An aldol-based approach to the asymmetric synthesis of L-callipeltose, the deoxyamino sugar of L-callipeltoside A.

Article abstract of DOI:10.1021/ol016416e

[reaction: see text] The L-callipeltose subunit of L-callipeltoside A has been synthesized in 10 steps and 13% overall yield from D-threonine. The key steps are a highly diastereoselective Felkin anti aldol addition to a methyl ketone and a selective methylation of a secondary alcohol in the presence of a secondary carbamate.

Full text of DOI:10.1021/ol016416e

ORGANIC
LETTERS
2001
Vol. 3, No. 20
3133-3136
An Aldol-Based Approach to the
Asymmetric Synthesis of L-Callipeltose,
the Deoxyamino Sugar of
L-Callipeltoside A
David A. Evans,* Essa Hu, and Jason S. Tedrow
Department of Chemistry and Chemical Biology, HarVard UniVersity,
Cambridge, Massachusetts 02138
eVans@chemistry.harVard.edu
Received July 10, 2001
ABSTRACT
The L-callipeltose subunit of L-callipeltoside A has been synthesized in 10 steps and 13% overall yield from D-threonine. The key steps are a
highly diastereoselective Felkin anti aldol addition to a methyl ketone and a selective methylation of a secondary alcohol in the presence of
a secondary carbamate.
Isolated from the lithistid sponge Callipelta sp. by Minale
and co-workers in 1996, the callipeltosides A-C (1a-c,
Figure 1) became the first members of a new class of marine
natural products with unique structural complexity and
biological activity.1 While all three callipeltosides share the
same dienyne-trans-chlorocyclopropane side chain appended
to the 14-membered aglycon, they are differentiated by their
appended sugar subunits. In particular, callipeltoside A (1a)
contains a highly functionalized deoxyamino sugar (calli-
peltose). Callipeltoside A exhibits cytotoxic activity against
the NSCLC-N6 human bronchopulmonary non-small-cell
lung carcinoma and P388 cell lines.1a The relative stereo-
chemical relationships of callipeltose to the macrolactone
have been proposed on the basis of extensive 2D-NMR
spectroscopic studies; however, the relative stereochemical
relationships of the chlorocyclopropane side chain to the
macrolactone and the absolute stereochemistry of callipel-
toside A remain to be determined. In our effort to synthesize
callipeltoside A and to elucidate its stereochemical ambigu-
ities, the macrolide was divided into three subunits: the
sugar 2, the side chain 32, and the macrolactone 4 (Scheme
1). Practical synthetic routes have been designed in order to
(1) (a) Zampella, A.; D’Auria, M. V.; Minale, L.; Debitus, C.; Roussakis,
C. J. Am. Chem. Soc. 1996, 118, 11085-11088. (b) Zampella, A.;
D’Auria, M. V.; Minale, L.; Debitus, C. Tetrahedron 1997, 53, 3243-
3248.
Figure 1. Proposed structures of callipeltosides A-C.
10.1021/ol016416e CCC: $20.00 © 2001 American Chemical Society
Published on Web 09/12/2001

Scheme 1. Callipeltoside Subunits and Retrosynthesis
Scheme 2. Diastereoselective Aldol Bond Constructiona
access both enantiomers of each of the three subunits. In
this Letter, we describe an asymmetric synthesis of L-
callipeltose.
a Reaction conditions: (a) NaOH, Cbz-Cl, MeCN, H2O; (b) MeI,
K2CO3, DMF; (c) TsOH, dimethoxypropane, benzene; (d) iPrMgCl,
HN(OMe)Me‚HCl, THF; (e) MeMgBr, THF; (f) LDA, then 6; (g)
HF, THF; (h) AcOH, H2O; (i) Me3O‚BF4, 2,6-di-tert-butyl-4-
methylpyridine.
To date, reported syntheses of callipeltose have begun with
manno-pyranoside, rhamno-pyranoside, or D-glucal.3 The
need to synthesize both enantiomers of callipeltose prompted
us to design a synthesis plan around the utilization of either
of the readily available D- or L-threonine enantiomers. Using
both enolate geometry and Felkin control elements, we
planned to use a diastereoselective aldol addition to threonine
methyl ketone to establish both the C2′ and C3′ stereocenters
of callipeltose 2 (Scheme 1).
The synthesis began with the well-established protection
procedure for D-threonine to afford the N-Cbz-D-threonine
methyl ester 5 in three steps and 93% yield (eq 1). The
addition of the N,O-dimethylhydroxyamine-derived magne-
sium amide to ester 5 provided the Weinreb amide in 93%
yield,4 which was transformed to methyl ketone 6 with
methylmagnesium bromide in 69% yield (eq 1).5
The design of a stereoselective aldol addition to methyl
ketone 6 was the pivotal step to this synthesis. Due to
reactivity considerations, lithium and sodium enolates were
evaluated (Scheme 2). Unfortunately, aldol additions of metal
enolates of R-methoxy esters and R-hydroxy esters favored
the formation of the undesired syn aldol adducts such as 7,
presumably as a result of metal chelation with the R-alkoxy
substituent reinforcing the formation of (E) enolates. For
example, the lithium enolate derived from ethyl R-methoxy-
acetate afforded a 2:1 ratio of the Felkin aldol adducts 7
and 8. The structure of the syn Felkin adduct 7 was un-
equivocally established by X-ray crystallography (Figure 2).
Both aldol adducts were then individually transformed to
afford lactones 9 and 10a. NOE studies carried on these
lactone diastereomers established that both aldol adducts
possessed the desired absolute stereochemical relationship
at the 3′ tertiary carbinol center and a diastereomeric
relationship at the 2′ center. These results then directed our
attention to the related aldol reactions of the lithium enolate
derived from 1,4-dioxaspiro[4.5]decan-2-one 11,6 which
would afford the complementary (Z) enolate geometry. In
the event, the aldol addition with methyl ketone 6 afforded
the Felkin anti aldol adduct 12 with a 15:1 diastereomeric
ratio and 80% yield (Scheme 2). The stereochemical assign-
(2) Our synthesis of the side chain has been published: Evans, D. A.;
Burch, J. D. Org. Lett. 2001, 3, 503-505.
(3) (a) Smith, G. R.; Finley, J. J. IV; Giuliano, R. M. Carbohydr. Res.
1998, 308, 223-237. (b) Gurjar, M. K.; Reddy, R. Carbohydrate Lett. 1998,
3, 169-172. (c) Pihko, A. J.; Nicolaou, K. C.; Koskinen, A. M. P.
Tetrahedron: Asymmetry 2001, 12, 937-942.
(4) Williams, J. M.; Jobson, R. B.; Yasuda, N.; Marchesini, G.; Dolling,
U.-H.; Grabowski, E. J. J. Tetrahedron Lett. 1995, 36, 5461-5464.
(5) Other conditions and nucleophiles, such as methyllithium, lead to
lower yields.
Figure 2. X-ray structure of aldol adduct 7.
Org. Lett., Vol. 3, No. 20, 2001
3134

ment of 12 was made through the lactone 10b, which was
directly correlated with its methylated counterpart 10a.
Scheme 3 illustrates the transition state model proposed
that the (E) enolate still exhibits a preferred addition to the
si face of methyl ketone 6. This suggests that either the steric
repulsion predicted by the polar Felkin-Nguyen model was
not sufficiently significant for the (E) enolate to reverse its
si face preference or that there exists another low energy
pathway that can accommodate the R-methoxy substituent
on the (E) enolate while still favoring si face addition to the
ketone. It is also worth noting that the assertion that there
should be a reversal in carbonyl face selectivity for enolate
14 is based on trends established for R-methyl substituted
aldehydes.9 Aldol addition studies to R-alkoxy aldehydes
reported by Heathcock10 also suggest that carbonyl face
reversal with enolates such as 14 may not be general for all
chiral aldehyde substrates.
Scheme 3. Asymmetric Induction and Felkin-Nguyen Model
The conversion of R-hydroxy lactone 10b to lactone
methyl ether 10a required methylation conditions that would
differentiate the 2′-secondary alcohol from the 4′ secondary
carbamate. Strongly basic conditions, such as sodium hydride
or sodium tert-butoxide, either epimerized the C2′ alcohol
stereocenter or opened the lactone ring. The addition of silver
oxide and methyl iodide to lactone 10b afforded only 40%
of the desired product at room temperature, and elevated
temperatures induced N-methylation of the carbamate. The
use of methyl triflate and 2,6-di-tert-butyl-4-methylpyridine
resulted in decomposition of the lactone. Meerwein’s reagent
(Me3O‚BF4) and proton sponge (N,N,N′,N′-tetramethyl-1,8-
naphthalenediamine) afforded less than 50% of the desired
product, and separation of the base from the polar product
was tedious. Fortunately, a combination of Meerwein’s
reagent and 2,6-di-tert-butyl-4-methylpyridine in dichlo-
romethane provided the desired product 10a in 82% yield
(Scheme 2).11
Two transformations remained to complete the synthesis
of callipeltose: formation of the cyclic carbamate and
reduction of the lactone to lactol (Scheme 4). Initial attempts
to form the cyclic carbamate followed by reduction of the
lactone resulted in preferential opening of the more reactive
cyclic carbamate. Thus, our attention turned toward lactone
reduction followed by formation of a stable glycoside
intermediate prior to carbamate cyclization. Diisobutylalu-
minum hydride reduction of lactone 10b to lactol 19 followed
by acidic methanolysis, however, did not provide the desired
pyranoside 20a but yielded the pyrrolidine derivative 20b
instead. (eq 6).
to explain the stereochemical outcome. The polar Felkin-
Nguyen model7 predicts that nucleophilic addition to the si
face of the methyl ketone 6 should be expected (eq 2). The
Zimmerman-Traxler transition state model8 suggests that
(Z) enolate addition to the methyl ketone should afford the
anti aldol adduct 12 (eq 3). Accordingly, the stereochemical
outcome of this transformation is in full accord with the
indicated models.
Rationalization of the stereochemical outcome of the aldol
addition of (E) enolate 14 to methyl ketone 6 is less
straightforward. On the basis of the polar Felkin-Nguyen
model, (E) enolate 14 should experience steric repulsion from
the threonine side chain in the Zimmerman-Traxler transi-
tion state (eq 4). Minimization of this unfavorable interaction
should have resulted in an anti Felkin addition of the (E)
enolate to the re face of methyl ketone 6 (eq 5). However,
the X-ray structure of syn aldol adduct 7 (Figure 2) reveals
This undesired lactol rearrangement was circumvented
using Rychnovsky’s one-pot lactone reduction-acylation
procedure.12 Lactone 10a was treated with diisobutylalumi-
num hydride, and the resulting reduction product, without
isolation, was treated with acetic anhydride, pyridine, and
DMAP to give the six-membered anomeric acetate 16 in 93%
yield and 9:1 anomeric ratio (Scheme 4). Addition of sodium
hydride transformed the N-Cbz group on acetate 16 into the
cyclic carbamate, while the benzyl alkoxide byproduct
(6) (a) Pearson, W. H.; Cheng, M.-C. J. Org. Chem. 1987, 52, 1353-
1355. (b) Pearson, W. H.; Cheng, M.-C. J. Org. Chem. 1987, 52, 3176-
3178. (c) Pearson, W. H.; Hines, J. V. J. Org. Chem. 1989, 54, 4235-
4237. (c) Comins, D. L.; Kuethe, J. T.; Lakner, F. J. J. Am. Chem. Soc.
1999, 121, 2651-2652.
(7) One of our referees pointed out that the surname of author “Nguyen
Trong Anh" is actually Nguyen because surnames are listed first according
to Vietnamese custom. Therefore, we will refer to the polar model as the
Felkin-Nguyen model instead of the Felkin-Anh model. (a) Cherest, M.;
Felkin, H.; Prudent, N. Tetrahedron Lett. 1968, 2199. (b) Nguyen, T. A.;
Eisenstein, O. NouV. J. Chim. 1977, 1, 61-70. (c) Nguyen, T. A. Top.
Curr. Chem. 1980, 88, 146-162.
(9) For a discussion of (Z) enolate additions to chiral substituted
aldehydes, see: (a) Evans, D. A.; Nelson, J. V.; Taber, T. R. Top.
Stereochem. 1982, 13, 1-115. (b) Roush, W. R. J. Org. Chem. 1991, 56,
4151-4157.
(10) Heathcock, C. H.; Young, S. D.; Hagan, J. P.; Pirrung, M. C.; White,
C. T.; VanDerveer, D. J. Org. Chem. 1980, 45, 3846-3856.
(8) Zimmerman, H. E.; Traxler, M. D. J. Am. Chem. Soc. 1957, 79,
1920-1923.
Org. Lett., Vol. 3, No. 20, 2001
3135

nitrile afforded callipeltose glycosyl donor 17 in 51% yield
for the three transformations.14 The addition of methanol and
TMS-OTf transformed the glycosyl donor to callipeltose
methyl glycoside 18, whose NMR spectrum matched that
reported by Giuliano and co-workers.3a
Scheme 4. Assembly of L-Callipeltosea
In summary, an asymmetric synthesis of L-callipeltose has
been completed in 10 steps and 13% overall yield from
commercially available D-threonine. Both enantiomers of
callipeltose can be accessed by the use of either D- or
L-threonine. This synthetic route is also ideal for investigation
of glycosidation of sugar 2 to macrolactone 4, since the
callipeltose lactol formed can be readily converted into
various kinds of glycosyl donors. The total synthesis of
callipeltoside A and the elucidation of its absolute stereo-
chemical structure will be reported in due course.
Acknowledgment. Support has been provided by the
National Institutes of Health (GM 33327-16) and Merck.
Supporting Information Available: Experimental pro-
cedures and characterization of compounds 5, 6, 10a, 10b,
12, 16, and 17. This material is available free of charge via
the Internet at http://pubs.acs.org.
OL016416E
(11) Evans, D. A.; Trotter, B. W.; Coleman, P. J.; Cote, B.; Dias, L. C.;
Rajapakse, H. A.; Tyler, A. N. Tetrahedron 1999, 55, 8671-8726.
(12) (a) Dahanukar, V. H.; Rychnovsky, S. D. J. Org. Chem. 1996, 61,
8317-8320. (b) Kopecky, D. J.; Rychnovsky, S. D. J. Org. Chem. 2000,
65, 191-198.
(13) Benzyl acetate was a major byproduct in the reaction mixture,
supporting the proposed role of benzyl alkoxide in acetate cleavage.
(14) Sebesta, D. P.; Roush, W. R. J. Org. Chem. 1992, 57, 4799-4802.
cleaved the anomeric acetate in situ to give the callipeltose
lactol 2.13 The NMR spectrum of the lactol matched that
reported by Nicolaou and co-workers.3c Treatment of the
lactol with a catalytic amount of DBU and trichloroaceto-
3136
Org. Lett., Vol. 3, No. 20, 2001