Stereoselective synthesis of polyhydroxylated indolizidines based on pyridinium salt photochemistry and ring rearrangement metathesis

Article abstract of DOI:10.1021/jo040226a

Ruthenium-catalyzed ring rearrangement metathesis (RRM) reactions of stereochemically diverse, differentially protected 4-N- allylacetamidocyclopenten-3,5-diols, prepared by using pyridinium salt photochemistry, have been explored as part of a program to develop novel routes for the synthesis of polyhydroxylated indolizidines. The RRM reactions, which produce selectively protected 1-acetyl-2-allyl-3-hydroxy-1,2,3,6- tetrahydropyridines, were found to take in high yields and with high levels of regioselectivity. The significance of RRM reactions of 4-N- allylacetamidocyclopenten-3,5-diols in the context of polyhydroxylated indolizidine synthesis is demonstrated by an application to the concise preparation of the potent glycosidase inhibitor, (-)-swainsonine.

Full text of DOI:10.1021/jo040226a

Ster eoselective Syn th esis of P olyh yd r oxyla ted In d olizid in es Ba sed
on P yr id in iu m Sa lt P h otoch em istr y a n d Rin g Rea r r a n gem en t
Meta th esis
Ling Song, Eileen N. Duesler, and Patrick S. Mariano*
Department of Chemistry, University of New Mexico, Albuquerque, New Mexico 87131
mariano@unm.edu
Received J uly 1, 2004
Ruthenium-catalyzed ring rearrangement metathesis (RRM) reactions of stereochemically diverse,
differentially protected 4-N-allylacetamidocyclopenten-3,5-diols, prepared by using pyridinium salt
photochemistry, have been explored as part of a program to develop novel routes for the synthesis
of polyhydroxylated indolizidines. The RRM reactions, which produce selectively protected 1-acetyl-
2-allyl-3-hydroxy-1,2,3,6-tetrahydropyridines, were found to take in high yields and with high levels
of regioselectivity. The significance of RRM reactions of 4-N-allylacetamidocyclopenten-3,5-diols
in the context of polyhydroxylated indolizidine synthesis is demonstrated by an application to the
concise preparation of the potent glycosidase inhibitor, (-)-swainsonine.
SCHEME 1
In tr od u ction
In previous publications, we reported the results of
studies aimed at developing the synthetic potential of a
novel photochemical reaction of pyridinium salts.1,2 An
example of this excited-state process is found in the
photoinduced transformation of in situ generated pyri-
dinium perchlorate to trans,trans-3,5-dihydroxy-4-ami-
nocyclopentene (1), isolated as the triacetyl derivative 2
(Scheme 1).3 In an early effort, we demonstrated that the
meso-diester 2 is enzymatically (EEACE) desymmetrized,
producing the monoalcohol 3 in 80% ee.4 Also, by using
an acetamide-directed alcohol inversion protocol, the
trans,trans-alcohol 3 can be converted to the cis,trans-
analogue 4.4 Additional investigations probing the syn-
thetic potential of this chemistry demonstrated that the
aminocyclopentendiol derivatives prepared in this fashion
serve as ideal starting materials in concise sequences for
stereoselective preparation of aminocyclopentitol natural
and nonnatural products5 and selectively protected 3-ami-
noaldopentoses.6
SCHEME 2
The synthetic power of the pyridinium salt photochemi-
cal process is a consequence of the fact that it transforms
a simple, inexpensive starting material into a stereo-
(1) Kaplan, L.; Pavlik, J . W.; Wilzbach, K. E. J . Am. Chem. Soc.
1972, 94, 3283.
(2) Yoon, U. C.; Quillen, S. L.; Mariano, P. S.; Swanson, R.;
Stavinoha, J . L.; Bay, E. J . Am. Chem. Soc. 1983, 105, 1204.
(3) Ling, R.; Yoshida, M.; Mariano, P. S. J . Org. Chem. 1996, 61,
4439.
(4) Ling, R.; Mariano, P. S. J . Org. Chem. 1998, 63, 6072. A newly
developed procedure for hydrolytic ring opening of the oxazoline,
formed by treatment of 3 and related substances with the Burgess salt,
was used in this case (see the description of the process used to
transform 19 to 20).
(5) (a) Cho, S. J .; Ling, R.; Kim, A.; Mariano, P. S. J . Org. Chem.
2000, 65, 1574. (b) Lu, H.; Mariano, P. S.; Lam, Y. F. Tetrahedron Lett.
2001, 42, 4755.
(6) Lu, H.; Su, Z.; Song, L.; Mariano, P. S. J . Org. Chem. 2002, 67,
3525.
chemically defined, structurally complex product, which
is endowed with functionality that can be used to guide
a wide variety of secondary reactions. In line with this
theme, we recognized that aminocyclopentendiol deriva-
tives related to 2-4 might serve as starting points in
novel sequences for the stereocontrolled synthesis of
polyhydroxylated indolizidines. The key step in the
proposed routes (Scheme 2) involves the ruthenium-
catalyzed tandem ring opening-ring closing metathesis
(or ring-rearrangement metathesis, RRM) chemistry
10.1021/jo040226a CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/21/2004
7284
J . Org. Chem. 2004, 69, 7284-7293

Synthesis of Polyhydroxylated Indolizidines
SCHEME 3
SCHEME 4
SCHEME 5a
developed by Grubbs7 and others.8,9 Accordingly, we
envisaged that N-allylacetamidocyclopentenes 5 might
serve as substrates in RRM-promoted formation of 6-al-
lyl-1,2,3,6-tetrahydropyridines 6. Relative and absolute
stereochemistry at the three chiral centers in 6 would
be controlled by the choice of the amidocyclopentene used
as starting materials for these reactions. Also, terminal
functionalization of the exocyclic allyl moiety in 6 would
then set the stage for cyclization to produce the indolizi-
dine skeleton 7. Finally, the allylic hydroxyl functionality
in 6 could be used to control the course of dihydroxylation
reactions at either or both of the unsaturated centers
either prior to or following indolizidine ring construction.
To explore the feasibility of the chemistry embodied
in the synthetic plan outlined in Scheme 2, a number of
N-allylacetamidocyclopentene derivatives related to 5
were prepared and subjected to RRM reaction conditions.
An important question probed in the exploratory phase
of this effort concerns the regiochemical course of the
RRM process, which has potential relative and absolute
stereochemical implications. This issue is exemplified by
the tandem metathesis reaction of the cis,trans-N-allyl-
acetamidocyclopentene 8, which can generate either of
the structurally (R1 * R2) or diastereomerically (R1 ) R2)
related tetrahydropyridines 9 and 10 depending on its
regiochemical course (Scheme 3). Observations made in
this effort demonstrate that the RRM reactions of N-
allylacetamidocyclopentenes take place to generate 6-al-
lyltetrahydropyridines in modest to excellent yields and
with high levels of regiochemical/stereochemical control.
In a second phase of this study, we have demonstrated
that this chemistry can be used as the foundation for
concise, stereocontrolled syntheses of biologically signifi-
cant indolizidines.
a
Reagents and conditions: (a) Burgess salt; NaH2PO4, aq THF;
(b) NaH, BnBr, DMF; (c) NaH, allyl bromide, DMF; (d) TBAF,
THF; (e) TBSCl, imidazole.
enantiomerically enriched (ca. 80% ee) N-allylacetamido-
cyclopentenes were synthesized to probe the efficiency
and regiochemical/stereochemical course of the ruthenium-
catalyzed RRM process. Included in this series are
substrates 14-16 and 18 (Scheme 4), which are prepared
by N-allylation reactions of previously reported4-6 acet-
amidocyclopentenes 11-13 and 17. In addition, the
trans,cis- and cis,cis-amidocyclopentenes, 22 and 24, are
prepared by sequences starting with the known4,6 mono-
TBS blocked cyclopentendiol 19 (Scheme 5).
Tandem metathesis reactions of the N-allylacetamido-
cyclopentenes are conducted by using ethylene-saturated
CH2Cl2 solutions containing 10-15 mol % of the ruthe-
nium alkylidene 2710 at reflux for 2-28 h. Chromato-
graphic separation in each case (except for reaction of
24) yields a single tetrahydropyridine product along with
trace or minor amounts of cross metathesis products (e.g.,
29) (Scheme 6). A ca. 15:1 mixture of tetrahydropyridines
41 and 42 is generated in the tandem metathesis reaction
of the cis,cis-N-allyacetamidocyclopentene 24.
Resu lts a n d Discu ssion
P r ep a r a tion a n d RRM Rea ction s of N-Allyla cet-
a m id ocyclop en ten es. Several stereochemically diverse
Structural and stereochemical assignments to the
tetrahydropyridine products, generated in these reac-
tions, are made by using a combination of NMR spectro-
scopic methods and chemical transformations. Specifi-
(7) Zuercher, W. J .; Hashimoto, M.; Grubbs, R. H. J . Am. Chem.
Soc. 1996, 118, 6634.
(8) Stragies, R.; Blechert, S. Synlett 1998, 169. Stragies, R.; Blechert,
S. J . Am. Chem. Soc. 2000, 122, 9584.
(9) For a recent review see: Randl, S.; Blechert, S. In Hanbook of
Metathesis. Applications in Organic Synthesis; Grubbs, R. H., Ed.;
Wiley-VCH: Morlenbach, 2003; Vol. 2.
(10) Morgan, J . P.; Grubbs, R. H. Org. Lett. 2000, 2, 3153. Huang,
J .; Stevens, E. D.; Nolan, S. P.; Petersen, J . L. J . Am. Chem. Soc. 1999,
121, 2674.
J . Org. Chem, Vol. 69, No. 21, 2004 7285

Song et al.
SCHEME 6
TABLE 1. 1H NMR Ch em ica l Sh ift Da ta for th e
6-Allyl-1,2,5,6-tetr a h yd r op yr id in es
cally, the exocyclic (C-7) vs cyclic (C-5) locations of the
OTBS groups in 30, 35, 38, and 41 are assigned by (1)
using COSY and HMQC NMR methods to determine the
chemical shifts of H-7 and H-5, (2) transforming OTBS
to OAc groups, and (3) again using COSY and HMQC
methods to assign chemical shifts to H-7 and H-5 in the
derived acetates (Table 1). In this manner, locations of
the OTBS groups in the tetrahydropyridines are revealed
by downfield shifts of either H-7 or H-5 promoted by the
OTBS f OAc transformations. A similar method (OAc
f OBn) is used to assign the structure of tetrahydropyr-
idine 32 (Table 1).
Additional support for the structural assignments
comes from (1) X-ray crystallographic analysis of the
monoalcohol 36 derived from tetrahydropyridine 35
(Figure 1) and (2) the observation that the OBn derivative
34, formed from tetrahydropyridine 32, is not equivalent
to 30.
1H NMR chemical shifts (ppm)
compd
H-4
H-5
H-7
H-8
30
31
32
33
34
35
36
37
38
40
41
43
5.60
5.55
5.60
5.65
5.71
5.65
5.80
5.80
5.85
5.90
5.60
5.90
3.75
5.50
5.65
3.85
3.99
4.10
4.01
5.10
3.85
3.90
4.46
5.45
4.05
3.85
5.10
4.65
4.34
3.55
3.52
3.55
3.85
5.20
3.65
3.74
5.85
5.70
5.90
5.90
5.82
5.75
5.68
5.80
5.65
5.70
5.70
5.70
substrates requires comment. It is generally believed that
RRM reactions, like their RCM counterparts, are equi-
librium processes occurring between interconvertible
alkenylcycloalkene starting materials and products.11 An
Regioch em ica l/Ster eoch em ica l F ea tu r es of th e
RRM Rea ction s. The observations summarized above
demonstrate that RRM reactions of the N-allylacetamido-
cyclopentenes take place with modest to high efficiencies.
The exceptionally high levels of regiochemical control
observed for reactions of the unsymmetrically substituted
(11) Miller, S. J .; Kim, S. H.; Chen, Z. R.; Grubbs, R. H. J . Am. Chem.
Soc. 1995, 117, 2108. Smith, A. B.; Adams, C. M.; Kozmin, S. A. J .
Am. Chem. Soc. 2001, 123, 990.
7286 J . Org. Chem., Vol. 69, No. 21, 2004

Synthesis of Polyhydroxylated Indolizidines
TABLE 2. Ma cr om od el (MM2 a n d MM3) Globa lly
Min im ized En er gies of P ossible P r od u cts of RRM
Rea ction s of N-Allyla ceta m id ocyclop en ten es
minimized energies (macromodel) of
tetrahydropyridine products
N-allylacet-
amidocyclo- product
energya
energya
product
pentene
obsd
MM2 MM3 not obsd MM2 MM3
15
16
18
22
24
32
30
35
38
+31
-46
-46
-48
122 32a
182 30a
178 35a
180 38a
+68
-21
-43
-41
167
158
181
181
178
F IGURE 1. Chem-3D plot of the X-ray crystallographically
determined atomic coordinates of 36.
SCHEME 7
41 (major) -50
174 42 (minor) -53
a
Energy in kJ /mol.
and 24. The results, compiled in Table 2, demonstrate
that in most, but not all, cases the exclusive or major
tetrahydropyridine product is calculated to have the
lowest energy. Importantly, in some instances, the en-
ergetic preferences are only marginal (e.g., 35 vs 35a )
and in one case (41 vs 42) the regioisomer formed in a
minor amount is lower in energy than the major product.
Finally, in several cases (30 vs 30a and 41 vs 42) the
relative energies of the regioisomeric tetrahydropyridines
are dependent on the choice of force constant parameters
used (MM2 vs MM3).
SCHEME 8
It is clear that the high degrees of regiochemical control
observed in RRM reactions of the 4-N-allylacetamidocy-
clopentenes are not well correlated with the calculated
energies of the possible tetrahydropyridine products.
Although the cause of this might be due to the calculation
method, an additional observation suggests that the RRM
reactions are not controlled by thermodynamics. Specif-
ically, the major and minor tetrahydropyridines 41 and
42, formed from acetamidocyclopentene 24, do not inter-
convert when they are independently subjected to the
RRM reaction conditions under which they are produced.
Several observation made in earlier studies of Ru-
catalyzed metathesis reactions suggest that kinetics can
govern features of the process. For example, a kinetic
argument was offered by Hoveyda and co-workers13 to
explain the surprising lack of reactivity of the cyclopen-
tenylstyrenyl ether 46 (Scheme 9) under RRM reaction
conditions. Ring strain associated with transformation
of an initially formed, exocyclic ruthenium-alkylidene
intermediate to a ruthenacyclobutane is believed to block
reaction of this substrate (Scheme 9). In a similar fashion,
the regiochemical course of RRM reactions of the unsym-
metrically substituted acetamidocyclopentenes might be
governed by the relative rates of formation of the regio-
example of this is found in studies of RRM reactions of
3-sulfonamidocyclopentenes (Scheme 7) by Blechert.12
The observed substituent effects on percent conversion
of starting materials 44 to products 45 is in accord with
thermodynamic arguments based on the influence sub-
stituent size on the equilibrium constant for intercon-
version of 44 to 45.
Thus, the regioselectivities of RRM reactions of the
unsymmetrically substituted 4-N-allylacetamidocyclo-
pentenes observed in the current study could be a
consequence of energy differences between two equili-
brating 2-allyl-1,2,3,6-tetrahydropyridine products. This
is depicted in Scheme 8 for reaction of cyclopentene 18,
which could generate either the observed tetrahydropyr-
idine 35 or the unobserved regioisomer 35a . To evaluate
whether the observed results are consistent with ther-
modynamic expectations, molecular mechanics calcula-
tions (Macromodel) were performed on all possible tet-
rahydropyridine regioisomers that could have been
produced in RRM reactions of substrates 15, 16, 18, 22,
(13) Harrity, J . P. A.; Visser, M. S.; Gleason, J . D.; Hoveyda, A. H.
J . Am. Chem. Soc. 1997, 119, 1488. Harrity, J . P. A.; La, D. S.; Cefalo,
D. R.; Visser, M. S.; Hoveyda, A. H. J . Am. Chem. Soc. 1998, 120, 2343.
(12) Ovaa, H.; Stapper, C.; van der Marel, G. A.; Overkleeft, H. S.;
van Boom, J . H.; Blechert, S. Tetrahedron 2002, 58, 7503.
J . Org. Chem, Vol. 69, No. 21, 2004 7287

Song et al.
polyhydroxylated indolizidines.16 As shown in Scheme 2,
the endocyclic and exocyclic alkene moieties in 2-allyl-
1,2,3,6-tetrahydropyridines, generated by RRM reaction
of 4-N-allylacetamidocyclopentenes, can serve as sites for
installation of hydroxyl functionality present in the
targets and required for key C-N bond forming cycliza-
tions to construct the indolizidine ring system. Clearly,
a critical feature of this general strategy is the control
of stereochemistry and regiochemistry in reactions used
to selectively introduce the hydroxyl groups. For example,
application of this methodology to the synthesis of the
potent glycosidase inhibitor, (-)-swainsonine (59),17 or
its analogues (Scheme 13) requires the execution of a
selective dihyroxylation and reduction of the respective
exocyclic and endocyclic alkene moieties in 2-allyltet-
rahydropyridines 56 generated by the RRM process. On
the other hand, selective dihydroxylation and hydrobo-
ration-oxidation is required in sequences used for the
conversion of tetrahydropyridines 56 to (+)-castanosper-
mine (60) or its analogues.
(-)-Sw a in son in e. Issues associated with selective
ring versus side chain olefin hydroxylation/dihydroxyla-
tion were addressed in a preliminary manner in a route
developed for synthesis of (-)-swainsonine starting with
tetrahydropyridine 38. This substance is efficiently formed
by RRM reaction of the acetamidocyclopentene 22 (Scheme
6). The challenge here resides in the requirement for
regioselective and stereoselective dihydroxylation of the
exocyclic olefin group in 38. Molecular modeling of 38 as
well as its alcohol and acetate derivatives, 39 and 40,
confirms the prediction (based on A1,3-strain consider-
ations) that there is a strong preference for these
substances to exist in 5,6-diaxial conformations. Conse-
quently, approach to either face of the endocyclic π-moi-
eties in 38-40 by hydroxylation reagents would be
SCHEME 9
isomeric tetrahydropyridine products. In these processes,
internal cross-metathesis reactions of an exocyclic Ru-
alkylidene,14 formed initially from the cyclopentene sub-
strates, can occur in two regiochemically different direc-
tions to competitively generate two polycyclic ruthenacyclo-
butanes. This is shown in Scheme 10 for reaction of
exocyclic ruthenium-alkylidene derived from 18, which
generates ruthenacyclobutanes 47 or 48, and for reaction
of the exocyclic ruthenium-alkylidene derived from 22,
which forms 49 and 50. Steric interactions between the
large group ligated, ruthenium ,and endo-positioned
OTBS and OBn groups in 48 and 50, respectively, would
make production of the polycyclic ruthenacyclobutanes
47 and 49 preferred. Thus, selective formation of the
respective tetrahydropyridines 35 and 38 in RRM reac-
tions of 18 and 22 could be the result of steric effects
governing competitive rates of formation of ruthenacy-
clobutane intermediates.13,15 A thermodynamic argument,
based on energetically biased (47 favored over 48)
equilibration of the tricyclic ruthenacyclobutane inter-
mediates, could also be used to explain these results.
Interestingly, a similar analysis can be applied to the
RRM reaction of cis,cis-acetamidocyclopentene 24 (Scheme
11). Here, the preference for metallacyclobutane 51 over
52, resulting from the size difference between the OBn
and OTBS groups, leads to selective formation of 41 as
the major tetrahydropyridine product.
Despite these successes, the kinetic/thermodynamic
argument presented above cannot be used to explain the
high degrees of regioselectivity attending RRM reactions
of the trans,trans-4-N-allylacetamidocyclopentenes 15
and 16. In each of these systems, the ruthenacyclobutane
intermediates (53 and 54, Scheme 12), derived from
initially formed exocyclic alkylidenes, are expected to
have nearly equal energies. Thus, kinetic considerations
suggest that, in opposition to the experimental findings,
these processes should be nonselective. However, it is
interesting to note that RRM reactions of 15 and 16 occur
with much lower yields than those of the corresponding
amidocyclopentenes 18, 22, and 24. At the current time,
we are unable to offer an explanation for these differ-
ences.
sterically blocked. Thus, mono dihydroxylation of these
substrates should occur at the exocyclic double bond,
preferentially. Moreover, substrate18 and/or catalyst19
control should serve to guide the stereochemical course
of the dihydroxylation process in order to derive exocyclic
erythro diol derivatives required for a (-)-swainsonine
synthesis.
Observations made in studies of the dihydroxylation
reactions of 38-40 indicate that these expectations are
only partially fulfilled. For example, treatment of 38
under the Sharpless AD-mix-R conditions19 or with the
catalytic OsO4/NMO reagent pair leads to predominant
formation of the ring-hydroxylated product 60 along with
Syn th esis of P olyh yd r oxyla ted In d olizid in es. The
RRM processes described above appear to be ideally
suited for applications to concise, stereocontrolled routes
for the synthesis of naturally occurring and nonnatural
(16) For other recent approaches to the synthesis of N-heterocycles
relying on RRM reactions, see: Stragies, R.; Blechert, S. Tetrahedron
1999, 55, 8179. Voigtmann, U.; Blechert, S. Org. Lett. 2000, 2, 3971.
Buschmann, N.; Ruckert, A.; Blechert, S. J . Org. Chem. 2002, 67, 4325.
(17) For a description of a recent synthesis of (-)-swainsonine as
well as references to earlier work, see: Pearson, W. H.; Ren, Y.; Powers,
J . D. Heterocycles 2002, 58, 421.
(18) Cha, J . K.; Christ, W. J .; Kishi, Y. Tetrahedron Lett. 1983, 24,
3943. Cha, J . K.; Kim, N. S. Chem. Rev. 1995, 95, 1761.
(19) Sharpless, K. B.; Amberg, W.; Bennani, Y. L.; Crispino, G. A.;
Hartung, J .; J eong, K. S.; Kwong, H. L.; Morikawa, K.; Wang, Z. M.;
Xu, D.; Zhang, X. L. J . Org. Chem. 1992, 57, 2768.
(14) Observation suggesting that RRM reactions of terminal-alkenyl-
substituted cycloalkenes are intiated by metathesis at the terminal
olefinic center is found in refs 7 and 13.
(15) Steric arguments such as the one presented here have been
used by Snapper (Snapper, M. L.; Tallarico, J . A.; Randall, M. L. J .
Am. Chem. Soc. 1997, 119, 1478) to explain the regiochemical course
of cross metathesis reactions of fused cyclobutenes.
7288 J . Org. Chem., Vol. 69, No. 21, 2004

Synthesis of Polyhydroxylated Indolizidines
SCHEME 10
SCHEME 11
SCHEME 13
SCHEME 14
SCHEME 12
the regiochemical course of this process. Thus, reaction
of alcohol 39 with catalytic OsO4/NMO gives an ca. 1:1
mixture of the erythro and threo triols, 62 and 64
(Scheme 15). Similarly, AD-mix-R treatment of acetate
40 leads to production of a ca. 1:1 mixture of exocyclic
diols 63 and 65 (Scheme 15).
Fortunately, in accord with earlier observations,20
catalytic OsO4/NMO dihydroxylation of acetate 40 ef-
ficiently (81%) generates a mixture of exocyclic diols 63
a minor amount of the desired exocyclic erythro diol
derivative 61 (Scheme 14). Reduction in the size of the
exocyclic allylic substituent has a pronounced effect on
(20) Marshall, J . A.; Tang, Y. J . Org. Chem. 1994, 59, 1457.
J . Org. Chem, Vol. 69, No. 21, 2004 7289

Song et al.
tene 11 (750 mg, 3.1 mmol) in 4 mL of DMF at 0 °C containing
NaH (94 mg, 3.7 mmol, 95%) was stirred for 20 min at 0 °C.
Allyl bromide (1.3 mL, 15.5 mmol) was added and the mixture
was stirred at 0°C for 1 h, diluted with water, and extracted
with EtOAc. The EtOAc extracts were washed with water and
then brine and concentrated in vacuo giving a residue, which
was subjected to column chromatography (silica gel, 3:1
hexane-acetone) to afford 14 (320 mg, 37%). 1H NMR δ 6.05-
5.65 (m, 5H), 5.20-5.01 (m, 2H), 3.93 (t, J ) 4.9 Hz, 1H), 3.86-
3.84 (m, 2H), 2.00 (s, 3H), 1.96 (s, 6H); 13C NMR δ 171.4, 170.4,
133.4, 132.8, 116.9, 78.5, 70.1, 52.6, 22.0, 20.8; HRMS (FAB)
m/z 282.1354 (M + 1) (calcd for C14H20NO5 282.1341).
SCHEME 15
SCHEME 16
N-Allyla ceta m id ocyclop en ten es 15, 16, 18, 22, a n d 24.
By using the N-allylation procedure described above, the
acetamidocyclopentenes 125a (370 mg, 1.2 mmol), 136 (40 mg,
0.11 mmol), 176 (169 mg, 0.47 mmol), 21 (100 mg, 0.28 mmol),
and 25 (199 mg, 0.55 mmol) were transformed into the
respective N-allylacetamidocyclopentenes 15 (338 mg, 81%),
16 (34 mg, 77%), 18 (175 mg, 93%), 22 (92 mg, 83%), and 24
(160 mg, 72%).
15: 1H NMR (mixture of rotamers) δ 5.83 (d, J ) 5.5 Hz,
0.6H), 5.82-5.62 (m, 3H), 5.35 (d, J ) 6.4 Hz, 0.2H), 5.04-
4.91 (m, 2.7H), 4.51 (d, J ) 5.5 Hz, 0.3H), 4.16 (t, J ) 6.3 Hz,
0.3H), 3.95-3.82 (m, 1.2H), 3.70-3.65 (m, 0.7H), 3.58-3.62
(m, 0.3H), 3.36 (t, J ) 5.5 Hz, 0.7H), 2.04 (s, 0.9H), 1.91 (s,
2.1H), 1.88 (s, 0.9H), 1.83 (s, 2.1H), 0.71 (s, 9H), -0.13 (s, 6H);
13C NMR, rotamer A δ 170.9, 170.2, 136.8, 133.4, 130.0, 117.0,
77.4, 75.9, 75.4, 54.4, 25.5, 22.2, 20.8, 17.7, -4.9, -5.1, rotamer
B δ 170.6, 170.0, 136.5, 133.6, 130.2, 115.6, 76.1, 75.2, 72.8,
45.3, 25.5, 21.6, 20.5, 17.7-4.9, -5.1; HRMS (FAB) m/z
354.2096 (M + 1) (calcd for C18H32NO4Si 354.2101).
16: 1H NMR (mixture of rotamers) δ 7.33-7.25 (m, 5H),
5.92-5.78 (m, 3H), 5.25 (s, 0.5H), 5.19-5.10 (m, 2H), 4.95-
4.90 (m, 0.5H), 4.61-4.58 (m, 0.5H), 4.57-4.54 (m, 2H), 4.45-
4.40 (m, 0.5H), 4.35-4.25 (m, 0.5H), 4.20-4.05 (m, 0.5H), 4.0-
3.85 (m, 1H), 3.65-3.50 (m, 0.5H), 3.48 (t, J ) 11.2 Hz, 0.5H),
2.23 (s, 1.5H), 2.07 (s, 1.5H), 0.874 (s, 9H), 0.053-0.025 (m,
6H); 13C NMR, rotamer A δ 171.1, 138.8, 135.7, 134.3, 131.3,
128.2, 127.6, 127.5, 117.2, 82.6, 77.1, 75.7, 71.3, 55.1, 25.7, 22.7,
22.0, -4.7, -4.8, rotamer B δ 171.1, 138.5, 135.3, 133.6, 131.2,
128.4, 127.8, 126.8, 115.9, 81.1, 75.6, 75.1, 71.6, 45.7, 25.6, 22.7,
22.0, 18.2, -4.7, -4.8; HRMS (FAB) m/z 402.2448 (M + 1)
(calcd for C23H36NO3Si 402.2464).
18: 1H NMR (mixture of rotamers) δ 7.32-7.24 (m, 5H),
6.03-6.01 (m, 1H), 5.95-5.93 (m, 1H), 5.93-5.75 (m, 1H),
5.30-5.20 (m, 1H), 5.15-5.0 (m, 2H), 4.95-4.85 (m, 1H), 4.85-
4.75 (m, 0.3H), 4.75-4.65 (m, 0.7H), 4.55-4.45 (m, 2H), 4.15-
4.10 (m, 0.3H), 4.10-3.95 (m, 1.7H), 2.11 (s, 1H), 2.09 (s, 1H),
0.85 (s, 9H), 0.07-0.01 (m, 6H); 13C NMR (mixture of rotamers)
δ 171.9, 171.1, 138.2, 137.9, 135.8, 135.3, 134.7, 133.7, 133.5,
128.4, 128.2, 127.7, 127.6, 127.5, 115.7, 114.8, 84.2, 83.8, 74.8,
74.3, 71.9, 71.0, 60.9, 50.0, 25.7, 21.8, 17.9, -4.9, -5.0; HRMS
(FAB) m/z 402.2468 (M + 1) (calcd for C23H36NO3Si 402.2464).
and 65, in which the erythro isomer 63 dominates by
4.4:1 (Scheme 15). Without separation, 63 and 65 are
converted to the separable indolizidine diacetates 66
(47%) and 67 (11%) by sequential treatment with 6 N
HCl, DEAD/Ph3P, and Ac2O/DMAP (Scheme 16). Al-
though hydrogenation and hydrogenolysis at the benzyl-
oxy-allyl moiety in 66 can be performed in a stepwise
manner, it is more convenient to carry out the two
operations simultaneously. Accordingly, reaction of 66
with H2/PdCl2 in MeOH, followed by acid-promoted ester
hydrolysis and ion exchange chromatography leads to the
formation of (-)-swainsonine (59) (80%, ca. 80% ee). The
physical21 (mp 139-140 °C [lit.22 mp 141-143 °C]; [R]25
D
-60 (c 0.84, MeOH) [lit.21[R]25D -71 (c 0.56, MeOH)]) and
spectroscopic properties of the synthetic material matched
those previously reported for the natural product.
In a similar fashion, the indolizidine diacetate 67 is
converted to (-)-2-epi-swainsonine (68) (63%; mp 168-
169 °C [lit.23 mp 169.5-172 °C]; [R]25D -35 (c 0.61, MeOH)
22: 1H NMR (mixture of rotamers) δ 7.29-7.24 (m, 5H),
6.01-5.90 (m, 3H), 5.24-4.93 (m, 3.5H), 4.68-4.66 (m, 0.5H),
4.48-4.43 (m, 3H), 4.15-4.05 (m, 0.25H), 3.94-3.92 (m, 1.5H),
3.85-3.75 (m, 0.25H), 2.11 (s, 1H), 2.05 (s, 2H), 0.83 (s, 9H),
0.03-0.02 (m, 6H); 13C NMR (mixture of rotamers) δ 171.7,
170.5, 137.9, 137.3, 137.1, 135.2, 134.6, 132.2, 131.5, 127.8,
127.2, 127.0, 126.9, 126.8, 115.4, 114.8, 80.6, 80.3, 77.2, 77.0,
71.8, 67.9, 63.0, 49.3, 25.1, 21.6, 21.5, 17.3, -4.6, -4.7, -5.0,
[lit.23 [R]25 -61 (c 0.12, EtOH)]) by sequential hydroge-
D
nation, hydrogenolysis, and ester hydrolysis (Scheme 16).
-5.2; HRMS (FAB) m/z 402.2470 (M + 1) (calcd for C23H36
NO3Si 402.2464).
-
Exp er im en ta l Section
24: 1H NMR δ 7.31-7.24 (m, 5H), 6.05-5.95 (m, 1H), 5.94-
5.93 (m, 1H), 5.88-5.85 (m,2H), 5.01-4.96 (m, 2H), 4.66-4.64
(m, 1H), 4.60-4.40 (ABq, J ) 11.8 Hz, 2H), 4.40-4.38 (m, 1H),
3.91-3.89 (m, 1H), 3.85-3.84 (m, 1H), 2.07 (s, 3H), 0.87 (s,
9H), 0.077 (d, J ) 3.3 Hz, 3H), 0.04 (s, 3H); 13C NMR δ 173.2,
137.9, 136.6, 135.7, 132.8, 128.2, 127.7, 127.5, 115.3, 80.3, 74.3,
71.9, 57.6, 51.2, 25.7, 22.8, 18.0, -5.0, -5.3.; HRMS (FAB) m/z
402.2456 (M + 1) (calcd for C23H36NO3Si 402.2464).
tr a n s,tr a n s-4-[N-Allyla cet a m id o]-3,5-d ia cet oxycyclo-
p en ten e (14): A Rep r esen ta tive N-Allyla tion P r oced u r e.
A solution of the known4 4-acetamido-3,5-diacetoxycyclopen-
(21) Lindsay, K. B.; Pyne, S. G. J . Org. Chem. 2002, 67, 7774.
(22) Naruse, M.; Aoyagi, S.; Kibayashi, C. J . Org. Chem. 1994, 59,
1358.
(23) Elbein, A. D.; Szumilo, T.; Sanford, B. A.; Sharpless, K. B.;
Adams, C. Biochemistry 1987, 26, 2502.
7290 J . Org. Chem., Vol. 69, No. 21, 2004

Synthesis of Polyhydroxylated Indolizidines
tr a n s,cis-3-ter t-Bu tyld im eth ylsilyloxy-4-a ceta m id ocy-
clop en ten -5-ol (20). To a solution of the known6 alcohol 19
(55 mg, 0.20 mmol) in 5 mL of THF at 70 °C was added a
solution of the Burgess reagent (58 mg, 0.24 mmol) in 5 mL of
THF. The resulting solution was stirred for 3 h at 70 °C and
10 mL of aqueous NaH2PO4 (pH 5.2) was added. The mixture
was stirred at 25 °C for 3 d, concentrated in vacuo, and
extracted with CH2Cl2. The CH2Cl2 extracts were concentrated
in vacuo to give a residue that was subjected to column
chromatography (silica gel, 2:1 hexane-acetone) to afford 20
25.7, 23.2, 18.1, -4.7, -4.9; HRMS (FAB) m/z 362.2168 (M +
1) (calcd for C20H32NO3Si 362.2151).
Ru th en iu m -Ca ta lyzed Ta n d em Meta th esis Rea ction
of N-Allyla ceta m id ocyclop en ten e 14 F or m in g 2-Allyltet-
r a h yd r op yr id in e 28: A Rep r esen ta tive RRM P r oced u r e.
A solution of N-allylacetamidocyclopentene 14 (72 mg, 0.25
mmol) and ruthenium catalyst 27 (21 mg, 0.025 mmol) in 13
mL of CH2Cl2 was stirred at reflux for 16 h while being purged
with C2H4. Concentration in vacuo provided a residue, which
was subjected to column chromatography (silica gel, 2:1
hexane-acetone) to afford the 2-allyltetrahydropyridine 28 (50
mg, 69%) and dimer 29 (6 mg, 4%).
28: 1H NMR (rotamer mixture) δ 5.90-5.70 (m, 2H), 5.65-
5.56 (m, 1.5H), 5.50-5.30 (m, 1.7H), 5.30-5.20 (m, 0.8H),
5.20-5.15 (m, 0.6H), 5.12-5.05 (m, 0.8H), 4.45-4.42 (m, 1H),
3.92-2.89 (m, 0.5H), 3.83-3.79 (m, 0.5H), 3.30-3.26 (m, 0.6H),
2.09 (s, 2.2H), 2.02(s, 1.3H), 1.98 (s, 2.8H), 1.93-1.91 (m, 2.7H);
13C NMR, rotamer A δ 170.9, 169.1, 133.5, 126.2, 123.2, 118.6,
71.1, 67.6, 55.3, 39.6, 21.7, 21.0, rotamer B δ 170.3, 169.8,
133.8, 125.0, 124.6, 115.9, 72.2, 66.0, 48.4, 43.7, 21.9, 20.9;
HRMS (FAB) m/z 282.1354 (M + 1) (calcd for C14H20NO5
282.1341).
29: 1H NMR δ 5.97-5.91 (m, 4H), 5.79-5.74 (m, 3H), 5.60-
5.58 (m, 3H), 4.20-4.15 (m, 2H), 3.88 (s, 2H), 3.82 (d, J ) 3.5
Hz, 2H), 2.15-2.14 (m, 4H), 2.06-2.00 (m, 14H); 13C NMR δ
171.4, 170.3, 170.2, 132.8, 127.7, 127.3, 78.4, 76.6, 69.4, 69.1,
50.5, 44.4, 21.9, 21.6, 20.8, 20.6; HRMS (FAB) m/z 535.2305
(M + 1) (calcd for C26H35N2O10 535.2292).
F or m a tion of 2-Allyltetr a h yd r op yr id in es 30, 32, 35, 38,
40, a n d 41 by Ta n d em Meta th esis. By using the general
procedure described above, N-allylacetamidocyclopentene 16
(34 mg, 0.085 mmol) was converted to tetrahydropyridine 30
(10 mg, 30%) and a dimer (9 mg, 14%); 15 (37 mg, 0.1 mmol)
was converted to tetrahydropyridine 32 (19 mg, 51%) and
dimer (10 mg, 14%), 18 (25 mg, 0.062 mmol) was converted to
tetrahydropyridine 35 (18 mg, 72%), 22 (64 mg. 0.16 mmol)
was converted to tetrahydropyridine 38 (58 mg, 90%), and 24
(46 mg, 0.11 mmol) was converted to tetrahydropyridines 41
(20 mg, 44%) and 42 (1 mg, 3%).
1
(42 mg, 76%) as a solid, mp 111-112 °C. H NMR δ 6.29 (d, J
) 7.5 Hz, 1H), 5.86-5.92 (m, 2H), 4.76-4.78 (m, 1H), 4.69 (s,
1H), 4.05-4.18 (m, 1H), 3.31 (s, 1H), 1.99 (s, 3H), 0.85 (s, 9H),
0.04 (s, 6H); 13C NMR δ 170.9, 138.4, 133.2, 80.3, 73.4, 60.3,
25.7, 23.3, 18.1, -4.7, -4.9; HRMS (FAB) m/z 272.1684 (M +
1) (calcd for C13H26NO3Si 272.1682).
tr a n s,cis-3-ter t-Bu tyld im eth ylsilyloxy-4-a ceta m id o-5-
ben zyloxycyclop en ten e (21). A solution of alcohol 20 (168
mg, 0.62 mmol) and NaH (19 mg, 0.74 mmol, 95%) in 3 mL of
DMF was stirred at 0 °C for 20 min. Benzyl bromide (30 µL,
0.25 mmol) was added and the solution was stirred for 2 h at
25 °C, diluted with water, and extracted with EtOAc. The
EtOAc extracts were washed with water and brine and
concentrated in vacuo providing a residue, which was subjected
to column chromatography (silica gel, 2:1 hexanes-ethyl
acetate) to afford the 21 (210 mg, 94%). Mp 77-78 °C; 1H NMR
δ 7.34-7.26 (m, 5H), 6.04 (d, J ) 8.2 Hz, 1H), 5.97-5.92 (m,
2H), 4.68 (s, 1H), 4.51 (d, J ) 5.7 Hz, 1H), 4.51-4.43 (ABq, J
) 11.6 Hz, 2H), 4.37-4.32 (m, 1H), 1.94 (s, 3H), 0.86 (s, 9H),
0.06 (s, 3H), 0.04 (s, 3H); 13C NMR δ 169.8, 138.8, 137.9, 131.7,
128.4, 127.9, 127.7, 80.7, 80.6, 72.0, 58.6, 25.7, 23.3, 18.0, -4.8,
-5.0; HRMS (FAB) m/z 362.2166 (M + 1) (calcd for C20H32
NO3Si 362.2151).
-
tr a n s,cis-4-Acetam ido-5-ben zyloxycyclopen ten -3-ol (23).
A solution of 21 (19 mg, 0.052 mmol) and TBAF (16 mg, 0.062
mmol) in 3 mL of THF was stirred at 25 °C for 3 h and
concentrated in vacuo to provide a residue, which was sub-
jected to column chromatography (silica gel, 1:2 hexane-
acetone) to afford 23 (12 mg, 100%). Mp. 130-131°C; 1H NMR
δ 7.32-7.25 (m, 5H), 6.54 (br s, 1H), 6.04-6.02 (m, 1H), 5.96-
5.94 (m, 1H), 4.71 (s, 1H), 4.69 (s, 1H), 4.50 (m, 2H), 4.43-
4.41 (m, 1H), 3.92-3.90 (m, 1H), 1.93 (s, 3H); 13C NMR δ 171.9,
138.9, 137.5, 130.2, 128.3, 127.7, 127.6, 81.5, 80.1, 71.2, 60.6,
30: 1H NMR (mixture of rotamers) δ 7.30-7.21 (m, 5H),
5.95-5.75 (m, 1H), 5.75-5.45 (m, 2H), 5.4-5.1 (m, 2H), 5.05-
4.95 (m, 0.5H), 4.6-4.55 (m, 0.5H), 4.55-4.5 (m, 0.5H), 4.5-
4.35 (m, 1H), 4.4-4.35 (m, 0.5H), 4.3-4.25 (m, 0.5H), 4.25-
4.2 (m, 0.5H), 4.05-4.0 (m, 0.5H), 4.00-3.95 (m, 0.5H), 3.8-
3.75 (m, 0.5H), 3.3-3.2 (m, 0.5H), 2.12 (s, 1.5H), 2.04 (s, 1.5H),
0.90 (s, 9H), 0.1(m, 6H); 13C NMR, rotamer A δ 170.5, 136.9,
128.7, 128.2, 127.4, 127.1, 123.9, 116.4, 77.0, 70.4, 67.5, 61.1,
40.4, 25.9, 22.1, 18.2, -4.76, rotamer B δ 171.5, 136.8, 129.7,
128.0, 127.7, 127.1, 122.9, 115.6, 78.7, 71.2, 65.9, 52.9, 44.6,
25.9, 22.1, -4.8; HRMS (FAB) m/z 408.2541 (M + 7) (calcd for
22.6; HRMS (FAB) m/z 248.1296 (M + 1) (calcd for C14H18
NO3 248.1287).
-
cis,cis-4-Aceta m id o-5-ben zyloxycyclop en ten -3-ol (26).
To a solution of 23 (310 mg, 1.3 mmol) in 20 mL of THF at 70
°C was added a solution of the Burgess reagent (58 mg, 0.24
mmol) in 10 mL of THF. The resulting solution was stirred
for 3 h at 70 °C. An aqueous NaH2PO4 solution (30 mL, pH
5.2) was added and the mixture was stirred at 25 °C for12 h,
concentrated in vacuo, and extracted with CH2Cl2. The CH2-
Cl2 extracts were concentrated in vacuo to give a residue that
was subjected to column chromatography (silica gel, 1:2
hexane-acetone) to afford 26 (240 mg, 77%). 1H NMR δ 7.34-
7.28 (m, 5H), 6.39 (d, J ) 6.3 Hz, 1H), 6.07-6.00 (m, 2H), 4.49
(s, 2H), 4.41-4.38 (m, 1H), 4.34-4.27 (m, 2H), 3.19 (br s, 1H),
2.02 (s, 3H); 13C NMR δ 170.6, 137.8, 136.5, 133.7, 128.2, 127.6,
127.6, 79.5, 72.9, 72.0, 52.9, 22.9; HRMS (FAB) m/z 248.1290
(M + 1) (calcd for C14H18NO3 248.1287).
cis,cis-3-ter t-Bu tyld im eth ylsiloxy-4-a ceta m id o-5-ben -
zyloxycyclop en ten e (25). A solution of 26 (229 mg, 0.99
mmol), imidazole (162 mg, 2.4 mmol), and TBSCl (179 mg, 1.2
mmol) in 10 mL of CH2Cl2 was stirred at 25 °C for 12 h, diluted
with water, and extracted with CH2Cl2. The CH2Cl2 extracts
were concentrated in vacuo to provide a residue, which was
subjected to column chromatography (silica gel, 2:1 hexane-
acetone) to afford 25 (304 mg, 85%). 1H NMR δ 7.30-7.24 (m,
5H), 6.0-5.9 (m, 3H), 4.6-4.5 (m, 4H), 4.33-4.32 (m, 1H), 1.94
(s, 3H), 0.88 (9H), 0.068 (d, J ) 5.4 Hz, 6H); 13C NMR δ 169.7,
138.6, 136.2, 133.2, 128.3, 127.4, 127.3, 79.2, 73.2, 71.6, 52.3,
C
23H35NO3SiLi 408.2546).
32: 1H NMR, rotamer A δ 5.89-5.93 (m, 1H), 5.73-5.59
(m, 3H), 5.19 (d, J ) 17.3 Hz, 1H), 5.12-5.08 (m, 3H), 4.44-
4.41 (m, 2H), 3.33 (d, J ) 19.5 Hz, 1H), 2.06 (s, 3H), 1.95 (s,
3H), 0.89 (s, 9H), 0.067 (s, 6H), rotamer B δ 6.02-5.98 (m,
1H), 5.73-5.59 (m, 3H), 5.12-5.08 (m, 2H), 4.57 (s, 1H), 4.00
(t, J ) 6.6 Hz, 1H), 3.82-3.89 (m, 2H), 2.12 (s, 3H), 1.97 (s,
3H), 0.87 (s, 9H), 0.052 (s, 6H); 13C NMR, rotamer A δ 170.3,
169.0, 134.8, 129.9, 122.1, 115.9, 72.7, 65.1, 59.8, 44.0, 25.7,
22.0, 21.0, 17.9, -4.9, rotamer B δ 170.7, 169.2, 135.2, 128.3,
123.8, 114.8, 71.2, 66.8, 51.2, 40.1, 25.6, 21.9, 20.9, 18.0, -4.9;
HRMS (FAB) m/z 354.2101 (M + 1) (calcd for C18H32NO4Si
354.2101).
35: 1H NMR δ 7.32-7.19 (m, 5H), 5.80-5.65 (m, 3H), 5.43-
5.40 (m, 1H), 5.35-5.32 (m, 1H), 4.66-4.6 0(m, 1H), 4.58-
4.51 (m, 1H), 4.24-4.18 (m, 1H), 4.19-4.10 (m, 1H), 4.12-
4.10 (m, 1H), 3.57-3.52 (m, 1H), 3.08 (d, J ) 20 Hz, 1H), 2.15
(s, 3H), 0.86 (s, 9H), 0.02 (d, J ) 14.6 Hz, 6H); 13C NMR δ
171.8, 137.4, 135.6, 128.4, 127.9, 126.2, 124.1, 120.9, 77.7, 70.0,
64.0, 63.8, 56.7, 38.9, 25.8, 21.6, 19.2, 18.1, -4.2, -4.6; HRMS
(FAB) m/z 402.2466 (M + 1) (calcd for C23H36NO3Si 402.2464).
J . Org. Chem, Vol. 69, No. 21, 2004 7291

Song et al.
38: 1H NMR δ 7.24-7.31 (m, 5H), 6.15-6.05 (m, 1H), 5.95-
5.80 (m, 1H), 5.75-5.60 (m, 1H), 5.20-5.05 (m, 2H), 4.80-
4.75 (m, 1H), 4.75-4.35 (m, 2H), 4.15-4.00 (m, 1H), 4.00-
3.80 (m, 2H), 3.45-3.35 (m, 1H), 2.12 (s, 2.4H), 2.10 (s, 0.6H),
0.84 (s, 3.6H), 0.81 (s, 5.4H), -0.02-0.04 (m, 6H); 13C NMR δ
171.4, 138.7, 138.2, 129.9, 128.3, 128.0, 127.6, 127.4, 124.2,
121.9, 118.3, 116.7, 73.4, 72.9, 70.2, 69.5, 69.1, 61.7, 54.2, 43.5,
39.1, 25.7, 25.5, 22.3, 21.7, 17.7, -4.0, -4.3, -5.0; HRMS (FAB)
m/z 402.2461 (M + 1) (calcd for C23H36NO3Si 402.2464).
mg, 90%). 1H NMR (mixture of rotamers) δ 7.34-7.26 (m, 5H),
6.05-5.9 (m, 0.6H), 5.9-5.75 (m, 1.2H), 5.75-5.65 (m, 0.4H),
5.6-5.55 (m, 0.4H), 5.30-5.09 (m, 2.4H), 4.67-4.63 (m, 0.6H),
4.58 (s, 1.8H), 4.50-4.46 (m, 0.6H), 4.25-4.24 (m, 0.6H), 4.35-
4.25 (m, 0.6H), 4.20-4.15 (m, 0.4H), 4.1-3.98 (m, 1H), 3.85-
3.75 (m, 0.4H), 3.6-3.54 (m, 1H), 2.07 (s, 3H), 0.83 (s, 9H),
-0.023 to -0.0881 (m, 6H); 13C NMR, rotamer A δ 171.0, 139.8,
137.8, 128.4, 127.6, 125.8, 115.5, 73.5, 71.3, 71.3, 59.2, 41.1,
25.7, 22.1, 17.9, -4.0, -4.8, rotamer B δ 170.2, 139.7, 137.9,
128.3, 127.7, 124.5, 114.6, 71.6, 71.6, 71.3, 50.9, 44.9, 25.9, 22.2,
17.8, -4.4, -4.5; HRMS (FAB) m/z 402.2468 (M + 1) (calcd
for C23H36NO3Si 402.2464).
Tetr a h yd r op yr id in e 36. A solution of 35 (129 mg, 0.32
mmol) and TBAF (101 mg, 0.38 mmol) in 10 mL of THF was
stirred at 25 °C for 2 h and concentrated in vacuo giving a
residue that was subjected to column chromatography (silica
gel, 1:1 acetone-hexane) to afford 36 (90 mg, 98%). Mp 127-
128 °C; 1H NMR δ 5.90-5.80 (m, 1H), 5.80-5.70 (m, 1H),
5.70-5.60 (m, 1H), 5.37 (d, J ) 10.2 Hz, 1H), 5.22 (d, J ) 17.1
Hz, 1H), 4.54-4.48 (m, 2H), 4.19 (d, J ) 11.6 Hz, 1H), 4.01-
3.96 (m, 2H), 3.52 (t, J ) 9.0 Hz, 1H), 3.03 (d, J ) 20 Hz, 1H),
2.14 (s, 3H); 13C NMR δ 172.6, 137.6, 135.2, 128.3, 127.8, 127.6,
127.4, 124.4, 121.0, 77.5, 69.8, 63.4, 63.2, 39.0, 21.6; HRMS
(ES) m/z 310.1418 (M + Na) (calcd for C17H21NO3Na 310.1414).
41: 1H NMR δ 7.34-7.25 (m, 5H), 5.78-5.76 (m, 1H), 5.76-
5.72 (m, 1H), 5.70-5.69 (m, 1H), 5.33-5.2 8 (m, 2H), 4.75-
4.70 (m, 1H), 4.58 (ABq, J ) 12 Hz, 1H), 4.42 (Abq, J ) 12
Hz, 1H), 4.46 (d, J ) 5.4 Hz, 1H), 3.86 (d, J ) 9.8 Hz, 1H),
3.67(t, J ) 6.01 Hz, 1H), 3.15 (d, J ) 19.7 Hz, 1H), 2.11 (s,
3H), 0.85 (s, 9H), 0.07 (d, J ) 2.75 Hz, 3H), 0.03 (s, 3H); 13C
NMR δ 170.6, 137.6, 135.5, 128.3, 127.8, 127.7, 127.3, 124.5,
120.1, 78.7, 70.6, 64.0, 63.1, 38.8, 25.9, 25.7, 21.6, 18.0, -4.4,
-4.7; HRMS (FAB) m/z 402.2456 (M + 1) (calcd for C23H36
NO3Si 402.2464).
-
42: 1H NMR δ 6.02-5.98 (m, 0.6H), 5.92-5.84 (m, 1.4H),
5.79-5.71 (m, 1H), 5.23-5.06 (m, 2H), 4.80-4.69 (m, 1.2H),
4.62-4.55 (m, 1.4H), 4.50-4.44 (m, 0.4H), 4.20 (d, J ) 5.4 Hz,
1H), 4.07-4.03 (m, 0.4H), 3.98-3.93 (m, 0.7H), 3.91-3.87 (m,
0.9H), 3.75-3.70 (m, 0.4H), 3.29 (d, 19.7 Hz, 0.6H), 2.08 (s,
2H), 2.05 (s, 1H), 0.82 (s, 6H), 0.79 (s, 3H), -0.04 to -0.05 (m,
3H), -0.08 to -0.10 (m, 3H); 13C NMR δ 170.6, 140.3, 138.4,
129.3, 128.3, 127.9, 127.6, 127.4, 127.1, 123.8, 122.1, 117.4,
116.3, 73.9, 72.3, 70.5, 69.5, 69.1, 61.3, 53.3, 43.1, 39.3, 25.6,
22.3, 21.8, 17.9, -4.21, -5.13; HRMS (FAB) m/z 402.2480 (M
+ 1) (calcd for C23H36NO3Si 402.2464).
Tetr a h yd r op yr id in e 37. To a solution of 35 (10 mg, 0.024
mmol) in 5 mL of THF was added TBAF (7 mg, 0.027 mmol),
at
and the reaction mixture was stirred
25 °C for 2 h and
concentrated. To the residue was added CH2Cl2 (5 mL),
pyridine (0.05 mL), DMAP (1 mg, 0.013 mmol), and acetic
anhydride (0.0 8 mL) and the resulting mixture was stirred
at 25 °C for 12 h and concentrated in vacuo to provide a
residue, which was subjected to column chromatography (silica
gel, 2:1 hexane-acetone) to afford the product 37 (5 mg, 66%).
1H NMR (mixture of rotamers) δ 7.33-7.18 (m, 5H), 5.97-
5.94 (m, 1H), 5.94-5.76 (m, 2H), 5.46-5.43 (m, 0.8H), 5.40-
5.30 (m, 0.2H), 5.30-5.26 (m, 0.8H), 5.24-5.28 (m, 0.2H),
5.11-5.09 (m, 0.8H), 5.05-5.00 (m, 0.2H), 4.66-4.61 (m, 0.8H),
4.56-4.53 (m, 0.8H), 4.10-4.05 (m, 0.4H), 4.04-4.02 (m, 0.8H),
3.95-3.90 (m, 0.2H), 3.7-3.65 (m, 0.2H), 3.59(t, J ) 9.4 Hz,
0.8H), 3.4-3.3 (d, J ) 20 Hz, 0.2H), 3.09 (d, J ) 20 Hz, 0.8H),
Tetr a h yd r op yr id in e 31. A solution of 30 (10 mg, 0.025
mmol) in 5 mL of THF and TBAF (6.8 mg, 0.027 mmol) was
stirred at 25 °C for 2 h. The reaction mixture was concentrated,
CH2Cl2 (5 mL), pyridine (0.05 mL), DMAP (1 mg, 0.013 mmol),
and acetic anhydride (0.08 mL) were added, and stirring was
continued overnight at 25 °C. The solvent was removed in
vacuo to provide a residue, which was subjected to column
chromatography (silica gel,2:1 hexane-acetone) to afford 31
(7 mg, 88%). 1H NMR δ 7.31-7.20 (m, 5H), 5.80-5.65 (m,
2.3H), 5.65-5.55 (m, 0.7H), 5.45 (m, 1H), 5.35-5.25 (m, 2H),
4.54-4.50 (m, 1.4H), 4.47 (d, J ) 2.8 Hz, 0.4H), 4.42 (t, J )
6.9 Hz, 0.6H), 4.30-4.20 (m, 1.4H), 4.10-4.00 (m, 0.3H), 3.86
(t, J ) 7.8 Hz, 1H), 3.85-3.60 (m, 0.2H), 3.26-3.20 (m, 0.7H),
2.14 (s, 2H), 2.06 (s, 1H), 1.99 (s, 2H), 1.98 (s, 1H); 13C NMR
δ 170.9, 170.4, 137.8, 136.1, 135.7, 128.3, 128.1, 127.7, 127.6,
127.3, 126.3, 125.4, 124.9, 123.7, 118.8, 117.1, 78.5, 76.6, 70.4,
69.7, 68.4, 67.0, 56.6, 49.7, 44.2, 39.9, 22.0, 21.2, 21.0; HRMS
(FAB) m/z 330.1715 (M + 1) (calcd for C19H24NO4 330.1705).
2.13 (s, 2.5H), 2.10 (s, 0.5H), 2.0 1(s, 2.5H), 1.98 (s, 0.5H); 13
C
NMR (mixture of rotamers) δ 171.6, 170.5, 137.5, 135.3, 134.8,
131.6, 129.6, 128.4, 128.3, 127.9, 127.8, 127.6, 122.1, 121.7,
120.3, 120.1, 78.1, 77.4, 70.0, 69.7, 66.1, 59.9, 53.5, 43.0, 39.0,
21.7, 21.0; HRMS (FAB) m/z 330.1711 (M + 1) (calcd for C19H24
NO4 330.1705).
-
Tetr a h yd r op yr id in e 40. A solution of 38 (220 mg, 0.55
mmol) in 5 mL of THF and TBAF (172 mg, 0.66 mmol) was
stirred at 25 °C for 2 h and concentrated in vacuo giving a
residue, which was subjected to column chromatography to
yield 154 mg (94%) of alcohol 39. 1H NMR δ 7.30-7.23 (m,
5H), 6.0-5.76 (m, 3H), 5.25-5.15 (m, 2H), 5.01 (d, J ) 11.0
Hz, 0.3 H), 4.9-4.2 (m, 2.5H), 4.11 (d, J ) 16.5 Hz, 0.5H), 4.0-
3.7 (m, 2H), 3.54 (d, J ) 19.9 Hz, 0.7H), 2.12 (s, 3H); 13C NMR
δ 172.3, 138.1, 138.0, 137.4, 128.4, 128.3, 128.0, 127.8, 127.7,
127.5, 121.9, 118.5, 72.1, 70.6, 70.4, 69.7, 69.1, 62.1, 54.4, 43.0,
39.1, 22.1, 21.6; HRMS (FAB) m/z 288.1608 (M + 1) (calcd for
Tetr a h yd r op yr id in e 33. A solution of 32 (74 mg, 0.21
mmol) and NaOMe (3 mg, 0.04 mmol) in 10 mL of MeOH was
stirred at 25 °C for 12 h, diluted with water, and extracted
with CH2Cl2. Concentration in vacuo of the CH2Cl2 extracts
gave a residue, which was subjected to column chromatogra-
phy (silica gel, 2:1 hexane-acetone) to afford the product 33
(51 mg, 78%). 1H NMR (mixture of rotamers) 6.0-5.85 (m, 1H),
5.7-5.5 (m, 2H), 5.4-5.3 (m, 1H), 5.1-5.05 (m, 1H), 5.02-
5.01 (m, 0.6H), 4.7-4.5 (m, 2.4H), 3.94-3.90 (m, 1H), 3.86-
3.82 (m, 0.6H), 3.51 (d, J ) 18.9 Hz, 0.4H), 2.08 (s, 1.2H), 2.05
(s, 1.8H), 0.90 (s, 4.5H), 0.88 (s, 4.5H), 0.09 (s, 6H); 13C NMR,
rotamer A δ 170.5, 137.1, 129.3, 124.3, 113.9, 72.1, 65.4, 52.7,
44.0, 25.7, 22.0, 17.9, -4.87, -4.92, rotamer B δ 170.4, 137.7,
128.1, 125.4, 115.6, 71.9, 66.8, 60.3, 40.2, 25.7, 21.9, 17.9, -4.9,
C
17H22NO3 288.1600).
A solution of alcohol 39 (52 mg, 0,13 mmol) in THF (5 mL)
containing acetyl chloride (22 µL, 0.3 mmoL) and Et3N (32uL,
0.23 mmol) was stirred for 12 h at 25 °C. Concentration in
vacuo gave a residue that was subjected to column chroma-
tography (silica gel, 1:1 hexane-acetone) to afford the product
acetate 40 (34 mg, 58%). 1H NMR δ 7.34-7.24 (m, 5H), 6.15-
6.05 (m, 1H), 6.0-5.8 (m, 1H), 5.8-5.6 (m, 1H), 5.35-5.15 (m,
3H), 4.72-4.63 (m, 1H), 4.60-4.40 (m, 2H), 4.09-4.06 (m, 1H),
4.0-3.8 (m, 1H), 3.39-3.34 (m, 1H), 2.12 (s, 2H), 2.10 (s, 1H),
1.99 (s, 1.5H), 1.98 (s, 1.5H); 13C NMR δ 171.1, 170.5, 169.9,
169.3, 137.9, 137.8, 132.8, 132.9, 130.4, 128.4, 128.3, 128.1,
127.8, 127.5, 123.5, 121.3, 121.2, 120.1, 72.9, 72.0, 70.5, 70.4,
69.2, 59.5, 51.5, 43.2, 39.2, 22.0, 21.3, 21.0, 20.8; HRMS (FAB)
m/z 330.1715 (M + 1) (calcd for C19H24NO4 330.1705).
-4.9; HRMS (FAB) m/z 312.1998 (M + 1) (calcd for C16H30
NO3Si 312.1995).
-
Tetr a h yd r op yr id in e 34. To a solution of 33 (31 mg, 0.1
mmol) in 2 mL of DMF at 0 °C was added NaH (5 mg, 0.2
mmol, 95%). After 10 min of stirring at 0 °C, benzyl bromide
(30 uL,0.25 mmol) was added and the mixture was stirred for
12 h at 25 °C, diluted with water, and extracted with EtOAc.
The EtOAc extracts were dried and concentrated in vacuo to
provide a residue, which was subjected to column chromatog-
raphy (silica gel, 2:1 hexanes-ethyl acetate) to afford 34 (36
7292 J . Org. Chem., Vol. 69, No. 21, 2004

Synthesis of Polyhydroxylated Indolizidines
Tetr a h yd r op yr id in e 43. To a solution of 41 (20 mg, 0.05
mmol) in 5 mL of THF was added TBAF (14 mg, 0.06 mmol),
and the reaction mixture was stirred at 25 °C for 2 h. The
reaction mixture was concentrated, CH2Cl2 (5 mL), pyridine
(0.05 mL), DMAP (2 mg), and acetic anhydride (0.08 mL) were
added, and stirring was continued for 12 h at 25 °C. Concen-
tration in vacuo provided a residue, which was subjected to
column chromatography (silica gel, 2:1 hexane-acetone) to
afford product 43 (10 mg, 50%). 1H NMR (mixture of rotamers)
7.35-7.24 (m, 5H), 5.96-5.92 (m, 1H), 5.87-5.85 (m, 1H),
5.85-5.67 (m, 1H), 5.56-5.55 (m, 0.35H), 5.50-5.48 (d, J )
5.6 Hz, 0.65H), 5.5-5.3 (m, 1.3H), 5.3-5.25 (m, 0.35H), 5.19
(d, J ) 17.1 Hz, 0.35H), 5.04 (d, J ) 9.4 Hz, 0.35H), 4.78-
4.72 (m, 0.65H), 4.61-4.58 (m, 1H), 4.28-4.22 (m, 1H), 4.04-
3.99 (m, 1H), 3.76-3.62 (m, 1.35H), 3.22 (d, J ) 20 Hz, 0.65H),
2.10 (s, 2H), 2.05 (s, 1H), 2.00 (s, 2H), 1.98 (s, 1H); 13C NMR
(mixture of rotamers) δ 170.5, 170.4, 170.3, 170.2, 137.8, 137.2,
136.2, 134.9, 130.4, 128.4, 128.3, 127.9, 127.8, 127.7, 122.3,
120.9, 120.7, 119.8, 79.3, 78.1, 70.7, 70.1, 66.3, 65.5, 59.6, 53.2,
43.0, 39.0, 22.2, 21.7, 21.0, 20.9; HRMS (FAB) m/z 330.1696
(M + 1) (calcd for C19H24NO4 330.1705).
Con ver sion of 66 to (-)-Sw a in son in e (59). A solution
of 66 (25 mg, 0.072 mmol) in 2 mL of CH3OH containing PdCl2
(9.7 mg, 0.055 mmol) was stirred under a H2 atmosphere for
1 h, filtered through Celite, and concentrated in vacuo. A
solution of the residue in 6 mL of 3 N HCl/THF (1:1) was
stirred at 25 °C for 24 h and extracted with ether. Concentra-
tion of the ether extracts provided a residue that was subjected
to ion-exchange chromatography (Dowex 1-X8, OH- form,
100-200 mesh, eluting with water) to yield 10 mg (80%) of
(-)-swainsonine (59) as a solid; mp 139-140 °C [lit.22 mp 141-
143 °C]; [R]25 -60 (c 0.84, MeOH) [lit.21 [R]25 -71 (c 0.56,
D
D
MeOH)]; HRMS (FAB) m/z 174.1128 (M + 1) (calcd for C8H16
-
NO3 174.1130). The 1H NMR and 13C NMR spectra of the
synthetic material matched those reported earlier.21
Con ver sion of 67 to (-)-2-Ep isw a in son in e (68). A
solution of 67 (26 mg, 0.074 mmol) in 2 mL of ethanol
containing 10% Pd/C (12 mg, 0.011 mmol) was stirred under
a H2 atmosphere at 25 °C for 2 h, filtered through Celite, and
concentrated in vacuo to provide 24 mg (93%) of the reduced
product, which was used in the next step without further
purification; [R]25 -27 (c 0.48, CHCl3); 1H NMR δ 7.31-7.24
D
Con ver sion of 40 to In d olizid in es 66 a n d 67. To a
solution of 40 (425 mg, 1.33 mmol) in 16 mL of acetone was
added NMO (148 mg, 1.26 mmol), with stirring for 10 min.
Then a solution of OsO4 (17 mg, 0.07 mmol) in 1:1 acetone/
H2O was added and the resulting mixture was stirred at 25
°C for 3 h and diluted with 3 mL of saturated aqueous Na2S2O3,
and then mixture was stirred at 25 °C for 1 h. MgSO4 was
added and the mixture was filtered. The filtrate was concen-
trated in vacuo to provide a residue, which was subjected to
column chromatography (silica gel, 1:1 hexane-acetone) to
afford erythro and threo diols 63 and 65 (389 mg, 81%) as a
4.4:1 mixture. A solution of this mixture (99 mg, 0.27 mmol)
in 8 mL of 6 N HCl/THF (1:1) was stirred at 70 °C for 4 h and
concentrated in vacuo to give a residue that was dissolved in
2 mL of pyridine-containing Ph3P (89 mg, 0.34 mmol), 4A
molecular sieves, and DEAD (57 µl, 0.34 mmol). This solution
was stirred at 0 °C for 3 h and diluted with 3 mL of CH2Cl2
before adding 0.5 mL of Ac2O and DMAP with stirring for an
additional 12 h at 25 °C. Concentration of the mixture in vacuo
gave a residue that was partitioned between aq NaHCO3 and
CH2Cl2. Concentration of the CH2Cl2 layer in vacuo provided
a residue that was subjected to column chromatography (silica
gel, 8:1 hexane-acetone) to yield 66 (44 mg, 47%, 79% ee by
(chiral HPLC)) and 67 (11 mg, 11%, 78% ee (chiral HPLC)).
66: [R]25D -85 (c 0.51, CHCl3); 1H NMR δ 7.32-7.24 (m, 5H),
5.93-5.88 (m, 1H), 5.83-5.80 (m, 1H), 5.57-5.52 (m, 1H),
5.37-5.30 (m, 1H), 4.59 (ABq, J ) 28 Hz, 1H), 4.49 (Abq, J )
28 Hz, 1H), 4.32-4.28 (m, 1H), 3.48-3.39 (m, 1H), 3.15-3.08
(m, 1H), 2.82-2.66 (m, 2H), 2.52-2.46 (m, 1H), 1.99 (s, 3H),
1.93 (s, 3H); 13C NMR δ 170.0, 169.9, 137.9, 128.4, 127.8, 127.7,
126.4, 126.1, 70.9, 70.8, 70.7, 69.9, 65.4, 58.3, 51.9, 20.6, 20.4;
HRMS (FAB) m/z 346.1641 (M + 1) (calcd for C19H24NO5
346.1654).
(m, 5H), 5.36 (d, J ) 4.2 Hz, 1H), 4.89 (t, J ) 6.3 Hz, 1H), 4.5
(ABq, J ) 14 Hz, 1H), 4.37 (Abq, J ) 14 Hz, 1H), 3.67 (t, J )
9.4 Hz, 1H), 3.54-3.51 (m, 1H), 3.0 (d, J ) 10.3 Hz, 1H), 2.3-
2.1 (m, 2H), 2.07 (s, 3H), 1.95 (s, 3H), 1.76-1.72 (m, 1H), 1.58-
1.54 (m, 1H), 1.23-1.18 (m, 1H); 13C NMR δ 169.9, 169.7,
138.1, 128.3, 127.7, 127.6, 76.7, 76.6, 73.1, 70.5, 69.1, 59.9, 52.0,
29.4, 23.5, 20.8, 20.7; HRMS (FAB) m/z 348.1813 (M + 1) (calcd
for C19H26NO5 348.1811).
A solution of the reduction product (20 mg, 0.057 mmol) in
2 mL of CH3OH containing PdCl2 (8 mg, 0.045 mmol) was
stirred under a H2 atmosphere at 25 °C for 1 h, filtered through
Celite, and concentrated in vacuo. A solution of the residue in
6 mL of 3 N HCl/THF (1:1) was stirred at 25 °C for 48 h and
washed with ether. Concentration of the aqueous layer pro-
vided a residue that was subjected to ion-exchange chroma-
tography (Dowex 1-X8, OH form, 100-200 mesh, eluting with
water) to yield 7 mg (68%) of (-)-2-episwainsonine (68) as a
solid; mp 168-169 °C [lit.23 mp 169.5-172 °C], [R]25 -35 (c
D
0.61, MeOH) [lit.23 [R]25D -61 (c 0.12, absolute EtOH)]; HRMS
(FAB) m/z 174.1130 (M + 1) (calcd for C8H16NO3 174.1130).
1H NMR spectroscopic data for this substance matches those
previously reported.24 13C NMR data not previously reported
are as follows: δ 78.0, 77.5, 72.5, 66.9, 61.4, 52.6, 33.3, 24.1.
Ack n ow led gm en t. These studies were conducted
with financial support provided by the National Insti-
tutes of Health (GM-27251). We thank Professor Robert
Grubbs for providing us with information about possible
factors governing the regiochemical course of RRM
reactions.
67: [R]25 -57 (c 0.13, CHCl3); 1H NMR δ 7.32-7.28 (m,
D
Su p p or tin g In for m a tion Ava ila ble: General experimen-
5H), 5.91 (ABq, J ) 12 Hz, 1H), 5.81 (ABq, J ) 12 Hz, 1H),
5.36 (d, J ) 4.5 Hz, 1H), 4.95 (t, J ) 5.9 Hz, 1H), 4.59 (ABq,
J ) 14 Hz, 1H), 4.44 (Abq, J ) 14 Hz, 1H), 4.22 (d, J ) 8.0
Hz, 1H), 3.73-3.69 (m, 1H), 3.42 (d, J ) 15.9 Hz, 1H), 2.82 (d,
J ) 16.3 Hz, 1H), 2.63 (t, J ) 4.7 Hz, 1H), 2.22-2.18 (m, 1H),
2.04 (s, 3H), 1.91 (s, 3H); 13C NMR δ 169.9, 169.7, 138.0, 128.4,
127.9, 127.7, 126.5, 125.8, 77.5, 77.0, 76.6, 76.4, 76.1, 71.0, 70.7,
64.9, 59.3, 51.9, 20.7; HRMS (FAB) m/z 346.1650 (M + 1) (calcd
for C19H24NO5 346.1654).
1
tal information, H and 13C NMR spectra for 14-16, 18, 20-
26, 28-43, and 66-68, and a summary of crystallographic
parameters for 36. This material is available free of charge
via the Internet at http://pubs.acs.org.
J O040226A
(24) The 1H NMR spectrum of 68 was kindly provided by C. Adams.
J . Org. Chem, Vol. 69, No. 21, 2004 7293