A synthesis of (1S,2R,8R,8aR)-8-hydroxy-1,2-(isopropylidinedioxy)indolizidin-5-one from D-ribose: Improved access to (-)-swainsonine and its analogs

Article abstract of DOI:10.3987/com-02-s(m)40

A short synthesis of the lactam (4), (1S,2R,8R,8aR)-8-hydroxy-1,2-(isopropylidenedioxy)indolizidin-5-one, from D-ribose is reported. This compound is a useful intermediate for the synthesis of analogs of the anticancer agent swainsonine (1) as well for sw

Full text of DOI:10.3987/com-02-s(m)40

HETEROCYCLES, Vol. 58, 2002, pp. 421-430, Received, 8th April, 2002
A SYNTHESIS OF (1S,2R,8R,8aR)-8-HYDROXY-1,2-
(ISOPROPYLIDINEDIOXY)INDOLIZIDIN-5-ONE FROM D-RIBOSE:
IMPROVED ACCESS TO (-)-SWAINSONINE AND ITS ANALOGS†
William H. Pearson,* Yi Ren,‡ and Jeremiah D. Powers
Department of Chemistry, University of Michigan, Ann Arbor, Michigan, 48109-
1055, USA
e-mail: wpearson@umich.edu
Abstract – A short synthesis of the lactam (4), (1S,2R,8R,8aR)-8-hydroxy-1,2-
(isopropylidenedioxy)indolizidin-5-one, from D-ribose is reported. This compound
is a useful intermediate for the synthesis of analogs of the anticancer agent
swainsonine (1) as well for swainsonine itself.
Swainsonine (1) (Scheme 1) is a potent inhibitor of certain α-mannosidases, and has proven useful as a
biochemical tool for the study of glycoprotein processing, since it inhibits a key late-stage enzyme in the
biosynthesis of glycoproteins.1-6 That enzyme, Golgi α-mannosidase II (GMII), is necessary for the
formation of so-called "complex glycoproteins". The altered distribution of such glycoproteins on the
surface of cancer cells is associated with metastasis and disease progression, hence inhibitors of GMII are
potentially useful for cancer treatment.7-10 More selective inhibition of GMII over other mannosidases
is a desirable goal for cancer drug development, and makes the synthesis of analogs of swainsonine a
significant undertaking.11-13 Further, swainsonine itself is expensive and difficult to obtain in large
quantities, whether by total synthesis or isolation from natural sources.
Many syntheses of swainsonine have been published,14 including the relatively practical routes of
Fleet15,16 and Cha,17 as well as our own.18,19 Nonetheless, improved routes to swainsonine are still
needed, as this important compound is still relatively difficult to make. Perhaps more significantly,
† Dedicated to Professor Albert I. Meyers on the occasion of his 70th birthday.
‡ Current address: Building 66, Room 214, Hoffmann-LaRoche, Inc., 340 Kingsland St., Nutley, NJ
07110-1199.

routes to analogs of swainsonine are needed, including those bearing substituents attached at positions 5,
6, and 7 of the swainsonine core.11,12,20 An attractive compound for preparing such analogs is the
lactam (4), which is often encountered in syntheses of swainsonine itself. We are exploring the
functionalization of 4 at positions 5, 6, and 7, and thus need significant quantities of this compound.11,12
Herein we report an improved route to 4, and thus swainsonine (1) and certain of its analogs, whereby D-
ribose is used as a starting material.
Our previous synthesis of swainsonine (including 4) was quite practical, allowing multigram quantities of
4 and 1 to be prepared (Scheme 1).18 The route began with D-isoascorbic acid, which was converted by
the literature procedure21 to the lactone (3), which can also be purchased (ca. $10/g, Aldrich). This
lactone has been used in other routes to swainsonine, e.g. Cha's.17 After a sequence of 9-10 steps, we
were able to convert 3 into 4. Our need for large quantities of 4 for analog work forced us to examine
improvements to this route, since it still required numerous steps and involved DIBAL-H, t-BuMe2SiCl,
Bu4NF, azide chemistry, five chromatographies, and three crystallizations. We have now developed a
route that uses less exotic and costly reagents, requires only two chromatographies and one crystallization,
and involves fewer steps. Further, the inexpensive starting material D-ribose (5) is used.
Previous route:
O
O
OH
O
O
OH
H
N
O
HO
HO
O
7
O
2 steps
9-10 steps
OH
6
5
O
4
O
D-isoascorbic acid (2)
3
key intermediate for analog synthesis
2 steps
8 steps
New route:
OH
OH
HO
H
HO
OH
7
HO
6
O
N
5
HO
D-ribose (5)
(-)-swainsonine (1)
Scheme 1. Overview of (-)-swainsonine syntheses.

RESULTS AND DISCUSSION
The acetonide (6) of D-ribose (5) was prepared by a modification of the procedure of Levene,22 and was
used without purification in a reaction with vinylmagnesium bromide, providing the triol (7). Mekki23,24
has studied the diastereoselectivity of the addition of organomagnesium reagents to similar carbohydrate-
derived lactols. The oxidative cleavage of the 1,2-diol of 7 to the lactol (8) was challenging, but was
eventually found to be efficient when Shing and Zhong's sodium periodate on silica gel method was
used.25 A reductive amination26 of 8 with dibenzylamine gave the amino alcohol (9), which was
subjected to a Johnson orthoester Claisen rearrangement27 to afford the unsaturated ester (10). At this
OH
O
O
O
BrMg
5 equiv.
O
O
HO
HO
acetone
OH
OH
conc. HCl
THF
HO
OH
O
HO
OH
D-ribose (5)
6
7
O
O
NaIO4
on SiO2
O
O
Bn2NH, AcOH
MeC(OMe)3
cat. EtCO2H
CH2Cl2
NaBH3CN, MeOH
OH
O
toluene
reflux
Bn2N
HO
8
9
O
O
O
O
K3Fe(CN)6
K2OsO4•2H2O
O
O
MsCl, NEt3
K2CO3, MeSO2NH2
(DHQD)2PHAL
H2O, t-BuOH
(58%)
CH2Cl2
(60%)
OH
11
Bn2N
CO2Me
Bn2N
10
43% from D-ribose
O
OH
OH
Pd(OH)2
HCO2NH4
HO
O
H
N
H
O
O
1) BH3•THF
O
O
HO
AcOH
MeOH
reflux
N
2) aq. HCl
(96%)
OMs
Bn2N
O
(80%)
12
4
1
Scheme 2. Synthesis of swainsonine from D-ribose.

point, the first purification in the sequence was employed, providing 10 in 43% overall yield from
D-ribose. A diastereoselective catalytic osmylation of 10 was then carried out using the Sharpless
protocol, affording 60% of 11, 5% of its bis-epimer (not shown), and 4% of starting material,
all easily separated by column chromatography. Mesylation of 11 afforded 12 in 60% yield.
It was important to rid 12 of impurities by repeated washes with aqueous base before a pass through
a short plug of silica gel, since these impurities were found to negatively affect the next reaction. With
pure mesylate in hand, hydrogenolysis of the N-benzyl groups was carried out using the method of
Ram and Spicer28 to produce the desired crystalline lactam (4), a key intermediate in our work on
the synthesis of analogs of swainsonine. The lactam (4) has previously been converted to swainsonine
(1) in good yield.18
In summary, we have achieved a short and efficient synthesis of the lactam (4), and thus swainsonine (1)
and its analogs, from the inexpensive starting material D-ribose (5). Fewer and simpler reactions in
conjunction with less purification steps makes for a significant improvement over our previous
synthesis,18 which was already relatively short and practical.
EXPERIMENTAL
2,3-O-Isopropylidene-D-ribofuranose (6). This compound was prepared by a modification of the
procedure reported by Levene22 wherein the addition of anhydrous copper sulfate was omitted.
Concentrated sulfuric acid (1.3 mL) was added to a suspension of D-ribose (50.0 g) in acetone (500 mL).
The mixture was stirred for 4 h until the white solid dissolved, resulting in a brown solution, which was
then neutralized to pH 7 by the addition of Ca(OH)2 (~10 g) in small portions. This mixture was filtered
through Celite and the filtrate was concentrated to give 63.3 g of crude product, which was dissolved in
ether and filtered. Concentration of the filtrate gave the title compound (63.0 g, 99%), which was used
without further purification. Although this compound has been prepared numerous times, methods for
its purification have not been reported and attempts at purification were unsuccessful in our hands.23,29-34
(2R,3R,4S,5S)-3,4-Isopropylidenedioxy-6-heptene-1,2,5-triol (7). Vinylmagnesium bromide (600 mL
of a 0.96 M solution in THF, 576 mmol) was added in a dropwise fashion via addition funnel to a
solution of 6 (21.9 g, 115 mmol) in THF (500 mL) at 0 °C. The mixture was stirred at rt for 12 h, then

cooled to 0 °C and treated with saturated aqueous NH4Cl (400 mL). After 15 min, the mixture was
concentrated to leave a pink solid, which was suspended in aqueous HCl (0.25 N) and extracted with
EtOAc. The aqueous layer was then acidified to pH 6 with aqueous HCl (6 N) and re-extracted with
EtOAc. The combined organic layers were washed with brine, dried (MgSO4), and concentrated to give
20.2 g (80%) of the title compound as a yellow oil that was used without further purification. A small
portion was purified by chromatography (2.5% MeOH/CH2Cl2) to provide a pure sample of 7 for
characterization. The purified material was found to be a low-melting solid, which was recrystallized
with hexanes/EtOAc. Rf to give white needles = 0.34 (2.5% MeOH/CH2Cl2); mp 62-63 ºC; [α]23D –29.5°
(c 1.03, CHCl3); 1H NMR (CDCl3, 400 MHz) δ 6.00 (ddd, J = 5.5, 10.3, 16.1 Hz, 1 H), 5.39 (dt, J = 1.5,
15.7 Hz, 1 H), 5.28 (dt, J = 1.5, 10.3 Hz, 1 H), 4.77 (s, 1 H), 4.45 (s, 1 H), 4.31 (br t, J = 7.7, 1 H), 4.13
(dd, J = 5.5, 9.5 Hz, 1 H), 4.04 (dd, J = 5.3, 9.9 Hz, 1 H), 3.90 (m, 2 H), 3.69 (m, 1 H), 3.44 (br t, 1 H),
13
1.40 (s, 3 H), 1.33 (s, 3 H); C NMR (100 MHz, CDCl3) δ 137.5, 116.9, 109.0, 80.0, 77.4, 70.5, 69.6,
64.3, 27.8, 25.4; IR (neat) 3682 (m), 3600 (m) cm–1; MS (CI, NH3) m/z (rel intensity) 219 [(M + H)+,
11], 201 (9), 161 (23), 143 (100), 125 (29); HRMS (CI, NH3) calcd for C10H19O5 [(M + H)+] 219.1154
found 219.1238.
(2R/S,3S,4S,5S)-2-Hydroxy-3,4-isopropylidenedioxy-5-vinyltetrahydrofuran (8). Silica gel-supported
sodium periodate was prepared according to the procedure reported by Shing and Zhong.25 Thus,
sodium periodate (114.6 g, 536 mmol) was dissolved in water (215 mL) and preheated to 70 °C, then
silica gel (400 g) was stirred into the solution to form a free flowing solid, which was added to a solution
of 7 (58.4 g 268 mmol) in CH2Cl2 (2.5 L) at rt. After 1.5 h, the mixture was filtered and the filtrate was
concentrated to give 47.7 g (96 %) of the title compound as a yellow oil that was used without further
purification. A small portion was purified by chromatography (2.5% MeOH/CH2Cl2) to provide a pure
sample of 8 as a yellow oil for characterization. This material was found to be a 6:1 mixture of
diastereomers as determined by 400 MHz 1H NMR. Rf = 0.38 (10% MeOH/CHCl3). 1H NMR (400 MHz,
CDCl3; selected data for major diastereomer) δ 6.00 (ddd, J = 8.1, 10.3, 17.2 Hz, 1 H), 5.48 (d, J = 2.9
Hz, 1 H), 5.29 (dt, J = 1.1, 17.6 Hz, 1 H), 5.16 (dt, J = 1.1, 10.3 Hz, 1 H), 4.65 (m, 3 H), 4.12 (br d, 1 H),
1.50 (3 H, s), 1.32 (3 H, s); 13C NMR (100 MHz, CDCl3), δ 137.8, 117.3, 112.4, 102.8, 88.4, 85.9, 84.5,
77.3, 76.7; IR (neat) 3421 (br s) cm–1; MS (CI, NH3) m/z (rel intensity) 204 [(M + NH4)+, 8], 186 (M+,
100), 169 (60); HRMS (CI, NH3) calcd for C9H18NO4 [(M + NH4)+] 204.1236 found 204.1226.

(2R,3S,4S)-1-(N,N-Dibenzylamino)-2,3-isopropylidenedioxyhex-5-ene-4-ol (9). A modification of
the method reported by Borch26 for the reductive amination of aldehydes was used to prepare the title
compound. Dibenzylamine (57.5 g, 56 mL, 291 mmol), glacial acetic acid (31.7 g, 26 mL, 522 mmol),
and sodium cyanoborohydride (44.4 g, 706 mmol) were added to a solution of 8 (45.2 g, 243 mmol) in
MeOH (550 mL) at rt. After 3 d, the mixture was concentrated and the resulting solid was dissolved in
saturated aqueous NaHCO3 and extracted with ether. The organic extract was dried (NaSO4) and
concentrated to ~150 mL, then treated with pyridine (28.8 g, 29.5 mL, 365 mmol) and acetic anhydride
(37.2 g, 34.4 mL, 365 mmol). After 3 h at rt, the solution was concentrated and the resulting oil was
dissolved in 800 mL of a 1:1 MeOH/saturated aqueous K2CO3 solution. After 12 h, the MeOH was
evaporated and the remaining liquid partitioned between ether and water. The aqueous layer was
extracted with ether and the combined organic layers were washed with water and aqueous HCl (0.1 N).
The acidic aqueous layers were collected and extracted with ether, then made basic (pH 8) with aqueous
KOH (2 N). The resulting cloudy white solution was extracted with ether, and the organic phase was
washed with brine, dried (NaSO4), and concentrated to provide 91.9 g (99%) of the title compound as a
yellow oil that was used without further purification. A small portion was purified by chromatography
(5% EtOAc/toluene) to provide a pure sample of 9 as a clear oil for characterization, Rf = 0.5 (25%
1
EtOAc/hexanes); [α]23 –71.02° (c = 1.56, CHCl3); H NMR (400 MHz, CDCl3) δ 7.37-7.26 (m, 10 H),
D
6.67 (s, 1 H), 5.85 (ddd, J = 5.9, 10.6, 17.2 Hz, 1 H), 5.18 (m, 2 H), 4.49 (ddd, J = 2.6, 5.5, 11.0 Hz, 1 H),
4.02 (dd, J = 5.5, 9.5 Hz, 1 H), 3.94 (d, J = 12.9 Hz, 2 H), 3.36 (m, 1 H), 3.27 (d, J = 13.2 Hz, 2 H), 3.09
(dd, J = 11.4, 12.8 Hz, 1 H), 2.36 (dd, J = 2.2, 12.8 Hz, 1 H), 1.35 (s, 3 H), 1.29 (s, 3 H); 13C NMR (100
MHz, CDCl3) δ 137.6, 136.3, 129.8, 128.6, 127.8, 116.5, 108.7, 80.9, 74.1, 69.5, 59.1, 54.1, 28.0, 25.3;
IR (neat) 3172 (br s) cm-1; MS (CI, NH3) m/z (rel intensity) 368 [(M + H)+, 100], 310 (1), 277 (3), 210
(14); HRMS (CI, NH3) calcd for C23H30NO3 [(M + H)+] 368.2226 found 368.2208.
Methyl (E)-(6S,7R)-8-(N,N-dibenzylamino)-6,7-isopropylidenedioxy-4-octenoate (10). Trimethyl
orthoacetate (160.6 g, 170.8 mL, 1.34 mol) and propionic acid (3.9 g, 4.0 mL, 53 mmol) were added to
9 (89.2 g, 243 mmol) in toluene (500 mL) and the mixture was heated at reflux for 16 h. The progress
1
of the reaction was monitored by H NMR spectroscopy. The mixture was then concentrated to provide
a yellow oil which was purified by chromatography (15% EtOAc/hexanes) to give 44.1 g (43% overall
yield from D-ribose) of 10 as a yellow oil, Rf = 0.45 (10% EtOAc/hexanes); [α]23 -33.4° (c = 1.4,
D
CDCl3); 1H NMR (400 MHz, CDCl3) δ 7.36-7.20 (m, 10 H), 5.62 (m, 1 H), 5.34 (dd, J = 8.1, 15.4 Hz, 1

H), 4.42 (t, J = 7.1 Hz, 1 H), 4.32 (dt, J = 4.4, 7.3 Hz, 1 H), 3.77 (d, J = 13 Hz, 2 H), 3.65 (s, 3 H), 3.52 (d,
J = 12.9 Hz, 2 H), 2.61 (dd, J = 4.0, 13.6 Hz, 1 H), 2.52 (dd, J = 7.3, 13.9 Hz, 1 H), 2.30 (br s, 4 H), 1.42
13
(s, 3 H), 1.31 (s, 3 H); C NMR (100 MHz, CDCl3) δ 172.9, 139.2, 132.4, 128.8, 128.0, 127.1, 126.7,
108.0, 78.8, 76.8, 58.7, 53.9, 51.5, 33.4, 28.1, 27.4, 25.6; IR (neat) 1739 (s) cm–1; MS (EI, 70 eV) m/z (rel
intensity) 423 (M+, 11), 306 (10), 210 (100), 181 (5), 91 (70); HRMS (EI, 70 eV) calcd for C26H33NO4
423.2409 found 423.2395.
(5R)-5-[(1S,2R,3R)-4-(N,N-Dibenzylamino)-1-hydroxy-2,3-isopropylidenedioxybutyl]tetrahydro-
furan-2-one (11) and (5S)-5-[(1R,2R,3R)-4-(N,N-Dibenzylamino)-1-hydroxy-2,3-isopropylidene-
dioxybutyl]tetrahydrofuran-2-one. The method of Sharpless35 for the asymmetric dihydroxylation of
alkenes was used to prepare the title compound. Potassium ferricyanide (9.58 g, 29.1 mmol), potassium
carbonate (4.02 g, 29.1 mmol), methanesulfonamide (0.92 g, 9.7 mmol), potassium osmate dihydrate (36
mg, 0.097 mmol), and (DHQD)2PHAL (75 mg, 0.097 mmol) were added to a mixture of 3:1 water/t-
butanol (85 mL) at 0 °C, followed by a solution of 10 (4.11 g, 9.7 mmol) in tert-butanol (45 mL). After
48 h at rt, sodium sulfite (15.9 g, 126 mmol) was added and the mixture was stirred for 1 h then extracted
with EtOAc. The combined organic layers were washed with aqueous KOH (2 M), dried (Na2SO4), and
concentrated. Chromatography (20% EtOAc/hexanes) gave 2.44 g (60%) of 11 as a white foam, 0.22g
(5%) of (5S)-5-[(1R,2R,3R)-4-(N,N-dibenzylamino)-1-hydroxy-2,3-isopropylidenedioxybutyl]tetrahydro-
furan-2-one as a white foam, and 0.15 (4%) of recovered starting material. Data for 11: Rf = 0.45
(40% EtOAc/hexanes); [α]23 –63.2° (c = 0.73, CDCl3); 1H NMR (400 MHz, CDCl3) δ 7.38-7.24 (m, 10
D
H), 4.52 (m, 2 H), 4.36 (dd, J = 5.9, 9.9 Hz, 1 H), 3.96 (d, J = 12.9 Hz, 2 H), 3.18 (d, J = 13.2 Hz, 2 H),
2.99 (dd, J = 11.4, 12.8 Hz, 1 H), 2.71 (m, 2 H), 2.37 (m, 2 H), 2.06 (m, 2 H), 1.33 (s, 3 H), 1.30 (s, 3 H);
13C NMR (100MHz, CDCl3) δ 178.1, 136.2, 130.0, 128.6, 127.9, 108.7, 79.4, 76.8, 73.4, 70.0, 59.2, 53.6,
37.1, 28.5, 28.0, 25.2, 23.6; IR (neat) 3200 (br s), 1771 (s) cm-1; MS (EI, 70 eV) m/z (rel intensity) 426
[(M + H)+, 1], 340 (8), 210 (100), 91 (55); HRMS (CI, NH3) calcd for C25H32NO5 [(M + H)+] 426.2280
found 426.2283. Data for (5S)-5-[(1R,2R,3R)-4-(N,N-dibenzylamino)-1-hydroxy-2,3-isopropylidene-
dioxybutyl]tetrahydrofuran-2-one: Rf = 0.28 (40% EtOAc/hexanes); [α]23D –24.9° (c = 2.25, CDCl3); 1H
NMR (400 MHz, CDCl3) δ 7.38-7.24 (m, 10 H), 4.42 (dd, J = 7.1, 12.9 Hz, 1 H), 4.36 (dt, J = 3.8, 6.6 Hz,
1 H), 3.93 (dd, J = 2.7, 6.6 Hz, 1 H), 3.82 (d, J = 12.8 Hz, 2 H), 3.54 (d, J = 13.1 Hz, 2 H), 2.95 (m, 2 H),
2.77 (dd, J = 3.8, 14.0 Hz, 1 H), 2.58-2.34 (m, 2 H), 2.14 (m, 1 H), 1.93 (m, 1 H), 1.51 (s, 3 H), 1.30 (s, 3
H); 13C NMR (100 MHz, CDCl3) δ 177.1, 137.9, 129.6, 128.4, 127.4, 108.6, 80.7, 76.5, 75.7, 70.4, 59.9,

51.7, 28.4, 26.7, 25.1, 23.9; IR (CHCl3) 3532 (br s), 1773 (s), 1732 (m) cm-1; MS (CI, NH3) m/z (rel
intensity) 426 [(M + H)+, 100], 334 (3), 210 (14), 198 (17), 91 (4); HRMS (CI, NH3) calcd for C25H32NO5
[(M + H)+] 426.2280 found 426.2265.
(5S)-5-[(1R,2R,3R)-4-(N,N-Dibenzylamino)-2,3-isopropylidenedioxy-1-
methanesulfonyloxybutyl]tetrahydrofuran-2-one (12). Methanesulfonyl chloride (3.68 g, 2.48 mL, 32
mmol) and triethylamine (3.69 g, 5.08 mL, 36.5 mmol) were added sequentially to a solution of 11 (6.20
g, 14.6 mmol) in CH2Cl2 (100 mL) at 0 °C. After warming to rt for 4 h, the mixture was diluted with
water and extracted with EtOAc. The organic layers were combined, washed with saturated aqueous
NaHCO3 (15x), dried (Na2SO4), and concentrated to give a pale yellow foam. Before the product could
be carried on to the next reaction, all of the mesylate impurities had to be removed with basic aqueous
1
washes, which were done until the singlet at δ 3.14 (s) in the H NMR (400 MHz, CDCl3) spectrum had
disappeared. Finally, passing the crude product through a silica plug (25% EtOAc/hexanes) gave 4.19 g
(60%) of the title compound as a white foam, Rf = 0.27 (40% EtOAc/hexanes); [α]23 +17.6° (c = 1.07,
D
CHCl3); 1H NMR (400 MHz, CDCl3) δ 7.38 (br d, 4 H), 7.26 (br t, 4 H), 7.17 (br t, 2 H), 4.80 (dd, J = 5.5,
4.0 Hz, 1 H), 4.60 (m, 1 H), 4.45 (br s, 1 H), 4.17 (t, J = 5.7 Hz, 1 H), 3.66 (br s, 4 H), 3.00 (s, 1 H), 2.78
(s, 3 H), 2.64 (dd, J = 9.2, 13.2 Hz, 1 H), 2.52 (ddd, J = 6.2, 10.3, 17.9 Hz, 1 H), 2.35 (m, 1 H), 2.16 (m,
1 H), 2.06 (m, 1 H), 1.28 (s, 3 H), 1.22 (s, 3 H); 13C NMR (100 MHz, CDCl3) δ 176.0, 129.1, 128.2, 127.0,
108.7, 79.2, 77.8, 76.0, 58.8, 53.3, 38.9, 37.1, 31.5, 27.6, 27.5, 25.5, 24.0; IR (neat) 2800 (s), 1784 (s),
1360 (s), 1178 (s); MS (CI, NH3) m/z (rel intensity) 504 [(M + H)+, 4], 318 (100), 293 (6), 228 (7), 204
(46), 108 (8); HRMS (CI, NH3) calcd for C26H34NO7S [(M + H)+] 504.2056 found 504.2040.
(1S,2R,8R,8aR)-8-Hydroxy-1,2-(isopropylidenedioxy)indolizidin-5-one (4). A modification of the
method reported by Ram and Spicer28 for the debenzylation of N-benzylamines was used to prepare the
title compound. Ammonium formate (4.53 g, 72 mmol), 20 wt.% Pd(OH)2/C (500 mg), and glacial
acetic acid (3.15 g, 3 mL) were added to a solution of 12 (4.55 g) in MeOH (200 mL). After heating at
reflux for 18 h, the mixture was cooled and filtered through Celite. The filtrate was concentrated and the
resulting solid was dissolved in a minimal amount of water and the pH was adjusted to 8 with aqueous
NaOH (1 N). The water was removed in vacuo and the residue was extracted with CHCl3. The
remaining solid was dissolved in a minimal amount of water and also extracted with CHCl3. The
organic layers were then combined, dried (NaSO4), and concentrated. Upon standing at rt for 12 h, the

resulting oil became a low-melting orange solid, which was recrystallized from EtOAc to provide 1.63 g
(80%) of the title compound as white rhomboid crystals, Rf = 0.3 (10% MeOH/EtOAc); mp 129-130 ºC
(lit., 126-128 ºC,15 129 ºC18); 1H NMR (400 MHz, CDCl3) δ 4.81 (dd, J = 4.5, 6.0 Hz, 1 H), 4.75 (dd, J =
5.0, 5.2 Hz, 1 H), 4.20 (d, J = 11.0 Hz, 1 H), 4.15 (m, 1 H), 3.34 (dd, J = 4.8, 8.4 Hz, 1 H), 3.14 (dd, J =
4.8, 13.5 Hz, 1 H), 2.53 (ddd, J = 2.5, 6.2, 17.9 Hz, 1 H), 2.43 (m, 1 H), 2.14 (m, 1 H), 2.05 (br s, 1 H),
1.88 (m, 1 H), 1.43 (s, 3 H), 1.34 (s, 3 H); 13C NMR (100 MHz, CDCl3) δ 168.2, 112.4, 80.0, 77.8, 66.3,
65.7, 50.8, 29.9, 29.9, 26.5, 24.9.
ACKNOWLEDGEMENTS
We thank the National Institutes of Health (CA77365) for support of this work.
REFERENCES AND NOTES
1.
2.
3.
4.
5.
6.
A. D. Elbein, FASEB J., 1991, 5, 3055.
G. P. Kaushal and A. D. Elbein, Methods in Enzymology, 1994, 230, 316.
K. W. Moremen, R. B. Trimble, and A. Herscovics, Glycobiology, 1994, 4, 113.
B. Ganem, Acc. Chem. Res., 1996, 29, 340.
Iminosugars as Glycosidase Inhibitors; ed. by A. E. Stütz, Wiley-VCH: Weinheim, 1999.
N. Asano, R. J. Nash, R. J. Molyneux, and G. W. J. Fleet, Tetrahedron: Asymmetry, 2000, 11,
1645.
7.
P. E. Goss, M. A. Baker, J. P. Carver, and J. W. Dennis, Clin. Cancer Res., 1995, 1, 935.
P. E. Goss, C. L. Reid, D. Bailey, and J. W. Dennis, Clin. Cancer Res., 1997, 3, 1077.
J. W. Dennis, M. Granovsky, and C. E. Warren, BioEssays, 1999, 21, 412.
J. W. Dennis, M. Granovsky, and C. E. Warren, Biochim. Biophys. Acta, 1999, 1473, 21.
W. H. Pearson, L. Guo, and T. M. Jewell, Tetrahedron Lett., 2002, 43, 2175.
W. H. Pearson and E. J. Hembre, Tetrahedron Lett., 2001, 42, 8274.
W. H. Pearson and L. Guo, Tetrahedron Lett., 2001, 42, 8267.
8.
9.
10.
11.
12.
13.
14.
Dozens of syntheses of swainsonine have been published; too many to review here. For a recent
example, see (a) H. Zhao, S. Hans, X. Cheng, and D. R. Mootoo, J. Org. Chem., 2001, 66, 1761.
For a review of the synthesis of swainsonine and its analogs, see A. El Nemr, Tetrahedron, 2000,
56, 8579.
15.
N. M. Carpenter, G. W. J. Fleet, I. Cenci di Bello, B. Winchester, L. E. Fellows, and R. J. Nash,

Tetrahedron Lett., 1989, 30, 7261.
16.
17.
G. W. J. Fleet, U. S. Patent 5,023,340, 1991 (Chem. Abstr., 1991, 115, 136472).
R. B. Bennett, III, J.-R. Choi, W. D. Montgomery, and J. K. Cha, J. Am. Chem. Soc., 1989, 111,
2580.
18.
19.
20.
W. H. Pearson and E. J. Hembre, J. Org. Chem., 1996, 61, 7217.
W. H. Pearson and E. J. Hembre, U. S. Patent 5,919,952, 1999 (Chem. Abstr., 1999, 131, 73836).
For related work on analogs of swainsonine, see ref. 13 above and: (a) E. J. Hembre and W. H.
Pearson, Tetrahedron, 1997, 53, 11021. (b) W. H. Pearson and E. J. Hembre, J. Org. Chem.,
1996, 61, 5537.
21.
N. Cohen, B. L. Banner, A. J. Laurenzano, and L. Carozza, In Organic Syntheses, Collective Vol.
IV; ed. by J. P. Freeman, Wiley, New York, 1990; p. 297.
22.
23.
24.
P. A. Levene and R. S. Tipson, J. Biol. Chem., 1936, 115, 73.
B. Mekki, G. Singh, and R. H. Wightman, Tetrahedron Lett., 1991, 32, 5143.
Stork and Raucher added vinylmagnesium bromide to 2,3-isopropylidene-L-erythrose, then
carried out a Johnson orthoester Claisen rearrangement on the resultant allylic alcohol in a fashion
similar to the strategy we have used in the current work. They did not study the stereoselectivity
of the Grignard reaction. See: G. Stork and Raucher, J. Am. Chem. Soc., 1976, 98, 1583.
T. K. M. Shing and Y.-L. Zhong, J. Org. Chem., 1997, 62, 2622.
25.
26.
27.
R. F. Borch, M. D. Bernstein, and H. D. Durst, J. Am. Chem. Soc., 1971, 93, 2897.
W. S. Johnson, L. Werthemann, W. R. Bartlett, T. J. Brocksom, T. Li, D. J. Faulkner, and M. R.
Petersen, J. Am. Chem. Soc., 1970, 92, 741.
28.
29.
30.
31.
32.
33.
S. Ram and L. D. Spicer, Tetrahedron Lett., 1987, 28, 515.
K. Mori and H. Kikuchi, Liebigs Ann. Chem., 1989, 1267.
B. Kaskar, G. L. Heise, R. S. Michalak, and B. R. Vishnuvajjala, Synthesis, 1990, 1031.
M. S. Cai, D. X. Qiu, D. Y. He, Z. J. Li, and Y. Q. Liu, Synth. Commun., 1992, 22, 2653.
Y. Nishimura, W.-m. Wang, T. Kudo, and S. Kondo, Bull. Chem. Soc. Jpn., 1992, 65, 978.
A. P. Rauter, F. Ramoa-Ribeiro, A. C. Fernandes, and J. A. Figueiredo, Tetrahedron, 1995, 51,
6529.
34.
35.
R. Hunter and R. Clauss, J. Chem. Soc., Perkin Trans. 1, 1997, 71.
H. C. Kolb, M. S. VanNieuwenhze, and K. B. Sharpless, Chem. Rev., 1994, 94, 2483.