MHz) δ 1.46-2.02 (m, 15 H), 2.20 (ddd,
;
(61) See Supporting Information for the preparation and subsequent
Claisen rearrangement, in the presence of butyl vinyl ether, of 18 and 20a.
7426 J. Org. Chem. Vol. 74, No. 19, 2009

Liu et al.
JOCArticle
J = 12.2, 5.7, 2.7 Hz, 1 H), 2.42 (dd, J = 16.7, 7.3 Hz, 1 H), 2.67
(dd, J = 16.7, 3.7 Hz, 1 H), 3.36 (s, 3 H), 3.77 (dd, J = 9.7, 4.1 Hz,
1 H), 3.81 (dd, J = 8.6, 8.6 Hz, 1 H), 3.91 (ddd, J = 10.6, 8.2,
4.1 Hz, 1 H), 4.63 (AB q, J = 6.4 Hz, 4νAB = 17.0 Hz, 2 H), 4.92
(ddd, J = 6.8, 6.4, 3.7 Hz, 1 H); 13C NMR (CDCl3, 125 MHz) δ
17.99 (t), 21.11 (q), 22.91 (t), 23.69 (t), 28.52 (t), 36.40 (t), 36.47 (t),
38.14 (t), 50.29 (d), 53.80 (d), 55.26 (q), 66.35 (s), 68.95 (t), 70.39
(d), 96.62 (t), 168.30 (s), 170.26 (s); exact mass m/z calcd for
C17H27NO5 325.1889, found 325.1888.
(4R,8S,8aR,11aS)-rel-Decahydro-β-hydroxy-8-methyl-r-met-
hylene-6-oxo-6H-cyclopenta[i]quinolizine-4-butanenitrile (5.5a,
b). A mixture of 5.4 (0.20 g, 0.80 mmol), DABCO (0.72 g,
6.4 mmol), and acrylonitrile (5 mL) was stirred at room tem-
perature for 5 days. Evaporation of the solvent and flash
chromatography of the residue over silica gel (2 ꢁ 10 cm), using
1.5:1 EtOAc/hexane, gave the adduct 5.5a,b as a mixture of
epimers which was used directly in the next step.
(1R,2S,90aS)-rel-70-[1,3-Dioxolan-2-yl]-10,20,30,60,90,90a-hexa-
hydro-2-[(1R-rel)-1-methyl-3-(phenylseleno)-3-(tributylstannyl)-
propyl]spiro[cyclopentane-1,4’-[4H]quinolizine] (6.4). n-BuLi
(2.5 M in hexane, 0.14 mL, 0.35 mmol) was added dropwise
to a stirred and cooled (0 °C) solution of i-Pr2NH (0.055 mL,
0.390 mmol) in THF (0.5 mL). Stirring was continued at this
temperature for 10 min, and then Bu3SnH (0.11 mL, 0.39 mmol)
was added. Stirring was continued for 15 min, and then the ice
bath was replaced by a dry ice-acetone bath. A solution of 6.3
(23.6 mg, 0.071 mmol) in THF (0.5 mL plus 0.5 mL as a rinse)
was added, and stirring was continued at -78 °C for 1 h. The
mixture was quenched by addition of saturated aqueous NH4Cl
(2 mL), the cold bath was removed, and the mixture was
extracted with EtOAc (3 ꢁ 10 mL). The combined organic
extracts were dried (Na2SO4) and evaporated to give an oily
residue which was kept under oil pump vacuum for 1 h and used
directly in the next step without purification.
Acetic Acid 2-Cyano-1-[(4R,8S,8aR,11aS)-rel-[decahydro-8-
methyl-6-oxo-6H-cyclopenta[i]quinolizin-4-yl]methyl]-2-propenyl
ester (5.6a,b). The above mixture of the alcohols 5.5a,b was
dissolved in CH2Cl2 (7 mL), and the solution was cooled to 0 °C.
Pyridine (0.39 mL, 4.8 mmol) and AcCl (0.28 mL, 3.9 mmol)
were added successively, and the mixture was stirred at 0 °C for
30 min. The ice bath was removed, and stirring was continued
for 2 h. Evaporation of the solvent and flash chromatography of
the residue over silica gel (2 ꢁ 15 cm), using 1:1 EtOAc/hexane,
gave the acetates 5.6a,b (0.255 g, 89% over two steps) as a
colorless oil, which was an inseparable mixture of two isomers:
FTIR (CH2Cl2 cast) 2223, 1746, 1649 cm-1; 1H NMR (CDCl3,
300 MHz) (mixture of two isomers) δ 0.88 (d, J = 6.2 Hz, 1.5 H),
0.90 (d, J = 6.2 Hz, 1.5 H), 1.24-2.03 (m, 15 H), 2.067 (s, 1.5 H),
2.071 (s, 1.5 H), 2.14-2.26 (m, 1.5 H), 2.33 (t, J = 4.5 Hz, 0.5 H),
2.39 (t, J = 4.2 Hz, 0.5 H), 2.74-2.84 (m, 0.5 H), 2.97-3.06 (m,
0.5 H), 3.19-3.37 (m, 1 H), 5.21-5.32 (m, 1 H), 5.97-6.05 (m,
2 H); 13C NMR (CDCl3, 125 MHz) (mixture of two isomers) δ
19.1 (d), 19.2 (d), 20.9 (t), 21.0 (q), 21.1 (q), 21.7 (t), 23.5 (t), 27.4
(t), 28.4 (t), 30.6 (t), 30.9 (t), 31.6 (q), 31.9 (q), 35.8 (t), 36.0 (t),
37.1 (t), 37.4 (t), 38.4 (t), 39.0 (t), 41.3 (t), 41.8 (t), 52.8 (d), 53.51
(d), 53.52 (d), 53.8 (d), 68.9 (s), 69.3 (s), 71.4 (d), 71.8 (d), 116.25
(s), 116.28 (s), 122.90 (s), 123.0 (s), 133.0 (t), 133.4 (t), 169.78 (s),
169.85 (s), 174.8 (s), 175.2 (s); exact mass (electrospray) m/z
calcd for C20H28N2NaO3 (M þ Na) 367.1992, found 367.1992.
(βR,1R,2S,90aS)-rel-70-Cyano-10,20,30,60,90,90a-hexahydro-β-
methylspiro[cyclopentane-1,40-[4H]quinolizine]-2-propanoic acid
methyl ester (1.4). Me3OþBF4- (0.52 g, 3.5 mmol) was added in
one portion to a stirred and cooled (0 °C) solution of 2,6-di-tert-
butyl-4-methylpyridine (0.88 g, 4.3 mmol) and 5.6a,b (mixture
of two isomers, 0.245 g, 0.712 mmol) in CH2Cl2 (7 mL). The ice
bath was removed, and stirring was continued for 2 h. The
solvent was then evaporated. The residue was taken up in
MeCN (14 mL) and cooled to 0 °C, and aqueous Na2CO3
(20% w/v, 7 mL) was added. The ice bath was removed, and
the mixture was stirred vigorously for 3 h. The aqueous phase
was extracted with EtOAc (3 ꢁ 15 mL), and the combined
organic extracts were dried (Na2SO4) and evaporated. Flash
chromatography of the residue over silica gel (2 ꢁ 15 cm), using
1:5 EtOAc/hexane, gave 1.4 (0.176 g, 83%) as a colorless solid:
mp 67-69 °C; FTIR (CH2Cl2 cast) 2217, 1735, 1654 cm-1; 1H
NMR (C6D6, 500 MHz) δ 0.70 (m, 1 H), 0.90 (d, J = 6.5 Hz,
3 H), 0.96-1.26 (m, 8 H), 1.39-1.60 (m, 6 H), 1.81-1.84 (m,
1 H), 2.06-2.12 (m, 1 H), 2.16-2.21 (m, 1 H), 2.77 (d, J =
16.1 Hz, 1 H), 3.10 (d, J = 16.0 Hz, 1 H), 3.16 (d, J = 13.6 Hz,
1 H), 3.39 (s, 3 H), 5.87-5.89 (m, 1 H); 13C NMR (C6D6,
125 MHz) δ 20.2 (d), 21.5 (t), 21.6 (t), 23.5 (t), 29.3 (t), 31.9 (q),
35.5 (t), 41.3 (t), 49.13 (t), 49.17 (t), 50.8 (d), 52.3 (d), 56.6 (q),
67.4 (s), 111.5 (s), 118.2 (s), 141.1 (d), 173.1 (s); exact mass
(electrospray) m/z calcd for C19H29N2O2 (M þ H) 317.2224,
found 317.2219.
Pyridine (0.1 mL), PhSeCN (52 μL, 0.42 mmol), and Bu3P
(0.11 mL, 0.42 mmol) were added successively dropwise to a
stirred and cooled (0 °C) solution of the above crude hydro-
xystannanes in THF (1 mL). The ice bath was removed, and the
mixture was stirred for 4 h. Evaporation of the solvent and flash
chromatography of the residue over silica gel (1.5 ꢁ 15 cm),
using hexane to 10:1 hexane/EtOAc, gave selenides 6.4 con-
taminated by impurities (33.1 mg, ca. 61% over two steps) as a
1
yellowish oil: FTIR (CH2Cl2 cast) 2954, 2927, 1463 cm-1; H
NMR (CDCl3, 500 MHz) δ 0.91 (t, J = 7.4 Hz, 12 H), 1.01 (dd,
J = 15.9, 7.4 Hz, 6 H), 1.33 (dd, J = 14.8, 7.4 Hz, 9 H),
1.49-1.55 (m, 11 H), 1.60-1.70 (m, 3 H), 1.78-2.10 (m, 6 H),
2.35 (td, J = 12.4, 2.6 Hz, 2 H), 2.79 (dd, J = 15.6, 1.8 Hz, 1 H),
3.02 (d, J = 16.1 Hz, 1 H), 3.12 (dd, J = 12.2, 4.7 Hz, 1 H),
3.86-3.96 (m, 4 H), 5.13 (s, 1 H), 5.81 (s, 1 H), 7.18-7.25 (m,
3 H), 7.48 (d, J = 7.1 Hz, 2 H); exact mass m/z calcd for
C38H62NO280Se120Sn (M - H) 764.2962, found 764.2966.
(1R,2S,90aS)-rel-2-[(1R-rel,5Z)-6-Chloro-8-[[(1,1-dimethylethyl)-
dimethylsilyl]oxy]-4-hydroxy-1-methyl-2-(phenylseleno)-5-octenyl]-
70-[1,3-dioxolan-3-yl]-10,20,30,60,90,90a-hexahydrospiro[cyclopen-
tane-1,40-[4H]quinolizine (6.6). n-BuLi (2.5 M in hexane,
0.08 mL, 0.20 mmol) was added dropwise to a stirred and cooled
(-78 °C) solution of the tin selenides 6.4 (38.5 mg, 0.050 mmol)
in THF (0.5 mL). Stirring at -78 °C was continued for 30 min.
Then aldehyde 6.58 (53 mg, 0.21 mmol) in THF (0.5 mL plus
0.5 mL as a rinse) was added dropwise. Stirring at -78 °C was
continued for 30 min. Saturated aqueous NH4C1 (2 mL) was
then added, and the mixture was extracted with EtOAc (3 ꢁ
7 mL). The combined organic extracts were dried (Na2SO4) and
evaporated. Flash chromatography of the residue over silica gel
(1 ꢁ 10 cm), using 1:4 to 1:1 EtOAc/hexane, gave phenylseleno
alcohols 6.6 which were dissolved in MeOH (2 mL) and used
directly for the next step.
(1R,2S,90aS)-rel-2-[(1R-rel,2E,5Z)-6-Chloro-8-[[(1,1-dimeth-
ylethyl)dimethylsilyl]-oxy]-4-hydroxy-1-methyl-2,5-octadienyl]-
70-[1,3-dioxolan-3-yl]-10,20,30,60,90,90a-hexahydrospiro[cyclope-
ntane-1,40-[4H]quinolizine] (6.7). Powdered NaHCO3 (32 mg),
NaIO4 (32 mg, 0.15 mmol), and water (0.4 mL) were added
successively to the above solution, and stirring was continued
for 24 h. The mixture was extracted with EtOAc (3 ꢁ 10 mL),
and the combined organic extracts were dried (Na2SO4) and
evaporated. Flash chromatography of the residue over silica
gel (1 ꢁ 10 cm), using 1:4 to 1:1 EtOAc/hexane, gave alcohols
6.7 as a colorless oil, which was used directly in the next step:
1H NMR (CDCl3, 500 MHz) (major isomer only) δ 0.059 (s, 6
H), 0.89 (s, 9 H), 1.00 (d, J = 6.8 Hz, 3 H), 1.10-1.26 (m, 3 H),
1.33-1.52 (m, 7 H), 1.71-1.95 (m, 4 H), 2.05-2.14 (m, 2 H),
2.35-2.39 (m, 1 H), 2.46-2.60 (m, 3 H), 2.84 (dd, J = 15.6,
2.7 Hz, 1 H), 3.10 (d, J = 15.6 Hz, 1 H), 3.78-3.82 (m, 2 H),
3.88-4.00 (m, 4 H), 5.03 (t, J = 7.2 Hz, 1 H), 5.16 (s, 1 H), 5.43
(dd, J = 15.6, 6.6 Hz, 1 H) 5.63 (d, J = 7.8 Hz, 1 H), 5.79-5.85
J. Org. Chem. Vol. 74, No. 19, 2009 7427

Liu et al.
JOCArticle
(m, 1 H), 6.04 (dd, J = 15.6, 7.6 Hz, 1 H); exact mass m/z calcd
for C31H5335ClNO4Si (M þ H) 566.3427, found 566.3426.
(4R,12Z,14R,15E,17S,17aR,20aS)-rel-12-Chloro-1,2,3,4,10,
11,14,17,17a,18,19,20-dodecahydro-14-hydroxy-17-methyl-8H-
4,7-ethanylylidene-6H-cyclopenta[f]pyrido[1,2-e][1,5]oxaazacy-
clopentadecin-8-one [(()-Halichlorine (7.5)]. N-Ethyl-N0-(3-
dimethylaminopropyl)carbodiimide hydrochloride (68 mg,
0.35 mmol) was added to a stirred solution of DMAP (77 mg,
114.0 (d), 128.8 (s), 129.8 (d), 159.6 (s), 161.8 (s), 191.6 (s); exact
mass (electrospray) m/z calcd for C15H20NO4 (M þ H) 278.1387,
found 278.1387.
(2R,4S)-2-[[[(1,1-Dimethylethyl)dimethylsilyl]oxy]methyl]-3,4-
dihydro-4-hydroxy-6-[[(4-methoxyphenyl)methoxy]methyl]-1(2H)-
pyridinecarboxylic acid phenylmethyl ester (8.10). NaBH4
(33 mg, 0.88 mmol) was added to a stirred and cooled
(-40 °C) slurry of 8.9 (0.23 g, 0.44 mmol) and CeCl3•7H2O
(326 mg, 0.875 mmol) in MeOH (9 mL). Stirring at -40 °C was
continued for 45 min, and the reaction was quenched with
acetone (0.2 mL) and saturated aqueous NH4Cl. The dry ice
bath was removed and replaced with an ice bath, and stirring
was continued for 30 min. The aqueous phase was extracted with
Et2O (3 ꢁ 5 mL), and the combined organic extracts were dried
(MgSO4) and evaporated. Flash chromatography of the color-
less residue over silica gel (1 ꢁ 10 cm), using 3:1 hexane/EtOAc,
gave 8.10 (0.198 g, 86%) as an oil: [R]25D þ44.2 (c 1.35, CHCl3);
0.63 mmol) and DMAP HCl (88 mg, 0.55 mmol) in CHCl3
3
(10 mL), and the resulting solution was refluxed. A solution of
7.3 [a 1.7:1 mixture of C(17) epimers] (3.4 mg, 6.3 μmol) in
CHCl3 (1 mL) was added to the above solution over 5 h (syringe
pump with gastight syringe). After two additional rinses of
CHCl3 (0.1 mL plus 0.1 mL) were added, the solution was
refluxed for a further 1 h. Heating was stopped, and stirring
was continued overnight. Evaporation of the solvent and flash
chromatography of the residue over silica gel (1 ꢁ 6 cm), using
1:3 EtOAc/hexane, gave 7.4 (1.9 mg, 58%) as a mixture of two
isomers which was used directly in the next step.
FTIR (CHCl3 cast) 3420, 1711, 1514, 1328, 1250 cm-1 1H
;
NMR (CDCl3, 400 MHz) δ -0.001 (s, 3 H), -0.005 (s, 3 H),
0.86 (s, 9 H), 1.38 (d, J = 7.2 Hz, 1 H), 1.73 (ddd J = 13.2, 10.0,
5.0 Hz, 1 H), 2.40-2.50 (m, 1H), 3.56 (t, J = 9.6 Hz, 1 H), 3.69
(dd, J = 9.6, 6.0 Hz, 1 H), 3.81 (s, 3 H), 4.28-4.54 (m, 6 H),
5.10-5.18 (m, 2 H), 5.33 (s, 1 H), 6.86 (d, J = 8.5 Hz, 2 H), 7.22
(d, J = 8.5 Hz, 2 H), 7.30-7.40 (m, 5 H); 13C NMR (CDCl3,
100 MHz) δ -5.53 (q), -5.47 (q), 18.2 (s), 25.8 (q), 31.9 (t), 55.3
(q), 55.8 (d), 60.7 (t), 62.2 (d), 67.8 (t), 70.3 (t), 71.9 (t), 112.9 (d),
113.7 (d), 128.1 (d), 128.2 (d), 128.5 (d), 129.2 (d), 130.3 (s), 135.9
(s), 136.2 (s), 153.7 (s), 159.1 (s); exact mass (electrospray) m/z
calcd for C29H41NNaO6Si (M þ Na) 550.2595, found 550.2593.
(2R,6S)-2-[[[(1,1-Dimethylethyl)dimethylsilyl]oxy]methyl]-3,6-
dihydro-6-[[(4-methoxyphenyl)methoxy]methyl]-6-(2-oxoethyl)-
1(2H)-pyridinecarboxylic acid phenylmethyl ester (9.1). A mix-
ture of 8.10 (0.153 g, 0.290 mmol), Hg(OAc)2 (9 mg, 0.03 mmol),
and Et3N (4 μL, 0.03 mmol) in n-butyl vinyl ether (3 mL) was
heated at 110 °C (oil bath temperature) in a sealed glass tube
(Teflon seal) for 36 h. Evaporation of volatile material and flash
chromatography of the residue over silica gel (1 ꢁ 10 cm), using
1:5 EtOAc/hexane, gave 9.1 (0.127 g, 79%) as a colorless oil:
HF pyridine (ca. 1.2 N, 0.2 mL) was added to a stirred
3
solution of 7.4 (1.9 mg, 3.7 μmol) in THF (2.5 mL), and stirring
was continued for 2 h. Saturated aqueous NaHCO3 (5 mL) was
added, and the mixture was extracted with EtOAc (3 ꢁ 5 mL).
The combined organic extracts were washed with brine, dried
(Na2SO4), and evaporated. Flash chromatography of the resi-
due over silica gel (1 ꢁ 6 cm), using 2:1 EtOAc/hexane, gave 7.5
as the more polar product [0.8 mg, 54%; this corresponds to
86% based on the proportion (ca. 63%) of the correct C(17)
isomer in the starting material] as a colorless glass: FTIR
(CH2Cl2 cast) 3407, 2929, 2872, 1715, 1659 cm-1 1H NMR
;
(CD3OD, 500 MHz) δ 1.01 (d, J = 6.8 Hz, 3 H), 1.13 (br d, J =
12.2 Hz, 1 H), 1.20-1.36 (m, 1 H), 1.43 (ddd, J = 12.0, 10.0, 3.0
Hz, 1 H), 1.50 (dddd, J = 13.0, 13.0, 13.0, 4.5 Hz, 1 H),
1.62-1.80 (m, 6 H), 1.95-2.03 (m, 1 H), 2.15-2.21 (m, 1 H),
2.55 (br d, J = 14.5 Hz, 1 H), 2.63 (dd, J = 19.5, 2.0 Hz, 1 H),
2.73 (dq, J = 7.1, 6.8 Hz, 1 H), 2.86 (ddd, J = 14.5, 12.5, 4.5 Hz,
1 H), 3.10-3.14 (m, 1 H), 3.21 (d, J = 17.5 Hz, 1 H), 3.44 (br d,
J = 17.5 Hz, 1 H), 3.98 (ddd, J = 11.5, 4.5, 2.0 Hz, 1 H), 4.62
(ddd, J = 14.5, 11.5, 3.0 Hz, 1 H), 5.03 (dd, J = 6.5, 4.5 Hz, 1 H),
5.35 (dd, J = 15.5, 4.0 Hz, 1 H), 5.57 (d, J = 7.5 Hz, 1 H), 5.75
(dd, J = 15.0, 8.5 Hz, 1 H), 7.03 (br s, 1 H); 13C NMR (CD3OD,
125 MHz) δ 18.1, 22.1, 22.3, 24.6, 24.9, 27.1, 32.1, 33.5, 33.7,
38.7, 41.8, 51.8, 51.9, 62.3, 69.5, 70.9, 128.3, 128.8, 129.7, 133.1,
136.9, 139.2, 167.6; exact mass m/z calcd for C23H3335ClNO3
(M þ H) 406.2144, found 406.2144.
[R]25 -42.0 (c 0.84, CHCl3); FTIR (CHCl3 cast) 1722, 1698,
D
1
1397, 1286 cm-1; H NMR (CDCl3, 400 MHz) δ -0.08 (br s,
6 H), 0.83 (s, 9 H), 2.14-2.24 (m, 1 H), 2.39 (dd, J = 17.4,
6.8 Hz, 1 H), 2.54 (dd, J = 16.0, 2.5 Hz, 1 H), 3.34-3.43 (m,
1 H), 3.46-3.86 (m, 4 H), 3.80 (s, 3 H), 4.30-4.48 (m, 3 H), 5.13
(AB q, J = 12.4 Hz, ΔνΑΒ = 19.2 Hz, 2 H), 5.79 (dd, J = 10.4,
3.0 Hz, 1 H), 5.84-5.92 (m, 1 H), 6.85 (d, J = 8.5 Hz, 2 H), 7.18
(d, J = 8.5 Hz, 2 H), 7.28-7.40 (m, 5 H), 9.61 (s, 1 H); 13C NMR
(CDCl3, 100 MHz) δ -5.53 (q), -5.48 (q), 18.1 (s), 23.9 (t), 25.8
(q), 48.1 (t), 52.1 (d), 55.2 (q), 58.6 (t), 63.2 (t), 67.1 (t), 72.6 (s),
73.1 (t), 113.8 (d), 123.3 (d), 127.8 (d), 128.0 (d), 128.5 (d), 129.6
(d), 130.1 (s), 130.5 (d), 136.4 (s), 155.1 (s), 159.2 (s), 201.1 (d);
exact mass (electrospray) m/z calcd for C31H44NO6Si (M þ H)
554.2932, found 554.2929.
The incorrect isomer of the starting material was separated,
but no attempt was made to recycle it (by oxidation and
reduction).
(2R)-2,3-Dihydro-2-(hydroxymethyl)-6-[[(4-methoxyphenyl)-
methoxy]methyl]-4(1H)-pyridinone (8.7). Cs2CO3 (0.65 g,
2.0 mmol) was added in three portions over ca. 15 min to a
stirred solution of 8.6 (0.201 g, 0.663 mmol) in MeOH (12 mL).
The mixture was stirred for 1 h at room temperature and then at
80 °C. When the reaction was complete (usually within 6 h, TLC
control), the solution was cooled and evaporated. The residue
was diluted with water (10 mL) and EtOAc (10 mL), and the
aqueous phase was extracted with EtOAc (3 ꢁ 10 mL). The
combined organic extracts were dried (MgSO4) and evaporated.
Flash chromatography of the yellow residue over silica gel (2 ꢁ
10 cm), using 1:9 MeOH/EtOAc, gave 8.7 (0.147 g, 80% yield) as
Acknowledgment. We thank NSERC for financial
support, Dr. R. McDonald for X-ray measurements, and
Professor H. Arimoto (Tohoku University) for a copy of the
13C NMR spectrum of natural halichlorine. D.L. and M.Y.
held Province of Alberta Graduate Fellowships.
a greasy solid: [R]25 -158.3 (c 0.78, CHCl3); FTIR (CHCl3
Supporting Information Available: Experimental proce-
dures, copies of NMR spectra for new compounds and crystal-
lographic information in CIF format. The X-ray data has been
deposited with the Cambridge Crystallographic Data Centre
and assigned the registry numbers CCDC 733408 for 4.4a and
CCDC 733409 for 4.8a. This material is available free of charge
via the Internet at http://pubs.acs.org.
D
cast) 3397, 3290, 1612, 1572, 1531, 1515 cm-1 1H NMR
;
(CDCl3, 400 MHz) δ 2.20-2.50 (m, 3 H), 3.60-3.85 (m, 3 H),
3.81 (s, 3 H), 4.13 (AB q, J = 14.4 Hz, ΔνΑΒ = 36.8 Hz, 2 H),
4.48 (s, 2 H), 4.93 (s, 1 H), 6.04 (br s, 1 H), 6.89 (d, J = 8.5 Hz,
2 H), 7.26 (d, J = 8.5 Hz, 2 H); 13C NMR (CDCl3, 100 MHz)
δ 37.9 (t), 54.2 (d), 55.3 (q), 64.1 (t), 68.0 (t), 72.9 (t), 95.9 (d),
7428 J. Org. Chem. Vol. 74, No. 19, 2009