Synthesis of salacinol

Article abstract of DOI:10.1016/S0040-4039(00)01129-1

Salacinol, a new type of α-glucosidase inhibitor discovered from the antidiabetic herb, was synthesized for the first time. Under the strategy that salacinol would be synthesized by the coupling reaction between 1,4-epithio-D-arabinitol and the cyclic sul

Full text of DOI:10.1016/S0040-4039(00)01129-1

Tetrahedron Letters 41 (2000) 6615±6618
Synthesis of salacinol
Hideya Yuasa,* Jun Takada and Hironobu Hashimoto
Department of Life Science, Graduate School of Bioscience and Biotechnology, Tokyo Institute of Technology,
4259 Nagatsuta, Midori-ku, Yokohama 226-8501, Japan
Received 24 May 2000; revised 29 June 2000; accepted 30 June 2000
Abstract
Salacinol, a new type of a-glucosidase inhibitor discovered from the antidiabetic herb, was synthesized
for the ®rst time. Under the strategy that salacinol would be synthesized by the coupling reaction between
1,4-epithio-d-arabinitol and the cyclic sulfate of an erythritol derivative, the model coupling reactions
between tetrahydrothiophene and versatile cyclic sulfate derivatives were undertaken. These experiments
indicated that the 1,3-diol of the cyclic sulfate should be protected with the isopropylidene group, otherwise,
even the benzylidene-protected cyclic sulfate decomposed during the reaction. Thus, the salacinol was
synthesized using the cyclic sulfate of 1,3-O-isopypropylidene-d-erythritol. The resulting coupling product
was deisopropylidenated to a€ord salacinol. A diastereomer of salacinol was also synthesized. # 2000
Elsevier Science Ltd. All rights reserved.
Keywords: salacinol; glycosidase inhibitors; sulfonium compounds; cyclic sulfates.
Thiacyclopentane derivative with the sulfur atom being in trivalent state is a new class of
glucosidase inhibitor. This class of compound was ®rst created as the sul®mide derivative 1,
which has shown a weak inhibitory e€ect toward b-glucosidase.1 The molecular model of the
compound 1 superimposed on an assumed transition state (TS) of the glucoside hydrolysis
showed that a cationic character of the trivalent sulfur atom resembles that of the anomeric
carbon of the structure TS. More recently, the sulfonium ion derivative (salacinol; 2) was dis-
covered from Salacia reticulata, the herb used in Indian traditional medicine for diabetes, and
was shown to be a strong inhibitor of ꢀ-glucosidases.2 The structure of salacinol is unique in that
the ring sulfonium ion is stabilized by a sulfate counter anion tethered with an erythritol chain,
constructing spirobicyclic-like structure. This characteristic structure may account for the strong
inhibitory e€ect of salacinol. In continuous studies on inhibitory activities of the ring sulfur
analogs of aldose toward glycosidases,3 we were interested in the e€ect of the tethered sulfonium
salt. We thus started studies on the construction of the tethered sulfonium salt on thiacyclopetane.
* Corresponding author. Tel: +81 45 924 5705; fax: +81 45 924 5805; e-mail: hyuasa@bio.titech.ac.jp
0040-4039/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)01129-1

6616
We planned the synthesis as shown in Fig. 1. A key step of the synthetic strategy is the ring
opening of the cyclic sulfate 5 by the nucleophilic attack of 1,4-epithio-d-arabinitol (4). While the
reaction of cyclic sulfates is well studied,4 we could not ®nd any examples in which a tethered
sulfonium salt is produced. Moreover, we were so uninformed about chemical properties of
salacinol that we were not con®dent in the conditions for the deprotection of compound 3.
Figure 1. A strategy for the synthesis of salacinol 2
Scheme 1. (a) NaIO4 (2 equiv.), NaHCO3, H2O, bromocresol green, 5 h; (b) NaBH4, CH3OH, 10 min; (c) SOCl2 (1.2
equiv.), Et3N (2.5 equiv.), CH2Cl2, 0ꢀC 10 min; (d) RuCl3 (cat.), NaIO4 (2 equiv.), CH3CN, CCl4, H2O, 30 min;
(e) BnBr (4 equiv.), NaH (4.5 equiv.), DMF, 1 h; (f) 80% AcOH, 60ꢀC, 3 h; (g) NaBrO3 (6 equiv.), Na2S2O4 (6 equiv.),
EtOAc, H2O, 12 h

6617
The compound 4 was prepared by the method of Yuasa and co-workers.1 The tethering arm of
salacinol is a derivative of erythritol and the cyclic sulfate precursor could be derived either from
d- or l-glucose, depending on the protecting group for the hydroxyl groups as shown in Scheme
1. Periodate oxidation of 4,6-O-isopropylidene-d- or l-glucose (6)5 followed by reduction of the
resulting aldehyde gave the 2,4- or 1,3-O-isopropylidene-d-erythritol (7 or 9). Formation of
the cyclic sulfate at the remaining diol function was carried out by the reported method to give
the compound 8 or 10. Benzylation of hydroxyl groups of the compound 7 followed by
deisopropylidenation gave the precursor (12) of the cyclic sulfate (13). Oxidative debenzylation6
of the compound 13 a€orded the cyclic sulfate (14) with free hydroxyl groups. Benzylidenated
cyclic sulfate (16) was deduced from 2,4-O-benzylidene-d-erythritol (15).7
Model experiments simulating the coupling reactions between the compound 4 and the cyclic
sulfates were undertaken by using tetrahydrothiophene (THT) as a nucleophile. While the cyclic
sulfate (8) protected with the isopropylidene group reacted with THT at 45ꢀC in dimethylformamide
(DMF) giving the coupled product (17),8 that protected with benzyl group (13) or with the
benzylidene group (16) or unprotected cyclic sulfate (14) did not react under the same conditions
used for the formation of 17 (Scheme 2). These cyclic sulfates only decomposed when the bath
temperature was increased to 60±70ꢀC. Use of methanol, water, or acetonitrile as solvent gave no
products, while dimethyl sulfoxide can substitute for DMF. Deisopropylidenation of the com-
pound 17 was carried out by 0.01% HCl to give the compound 18, which is the assumed product
of the coupling reaction between THT and the unprotected cyclic sulfate (14). The compound 18
was stable under the conditions for the coupling reaction (45ꢀC in DMF for 9 h), indicating that
the reason for the unsuccessful coupling reaction with the unprotected cyclic sulfate 14 was not
due to instability of the product that might exist temporarily. Protons of the hydroxyl groups
might interfere with the nucleophilic attack of THT. In the same manner, the nucleophilic attack
can be blocked by a benzyl or benzylidene group.
ꢀ
Scheme 2. (a) DMF, 45 C, 13 h; (b) 0.01% HCl, 40ꢀC, 4 h
The coupling reaction between the compound 4 and 1,3- or 2,4-O-isopropylidene-d-erythritol
(8 or 10) was successful and a€orded the coupling product (19 or 20). The coupling products were
then deisopropylidenated to give salacinol 29 and its diastereomer (21),10 respectively. The NMR

6618
spectra and optical rotation of salacinol 2 were in good accordance with those reported.2
Yoshikawa and co-workers reported the absolute con®guration of salacinol di€erently in the
®rst2a and the latter2b,c papers. Our result supported the structure depicted in the latter papers.
Stereochemistry at the sulfonium center of the diastereomer 21 was determined to be S, the same
con®guration as that of salacinol, from the existence of strong NOEs between one of two H-1s
and H-40. Syntheses of the derivatives of salacinol, the new candidates as glycosidase inhibitors,
using the strategy developed in this study are underway.
Acknowledgements
This work was supported by a Grant-in-Aid for Encouragement of Young Scientists No.
11780416 from the Japanese Ministry of Education, Science, Sports and Culture.
References
1. Yuasa, H.; Kajimoto, T.; Wong, C.-H. Tetrahedron Lett. 1994, 35, 8243±8246.
2. (a) Yoshikawa, M.; Murakami, T.; Shimada, H.; Matsuda, H.; Yamahara, J.; Tanabe, G.; Muraoka, O.
Tetrahedron Lett. 1997, 38, 8367±8370. (b) Yoshikawa, M.; Murakami, T.; Yashiro, K.; Matsuda, H. Chem.
Pharm. Bull. 1998, 46, 1339±1340. (c) Matsuda, H.; Murakami, T.; Yashiro, K.; Yamahara, J.; Yoshikawa, M.
Chem. Pharm. Bull. 1999, 47, 1725±1729.
3. Yuasa, H.; Hashimoto, H. Reviews on Heteroatom Chemistry 1999, 19, 35±65.
4. Calvo-Flores, F. G.; Garcõa-Mendoza, P.; Hernandez-Mateo, F.; Isac-Garcõa, J.; Santoyo-Gonzalez, F. J. Org.
Chem. 1997, 62, 3944±3961.
5. Wolfrom, M. L.; Diwadkar, A. B.; Gelas, J.; Horton, D. Carbohydr. Res. 1974, 35, 87±96.
6. Adinol®, M.; Barone, G.; Guariniello, L.; Iadonisi, A. Tetrahedron Lett. 1999, 40, 8439±8441.
7. MacDonald, D. L.; Fischer, H. O. L.; Ballou, C. E. J. Am. Chem. Soc. 1956, 78, 3720±3722.
23
8. Compound 17: ‰ꢀŠ ^15.4 (c 1.05, methanol); 1H NMR (400 MHz, CDCl3) ꢁ 4.45 (ddd, J=5.7, 9.6, 9.8 Hz, 1H, H-
D
3), 4.30 (ddd, J=3.0, 3.8, 9.8 Hz, 1H, H-2), 4.13 (dd, J=5.7, 11.5 Hz, 1H, H-4), 3.94 (dd, J=3.0, 13.7 Hz, 1H, H-
1a), 3.85 (m, 1H, -CHH-S), 3.80 (dd, J=9.6, 11.5 Hz, 1H, H-4), 3.68 (m, 1H, -CHH-S), 3.56 (dd, J=3.8, 13.7 Hz,
1H, H-1b), 3.56 (m, 1H, -CHH-S), 3.45 (m, 1H, -CHH-S), 2.36 (m, 4H, -CH2-C-S), 1.49, 1.41 (each s, each 3H,
CH3); 13C NMR (67.8 MHz, CDCl3) ꢁ 99.9, 77.2, 69.6, 67.6, 62.3, 46.7, 45.2, 44.8, 28.6, 28.4, 19.1; HRMS (ESI)
calcd for [C11H20O6S2+H]+: 313.0780; found: 313.0784.
23
D
28
D
9. Compound 2: ‰ꢀŠ +5.1 (c 1.02, methanol), lit.2 ‰ꢀŠ +4.9 (c 0.35, methanol); HRMS (ESI) calcd for
[C9H18O9S2+H]+: 335.0470; found: 335.0466.
23
D
1
10. Compound 21: ‰ꢀŠ ^25.2 (c 1.17, methanol); H NMR (400 MHz, pyridine-d5 with a drop of D2O) ꢁ 5.19 (ddd,
J=3.4, 3.7, 8.2 Hz, 1H, H-30), 5.15 (m, 1H, H-2), 5.14 (m, 1H, H-3), 4.97 (ddd, J=5.1, 4.0, 8.2 Hz, 1H, H-20), 4.81
(dd, J=4.0, 13.0 Hz, 1H, H-10), 4.76 (m, 1H, H-4), 4.67 (dd, J=3.4, 11.9 Hz, 1H, H-40), 4.59 (dd, J=6.7, 6.1 Hz,
2H, H-5), 4.51 (dd, J=5.1, 13.0 Hz, 1H, H-10), 4.41 (dd, J=3.7, 11.9 Hz, 1H, H-40), 4.28 (dd, J=3.7, 12.7 Hz, 1H,
H-1), 4.16 (dd, J=2.7, 12.7 Hz, 1H, H-1); 13C NMR (67.8 MHz, pyridine-d5) ꢁ 79.3, 79.2, 79.0, 71.8, 67.5, 62.1,
60.2, 52.8, 51.0; HRMS (ESI) calcd for [C9H18O9S2+H]+: 335.0470; found: 335.0477.