Asymmetric synthesis of (-)-swainsonine

Article abstract of DOI:10.1071/CH04009

This paper describes a new synthesis of (-)-swainsonine via the ring-closing metathesis reaction of a substituted 3-allyl-4-vinyloxazolindin-2- one and subsequent diastereoselective syn-dihydroxylation of the resulting pyrrolo[1,2-c]oxazol-3-one.

Full text of DOI:10.1071/CH04009

Rapid Communication
CSIRO PUBLISHING
www.publish.csiro.au/journals/ajc
Aust. J. Chem. 2004, 57, 669–672
Asymmetric Synthesis of (−)-Swainsonine†
Karl B. LindsayA and Stephen G. PyneA,B
A Department of Chemistry, University of Wollongong, Wollongong NSW 2522, Australia.
B Author to whom correspondence should be addressed (e-mail: spyne@uow.edu.au).
This paper describes a new synthesis of (−)-swainsonine via the ring-closing metathesis reaction of a substituted
3-allyl-4-vinyloxazolindin-2-one and subsequent diastereoselective syn-dihydroxylation of the resulting pyrrolo-
[1,2-c]oxazol-3-one.
Manuscript received: 15 January 2004.
Final version: 3 March 2004.
RO
H8a
While numerous total syntheses of naturally occurring
(−)-swainsonine 1[1] and its non-natural enantiomer (+)-
swainsonine[2–4] have been reported, these molecules and
theiranaloguesarestillpopulartargetstodevelopnewsynthe-
tic strategies and methodologies. This is partially driven by
theabilityof(−)-swainsoninetoinhibitGolgiα-mannosidase
II, an enzyme involved in the processing of glycoproteins on
the surface of cancer cells. This process has been associ-
ated with cancer metastasis, and thus (−)-swainsonine and
analogues are potentially useful anti-metastasis drugs for
the treatment of cancer.[5,6] Unfortunately, (−)-swainsonine
is not selective for Golgi α-mannosidase II over other
α-mannosidases (e.g. lysosomal α-mannosidase)[7] resulting
in undesired side effects (e.g. inhibition of the catabolism
of oligosaccharides), and thus creating the need for more
potent and selective (−)-swainsonine analogues. A com-
prehensive review of the synthesis of swainsonine and its
analogues was published in 2000 by Nemr.[1] Since this
review, seven total syntheses of (−)-swainsonine have been
reported[8–14] along with several papers describing the syn-
thesis of analogues.[14–21]
H
RO
H3␤
N
H
N
2
2
‘‘OsO4L*’’
RO
OH
H
8
‘‘OsO4L*’’
1
3
8a
OH
5
N
3 R ϭ PG
1 R ϭ H
Scheme 1.
analogue of 2 gave the desired diol 3 (R = Bn) in excellent
diastereomeric ratios (d.r. 95 : 5 to 98 : 2). The stereochemi-
caloutcomesforthesereactionswereconsistentwithaddition
of the osmium reagent to the less hindered α-face of 2,[14,22]
approach to the β-face being hindered by the pseudo-axial
protons H8a and H3β (Scheme 1).
While the Sharpless AD reagents allowed us access to the
desired diol 3 (R = Bn), the yields for these reactions were
disappointingly low (44–49%).[14] We suspect that this low
yield was due to a competing oxidation reaction occurring at
thetertiarynitrogenatomof 2(R = Bn)or 3(R = Bn). Inview
of this difficulty, we have recently examined pyrrolo[1,2-
c]oxazol-3-ones (e.g. 4) as more suitable bicyclic substrates
to control the DH of substituted pyrrolidines and to simul-
taneously protect the nitrogen atom from oxidation.[24,25]
For example, the unsubstituted pyrrolo[1,2-c]oxazol-3-one
4 underwent DH with catalytic osmium tetraoxide to
give exclusively the diol 5 in good yield (Scheme 2). This
diol resulted from attack of the oxidizing agent from the
Blechert,[12] Carretero,[13] andPyne[14] haveeachreported
a synthesis of (−)-swainsonine via the syn-dihydroxylation
(DH) of the indolizidine derivative 2 (Scheme 1).
Syn-dihydroxylation of racemic 2 (R =TBS) using cata-
lytic osmium tetraoxide/stoichiometric N-methylmorpholine
oxide (NMO) has been reported[22] to give an 88 : 12 mixture
of diols in favour of the desired diastereomer 3 (Scheme 1).
Blechert,[12] however, found that this method using enantio-
merically pure 2 (R =TBS) gave an almost equal mixture of
the two diastereomers. However, the use of the Sharpless
asymmetric dihydroxylation (AD) reaction conditions[23]
usingAD-mix-α gave a 20 : 1 mixture of diols in favour of the
desired diol 3 (R =TBS). Independently, we[14] also found
that the use of eitherAD-mix-α orAD-mix-β on the O-benzyl
† Dedicated to Professor Lew Mander on the occasion of his 65th birthday. Lew, thanks for teaching me carbocyclic chemistry. I will never forget the structure
of gibberellic acid (GA3).
© CSIRO 2004
10.1071/CH04009
0004-9425/04/070669

670
K. B. Lindsay and S. G. Pyne
concave face of the molecule (Fig. 1) due to the pseudo-
axial protons H5β and H7a that sterically hinder the β-face
to attack by the osmium reagent (Fig. 1).[24] This argument is
similar to that proposed to account for the facial selectivity
of DH reactions on the related indolizines 2 (Scheme 1). This
methodology allowed the diastereoselective synthesis of the
triol 6 having the desired relative stereochemistry required
for the synthesis of swainsonine.[24] In this paper, we report
the application of this methodology to the synthesis of (−)-
swainsonine 1 (Schemes 3 and 5).
The requisite pyrrolo[1,2-c]oxazol-3-one 11 was prepared
according to Scheme 3 from the vinyl epoxide (R,S)-7
(e.e. 92%) that was used in our earlier synthesis of (−)-
swainsonine.[14] Aminolysis of 7 with allylamine in the pres-
ence of lithium triflate with microwave heating at 110◦C for
1 h[26] gave the desired amino alcohol 8 in 88% yield. While
this compound had been prepared previously by us by means
of conventional heating in a sealed tube using acid catalysis
(0.1 equiv. of p-TsOH, 105◦C, 5 days), this method was a sig-
nificant improvement in terms of reaction time. The amino
alcohol 8 was converted into the N-Boc 2,5-dihydropyrrole
10 as described by us earlier.[14] Upon treatment with sodium
hydride in toluene solution at 50◦C, compound 10 was con-
verted into the desired pyrrolo[1,2-c]oxazol-3-one 11 in 74%
yield. Other solvents were much less effective in this trans-
formation; for example, the use of tetrahydrofuran (THF)
and dimethylformamide (DMF) resulted in very poor con-
versions. We suspect that the sodium tert-butoxide that is
generated in these reactions can react reversibly with 11
giving an equilibrium mixture of 10 and 11. In toluene,
however, sodium tert-butoxide would be expected to be less
soluble, forcing the equilibrium to favour the desired ring-
closed product 11. Alternatively, 11 could be prepared by
first treatment of the amino alcohol 8 with triphosgene in
the presence of base (Et3N) to give the oxazolidinone 16
in 77% yield. Treatment of 16 with Grubbs I catalyst (ben-
zylidene bis(tricyclohexylphosphine)dichlororuthenium)[24]
in refluxing dichloromethane solution for 18 h gave 11 in
77% yield (Scheme 4).
OH
H
N
H
N
4
.
K2OsO4 2H2O, NMO
OH
O
O
acetone/H2O,
room temp., 20h
[76%]
O
O
5
OH
H
(a) NaOH, MeOH, H2O
reflux, 24h
OH
HO
HN
(b) ion-exchange
[96%]
6
Scheme 2.
H7a
OsO4
H5␤
1
N
O
O
With the pyrrolo[1,2-c]oxazol-3-one 11 in hand, we next
examined its DH reactions; the results of this study are
summarized in Table 1. To this end, 11 was treated with
catalytic potassium osmate dihydrate (K2OsO4 · 2H2O) and
stoichiometric NMO in acetone–water,[14] which gave a 3 : 1
OsO4
Fig. 1. Stereochemical model of 4 (Spartan PC AM1).
OH
H
OH
H
PMBO
H
H
N
O
(a)
(c)
(d )
O
BocN
RN
OPMB
OPMB
O
OPMB
8 R ϭ H
7
10
11
(b)
9 R ϭ Boc
PMBO
OR
OR
PMBO
H
H
H
H
(e)
ϩ
OR
OR
O
O
N
N
O
O
12 R ϭ H
13 R ϭ H
15 R ϭ Bn
(f )
(f )
14 R ϭ Bn
Scheme 3. (a) CH2=CHCH2NH2, LiOTf, microwave, 110◦C, 1 h (88%); (b) Boc2O, NEt3, THF, room temp., 24 h (98%); (c) Grubbs’ catalyst,
CH2Cl2, reflux, 18 h (95%); (d) NaH, toluene, 50◦C, 24 h (74%); (e) dihydroxylation (see Table 1); (f) NaH, BnBr, Bun4NI, THF, room temp.,
2 days (100%).

Asymmetric Synthesis of (−)-Swainsonine
671
1
inseparable mixture of the diols 12 and 13 from H NMR
analysis (Scheme 3).[27] Complete conversion was obtained
within 18 h at room temperature giving an 85% yield. A pure
sample of the major diol 12 [mp 146◦C, [α]D25 −31.0 (c 1.77
in CHCl3)] was isolated by careful crystallization from hot
dichloromethane and petroleum spirit. This crystallization
was not necessary, however, because benzylation of the mix-
ture of diols gave the corresponding bis(benzyl ethers) 14 and
15 in quantitative yield, and these were readily separable by
column chromatography. Conducting the same DH reaction
at 0◦C resulted in an improved diastereoselectivity and yield
(3.5 : 1 and 92%, respectively).Thus the diastereoselectivities
for the DH reaction of 11 were similar to those found for the
DH reactions of indolizidine 2.These DH reactions, however,
were far less diastereoselective than those of the unsub-
stituted pyrrolo[1,2-c]oxazol-3-one 4. The C1 α-substituent
present in 11 is most likely responsible for this reduced
α-face diastereoselectivity. With the aim of increasing the
steric bulk of the oxidant, and perhaps the diastereoselec-
tivity, the DH reaction was repeated in the presence of the
coordinating ligand pyridine (10 equiv.). Unfortunately, the
use of pyridine extended the reaction time to 7 days and
resulted in a significant reduction in the diastereoselectiv-
ity to 1.5 : 1 (Table 1). When AD-mix-α was used at room
temperature, the reaction did not go to completion within 6
days, and the diastereoselectivity was only slightly improved
(3.7 : 1). Surprisingly, whenAD-mix-β was used, a 20 : 1 ratio
of diastereoisomers was obtained, albeit at low conversion
after 6 days at room temperature, giving a 46% yield of prod-
uct diols 14 and 15 (and 45% recovered starting material). We
attribute the discrepancies between the α- and β-AD-mixes
to be a result of a matched/mismatched situation.[23]
to protect the secondary alcohol of 17, as it might inter-
fere with the cyclization of the piperidine ring, by providing
an alternative nucleophile to the nitrogen atom. Conse-
quently, 17 was treated with tert-butylchlorodiphenylsilane
(TBDPSCl) and imidazole at 60◦C to give the silyl ether
18 in 97% yield. Oxidative removal of the O-PMB (para-
methoxybenzyl) protecting group was then conducted by
reaction of 18 with ceric(iv) ammonium nitrate (CAN),
giving the amino alcohol 19 in 92% yield.[14,25] Cycliza-
tion of 19 by activation of the primary hydroxyl and then
intramolecular N-alkylation (Ph3P, CBr4, Et3N)[14] gave
the indolizidine derivative 20 in excellent yield (93%).
Deprotection of the silyl ether by reaction with tetra-n-
butylammonium fluoride (TBAF) was slow, requiring 5 days
to complete; however, the desired alcohol 21 was obtained
in 76% yield. Finally, the two benzyl ethers were removed
by catalytic hydrogenolysis under acidic conditions (PdCl2,
H2) and the resulting product was purified and neutralized
by ion-exchange chromatography to give (−)-swainsonine 1
in excellent overall yield (93%) and purity.
This synthetic compound had identical 1H and 13C NMR
spectra and TLC mobility to an authentic sample of (−)-
swainsonine[28] and had a specific rotation, [α]2D6 −71 (c 0.56
RO
OBn
H
OBn
PMBO
H
H
N
(a)
OBn
OBn
O
HN
PMBO
O
14
17 R ϭ H
(b)
18 R ϭ TBDPS
Conversion of 14 into (−)-swainsonine proved to be
relatively straightforward (Scheme 5). Hydrolysis of the
oxazolidinone with NaOH in MeOH–water[25] gave the
amino alcohol 17 in good yield (84%). We thought it prudent
TBDPSO
RO
OBn
OBn
H
N
H
(c)
(d )
(f )
OBn
OBn
1
HN
PMBO
H
HO
H
N
19
20 R ϭ TBDPS
21 R ϭ H
(a)
(b)
(e)
8
11
O
Scheme 5. (a) NaOH, MeOH, H2O, 110◦C, microwave, 2 h (84%);
(b) TBDPSCl, imidazole, 65◦C, 3 days (97%); (c) CAN, CH3CN, H2O,
room temp., 3 h (92%); (d) PPh3, CBr4, NEt3, CH2Cl2, 0◦C, 16 h (93%);
(e) TBAF, THF, room temp., 5 days (76%); (f) PdCl2, H2 (1 atm),
MeOH, room temp., 2 h; ion-exchange (93%).
O
16
Scheme 4. (a) triphosgene, NEt3, CH2Cl2, 0◦C, 2 h (77%);
(b) Grubbs’ catalyst, CH2Cl2, reflux, 18 h (77%).
Table 1. Summary of results for the dihydroxylation reactions of 11
Reagent usedA
Reaction
temp.B
Combined yield [%]
Recovered
11 [%]
Ratio of
12 to 13
Reaction
time [days]
of 12 and 13
K2OsO4 · 2H2O/NMO
K2OsO4 · 2H2O/NMO
K2OsO4 · 2H2O/NMO,
pyridine (10 equiv.)
AD-mix-α
RT
85
92
90
–
–
5
3 : 1
3.5 : 1
1.5 : 1
1
2
7
0◦C
RT
RT
RT
24
46
56
45
3.7 : 1
20 : 1
6
6
AD-mix-β
A See references [14] and [25] for general experimental procedures.
B RT refers to room temperature.

672
K. B. Lindsay and S. G. Pyne
in MeOH) [lit.[29] −82.6 (c 1.03 in MeOH); lit.[8] [α]D23 −86
(c 0.30 in MeOH); lit.[30] [α]D23 −74 (c 0.98 in MeOH); lit.[31]
[α]2D6 −85.5 (c 0.42 in MeOH); lit.[32] [α]2D3 −87.2 (c 2.1 in
MeOH)], that was consistent with the enantiomeric purity of
its precursor 7.
[9] B. M. Trost, D. E. Patterson, Chem. Eur. J. 1999, 5, 3279. doi:
10.1002/(SICI)1521-3765(19991105)5:11<3279::AID- CHEM-
3279>3.3.CO;2-W
[10] T. Punniyamurthy, R. Irie, T. Katsuki, Chirality 2000,
12, 464. doi:10.1002/(SICI)1520-636X(2000)12:5/6<464::AID-
CHIR28>3.3.CO;2-Q
In conclusion, an alternative synthetic strategy has been
developed that allows the synthesis of (−)-swainsonine. The
oxazolidinone group was found to be a useful protecting
group in the ring-closing metathesis reaction and, as part of
a pyrrolo[1,2-c]oxazol-3-one ring system, has functioned as
a stereodirecting group in the syn-DH reaction.
[11] W. H. Pearson, Y. Ren, J. D. Powers, Heterocycles 2002, 58, 421.
[12] N. Buschmann, A. Rückert, S. Blechert, J. Org. Chem. 2002,
67, 4325. doi:10.1021/JO025589U
[13] J. de Vicente, R. G. Arrayás, J. Canˇada, J. C. Carretero, Synlett
2000, 53.
[14] K. B. Lindsay, S. G. Pyne, J. Org. Chem. 2002, 67, 7774. doi:
10.1021/JO025977W
[15] H. Razavi, R. Polt, J. Org. Chem. 2000, 65, 5693. doi:10.1021/
JO000527U
[16] V. Cerè, S. Pollicino, A. Ricci, J. Org. Chem. 2003, 68, 3311.
doi:10.1021/JO020511Z
[17] P. Diaz-Pérez, M. I. Garcia-Moreno, C. Ortiz Mellet,
J. M. Garcia Fernández, Synlett 2003, 341.
[18] W. H. Pearson, L. Guo, Tetrahedron Lett. 2001, 42, 8267.
doi:10.1016/S0040-4039(01)01777-4
Accessory Materials
General experimental details, and syntheses of compounds 1,
11–15, and 17–21 are available from the author or, until July
2009, the Australian Journal of Chemistry.
[19] W. H. Pearson, E. J. Hembre, Tetrahedron Lett. 2001, 42, 8273.
doi:10.1016/S0040-4039(01)01778-6
[20] W. H. Pearson, L. Guo, T. M. Jewell, Tetrahedron Lett. 2002,
43, 2175. doi:10.1016/S0040-4039(02)00254-X
[21] For a recent review see: S. G. Pyne, Curr. Org. Synth. 2004,
in press.
[22] C. Mukai, Y.-i. Sugimoto, K. Miyazawa, S. Yamaguchi,
M. Hanaoka, J. Org. Chem. 1998, 63, 6281. doi:10.1021/
JO980598H
Acknowledgments
We thank the Australian Research Council for financial sup-
portandtheUniversityof WollongongforaPh.D. scholarship
to K.B.L. We thank Dr Reg Smith from Phytex Australia for
an authentic sample of (−)-swainsonine.
[23] H. C. Kolb, M. S. Van Nieuwenhze, K. B. Sharpless, Chem. Rev.
1994, 94, 2483.
References
[1] A. E. Nemr, Tetrahedron 2000, 56, 8579. doi:10.1016/S0040-
4020(00)00790-0
[24] A. S. Davis, N. J. Gates, K. B. Lindsay, M. Tang, S. G. Pyne,
Synlett 2004, 49.
[2] B. Davis, A. A. Bell, R. J. Nash, A. A. Watson, R. C. Griffiths,
M. G. Jones, C. Smith, G. W. J. Fleet, Tetrahedron Lett. 1996,
37, 8565. doi:10.1016/0040-4039(96)01957-0
[3] J. P. Shivlock, J. R. Wheatley, R. J. Nash, A. A. Watson,
R. C. Griffiths, T. D. Butters, M. Müller, D. J. Watkin,
D. A. Winkler, G. W. J. Fleet, J. Chem. Soc., Perkin Trans. 1
1999, 2735.
[25] M. Tang, S. G. Pyne, J. Org. Chem. 2003, 68, 7818. doi:10.1021/
JO034914Q
[26] K. B. Lindsay, M. Tang, S. G. Pyne, Synlett 2002, 731. doi:
10.1055/S-2002-25337
[27] The structure of 15, and hence 13, was established by its
formation, as a by-product, in the benzylation of compound 17
in reference [14].
[4] T. Oishi, T. Iwakuma, M. Hirama, S. Itô, Synlett 1995, 404.
doi:10.1055/S-1995-5002
[28] Kindly provided by Dr Reg Smith, from Phytex Australia,
Sydney.
[5] P. E. Goss, M. A. Baker, J. P. Carver, J. W. Dennis, Clin. Cancer
Res. 1995, 1, 935.
[29] M. Naruse, S. Aoyagi, C. Kibayashi, J. Org. Chem. 1994,
59, 1538.
[6] P. E. Goss, C. L. Reid, D. Bailey, J. W. Dennis, Clin. Cancer
Res. 1997, 3, 1077.
[30] W. H. Pearson, E. J. Hembre, J. Org. Chem. 1996, 61, 7217.
doi:10.1021/JO961101B
[7] P. R. Dorling, C. R. Huxtable, S. M. Colegate, Biochem. J. 1980,
191, 649.
[31] J. A. Hunt, W. R. Roush, J. Org. Chem. 1997, 62, 1112.
doi:10.1021/JO961840S
[8] H. Zhao, S. Hans, X. Cheng, D. R. Mooto, J. Org. Chem. 2001,
66, 1761. doi:10.1021/JO001447T
[32] M. J. Schneider, F. S. Ungemach, H. P. Broquist, T. M. Harris,
Tetrahedron 1983, 39, 29. doi:10.1016/S0040-4020(01)97625-2