4,5-Diisopropyl-B-[(E)-[(3'-menthofuryl)dimethylsilyl]allyl]-1,3,2-dio xaborolane, an improved chiral reagent for the anti α-hydroxyallylation of aldehydes: Application to the enantioselective synthesis of (-)-swainsonine

Article abstract of DOI:10.1021/jo961840s

Diisopropyl tartrate-modified (E)-[γ-(furyldimethylsilyl)allyl]boronates 16 and 11 were designed for the enantioselective synthesis of substituted anti-3,4-dihydroxy-1-butenes 9 via the anti-α-hydroxyallylation of aldehydes. The reactions of aldehydes with 16 and 11 provided furyldimethylsilyl-substituted anti-1,2-silanols 12 and 13 with good enantioselectivity (74-82% ee), and the silanols were oxidized to the corresponding anti-1,2-diols 9 by a modified Fleming-Tamao oxidation protocol. The high reactivity of the (menthofuryl)silane unit toward electrophilic substitution allowed complete protodesilylation of menthofuryl-substituted silanols 13 and subsequent oxidation to diols 9 in a one-pot operation without competing protodesilylation of the allylsilane unit. However, a two-stop protocol was required for the protodesilylation-oxidation of the less reactive 2-(5-methyl)furyl-substituted silanols 12. Reagents 16 and 11 exhibited similar diastereoselectivity in double asymmetric reactions with three chiral aldehydes 23-25 (80 to >92% de, with the exception of the mismatched reactions with aldehyde 25, which were essentially unselective). However, [[[2-(5-methylfuryl)]dimethylsilyl]allyl]boronate 16 could only be synthesized in ≤15% yield, and oxidations of the 2-(5-methylfuryl)-substituted silanols 12 were lower-yielding than oxidations of the corresponding menthofuryl-substituted silanols 13. Thus, diisopropyl tartrate-modified (E)-[γ-[(menthofuryl)dimethylsilyl]allyl]boronate 11 proved to be the more useful of the two reagents. As a demonstration of the utility of 11 in the synthesis of polyhydroxylated natural products, this new reagent was used in a brief and highly stereoselective synthesis of the naturally occurring alkaloid (-)-swainsonine (42).

Full text of DOI:10.1021/jo961840s

1112
J . Org. Chem. 1997, 62, 1112-1124
4,5-Diisop r op yl-B-[(E)-[(3′-m en th ofu r yl)d im eth ylsilyl]a llyl]-1,3,2-
d ioxa bor ola n e, a n Im p r oved Ch ir a l Rea gen t for th e An ti
r-Hyd r oxya llyla tion of Ald eh yd es: Ap p lica tion to th e
En a n tioselective Syn th esis of (-)-Sw a in son in e
J . A. Hunt and William R. Roush*
Department of Chemistry, Indiana University, Bloomington, Indiana 47405
Received September 26, 1996X
Diisopropyl tartrate-modified (E)-[γ-(furyldimethylsilyl)allyl]boronates 10 and 11 were designed
for the enantioselective synthesis of substituted anti-3,4-dihydroxy-1-butenes 9 via the anti-R-
hydroxyallylation of aldehydes. The reactions of aldehydes with 10 and 11 provided furyldimeth-
ylsilyl-substituted anti-1,2-silanols 12 and 13 with good enantioselectivity (74-82% ee), and the
silanols were oxidized to the corresponding anti-1,2-diols 9 by a modified Fleming-Tamao oxidation
protocol. The high reactivity of the (menthofuryl)silane unit toward electrophilic substitution
allowed complete protodesilylation of menthofuryl-substituted silanols 13 and subsequent oxidation
to diols 9 in a one-pot operation without competing protodesilylation of the allylsilane unit. However,
a two-step protocol was required for the protodesilylation-oxidation of the less reactive 2-(5-methyl)-
furyl-substituted silanols 12. Reagents 10 and 11 exhibited similar diastereoselectivity in double
asymmetric reactions with three chiral aldehydes 23-25 (80 to >92% de, with the exception of the
mismatched reactions with aldehyde 25, which were essentially unselective). However, [[[2-(5-
methylfuryl)]dimethylsilyl]allyl]boronate 10 could only be synthesized in e15% yield, and oxidations
of the 2-(5-methylfuryl)-substituted silanols 12 were lower-yielding than oxidations of the
corresponding menthofuryl-substituted silanols 13. Thus, diisopropyl tartrate-modified (E)-[γ-
[(menthofuryl)dimethylsilyl]allyl]boronate 11 proved to be the more useful of the two reagents.
As a demonstration of the utility of 11 in the synthesis of polyhydroxylated natural products, this
new reagent was used in a brief and highly stereoselective synthesis of the naturally occurring
alkaloid (-)-swainsonine (42).
The stereoselective synthesis of carbohydrates and
other polyoxygenated natural products from acyclic pre-
cursors remains a topic of considerable interest. The
challenge of creating intermediates with multiple con-
tiguous oxygenated stereogenic centers has received
substantial attention from organic chemists, and a
variety of strategies for meeting this challenge have been
described.1-6 The R-hydroxyallylation of aldehydes with
(γ-alkoxyallyl)metal reagents (such as 1 or 2) is a
particularly attractive strategy, as these R-hydroxyally-
lation reactions generate syn- or anti-1,2-diols in concert
with the formation of a carbon-carbon bond.7-9
developed highly enantioselective procedures for the
synthesis of syn diol monoethers 4 via the reactions of
aldehydes with (Z)-(γ-alkoxyallyl)boranes10 or (Z)-(γ-
alkoxyallyl)stannanes.9,11-13 However, efforts to synthe-
size substituted anti-3,4-dihydroxy-1-butenes 3 from (E)-
(γ-alkoxyallyl)metal reagents have not met with as much
success. (E)-(γ-Alkoxyallyl)boron reagents have proved
difficult to synthesize because of the configurational
instability of the (E)-γ-alkoxyally anion precursors.14,15
Takai’s (γ-alkoxyallyl)chromium reagent, generated in
situ by the reduction of an acrolein acetal with CrCl2,
offers direct and often high-yielding access to anti-1,2-
diol monoethers 3, but this reagent suffers from poor
diastereoselectivity in a number of cases.16 Recently,
Marshall has circumvented some of these difficulties by
transmetalating enantioenriched R-alkoxyallylic stan-
nanes with InCl3, presumably giving rise to allylindium
reagents, which react with aldehydes enantioselectively
to provide anti-1,2-diol monoethers 3 as the major
products.17,18
Brown,10 Marshall,9 Yamamoto,11,12 and we13 have
X Abstract published in Advance ACS Abstracts, J anuary 15, 1997.
(1) Zamojski, A.; Banaszek, A.; Grynkiewicz, C. Adv. Carbohydr.
Chem. Biochem. 1982, 40, 400.
(2) McGarvey, G. J .; Kimura, M.; Oh, T.; Williams, J . M. J .
Carbohydr. Chem. 1984, 3, 125.
(3) Danishefsky, S. J .; DeNinno, M. P. Angew. Chem., Int. Ed. Engl.
1987, 26, 15.
(4) Trends in Synthetic Carbohydrate Chemistry; Horton, D., Hawk-
ins, L. D., McGarvey, G. J ., Eds.; American Chemical Society: Wash-
ington, D. C., 1989.
(5) Ager, D. J .; East, M. B. Tetrahedron 1992, 48, 2803.
(6) Hudlicky, T.; Entwistle, D. A.; Pitzer, K. K.; Thorpe, A. J . Chem.
Rev. 1996, 96, 1195.
(7) Roush, W. R. In Comprehensive Organic Synthesis; Heathcock,
C. H., Ed.; Pergamon Press: Oxford, 1991; Vol. 2; p 1.
(8) Yamamoto, Y.; Asao, N. Chem. Rev. 1993, 93, 2207.
(9) Marshall, J . A. Chem. Rev. 1996, 96, 31 and references cited
therein.
(10) Brown, H. C.; J adhav, P. K.; Bhat, K. S. J . Am. Chem. Soc.
1988, 110, 1535.
(11) Kadota, I.; Kobayashi, K.; Okano, H.; Asao, N.; Yamamoto, Y.
Bull. Soc. Chim. Fr. 1995, 132, 615.
(12) Yamamoto, Y.; Kobayashi, K.; Okano, H.; Kadota, I. J . Org.
Chem. 1992, 57, 7003.
(14) Hoffmann, R. W.; Kemper, B. Tetrahedron Lett. 1980, 21, 4883.
(15) Hoffmann, R. W.; Metternich, R.; Lanz, J . W. Liebigs Ann.
Chem. 1987, 881.
(13) Roush, W. R.; VanNieuwenhze, M. S. J . Am. Chem. Soc. 1994,
116, 8536.
(16) Takai, K.; Nitta, K.; Utimoto, K. Tetrahedron Lett. 1988, 29,
5263.
S0022-3263(96)01840-3 CCC: $14.00 © 1997 American Chemical Society

Enantioselective Synthesis of (-)-Swainsonine
J . Org. Chem., Vol. 62, No. 4, 1997 1113
The initial difficulties associated with the stereoselec-
tive synthesis of (E)-γ-alkoxyallyl metal reagents14,15
prompted the development of indirect methods for syn-
thesis of substituted anti-3,4-dihydroxy-1-butenes. The
(E)-[γ-[(dialkylamino)silyl]allyl]zinc reagent developed by
Tamao and Ito reacts diastereoselectively with aldehydes
to give anti-silanol intermediates that are then oxidized
with retention of C-Si stereochemistry to the correspond-
ing anti-3,4-diols 3 (R ) H).19 We20-22 and, subsequently,
Barrett23 extended the Tamao strategy to the asymmetric
synthesis of anti-1,2-diols with the use of (E)-γ-silyl-
substituted allylboron reagents containing, respectively,
diisopropyl tartrate and B-diisopinocampheyl24 chiral
auxiliaries. Brown has recently developed a related
strategy for the indirect R-hydroxyallylation of aldehydes
via reactions with an (E)-γ-boryl-substituted allyldiiso-
pinocampheylborane.25
sequently, allylboronate 6 did not prove to be a useful
reagent for the synthesis of anti-1,2-diols 9.
Because furans undergo a variety of electrophilic
substitution reactions, including protodesilylation, much
faster than do correspondingly substituted phenyl groups,31
we anticipated that replacement of the phenyldimethyl-
silyl group in silanols 8 with a furyldimethylsilyl group
would allow us to convert the resulting silanols 12 or 13
to diols 9 via a modified Tamao-Fleming oxidation.
Stork had previously demonstrated that furylsilanes can
function as hydroxyl surrogates, via cleavage to fluoro-
silanes (which are suitable substrates for the Tamao
oxidation), upon treatment with n-Bu4NF in THF.32 Also,
subsequent to the initiation of our work, Kocienski
demonstrated that allylic furyldimethylsilanes can be
converted to allylic alcohols by a sequence involving
photooxidation of the furylsilane unit prior to the Tamao-
Fleming oxidation.33
We report herein the synthesis of the furyldimethyl-
silyl-substituted allylboronates 10 and 11 and the devel-
opment of mild protodesilylation-oxidation conditions
suitable for elaboration of allylic silanes 12 and 13 to diols
9 in good to excellent yields. As a demonstration of the
utility of these new reagents in the synthesis of polyhy-
droxylated natural products, we also describe the use of
reagent 11 in a brief and highly stereoselective synthesis
of the naturally occurring alkaloid (-)-swainsonine. A
preliminary account of these investigations has been
reported.34
We previously developed the diisopropyl tartrate-
modified (E)-γ-silyl-substituted allylboronate reagents 5
and 6 for the synthesis of anti-1,2-diols 9.20-22 (Cyclo-
hexyloxy)dimethylsilyl-substituted allylboronate 5 reacts
with achiral aldehydes (R′CHO) to provide anti-1,2-
silanols 7 in good yield, though with only moderate
enantioselectivity (64-72% ee);20,22 silanols 7 were then
oxidized to anti-1,2-diols 9 via the Tamao procedure.26,27
Unfortunately, although the phenyldimethylsilyl-substi-
tuted allylboronate 6 proved to be more enantioselective
than 5 (81-87% ee), the resulting anti-1,2-silanols 8
could not be oxidized to the corresponding diols 9.21,22
Fleming and co-workers have shown that the phen-
yldimethylsilyl group can be converted to a hydroxyl with
retention of stereochemistry via electrophilic substitution
of the phenyl group to provide a heteroatom-substituted
silane, which can then be oxidized to the corresponding
alcohol.26,28,29 However, electrophilic substitution reac-
tions of the allylsilane unit are faster than those of the
phenylsilane group; thus, the Fleming procedure cannot
be used to convert a phenyldimethylsilyl unit to a
hydroxyl group in the presence of an allylsilane.30 Con-
Diisop r op yl Ta r tr a te-Mod ified (E)-[γ-[[2-(5-Meth -
ylfu r yl)]d im et h ylsilyl]a llyl b or on a t e. The proto-
desilylation of aryltrimethylsilanes has been thoroughly
studied by Earborn and co-workers.31,35,36 Various sub-
stituted phenyltrimethylsilanes [e.g., anisyltrimethyl-
silane (krel ) 1500) and mesityltrimethylsilane (krel
)
53 600)] and heteroaryltrimethylsilanes [e.g., 2-(phe-
nylthio)trimethylsilane (krel ) 4800) and 2-furyltrimeth-
ylsilane (krel ) 17 200)] undergo protodesilylation much
more readily than does phenyltrimethylsilane.15a,16 The
very large difference (>104) in the rate of protodesilyla-
tion of furyltrimethylsilane compared to phenyltrimeth-
ylsilane suggested that it might be possible to protode-
silylate a furylsilane in the presence of an allylsilane.
We initially chose to use the 5-methylfuryl-substituted
allylsilane 15 rather than one containing a simpler
(unsubstituted) furan because the 5-methyl group would
prevent metalation at the C(5) position during the
synthesis of allylboronate 10.
(17) Marshall, J . A.; Hinkle, K. W. J . Org. Chem. 1995, 60, 1920.
(18) Marshall, J . A.; Hinkle, K. W. J . Org. Chem. 1996, 61, 105.
(19) Tamao, K.; Nakajo, E.; Ito, Y. J . Org. Chem. 1987, 52, 957.
(20) Roush, W. R.; Grover, P. T.; Lin, X. Tetrahedron Lett. 1990,
31, 7563.
(21) Roush, W. R.; Grover, P. T. Tetrahedron Lett. 1990, 31, 7567.
(22) Roush, W. R.; Grover, P. T. Tetrahedron 1992, 48, 1981.
(23) Barrett, A. G. M.; Malecha, J . W. J . Org. Chem. 1991, 56, 5243.
(24) Racherla, U. S.; Brown, H. C. J . Org. Chem. 1991, 56, 401.
(25) Brown, H. C.; Narla, G. J . Org. Chem. 1995, 60, 4686.
(26) J ones, G. R.; Landais, Y. Tetrahedron 1996, 52, 7599.
(27) Tamao, K.; Ishida, N.; Kumada, M. J . Org. Chem. 1983, 48,
2120.
(28) Fleming, I.; Henning, R.; Plaut, H. J . Chem. Soc., Chem.
Commun. 1984, 29.
(29) Fleming, I.; Henning, R.; Parker, D. C.; Plaut, H. E.; Sanderson,
P. E. J . J . Chem. Soc., Perkin Trans. 1 1995, 317.
(30) Fleming has recently reported the use of the (2-methylbut-2-
enyl)diphenylsilyl group as a hydroxyl surrogate and has shown that
this new unit can be converted to a hydroxyl in the presence of a 1,2-
disubstituted allylsilane (Fleming, I.; Winter, S. B. D. Tetrahedron Lett.
1993, 45, 7287). However, we suspect that this new silyl group cannot
be used in the synthesis of allylboronates like 6, 10, or 11 because the
2-methylbut-2-enyl moiety might be metalated competitively with the
allyl unit under the reaction conditions.
(31) Earborn, C.; Sperry, J . A. J . Chem. Soc. 1961, 4921.
(32) Stork, G. Pure Appl. Chem. 1989, 61, 439.
(33) Norley, M. C.; Kocienski, P. J .; Faller, A. Synlett 1994, 77.
(34) Hunt, J . A.; Roush, W. R. Tetrahedron Lett. 1995, 36, 501.
(35) Earborn, C. J . Chem. Soc. 1956, 4858.
(36) Deans, F. B.; Earborn, C. J . Chem. Soc. 1959, 2299.

1114 J . Org. Chem., Vol. 62, No. 4, 1997
Hunt and Roush
Allylsilane 15 was prepared from 2-methylfuran by
metalation with n-BuLi in THF and subsequent silylation
with allyldimethylchlorosilane (14).37 As a test of our
hypothesis that protodesilylation of the furylsilane would
be faster than the allylsilane, a solution of 15 in CH2Cl2
at 0 °C was treated with 1.1 equiv of trifluoroacetic acid.
Although we were not able to isolate the intermediate
product, presumably the trifluoroacetate-substituted di-
methylallylsilane, addition of cyclohexanol and triethyl-
amine to the crude product provided (cyclohexyloxy)-
dimethylallylsilane (16) in 76% yield.
ate esterification of the allylboronic acid with either
(R,R)- or (S,S)-diisopropyl tartrate provided reagent 10
that was contaminated with ca. 20% of DIPT; the yield
was generally 50-65%. Reagent 10 was stored over 4 Å
molecular sieves at -20 °C as a 0.5 M solution in toluene.
Although the reagent was sufficiently pure to use in the
R-hydroxyallylation of a variety of aldehydes, the yield
of 10 synthesized by this two-step procedure was very
low (e15% from 15).
Allylboronate 10 was prepared from allylsilane 15
using the procedure previously employed for the synthe-
sis of γ-silyl-substituted allylboronates 5 and 6.22,38 Thus,
a solution of 15 in THF was treated with potassium tert-
butoxide and n-butyllithium (1.0 equiv each) at -40 °C
for 15-20 min and then cooled to -78 °C and treated
with triisopropyl borate (1.0 equiv) for 15 min. The
mixture was poured into aqueous NH4Cl, extracted with
Et2O, treated with either (R,R)- or (S,S)-diisopropyl
tartrate (1.0 equiv), and then dried over MgSO4 and
concentrated in vacuo.
We examined the possibility of replacing the 5-meth-
ylfuryl unit in 10 with other aryl groups, in the hope of
improving the yield of the allylboronate synthesis. We
found that allylsilanes 19a and 19b derived from anisole
(18a ) and m-methylanisole (18b) were readily converted
to the corresponding allylboronate reagents 20a and 20b
under the conditions described for the synthesis of
reagent 10, but the â-hydroxyallylsilanes resulting from
addition of these reagents to a simple aldehyde could not
be protodesilylated with CF3CO2H. Mesityldimethyl-
allylsilane (19c) was compatible with CF3CO2H proto-
desilylation, but we had difficulty synthesizing this
allylsilane in acceptable yield or purity.
Both the yield and the purity of reagent 10 were low,
and our attempts to purify 10 were unsuccessful. Reac-
tions of crude 10 with various aldehydes were low yield-
ing, presumably because of the impurity of the reagent.
We attempted to improve the yield and purity of 10 by
varying the reaction temperature and duration, and by
experimenting with different bases (including n-BuLi,
n-BuLi with TMEDA, s-BuLi, and s-BuLi with TMEDA),
but the yield (< 30%) and the purity remained low.
To improve the purity of the reagent, we turned to a
two-step procedure that proceeds by way of the crystal-
line diethanolamine complex 17.38 Treatment of the
crude allylboronic acid with 0.3 equiv of diethanolamine
provided crude 17 that was purified by recrystallization
(23% yield from 15). The amount of diethanolamine used
in this experiment was dictated by the amount of
allylboronic acid estimated to be present in the product
of the metalation procedure; use of larger amounts of
diethanolamine gave an oily product from which crystal-
lization of 17 was exceedingly difficult. Hydrolysis of 17
in a two-phase mixture of EtOAc and brine and immedi-
We also studied the metalation of several furyl-
substituted allylsilanes (15, 19d , and 19e) and one
phenylthio-substituted allylsilane (19f). We subjected
these allylsilanes to the metalation conditions for the
synthesis of boronate reagent 10 and then quenched the
(E)-γ-silyl allyl anions with deuterated methanol instead
of triisopropyl borate. With allylsilane 15, we suspected
that the furan C(5)-methyl group might be metalated
under the reaction conditions, but when only 1 equiv each
(37) Ha¨bich, D.; Effenberger, F. Synthesis 1979, 841.
(38) Roush, W. R.; Ando, K.; Powers, D. B.; Halterman, R. L.;
Palkowitz, A. D. J . Am. Chem. Soc. 1990, 112, 6339.
2
of n-BuLi and KO-t-Bu were used, H NMR analysis of

Enantioselective Synthesis of (-)-Swainsonine
Ta ble 1. r-Hyd r oxya llyla tion of Ach ir a l Ald eh yd es
allylation stepa
J . Org. Chem., Vol. 62, No. 4, 1997 1115
oxidation stepb,c
product
yieldd (%)
entry
R′CHO
reagent
R
product
yieldd (%)
% eee
1
c-C6H11CHO
(S,S)-10
(S,S)-11
(R,R)-5
(R,R)-6
5-methylfuryl-
menthofuryl-
C6H11O-
12a
13a
7a
83
87
95
88
(R,S)-9a
(R,S)-9a
(S,R)-9a
61
79
95
82
81
72
87g
2
3f
4f
Ph-
8a
5
n-C5H11CHO
(S,S)-10
(S,S)-11
(R,R)-5
(R,R)-6
5-methylfuryl-
menthofuryl-
C6H11O-
12b
13b
7b
77
86
86
95
(R,S)-9b
(R,S)-9b
(S,R)-9b
43
82
95
74
74
64
81g
6
7f
8f
Ph-
8b
a
The allylboration reactions of 10 and 11 were performed by treatment of a 0.5 M solution of aldehyde in toluene with a 0.5 M solution
of each allylboronate reagent in toluene. The reactions were run at -78 °C, in the presence of activated, crushed 4 Å molecular sieves.
The reactions with 10 were quenched after 4 h by filtration through a plug of silica gel, whereas the reactions with 11 were quenched
b
after four hours by addition of a -78 °C, 1 M solution of NaBH4 in EtOH. A two-step protodesilylation/oxidation sequence was used for
silanols 12a and 12b: the silanols were treated with CH3CO2H in CH2Cl2 at 0 °C for 15 min, then solvent was removed in vacuo, the
residue was dissolved in 1:1 THF-MeOH, and H2O2 (20 equiv), KHCO3 (2 equiv), and KF (2 equiv) were added at ambient temperature.
c A one-pot protodesilylation/oxidation sequence was used for silanols 13a and 13b: these silanols were treated with CH3CO2H in THF
d
at -0 °C to room temperature (ca. 1 h), then MeOH (cosolvent), H2O2 (20 equiv), KHCO3 (2 equiv), and KF (2 equiv) were added. Yield
of product isolated chromatographically. e Determined by NMR analysis of the bis-MTPA esters prepared from 9a and 9b, unless indicated
g
otherwise. f The data summarized in entries 3, 4, 7, and 8 are reproduced from ref 20. % Ee’s for the vinylogous oxidation products of
8a and 8b determined as described in ref 22.
the recovered allylsilane showed that no metalation had
occurred at that site. However, the mass balance of the
reaction was low, indicating that decomposition of the
allylsilane was occurring under the reaction conditions.
Similar metalation studies with heteroaryl-substituted
allylsilanes 19d -f and a variety of bases (n-BuLi-KO-
t-Bu, n-BuLi, and s-BuLi) also resulted in extensive
decomposition of the starting materials.
onate 11 was then prepared from allylsilane 22 following
the general outline of our synthesis of reagent 10 (see
Experimental Section for complete details). The crude
product, consisting primarily of reagent 11, DIPT, and
1
residual 22, was analyzed by H NMR spectroscopy to
determine the weight percentage of 11 in the mixture;
the yield of 11 was generally 65-75%. Reagent 11 was
stored over 4 Å molecular sieves at -20 °C as a 0.5 M
solution in toluene.43
We concluded that deprotonation at C(3) of the furan
or thiophene nucleus and subsequent eliminative ring-
opening39 was most likely responsible for the decomposi-
tion of furyl- and phenylthio-substituted allylsilanes 15
and 19d -f during the metalation reactions, especially
as it is known that a C(2)-silyl substituent favors this
type of ring-opening reaction.40,41 Accordingly, we syn-
thesized allylsilane 19g, which bears a methyl at the
furan C(3) and thus cannot be metalated at that site. This
allylsilane was efficiently metalated and converted to the
corresponding allylboronate 20g, and it was compatible
with CF3CO2H protodesilylation. However, 2,3,4-alkyl-
substituted furans such as 18g, the precursor for allyl-
silane 19g, are not readily available from simple starting
materials; in this case, 18g was prepared in five steps
from cyclohexene-1-acetonitrile.42 Because the furan
moiety of allylboronate 20g is destined to be discarded
immediately after the allylation reaction, we decided that
the synthesis of furan 18g was, for our purposes, pro-
hibitively long.
Fortunately, at this point we made the serendipitous
discovery that the 2,3,4-alkyl-substituted furan 21 is
commercially available from the Flavors and Fragrances
Division of Aldrich for less than $1 per gram. Known as
menthofuran, 21 is a natural product, one of the fragrant
components of peppermint oil.
Diisop r op yl Ta r tr a te-Mod ified (E)-[γ-[(Men th o-
fu r yl)d im et h ylsilyl]a llyl]b or on a t e 11. [(Mentho-
furyl)dimethylallyl]silane 22 was prepared from men-
thofuran 21 by metalation and subsequent silylation with
allyldimethylchlorosilane (14).37 (E)-(γ-Silylallyl)bor-
An ti r-Hyd r oxya llyla tion of Ach ir a l Ald eh yd es.
The enantioselectivity of the new reagents 10 and 11 was
assessed by the allylation of two representative achiral
aldehydes, cyclohexanecarboxaldehyde and hexanal (see
Table 1). Following our standard procedure for the
allylation of aldehydes with diisopropyl tartrate-modified
allylboronates,38,44 we treated a 0.5 M solution of each
aldehyde in toluene with a 0.5 M solution of each
allylboronate reagent in toluene. The reactions were run
at -78 °C, in the presence of activated, crushed 4 Å
molecular sieves. The reactions with reagent 10 were
quenched after four hours by filtration through a plug of
silica gel, whereas the reactions with reagent 11 were
quenched after 4 h by addition of a -78 °C, 1 M solution
of NaBH4 in EtOH.44 The 3,4-anti-silanols 12a ,b and
13a ,b were the only products observed in all cases.
(43) Owing to the stereocenter in the menthofuryl residue, (R,R)-
and (S,S)-11 are in fact diastereomers, and not enantiomers. However,
because this stereocenter is so far removed from the site of developing
stereochemistry in the allylboration transition state, we assume that
it has an insignificant influence on the allylboration stereoselectivity
and feel safe in treating (R,R)-11 and (S,S)-11 as “pseudoenantiomers.”
The enantiomeric purity of the menthofuran used in the preparation
of (R,R)-11 and (S,S)-11 was not determined.
(39) Gilchrist, T. L. Adv. Heterocycl. Chem. 1987, 41, 41.
(40) Karlsson, J . O.; Gronowitz, S.; Frejd, T. J . Org. Chem. 1982,
47, 374.
(41) Gronowitz, S.; Frejd, T.; Karlsson, J . O.; Lawitz, K.; Pedaja,
P.; Pettersson, K. Chem. Scr. 1981, 18, 192.
(42) Grieco, P. A.; Pogonowski, C. S.; Burke, S. J . Org. Chem. 1975,
40, 542.
(44) Roush, W. R.; Hoong, L. K.; Palmer, M. A. J .; Park, J . C. J .
Org. Chem. 1990, 55, 4109.

1116 J . Org. Chem., Vol. 62, No. 4, 1997
Hunt and Roush
The silanols 12a ,b and 13a ,b were oxidized to the
corresponding anti-1,2-diols 9a and 9b by a modified
Fleming-Tamao oxidation.19,26-29 The 5-methylfuryl-
substituted silanols 12a and 12b were dissolved in CH2-
Cl2 and cooled to 0 °C. Addition of CF3CO2H resulted in
rapid (>15 min) protodesilylation of the furyl nucleus,
and each of the subsequent oxidations was effected by
removal of the solvent in vacuo, dissolution of the residue
in 1:1 THF-MeOH, and addition of H2O2 (20 equiv),
KHCO3 (2 equiv), and KF (2 equiv).
data for the analogous reactions with boronate reagents
5 and 6 are also presented in Table 1 (entries 3, 4, 7,
and 8).22
An ti r-Hyd r oxya llyla tion of Ch ir a l Ald eh yd es.
Double asymmetric reactions of R-hydroxyallylation re-
agents 10 and 11 with three representative chiral alde-
hydes, glyceraldehyde acetonide 23, (2S,3R)-6-bromo-2,3-
epoxyhexanal (24), and (2R)-3-[(tert-butyldimethylsilyl)-
oxy]-2-methylpropanal (25), were also examined in order
to more fully document the stereoselectivity of these new
reagents.47 Results of these experiments are summarized
in Table 2, along with comparative data for the reactions
of 23, 24, and 25 with the first-generation allylboronate
reagent 5.22
The reactions of glyceraldehyde acetonide (23) with
both enantiomers of reagents 10 and 1143 (Table 2,
entries 1-4) proceeded with excellent diastereoselectiv-
ity. This is perhaps not surprising, since 23 is an
excellent (high selectivity) substrate for the tartrate
ester-modified allylborating reagents.48 The matched
double asymmetric reactions of 23 with (S,S)-10 and
(S,S)-11, like the previously reported reaction with (S,S)-
5,22 provided g95:5 selectivity for diastereomer 26, while
the mismatched double asymmetric reactions with (R,R)-
10 and (R,R)-11 showed selectivity of 92-94:8-6 favoring
diastereomer 27.49 This is a considerable improvement
over the 85:15 selectivity obtained in the mismatched
reaction of 23 with (R,R)-5 (Table 2, entry 6).22
Protodesilylation of the menthofuryl-substituted sil-
anols 13a and 13b with CF3CO2H in CH2Cl2 was nearly
instantaneous even at -40 °C. This very rapid proto-
desilylation of the menthofuran group, as compared to
the 5-methylfuran group, is no doubt due to steric and
hyperconjugative effects.31 Because of this increased
rate, we were able to perform the protodesilylation
reactions of the menthofuryl-substituted silanols in THF,
which made the protodesilylation-oxidation sequence a
very simple one-pot procedure.45 Once the protodesily-
lation reactions were complete (ca. 1 h at 0 °C to room
temperature), the oxidation reactions were performed in
situ by addition of H2O2 (20 equiv), KHCO3 (2 equiv), KF
(2 equiv), and MeOH (cosolvent). The protodesilylation-
oxidation sequence is clearly more efficient with the
menthofuryl-substituted silanols 13a and 13b than with
the corresponding 2-(5-methylfuryl)-substituted silanols
12a and 12b (see Table 1), most likely because of the
milder protodesilylation conditions.
Like 23, the double asymmetric reactions of epoxy
aldehyde 24 gave consistently excellent results with both
enantiomers of 10 and 1143 (Table 2, entries 7-10).50 This
is in striking contrast to the mismatched reaction with
(S,S)-5, which provided only a 2:1 mixture of diastereo-
mers 28c:29c (Table 2, entry 11). However, the enan-
tiomeric purity of 24 (92% ee; determined by Mosher ester
analysis of the corresponding epoxy alcohol) puts an
upper limit on the diastereoselectivity that can be
achieved in these reactions. Assuming that enantiomeri-
cally pure 24 had been used, it can be calculated that
the diastereoselectivity of the reactions with (S,S)-10
(Table 2, entry 7), (R,R)-10 (Table 2, entry 8), and (S,S)-
11 (Table 2, entry 9) would have been 93:7, 7:93, and 93:
7, respectively, whereas the matched double asymmetric
reaction of 24 and (R,R)-11 would have provided a 1:99
We were unable to determine the ee’s of silanols 12
and 13 directly by Mosher ester analysis46 because the
esterification reactions were not efficient (presumably
due to steric hindrance by the neighboring trialkylsilyl
groups). However, the diols 9a and 9b were readily
converted into the bis-MTPA esters, and these diesters
were used to determine the enantioselectivity of the
R-hydroxyallylation for each aldehyde. For comparison,
(47) Masamune, S.; Choy, W.; Petersen, J . S.; Sita, L. R. Angew.
Chem., Int. Ed. Engl. 1985, 24, 1.
(48) Roush, W. R.; Hoong, L. K.; Palmer, M. A. J .; Straub, J . A.;
Palkowitz, A. D. J . Org. Chem. 1990, 55, 4117.
(49) The intrinsic diastereofacial selectivity of 23 was assessed via
reaction with pinacol (E)-[γ-(dimethylphenylsilyl)allyl]boronate (see ref
22), which provided a 58:42 mixture of diastereomers 26d :27d (R )
Ph). Therefore, diastereomer 26 is the product of matched double
asymmetric reactions with the chiral allylboronates 10 and 11.
(50) We did not determine the instrinsic face selectivity of epoxy
aldehyde 24. However, it is reasonable to expect that diastereofacial
selectivity of 24 will be similar to the structurally related aldehyde
(2R,3S)-4-(benzyloxy)-2,3-epoxybutanal, for which the facial selectivity
has been determined [60:40 favoring the diastereomer corresponding
to 29d (R ) Ph) in reaction with pinacol (E)-[γ-(dimethylphenylsilyl)-
allyl]boronate; see ref 22]. On this basis, diastereomer 29 should be
the product of the matched reactions of 24 with (R,R)-10 and (R,R)-
11.
(45) The Lewis basicity of THF significantly attenuates the reactiv-
ity of the CF3CO2H, and thus, THF is not a suitable solvent for the
protodesilylation of the less reactive [[2-(5-methylfuryl)]allyl]silanes,
which were protodesilylated in CH2Cl2. However, the subsequent
Tamao oxidation is not efficient in either neat CH2Cl2 or CH2Cl2-
MeOH mixtures, so the CH2Cl2 was removed in vacuo and replaced
with THF-MeOH for the oxidation. With the very reactive (mentho-
furyl)allylsilanes, we were able to perform the protodesilylation in THF
and simply add the MeOH cosolvent at the start of the Tamao
oxidation.
(46) Dale, J . A.; Dull, D. L.; Mosher, H. S. J . Org. Chem. 1969, 34,
2543.

Enantioselective Synthesis of (-)-Swainsonine
J . Org. Chem., Vol. 62, No. 4, 1997 1117
Ta ble 2. Dou ble Asym m etr ic r-Hyd r oxya llyla tion of Ch ir a l Ald eh yd es
allylation stepa
ratioe
oxidation stepb,c,d
product
yieldf
entry
aldehyde
23
reagent
R
products
yieldf
1
2
3
(S,S)-10
(R,R)-10
(S,S)-11
(R,R)-11
(S,S)-5
5-methylfuryl-
5-methylfuryl-
menthofuryl-
menthofuryl-
C6H11O-
26a :27a
26a :27a
26b:27b
26b:27b
26c:27c
26c:27c
95:5
6:94
93
90
79
85
82
88
32c
33c
32d
33d
32g
33g
65
65
77
82
92
95
>96:4
4
8:92
5g
6g
>95:5
15:85
(R,R)-5
C6H11O-
7
8
9
10
11
24h
(S,S)-10
(R,R)-10
(S,S)-11
(R,R)-11
(S,S)-5
5-methylfuryl-
5-methylfuryl-
menthofuryl-
menthofuryl-
C6H11O-
28a :29a
28a :29a
28b:29b
28b:29b
28c:29c
90:10i
10:90i
90:10j
4:96j
69
68
81
82
78
34c
57
h
h
h
h
34d
35d
34k
85
79
44
67:33
12
13
14
25
(S,S)-10
(R,R)-10
(S,S)-11
(R,R)-11
(S,S)-5
5-methylfuryl-
5-methylfuryl-
menthofuryl-
menthofuryl-
C6H11O-
30a :31a
30a :31a
30b:31b
30b:31b
30c:31c
30c:31c
>96:4
33:67
94:6
45:55
>95:5
36:64
67
58
82
80
76
70
36c
37c
36d
37d
36g
37g
56
53
82
76
88
86
15
16g
17g
(R,R)-5
C6H11O-
a
The allylboration reactions with reagents 10 and 11 were performed by treatment of a 0.5 M solution of aldehyde in toluene with a
0.5 M solution of each allylboronate reagent in toluene. The reactions were run at -78 °C, in the presence of activated, crushed 4 Å
molecular sieves. The reactions with 10 were quenched after 4 h by filtration through a plug of silica gel, whereas the reactions with 11
b
were quenched after 4 h by addition of a -78 °C, 1 M solution of NaBH4 in EtOH. All modified Tamao oxidations were performed on the
purified, major product of the allylboration experiments summarized in each entry. c The two-step protodesilylation/oxidation sequence
described in footnote (b) of Table 1 was used for silanols 26a , 27a , 28a , 30a and 31a (i.e., products of allylboration experiments with
d
5-methylfuryl-substituted reagent 10). The one-pot protodesilylation/oxidation sequence described in footnote (c) of Table 1 was used
for silanols 26b, 27b, 28b, 29b, 30b and 31b (i.e., products of allylboration experiments with menthofuryl-substituted reagent 11). e Product
ratios determined by 1H NMR analysis of the crude allylboration reaction mixture after filtration through a short plug of silica gel, with
g
care taken not to separate the reaction diastereomers. f Yield of product(s) isolated chromatographically. The data in entries 5, 6, 16,
h
and 17 are reproduced from ref 22. The enantiomeric purity of epoxyalcohol 24 used in these experiments was 92% ee. i If enantiomerically
pure epoxyaldehyde 24 had been used in this experiment (see footnote (g)), the ratio of 28a :29a would have been 93:7 and 7:93 in the
j
double asymmetric reactions with (S,S)- and (R,R)-10, respectively (see text for discussion). If enantiomerically pure epoxyaldehyde 24
had been used in this experiment (see footnote (g)), the ratio of 28b:29b would have been 93:7 and 1:>99 in the double asymmetric
k
reactions with (S,S)- and (R,R)-11, respectively (see text for discussion). The Tamao oxidation of 28c was performed under standard
conditions (2 equiv of KF‚2H2O, 2 equiv of KHCO3 and 20 equiv of 30% H2O2 in 1:1 THF-MeOH).
mixture favoring diastereomer 29b (Table 2, entry 10).51
Consistent with this analysis is the fact that the enan-
tiomeric purity of the minor product (i.e., 29b) from the
reaction of 24 with (S,S)-11 (Table 2, entry 9) was only
58% ee,52 which corresponds well with the % ee we
expected (e60% ee) for this minor product on the basis
of diastereoselectivities calculated for the reactions with
the enantiomerically pure epoxy aldehyde.
The matched double asymmetric reactions of R-methyl-
â-alkoxy aldehyde 25 with (S,S)-10 and (S,S)-11 pro-
ceeded with excellent stereoselectivity (g94:6; Table 2,
entries 12 and 14), but the mismatched reaction with
(R,R)-10 and (R,R)-11 did not display any significant
diastereoselectivity (Table 2, entries 13 and 15).53 These
results track the stereoselectivity pattern previously
documented for the reactions of 25 with both enantiomers
of 5 (Table 2, entries 16 and 17).22
The modified Fleming-Tamao oxidations of the 2-(5-
methyl)furyl-substituted silanols 26a , 27a , 28a , 30a ,
Table 2, and 31a (Table 2, entries 1, 2, 7, 12 and 13) were
performed as described for the achiral aldehydes. The
yields of the protodesilylation-oxidation reactions of
these silanols were moderate in all cases (53-65%). The
oxidation of silanol 29a (Table 2, entry 8) was not
performed, though the corresponding diol 35 was subse-
quently obtained by the oxidation of menthofuryl-
substituted silanol 29b.
The menthofuryl-substituted silanols 26b, 27b, 28b,
29b, 30b, and 31b (Table 2, entries 3, 4, 9, 10, 14, and
15) proved to be excellent substrates for the one-pot
protodesilylation-oxidation sequence described previously
for achiral aldehydes 13a and 13b, and in this way diols
32-37 were obtained in 76-85% yield. Here, as with
the achiral aldehydes, the protodesilylation-oxidation
sequence is more efficient with the menthofuryl-substi-
tuted silanols than with the corresponding 2-(5-methyl-
furyl)-substituted silanols, most likely because of the
milder protodesilylation conditions.45
In summary, the data summarized in Table 2 demon-
strate that reagents 10 and 11 are both more enantiose-
lective than the first-generation [(alkoxysilyl)allyl]bor-
onate 5, as clearly demonstrated in the mismatched
double asymmetric reactions with glyceraldehyde ac-
etonide 23 (Table 2, entries 2, 4, and 6), and especially
in reactions with epoxyaldehyde 25 (Table 2, entries 7,
9, and 11). Although reagents 10 and 11 display com-
parable selectivity in most of the reactions examined, the
greater ease of preparation of 11 and the considerably
higher yields of diols obtained in the one-pot protodesi-
lylation-oxidation sequence for the menthofuryl-substi-
tuted silanols clearly define 11 as the superior reagent
for these applications.
(51) For a discussion of this point, and other examples, see: Roush,
W. R.; Straub, J . A.; VanNieuwenhze, M. S. J . Org. Chem. 1991, 56,
1636.
(52) The enantiomeric purity of 29 was determined by Mosher ester
analysis of the derived diol 35.
(53) The intrinsic diastereofacial selectivity of 25 was assessed via
reaction with pinacol (E)-[γ-(dimethylphenylsilyl)allyl]boronate (see ref
22), which provided a 73:27 mixture of diastereomers 30d :31d (R )
Ph). Therefore, diastereomer 30 is the product of matched double
asymmetric reactions with the chiral allylboronates 10 and 11.
En a n tioselective Syn th esis of Sw a in son in e. As
a demonstration of the utility of our new anti-R-hydroxy-
allylation reagents in the synthesis of polyhydroxylated
natural products, we used reagent 11 in a brief and
highly stereoselective total synthesis of the trihydroxy-

1118 J . Org. Chem., Vol. 62, No. 4, 1997
Hunt and Roush
been reported in the past decade.55,58 Most of the
approaches to swainsonine originate from carbohydrate
precursors, with many or all of the four chiral centers
required in the product already present in the starting
material; other approaches begin with R-glutamic acid,
D-tartaric acid, D-malic acid, and D-isoascorbic acid as the
chiral precursors.55,58 The first synthesis of swainsonine
from achiral precursors was published by Sharpless in
1985, but this route is fairly lengthy (20 steps).63 To our
knowledge, Zhou and co-workers’ formal synthesis of
swainsonine is the only other synthesis that begins from
achiral precursors. This synthesis provided 8-O-benzyl-
1,2-di-O-isopropylidene swainsonine in nine steps and 8%
overall yield from an achiral R-furfuryl sulfonamide
(which, in turn, was synthesized in two steps from
p-toluenesulfonamide and 2-furaldehyde dialkyl acetal).59
Our synthesis of swainsonine began with 4-bromo-
butanal (38), which is available in two steps from THF
(Scheme 1).64,65 The Horner-Wadsworth-Emmons reac-
tion of 38 with triethyl phosphonoacetate provided
exclusively the trans R,â-unsaturated ester, and DIBAL
reduction of this ester gave the known allylic alcohol 39.64
Sharpless asymmetric epoxidation66 of 39 provided epoxy
alcohol 40 with an enantiomeric purity of 92% ee (as
determined by Mosher ester analysis),46 and oxidation
of 40 with SO3-pyridine in DMSO provided epoxy
aldehyde 24.67 The R-hydroxyallylation of 24 with (S,S)-
11, as previously described, proceeded with 90:10 dia-
stereoselectivity; the major product, silanol 28b, was
obtained in 73% yield simply by recrystallization of the
mixture of diastereomers from hexanes (the minor dias-
tereomer 29b is an oil). Silanol 28b was then subjected
to our one-pot protodesilylation-oxidation protocol to give
diol 34 in 85% yield, thus completing the R-hydroxy-
allylation of aldehyde 24.
Protection of diol 34 as an acetonide and ozonolysis of
the terminal alkene provided aldehyde 41. Reductive
amination68,69 of this aldehyde was accompanied by
spontaneous ring closure to form the indolizidine nucleus
of swainsonine acetonide, which was deprotected by
acidic hydrolysis to provide (-)-swainsonine (42) [mp
indolizidine alkaloid (-)-swainsonine (42). Since its
initial isolation from fungal cultures more than two
decades ago,54 swainsonine has attracted considerable
attention because of its potent biological activity as an
inhibitor of glycoprotein-processing R-mannosidases and
because of its possible therapeutic value.55-57 Swainso-
nine, its epimers, and its analogs, are widely used in the
study of glycosidases and glycoprotein processing; ad-
ditionally, the natural product reached phase I clinical
trials as a potential anticancer drug, and various 2-O-
and 8-O-esters of are being examined as drug candidates
with improved pharmacokinetics and reduced side ef-
fects.55
Swainsonine’s combination of significant biological
properties and moderate structural complexity make it
an attractive synthetic target. More than a dozen
syntheses of swainsonine, its epimers, and its analogs
have been published in the past several years,58-63 and
many more syntheses of this family of compounds have
137-140 °C; [R]26 ) -85.5° (c 0.42, CH3OH) [lit.54 mp
D
144-145 °C; [R]25 ) -87.2° (c 2.1, CH3OH); lit.58 mp
D
141-143 °C; [R]26 ) -82.6° (c 1.03, CH3OH); lit.62 mp
D
138-140 °C; [R]27D ) -85.2° (c 0.5, CH3OH)]] in 10 steps
and 16% overall yield from 4-bromobutanal 38.
Su m m a r y. We have developed the diisopropyl tar-
trate-modified (E)-[γ-[(menthofuryl)dimethylsilyl]allyl]-
boronate 11 as an improved chiral reagent for the anti-
R-hydroxyallylation of aldehydes. Reactions of 11 with
two achiral aldehydes provided the corresponding anti-
(59) Zhou, W.-S.; Xie, W. G.; Lu, Z.-H.; Pan, X.-F. J . Chem. Soc.,
Perkin Trans. 1 1995, 2599.
(60) Kang, S. H.; Kim, G. T. Tetrahedron Lett. 1995, 36, 5049.
(61) Bennett, R. B., III; Choi, J .-R.; Montgomery, W. D.; Cha, J . K.
J . Am. Chem. Soc. 1989, 111, 2580.
(62) Miller, S. A.; Chamberlin, A. R. J . Am. Chem. Soc. 1990, 112,
8100.
(63) Adams, C. E.; Walker, F. J .; Sharpless, K. B. J . Org. Chem.
1985, 50, 420.
(64) Vedejs, E.; Arnost, M. J .; Hagen, J . P. J . Org. Chem. 1979, 44,
3230.
(54) Guengerich, F. P.; DiMari, S. J .; Broquist, H. P. J . Am. Chem.
Soc. 1973, 95, 2055.
(55) Michael, J . P. Nat. Prod. Rep. 1995, 535 and literature cited
therein.
(56) Elbein, A. D. FASEB J . 1991, 5, 3055 and references therein.
(57) Cenci di Bello, I.; Fleet, G.; Namgoong, S. K.; Tadano, K.-I.;
Winchester, B. Biochem. J . 1989, 259, 855 and references therein.
(58) Naruse, M.; Aoyagi, S.; Kibayashi, C. J . Org. Chem. 1994, 59,
1358 and references therein.
(65) Roush, W. R.; Hall, S. E. J . Am. Chem. Soc. 1981, 103, 5200.
(66) Gao, Y.; Hanson, R. M.; Klunder, J . M.; Ko, S. Y.; Masamune,
H.; Sharpless, K. B. J . Am. Chem. Soc. 1987, 109, 5765.
(67) Parikh, J . R.; von Doering, E. W. J . Am. Chem. Soc. 1967, 89,
5505.
(68) Sharma, S. K.; Songster, M. F.; Colpitts, T. L.; Hegyes, P.;
Barany, G.; Castellino, F. J . J . Org. Chem. 1993, 58, 4993.
(69) Borch, R. G.; Bernstein, M. D.; Durst, H. D. J . Am. Chem. Soc.
1971, 93, 2897.

Enantioselective Synthesis of (-)-Swainsonine
J . Org. Chem., Vol. 62, No. 4, 1997 1119
Sch em e 1
to 0 °C. Allylchlorodimethylsilane (14) (5.4 mL, 37.1 mmol)
was added to the cooled solution dropwise, and the mixture
was warmed to room temperature and stirred overnight (20
h) and then filtered through Celite and concentrated to a dark
red oil. This crude product was filtered through a plug of silica
gel (eluting with hexanes) and then distilled under vacuum
(5 mmHg, bp 45 °C) to yield 4.91 g (73%) of 15 as a colorless
1,2-diols 9a and 9b in good yields (64-70% over two
steps) and with respectable enantioselectivities (74-81%
ee). Double asymmetric reactions of both enantiomers
of reagent 11 with each of three chiral aldehydes (23-
25) provided the corresponding anti-1,2-diols 32-37 in
good yields (60-79% over two steps) and with respectable
diastereoselectivities (80-90% de, with the exception of
the mismatched reaction of 25 with (R,R)-11, which was
essentially unselective). Reagent 11 is thus a clear
improvement over the two anti-R-hydroxyallylation re-
agents developed previously in our laboratory: reagent
5, which showed poor enantioselectivity, and reagent 6,
which was incompatible with the protodesilylation-
oxidation sequence.20-22
1
liquid: Rf ) 0.82 (5:1 hexanes-CH2Cl2); H NMR (400 MHz,
CDCl3) δ 6.54 (d, J ) 3.0 Hz, 1 H), 5.96 (dq, J ) 0.8, 3.0 Hz,
1 H), 5.78 (ddd, J ) 8.1, 10.2, 16.1 Hz, 1 H), 4.91-4.84 (m, 2
H), 2.33 (d, 0.6 Hz, 3 H), 1.75 (dt, J ) 1.2, 8.1 Hz, 2 H), 0.25
(s, 6 H); 13C NMR (100 MHz, CDCl3) δ 156.6, 134.3, 121.4,
113.5, 105.7, 26.4, 23.1, 13.7, 0.1, -3.7; IR (neat) 3060, 2955,
2920, 1625, 1590, 1490 cm-1; HRMS calcd for C10H16OSi
180.0960, found 180.0971.
(E)-[γ-[[2′-(5′-Meth ylfu r yl)]dim eth ylsilyl]allyl]bor on ate-
Dieth a n ola m in e Com p lex (17). A 1.0 M solution of KO-t-
Bu in THF (16.6 mL, 16.6 mmol) was cooled to -78 °C, and
allylsilane 15 (3.0 g, 16.6 mmol) in 5 mL of THF was added.
n-BuLi (6.6 mL of a 2.5 M solution, 16.6 mmol) was added
dropwise to this mixture at a rate such that the internal
temperature of the solution did not rise above -40 °C. The
mixture was allowed to warm to -40 °C, stirred at this
temperature for 15 min, and then cooled to -78 °C. Triiso-
propyl borate (3.8 mL, 16.6 mmol) was added dropwise, and
the mixture was stirred for 15 min at -78 °C and then poured
into a separatory funnel containing 20 mL of a saturated NH4-
Cl solution. The phases were separated and the aqueous phase
was extracted with 4 × 10 mL of Et2O. The combined organics
were treated with MgSO4, and a 3 M solution of diethanol-
amine (0.52 g, 5.0 mmol)71 in i-PrOH was added. The mixture
was stirred overnight and then filtered and concentrated to a
dark yellow oil, which was dissolved in toluene and concen-
trated to a dark yellow solid. This solid was dissolved in hot
benzene and filtered and then recrystallized from benzene-
hexanes to yield 1.11 g of 17 (23% based on 15) as a pale yellow
solid: mp 130-135 °C; 1H NMR (400 MHz, CDCl3) δ 6.82 (dt,
J ) 7.8, 18.5 Hz, 1 H), 6.63 (d, J ) 3.0 Hz, 1 H), 5.90 (dd, J )
3.0, 0.8 Hz, 1 H), 5.74 (dt, J ) 1.1, 17.5 Hz, 1 H), 4.66 (br s, 1
H), 3.67 (m, 2 H) 3.58 (m, 2 H), 2.55 (m, 2 H), 2.07 (d, J ) 0.5
Hz, 3 H), 1.95 (d, J ) 7.8 Hz, 2 H), 1.84 (m, 2 H), 0.41 (s, 6 H);
13C NMR (100 MHz, CDCl3) δ 158.7, 156.3, 155.2, 121.7, 106.4,
62.8, 51.3, 13.7, -2.2; IR (CH2Cl2) 3320-3280, 3160-3100,
3040, 2960, 2860, 2920, 2860, 1600 cm-1; HRMS calcd for
We have also synthesized a related anti-R-hydroxy-
allylation reagent, tartrate ester-modified (E)-[γ-[2-(5-
methylfuryl)dimethylsilyl]allyl]boronate 10. This re-
agent was slightly more stereoselective than reagent 11
in some, though not all, of the double asymmetric
reactions examined in this study (see Table 2). However,
we were unable to synthesize this reagent in good yield
(e15%), and the overall yields for the two-step R-hy-
droxyallylation sequence were consistently lower with
reagent 10 than were the analogous reactions with
reagent 11, owing to the more vigorous conditions
required to promote the protodesilylation of the 2-(5-
methylfuryl)silane unit.45 On this basis we concluded
that reagent 11 is the more efficient of our two new
reagents for the anti-R-hydroxyallylation of aldehydes.
As a demonstration of the application of reagent 11 to
the synthesis of polyhydroxylated natural products from
achiral precursors, we used our new reagent in a brief,
enantioselective synthesis of the trihydroxyindolizidine
alkaloid (-)-swainsonine (42).
Exp er im en ta l Section 70
Allyld im eth yl[2-(5-m eth ylfu r yl)]sila n e (15). 2-Methyl-
furan (3.05 g, 37.1 mmol) was added dropwise to BuLi (14.8
mL of a 2.5 M solution in hexanes, 37.1 mmol) in 15 mL of
THF. The mixture was heated to 50 °C for 1 h and then cooled
C
14H2510BO3Si 293.1726, found 293.1732.
(70) For general experimental details, see: Roush, W. R.; Hunt, J .
A. J . Org. Chem. 1995, 60, 798. Unless otherwise noted, all compounds
purified by chromatography are sufficiently pure (>95% by 1H NMR
analysis) for use in subsequent reactions
(71) Diethanolamine complex 17 would not crystallize if larger
quantities of diethanolamine were used. The stoichiometry (0.3 equiv
of diethanolamine) here is consistent with the efficiency of the
preparation of allylboronate 10 directly from allylsilane 15.

1120 J . Org. Chem., Vol. 62, No. 4, 1997
Hunt and Roush
Diisop r op yl Ta r tr a te-Mod ified (E)-[γ-[[2′-(5′-Meth yl-
fu r yl)]d im eth ylsilyl]a llyl]bor on a te [(R,R)-10 a n d (S,S)-
10]. A two-phase mixture of diethanolamine complex 17 (1.0
g, 3.41 mmol) and L-(R,R)-DIPT (0.73 mL, 3.41 mmol) in 20
mL of brine and 10 mL of EtOAc was stirred vigorously for 1
h. The mixture was poured into a separatory funnel contain-
ing 20 mL of EtOAc, 20 mL of brine, and 2 mL of 1 M NaHSO4
and shaken well. The phases were separated, and the aqueous
phase was extracted with 4 × 20 mL of EtOAc. The combined
organics were stirred over MgSO4 overnight (20 h), filtered,
concentrated, and then pumped to a constant weight to yield
1.25 g of a light yellow oil. This crude product consisted of
(R,R)-10 contaminated with ca. 20% of DIPT, as determined
by 1H NMR analysis; the yield of (R,R)-10 was calculated to
be 64%. Crude (R,R)-10 was dissolved in toluene (ca. 0.5 M)
and stored at -20 °C over crushed 4 Å molecular sieves: 1H
NMR (400 MHz, CDCl3) δ 6.52 (d, J ) 3.5, 1 H), 6.21 (dt, J )
7.3, 18.3 Hz, 1 H), 5.94 (m, 1 H), 5.76 (dt, J ) 1.6, 18.3 Hz, 1
H), 5.20-5.08 (m, 2.5 H), 4.78 (s, 2 H), 2.32 (d, J ) 0.8, 3 H),
2.08 (dd, J ) 1.6, 7.3 Hz, 2 H), 1.32-1.28 (m, 13 H), 0.30 (s, 6
H).
Rep r esen ta tive P r oced u r e for Rea ction s of Rea gen t
10 a n d Ald eh yd es: 1-Cycloh exyl-2-[2′-(5′-m eth ylfu r yl)]-
d im eth ylsilyl]bu t-3-en -1-ol (12a ). A solution of freshly
distilled cyclohexanecarboxaldehyde (63 µL, 0.52 mmol) in 0.5
mL of toluene was stirred over crushed 4 Å molecular sieves
for 15 min and then cooled to -78 °C and added dropwise over
5 min to a -78 °C solution of (S,S)-10 in toluene (1.25 mL of
a 0.5 M solution, 0.62 mmol). The resulting mixture was
stirred at -78 °C for 4 h, at which point TLC analysis showed
that no starting material remained. The reaction mixture was
added directly to a silica gel column and eluted with 9:1
hexanes-Et2O to yield 126 mg (83%) of 12a as a colorless oil:
Rf ) 0.51 (6:1 hexanes-Et2O); [R]23 ) 15.5° (c 0.44, CHCl3);
D
1H NMR (400 MHz, CDCl3) δ 6.56 (d, J ) 3.2 Hz, 1 H), 5.97
(dq, J ) 1.1, 2.9 Hz, 1 H), 5.85 (ddd, J ) 10.5, 10.5, 17.2 Hz,
1 H), 5.03 (dd, J ) 2.2, 10.2 Hz, 1 H), 4.89 (ddd, J ) 0.8, 2.2,
17.2 Hz, 1 H), 3.43 (dt, J ) 4.0, 7.8 Hz, 1 H), 2.33 (d, J ) 1.1
Hz, 3 H), 2.12 (dd, J ) 4.0, 10.5, 1 H), 1.91-1.84 (m, 1 H),
1.78 (d, J ) 4.3 Hz, 1 H), 1.74-1.58 (m, 3 H), 1.42-0.82 (m, 7
H), 0.29 (s, 3 H), 0.28 (s, 3 H); 13C NMR (100 MHz, CDCl3) δ
156.8, 156.4, 134.8, 122.2, 115.1, 105.9, 75.8, 41.9, 39.0, 29.2,
28.6, 26.4, 26.2, 25.9, 13.7, -3.8, -4.2; IR (neat) 3600-3450,
The same procedure was followed for the synthesis of (S,S)-
10 from diethanolamine complex 16 and D-(S,S)-DIPT.
Allyld im et h yl[2-(4,5,6,7-t et r a h yd r o-3,6-d im et h ylb en -
zofu r yl)]sila n e (22). 2-4,5,6,7-Tetrahydro-3,6-dimethylben-
zofuran [21 (menthofuran), 10.0 g, 66.6 mmol] in 10 mL of THF
was added dropwise to a solution of n-BuLi (34 mL of a 1.96
M solution in hexanes, 66.6 mmol) in 30 mL of THF. The
mixture was heated to 50 °C for 1 h then cooled to 0 °C.
Allylchlorodimethylsilane (14) (4.9 mL, 4.5 mmol) was added
to the cooled solution dropwise, and the mixture was warmed
to room temperature and stirred overnight (20 h) and then
filtered through Celite and concentrated to a dark yellow oil.
This oil was filtered through a plug of silica gel (eluting with
hexanes) and then distilled under vacuum (5 mmHg, bp 136-
139 °C) to yield 11.0 g (67%) of 22 as a colorless oil: Rf ) 0.90
3060, 2920, 2830, 1625, 1590 cm-1; HRMS calcd for C17H27
-
OSi (M - OH) 275.1824, found 275.1818. Anal. Calcd for
C17H28O2Si: C, 69.81; H, 9.65. Found: C, 69.71; H, 9.86.
3-[[2′-(5′-Meth ylfu r yl)]dim eth ylsilyl]n on -1-en -4-ol (12b).
12b was prepared in 77% yield from freshly distilled hexanal
and (S,S)-10: Rf ) 0.50 (5:1 hexanes-Et2O); [R]23 ) 4.8° (c
D
0.49, CHCl3); 1H NMR (400 MHz, C6D6) δ 6.61 (d, J ) 3.2 Hz,
1 H), 5.96 (ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.87 (dq, J )
0.8, 3.0 Hz, 1 H), 5.02 (dd, J ) 2.4, 10.5 Hz, 1 H), 4.91 (ddd J
) 0.8, 2.4, 17.2 Hz, 1 H), 3.78 (m, 1 H), 2.07 (s, 3 H), 1.94 (dd,
J ) 3.5, 10.5, 1 H), 1.46-1.16 (m, 9 H), 0.86 (m, 3 H), 0.41 (s,
3 H), 0.38 (s, 3 H); 13C NMR (100 MHz, C6D6) δ 158.8, 135.4,
122.7, 115.3, 106.4, 71.5, 41.9, 37.5, 32.2, 25.9, 23.0, 14.3, 13.6,
-3.4, -3.8; IR (neat) 3550-3400, 2960, 2920, 2850, 1625, 1590
cm-1; HRMS calcd for C16H27OSi (M - OH), 263.1824, found
263.1823. Anal. Calcd for C16H28O2Si: C, 68.51; H, 10.19.
Found: C, 68.53; H, 10.19.
1
(100% hexanes); H NMR (400 MHz, CDCl3) δ 5.81 (m, 1 H),
4.82 (m, 2 H), 2.67 (br dd, J ) 5.4, 16.4 Hz, 1 H), 2.41-2.25
(m, 2 H), 2.18 (qt, J ) 2.2, 9.4 Hz, 1 H), 1.99 (s, 3 H), 1.95-
1.74 (m, 4 H), 1.40-1.28 (m, 2 H), 1.07 (d, J ) 6.8 Hz, 3 H),
0.27 (s, 6 H); 13C NMR (100 MHz, CDCl3) δ 154.7, 150.4, 134.7,
130.8, 117.9, 31.9, 31.4, 29.6, 23.6, 21.5, 20.0, 9.3, -3.2; IR
(neat) 3070, 2950, 2920, 2840, 1630 cm-1; HRMS calcd for
C15H25OSi (M + 1) 249.1668, found 249.1667.
(3R,4S,5R)-5,6-O-Isopr opyliden e-3-[[2′-(5′-m eth ylfu r yl)]-
d im eth ylsilyl]h ex-1-en -4-ol (26a ). 26a was prepared in 93%
yield (86% major diastereomer 26a , 7% minor diastereomer
27a ) from freshly distilled glyceraldehyde acetonide 23 and
(S,S)-10. The reaction stereoselectivity was determined to be
95:5 by 1H NMR analysis of the reaction mixture after
filtration of the mixture through a silica gel plug, eluting with
Diisop r op yl Ta r tr a te-Mod ified (E)-[γ-[[2′-(4′,5′,6′,7′-Tet-
r a h yd r o-3′,6′-d im et h ylb en zofu r yl)]d im et h ylsilyl]a llyl]-
bor on a te [(S,S)-11 a n d (R,R)-11]. KO-t-Bu (2.26 g, 20.1
mmol) was weighed out under an atmosphere of dry N2 and
dissolved in 30 mL of THF; the solution was then cooled to
-78 °C. Allylsilane 22 (5.00 g, 20.1 mmol) in 10 mL of THF
was added dropwise. n-BuLi (10.7 mL of a 1.88 M solution in
hexanes, 20.1 mmol) was added dropwise to this mixture at a
rate such that the internal temperature of the solution did
not rise above -25 °C. The mixture was transferred to -25
°C bath, stirred at this temperature for 4 h, and then recooled
to -78 °C. Triisopropyl borate (4.6 mL, 20.1 mmol) was added
dropwise, and the mixture was stirred for 15 min at -78 °C
and then poured into a separatory funnel containing 30 mL
of a saturated NH4Cl solution and D-(S,S)-DIPT (4.3 mL, 20.1
mmol) in 10 mL of Et2O. The phases were separated, and the
aqueous phase was extracted with 4 × 15 mL of Et2O. The
combined organics were stirred over MgSO4 overnight (20 h),
filtered, and concentrated in vacuo to a constant weight of 9.30
g of a colorless oil. This crude product consisted of (S,S)-11
contaminated with ca. 20% of DIPT and ca. 10% of allylsilane
5:1 hexanes-Et2
O, to remove excess allylboronate reagent.
Data for 26a follow: Rf ) 0.25 (3:1 hexanes-Et2O); [R]26
)
D
1
1.5° (c 1.81, C6H6); H NMR (400 MHz, C6D6) δ 6.65 (d, J )
3.0 Hz, 1 H), 6.12 (ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.87 (dq,
J ) 0.8, 3.0 Hz, 1 H), 4.91 (dd, J ) 1.9, 10.2 Hz, 1 H), 4.75
(dd, J ) 2.2, 17.2 Hz, 1 H), 4.12 (dt, J ) 6.4, 8.1 Hz, 1 H),
3.78-3.73 (m, 2 H), 3.36 (dd, J ) 6.7, 8.1 Hz, 1 H), 2.36 (t, J
) 2.2 Hz, 1 H), 2.08 (d, J ) 0.8 Hz, 3 H), 1.79 (dt, J ) 2.2,
10.8 Hz, 1 H), 1.29 (s, 3 H), 1.22 (s, 3 H), 0.49 (s, 3 H), 0.43 (s,
3 H);
13C NMR (100 MHz, C6D6) δ 156.6, 156.5, 134.6, 128.5,
122.8, 114.8, 109.5, 106.4, 79.7, 73.8, 65.9, 38.2, 27.0, 25.7, 13.6,
-3.7; IR (neat) 3520-3460, 1625, 1595 cm-1; HRMS calcd for
C16H26O4Si 310.1593, found 310.1607. Anal. Calcd for
C16H26O4Si: C, 61.90; H, 8.44. Found: C, 61.82; H, 8.44.
(3S,4R,5R)-5,6-O-Isopr opyliden e-3-[[2′-(5′-m eth ylfu r yl)]-
d im eth ylsilyl]h ex-1-en -4-ol (27a ). 27a was prepared in 90%
yield (83% major diastereomer 27a , 7% minor diastereomer
26a ) from freshly distilled glyceraldehyde acetonide 23 and
(R,R)-10. The reaction stereoselectivity was determined to be
94:6 by 1H NMR analysis of the reaction mixture after
filtration of the mixture through a silica gel plug, eluting with
5:1 hexanes-Et2O, to remove excess allylboronate reagent.
1
22, as determined by H NMR analysis; the yield of (S,S)-11
was calculated to be 67%. Crude (S,S)-11 was dissolved in
toluene (ca. 0.5 M) and stored at -20 °C over crushed 4 Å
molecular sieves: 1H NMR (400 MHz, CDCl3) δ 6.19 (dt, J )
7.0, 18.3 Hz, 1 H), 5.79 (dt, J ) 1.6, 18.3 Hz, 1 H), 5.11 (m, 2
H), 4.78 (s, 2 H), 2.67 (br dd, J ) 4.3, 15.3 Hz, 1 H), 2.40-2.12
(m, 3 H), 2.05 (dd, J ) 1.4, 7.3 Hz, 1 H), 1.99-1.76 (m, 6 H),
1.30 (d, J ) 6.4 Hz, 12 H), 1.06 (d, J ) 6.4 Hz, 3 H), 0.31 (s,
6 H).
Data for 27a follow: Rf ) 0.13 (3:1 hexanes-Et2O); [R]25
)
D
1
5.7° (c 1.08, C6H6); H NMR (400 MHz, C6D6) δ 6.58 (d, J )
3.0 Hz, 1 H), 5.96 (ddd, J ) 10.5, 10.5, 16.9 Hz, 1 H), 5.84 (dq,
J ) 0.8, 3.0 Hz, 1 H), 4.98-4.89 (m, 2 H), 4.03 (m, 1 H), 3.95-
3.85 (m, 3 H), 2.17 (dd, J ) 2.0, 10.8 Hz, 1 H), 2.04 (d, J ) 0.8
Hz, 3 H), 1.80 (d, J ) 3.6 Hz, 1 H), 1.33 (s, 3 H), 1.27 (s, 3 H),
0.37 (s, 3 H), 0.33 (s, 3 H); 13C NMR (100 MHz, C6D6) δ 157.0,
156.3, 134.6, 128.5, 122.9, 115.8, 108.3, 106.4, 78.9, 71.8, 66.0,
The same procedure was followed for the synthesis of (R,R)-
11 from allylsilane 22, triisopropyl borate, and L-(R,R)-DIPT.

Enantioselective Synthesis of (-)-Swainsonine
J . Org. Chem., Vol. 62, No. 4, 1997 1121
38.4, 26.9, 25.6, 13.5, -3.8, -4.1; IR (neat) 3550-3450, 1625,
1595 cm-1; HRMS calcd for C15H23O4Si (M - CH3) 295.1343,
found 295.1354. Anal. Calcd for C16H26O4Si: C, 61.90; H, 8.44.
Found: C, 61.73; H, 8.50.
mixture through a silica gel plug, eluting with 20:1 hexanes-
Et2O, to remove excess allylboronate reagent. Data for 31a
follow: Rf ) 0.55 (9:1 hexanes-Et2O); [R]26 ) -21.9° (c 1.24,
D
1
C6H6); H NMR (400 MHz, C6D6) δ 6.73 (d, J ) 3.2 Hz, 1 H),
6.30 (ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.91 (dq, J ) 0.8, 3.0
Hz, 1 H), 5.07 (dd, J ) 2.4, 10.5 Hz, 1 H), 4.96 (dd, J ) 2.4,
17.5 Hz, 1 H), 4.12 (t, J ) 1.9 Hz, 1 H), 3.86 (br d, J ) 9.1 Hz,
2 H), 3.56 (dd, J ) 4.0, 9.9 Hz, 1 H), 3.38 (t, J ) 9.1 Hz, 1 H),
2.21 (dt, J ) 2.2, 10.5 Hz, 1 H), 2.10 (s, 3 H), 1.98-1.87 (m, 1
H), 0.87 (s, 9 H), 0.61 (s, 3 H), 0.58 (d, J ) 7.2 Hz, 3 H), 0.54
(s, 3 H), 0.05 (s, 6 H); 13C NMR (100 MHz, CDCl3) δ 157.1,
156.3, 134.9, 121.8, 114.2, 105.7, 69.3, 39.1, 38.9, 25.9, 18.2,
13.7, 12.8, -3.9, -5.5, -5.6; IR (neat) 3500 (br), 1625, 1595
cm-1; HRMS calcd for C20H37O2Si2 (M - OH), 365.2322, found
365.2343. Anal. Calcd for C20H38O3Si2: C, 62.77; H, 10.01.
Found: C, 62.50; H, 10.22.
(3R,4S,5S,6R)-9-br om o-3-[[2′-(5′-m eth ylfu r yl)]d im eth -
ylsilyl]-5,6-ep oxyn on -1-en -4-ol (28a ). 28a was prepared in
69% yield (62% major diastereomer 28a , 7% minor diastere-
omer 29a ) from epoxy aldehyde 24 and (S,S)-10. The reaction
1
stereoselectivity was determined to be 90:10 by H NMR and
HPLC analysis of the reaction mixture after filtration of the
mixture through a silica gel plug, eluting with 5:1 hexanes-
Et2O, to remove excess allylboronate reagent. Data for 28a
follow: HPLC tR 9.70 min (10 mm column, 4 mL/min, 20%
hexanes-EtOAc); Rf ) 0.39 (1:1 hexanes-Et2O); [R]26D ) -5.5°
(c 0.65, CHCl3); 1H NMR (400 MHz, CDCl3) δ 6.59 (d, J ) 3.0
Hz, 1 H), 5.97 (dq, J ) 3.0, 0.8 Hz, 1 H), 5.87 (ddd, J ) 10.2,
10.2, 17.2 Hz, 1 H), 5.02 (dd, J ) 2.14, 10.2 Hz, 1 H), 4.96
(ddd, J ) 0.8, 2.2, 17.2 Hz, 1 H), 3.99 (dd, J ) 3.2, 5.9 Hz, 1
H), 3.43 (m, 2 H) 3.03 (ddd, J ) 2.4, 5.1, 7.0 Hz, 1 H), 2.72 (t,
J ) 2.4 Hz, 1 H), 2.33 (d, J ) 1.0 Hz, 3 H), 2.07 (dd, J ) 3.5,
10.5, 1 H), 2.02-1.92 (m, 3 H), 1.76-1.68 (m, 1 H), 1.62-1.52
(m, 2 H) 0.30 (s, 6 H); 13C NMR (100 MHz, CDCl3) δ 156.9,
155.7, 134.2, 122.5, 115.4, 105.9, 68.4, 60.8, 54.6, 38.8, 33.2,
30.2, 29.1, 13.7, -4.0, -4.4; IR (neat) 3540-3380, 1620, 1590
cm-1; HRMS calcd for C16H26O3Si81Br (M + 1) 375.0807, found
375.0826. Anal. Calcd for C16H25O3SiBr: C, 51.47; H, 6.75.
Found: C, 51.71; H, 7.00.
Rep r esen ta tive P r oced u r e for Rea ction s of Rea gen t
11 an d Aldeh ydes: (3R,4S,5S,6R)-9-Br om o-3-[[2′-(4′,5′,6′,7′-
tetr a h yd r o-3′,6′-d im eth ylben zofu r yl)]d im eth ylsilyl]-5,6-
ep oxyn on -1-en -4-ol (28b). Aldehyde 24 (350 mg, 1.81 mmol)
was dissolved in 2 mL of toluene, stirred over 100 mg of
crushed 4 Å molecular sieves for 15 min at room temperature,
and then cooled to -78 °C. A 0.78 M solution of allylboronate
(S,S)-11 in toluene (3.5 mL, 2.72 mmol) was stirred over 200
mg of crushed 4 Å molecular sieves for 15 min at room
temperature and then cooled to -78 °C. The aldehyde solution
was added to the solution of reagent 11 dropwise via cannula,
and the resulting mixture was stirred at -78 °C for 4 h, at
which point TLC analysis showed that no starting material
remained. A 1 M solution of NaBH4 in EtOH (3 mL) was
cooled to -78 °C and added dropwise via cannula to the
reaction mixture. The mixture was allowed to warm to room
temperature and then partitioned between 30 mL of Et2O and
10 mL of pH 7 buffer saturated with NaCl. The aqueous phase
was extracted with 3 × 15 mL of Et2O, and the combined
organics were dried over MgSO4, filtered, and concentrated
to a colorless oil. The reaction stereoselectivity was deter-
mined to be 90:10 (28b:29b) by 1H NMR analysis of the
reaction mixture after filtration of the mixture through a silica
gel plug, eluting with 5:1 hexanes-Et2O. The major diaster-
omer 28b was separated from the minor isomer 29b by
recrystallization from hexanes to yield 583 mg (73%) of 28b
as a white solid; concentration of the mother liquor provided
65 mg (8%) of 29b as a colorless oil (81% total). Data for 28b
(3S,4R,5S,6R)-9-br om o-3-[[2′-(5′-m eth ylfu r yl)]d im eth -
ylsilyl]-5,6-ep oxyn on -1-en -4-ol (29a ). 29a was prepared in
68% yield (63% major diastereomer 29a , 5% minor diastere-
omer 28a ) from epoxy aldehyde 24 and (R,R)-10. The reaction
1
stereoselectivity was determined to be 90:10 by H NMR and
HPLC analysis of the reaction mixture after filtration of the
mixture through a silica gel plug, eluting with 5:1 hexanes-
Et2O, to remove excess allylboronate reagent. Data for 29a
follow: HPLC tR 12.21 min (10 mm column, 4 mL/min, 20%
hexanes-EtOAc); Rf ) 0.30 (1:1 hexanes-Et2O); [R]26D ) 21.1°
(c 1.39, CHCl3); 1H NMR (400 MHz, CDCl3) δ 6.58 (d, J ) 3.2
Hz, 1 H), 6.00-5.90 (m, 2 H), 5.07 (dd, J ) 2.2, 10.5 Hz, 1 H),
4.97 (dd, J ) 1.9, 16.9 Hz, 1 H), 3.61 (dd, J ) 4.8, 9.4 Hz, 1
H), 3.42 (m, 2 H), 2.83 (m, 2 H), 2.32 (d, J ) 0.5 Hz, 3 H),
2.09-1.88 (m, 4 H), 1.61-1.51 (m, 1 H), 0.35 (s, 3 H), 0.30 (s,
3 H); 13C NMR (100 MHz, CDCl3) δ 156.9, 155.6, 134.2, 122.5,
115.7, 106.0, 71.4, 61.4, 55.9, 40.5, 33.0, 30.1, 29.1, 13.7, -3.8,
-4.2; IR (neat) 3500-3300, 2960, 2860, 2820 cm-1; HRMS
calcd for C16H26O3Si81Br (M + 1) 375.0807, found 375.0826.
Anal. Calcd for C16H25O3SiBr: C, 51.47; H, 6.75. Found: C,
51.74; H, 6.68.
follow: Rf ) 0.69 (1:1 hexanes-Et2O); mp 114-116 °C; [R]26
D
) 31.9° (c 0.26, CHCl3); 1H NMR (400 MHz, CDCl3) δ 5.88 (ddd,
J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.03 (dd, J ) 1.9, 10.2 Hz, 1 H),
4.98 (dd, J ) 1.3, 17.2 Hz, 1 H), 3.97 (dd, J ) 3.2, 5.6 Hz, 1
H), 3.43 (m, 2 H), 3.02 (ddd, J ) 2.2, 4.6, 7.0 Hz, 1 H), 2.71-
2.64 (m, 2 H), 2.41-1.65 (m, 13 H), 1.58-1.48 (m, 1 H), 1.40-
1.28 (m, 1 H), 1.07 (d, J ) 6.4 Hz, 3 H), 0.36 (s, 3 H), 0.32 (s,
3 H); 13C NMR (100 MHz, C6D6) δ 155.0, 150.0, 135.1, 131.9,
118.3, 115.0, 68.8, 60.9, 54.4, 39.0, 33.2, 32.2, 31.6, 30.4, 29.8,
29.4, 21.4, 20.2, 9.6, -3.1, -3.3; IR (CHCl3) 2980, 2925, 2900,
1615 cm-1; HRMS calcd for C21H33O3Si79Br 440.1373, found
440.1362. Anal. Calcd for C21H33O3SiBr: C, 57.13; H, 7.53.
Found: C, 56.46; H, 7.67.
(3R,4S,5R)-6-[(ter t-Bu t yld im et h ylsilyl)oxy]-3-[[2′-(5′-
m eth ylfu r yl)]d im eth yl silyl]-5-m eth ylh ex-1-en -4-ol (30a ).
30a was prepared in 67% yield (64% major diastereomer 30a ,
3% minor diastereomer 31a ) from 25 and (S,S)-10. The
reaction stereoselectivity was determined to be >96:4 by 1H
NMR analysis of the reaction mixture after filtration of the
mixture through a silica gel plug, eluting with 19:1 hexanes-
Et2O, to remove excess allylboronate reagent. Data for 30a
follow: Rf ) 0.36 (9:1 hexanes-Et2O); [R]25 ) -4.7° (c 0.36,
D
1
C6H6); H NMR (400 MHz, C6D6) δ 6.62 (d, J ) 3.0 Hz, 1 H),
(3S,4R,5S,6R)-9-Br om o-3-[[2′-(4′,5′,6′,7′-tetr a h yd r o-3′,6′-
d im eth yl ben zofu r yl)]d im eth ylsilyl]-5,6-ep oxyn on -1-en -
4-ol (29b). 29b was prepared in 82% yield (79% major
diastereomer 29b, 5% minor diastereomer 28b) from epoxy
aldehyde 24 and (R,R)-11. The reaction stereoselectivity was
determined to be 96:4 by 1H NMR analysis of the crude
reaction mixture after filtration of the mixture through a silica
gel plug, eluting with 5:1 hexanes-Et2O, to remove excess
5.96 (ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.86 (dq, J ) 0.8, 3.0
Hz, 1 H), 4.99 (dd, J ) 2.2, 10.5 Hz, 1 H), 4.92 (ddd, J ) 0.8,
2.2, 17.2 Hz, 1 H), 4.01 (ddd, J ) 4.3, 4.6, 3.8 Hz, 1 H), 3.52
(dq, J ) 5.1, 9.9 Hz, 2 H), 2.21 (dd, J ) 4.8, 10.5 Hz, 1 H),
2.08 (s, 3 H), 1.97 (d, J ) 4.3 Hz, 1 H), 1.86-1.77 (m, 1 H),
0.98 (d, J ) 7.0 Hz, 3 H), 0.95 (s, 9 H), 0.43 (s, 3 H), 0.39 (s, 3
H), 0.03 (s, 6 H); 13C NMR (100 MHz, CDCl3) δ 156.7, 156.1,
136.4, 122.2, 114.7, 105.8, 73.0, 67.1, 40.0, 39.4, 25.9, 18.2, 13.7,
allylboronate reagent. Data for 29b follow: Rf ) 0.63 (1:1
11.2, -3.5, -4.4, -5.6; IR (neat) 3600-3480, 1650, 1590 cm-1
;
1
hexanes-Et2O); [R]27 ) 50.3° (c 0.61, CHCl3); H NMR (400
D
HRMS calcd for C19H35O3Si2 (M - CH3), 367.2115, found
367.2141. Anal. Calcd for C20H38O3Si2: C, 62.77; H, 10.01.
Found: C, 63.24; H, 10.26.
MHz, CDCl3) δ 5.96 (ddd, J ) 10.2, 10.2, 17.2 Hz, 1 H), 5.07
(dd, J ) 1.9, 10.2 Hz, 1 H), 4.99 (ddd, J ) 0.5, 1.9, 17.4 Hz, 1
H), 3.60 (dd, J ) 4.6, 9.7 Hz, 1 H), 3.42 (m, 2 H), 2.82 (m, 2
H), 2.65 (br dd, J ) 5.1, 16.1 Hz, 1 H), 2.40-1.69 (m, 13 H),
1.60-1.48 (m, 1 H), 1.39-1.27 (m, 1 H), 1.07 (d, J ) 6.8 Hz, 3
H), 0.35 (s, 3 H), 0.31 (s, 3 H); 13C NMR (100 MHz, CDCl3) δ
155.0, 149.0, 134.6, 132.0, 118.2, 115.5, 71.6, 61.5, 56.0, 40.8,
33.0, 31.9, 31.3, 30.0, 29.6, 29.1, 21.5, 19.9, 9.4, -3.1, -3.7; IR
(neat) 3580-3300, 2960, 2920, 1660 cm-1; HRMS calcd for
(3S,4R,5R)-6-[(ter t-Bu tyld im eth ylsilyl)oxy]-3-[2′-(5′-m e-
t h ylfu r yl)]d im et h yl silyl]-5-m et h ylh ex-1-en -4-ol (31a ).
31a was prepared in 58% yield (40% major diastereomer 31a ,
18% minor diastereomer 30a ) from 25 and (R,R)-10. The
reaction stereoselectivity was determined to be 67:33 by 1H
NMR analysis of the reaction mixture after filtration of the

1122 J . Org. Chem., Vol. 62, No. 4, 1997
Hunt and Roush
C21H33O3Si81Br 42.1373, found 442.1367. Anal. Calcd for
C21H33O3SiBr: C, 57.13; H, 7.53. Found: C, 56.37; H, 7.73.
1-Cycloh exyl-2-[[2′-(4′,5′,6′,7′-tetr a h yd r o-3′,6′-d im eth yl-
ben zofu r yl)] d im eth ylsilyl]bu t-3-en -1-ol (13a ). 13a was
prepared in 87% yield from freshly distilled cyclohexanecar-
determined to be 94:6 by 1H NMR analysis of the reaction
mixture after filtration of the mixture through a silica gel plug,
eluting with 9:1 hexanes-Et2O, to remove excess allylboronate
reagent. Data for 30b follow: Rf ) 0.75 (9:1 hexanes-Et2O);
[R]26 ) 28.2° (c 1.19, C6H6); 1H NMR (400 MHz, C6D6) δ 6.10
D
(ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.02 (dd, J ) 2.2, 10.2 Hz,
1 H), 4.98 (dd, J ) 2.2, 17.2 Hz, 1 H), 4.07 (dd, J ) 4.8, 9.1
Hz, 1 H), 3.57 (m, 2 H), 2.62 (br dd, J ) 5.2, 16.0 Hz, 1 H),
2.32-2.06 (m, 5 H), 2.03 (s, 3 H), 1.85 (m, 1 H), 1.72-1.53 (m,
2 H), 1.20-1.08 (m, 1 H), 1.01 (d, J ) 6.4 Hz, 3 H), 0.96 (s, 9
H), 0.83 (d, J ) 6.4 Hz, 3 H), 0.53 (s, 3 H), 0.49 (s, 3 H), 0.04
(s, 6 H); 13C NMR (100 MHz, C6D6) δ 154.9, 150.2, 137.0, 131.6,
118.2, 114.5, 72.3, 66.6, 41.0, 40.8, 32.2, 31.6, 29.8, 26.1, 21.4,
20.2, 18.4, 12.3, 9.6, -2.7, -3.4, -5.4; IR (neat) 3600-3480,
2960, 2930, 2860, 1625 cm-1; HRMS calcd for C25H47O3Si2 (M
+ 1), 451.3051, found 451.3052. Anal. Calcd for C25H46O3Si2:
C, 66.60; H, 10.29. Found: C, 66.76; H, 10.43.
boxaldehyde and (S,S)-11: Rf ) 0.50 (5:1 hexanes-Et2O);
1
[R]27 ) 0.8° (c 1.03, C6H6); H NMR (400 MHz, C6D6) δ 6.05
D
(ddd, J ) 10.2, 10.2, 17.2 Hz, 1 H), 5.04 (dd, J ) 2.2, 10.2 Hz,
1 H), 4.95 (dd, J ) 2.2, 17.2 Hz, 1 H), 3.52 (m, 1 H), 2.30-1.84
(m, 7 H), 1.22-1.00 (m, 4 H), 0.92-0.80 (m, 5 H), 0.52 (s, 3
H), 0.48 (s, 3 H); 13C NMR (100 MHz, C6D6) δ 154.9, 150.4,
135.8, 131.5, 118.3, 114.8, 76.0, 42.5, 39.7, 32.2, 31.6, 29.8, 29.4,
29.1, 26.8, 26.6, 26.3, 21.4, 20.2, 9.6, -2.8, -3.3; IR (neat)
3600-3500, 2930, 2860, 1660 cm-1; HRMS calcd for C22H36O2-
Si 360.2475, found 360.2468.
3-[2′-(4′,5′,6′,7′-Tet r a h yd r o-3′,6′-d im et h ylb en zofu r yl)-
d im eth yl]silyln on -1-en e-4-ol (13b). Prepared in 80% yield
from freshly distilled hexanal and (S,S)-11: Rf ) 0.60 (9:1
(3S ,4R ,5R )-6-[(t er t -B u t y ld im e t h y ls ily l)o x y ]-3-[[2′-
(4′,5′,6′,7′-t et r a h yd r o-3′,6′-d im et h ylb en zofu r yl)]d im et h -
ylsilyl]-5-m eth ylh ex-1-en -4-ol (31b). Prepared in 80% yield
(44% of major diastereomer 31b, 36% minor diastereomer 30b)
from 25 and (R,R)-11. The reaction stereoselectivity was
determined to be 55:45 by 1H NMR analysis of the reaction
mixture after filtration of the mixture through a silica gel plug,
eluting with 9:1 hexanes-Et2O, to remove excess allylboronate
reagent. Data for 31b follow: Rf ) 0.60 (9:1 hexanes-Et2O);
hexanes-Et2O); [R]25 ) 41.6° (c 1.00, C6H6); 1H NMR (400
D
MHz, C6D6) δ 6.04 (ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.06
(dd, J ) 2.2, 10.5 Hz, 1 H), 4.97 (m, 1 H), 3.88 (m, 1 H), 2.61
(br dd, J ) 5.4, 16.1 Hz, 1 H), 2.28-1.98 (m, 7 H), 1.68-1.09
(m, 12 H), 0.86 (t, J ) 6.4 Hz, 3 H), 0.82 (d, J ) 6.8 Hz, 3 H),
0.51 (s, 3 H), 0.47 (s, 3 H); 13C NMR (100 MHz, C6D6) δ 155.0,
150.3, 135.9, 131.6, 128.5, 118.3, 115.2, 71.5, 42.5, 37.5, 32.2,
32.1, 31.6, 29.8, 25.9, 23.0, 21.4, 20.2, 14.3, 9.6, -2.7, -3.2; IR
(neat) 3600-3450, 2960, 2930, 2870, 1625 cm-1; HRMS calcd
for C21H36O2Si 348.2475, found 348.2481.
[R]28 ) 15.8° (c 0.19, C6H6); 1H NMR (400 MHz, C6D6) δ 6.35
D
(ddd, J ) 10.5, 10.5, 17.2 Hz, 1 H), 5.09 (dd, J ) 2.2, 10.2 Hz,
1 H), 5.02 (dd, J ) 2.4, 17.5 Hz, 1 H), 4.03 (t, J ) 1.6 Hz, 1 H),
3.89 (br d, J ) 9.1 Hz, 1 H), 3.56 (dd, J ) 4.0, 9.9 Hz, 1 H),
3.38 (dd, J ) 8.6, 9.7 Hz, 1 H), 2.64 (br dd, J ) 5.2, 15.6 Hz,
1 H), 2.32-2.10 (m, 6 H), 2.00-1.88 (m, 1 H), 1.70-1.52 (m, 2
H), 1.40-1.08 (m, 3 H), 0.88 (s, 9 H), 0.82 (d, J ) 6.8 Hz, 3 H),
0.70 (s, 3 H), 0.62 (d, J ) 6.4 Hz, 3 H), 0.61 (s, 3 H), 0.06 (s, 6
H); 13C NMR (100 MHz, C6D6) δ 154.6, 150.9, 135.8, 131.5,
118.2, 114.3, 77.3, 69.7, 40.0, 39.4, 32.3, 31.7, 29.8, 26.0, 21.5,
20.3, 18.3, 12.7, 9.7, -2.8, -5.6, -5.7; IR (neat) 3520-3480,
2960, 2930, 2860, 1625 cm-1; HRMS calcd for C25H46O3Si2
450.2973, found 450.2967. Anal. Calcd for C25H46O3Si2: C,
66.60; H, 10.29. Found: C, 66.53; H, 10.46.
(3R,4S,5R)-5,6-O-Isopr opyliden e-3-[2′-(4′,5′,6′,7′-tetr ah y-
d r o-3′,6′-d im eth ylben zofu r yl)]d im eth ylsilyl]h ex-1-en -4-
ol (26b). 26b was prepared in 79% yield (77% of major
diastereomer 26b, 2% of minor diastereomer 27b) from freshly
distilled glyceraldehyde acetonide 23 and (S,S)-11. The reac-
1
tion stereoselectivity was determined to be >96:4 by H NMR
analysis of the reaction mixture after filtration of the mixture
through a silica gel plug, eluting with 5:1 hexanes-Et2O, to
remove excess allylboronate reagent. Data for 26b follow: Rf
1
) 0.60 (2:1 hexanes-Et2O); [R]28 ) 31.1° (c 1.15, C6H6); H
D
NMR (400 MHz, C6D6) δ 6.19 (ddd, J ) 10.2, 10.2, 17.2 Hz, 1
H), 4.94 (dd, J ) 2.4, 10.2 Hz, 1 H), 4.82 (dd, J ) 2.2, 17.2 Hz,
1 H) 4.15 (m, 1 H) 3.83 (dt, J ) 2.4, 8.3 Hz, 1 H), 3.80 (dd, J
) 6.4, 8.3 Hz, 1 H), 3.43 (dd, J ) 7.0, 8.0 Hz, 1 H), 2.63 (br dd,
J ) 5.1, 15.6 Hz, 1 H), 2.16 (t, J ) 2.2 Hz, 1 H), 2.28-2.08 (m,
3 H), 2.01 (s, 3 H), 1.91 (dt, J ) 2.2, 10.7 Hz, 1 H), 1.72-1.52
(m, 2 H), 1.27 (s, 3 H), 1.22 (s, 3 H), 1.18-1.08 (m, 1 H), 0.82
(d, J ) 6.8 Hz, 3 H), 0.59 (s, 3 H), 0.50 (s, 3 H); 13C NMR (100
MHz, C6D6) δ 154.8, 150.1, 135.0, 131.7, 118.2, 114.8, 109.5,
79.6, 73.9, 66.0, 38.5, 32.2, 31.6, 29.8, 25.7, 21.4, 20.2, 9.6, -2.9,
-3.2; IR (neat) 3600-3420, 2970, 2920, 1625 cm-1; HRMS
calcd for C21H34O4Si 378.2217, found 378.2239. Anal. Calcd
for C21H34O4Si: C, 66.63; H, 9.05. Found: C, 66.60; H, 9.25.
(3S,4R,5R)-5,6-O-Isopr opyliden e-3-[[2′-(4′,5′,6′,7′-tetr ah y-
d r o-3′,6′-d im eth ylben zofu r yl)]d im eth ylsilyl]h ex-1-en e-4-
ol (27b). Prepared in 85% yield (78% of major diastereomer
27b, 7% of minor diastereomer 26b) from freshly distilled
glyceraldehyde acetonide 23 and (R,R)-11. The reaction
stereoselectivity was determined to be 92:8 by 1H NMR
analysis of the reaction mixture after filtration of the mixture
through a silica gel plug, eluting with 5:1 hexanes-Et2O, to
R ep r esen t a t ive P r oced u r e for P r ot od esilyla t ion -
Oxidation of 2-(5-Meth ylfu r yl)-Su bstitu ted Silan ols: (R,S)-
1-Cycloh exylbu t-3-en e-1,2-d iol (9a ). Silanol 12a (40 mg,
0.14 mmol) was dissolved in 1 mL of CH2Cl2, and the solution
was cooled to 0 °C. CF3CO2H (11.6 µL, 0.15 mmol) was added
dropwise via syringe, and the mixture was stirred for 10 min
at 0 °C, at which point no starting material remained by TLC
analysis. The mixture was concentrated, and to the residue
was added 1 mL of THF-CH3OH (1:1), KHCO3 (28 mg, 0.28
mmol), KF‚2H2O (28 mg, 0.28 mmol), and 30% H2O2 (290 µL,
2.8 mmol). The mixture was stirred overnight (ca. 16 h) at
room temperature; the reaction was then quenched by addition
of Na2S2O3 (670 mg, 4.2 mmol). The mixture was stirred for
30 min with the Na2S2O3, and then the solids were filtered
and rinsed well with EtOAc. The filtrate was concentrated
and the crude product purified by flash chromatography,
eluting with 1:1 hexanes-EtOAc, to yield 14 mg (61%) of
known22 diol (R,S)-9a as a white solid (82% ee, as determined
by Mosher ester analysis46): mp 62-64 °C; [R]24 ) 16.3° (c
D
0.27, CHCl3) [lit.22 [R]23 ) 12.3 (c 2.3, CHCl3)].
remove excess allylboronate reagent. Data for 27b follow: Rf
D
1
) 0.50 (2:1 hexanes-Et2O); [R]28 ) 25.3° (c 1.19, C6H6); H
Diol 9a was also prepared in 79% yield and 81% ee (as
determined by Mosher ester analysis)46 from silanol 13a by
using the procedure subsequently described for preparation
D
NMR (400 MHz, C6D6) δ 6.06 (m, 1 H), 5.01 (m, 2 H), 4.11 (m,
1 H) 4.02-3.88 (m, 3 H), 2.58 (br dd, J ) 5.2, 16.4 Hz, 1 H),
2.29 (dd, J ) 2.0, 10.8 Hz, 1 H), 2.26-1.96 (m, 7 H), 1.66-
1.50 (m, 2 H), 1.36 (s, 3 H), 1.28 (s, 3 H), 1.20-1.06 (m, 1 H),
0.47 (s, 3 H), 0.42 (s, 3 H); 13C NMR (100 MHz, C6D6) δ 155.1,
149.9, 134.9, 132.1, 118.4, 115.6, 108.3, 78.8, 72.1, 66.2, 38.9,
32.1, 31.5, 29.7, 27.0, 25.6, 21.3, 20.1, 9.5, -3.1, -3.4; IR (neat)
3550-3400, 2960, 2920, 1620 cm-1; HRMS calcd for C21H34O4-
Si (M+) 378.2217, found 378.2213. Anal. Calcd for C21H34O4-
Si: C, 66.63; H, 9.05. Found: C, 66.84; H, 9.30.
of 34: mp 63-64 °C; [R]24 ) 14.3° (c 0.60, CHCl3).
D
(R,S)-Non -1-en e-3,4-d iol (9b). The known22 diol 9b was
prepared in 43% yield and 74% ee (as determined by Mosher
ester analysis46) from silanol 12b: mp 35-38 °C; [R]27D ) 8.4°
(c 0.62, CHCl3). Diol 9b was also prepared in 82% yield and
74% ee (as determined by Mosher ester analysis36) from silanol
13b by using the procedure subsequently described for prepa-
ration of 34: mp 37-39 °C; [R]29 ) 8.3° (c 0.45, CHCl3).
D
(3R ,4S ,5R )-6-[(t er t -B u t y ld im e t h y ls ily l)o x y ]-3-[[2′-
(4′,5′,6′,7′-t et r a h yd r o-3′,6′-d im et h ylb en zofu r yl)]d im et h -
ylsilyl]-5-m eth ylh ex-1-en -4-ol (30b). Prepared in 82% yield
(75% of major diastereomer 30b, 7% of minor diastereomer
31b) from 25 and (S,S)-11. The reaction stereoselectivity was
(3R,4R,5R)-5,6-O-Isop r op ylid en eh ex-1-en e-3,4-d iol (32).
The known22 diol 32 was prepared in 65% yield from silanol
26a . Diol 32 was also prepared in 77% yield from silanol 26b
by using the procedure subsequently described for preparation
of 34.

Enantioselective Synthesis of (-)-Swainsonine
J . Org. Chem., Vol. 62, No. 4, 1997 1123
(3S,4S,5R)-5,6-O-Isop r op ylid en eh ex-1-en e-3,4-d iol (33).
The known22 diol 33 was prepared in 65% yield from silanol
27a . Diol 33 was also prepared in 82% yield from silanol 27b
by using the procedure subsequently described for preparation
of 34.
(3R,4S,5R)-6-[(ter t-Bu tyld im eth ylsilyl)oxy]-5-m eth yl-
h ex-1-en e-3,4-d iol (36). The known22 diol 36 was prepared
in 56% yield from silanol 30a . Diol 36 was also prepared in
82% yield from silanol 30b by using the procedure subse-
quently described for preparation of 34.
(3R,4S,5R)-6-[(ter t-Bu t yld im et h ylsilyl)oxy]-5-m et h yl-
h ex-1-en e-3,4-d iol (37). The known22 diol 37 was prepared
in 53% yield from silanol 31a . Diol 37 was also prepared in
76% yield from silanol 31b by using the procedure subse-
quently described for preparation of 34.
(t, J ) 6.7 Hz, 2 H) 2.39 (dq, J ) 1.6, 7.0 Hz, 2 H), 2.02 (p, J
) 6.7, 2 H), 1.29 (t, J ) 7.3 Hz, 3 H).
To a 0 °C solution of ethyl 6-bromohex-2-enoate (11.3 g, 51.1
mmol) in 50 mL of Et2O was added 150 mL of a 1.0 M solution
of DIBAL in hexanes. The mixture was allowed to warm to
room temperature, stirred for 2 h at this temperature, and
then recooled to 0 °C and quenched by addition of 200 mL of
1 N HCl. The phases were separated, and the aqueous layer
was extracted with 3 × 100 mL of Et2O. The combined
organics were dried over MgSO4 filtered, and concentrated to
a colorless oil. The crude product was purified by vacuum
distillation (5 mmHg, bp 95-100 °C) to yield 7.92 g (87%) of
the known64,65 allylic alcohol 39: 1H NMR (400 MHz, CDCl3)
δ 5.68 (m, 2 H), 4.10 (d, J ) 4.6 Hz, 2 H) 3.41 (t, J ) 6.7 Hz,
2 H) 2.22 (dd, J ) 5.9, 7.0 Hz, 2 H), 1.95 (quintet, J ) 6.7 Hz,
2 H).
R ep r esen t a t ive P r oced u r e for P r ot od esilyla t ion -
Oxid a tion of Men th ofu r yl-Su bstitu ted Sila n ols: (3R,4S,-
5S,6R)-9-Br om o-5,6-ep oxyn on -1-en e-3,4-d iol (34). Silanol
28b (363 mg, 0.82 mmol) was dissolved in 1.5 mL of THF, and
the solution was cooled to 0 °C. CF3CO2H (70 µL, 0.90 mmol)
in 0.5 mL of THF was added dropwise via syringe, and the
mixture was warmed to room temperature and stirred for 2.5
h, at which point no starting material remained by TLC
analysis. To the solution was added 2 mL of CH3OH, KHCO3
(160 mg, 1.60 mmol), KF‚2H2O (160 mg, 1.60 mmol), and 30%
H2O2 (1.6 mL, 16.0 mmol). The mixture was stirred for 24 h;
then the reaction was quenched by addition of Na2S2O3 (3.8 g,
24.0 mmol). The mixture was stirred over Na2S2O3 for 30 min,
Na2SO4 was added, and the solids were filtered and rinsed well
with EtOAc. The filtrate was concentrated and the crude
product was purified by flash chromatography, eluting with a
gradient system of 1:1 hexanes-Et2O to 100% Et2O, to yield
(2S,3R)-6-Br om o-2,3-ep oxyh exa n ol (40). CH2Cl2 (100
mL) was stirred over crushed 4 Å molecular sieves for 15 min;
Ti(OiPr)4 (1.7 mL, 5.8 mmol) was then added, and the solution
was cooled to -30 °C in a H2O-ethylene glycol bath. D-DIPT
(1.6 mL, 7.7 mmol) was added, and the solution was stirred
for 5 min at -30 °C; allylic alcohol 39 (6.9 g, 38.5 mmol) in 10
mL of CH2Cl2 was then added, and the solution was stirred
for 20 min at -30 °C. tert-Butyl hydroperoxide (13.6 mL of a
4.27 M solution in toluene) was added, and the reaction
mixture was stored at -20 °C for 2 d, filtered through Celite,
and concentrated to a yellow oil. The crude product was
purified by flash chromatography, eluting with 1:1 hexanes-
EtOAc, to yield 5.48 g (73%) of epoxy alcohol 40: Rf ) 0.17
1
(1:1 hexanes-EtOAc); [R]23 ) 8.4° (c 1.02, CHCl3); H NMR
D
(400 MHz, CDCl3) δ 3.92 (ddd, J ) 2.7, 2.4, 12.6 Hz, 1 H),
3.66 (ddd, J ) 4.0, 7.3, 12.6 Hz, 1 H), 3.46 (m, 2 H), 3.00 (ddd,
J ) 2.4, 4.6, 7.0 Hz, 1 H), 2.96 (dt, J ) 2.4, 4.0 Hz, 1 H), 2.12-
1.95 (m, 2 H), 1.90-1.81 (m, 1 H), 1.72-1.60 (m, 2 H); 13C NMR
(100 MHz, CDCl3) δ 61.4, 58.2, 54.8. 33.0, 30.0, 29.2; IR (neat)
3500-3300, 2920, 2860 cm-1; HRMS calcd for C6H12O281Br (M
+ 1) 196.9997, found 196.9980; calcd for C6H12O279Br (M + 1)
195.0017, found 195.0013.
176 mg (85%) of diol 34: Rf ) 0.16 (2:1 Et2O-hexanes); [R]23
D
) 11.2° (c 0.73, CHCl3); 1H NMR (400 MHz, CDCl3) δ 5.97 (ddd,
J ) 6.2, 10.5, 17.2 Hz, 1 H), 5.32 (dt, J ) 1.3, 17.2 Hz, 1 H),
5.42 (dt, J ) 1.3, 10.7 Hz, 1 H), 4.35 (m, 1 H), 3.82 (t, J ) 4.0
Hz, 1 H), 3.47 (m, 2 H), 3.05 (ddd, J ) 2.2, 5.1, 6.7 Hz, 1 H),
2.91 (dd, J ) 2.2, 4.0 Hz, 1 H), 2.28-2.12 (br s, 1 H), 2.10-
1.98 (m, 2 H), 1.82-1.52 (m, 3 H); 13C NMR (100 MHz, CDCl3)
δ 135.5, 117.7, 74.0, 71.4, 57.3, 54.6, 33.0, 30.0, 29.1; IR (neat)
3500-3300, 2930 cm-1; HRMS calcd for C9H15O381Br, 252.0180,
found 252.0196. Anal. Calcd for C9H15O3Br: C, 43.04; H, 6.02.
Found: C, 42.89; H, 6.03.
En a n tiom er ic P u r ity An a lysis of 40. Mosher ester
analysis of 40 followed the standard literature procedure.46
(S)-MTP A ester : 1H NMR (400 MHz, C6D6) δ 7.66 (m, 2 H),
7.08-6.90 (m, 3 H), 3.95 (dd, J ) 3.8, 11.8 Hz, 1 H) 3.71 (dd,
J ) 6.5, 12.1 Hz, 1 H) 3.38 (s, 3 H) 2.82-2.70 (m, 2 H), 2.32
(ddd, J ) 2.2, 3.2, 5.6 Hz, 1 H), 2.12 (ddd, J ) 2.2, 4.6, 6.7 Hz,
1 H), 1.43-1.24 (m, 2 H). 1.18-1.24 (m, 1 H), 1.03-0.93 (m,
1 H). (R)-MTP A ester : 1H NMR (400 MHz, C6D6) δ 7.57 (m,
2 H), 7.45-7.36 (m, 3 H), 4.12 (dd, J ) 3.2, 12.0 Hz, 1 H), 3.67
(dd, J ) 5.6, 12.0 Hz, 1 H), 3.42 (s, 3 H), 2.85-2.75 (m, 2 H),
2.39 (ddd, J ) 2.2, 3.2, 5.6 Hz, 1 H), 2.25 (ddd, J ) 2.2, 5.6,
6.7 Hz, 1 H), 1.45-1.25 (m, 2 H). 1.21-1.09 (m, 1 H), 1.05-
0.97 (m, 1 H).
Diol 34 was also prepared in 57% yield from silanol 28a by
using the procedure described for preparation of 9a .
(3S,4R,5S,6R)-9-Br om o-5,6-epoxyn on -1-en e-3,4-diol (35).
Diol 35 was prepared in 79% yield from silanol 29b: Rf ) 0.15
(2:1 Et2O-hexanes); mp 58-59 °C; [R]26 ) 18.7° (c 0.54,
D
CHCl3); 1H NMR (400 MHz, CDCl3) δ 5.96 (ddd, J ) 5.6, 10.7,
17.2 Hz, 1 H), 5.43 (dt, J ) 1.6, 17.3 Hz, 1 H), 5.32 (dt, J )
1.6, 10.8 Hz, 1 H), 4.30 (br s, 1 H), 3.61 (m, 1 H) 3.50 (m, 2 H),
2.95 (m, 2 H), 2.37-2.25 (m, 2 H), 2.10 (m, 2 H), 1.88-1.78
(m, 1 H), 1.70-1.58 (m, 1 H); 13C NMR (100 MHz, CDCl3) δ
136.0, 117.4, 74.6, 72.2, 57.7, 55.1, 32.9, 30.0, 29.1; IR (CHCl3)
3600-3500, 3010, 2920 cm-1; HRMS calcd for C9H16O381Br (M
+ 1) 253.0258, found 253.0267. Anal. Calcd for C9H15O3Br:
C, 43.04; H, 6.02. Found: C, 43.10; H, 6.11.
(2R,3R)-6-Br om o-2,3-ep oxyh exa n a l (24). To a 0 °C solu-
tion of alcohol 40 (0.5 g, 2.6 mmol) in 4 mL of CH2Cl2 was
added 5 mL of dry DMSO, Et3N (1.8 mL, 12.8 mmol), and SO3-
pyridine (2.04 g, 12.8 mmol). The cooling bath was removed,
and the mixture was stirred for 20 min at room temperature,
at which point TLC analysis showed that no starting material
remained. The mixture was diluted with 35 mL of Et2O and
washed with 20 mL of saturated aqueous NaHCO3 and 20 mL
of brine. The aqueous washes were extracted with 4 × 10 mL
of EtOAc, and the combined organics were dried over MgSO4,
filtered, and concentrated to a yellow oil. The crude product
was purified by flash chromatography, eluting with 2:1 hex-
anes-EtOAc, to yield 392 mg (80%) of epoxy aldehyde 24: Rf
6-Br om oh ex-2-en ol (39). To a 1 L flask equipped with an
overhead mechanical stirrer was added 3.6 g (88.9 mmol) of a
60% dispersion of NaH in oil. The dispersion was rinsed with
dry THF to remove oil; 260 mL of THF was then added.
Triethyl phosphonoacetate (16.6 g, 74.1 mmol) in 10 mL of
THF was added, and after gas evolution ceased, the mixture
was cooled to -50 °C in a dry ice-i-PrOH bath. To the cold
mixture was added 4-bromobutanal (38)64,65 (12.3 g, 81.5 mmol)
in 10 mL of THF. The resulting mixture was allowed to warm
slowly to room temperature and stirred for 2.5 h. The mixture
was recooled to 0 °C and quenched by addition of H2O (150
mL). The phases were separated, and the aqueous phase was
extracted with 3 × 100 mL of Et2O. The combined organics
were washed with brine (150 mL), dried over MgSO4, filtered,
and concentrated to a yellow oil that was purified by vacuum
distillation (5 mmHg, bp 95-100 °C) to yield 13.9 g (85%) of
the known64,65 ethyl 6-bromohex-2-enoate as a colorless oil: 1H
NMR (400 MHz, CDCl3) δ 6.91 (dt, J ) 7.0, 15.6 Hz, 1 H),
5.88 (dt, J ) 1.6, 15.6 Hz, 2 H), 4.19 (q, J ) 7.3 Hz, 2 H) 3.41
) 0.50 (1:1 hexanes-EtOAc); [R]26 ) -24.8° (c 0.21, C6H6);
D
1H NMR (400 MHz, CDCl3) δ 9.18 (d, J ) 6.2 Hz, 1 H), 3.47
(m, 2 H), 3.28 (ddd, J ) 1.9, 4.3, 6.2 Hz, 1 H), 3.19 (dd, J )
1.9, 6.2 Hz, 1 H), 2.20-1.90 (m, 3 H), 1.80-1.70 (m, 1 H); 13
C
NMR (100 MHz, CDCl3) δ 197.9, 58.9, 55.8. 32.4, 29.8, 28.9;
81
IR (neat) 2960, 2920, 2860, 1725 cm-1; HRMS for C6H10O2
-
Br (M + 1) 193.9763, found 193.9853.
(3R ,4S ,5S ,6R )-9-Br om o-5,6-e p oxy-3,4-O-d ih yd r oxy-
n on a n -1-a l 3,4-Aceton id e (41). Diol 34 (99 mg, 0.39 mmol),
p-PTS (5 mg, 0.02 mmol), and 2,2-dimethoxypropane (0.5 mL,
44 mmol) were dissoved in 1 mL of CH2Cl2. The solution was
heated to reflux for 2 h, at which point no starting material

1124 J . Org. Chem., Vol. 62, No. 4, 1997
Hunt and Roush
remained by TLC analysis. The solution was cooled to room
temperature, diluted with 10 mL of of Et2O, and washed with
5 mL of saturated aqueous NaHCO3 solution The aqueous
phase was extracted with 3 × 5 mL of Et2O, and the combined
organics were dried over MgSO4, filtered, and concentrated.
The residue was purified by flash chromatography, eluting
with 9:1 hexanes-Et2O, to yield 107 mg of (3R,4S,5S,6R)-9-
bromo-5,6-epoxy-3,4-dihydroxynon-1-ene 3,4-acetonide (95%)
purified by flash chromatography, eluting with a gradient
system of 1:1 hexanes-EtOAc to 100% EtOAc to yield 15 mg
(71%) of swainsonine acetonide54,61 as an off-white solid: Rf )
0.10 (1:1 hexanes-EtOAc); mp 102-103 °C (lit.54 mp 105-
107 °C; lit.61 mp 100-103 °C); [R]26 ) -84.6° (c 1.53, CH3-
D
OH) [lit.54 [R]24D ) -75.1° (c 1.54, CH3OH); lit.61 [R]25D ) -72.7°
1
(c 0.43, CH3OH)]; H NMR (400 MHz, CDCl3) δ 4.70 (dd, J )
4.6, 6.2 Hz, 1 H), 4.61 (dd, J ) 4.3, 6.5 Hz, 1 H), 3.83 (m, 1 H),
3.15 (d, J ) 10.5 Hz, 1 H), 2.99 (dt, J ) 3.2, 10.5 Hz, 1 H),
2.13 (dd, J ) 4.3, 10.8 Hz, 1 H), 2.05 (m, 2 H), 1.85 (m, 1 H),
1.71-1.60 (m, 3 H), 1.50 (s, 3 H), 1.33 (s, 3 H), 1.30-1.17 (m,
1 H); 13C NMR (100 MHz, CDCl3) δ 111.5, 79.3, 78.3, 73.7,
as a colorless oil: Rf ) 0.72 (2:1 Et2O-hexanes); [R]25 ) 3.9°
D
1
(c 1.23, CHCl3); H NMR (400 MHz, CDCl3) δ 5.97 (ddd, J )
7.0, 10.5, 17.2 Hz, 1 H), 5.48 (dt, J ) 1.3, 17.2 Hz, 1 H), 5.36
(dt, J ) 1.3, 10.5 Hz, 1 H), 4.72 (t, J ) 6.7 Hz, 1 H), 3.80 (t, J
) 6.5 Hz, 1 H), 3.46 (m, 2 H), 2.90 (ddd, J ) 2.2, 4.8, 7.0 Hz,
1 H), 2.77 (dd, J ) 2.2, 7.5 Hz, 1 H), 2.10-1.97 (m, 2 H), 1.88-
1.80 (m, 1 H), 1.68-1.59 (m, 1 H), 1.52 (s, 3 H), 1.38 (s, 3 H);
13C NMR (100 MHz, CDCl3) δ 132.4, 118.8, 109.3, 78.6, 78.5,
56.5, 55.8, 32.9, 29.9, 29.2, 27.7, 25.1; IR (CHCl3) 2980, 2930,
1450, 1430 cm-1; HRMS calcd for C11H16O379Br (M - CH3),
275.0278, found 275.0291. Anal. Calcd for C12H19O3Br: C,
49.50; H, 6.58. Found: C, 49.69; H, 6.46.
67.6, 60.0, 51.6, 33.0, 26.0, 25.0, 24.1; HRMS calcd for C11H19
NO3 213.1360, found 213.1358.
-
(1S,2R,8R,8a R)-1,2,8-Tr ih yd r oxyin d olizin e (42) [(-)-
Sw a in son in e]. Swainsonine acetonide (15 mg, 0.07 mmol)
was dissolved in 1 mL of THF, and 1 mL of 6 N HCl was added.
The mixture was stirred vigorously for 20 h at ambient
temperature and then concentrated in vacuo. The crude
product was purified by ion-exchange chromatography with
Amberlite IRA-400(OH), eluting with distilled H2O, to yield
11 mg (92%) of (-)-swainsonine (42) as a white crystalline
solid: Rf ) 0.35 (4:4:4:1 n-BuOH-CHCl3-CH3OH-concd NH4-
OH); mp 137-140 °C (lit.54 mp 144-145 °C; lit.58 mp 141-
143 °C; lit.62 mp 138-140 °C); [R]26D ) -85.5° (c 0.42 CH3OH);
[lit.54 [R]25 ) -87.2° (c 2.1, CH3OH); lit.58 [R]26 ) -82.6° (c
(3R,4S,5S,6R)-9-Bromo-5,6-epoxy-3,4-dihydroxynon-1-ene 3,4-
acetonide from the preceding experiment (105 mg, 0.36 mmol)
was dissolved in 6 mL of CH2Cl2, and the solution was cooled
to -78 °C. A stream of O3 in O2 was bubbled through this
solution until the solution turned blue and TLC analysis of
the reaction mixture showed no starting material; the solution
was then flushed with O2 for 5 min, and Ph3P was added. The
mixture was allowed to warm to room temperature stirred for
1 h at this temperature, and then concentrated in vacuo. The
crude product was purified by flash chromatography, eluting
with 1:1 hexanes-EtOAc, to yield 92 mg (88%) of aldehyde
D
D
1.03, CH3OH); lit.62 [R]27 ) -85.2° (c 0.5, CH3OH)]; 1H NMR
D
(400 MHz, D2O) δ 4.33 (ddd, J ) 2.4, 6.2, 8.0 Hz, 1 H), 4.24
(dd, J ) 3.8, 5.9 Hz, 1 H), 3.78 (ddd, J ) 4.6, 9.7, 10.7 Hz, 1
H), 2.89 (br d, overlapping with dd δ 2.86, 1 H), 2.86 (dd,
overlapping with br d δ 2.89, J ) 2.1, 11.0 Hz, 1 H), 2.05 (m,
1 H), 1.93 (dt, overlapping with dd δ 1.89, J ) 3.0, 12.4 Hz, 1
H), 1.89 (dd, overlapping with dt δ 1.93, J ) 3.5, 9.4 Hz, 1 H),
1.71 (m, 1 H), 1.50 (qt, J ) 4.0, 13.4 Hz, 1 H), 1.22 (qd, J )
4.6, 12.4 Hz, 1 H); 13C NMR (100 MHz, D2O) δ 73.1, 70.0, 69.4,
66.7, 61.0, 52.0, 32.8, 23.6; HRMS calcd for C8H16NO3 (M + 1)
174.1126, found 174.1133.
41: Rf ) 0.26 (1:1 hexanes-EtOAc); [R]25 ) 27.6° (c 0.21,
D
1
C6H6); H NMR (400 MHz, C6D6) δ 9.56 (d, J ) 2.4 Hz, 1 H),
4.00 (dd, J ) 2.4, 7.0 Hz, 1 H), 3.71 (dd, J ) 5.9, 7.0 Hz, 1 H),
2.84 (m, 2 H), 2.58 (dd, J ) 2.4, 5.9 Hz, 1 H), 2.45 (ddd, J )
1.9, 4.6, 6.7 Hz, 1 H), 1.50-1.38 (m, 5 H), 1.30-1.20 (m, 1 H),
1.14-1.02 (m, 4 H); 13C NMR (100 MHz, CDCl3) δ 192.2, 111.7,
81.6, 77.5, 56.6, 55.3, 32.8. 29.7, 29.0, 27.2, 25.2; IR (CHCl3)
2985, 2930, 1735, 1450, 1435 cm-1; HRMS calcd for C10H14
-
Ack n ow led gm en t. This research was supported by
a grant from the National Institute of General Medical
Science (GM 38907). J .A.H. acknowledges support from
the Department of Education National Needs Fellow-
ship and the E. M. Kratz Research Fellowship.
O479Br (M - CH3) 277.0071, found 277.0067. Anal. Calcd for
C11H17O4Br: C, 45.07; H, 5.85. Found: C, 44.95; H, 5.98.
(1S,2R,8R,8a R)-8-Hyd r oxy-1,2-(isop r op ylid en ed ioxy)-
in d olizin e (Sw a in son in e Aceton id e). A mixture of alde-
hyde 41 (30 mg, 0.10 mmol), NH4OAc (10 mg, 0.15 mmol), and
25 mg of 3 Å molecular sieves in 1 mL of CH3OH was heated
to reflux for 6 h. The mixture was cooled to room temperature,
and NaBH3CN (13 mg, 0.20 mmol) was added.68 The mixture
was stirred for 12 h at room temperature; 1 mL of saturated
aqueous NaHCO3 was then added, and the CH3OH was
removed in vacuo. The remaining suspension was extracted
with 4 × 10 mL of EtOAc, and the combined extracts were
concentrated to a pale yellow oil. The crude product was
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra of
(R,R)-10, (R,R)-11, 13a ,b, 15, 17, 22, 24, 40, and 42 (10 pages).
This material is contained in libraries on microfiche, im-
mediately follows this article in the microfilm version of the
journal, and can be ordered from the ACS; see any current
masthead page for ordering information.
J O961840S