Total syntheses of (+)-australine and (-)-7-epialexine

Article abstract of DOI:10.1021/jo000689q

Three approaches were examined for the synthesis of 3-(hydroxymethyl)pyrrolizidines, a class of compounds that includes the polyhydroxylated pyrrolizidine alkaloids alexine (1), australine (2), and various stereoisomers of thereof. In the first approach, the intramolecular cycloaddition of an azide onto an electron-rich 1,3-diene bearing a terminal alkoxymethyl substituent (i.e., 21) afforded the dehydropyrrolizidines 22a and 22b, with 22a predominating. A rationale for this stereoselectivity was proposed. Transformation of the major diastereomer 22a into a natural 3-(hydroxymethyl)-pyrrolizidine was not possible due to difficulties encountered in transforming the phenyl vinyl sulfide functionality into other useful functional groups. A second approach was examined, Wherein the intramolecular cycloaddition of an azide with an optically pure S-t-Bu-substituted diene (i.e., 30) was found to produce the pyrrolizidine 31. In this case, the alkoxymethyl substituent was incorporated into the tether between the azide and the diene, rather than on the diene itself. A key transformation in the synthesis of the diene 30 was the use of the allylic borane R₂BCH₂CH=C(TMS)(StBu) for the stereoselective conversion of the D-arabinose-derived azido aldehyde 28 to the E-isomer of 30. The cyclization of 30 to 31 also produced the bicyclic triazene 32, the result of 1,3-dipolar cycloaddition of the azide onto the distal double bond of the diene. Again, difficulties in transformation of the vinyl sulfide functionality of 31 into useful oxygen functionality limited this approach to naturally occurring 3-(hydroxymethyl)pyrrolizidines. A third approach to these compounds was successful. The transformation of L-xylose into the azido epoxy tosylate 46 was accomplished using two Wittig reactions and an epoxidation, in addition to other standard functional group manipulations. Reductive double-cyclization of 46 afforded the pyrrolizidines 47a and 47b, which were debenzylated to afford (+)-australine 2 and (-)-7-epialexine 4, respectively. In the preliminary report of this work, erroneous spectroscopic data in the original literature on the structural assignment of australine led to the conclusion that the synthetic material obtained herein was actually (+)-7-epiaustraline. Recently corrected spectroscopic data have appeared which verify that (+)-australine 2 was indeed synthesized for the first time.

Full text of DOI:10.1021/jo000689q

J . Org. Chem. 2000, 65, 5785-5793
5785
Tota l Syn th eses of (+)-Au str a lin e a n d (-)-7-Ep ia lexin e
William H. Pearson* and J ennifer V. Hines1
Department of Chemistry, University of Michigan, Ann Arbor, Michigan, 48109-1055
wpearson@umich.edu
Received May 4, 2000
Three approaches were examined for the synthesis of 3-(hydroxymethyl)pyrrolizidines, a class of
compounds that includes the polyhydroxylated pyrrolizidine alkaloids alexine (1), australine (2),
and various stereoisomers of thereof. In the first approach, the intramolecular cycloaddition of an
azide onto an electron-rich 1,3-diene bearing a terminal alkoxymethyl substituent (i.e., 21) afforded
the dehydropyrrolizidines 22a and 22b, with 22a predominating. A rationale for this stereoselectivity
was proposed. Transformation of the major diastereomer 22a into a natural 3-(hydroxymethyl)-
pyrrolizidine was not possible due to difficulties encountered in transforming the phenyl vinyl sulfide
functionality into other useful functional groups. A second approach was examined, wherein the
intramolecular cycloaddition of an azide with an optically pure S-t-Bu-substituted diene (i.e., 30)
was found to produce the pyrrolizidine 31. In this case, the alkoxymethyl substituent was
incorporated into the tether between the azide and the diene, rather than on the diene itself. A
key transformation in the synthesis of the diene 30 was the use of the allylic borane R2BCH2CHd
C(TMS)(StBu) for the stereoselective conversion of the D-arabinose-derived azido aldehyde 28 to
the E-isomer of 30. The cyclization of 30 to 31 also produced the bicyclic triazene 32, the result of
1,3-dipolar cycloaddition of the azide onto the distal double bond of the diene. Again, difficulties in
transformation of the vinyl sulfide functionality of 31 into useful oxygen functionality limited this
approach to naturally occurring 3-(hydroxymethyl)pyrrolizidines. A third approach to these
compounds was successful. The transformation of L-xylose into the azido epoxy tosylate 46 was
accomplished using two Wittig reactions and an epoxidation, in addition to other standard functional
group manipulations. Reductive double-cyclization of 46 afforded the pyrrolizidines 47a and 47b,
which were debenzylated to afford (+)-australine 2 and (-)-7-epialexine 4, respectively. In the
preliminary report of this work, erroneous spectroscopic data in the original literature on the
structural assignment of australine led to the conclusion that the synthetic material obtained herein
was actually (+)-7-epiaustraline. Recently corrected spectroscopic data have appeared which verify
that (+)-australine 2 was indeed synthesized for the first time.
In tr od u ction
The legumes Alexa leiopetala and Castanospermum
australe have yielded a small class of pyrrolizidine
alkaloids bearing a C3 hydroxymethyl group,2 repre-
sented by alexine (1)3,4 and australine (2), Figure 1.5-7
Several other stereoisomers of alexine and australine,
e.g., 7-epiaustraline (3)8 and 7-epialexine (4),4,6a have
(1) Current address: Department of Chemistry, Ohio University,
Athens, OH 45701.
(2) (a) Fellows, L. E.; Fleet, G. W. In Natural Products Isolation;
Wagman, G. H., Copper, R., Eds.; Elsevier: Amsterdam, 1988; p 540.
(b) Nash, R. J .; Watson, A. A.; Asano, N. In Alkaloids: Chemical and
Biological Perspectives; Pelletier, S. W, Ed.; Elsevier: Oxford, 1996;
Vol. 11; pp 345-376. (c) Liddell, J . R. Nat. Prod. Rep. 1996, 187-193.
(3) Isolation of alexine: Nash, R. J .; Fellows, L. E.; Dring, J . V.;
Fleet, G. W. J .; Derome, A. E.; Hamor, T. A.; Scofield, A. M.; Watkin,
D. J . Tetrahedron Lett. 1988, 29, 2487-2490.
(4) Synthesis of alexine: Fleet, G. W. J .; Haraldsson, M.; Nash, R.
J .; Fellows, L. E. Tetrahedron Lett. 1988, 29, 5441-5444. The syntheses
of 3-epi- and 7-epialexine are also reported.
(5) Isolation of australine: Molyneux, R. J .; Benson, M.; Wong, R.
Y.; Tropea, J . E.; Elbein, A. D. J . Nat. Prod. 1988, 51, 1198-1206.
(6) Syntheses of australine: (a) Pearson, W. H.; Hines, J . V.
F igu r e 1.
Tetrahedron Lett. 1991, 32, 5513-5516 (the synthesis of 7-epiaustra-
been obtained from nature or by synthesis.2 The potential
of the polyhydroxylated pyrrolizidine alkaloids as selec-
line reported in this paper was actually a synthesis of australine; see
text for explanation). (b) White, J . D.; Hrnciar, P.; Yokochi, A. F. T. J .
Am. Chem. Soc. 1998, 120, 7359-7360. (c) Denmark, S. E.; Martin-
borough, E. A. J . Am. Chem. Soc. 1999, 121, 3046-3056. For
a
synthesis of australine from castanospermine, see: (d) Furneaux, R.
H.; Gainsford, G. J .; Mason, J . M.; Tyler, P. C. Tetrahedron 1994, 50,
2131-2160.
(7) For the synthesis of ring-expanded analogues of australine and
alexine, see: Pearson, W. H.; Hembre, E. J . J . Org. Chem. 1996, 61,
5545-5556.
10.1021/jo000689q CCC: $19.00 © 2000 American Chemical Society
Published on Web 08/15/2000

5786 J . Org. Chem., Vol. 65, No. 18, 2000
Pearson and Hines
Sch em e 1
tive glycosidase inhibitors and as antiviral and antiret-
roviral agents makes them attractive targets for synthe-
sis.9 Alexine (1) is an inhibitor of R-glucosidase, trehalase,
amyloglucosidase, fungal glucan 1,4-R-glucosidase,8a and
thioglucosidase.10 Australine (2) is an inhibitor of
amyloglucosidase8a,5,11 and the glycoprotein processing
enzyme glucosidase I11 and displays antiviral12 and anti-
HIV activity.13 We wish to report our efforts to synthesize
3-(hydroxymethyl)pyrrolizidines, ultimately culminating
in a synthesis of (+)-australine (2) and (-)-7-epialexine
(4).14
A retrosynthetic analysis of the 3-(hydroxymethyl)-
pyrrolizidines is shown in Figure 1. Our initial efforts
centered on the synthesis of the pyrrolines 5 and 7 by a
[4 + 1] cycloadditive approach15,16 involving the cycliza-
tion of the azidodienes 6 and 8, respectively (routes a and
b). A second approach involving the reductive double-
cyclization of azidoepoxides such as 9 proved to be most
practical for the total synthesis of 2 and 4 (route c).17
alexine numbering). We were confident that the relative
stereochemistry at C(7) and C(7a) would result based on
our prior work with dienes bearing an allylic alkoxy
group.15 However, since our prior work generally involved
azido dienes with no substituent at the diene terminus,
we were interested in the outcome of these cyclizations
with respect to the relative configuration at C(3) vs C(7a).
It is believed that the reaction proceeds via the triazoline
11, which should fragment with loss of nitrogen to the
zwitterion 12, which is probably in equilibrium with a
vinylaziridine (not shown). A 5-endo-trig closure of 12
would then produce 13. Examination of models suggested
that the zwitterion with the alkoxymethyl group exo to
the newly forming ring (12-exo) would be preferred to
the alternative zwitterion 12-en d o based on steric
grounds, although the preference might not be high.
Thus, we expected that pyrroline 13â would be produced
predominantly over 13R, potentially allowing access to
C(3)-epi analogues of alexine, for example, the known
synthetic compound 3-epialexine.18
The synthesis of an appropriate alkoxymethyl-substi-
tuted azido diene (i.e., 21) and its cyclization is shown
in Scheme 2. The enyne 1419 was protected as its tert-
butyldiphenylsilyl ether 15, then metalated and added
to the aldehyde 16.20 RedAl reduction of the resultant
propargyl alcohol 17 using Marshall’s method21 followed
by quenching of the aluminum intermediate with phe-
nylsulfenyl chloride according to our prior work15 gave
the phenylthiodiene 18. Protection of the secondary
alcohol followed by removal of the tert-butyldimethylsilyl
group gave the primary alcohol 20, which was converted
to the key azide 21 by a Mitsunobu reaction.22 Heating
21 at 90 °C in THF in a sealed tube followed by
purification and separation gave 22a and 22b in 35% and
15% yields, respectively. In both compounds, no NOE was
detected between the protons at C(7) and C(7a), consis-
tent with a trans relationship, as precedented by our
prior work on the cyclization of dienes bearing an allylic
alkoxy group.15 The relative configuration at C(3) rested
Resu lts a n d Discu ssion
In tr am olecu lar Dipolar Cycloaddition of an Azide
w ith a n Alk oxym eth yl-Su bstitu ted Dien e. Our initial
approach to 3-(alkoxymethyl)pyrrolizidines was based on
our work on the intramolecular 1,3-dipolar cycloaddition
of azides with electron-rich dienes, a method that pro-
duces bicyclic pyrrolines.15 Two disconnections of the
pyrrolizidine nucleus were explored, the first of which is
outlined in Schemes 1 and 2. We planned to study a
precursor such as 10, which bears an alkoxymethyl group
at the terminus of the diene (Scheme 1). Heating 10
should produce the bicyclic pyrroline 13 with an alkoxy-
methyl group in the desired C(3) position (australine/
(8) For a report of the isolation of 7-epiaustraline, later found to be
australine, see: (a) Nash, R. J .; Fellows, L. E.; Dring, J . V.; Fleet, G.
W. J .; Girdhar, A.; Ramsden, N. G.; Peach, J . M.; Hegarty, M. P.;
Scofield, A. M. Phytochemistry 1990, 29, 111-114. For a report of the
synthesis of 7-epiaustraline, later found to be australine, see ref 6a.
For an unambiguous synthesis of 7-epiaustraline, see: (b) Denmark,
S. E.; Herbert, B. J . Am. Chem. Soc. 1998, 120, 7357-7358. (c)
Denmark, S. E.; Herbert, B. J . Org. Chem. 2000, 65, 2887-2896. See
the text for a full explanation.
(9) For an excellent review of the chemistry and biochemistry of
alkaloidal glycosidase inhibitors, see: Iminosugars as Glycosidase
Inhibitors; Stu¨tz, A. E., Ed.; Wiley-VCH: Weinheim, 1999.
(10) Scofield, A. M.; Rossiter, J . T.; Witham, P.; Kite, G. C.; Nash,
R. J .; Fellows, L. E. Phytochemistry 1990, 29, 107-109.
(11) Tropea, J . E.; Molyneux, R. J .; Kaushal, G. P.; Pan, Y. T.;
Mitchell, M.; Elbein, A. D. Biochemistry 1989, 28, 2027-2034.
(12) Elbein, A. D.; Tropea, J . E.; Molyneux, R. J . U. S. Pat. Appl.
US 289,907; Chem. Abstr. 1990, 113, P91444p.
(13) Fellows, L.; Nash, R. PCT Int. Appl. WO GB Appl. 89/7,951;
Chem. Abstr. 1990, 114, 143777f.
(14) For a preliminary report of this work, see ref 6a.
(15) Pearson, W. H.; Bergmeier, S. C.; Degan, S.; Lin, K.-C.; Poon,
Y.-F.; Schkeryantz, J . M.; Williams, J . P. J . Org. Chem. 1990, 55, 5719-
5738.
(16) Hudlicky, T.; Seoane, G.; Price, J . D.; Gadamasetti, K. G. Synlett
1990, 433.
(17) For related one-pot double cyclizations using azido epoxides or
amino epoxides, see refs 6a, 7, and: (a) Setoi, H.; Takeno, H.;
Hashimoto, M. J . Org. Chem. 1985, 50, 3948-3950. (b) Setoi, H.;
Takeno, H.; Hashimoto, M. Tetrahedron Lett. 1985, 26, 4617-4620.
(c) Kim, Y. G.; Cha, J . K. Tetrahedron Lett. 1989, 30, 5721-5724. (d)
Pearson, W. H.; Bergmeier, S. C.; Williams, J . P. J . Org. Chem. 1992,
57, 3977-3987. (e) Ina, H.; Kibayashi, C. J . Org. Chem. 1993, 58, 52-
61. (f) J irousek, M. R.; Cheung, A. W.-H.; Babine, R. E.; Sass, P. M.;
Schow, S. R.; Wick, M. M. Tetrahedron Lett. 1993, 34, 3671-3674. (g)
Kim, N.-S.; Choi, J .-R.; Cha, J . K. J . Org. Chem 1993, 58, 7096-7099.
(h) Poitout, L.; Le Merrer, Y.; Depezay, J .-C. Tetrahedron Lett. 1994,
35, 3293-3296. (i) Lohray, B. B.; J ayamma, Y. and Chatterjee, M. J .
Org. Chem. 1995, 60, 5958-5960. (j) Pearson, W. H.; Hembre, E. J . J .
Org. Chem. 1996, 61, 5537-5545. (k) Hembre, E. J .; Pearson, W. H.
Tetrahedron 1997, 53, 11021-11032.
(18) Fleet, G. W. J .; Haraldsson, M.; Nash, R. J .; Fellows, L. E.
Tetrahedron Lett. 1988, 29, 5441-5444.
(19) Brandsma, L.; Verkuijsse, H. D. Synthesis of Acetylenes, Allenes,
and Cumulenes; Elsevier: Amsterdam, 1981; Vol. 8.
(20) Pirrung, M. C.; Webster, N. J . G. J . Org. Chem. 1987, 52, 3603-
3613.
(21) Marshall, J . A.; Shearer, B. G.; Crooks, S. L. J . Org. Chem.
1987, 52, 1236-1245.
(22) Lal, B.; Pramanick, B. M.; Manhas, M. S.; Bose, A. K.
Tetrahedron Lett. 1977, 1977-1980.

Syntheses of Australine and 7-Epialexine
J . Org. Chem., Vol. 65, No. 18, 2000 5787
Sch em e 2
Sch em e 3
hydrolysis of the vinyl sulfide cycloadducts to ketones,24
we incorporated this design modification into our sub-
strate. An attempted synthesis of 3-epiaustraline (34)6c,25
is shown in Scheme 3. The lactol 25 was synthesized from
D-arabinose by the literature procedure26 and then sub-
jected to a Wittig reaction to produce the alkene 26.
Formation of the triflate of 26 followed by displacement
with azide ion gave the alkene 27, which was best used
without manipulation because of its propensity to un-
dergo an intramolecular dipolar cycloaddition to give the
corresponding triazoline. Ozonolysis of 27 followed by
reductive workup gave the aldehyde 28, which was found
to decompose and/or epimerize upon attempted purifica-
tion. Thus, the crude aldehyde was subjected to olefina-
tion with the allylic borane reagent derived from the
hydroboration of the allene 29 by 9-BBN, producing the
diene 30 in 50% overall yield from the alkene 27 after
sodium hydroxide deoxysilylation of the intermediate
â-hydroxysilane. We had previously reported this method
for the stereocontrolled formation of either E- or Z-2-
arylthio- or alkylthio-1,3-butadienes, depending on
whether the intermediate â-hydroxysilane was decom-
posed by base or acid, respectively.24,27 Heating the azido
diene 30 in chloroform at 75 °C produced two cyclized
products in equal amounts. The first, isolated in 25%
yield (30% based on recovered 30), was found to be a
single stereoisomer of the desired pyrrolizidine 31. The
diastereoselectivity was found to be consistent with our
earlier studies on the cyclization of azido dienes bearing
allylic oxygenation.15 Elemental analysis of the other
cyclization product showed that it was isomeric with the
starting diene 30, still containing three nitrogen atoms.
upon the well-known deshielding of protons syn to
nitrogen lone pairs.23 Thus, the methine hydrogen at C(3)
in 22b appeared at δ 4.13, vs 3.43 for 22a . Our expecta-
tions about the stereoselectivity of this cyclization were
confirmed, although the level of selectivity was low.
Attempts to improve the yield and stereoselectivity of the
cyclization using different solvents and other alcohol
protecting groups (acetate, benzyl, tetrahydropyranyl)
were unsuccessful, generally resulting in lower yields
with similar stereoselectivities. Cyclization of O-benzyl
versions of 21 resulted in the formation of benzyl alcohol
by loss of the terminal benzyloxy group. Accepting 21 as
an optimized substrate for cyclization, we then proceeded
to examine the transformation of the major cycloadduct
into 3-epi analogues of alexine, e.g., 24. Desilylation of
22a afforded the diol 23, but all attempts to use the vinyl
sulfide functionality in 23, 22a , or 22b for the introduc-
tion of the remaining two hydroxyl groups of 24 failed.
Faced with the low selectivity and moderate yield of the
cycloaddition of 21 and the difficulty in transforming the
vinyl sulfide into the desired diol functionality, we
embarked on an alternative cycloaddition strategy in-
volving the assembly of the less hydroxylated ring of
these pyrrolizidines.
In tr am olecu lar Dipolar Cycloaddition of an Azide
w ith a Sim p lified Dien e. The studies outlined above
showed that a diene bearing a terminal alkoxymethyl
group was not an effective choice for the azide/diene
cycloaddition. Further, the vinyl sulfide produced in the
cycloadduct was difficult to use. We chose next to attempt
to generate the desired pyrrolizidines by construction of
the less-oxygenated ring by the cycloaddition of an azide
with a less complex diene, e.g., 30 in Scheme 3. Further,
since we had previously found that the use of a tert-
butylthio group instead of a phenylthio group allowed
(24) Bergmeier, S. C. Ph.D. Thesis, University of Michigan, 1991.
(25) Nash, R. J .; Fellows, L. E.; Plant, A. C.; Fleet, G. W. J .; Derome,
A. E.; Baird, P. D.; Hegarty, M. P.; Scofield, A. M. Tetrahedron 1988,
44, 5959-5964.
(26) (a) Barker, R.; Fletcher, H. G. J . Org. Chem. 1961, 26, 4605-
4609. (b) Tejima, S.; Fletcher, H. G. J . Org. Chem. 1963, 28, 2999-
3003.
(23) For studies on the deshielding effect in pyrrolizidine rings,
see: (a) Skvortsov, I. M.; Elvidge, J . A. J . Chem. Soc. B 1968, 1589-
1595. (b) J ones, T. H.; Blum, M. S. J . Org. Chem. 1980, 45, 4778-
4780.
(27) Pearson, W. H.; Lin, K.-C.; Poon, Y.-F. J . Org. Chem. 1989, 54,
5814-5819.

5788 J . Org. Chem., Vol. 65, No. 18, 2000
Pearson and Hines
Sch em e 4
a variety of reasons. Thus, we sought to use a more
conventional strategy to reach our objectives.
Red u ctive Dou ble-Cycliza tion of Azid oep oxid es.
Syn th esis of (+)-Au str a lin e (2) a n d (-)-7-Ep ia lexin e
(4). We and others have found that the reductive double-
cyclization of azides bearing an epoxide and one ad-
ditional electrophile (e.g., an alkylating or acylating
agent) is an effective way to generate the two rings of
the pyrrolizidine, indolizidine, and quinolizidine alka-
loids.17 For example, the reductive double-cyclization of
the azido epoxy tosylate 35 gave the indolizidine 36, a
precursor of slaframine, in good yield (eq 1).17d An
internal azide was used in a double cyclization involving
an epoxide and an acylating agent in the conversion of
37 to the indolizidine 38, a precursor of (3R)-3-(hy-
droxymethyl)swainsonine (eq 2).17k Ring expanded ana-
logues of alexine, australine, and swainsonine have also
been prepared by reductive double-cyclizations of azido
epoxides bearing internal acylating agents and alkylating
agents, respectively.7,17j This strategy has now proven
useful for the completion of the synthesis of australine
(2) and 7-epialexine (4), as shown in Scheme 4.
The IR spectrum showed no azide stretch. The single
vinylic hydrogen appeared as a clean doublet in the H
1
Our synthesis began with 2,3,5-tri-O-benzyl-L-xylo-
furanose 39,28 which was prepared from L-xylose in three
steps.26 Wittig olefination produced 40, which was con-
verted to the azide 41 via the triflate. The azide 41 was
relatively unstable, undergoing an intramolecular 1,3-
dipolar cycloaddition upon standing at room temperature.
Hence, it was best to carry 41 to the next step without
delay. Ozonolysis of 41 afforded the aldehyde 42, which
was not purified due to its sensitivity and was directly
converted to the Z-alkene 44 by a stereoselective Wittig
reaction with the silyloxy-substituted ylide 43.29-31 Again,
intramolecular 1,3-dipolar cycloaddition was observed if
this azidoalkene was allowed to stand at room temper-
ature. Epoxidation of 44 with m-CPBA produced a 1:1
mixture of epoxides 45R and 45â, which were not
separated.32 Tosylation of 45R/â gave 46R/â as a 2:1
mixture of isomers, since one of the isomers of 45
underwent tosylation faster than the other and the
reaction could not be driven to completion. Reduction of
46R/â without debenzylation was possible by hydro-
NMR spectrum, rather than the expected doublet of
doublets. These data led us to propose that dipolar
cycloaddition had taken place at the distal double bond
of the diene 30, producing the bicyclic triazoline 32. We
had previously found such materials in the cyclizations
of similar azido dienes bearing oxygen rather than sulfur
substituents.15 To avoid this problem, we attempted to
prepare the Z-isomer of 30 using the aforementioned
acidic workup of the allylboration/desilylation reaction.
Unfortunately, the acid sensitivity of these materials
thwarted these efforts. Nonetheless, we had the desired
cycloadduct 31 in hand. To complete a synthesis of
3-epiaustraline 34, we attempted to hydrolyze the vinyl
sulfide to the ketone 33. A survey of many different
hydrolysis conditions failed to produce more than a trace
of the desired ketone, leading mainly to decomposition.
Other strategies for transforming the vinyl sulfide into
the required functionality were also found to be unsuc-
cessful.
While the intramolecular cycloadditions of azides with
sulfur-substituted dienes was useful for the generation
of the late-stage pyrrolizidines 23 and 31 bearing the
crucial hydroxy- or alkoxymethyl substituent (Schemes
2 and 3), functional group manipulation issues and low-
yielding steps thwarted our efforts to complete the
syntheses of the target pyrrolizidines. We also studied
other diene substituents in an attempt to solve the
functional group problems (e.g., S-t-Bu, OCON-i-Pr2), but
these materials were found to be difficult to employ for
(28) MacCoss, M.; Chen, A.; Tolman, R. L. Tetrahedron Lett. 1985,
26, 4287-4290.
(29) Kunz, H. Leibigs Ann. Chem. 1973, 2001-2009.
(30) Salmond, W. G.; Barta, M. A.; Havens, J . L. J . Org. Chem. 1978,
43, 790-792.
(31) Takahashi, T.; Miyazawa, M.; Ueno, H.; Tsuji, J . Tetrahedron
Lett. 1986, 27, 3881-3884.
(32) For reports of similar epoxidations of allylic ethers, see: (a)
Wang, Z.; Schreiber, S. L. Tetrahedron Lett. 1990, 31, 31-34. (b)
Erickson, S. D.; Still, W. C. Tetrahedron Lett. 1990, 31, 4253-4256.
(c) Coutts, S. J .; Wittman, M. D.; Kallmerten, J . Tetrahedron Lett. 1990,
31, 4301-4304.

Syntheses of Australine and 7-Epialexine
J . Org. Chem., Vol. 65, No. 18, 2000 5789
genolysis. The resultant amine was not isolated but was
directly heated in refluxing EtOH containing K2CO3,
affording a 2:1 mixture of two pyrrolizidines 47a and 47b
which were easily separated by flash chromatography.
At this point, the configurations of 47a and 47b at C(7)
and C(7a) were unknown, but it was assumed that the
methine hydrogens at these positions were cis in both
isomers due to their cis relationship in alkene 44 and
epoxides 45 and 46. Hydrogenolytic debenzylation of 47b
using a larger amount of palladium catalyst afforded (-)-
7-epialexine 4, as expected. However, a similar deben-
zylation of 47a gave a pyrrolizidine that did not match
the literature data for (+)-australine 2 as expected, but
rather matched the literature data for (+)-7-epiaustraline
3.18,33 Since the structures of alexine 1, australine 2, and
7-epialexine 4 (three of the four possible diastereomers
at C-7 and C-7a) had been established by X-ray crystal-
lography3,18,34 and since our data for the debenzylation
product from 47a were not consistent with these struc-
tures, it appeared initially that this material was 7-epi-
australine 3.6a The literature assignment of 3 rested on
NOE experiments and comparison with known struc-
tures.33 The absolute stereochemistry was not deter-
mined, but was proposed to be as shown. The apparent
formation of 3 was unexpected, since the methine hy-
drogens at C-7 and C-7a were trans rather than cis.
Although we were confident that our chemistry would
give australine 2, the fact that our data did not match
the literature values for australine, the structure of which
was based on X-ray crystallography, led us to conclude
that an unusual epimerization had occurred at some
point. After extensive experimentation, we were not able
to find evidence for such an epimerization. Fortunately,
recent publications have allowed a resolution of this
problem. Denmark and co-workers synthesized 7-epiaus-
traline 3 and proved its structure by X-ray crystallog-
raphy.8b,c The properties of synthetic 3 did not match
those of natural 3, leading to a reevaluation of the
assignments of the various naturally occurring pyrroliz-
idines. It was determined by Wormald et al. that the
spectra for australine 2 were incorrectly reported in the
original work.35 We now find that our data for the
debenzylation product of 47a match the revised data for
(+)-australine 2. Thus, as we had originally intended, we
were successful in our attempt to prepare 2. In retrospect,
our work represents the first total synthesis of (+)-
australine.6
vinyl sulfide functionality was not possible. The novel
byproduct 32, the result of cyclization of the azide onto
the distal double bond of the diene, was formed in the
cyclization reaction. Finally, a successful route to (+)-
australine (2) and (-)-7-epiaustraline (4) was developed,
relying on the reductive double cyclization of the azido
epoxy tosylate 46 to the pyrrolizidines 47. Originally, it
was thought that an unusual epimerization had occurred
in the reductive cyclization, resulting in the eventual
production of 7-epiaustraline 3 rather than australine 2.
Due to recent work by Denmark, an error in the original
structural work on australine was found, allowing us to
now conclude that australine 2 had been synthesized as
was originally expected, resulting in what is, in retro-
spect, the first total synthesis of this alkaloid.
Exp er im en ta l Section
(E)-5-(ter t-Bu tyld ip h en ylsilyl)oxyp en t-3-en -1-yn e (15).
Imidazole (7.6 g, 112 mmol), tert-butyldiphenylsilyl chloride
(15.2 g, 55.4 mmol), and E-5-hydroxypent-3-en-1-yne 14 (4.14
g, 50.5 mmol),19 were stirred in dimethylformamide (90 mL)
at room temperature for 8 h. The mixture was then diluted
with ether and washed with saturated aqueous NH4Cl (150
mL). The aqueous wash was extracted with ether, and the
combined organic layers were washed with brine, dried over
Na2SO4 and concentrated in vacuo. Chromatography (5% ethyl
acetate/hexane) gave 13.28 g (82%) of the title compound as a
clear oil: Rf 0.62 (10% ethyl acetate/hexane); IR (neat) 3293
(s) cm-1; 1H NMR (360 MHz, CDCl3) δ 7.7-7.3 (m, 10 H), 6.28
(dt, 1 H, J ) 15.5, 3.5 Hz), 5.92 (app dq, 1 H, J ) 15.5, 1.8
Hz), 4.3-4.2 (m, 2 H), 2.85 (m, 1H), 1.2-1.0 (m, 9 H); 13C NMR
(90 MHz, CDCl3) δ 143.8, 135.45, 135.4, 129.8 127.8, 107.7,
82.1, 77.5, 63.4, 26.7, 19.2; MS (CI, CH4), m/e 321 (MH+, 48.9),
303 (10.6), 280 (14.9), 178 (100.0); HRMS (CI, CH4) Calcd for
C
C
21H24OSiH (MH+) 321.1675 found 321.1678. Anal. Calcd for
21H24OSi: C, 78.69; H, 7.55. Found: C, 78.71; H, 7.88.
(E)-1-(ter t-Bu t yld im et h yllsilyl)oxy-3-h yd r oxy-8-(ter t-
bu tyld ip h en ylsilyl)oxyoct-6-en -4-yn e (17). n-Butyllithium
(6.1 mL of a 2.5 M solution in hexane, 15.3 mmol) was added
in a dropwise fashion over 5 min to a solution of the enyne 15
(4.65 g, 14.5 mmol) in THF (200 mL) at -78 °C. After 10 min,
3-(tert-butyldimethylsilyl)oxypropanal 16 (3.28 g, 17.4 mmol)20
in THF (10 mL) was over a 50 min period. After an additional
40 min, saturated aqueous NH4Cl (2 mL) was added, and the
mixture was warmed to room temperature and poured into
saturated aqueous NH4Cl. The layers were separated, and the
aqueous layer was extracted three times with ether. The
organic layers were combined and washed with brine, dried
over Na2SO4, and concentrated in vacuo to give 7.3 g of crude
oil. Chromatography (10% ethyl acetate/hexane) gave 5.4 g
(73%) of the title compound as a clear oil: Rf 0.15 (10% ethyl
acetate/hexane); IR (neat) 3420 (br) cm-1; 1H NMR (360 MHz,
CDCl3) δ 7.7-7.4 (m, 10 H), 6.21 (dt, 1 H, J ) 15.5, 4.0 Hz),
5.94 (app dq, 1 H, J ) 15.5, 2.0 Hz), 4.75 (m, 1 H), 4.24 (dd, 2
H, J ) 4.0, 2.0 Hz), 4.05 (m, 1 H), 3.86 (m, 1 H), 3.45 (d, 1 H,
J ) 6.0 Hz, -OH), 1.06 (s, 9 H), 0.91 (s, 9 H), 0.10 (s, 3 H),
0.09 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 142.1, 135.6, 133.4,
129.8, 127.8, 108.5, 90.1, 83.3, 63.6, 62.3, 61.2, 38.9, 26.8, 25.9,
19.3, 18.2, -5.5; MS (CI, CH4), m/e 509 (M + H+, 5.1), 507 (M
- H+, 5.2), 493 (25.8), 491 (100.0); HRMS (CI, CH4) calcd for
Con clu sion
Three synthetic strategies for the assembly of 1,2,7-
trihydroxy-3-(hydroxymethyl)pyrrolizidines were exam-
ined. The cyclization of the azido diene 21 gave the
3-(hydroxymethyl)pyrrolizidine 22 in reasonable yield as
one major stereoisomer of four possibilities, although
further transformations to naturally occurring pyrroliz-
idines was not possible. The cyclization of the azido diene
30 produced the 3-(hydroxymethyl)pyrrolizidine 31 as a
single stereoisomer, but further transformation of the
C
30H43O3Si2 (M - H+) 507.2750. Found: 507.2751. Anal. Calcd
for C30H44O3Si2: C, 70.81; H, 8.72. Found: C, 70.50; H, 8.78.
(4Z,6E)-1-(ter t-Bu tyld im eth ylsilyl)oxy-8-(ter t-bu tyld i-
p h en ylsilyl)oxy-5-p h en ylth io-4,6-octa d ien e-3-ol (18). A
solution of the enyne 17 (4.5 g, 8.8 mmol) in ether (10 mL)
was added over 5 min dropwise to a solution of sodium bis(2-
methoxyethoxy)aluminum hydride (Red-Al) (4.21 g of a 70%
solution in toluene, 14.58 mmol) in ether (45 mL) at 0 °C. After
40 min, dry ethyl acetate (0.56 mL, 5.7 mmol) was added. After
5 min, the mixture was cooled to -78 °C and phenylsulfenyl
chloride (1.8 g, 12 mmol) in ether (10 mL) was added over 5
(33) Nash, R. J .; Fellows, L. E.; Dring, J . V.; Fleet, G. W. J .; Girdhar,
A.; Ramsden, N. G.; Peach, J . M.; Hegarty, M. P.; Scofield, A. M.
Phytochemistry 1990, 29, 111-114.
(34) Molyneux, R. J .; Benson, M.; Wong, R. Y.; Tropea, J . E.; Elbein,
A. D. J . Nat. Prod. 1988, 51, 1198-1206.
(35) Wormald, M. R.; Nash, R. J .; Hrnciar, P.; White, J . D.;
Molyneux, R. J .; Fleet, G. W. J . Tetrahedron: Asymmetry 1998, 9,
2549-2558.

5790 J . Org. Chem., Vol. 65, No. 18, 2000
Pearson and Hines
min. The solution was warmed to room temperature over 1 h
and then poured into saturated aqueous NH4Cl. The layers
were separated and the aqueous layer was extracted with ether
(6×). The combined organic layers were dried over Na2SO4 and
concentrated in vacuo to give 7.35 g crude oil. Chromatography
(10% ethyl acetate/hexane) gave 2.7 g (50%) of the title
compound as a pale yellow oil. In earlier runs, this compound
was found to decompose upon standing at room temperature
and was thus used immediately in the next step: Rf 0.29 (10%
r a h yd r o-1H-p yr r olizin e (22a ) a n d (1R*,5R*,7a R*)-1-(ter t-
Bu t yld ip h en ylsilyl)oxy-5-(ter t-b u t yld ip h en ylsilyl)oxy-
m e t h yl-7-p h e n ylt h io-2,3,5,7a -t e t r a h yd r o-1H -p yr r oli-
zin e (22b). The azide 21 (0.36 g, 0.487 mmol) in THF (4.8
mL) was degassed by four freeze-thaw cycles, and the glass
tube was sealed and heated behind a blast shield for 15 h at
90 °C. The vessel was then cooled to room temperature and
opened, and the contents were concentrated in vacuo. Chro-
matography (5% ethyl acetate/hexane to elute 22a and 25%
ethyl acetate/hexane to elute 22b) yielded 120 mg (35%) of 22a
and 51 mg (15%) of 22b (50% overall yield). Data for 22a : Rf
1
ethyl acetate/hexane); IR (neat) 3421(br) cm-1; H NMR (300
MHz, CDCl3) δ 7.8-7.0 (m, 15 H), 6.40 (bd, 1 H, J ) 15.1 Hz),
6.28 (d, 1 H, J ) 7.7 Hz), 6.10 (dt, 1 H, J ) 15.1, 4.2 Hz), 5.09
(m, 1 H), 4.19 (bd, 2 H, J ) 4.2 Hz), 3.9-3.8 (m, 2 H), 3.63 (d,
1 H, J ) 2.9 Hz, OH), 2.0-1.7 (m, 2 H), 1.02 (s, 9 H), 0.89 (s,
9 H), 0.07 (s, 6 H); 13C NMR (300 MHz, CDCl3) δ 143.7, 136.5,
135.8, 135.4, 134.8, 133.5, 133.4, 132.8, 131.3, 130.9, 129.6,
129.4, 128.8, 128.7, 127.64, 127.58, 127.53, 125.4, 70.3, 63.7,
62.1, 38.3, 26.8, 26.6, 25.8, 19.2, 18.1; MS (CI, CH4), m/e 635
(3.5), 601 (100.0).
(4Z,6E )-1-(t er t -Bu t yld im e t h ylsilyl)oxy-3,8-d i-[(t er t -
bu tyld ip h en ylsilyl)oxy]-5-p h en ylth io-4,6-octa d ien e (19).
The alcohol 18 (1.8 g, 2.9 mmol) was added to a mixture of
imidazole (0.43 g, 6.3 mmol) and tert-butyldimethylsilyl chlo-
ride (1.1 g, 3.8 mmol) in dimethylformamide (8 mL) at room
temperature. After 15 h, the mixture was diluted with ether
and washed with saturated aqueous NH4Cl. The aqueous
washes were then extracted with ether and the combined
organic layers were washed with brine, dried over Na2SO4 and
concentrated in vacuo to give 4.3 g of a crude oil. Chromatog-
raphy (5% ethyl acetate/hexane) gave 1.7 g (69%) of the title
compound: Rf 0.5 (10% ethyl acetate/hexane); IR (neat) 2929-
(s), 1472(s) cm-1; 1H NMR (360 MHz, CDCl3) δ 7.7-6.8 (m, 25
H), 6.27 (d, 2 H, J ) 7.5 Hz), 6.25 (bd, 1 H, J app) 14.5 Hz),
5.95 (dt, 1 H, J ) 14.5, 4.5 Hz), 5.19 (m, 1 H), 4.2-4.1 (m, 2
H), 3.8-3.5 (m, 2 H), 2.0-1.7 (m, 2 H), 1.12 (s, 9 H), 1.08 (s,
9 H), 0.89 (s, 9 H), 0.08 (s, 6 H); 13C NMR (300 MHz, CDCl3)
δ 144.0, 136.1, 135.6, 134.5, 133.7, 132.4, 129.7, 129.5, 129.3,
128.6, 127.7, 127.6, 127.4, 125.1, 69.6, 63.8, 59.6, 41.4, 27.1,
27.0, 26.9, 25.9, 19.4, 19.3, 18.3; MS (CI, CH4), m/e 857.4 (7.8,
M + H+), 856.4 (7.3, M+), 601.2 (100.0); HRMS (CI, CH4) calcd
for C52H68O3SSi3 (M+) 856.4197, found 856.4168.
(4Z,6E)-3,8-Di[(ter t-bu tyld ip h en ylsilyl)oxy]-5-p h en yl-
th io-4,6-octa d ien -1-ol (20). Diene 19 (1.32 g, 1.54 mmol) was
stirred in acetic acid/THF/water (20 mL of a 3:1:1 mixture) at
50 °C for 6h. The solution was then poured into a mixture of
ether and brine. The layers were separated, the aqueous layer
was extracted with ether, and the combined organic extracts
were washed with saturated aqueous NaHCO3 and brine, dried
over Na2SO4, and concentrated in vacuo. Chromatography (7%
ethyl acetate/hexane) gave 0.69 g (61%) of the title compound
as a thick oil. In earlier experiments, it was found that this
material decomposed upon standing at room temperature and
was thus used immediately in the next step: Rf 0.55 (25% ethyl
acetate/hexane); 1H NMR (300 MHz, CDCl3) δ 7.7-6.7 (m, 25
H), 6.26 (d, 1 H, J ) 14), 6.22 (d, 1 H, J ) 8.8 Hz), 5.90 (dt, 1
H, J ) 14, 6 Hz), 5.15 (m, 1 H), 4.2-4.0 (m, 2 H), 3.8-3.5 (m,
3 H, CH2 and OH), 2.0-1.7 (m, 2 H), 1.2-1.0 (m, 18 H).
(4Z,6E)-1-Azid o-3,8-d i[(ter t-bu tyld ip h en ylsilyl)oxy]-5-
p h en ylth io-4,6-octa d ien e (21). The alcohol 20 (0.61 g, 0.82
mmol) was added to a solution of triphenylphosphine (0.24 g,
0.92 mmol) in THF (8 mL) at 0 °C. Diethyl azodicarboxylate
(0.16 g, 0.92 mmol) was added in a dropwise fashion. After 5
min, diphenylphosphoryl azide (0.31 g, 1.1 mmol) was added
and the mixture was warmed to room temperature and stirred
for 48 h in the dark. The mixture was then concentrated in
vacuo and the resultant oil was purified by chromatography
(3% ethyl acetate/hexane) to give 0.5 g (71%) of the title
compound as an oil. Repeated attempts to obtain analytically
pure material were unsuccessful: Rf 0.69 (10% ethyl acetate/
hexane); 1H NMR (300 MHz, CDCl3) δ 7.7-6.8 (m, 25 H), 6.3-
6.1 (m, 2 H), 5.91 (dt, 1 H,J ) 15.2, 5.0 Hz), 5.03 (m, 1 H),
4.12 (bs, 2 H), 3.25 (m, 2 H), 1.9-1.6 (m, 2 H), 1.04 (s, 9H),
1.08 (s, 9 H).
1
0.6 (20% ethyl acetate/hexane); H NMR (360 MHz, CDCl3) δ
7.7-7.1 (m, 25 H), 5.47 (app t, 1 H, J ) 1.5 Hz), 4.29 (m, 1 H),
4.22 (m, 1 H), 3.69 (A), 3.50 (B) (AB of ABX, 2 H, J AB ) 8.5
Hz, J AX ) 6.0 Hz, J BX ) 16.0 Hz), 3.43 (m, 1 H, X of ABX),
3.25 (td, 1 H, J ) 10.5, 4.5 Hz), 2.68 (m, 1 H), 1.52 (m, 1 H),
1.39 (m, 1 H), 1.02 (s, 9 H), 0.98 (s, 9 H); 13C NMR (90 MHz,
CDCl3) δ 136.1, 135.9, 135.7, 131.8, 129.6, 129.1, 128.8, 127.6,
127.5, 127.3, 82.3, 75.5, 68.8, 54.2, 33.4, 26.9, 19.2, 19.1 Data
for 22b : Rf 0.55 (25% ethyl acetate/hexane); 1H NMR (360
MHz, CDCl3) δ 7.8-7.2 (m, 25 H), 5.12 (appt, 1 H, J ) 1.5
Hz), 4.39 (m, 1H), 4.32 (m, 1 H), 4.13 (m, 1 H, X of ABX), 3.66,
3.64 (ABof ABX, 2 H, J AB ) 10.0 Hz, J AX ) 4.5 Hz, J BX ) 5.5
Hz), 3.02 (t, 2 H, J ) 6.0 Hz), 1.66 (m, 1 H), 1.57 (m, 1 H),
1.05 (s, 9 H), 0.97 (s, 9 H). Proton NMR assignments were
made by COSY and NOE experiements. For both 22a /b, no
NOE was observed between the methine hydrogens at C-1 and
C-7a, implying a trans relationship between these two protons.
[In the earlier text, alexine/australine numbering was used
instead, i.e., the phenylthio group is at C(1), the silyloxymethyl
group at C(3), etc.] The relative configuration at C-5 vs C-7a
was assigned based on the chemical shifts of the C-5 methine
protons in 22a (δ 3.43) and 22b (δ 4.13), since methine
hydrogens syn to lone pairs on nitrogen appear further
downfield than those that are anti.23
(1R*,5S*,7a R*)-5-H yd r oxym et h yl-7-p h en ylt h io-2,3,5,-
7a -tetr a h yd r o-1H-p yr r olizin -1-ol (23). Tetra-n-butylam-
monium fluoride (320 µL of a 1 M solution in THF, 0.32 mmol)
was added to a solution of the tetrahydropyrrolizine 22a (40
mg, 0.054 mmol) in THF (2 mL) at room temperature. After
24 h, the mixture was concentrated in vacuo and the resultant
oil was purified by chromatography (95/4.5/0.5 CHCl3/MeOH/
NH4OH). The eluate was dissolved in CHCl3 and washed with
4% NaOH and concentrated to give 7.0 mg (50%) of the title
compound as a white solid, which was recrystallized from
ether: Rf 0.28 (90/9/1 CHCl3/MeOH/NH4OH); mp 97-102 °C;
1
IR (CHCl3) 3690 cm-1; H NMR (360 MHz, CDCl3) δ 7.5-7.3
(m, 5 H), 5.35 (appt, 1 H, J app) 2.0 Hz), 4.29 (ddd, 1 H, J )
8.0, 4.9, 2.5 Hz), 4.1 (m, 1 H), 3.61 (m, 1 H, X of ABX), 3.58
(A), 3.38 (B) (AB of ABX, 1 H, J AB ) 10.5 Hz, J AX ) 4.5 Hz,
J BX ) 5.5 Hz), 3.31 (ddd, 1 H, J ) 10.5, 9.0, 5.5 Hz), 2.80 (ddd,
1 H, J ) 11.5, 6.5, 4.5 Hz), 2.55 (bs, 2 H, OH), 1.90 (m, 1 H),
1.79 (m, 1 H); 13C NMR (90 MHz, CDCl3) δ 136.9, 132.7, 131.7,
129.4, 128.3, 125.7, 81.3, 75.8, 74.9, 64.7, 54.0, 33.4; MS (70
eV) m/e 263 (5), 232 (100); HRMS calcd for C14H17NO2S (M+)
263.0980, found 263.0966.
(3R,4R,5R)-5-Hydr oxy-3,4,6-tr iben zyloxy-1-h exen e (26).
n-Butyllithium (25.6 mL of a 1.87 M solution in hexanes, 49.6
mmol) was added to a solution of methyltriphenylphosphonium
bromide (16.3 g, 45.5 mmol) in THF (71 mL) at 0 °C. After 0.5
h, the solution was cooled to -78 °C for 15 min, and then 2,3,5-
tri-O-benzyl-D-arabinose 25 (7.8 g, 18 mmol)26 in THF (31 mL)
was added dropwise in a dropwise fashion. The resulting slurry
was warmed to -25 °C over 1 h and then to room temperature.
After 24 h, the reaction was quenched by the addition of water
(1.5 mL), and the mixture was poured into cold ether and
filtered. Concentration of the filtrate in vacuo followed by
chromatography of the residue (10% ethyl acetate/hexane) gave
6.16 g (80%) of the title compound as an oil: Rf 0.41 (25% ethyl
acetate/hexane); [R]25D -11.9 (c ) 1.26, CHCl3); IR (neat) 3480-
(br) cm-1 1H NMR (360 MHz, CDCl3) δ 7.4-7.2 (m, 15 H),
;
5.96 (ddd, 1 H J ) 16.7, 11.0, 7.5 Hz), 5.34 (bs, 1 H), 5.30 (bd,
1 H, J ) 7.5 Hz), 4.63, 4.35 (AB, 2 H, J AB ) 11.9 Hz), 4.62,
4.55 (AB, 2 H, J AB ) 11.4 Hz), 4.50 (s, 2 H), 4.07 (dd, 1 H, J )
7.5, 3.9 Hz), 4.01 (m, 1 H), 3.62 (dd, 1 H, J ) 6.9, 3.9 Hz), 3.59
(1R*,5S*,7a R*)-1-(ter t-Bu t yld ip h en ylsilyl)oxy-5-(ter t-
bu tyld ip h en ylsilyl)oxym eth yl-7-p h en ylth io-2,3,5,7a -tet-

Syntheses of Australine and 7-Epialexine
J . Org. Chem., Vol. 65, No. 18, 2000 5791
(bd, 2 H, J ) 4.5 Hz), 2.83 (d, 1 H, J ) 5.0 Hz, OH); 13C NMR
(90 MHz, CDCl3) δ 138.3, 138.2, 137.9, 135.2, 128.4, 128.3,
128.1, 127.8, 127.7, 127.6, 118.8, 80.7, 80.3, 74.1, 73.4, 71.0,
70.7, 70.4; MS (CI, CH4), m/e 436 (50.3, M + NH4+), 419 (49.1,
M + H+), 91 (100.0); HRMS (CI, CH4) calcd for C27H31O4 (M +
H+) 419.2222, found 419.2216. Anal. Calcd for C27H30O4: C,
77.48; H, 7.22. Found: C, 77.33; H, 7.12.
was allowed to warm to room temperature and treated with
saturated aqueous ammonium chloride. The mixture was
extracted twice with petroleum ether, and the organic extracts
were combined, dried (MgSO4), and concentrated. Distillation
gave 8.4 g (64%) of the title compound, 89% pure by GLC
analysis: bp 69-72 °C at 4 mmHg; IR (neat) 2960 (s), 1920
(s) cm-1 1H NMR (CDCl3, 300 MHz) δ 4.58 (s, 2H), 1.42 (s,
;
(3R,4S,5R)-5-Tr iflu or om eth an esu lfon yloxy-3,4,6-tr iben -
zyloxy-1-h exen e. Trifluoromethanesulfonic anhydride (1.5
mL, 2.5 g, 8.9 mmol) in CH2Cl2 (9 mL) was added to a solution
of pyridine (0.8 mL, 9.9 mmol) in CH2Cl2 (9 mL) at -40 °C.
After 10 min, 26 in CH2Cl2 (39 mL) was added over 5 min.
The slurry was warmed slowly to room temperature over 30
min and stirred at 0 °C for 2.5 h. The resulting mixture was
filtered through Celite, washing with cold CH2Cl2, and the
filtrate was concentrated in vacuo to give 5.6 g of the title
compound as a crude pale yellow oil. Because of its instability,
this crude material was used immediately without further
purification for the synthesis of 27. For characterization
purposes, a small portion was purified by chromatography
(10% ethyl acetate/hexane): Rf 0.29 (10% ethyl acetate/
hexane); IR (CHCl3) 1410(s) cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.4-7.2 (m, 15 H), 5.80 (ddd, 1 H, J ) 17.4, 10.3, 7.4 Hz),
5.32 (m, 2 H), 5.22 (m, 1 H), 4.70, 4.65 (AB, 2 H, J AB ) 11.3
Hz), 4.56, 4.34 (AB, 2 H, J AB ) 11.6 Hz), 4.47 (bs, 2 H), 3.95-
3.91 (m, 2 H), 3.83 (bd, 2 H, J ) 4.9 Hz); 13C NMR (75 MHz,
CDCl3) δ 137.7, 137.4, 137.3, 134.1, 128.4, 128.3, 128.0, 127.9,
127.85, 127.8, 127.7, 120.2, 88.6, 81.2, 80.5, 75.2, 73.4, 70.8,
67.6.
9H), 0.15 (s, 6H); 13C NMR (CDCl3, 75 MHz) δ 208.0, 89.9,
72.5, 47.1, 30.1, -1.7.
(3E,5R,6R,7S)-7-Azid o-5,6,8-tr iben zyloxy-3-ter t-bu tyl-
th io-1,3-octa d ien e (30). Allene 29 (1.05 g, 5.20 mmol) and
9-borabicyclo[3.3.1]nonane (10.5 mL of a 0.5 M solution in
THF, 5.3 mmol) were combined and heated at 70 °C for 24 h.
The solution was then cooled to room temperature and the
entire crude aldehyde 28 from the above experiment (5.1 mmol
theoretically) in THF (5 mL) was added via a cannula. The
mixture was stirred for 20 h, then 20% aqueous NaOH (25
drops) was added. After 48 h, the mixture was poured into
ether/petroleum ether (1:1, 300 mL), the resultant mixture was
filtered through Celite, and the filtrate was concentrated in
vacuo. Chromatography (2.5% ethyl acetate/hexane) gave 1.4
g (50% from 27) of the title compound: Rf 0.4 (10% ethyl
acetate/hexane); [R]25 -8.86 (c ) 1.14, abs EtOH); IR (neat)
D
2096(s) cm-1; 1H NMR (300 MHz, CDCl3) δ 7.4-7.1 (m, 15 H),
6.73 (dd, 1 H, J ) 16.6, 10.3 Hz), 6.06 (bd, 1 H, J ) 9.6 Hz),
5.97 (dd, 1 H, J ) 16.6, 1.8 Hz), 5.34 (dt, 1 H, J ) 10.3, 1.8
Hz), 4.86 (d, 1 H, J ) 11.4 Hz), 4.7-4.6 (m, 3 H), 4.39 (s, 2 H),
4.38 (d, 1 H, J ) 11.4 Hz), 3.65 (dd, 1 H, J ) 9.6, 2.9 Hz),
3.6-3.4 (m, 3 H), 1.42 (s, 9 H); 13C NMR (75 MHz, CDCl3) δ
139.7, 138.1, 137.9, 137.6, 136.8, 133.0, 128.4, 128.3, 128.2,
128.1, 127.7, 127.6, 121.3, 80.2, 75.3, 73.3, 71.1, 69.6, 61.4, 46.4,
(3R,4R,5S)-5-Azido-3,4,6-tr iben zyloxy-1-h exen e (27). Tet-
ra-n-butylammonium azide (10.1 g, 35.5 mmol) was dissolved
in benzene (65 mL) and the solution was cooled to -10 °C with
rapid strirring. A solution of the crude triflate from above in
benzene (30 mL) was added and the mixture was warmed to
0 °C over 30 min, then warmed to room temperature for 1 h
and concentrated in vacuo to approximately 10 mL. The
resulting red-orange oil was extracted with ether and the ether
extracts were concentrated in vacuo to give 5.8 g of a yellow
oil. This procedure separated the product from most of the
tetra-n-butylammonium azide. Chromatography (6% ethyl
acetate/hexane) gave 2.27 g (71% from 26) of the title com-
pound. This material proved to be unstable at room temper-
ature, and was therefore used as soon as possible for the
synthesis of 28: Rf 0.24 (10% ethyl acetate/hexane); IR (neat)
31.7; MS (CI, NH4), m/e 575 ((M + NH4)+, 25), 530 (100), 91
+
(45); HRMS (CI, NH4) calcd for C33H39N3O3SNH4 (M + NH4
575.3056, found 575.3052.
)
(1R,2R,3S,7a S)-1,2-Diben zyloxy-3-ben zyloxym eth yl-7-
ter t-bu tylth io-2,3,5,7a-tetr ah ydr o-1H-pyr r olizin e (31) an d
(6R,7R,8S)-6,7-Diben zyloxy-8-ben zyloxym eth yl-4-ter t-bu -
t ylt h io-3a ,6,7,8-t e t r a h yd r o-3H -[1,2,3]t r ia zolo[1,5-a ]-
a zep in e (32). A solution of the azidodiene 30 (0.5 g, 0.9 mmol)
in CHCl3 (4 mL) was degassed with 5 freeze/thaw cycles and
sealed in a thick-walled glass reaction tube fitted with a Teflon
sealing valve. The solution was then heated at 75 °C for 18 h
behind a blast shield and then cooled, removed from the tube,
and concentrated in vacuo. Chromatography (1.5% MeOH/
CHCl3) gave 120 mg of 31 (35%; 30% based on recovered
starting material), 124.6 mg of 32 (25%; 31% based on
recovered starting material), and 76 mg (15%) of starting
azidodiene 30. For 31: Rf 0.43 (4% MeOH/CHCl3); [R]25D +1.74
(c ) 1.18, abs EtOH), -5.3 (c 1.6 CHCl3); IR (CHCl3) 1496(s)
2096(s), cm-1
H), 5.82 (ddd, 1 H, J ) 17.4, 10.3, 7.8 Hz), 5.5-5.3 (m, 2 H),
4.84, 4.63 (AB, 2 H, J AB) 11.5 Hz), 4.59, 4.33 (AB, 2 H, J
;
1H NMR (360 MHz, CDCl3) δ 7.4-7.1 (m, 15
)
AB
11.7 Hz), 4.41 (s, 2 H), 4.08 (bt, 1 H, J ) 7.6 Hz), 3.7-3.6 (m,
2 H), 3.6-3.4 (m, 2 H); 13C NMR (75 MHz, CDCl3) δ 138.4,
137.9, 135.1, 128.5, 128.4, 128.3, 128.1, 127.9, 127.8, 127.7,
119.5, 82.1, 80.7, 74.9, 73.5, 70.9, 69.8, 61.7.
1
cm-1; H NMR (360 MHz, CHCl3) δ 7.5-7.1 (m, 15 H), 5.85
(bs, 1 H), 4.68, 4.57 (AB, 2 H, J AB ) 11.9 Hz), 4.60, 4.56 (AB,
2 H, J AB ) 12.1 Hz), 4.45, 4.28 (AB, 2 H, J AB ) 11.9 Hz), 4.24
(m, 1 H), 4.10 (ddd, 1 H, J ) 14.9, 5.7, 2.1 Hz), 4.0-3.93 (m,
2 H), 3.87 (A), 3.84 (B), (AB of ABX system, 2 H, J AB ) 9.5,
J AX ) 7.3, J BX ) 6.4 Hz), 3.7-3.6 (m, 2 H), 1.42 (s, 9 H); 13C
NMR (90 MHz, CHCl3) δ 138.3, 138.2, 133.9, 131.7, 128.3,
128.2, 127.8, 127.7, 127.6, 127.5, 127.4, 127.2, 85.6, 84.9, 82.2,
73.3, 71.4, 71.2, 66.1, 64.2, 55.3, 46.2, 31.5; MS (70 eV), m/e
529 (M+, 38), 91 (100); HRMS calcd for C33H39NO3S (M+)
529.2650, found 529.2626. The structural assignment of 31 was
based on COSY and NOESY spectroscopic studies. No NOE
was detected between the methine hydrogens at C-7 and C-7a,
indicating a trans relationship. For 32: Rf 0.75 (4% MeOH/
(2S,3R,4S)-4-Azid o-2,3,5-tr iben zyloxy-1-p en ta n a l (28).
Azidoalkene 27 (2.3 g, 5.1 mmol) in CH2Cl2 (27 mL) and
methanol (5.3 mL) was cooled to -78 °C, and one drop of
Sudan III (0.1% in CH2Cl2) was added. The mixture was
purged with argon, and then ozone was bubbled into the cooled
mixture for 17 min until the pink color disappeared. The
reaction mixture was then purged with argon for 15 min and
treated with dimethyl sulfide (0.93 g, 14.9 mmol). After 45 min
at -78 °C, 1 h at 0 °C, and 1.5 h room temperature, the yellow
solution was concentrated in vacuo to give 2.3 g of a golden
1
brown oil which was found to be pure aldehyde by H NMR.
This material was unstable to purification and was thus used
immediately in the following reaction: Rf 0.3 (25% ethyl
CHCl3); [R]25D -35.7 (c ) 1.56, CHCl3); IR (neat) 1454(s) cm-1
;
1
acetate/hexane); IR (neat) 2866(s), 2099(s), 1729(s) cm-1; H
1H NMR (300 MHz, CDCl3) δ 7.4-7.1 (m, 15 H), 6.20 (dd, 1 H,
J ) 6.0, 2.5 Hz), 5.05 (ddd, 1 H,J ) 8.0, 6.0, 4.5 Hz), 4.7-4.5
(m, 7 H), 4.5-4.2 (m, 3 H), 4.15 (dd, 1 H, J ) 6.5, 4.5 Hz),
NMR (300 MHz, CDCl3) δ 9.74 (d, 1 H, J ) 1.0 Hz), 7.4-7.1
(m, 15 H), 4.72, 4.45 (AB, 2 H, J AB ) 11.8 Hz), 4.57 (s, 2 H),
4.44 (s, 2 H), 4.0-3.7 (M, 2 H), 3.6-3.4 (m, 3 H); 13C NMR (75
MHz, CDCl3) δ 201.3, 137.6, 137.2, 136.9, 128.5, 128.4, 128.2,
128.1, 128.0, 127.8, 81.9, 78.7, 74.2, 73.46, 73.41, 69.1, 60.8.
1-Tr im eth ylsilyl-1-ter t-bu tylth io-1,2-p r op a d ien e (29).24
A solution of lithium diisopropylamide (66 mmol) in THF (300
mL) was cooled to -78 °C and treated with 1-tert-butylthi-
opropyne (8.37 g, 65.3 mmol).19 After 30 min, chlorotrimeth-
ylsilane (7.17 g, 66 mmol) was added rapidly, and the mixture
3.84 (A), 3.78 (B) (AB of ABX, 2 H, J AB) 9.0, J Ax) 8.5, J BX
)
6.0 Hz), 1.16 (s, 9 H); 13C NMR (90 MHz, CDCl3) δ 138.9, 138.2,
137.9, 128.5, 128.4, 128.1, 127.9, 127.8, 127.7, 75.4, 73.8, 73.4,
71.96, 71.90, 67.7, 59.0, 57.3, 47.2, 31.3; MS (70 eV), m/e 529
(M+, 6), 91 (100). Anal. Calcd for C33H39N3O3S: C, 70.31; H,
6.60; N, 7.62. Found: C, 70.68; H, 6.62; N, 7.25. The structural
assignment of 32 was based on elemental analysis (three
nitrogens), the absence of an azide stretch in the IR spectrum,

5792 J . Org. Chem., Vol. 65, No. 18, 2000
Pearson and Hines
and the fact that COSY and decoupling studies showed that
the vinyl hydrogen appeared as a clean doublet, coupled only
to the methine hydrogen at C-6. The configuration at C-3a was
not assigned, but only one stereoisomer was observed by 1H
NMR.
(300 MHz, CDCl3) δ 9.65 (d, 1 H, J ) 1 Hz), 7.4-7.0 (m, 15
H), 4.8-4.4 (m, 6 H), 4.01 (m, 1 H), 3.9-3.4 (m, 2 H), 3.5-3.3
(m, 2 H).
(2R,3R,4R,5Z)-2-Azido-8-h ydr oxy-1,3,4-tr iben zyloxyoct-
5-en e (44). A solution of 3-hydroxypropyltriphenylphos-
phonium bromide (2.31 g, 5.76 mmol)29-31 in THF (30 mL) was
cooled to 0 °C and treated with potassium hexamethyldisi-
lazide (23 mL of a 0.5 M solution in toluene, 11.5 mmol). The
solution was warmed to room temperature over 1 h, kept at
that temperature for another 1 h, and then cooled to 0 °C and
treated with freshly distilled chlorotrimethylsilane (0.74 mL,
5.83 mmol). After 10 min, the solution of the ylide 43 was
cooled to -78 °C and treated with a solution of the entire
quantity of aldehyde 42 prepared above (2.5 g, 5.4 mmol
theoretical) in THF (10 mL). After 1 h, the mixture was
warmed to room temperature. After 1 h, 1 M HCl (20 mL) was
added, and the mixture was concentrated in vacuo. The residue
was then taken up in THF (50 mL) and 1 M HCl (10 mL) and
stirred vigorously for 1 h. The layers were then separated and
the aqueous layer was extracted with ethyl acetate. The
combined organic extracts were dried over MgSO4 and con-
centrated in vacuo. Chromatography (25% ethyl acetate/
hexane) gave 0.92 g (35% from the alkene 41) of the title
(3R,4R,5S)-5-Hyd r oxy-3,4,6-tr iben zyloxy-1-h exen e (40).
n-Butyllithium (9.5 mL of 2.5 M solution in hexane, 23.7 mmol)
was added to a solution of methyltriphenylphosphonium
bromide (7.8 g, 22 mmol) in THF (50 mL) at 0 °C. After 15
min, the solution was cooled to -78 °C, and 2,3,5-tri-O-benzyl-
L-xylose 39 (4.14 g, 9.90 mmol) in THF (20 mL) was added in
a dropwise fashion. (The procedure used for the synthesis of
2,3,5-tri-O-benzyl-L-xylofuranose28 was the same as that re-
ported for the synthesis of 2,3,5-tri-O-benzyl-D-arabinofuranose
used in the synthesis of 26.26) The resultant slurry was
warmed slowly to room temperature over 3 h. After 24 h, water
(0.5 mL) was added and the mixture was poured into cold ether
and filtered, washing with cold ether. The combined filtrates
were concentrated in vacuo. Chromatography (10% ethyl
acetate/hexane) gave 2.72 g (66%) of the title compound: Rf
0.34 (25% ethyl acetate/hexane); [R]25 +5.8 (c ) 1.6, CHCl3);
D
IR (neat) 3449(br) cm-1 1H NMR (360 MHz, CDCl3) δ 7.4-
;
7.1 (m, 15 H), 5.86 (m, 1 H), 5.4-5.3 (m, 2 H), 4.85, 4.55 (AB
q, 2 H, J ) 11.2 Hz), 4.62, 4.37 (AB q, 2 H, J ) 11.8 Hz), 4.46,
4.41 (AB q, 2 H, J ) 11.9 Hz), 4.10 (appt, 1 H, J ) 7.2 Hz),
3.93 (m, 1 H), 3.61 (dd, 1 H, J ) 6.5, 2.7 Hz), 3.45 (A), 3.41 (B)
(ABq of ABX system, 2 H, J AB ) 9.4, J AX ) 6.0, J BX ) 6.3 Hz),
2.45 (d, 1 H, J ) 6.9 Hz, OH); 13C NMR (90 MHz, CDCl3) δ
138.4, 138.3, 138.1, 135.3, 128.3, 128.26, 128.24, 128.1, 127.79,
127.77, 127.6, 127.5, 119.1, 82.1, 80.4, 74.9, 73.3, 71.3, 70.7,
69.9; MS (CI, NH3), m/e 436 (M + NH4+, 100), 419 (M + H+,
2). Anal. Calcd for C27H30O4: C, 77.48; H, 7.22. Found: C,
77.59; H, 6.96.
compound: Rf 0.17 (25% ethyl acetate/hexane); [R]25 -10.8
D
(c ) 1.6, CDCl3); IR (neat) 2097(s) cm-1; H NMR (300 MHz,
1
CDCl3) δ 7.5-7.2 (m, 15 H), 5.8-5.5 (m, 2 H), 4.75, 4.60 (AB
q, 2 H, J ) 11.2 Hz), 4.66, 4.39 (AB q, 2 H, J ) 11.9 Hz), 4.57,
4.52 (AB q, 2 H, J ) 12.4 Hz), 4.40 (m, 1 H), 3.9-3.8 (m, 2 H),
3.70 (m, 1 H), 3.65-3.50 (m, 3 H), 2.3-2.1 (m, 2 H), 1.98 (bs,
1 H, OH); 13C NMR (75 MHz, CDCl3) δ 138.3, 137.9, 137.8,
131.5, 129.6, 128.3, 128.2, 127.9, 127.8, 127.6, 81.3, 75.0, 74.5,
73.4, 70.3, 69.7, 61.8, 61.6, 31.4; MS (CI, NH3), m/e 460 ((M -
N2)+, 100); HRMS (CI, NH3) calcd for C29H33NO4H ((M - N2)
+ H)+) 460.2488, found 460.2466. Decoupling of the methylene
protons at δ 2.3-2.1 allowed the determination of the coupling
constant between the vinyl protons as J ) 11.5 Hz, indicating
(3R,4R,5R)-5-Azido-3,4,6-tr iben zyloxy-1-h exen e (41). Tri-
fluoromethanesulfonic anhydride (1.5 mL, 2.5 g, 8.9 mmol) in
CH2Cl2 (9 mL) was added to a solution of pyridine (0.85 mL,
9.9 mmol) in CH2Cl2 (9 mL) at -40 °C. After 10 min, the
alcohol 40 (3.03 g, 7.24 mmol) in CH2Cl2 (39 mL) was added
over 5 min. The resultant slurry was warmed slowly to room
temperature over 30 min cooled to 0 °C. After 2.5 h, the
mixture was filtered through Celite, washing with cold CH2-
Cl2, and the filtrate was concentrated in vacuo to give the crude
triflate. This material was diluted with benzene (65 mL) and
added to a cold (0 °C) solution of tetra-n-butylammonium azide
(10.2 g, 36 mmol) in benzene (35 mL). The mixture was then
warmed to room temperature for 1 h and then concentrated
in vacuo to approximately 10 mL. The resulting red-orange
oil was extracted with ether. Without concentration, the
combined ethereal extracts were passed through a short
column of silica gel and concentrated without warming to give
2.4 g (75%) of the title compound. Due to the instability of this
material to cyclization, it was used immediately in the next
1
a cis relationship. No trans isomer was detected by H NMR.
(2R,3R,4S,5S,6S)-2-Azid o-8-h yd r oxy-1,3,4-tr iben zyloxy-
5,6-ep oxyoct a n e (45r) a n d (2R,3R,4S,5R,6R)-2-Azid o-8-
h yd r oxy-1,3,4-tr iben zyloxy-5,6-ep oxyocta n e (45â). A so-
lution of alkene 44 (0.61 g, 1.3 mmol) in CH2Cl2 (5 mL) was
cooled to 0 °C and treated with m-chloroperbenzoic acid (0.33
g, 1.9 mmol). After being warmed to room temperature for 24
h, the mixture was diluted with CH2Cl2 and washed with 1 M
NaOH and water. The organic layer was dried over MgSO4
and concentrated. Chromatography (40% ethyl acetate/hexane)
gave 0.41 g (65%) of an inseparable 1:1 mixture of 45R and
45â as determined by 1H NMR and HPLC: Rf 0.13 (25% ethyl
acetate/hexane); [R]25D -36.9 (c ) 2.3, CHCl3); IR (CHCl3) 3481-
1
(br), 2099(s) cm-1; H NMR (360 MHz, CDCl3) δ 7.4-7.1 (m,
15 H), 4.87, 4.48 (AB q, 1 H, J ) 11.5 Hz), 4.76 (1/2 of AB q,
0.5 H, J ) 11 Hz), 4.7-4.4 (m, 5.5 H), 4.0-3.8 (m, 3 H), 3.8-
3.7 (m, 3 H), 3.63 (dd, 0.5 H, J ) 8, 2 Hz), 3.55 (dd, 0.5 H, J
) 8, 2.5 Hz), 3.41 (dd, 0.5 H, J ) 8, 2 Hz), 3.3-3.2 (m, 1 H),
3.17 (dd, 0.5 H, J ) 8, 4.5 Hz), 2.69 (app q, 0.5 H, J ) 4 Hz),
2.0-1.4 (m, 4 H); 13C NMR (90 MHz, CDCl3) δ 138.0, 137.9,
137.8, 137.4, 137.3, 128.6, 128.5, 128.47, 128.41, 128.33,
128.30, 128.1, 128.0, 127.9, 127.8, 127.7, 79.3, 77.8, 75.0, 74.7,
74.1, 73.5, 73.4, 72.5, 72.3, 69.7, 69.6, 61.0, 60.9, 60.5, 57.6,
56.4, 55.1, 52.8, 31.3, 31.2; MS (CI, NH3), m/e 521 ((M + NH4)+,
100); HRMS (CI, NH3) calcd for C29H33N3O5NH4 (M + NH4)+
521.2763, found 521.2747.
step: Rf 0.25 (10% ethyl acetat/hexane); IR (neat) 2864(s) cm-1
;
1H NMR (300 MHz, CDCl3) δ 5.97 (m, 1 H), 5.5-5.3 (m, 2 H),
4.77, 4.64 (AB q, 2 H, J ) 11.2 Hz), 4.61, 4.58 (AB q, 2 H, J )
9.4 Hz), 4.71, 4.45 (ABq, 2 H, J ) 11.8 Hz), 4.21 (m, 1 H), 4.08
(m, 1 H), 3.9-3.8 (m, 2 H), 3.8-3.4 (m, 2 H); 13C NMR (75
MHz, CDCl3) δ 138.2, 138.0, 137.9, 135.3, 128.6, 128.4, 128.3,
128.1, 128.0, 127.7, 119.2, 81.1, 80.6, 75.0, 73.3, 70.7, 69.4, 61.5.
(2S,3R,4R)-4-Azid o-2,3,5-tr iben zyloxy-1-p en ta n a l (42).
Azidoalkene 41 (2.4 g, 5.4 mmol) in CH2Cl2 (30 mL) and
methanol (5.6 mL) was cooled to -78 °C and one drop of Sudan
III (0.1% in CH2Cl2) was added. The mixture was purged with
argon and then ozone was bubbled in for 17 min until the pink
color disappeared. After the mixture was purged with with
argon for 15 min, dimethyl sulfide was (1.0 g, 16 mmol) added.
After 45 min at -78 °C, 1 h at 0 °C, and 1.5 h at room
temperature, the yellow solution was concentrated in vacuo
to give 2.5 g golden brown oil which was found to be nearly
pure aldehyde by proton NMR. Due to its instability to
purification procedures, the entire productc was used im-
mediately in the next reaction: Rf 0.53 (30% ethyl acetate/
(2R,3R,4S,5S,6S)-2-Azido-8-(p-tolu en esu lfon yl)oxy-1,3,4-
tr iben zyloxy-5,6-epoxyoctan e (46r) an d (2R,3R,4S,5R,6R)-
2-Azid o-8-(p-tolu en esu lfon yl)oxy-1,3,4-tr iben zyloxy-5,6-
ep oxyocta n e (46â). A solution of the epoxides 45R and 45â
(0.34 g, 0.68 mmol) in CH2Cl2 (2 mL) was cooled to -15 °C
and treated with pyridine (0.17 mL, 2.1 mmol), (dimethylami-
no)pyridine (10 mg, 0.08 mmol), and p-toluenesulfonyl chloride
(0.27 g, 1.4 mmol). After 24 h at -15 °C and 24 h at -10 °C,
the solution was diluted with ether (150 mL) and washed with
water, 1 M HCl, saturated NaHCO3, and water. The organic
phase was dried over MgSO4 and concentrated in vacuo.
Chromatography (20% ethyl acetate/hexane) gave 40 mg (12%)
of starting material and 0.3 g (67%, 77% based on recovered
hexane); IR (neat) 2864(s), 2098(s), 1729(m) cm-1 1H NMR
;

Syntheses of Australine and 7-Epialexine
J . Org. Chem., Vol. 65, No. 18, 2000 5793
starting material) of an inseparable 2:1 mixture of 46R and
46â as determined by 1H NMR and HPLC: Rf 0.27 (20% ethyl
acetate/hexane); [R]25D -39.5 (c ) 1.1, CHCl3); IR (CHCl3) 2099-
(s), 1656(s) cm-1; 1H NMR (300 MHz, CDCl3) δ 7.78 (d, 2 H, J
) 8 Hz), 7.5-7.1 (m, 17 H), 4.82, 4.43 (AB q, 1.4 H, J ) 11.5
Hz), 4.7-4.5 (m, 4.6 H), 4.2-4.0 (m, 2 H), 3.9-3.6 (m, 4 H),
3.51 (d, 0.66 H, J ) 5 Hz), 3.45 (d, 0.33 H, J ) 5 Hz), 2.29 (d,
0.66 H, J ) 5 Hz), 3.20 (m, 0.33 H), 3.1-3.0 (m, 1 H), 2.52 (m,
0.66 H), 2.43 (s, 3 H), 2.0-1.4 (m, 2 H); 13C NMR (90 MHz,
CDCl3) δ 144.9, 137.9, 137.8, 137.2, 129.8, 128.6, 128.5, 128.47,
128.41, 128.3, 128.2, 128.0, 127.9, 127.8, 127.79, 127.74, 79.1,
77.7, 75.0, 74.5, 74.0, 73.5, 73.4, 72.7, 72.2, 69.6, 69.5, 67.5,
67.4, 61.0, 60.9, 57.8, 55.5, 54.8, 51.0, 28.6, 28.3, 21.5; MS (CI,
NH3), m/e 675 ((M + NH4+, 31), 246 (100); HRMS (CI, NH3)
calcd for C36H39N3O7SNH4 (M + NH4+) 675.2852, found
675.2842.
137.2, 128.7, 128.4, 128.3, 128.2, 127.9, 127.8, 127.7, 127.6,
85.8, 84.7, 73.3, 73.1, 72.5, 71.8, 68.9, 68.2, 62.9, 46.5, 35.7;
MS (CI, NH3), m/e 460 ((M + H)+, 100), 91 (30); HRMS (CI,
NH3) calcd for C29H33NO4H (M + H+) 460.2488, found 460.2490.
(1R,2R,3R,7S,7aR)-3-Hydr oxym eth yl-2,3,5,6,7,7a-h exah y-
d r o-1H-p yr r olizin e-1,2,7-tr iol [2, (+)-Au str a lin e]. Pyrroliz-
ine 47a (12 mg, 0.03 mmol) in EtOH (1 mL) was stirred
vigorously with 10% Pd on carbon (30 mg, 300 wt %) under a
balloon of hydrogen for 48 h at room temperature. The
hydrogen was evacuated and replaced by nitrogen, the mixture
was filtered through Celite, and the filtrate was concentrated
in vacuo to give 4.3 mg (88%) of the title compound, which
was found to be pure by 1H NMR: [R]25D +8.0 (c ) 0.35, H2O);
1H NMR (360 MHz, D2O) δ 4.21 (m, 1 H), 4.08 (t, 1 H, J ) 7.5
Hz), 3.73 (t, 1 H, J ) 8.5 Hz), 3.63 (dd, 1 H, J ) 11, 3 Hz),
3.45 (dd, 1 H, J ) 11, 6.5 Hz), 3.1-2.9 (m, 2 H), 2.7-2.5 (m,
2 H), 1.9-1.7 (m, 2 H); 13C NMR [90 MHz, D2O with dioxane
as an internal standard (δ 67.4)] δ 79.5, 73.9, 71.7, 71.4, 70.3,
63.1, 52.7, 35.9. These values were in good agreement with
literature data.6b,c
(1R,2R,3R,7S,7aR)-1,2-Diben zyloxy-3-ben zyloxym eth yl-
7-h ydr oxy-2,3,5,6,7,7a-h exah ydr o-1H-pyr r olizin e (47a) an d
(1R,2R,3R,7R,7a S)-1,2-Diben zyloxy-3-ben zyloxym eth yl-
7-h yd r oxy-2,3,5,6,7,7a -h exa h yd r o-1H-p yr r olizin e (47b). A
mixture of the epoxides 46R and 46â (240 mg, 0.36 mmol),
ether (2 mL), and EtOH (1 mL) was stirred with 10%
palladium on carbon (17 mg, 5 wt %) under a balloon of
hydrogen for 15 h at room temperature. The hydrogen was
evacuated and replaced by nitrogen, and then the catalyst was
removed by filtration through a short column of Celite. The
filtrate was diluted with EtOH (35 mL), and K2CO3 (0.30 g,
2.2 mmol) was added. The mixture was heated at reflux for
20 h and then filtered and concentrated. This residue was
taken up in CHCl3 and filtered. After concentration of the
filtrate, chromatography (5% MeOH/CHCl3) of the residue gave
116 mg (71%) of a 2:1 mixture of 47a and 47b. A second
chromatographic separation (20% i-PrOH/hexane followed by
7% MeOH/CHCl3) give 69 mg of 47a and 29 mg 47b as thick
(1R,2R,3R,7R,7aS)-3-Hydr oxym eth yl-2,3,5,6,7,7a-h exah y-
d r o-1H-p yr r olizin e-1,2,7-tr iol [4, (-)-7-Ep ia lexin e]. Pyr-
rolizine 47b (12 mg, 0.03 mmol) in EtOH (1.2 mL) was stirred
vigorously with 10% Pd on carbon (32 mg, 300 wt %) under a
balloon of hydrogen for 48 h at room temperature. The
hydrogen was evacuated and replaced by nitrogen, the mixture
was filtered through Celite, and the filtrate was concentrated
in vacuo to give 4.4 mg (87%) of the title compound, which
was found to be pure by 1H NMR: [R]25 -8.9 (c 0.42, H2O);
D
1H NMR (360 MHz, D2O) δ 4.32 (bt, 1 H, J ) 3 Hz), 4.09 (t, 1
H, J ) 8 Hz), 3.8-3.7 (m, 3 H), 3.39 (dd, 1 H, J ) 8, 4 Hz),
3.08 (m, 1 H), 2.85 (m, 2 H), 1.71 (m, 2 H); 13C NMR [(90 MHz,
D2O with dioxane as an internal standard (δ 67.4)] δ 78.3, 76.2,
72.8, 67.0, 64.5, 59.9, 46.6, 34.6. These values were in good
agreement with literature data.18
oils. Data for 47a : Rf 0.66 (20% i-PrOH/hexane); [R]25 +8.4
D
(c ) 1.6, EtOH); IR (CHCl3) 3438(br) cm-1; 1H NMR (300 MHz,
CDCl3) δ 7.4-7.1 (m, 15 H), 4.65, 4.57 (AB q, 2 H, J ) 11.5
Hz), 4.55-4.45 (m, 4 H), 4.29 (t, 1 H, J ) 4 Hz), 4.17 (t, 1 H,
J ) 4.5 Hz), 4.12 (m, 1 H), 3.6-3.4 (m, 3 H), 3.22 (m, 1 H),
3.39 (app q, 1 H, 1 H, J ) 4.5 Hz), 2.8-2.7 (m, 1H), 2.22 (bs,
1 H), 2.1-1.8 (m, 2 H); 13C NMR (90 MHz, CDCl3) δ 138.6,
138.4, 137.9, 128.4, 128.3, 127.9, 127.72, 127.70, 127.6, 127.4,
85.9, 81.7, 73.4, 73.3, 72.3, 72.1, 71.9, 71.8, 69.9, 52.7, 36.9;
MS (70 eV), m/e (%) 459 (M+, 1), 338 (100), 91 (95); HRMS
Calcd for C29H33NO4 (M+): 459.2409. Found: 459.2404. Data
Ack n ow led gm en t. This work was funded by the
National Institutes of Health (GM-35572). J .V.H. was
supported in part by National Research Service Award
T32 GM07767. We wish to thank Professors Scott E.
Denmark (University of Illinois) and George W. J . Fleet
(University of Oxford) for the useful exchange of infor-
mation.
for 47b: Rf 0.07 (20% i-PrOH/hexane); [R]25 +8.2 (c ) 0.62,
1
D
Su p p or tin g In for m a tion Ava ila ble: Copies of H NMR
EtOH); IR (CHCl3) 3488 (br) cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.4-7.1 (m, 15 H), 4.7-4.5 (m, 6 H), 4.3-4.2 (m, 2 H), 4.0-
3.9 (m, 2 H), 3.8-3.6 (m, 2 H), 3.38 (dd, 1 H, J ) 7.5, 3.5 Hz),
3.28 (app q, 1 H, J ) 7.5 Hz), 3.19 (m, 1 H), 2.80 (m, 1 H),
2.0-1.7 (m, 2 H); 13C NMR (90 MHz, CDCl3) δ 138.3, 138.2,
spectra for all new, stable compounds without elemental
analysis. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O000689Q