Improved syntheses of the naturally occurring glycosidase inhibitor salacinol

Article abstract of DOI:10.1055/s-2003-40353

Improved syntheses of the naturally occurring sulfonium ion, salacinol are described. Salacinol is one of the active principles in the aqueous extracts of Salacia reticulata that are traditionally used in Sri Lanka and India for the treatment of Type 2 Diabetes. The synthetic strategy relies on the nucleophilic attack of 2,3,5-triO-benzyl- or 2,3,5-tri-O-p-methoxybenzyl-1,4-anhydro-4-thio-D-arabinitol at the least hindered carbon of benzylidene-protected L-erythritol-1,3-cyclic sulfate in 1,1,1,3,3,3-hexafluoro-2-propanol as solvent. The reactions are compared to those with the benzyl-protected L-erythritol-1,3-cyclic sulfate and also to those in acetone and 2-propanol. Excellent yields are obtained for the reactions with the benzylidene-protected cyclic sulfate. The synthetic route employing p-methoxybenzyl ether protecting groups is advantageous since all protecting groups in the adduct may be removed with trifluoroacetic acid to yield salacinol, thereby obviating the problematic deprotection of benzyl ethers by hydrogenolysis.

Full text of DOI:10.1055/s-2003-40353

LETTER
1259
Improved Syntheses of the Naturally Occurring Glycosidase Inhibitor
Salacinol
I
mprove
d
Synthese
h
s
of
G
lycos
m
a
inol d Ghavami,a Kashinath S. Sadalapure,a Blair D. Johnston,a Mercedes Lobera,b Barry B. Snider,b
B. Mario Pinto*a
a
Department of Chemistry, Simon Fraser University, Burnaby, B.C., V5A 1S6 Canada
b
Department of Chemistry, Brandeis University, MS#015, 415 South Street, Box 9110 Waltham, Massachusetts 02454-9110, USA
Fax +1(604)2915424; E-mail: bpinto@sfu.ca
Received 9 October 2002
This work is dedicated, with respect, to the memory of Raymond U. Lemieux.
OH
Abstract: Improved syntheses of the naturally occurring sulfonium
ion, salacinol are described. Salacinol is one of the active principles
in the aqueous extracts of Salacia reticulata that are traditionally
OH
S
OSO3
used in Sri Lanka and India for the treatment of Type 2 Diabetes.
The synthetic strategy relies on the nucleophilic attack of 2,3,5-tri-
O-benzyl- or 2,3,5-tri-O-p-methoxybenzyl-1,4-anhydro-4-thio-D-
arabinitol at the least hindered carbon of benzylidene-protected
L-erythritol-1,3-cyclic sulfate in 1,1,1,3,3,3-hexafluoro-2-propanol
as solvent. The reactions are compared to those with the benzyl-pro-
tected L-erythritol-1,3-cyclic sulfate and also to those in acetone
and 2-propanol. Excellent yields are obtained for the reactions with
the benzylidene-protected cyclic sulfate. The synthetic route em-
ploying p-methoxybenzyl ether protecting groups is advantageous
since all protecting groups in the adduct may be removed with tri-
fluoroacetic acid to yield salacinol, thereby obviating the problem-
atic deprotection of benzyl ethers by hydrogenolysis.
HO
HO
OH
Salacinol 1
Figure 1
The key step in the published syntheses of salacinol (1)5,6
is the ring opening reaction of a cyclic sulfate by nucleo-
philic attack of the ring sulfur atom of 1,4-anhydro-4-thio-
D-pentitol (2) (Scheme 1). The alkylation reaction involv-
ing these partners is critically dependent on the protecting
groups on the cyclic sulfate. Thus, the unoptimized reac-
tion of the per-benzylated thioether 2 with benzylidene-
protected cyclic sulfate 3 in acetone, containing potassium
carbonate, proceeded in 33% yield (Scheme 1).5 A similar
yield was obtained in the reaction with the monobenzylat-
ed thioether 5.5 Reaction of the unprotected thioether 7
with the isopropylidenated-cyclic sulfate 8 in DMF pro-
ceeded in 61% yield to give 9, although its reaction with
the corresponding benzylated-cyclic sulfate 10 did not
proceed.6 The latter derivative 10 is clearly a much less re-
active alkylating agent than 8. Significant decomposition
of the cyclic sulfates 8 and 10 at temperatures of 60–70 ºC
in DMF was also observed.6 Deprotection of 9 proceeded
in 75% yield to afford salacinol 1 in 46% overall yield.6
Key words: glycosidase inhibitors, salacinol, efficient synthesis,
sulfonium salt, cyclic sulfate
Salacinol (1) is a potent glycosidase inhibitor isolated
from the aqueous extracts of Salacia reticulata that
are used in Sri Lanka and India for the treatment of diabe-
tes.1–3 The molecular structure of this inhibitor is unique
in that it contains a sulfonium ion (1,4-anhydro-4-thio-D-
pentitol cation) stabilized by an internal sulfate counterion
(1′-deoxy-L-erythrosyl-3′-sulfate anion). Glycosidase in-
hibitors containing sulfonium ions are of interest as mim-
ics of oxacarbenium intermediates in glycosidase
hydrolysis reactions.4–6 In this regard, we5,7 and others6
have previously reported the syntheses of salacinol (1)
and its stereoisomers. In the search for novel glycosidase
inhibitors, we have also reported the synthesis and gly-
cosidase inhibitory properties of the heteroatom conge-
ners of salacinol in which the ring sulfur atom has been
substituted by the cognate atoms nitrogen8 and selenium.9
We report here an improved method for the synthesis of
the natural product salacinol (1) that exploits an unusual
solvent effect provided by 1,1,1,3,3,3-hexafluoroisopro-
panol (HFIP). The previous syntheses of salacinol and its
stereoisomers employed either acetone5,7 or DMF6 as sol-
vents.
The biological importance of salacinol (1)1–3 prompted us
to investigate a more efficient method for its synthesis.
The Hughes–Ingold rules indicate that the SN2 reaction
between a neutral nucleophile, such as 2 or 5, and a neutral
electrophile, such as 3, 8 or 10, should show a large in-
crease in rate on increasing solvent polarity.10 1,1,1,3,3,3-
Hexafluoroisopropanol (HFIP) has a higher normalized
Dimroth–Reichardt
solvent
polarity
parameter,
ETN = 1.068, than water, ETN = 1.00.10 In contrast, the ET
values for acetone and DMF are only 0.355 and 0.404, re-
spectively. Furthermore, HFIP is volatile (bp = 59 °C),
thus facilitating product purification. Preliminary studies
indicated that tetrahydrothiophene reacted cleanly with 3
and 10 in HFIP at 45 °C for 2 days to give the desired
alkylation products in >90% yield.11
N
Synlett 2003, No. 9, Print: 11 07 2003.
Art Id.1437-2096,E;2003,0,09,1259,1262,ftx,en;S06102ST.pdf.
© Georg Thieme Verlag Stuttgart · New York
ISSN 0936-5214

1260
A. Ghavami et al.
LETTER
C6H5
O
O
3`
4`
1`
2`
5
OSO3
S
S
4
3
O
O
1
2
BnO
BnO
O
O
Ph
Acetone, K2CO3
O
+
S
O
Ref. 5
33%
BnO
OBn
BnO
OBn
3
4
2
C6H5
O
O
OSO3
S
S
O
O
HO
HO
O
O
Ph
Acetone, K2CO3
O
+
S
O
Ref. 5
32%
BnO
OH
BnO
OH
3
6
5
O
O
S
S
OSO3
O
HO
HO
O
HO
DMF
61%
O
O
O
+
O
S
O
Ref. 6
HO
HO
OH
HO
OH
9
8
7
S
O
O
BnO
+ BnO
DMF
Ref. 6
S
O
no reaction
OH
7
10
Scheme 1
Therefore, a systematic evaluation of the role of solvent in the cyclic sulfate containing an acetal protecting group
the alkylation reactions of 2 with benzyl- or benzylidene- (Scheme 1). Thus, the reactions of the benzylidene-pro-
protected cyclic sulfates 10 or 3, respectively was under- tected cyclic sulfate 3 in acetone and HFIP, containing
taken. The reactions were carried out in acetone and potassium carbonate, under identical conditions of con-
hexafluoroisopropanol (HFIP) concurrently under identi- centration, temperature, and duration were examined next
cal conditions of concentration, temperature, and duration (Scheme 2). The alkylation reaction of 2 with 3 in acetone
(Scheme 2). Reaction of the thioether 2 (1 equiv) and the proceeded with a dramatic increase in the yield (59%) of
cyclic sulfate 10 (1.2 equiv) in acetone containing K2CO3 the alkylated product 4 relative to the reaction with 10.
at 75–80 °C in a sealed tube proceeded very slowly and The improvement from our earlier reported unoptimized
yielded the desired alkylated product 11 in only 5% yield; yield of 33%5 is due to the use of a more concentrated re-
the remainder of the starting materials was recovered. action mixture.13
Prolonged heating and use of excess cyclic sulfate did not
More significantly, the desired product 4 was obtained in
improve the yields. In addition, when excess cyclic sulfate
94% yield when the reaction was performed in HFIP.13
8 was used, its slow decomposition complicated the puri-
Higher temperatures (>80 °C) and prolonged reaction
times led to the decomposition of the cyclic sulfate, al-
though the stability of the cyclic sulfate was greater in the
presence of K2CO3.
fication of the product 11 formed. However, the analo-
gous reaction between 2 and the cyclic sulfate 10 in HFIP
yielded the adduct 11 in 45% yield, with recovery of the
unreacted starting materials (Scheme 2).12 It is notewor-
The increased yields in HFIP may be accounted for by
better solvation of the transition states for the reactions
thy that the analogous reaction between 2 and the cyclic
sulfate 10 in the polar, protic solvent 2-propanol at 83 °C
and of the adducts. The increased reactivity of the cyclic
for 26 h did not yield any desired product, the starting ma-
sulfate with the benzylidene protecting group (3) may be
terials being recovered. It would appear, therefore, that it
accounted for by the relief of ring strain accompanying
is the highly polar nature of HFIP that is critical in facili-
the reaction, unlike in the corresponding reaction of the
tating this reaction.
benzyl-protected cyclic sulfate 10. Finally, the reaction of
the thioether 7 (not containing protecting groups) with the
benzylidene-protected cyclic sulfate 3 in HFIP was exam-
The previous results of Yuasa et al.6 had indicated a far
lesser reactivity of the benzylated cyclic sulfate relative to
Synlett 2003, No. 9, 1259–1262 ISSN 0936-5214 © Thieme Stuttgart · New York

LETTER
Improved Syntheses of Glycosidase Inhibitor Salacinol
1261
OH
OBn
OBn
OH
OSO3
S
S
OSO3
HO
BnO
Pd(OH)2
a) 5%
2
+
10
b) 45%
80% aq.AcOH
HO
OH
BnO
OBn
1
11
C6H5
O
OH
O
OH
OSO3
S
S
OSO3
HO
BnO
Pd(OH)2
a) 59%
b) 94%
2
+
3
80% aq.AcOH
HO
OH
BnO
OBn
1
4
a) Acetone, K2CO3 b) HFIP, K2CO3
Scheme 2
ined. At 60 °C, decomposition of the cyclic sulfate was chemistry was confirmed by means of NOESY
observed, with no significant formation of the desired experiments that showed clear correlations between H-4
coupled product. Hydrogenolysis of the protected deriva- and H-1′, thus indicating the presence of the isomer with
tives 45 and 11 afforded salacinol (1), although this step a trans relationship between C-5 and C-1′. The barrier to
was problematic because of poisoning of the catalyst, and inversion at the sulfonium ion center must be substantial
only afforded the product in 65% yield.
since no evidence for isomerization in these and related
derivatives7 has been noted.
In order to obviate the problematic hydrogenolysis step,
we next chose to examine the reaction of the thioether
containing p-methoxybenzyl ether protecting groups with
the benzylidene-protected L-erythritol-1,3-cyclic sulfate;
we reasoned that the removal of all protecting groups by
acid hydrolysis would be facile. Thus, 2,3,5-tri-O-p-meth-
oxybenzyl-1,4-anhydro-4-thio-D-arabinitol (12),14 syn-
thesized in 87% yield from 7, was reacted with cyclic
sulfate 3 in HFIP to afford the sulfonium salt 13 in quan-
titative yield (Scheme 3).15 Deprotection of 13 proceeded
smoothly (86%) in aqueous trifluoroacetic acid to afford
salacinol 1 in 75% overall yield.16 The latter sequence rep-
resents, therefore, an efficient synthesis of the biological-
ly important natural product salacinol 1.
Acknowledgment
We are grateful to the Natural Sciences and Engineering Research
Council of Canada for financial support and to the Michael Smith
Foundation for Health Research for a postgraduate traineeship (to
A.G.).
References
(1) Yoshikawa, M.; Murakami, T.; Shimada, H.; Matsuda, H.;
Yamahara, J.; Tanabe, G.; Muraoka, O. Tetrahedron Lett.
1997, 38, 8367.
(2) Yoshikawa, M.; Murakami, T.; Yashiro, K.; Matsuda, H.
Chem. Pharm. Bull. 1998, 46, 1339.
As a final point of interest, we comment on the stereo-
chemistry at the stereogenic sulfonium center in 4, 11, and
13. These reactions proceeded stereoselectively irrespec-
tive of the solvent used in the reaction. The stereo-
(3) Yoshikawa, M.; Morikawa, T.; Matsuda, H.; Tanabe, G.;
Muraoka, O. Bioorg. Med. Chem. 2002, 10, 1547.
(4) Svansson, L.; Johnston, B. D.; Gu, J.-H.; Patrick, B.; Pinto,
B. M. J. Am. Chem. Soc. 2000, 122, 10769.
C6H5
O
O
OH
OH
S
TFA
86%
HFIP
S
OSO3
S
OSO3
O
PMBO
+
C6H5
O
O
PMBO
HO
O
O
quant.
S
O
PMBO
OPMB
PMBO
OPMB
HO
OH
13
12
1
3
PMB =
OCH3
Scheme 3
Synlett 2003, No. 9, 1259–1262 ISSN 0936-5214 © Thieme Stuttgart · New York

1262
A. Ghavami et al.
LETTER
attain room temperature and further stirred for 1 h before
heating to 55 °C for 12 h. The reaction mixture was cooled
and poured in to ice-water (150 mL) and extracted with Et2O
(150 mL). The organic solution was dried (Na2SO4) and
concentrated. The product was purified by column
(5) Ghavami, A.; Johnston, B. D.; Pinto, B. M. J. Org. Chem.
2001, 66, 2312.
(6) Yuasa, H.; Takada, J.; Hashimoto, H. Tetrahedron Lett.
2000, 41, 6615.
(7) Ghavami, A.; Johnston, B. D.; Maddess, M. D.; Chinapoo,
S. M.; Jensen, M. T.; Svensson, B.; Pinto, B. M. Can. J.
Chem. 2002, 80, 937.
(8) Ghavami, A.; Johnston, B. D.; Jensen, M. T.; Svensson, B.;
Pinto, B. M. J. Am. Chem. Soc. 2001, 123, 6268.
chromatography [hexanes–EtOAc, 7:3] to give a colorless
syrup (2.96 g, 87%). [α]D +6 (c 1, CHCl3); 1H NMR
(CDCl3): δ = 7.20–6.80 (12 H, m, Ar), 4.55 (2 H, s, CH2Ph),
4.48 and 4.45 (2 H, 2d, JA,B = 11.7 Hz, CH2Ph), 4.42 and
4.39 (2 H, 2d, JA,B = 12.0 Hz, CH2Ph), 4.13 (1 H, dd,
J1a,2 = 4.6, J2,3 = 9.1 Hz, H-2), 4.05 (1 H, dd, J2,3 = J3,4 = 3.7
Hz, H-3), 3.81 (3 H, s, OCH3), 3.79 (3 H, s, OCH3), 3.76 (3
H, s, OCH3), 3.64 (1 H, dd, J5a,5b = 8.9, J4,5a = 7.5 Hz, H-5a),
3.50 (1 H, ddd, J4,5b = 6.3 Hz, H-4), 3.45 (1 H, dd, H-5b),
3.04 (1 H, dd, J1a,1b = 11.4, J1a,2 = 5.2 Hz, H-1a), 2.85 (1 H,
dd, H-1b). 13C NMR (CDCl3): δ = 159.24, 159.16 (3 × Cpara),
130.31, 130.19, 130.01 (3 Cipso), 129.48, 129.28, 129.22 (6 ×
Cortho), 113.80, 113.74 (6 × Cmeta), 84.77 (C-3), 84.70 (C-2),
72.66, 71.49, 71.20 (3 × CH2Ph), 72.15 (C-5), 55.24
(9) Johnston, B. D.; Ghavami, A.; Jensen, M. T.; Svensson, B.;
Pinto, B. M. J. Am. Chem. Soc. 2002, 124, 8245.
(10) (a) Reichardt, C. Solvents and Solvent Effects in Organic
Chemistry; VCH: Weinheim, 1996, 2nd Ed., 137–147.
(b) Reichardt, C. Solvents and Solvent Effects in Organic
Chemistry; VCH: Weinheim, 1996, 2nd Ed., 359–384.
(11) Snider, B. B.; Lobera, M. unpublished results.
(12) 2,3,5-Tri-O-benzyl-1,4-dideoxy-1,4-{[(2S,3S)-2,4-di-
(benzyloxy)-3-sulfooxy)butyl]- episulfoniumylidene}-D-
arabinitol Inner Salt (11).
A mixture of the thioether 25 (270 mg, 0.64 mmol)and 2,4-
di-O-benzyl-1,3-cyclic sulfate(10)6,9 (280 mg, 0.77 mmol) in
either acetone or HFIP (0.5 mL), containing anhydrous
K2CO3 (16 mg, 0.10 mmol) was stirred in a sealed tube in an
oil-bath (75–80 °C) for 14h. The solvent was removed under
reduced pressure and the residue was purified by column
chromatography using (CH2Cl2–MeOH, 10:1) as eluant to
give the title compound 11, as an amorphous solid (29 mg,
5%) in acetone and (229 mg, 45%) in HFIP. Rf 0.40
(CH2Cl2–MeOH, 10:1); [α]D – 26 (c 1.3, CHCl3); 1H NMR
(CDCl3): δ = 7.38–7.05 (25 H, m, Ar), 4.67 and 4.45 (2 H,
2d, JA,B = 11.8 Hz, CH2Ph), 4.60 and 4.45 (2 H, 2d,
JA,B = 9.5 Hz, CH2Ph), 4.59 and 4.44 (2 H, 2d, JA,B = 11.2
Hz, CH2Ph), 4.58 (1 H, dt, J2′,3′ = 5.0 Hz, H-3′), 4.42 and
4.28 (2 H, 2d, JA,B = 11.0 Hz, CH2Ph), 4.36 (1 H, m, H-2),
4.32 (1 H, ddd, J = 1.7, 4.1, 6.3 Hz, H-2′), 4.30 and 4.20 (2
H, 2d, JA,B = 11.7 Hz, CH2Ph), 4.23 (1 H, m, H-3), 4.13 (1
H, dd, J1′a,1′b = 13.4, J1′a,2′ = 2.0 Hz, H-1′a), 4.05 (1 H, d,
(3 × OCH3), 48.96 (C-4), 33.07 (C-1). Anal. Calcd for
C29H34O6S: C, 68.21; H, 6.71. Found: C, 67.99; H, 6.69.
(15) 2,3,5-Tri-O-p-Methoxybenzyl-1,4-dideoxy-1,4-{[(2S,3S)-
2,4-benzylidenedioxy-3-(sulfooxy)butyl]-episulfonium-
ylidene}-D-arabinitol Inner Salt (13). A mixture of the
thioether 12 (1.50 g, 2.94 mmol), and the cyclic sulfate 3
(0.96 g, 1.2 equiv) in HFIP (2.5 mL) containing anhydrous
K2CO3 (30 mg) was stirred in a sealed tube in an oil-bath
(55°C) overnight. TLC analysis (CH2Cl2–MeOH, 10:1)
showed that the thioether 12 was completely consumed. The
solvent was removed under reduced pressure and the product
was purified by column chromatography (gradient of
CH2Cl2 to CH2Cl2–MeOH, 10:1) to give compound 13 (2.3
g, 100%) as a colorless foam. [α]D –10.5 (c 1.1, CH2Cl2); 1H
NMR (CD2Cl2): δ = 7.51–6.81 (17 H, m, Ph), 5.53 (1 H, s,
C6H5CH), 4.57 (1 H, ddd, J2′,3′ = J3′,4′ax = 10.0, J3′,4′eq = 5.5
Hz, H-3′), 4.49 (1 H, dd, J4′ax,4′eq = 10.8 Hz, H-4′eq), 4.44 (2
H, s, CH2Ph), 4.42–4.39 (1 H, m, H-2), 4.39 and 4.29 (2 H,
2d, JA,B = 11.4 Hz, CH2Ph), 4.33 (1 H, dd, J1′a,1′b = 13.4,
J2,3 = 13.3 Hz, H-1a), 4.00 (1 H, dd, J4′a,4′b = 11.1, J3′,4′a = 2.7
Hz, H-4′a), 3.86 (1 H, dd, J3′,4′b = 2.4, J4′a,4′b = 11.3 Hz, H-
J1′a,2′ = 2.6 Hz, H-1′a), 4.29–4.26 (1 H, m, H-3), 4.26 (1 H,
4′b), 3.71 (1 H, br t, J = 9.2 Hz, H-4), 3.69 (1 H, dd,
ddd, H-2′), 4.19 and 4.09 (2 H, 2d, JA,B = 11.5 Hz, CH2Ph),
4.03 (1 H, br d, J1a,2 <1 Hz, H-1a), 3.96–3.89 (2 H, m, H-4,
H-1′b), 3.80 (3 H, s, OCH3), 3.79 (3 H, s, OCH3), 3.78 (3 H,
s, OCH3), 3.77 (1 H, dd, H-4′ax), 3.63 (1 H, dd, J1a,1b = 13.3,
J1b,2 = 3.8 Hz, H-1b), 3.58 (1 H, dd, J5a,5b = 9.9, J4,5a = 8.5
Hz, H-5a), 3.49 (1 H, dd, J4,5b = 7.3 Hz, H-5b); 13C NMR
(CD2Cl2): δ 160.30, 160.23, 159.97, 137.20 and 130.27-
J
1′b,2′ = 3.8, J1′b,1′a = 9.2 Hz, H-1′b), 3.60 (1 H, dd,
J1a,1b = 13.5, J1b,2 = 3.8 Hz, H-1b), 3.51 (1 H, dd,
5a,5b = 13.6, J4,5a = 9.7 Hz, H-5a), 3.49 (1 H, dd, J4,5b = 9.7
J
Hz, H-5b); 13C NMR (CDCl3): δ = 137.97, 136.77, 136.71,
136.05 and 135.77 (5 × Cipso Ph), 128.81–127.66 (25 C, Ph),
83.14 (C-3), 81.65 (C-2), 74.59 (C-3′), 73.81, 73.53, 3.39,
72.12, 71.84 (5 × CH2Ph), 73.10 (C-2′), 68.79 (C-4′), 66.62
(C-5), 65.53 (C-4), 50.89 (C-1′), 48.07 (C-1). MALDI–TOF
MS: m/z 785.41 (M+ + H), 808.32 (M+ + Na). Anal. Calcd for
C44H48O9S2: C, 67.32; H, 6.16. Found: C, 67.36; H, 6.10.
(13) 2,3,5-Tri-O-benzyl-1,4-dideoxy-1,4-{[(2S,3S)-2,4-
benzylidenedioxy-3-(sulfooxy)butyl]-episulfonium-
ylidene}-D-arabinitol Inner Salt (4).
126.61 (21 × C, Ph), 114.45, 114.36 and 114.18 (3 × Cipso
,
OMBn), 101.96 (PhCH), 83.29 (C-3), 82.37 (C-2), 76.76 (C-
2′), 73.36, 72.43, and 72.14 (3 × CH2Ph), 69.50 (C-4′), 66.71
(C-5), 66.55 (C-4), 66.45 (C-3′), 55.61 (3 C, 3 × OCH3),
49.55 (C-1′), 48.48 (C-1). Anal. Calcd for C40H46O12S2: C,
61.36; H, 5.92. Found: C, 61.13; H, 6.00.
(16) 1,4-Dideoxy-1,4-{[(2S,3S)-2,4-dihydroxy-3-
(sulfooxy)butyl]-episulfoniumylidene}-D-arabinitol
Inner Salt(1). Compound 13 (2.30 g, 2.94 mmol) was
dissolved in trifluoroacetic acid (24 mL) and while stirring,
water (2.4 mL) was added. The mixture was stirred at room
temperature for 0.5 h. The solvent was removed under
reduced pressure and the gummy residue was washed with
CH2Cl2 (3 × 20 mL). Water (15 mL) was added to dissolve
the crude product, and then evaporated under reduced
pressure to remove the traces of remaining acid. Salacinol
(1) (0.67 g, 68%) was crystallized from MeOH. The mother
liquor was concentrated and purified by column
A mixture of the thioether 25 (260 mg, 0.62 mmol)and 2,4-
di-O-benzylidene-1,3-cyclic sulfate(3)5 (200 mg, 0.74
mmol) in either acetone or HFIP (0.5 mL) containing K2CO3
(13 mg, 0.09 mmol) was treated as described above to yield
the title compound 45 as an amorphous solid (252 mg, 59%
in acetone) and (406 mg, 94% in HFIP).
(14) 1,4-Anhydro-2,3,5-tri-O-( p-methoxybenzyl)-4-thio-D-
arabinitol (12).
To an ice-cold mixture of 1,4-anhydro-4-thio-D-arabinitol 75
(0.98 g, 6.52 mmol)and 60% NaH (1.56 g, 39.15 mmol, 6
equiv) in THF (15 mL), a solution of p-methoxybenzyl
chloride (4.59 g, 29.34 mmol, 4.5 equiv) in THF (10 mL)
was added over 30 min. The reaction mixture was allowed to
chromatography (EtOAc–MeOH–H2O, 7:3:1) to give more
salacinol 1 as a white solid (0.18 g, 18%).
Synlett 2003, No. 9, 1259–1262 ISSN 0936-5214 © Thieme Stuttgart · New York