Asymmetric synthesis of (+)-castanospermine through enol ether metathesis-hydroboration/oxidation

Article abstract of DOI:10.1039/b901488h

An asymmetric synthesis of (+)-castanospermine is presented in which enol ether metathesis-hydroboration/oxidation is used for stereoselective installation of the trans-trans hydroxyl groups on the piperidine ring of the alkaloid.

Full text of DOI:10.1039/b901488h

View Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/obc | Organic & Biomolecular Chemistry
Asymmetric synthesis of (+)-castanospermine through enol ether
metathesis–hydroboration/oxidation†
Julien Ceccon, Gre´gory Danoun, Andrew E. Greene and Jean-Franc¸ois Poisson*
Received 23rd January 2009, Accepted 12th March 2009
First published as an Advance Article on the web 23rd March 2009
DOI: 10.1039/b901488h
An asymmetric synthesis of (+)-castanospermine is presented
in which enol ether metathesis–hydroboration/oxidation is
used for stereoselective installation of the trans-trans hydroxyl
groups on the piperidine ring of the alkaloid.
(+)-Castanospermine (1, Fig. 1) is a polyhydroxylated indolizidine
first isolated in 1981 from Castanospermum australe,1 and later
from Alexa leiopetala.2 This alkaloid, because of its resemblance
to glucose, is a strong inhibitor of a- and b-glucosidases.
(+)-Castanospermine and certain of its derivatives have been
advanced for possible use in the treatment of cancers, diabetes,
viral infections, and immunosuppressive deficiencies.3
Fig. 2 Retrosynthetic analysis.
bond is far from obvious: epoxidation, followed by ring opening,
Fig. 1 (+)-Castanospermine (1).
would be expected to lead to the alternative, undesired trans diol,
embedded in a trans-cis array, as the major product through a
more favorable transition state.6 On the other hand, selective
protection of the hydroxyl group(s) after cis dihydroxylation5d
to permit subsequent inversion at C-6 would, if in fact feasible,
most likely require several unattractive protection/deprotection
steps. There was, though, an option that seemed worth exploring:
modification of the above-mentioned ring-closing metathesis in
which an alkoxy-substituted olefin would now be used as one
of the olefinic partners.7 We postulated that the hydroboration
of the resultant cyclic enol ether (III, X = OR¢¢) would lead
directly to the required trans-trans hydroxyl stereochemistry on
the six-membered ring because of steric shielding by the allylic
substituent, which would be expected to force hydroboration to
occur on the concave face of the molecule.8
Hydroboration had never been applied to an alkoxy-substituted
cyclic olefin with an allylic oxygen substituent. Therefore, to
study the selectivity and efficiency of the proposed hydrob-
oration, a model substrate was first prepared (Scheme 1).
3,5-Dimethoxypyridine9 was reduced in the presence of benzyl
chloroformate to afford an intermediate bis-enol ether, which
was hydrolyzed with aqueous acid. The resulting vinylogous acid
was etherified with diazomethane to yield the corresponding
vinylogous ester 3, which was reduced with lithium aluminium
hydride to give piperidine 4 in good overall yield from 2. A sample
of 4 was converted into benzyl ether 5 to study the influence, if
any, of hydroxyl-group protection on the stereochemical outcome
of the hydroboration.
Due to its biological activity and intrinsic complexity—
an amino tetraol with five contiguous stereocenters—
(+)-castanospermine has been the subject of numerous total
synthesis studies over the past 30 years.4 As four of the five
stereocenters in this indolizidine correspond to those in D-glucose,
it is not surprising that a majority of the syntheses to date have
started from sugars. What is surprising, though, is that there are
only four syntheses that are non chiral-pool,4a-d an approach
that generally offers more flexibility for analog preparation. In
a program aimed at the total synthesis of polyhydroxylated
indolizidines from non chiral-pool starting material through
[2 + 2] cycloaddition of dichloroketene to chiral enol ethers,
(+)-castanospermine was targeted. Its efficient synthesis, which
showcases a new approach to the often problematic introduction
of the trans-trans triol array on six-membered rings, is described
below.
The six membered-ring of indolozidines has been obtained5
through ring-closing metathesis, but this approach has to date
never been used for the synthesis of (+)-castanospermine. The for-
mation of the trans-trans (all equatorial) triol arrangement found
in (+)-castanospermine from a 6,7-dehydroindolizidine (e.g., III,
X = H, Fig. 2) through overall trans dihydroxylation of the double
Universite´ Joseph Fourier, De´partement de Chimie Mole´culaire (SERCO)
UMR-5250, ICMG FR-2607, CNRS, BP 53, 38041, Grenoble Cedex 9,
France. E-mail: jean-francois.poisson@ujf-grenoble.fr; Fax: +33 4 76 63 57
54; Tel: +33 4 76 63 59 86
† Electronic supplementary information (ESI) available: Complete char-
acterization data and 1H and 13C NMR spectra for new compounds. See
DOI: 10.1039/b901488h
The hydroboration/oxidation was first tested by treatment of
4 with the borane-dimethyl sulfide complex in THF at 0 ◦C,
followed by oxidation of the resulting organoborane using sodium
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2029–2031 | 2029
©

View Online
With the completely diastereoselective model-system re-
sults as support for our synthetic strategy, the synthesis of
(+)-castanospermine was undertaken (Scheme 3). The preparation
of intermediate 15a began, as earlier,5d with the synthesis of
dichloro enol ether 10, which was generated by reaction of the
potassium alkoxide of (R)-stericol10 with trichloroethylene (79%
yield). Treatment of 10 with butyllithium produced sequentially
deprotonation, cis b-elimination, and chlorine–lithium exchange11
to afford the intermediate lithio-ynol ether, which was allylated
to give 11. The ynol ether was next reduced by using Dibal-H
at 50 ◦C to the corresponding enol ether 12, which underwent
diastereoselective cycloaddition with dichloroketene to provide
dichlorocyclobutanone 13 (dr = 96:4). Beckmann ring expansion
with Tamura’s reagent and dechlorination then afforded g-lactam
14 in 35% overall yield (5 steps, 81%/step). Allylic oxidation of
14 under Sharpless conditions afforded the corresponding allylic
alcohols 15a,b in 56% yield (75% brsm) as a 1:1 diastereomeric
mixture.12 The desired cis-anti isomer 15a could be obtained as the
major product (92:8 diastereoisomeric ratio) in 82% yield (2 steps)
by applying an oxidation/reduction sequence to 15a,b.5d
Scheme 1 Model substrate synthesis.
Table 1 Hydroboration of piperidines 4 and 5
Entry Cpd Hydroboration
Oxidation
Yielda drb
1
2
3
4
5
6
7
8
4
4
4
4
4
4
5
5
BH3·SMe2, THF NaOH/H2O2
BH3·SMe2, THF Me3NO
BH3·SMe2, THF NaBO3
51%
> 98 : 2
c
66%
> 98 : 2
> 98 : 2
BH3·SMe2, Et2O NaOH/H2O2/EtOH 65%
d
9-BBN
NaOH/H2O2
d
Catecholborane NaOH/H2O2
BH3·SMe2, THF NaOH/H2O2
BH3·SMe2, THF NaBO3
61%
68%
> 98 : 2
> 98 : 2
a Isolated yield. b Determined by 1H NMR. c Only decomposition. d No
hydroboration (starting material recovered).
hydroperoxide. Encouragingly, the triol derivative 6 was isolated
as a single isomer in 51% yield (Table 1, entry 1). Other oxidants
that could improve the overall yield of the transformation were
then tested: the use of trimethylamine N-oxide surprisingly led to
complete decomposition of the intermediate borane adduct (entry
2), but sodium perborate did produce an improved yield (66%,
entry 3). Performing the hydroboration in diethyl ether instead of
THF, followed by sodium hydroperoxide oxidation in the presence
of ethanol, was nearly as effective, however (entry 4). Neither
9-BBN nor catecholborane led to hydroboration (entries 5 and 6).
The hydroboration/oxidation sequence (of entries 1 and 3) applied
to the benzyl-protected derivative 5 again provided a single isomer
7 in good yield (entries 7 and 8), a better result once more being
obtained with sodium perborate as the borane oxidant.
The relative configuration of triol derivatives 6 and 7 could
be assigned by high-temperature NMR (to eliminate signal
broadening due to conformers) of acetates 8 and 9 (Scheme 2).
The two 3J coupling constants of 8.4 Hz in 8 and 9 are diagnostic
of diaxial relationships of the three contiguous carbinolic protons.
Noe experiments provided additional proof of the all equatorial
arrangement in both molecules.
a
Scheme 3 Intermediate lactam 15a synthesis.
DCK = dichloro-
b
ketene. MSH = O-mesitylenesulfonylhydroxylamine. St = 1-(2,4,6-
triisopropylphenyl)ethyl. brsm = based on recovered starting material.
The allylic hydroxyl group was next protected as its tri-
isopropylsilyl ether, which allowed N-alkylation with 3-iodo-2-
(methoxymethoxy)prop-1-ene13 under phase transfer conditions
to proceed smoothly to give 16 in 77% overall yield (Scheme 4). The
key ring-closing metathesis of 16 could not be achieved, though,
despite considerable effort with a number of different catalysts
under various reaction conditions. This failure can reasonably
be ascribed, at least to some extent, to steric hindrance by the
TIPS ether, which would favor formation of the ruthenium Fischer
carbene, non-reactive in this system.14 Pleasingly, after removal
of the TIPS protecting group (TBAF, 87%), the diene readily
underwent ring-closing metathesis in the presence of Grubbs II
catalyst to produce 17 in 78% yield. This RCM product was
accompanied by 9% of the ethyl ketone, the result of olefin
isomerization in the allylic alcohol; the presence of benzoquinone
did not significantly suppress its formation.15
With the cyclic enol ether 17 in hand, the above studied
hydroboration/oxidation sequence could now be applied. To
our delight, addition of excess borane-dimethyl sulfide complex
followed by sodium perborate to 17 in THF afforded the product
Scheme 2 Relative configuration assignments.
2030 | Org. Biomol. Chem., 2009, 7, 2029–2031
This journal is
The Royal Society of Chemistry 2009
©

View Online
Notes and references
1 L. D. Hohenschutz, E. A. Bell, P. J. Jewess, D. P. Leworthy, R. J.
Pryce, E. Arnold and J. Clardy, Phytochemistry, 1981, 20, 811–
814.
2 R. J. Nash, L. E. Fellows, J. V. Dring, C. H. Stirton, D. Carter,
M. P. Hegarty and E. A. Bell, Phytochemistry, 1988, 27, 1403–
1404.
3 For a review, see: (a) A. A. Watson, G. W. J. Fleet, N. Asano, R. J.
Molyneux and R. J. Nash, Phytochemistry, 2001, 56, 265–295; See also:
(b) G. K. Ostrander, N. K. Scribner and L. R. Rohrschneider, Cancer
Res., 1988, 48, 1091–1094; (c) P. S. Sunkara, D. L. Taylor, M. S. Kang,
T. L. Bowlin, P. S. Liu, A. S. Tyms and A. Sjoerdsma, Lancet, 1989,
1, 1206; (d) V. A. Johnson, B. D. Walker, M. A. Barlow, T. J. Paradis,
T. C. Chou and M. S. Hirsch, Antimicrob. Agents Chemother., 1989,
33, 53–57; (e) D. L. Taylor, P. S. Sunkara, P. S. Liu, M. S. Kang, T. L.
Bowlin and A. S. Tyms, AIDS, 1991, 5, 693–8; (f) D. O. Willenborg,
C. R. Parish and W. B. Cowden, Immunol. Cell Biol., 1992, 70, 369–
377.
4 For non chiral-pool syntheses, see: (a) J.-L. Reymond, A. A. Pinkerton
and P. Vogel, J. Org. Chem., 1991, 56, 2128–2135; (b) S. E. Denmark
and E. A. Martinborough, J. Am. Chem. Soc., 1999, 121, 3046–
3056; (c) P. Somfai, P. Marchand, S. Torsell and U. M. Lindstrom,
Tetrahedron, 2003, 59, 1293–1299; (d) Z. Zhao, L. Song and P. S.
Mariano, Tetrahedron, 2005, 61, 8888–8894; For other recent strategies,
see:; (e) L. Cronin and P. V. Murphy, Org. Lett., 2005, 7, 2691–2693;
(f) N. S. Karanjule, S. D. Markad, V. S. Shind and D. D. Dhavale,
J. Org. Chem., 2006, 71, 4667–4670; (g) T. Machan, A. S. Davis, B.
Liawruangrath and S. G. Pyne, Tetrahedron, 2008, 64, 2725.
5 For recent examples, see: (a) A. O. H. El-Nezhawy, H. I. El-Diwani
and R. R. Schmidt, Eur. J. Org. Chem., 2002, 4137–4142; (b) K. L.
Chandra, M. Chandrasekhar and V. K. Singh, J. Org. Chem., 2002, 67,
4630–4633; (c) F. Chevallier, E. Le, Grognec, I. Beaudet, F. Fliegel, M.
Evain and J.-P. Quintard, Org. Biomol. Chem., 2004, 2, 3128–3133; (d) J.
Ceccon, A. E. Greene and J.-F. Poisson, Org. Lett., 2006, 8, 4739–4742.
6 (a) A. Fu¨rst and P. A. Plattner, Helv. Chim. Acta, 1949, 32, 275–283;
(b) H. O. House and W. A. B. Reading, Modern Synthetic Reactions,
1972, pp. 300–301; For a recent example, see: (c) K. C. Nicolaou, Y.-P.
Sun, X.-S. Peng, D. Ploet and D. Y.-K. Chen, Angew. Chem., Int. Ed.,
2008, 47, 7310–7313.
Scheme 4 Completion of the synthesis. a GII = Grubbs 2nd generation
catalyst. b MP = 2-methoxypropene.
of hydroboration/oxidation and lactam reduction 18, together
with the corresponding unreduced lactam 19 (1:1.5 ratio), both
as single isomers (dr > 95:5), in a combined yield of 58%.
Since efforts to increase the transformation of 17 into 18 at
the expense of 19 proved unrewarding, diol 19 was converted
into its acetonide derivative, which underwent smooth reduction
with BH3 to provide the protected indolizidine (70%, 2 steps).
The cleavage of the numerous protecting groups in the two
indolizidines with hydrochloric acid in ethanol led, in each case, to
(+)-castanospermine in > 90% yield. The synthetically derived
7 For
a review on the formation of alkoxy-substituted cyclic
olefins through metathesis, see: P. Van de Weghe, P. Bisseret, N.
Blanchard and J. Eustache, J. Organomet. Chem., 2006, 691, 5078–
5108.
8 For examples of hydroboration of alkoxy-substituted cyclic olefins, see:
(a) M. E. Jung and L. A. Light, J. Am. Chem. Soc., 1984, 106, 7614–
7618; (b) H. Mattes and C. Benezra, J. Org. Chem., 1988, 53, 2732–2737;
(c) P. E. Peterson and M. Stepanian, J. Org. Chem., 1988, 53, 1903–1907;
(d) G. Stork, P. C. Tang, M. Casey, B. Goodman and M. Toyota, J. Am.
Chem. Soc., 2005, 127, 16255–16262.
9 L. Testaferri, M. Tiecco, M. Tingoli, D. Bartoli and A. Massoli,
Tetrahedron, 1985, 41, 1373–1384.
10 (R)-1-(2,4,6-Triisopropylphenyl)ethanol. See: P. Delair, A. M.
Kanazawa, M. M. B. de Azevedo and A. E. Greene, Tetrahedron:
Asymmetry, 1996, 7, 2707–2710.
(+)-castanospermine ([a]20 +78.0 (c 0.75, H2O), mp 204–206 ◦C
D
(dec)) was spectroscopically and chromatographically identical
with an authentic sample of the natural product.16
In summary, we have developed a new, efficient strategy, based
on sequential enol ether metathesis—hydroboration/oxidation for
the introduction of the all equatorial hydroxyl substituents on
the six membered ring of (+)-castanospermine. Using this novel
approach, a highly stereoselective non chiral-pool synthesis of
(+)-castanospermine could be realized in 3.7% overall yield and
17 steps. Since several other biologically active polyol natural
products share this trans-trans hydroxyl arrangement (inter alia,
deoxynojirimycin, pancratistatin, siculinine, calystegin B2), this
metathesis–hydroboration/oxidation tandem should find addi-
tional application.
11 B. Darses, A. Milet, C. Philouze, A. E. Greene and J.-F. Poisson, Org.
Lett., 2008, 10, 4445–4447.
12 J. Ceccon, A. E. Greene and J.-F. Poisson, Synlett, 2005, 1413–
1416.
13 (a) X.-P. Gu, I. Ikeda and M. Okahara, Bull. Chem. Soc. Jpn., 1987, 60,
397–398; (b) X.-P. Gu, N. Nishida, I. Ikeda and M. Okahara, J. Org.
Chem., 1987, 52, 3192–3196; (c) S. Z. Janicki, J. M. Fairgrieve and P. A.
Petillo, J. Org. Chem., 1998, 63, 3694–3700.
14 (a) V. K. Aggarwal and A. M. Daly, Chem. Commun., 2002, 2490–2491;
(b) J. Louie and R. H. Grubbs, Organometallics, 2002, 21, 2153–2164;
(c) U. Majumder, J. M. Cox, H. W. B. Johnson and J. D. Rainier,
Chem.–Eur. J., 2006, 12, 1736–1746.
15 S. H. Hong, D. P. Sanders, C. W. Lee and R. H. Grubbs, J. Am. Chem.
Soc., 2005, 127, 17160–17161.
16 Purchased from Sigma-Aldrich.
Acknowledgements
Financial support from the CNRS and the Universite´ Joseph
Fourier (UMR 5250, ICMG FR 2607) is gratefully acknowledged.
We thank Professor P. Dumy for his interest in our work.
This journal is
The Royal Society of Chemistry 2009
Org. Biomol. Chem., 2009, 7, 2029–2031 | 2031
©