sulfonates, excellent yields and high «-selectivity were obtained upon coupling to a range of alcohols in various solvents. It was noted that the «-selectivity was minimized with the use of the triflate leaving group, albeit for reactions conducted at — 78°C, in contrast to the other members of the series when room temperature was used. In the case of the reaction of the 4,6-di-O-(N-phenylcarbamoyl)-2,3-di-O-ben-zylgalactopyranosyl triflate with methanol, high ^-selectivity was exceptionally obtained. These results were again all interpreted in terms of more or less tight ion pairs [4,5].

Finally, Srivastava and Schuerch studied the formation of ^-mannopyranosides and the related ^-rhamnopyranosides, using a range of glycosyl sulfonates as donors [6,7]. In this most difficult series, the authors emphasized the increased anomeric effect resulting from the antiparallel dipoles of the C1—O and C2—O bonds and sought to maximize this contribution by installing a strongly electron-withdrawing, nonparticipating protecting group on O2. Thus, 3,4,6-tri-O-benzyl-2-O-mesyl-«-d-mannopyranosyl chloride (4) was prepared and reacted with a range of silver sulfonates to give the corresponding a-mannosyl sulfonates. Subsequent exposure to methanol or cyclohexanol, usually in acetonitrile, then gave the mannopyranosides in high yield and excellent ^-selectivity. Only one carbohydrate-based alcohol (6) was used as glycosyl acceptor, but this too gave excellent ^-selectivity especially with the 2,2,2-trifluoroethanesulfonate 5 as donor (Scheme 2).

In considering the mechanism of their mannosylation reaction, Srivastava and Schuerch suggested that the electron-withdrawing sulfonate ester at the 2-position served to render the reactive ion pair in the SN1 mechanism tighter on the a face than was the case in the glucose and galactose series. In this manner the a face was considered to be highly shielded toward approach of the nucleophile, with the displacement taking on the high stereoselectivity normally associated with SN2 processes [6,7]. Similar ^-selectivities were obtained from a 3,4-di-O-benzyl-2-O-mesyl-a-l-rhamnopyranosyl tosylate [6,7]. Although the authors obtained excellent

^-selectivity and yields in their mannosylations and rhamnosylations, they repeatedly alluded to the high moisture sensitivity of the anomeric sulfonates, as in the earlier glucose and galactose series, and the need to work on vacuum lines. Presumably this, and the need to remove the somewhat unconventional 2-O-mesylate protecting group subsequent to coupling, prevented them from developing the full potential of the method and applying it in the synthesis of oligosaccharides.

More or less contemporaneously with the work of Schuerch, Leroux and Perlin studied the sulfonylation of pyranoses. Thus, 2,3,4,6-tetra-O-benzyl-a-d-glucopyra-nose was treated in cold dichloromethane with triflic anhydride in the presence of 2,4,6-collidine, leading to a purported anomeric triflate. However, the authors noted that subsequent addition of methanol or ethanol did not lead to the formation of the glycoside in acceptable yields [8,9]. Subsequently, it was discovered that conducting the triflation in the presence of tetrabutylammonium bromide resulted in the formation of the a-glucopyranosyl bromide and that addition of an alcohol then led cleanly to the glycoside. In these reactions Perlin and Leroux typically carried out the triflation/bromide displacement at — 70°C, noting that bromide formation was complete after 15-30 min; the reaction mixture was warmed to room temperature before addition of the alcohol. However, it was also noted that bromide formation could be conducted at room temperature without apparent detriment. These workers also investigated the use of 2,3,4,6-tetra-O-acetyl-fi-d-glucopyranose (8), with its potentially participating protecting groups, as substrate. In the presence of collidine but the absence of bromide ion, an orthoester (10) was isolated in 44% yield. When bromide ion was included in the reaction mixture, acetobromoglucose (11) was isolated in 50% yield and there was no indication of orthoester formation. It was suggested that both reactions proceeded by way of the fi-triflate (9), which was trapped either by further pyranose or by bromide ion according to the conditions employed (Scheme 3) [8,9].

The reaction of methanesulfonic anhydride with 2,3,4,6-tetra-O-benzyl-a-d-glucopyranose in dichloromethane in the presence of collidine, followed by addition of methanol, resulted in the formation of a 3: 2 a/fi mixture of the methyl glycosides, isolated in 87% yield. Since glycosylation was achieved without the need for addition of the quaternary ammonium bromide, unlike the case of triflic anhydride, it was concluded that the glucosyl mesylate was considerably more stable and allowed for the displacement reaction to take place. Moreover, the anomeric ratio suggested that the a and fi anomers of this mesylate were in equilibrium, with displacement of the fi anomer occurring more rapidly. When 2,3,4,6-tetra-O-acetyl-fi-d-glucopyranose was treated with methanesulfonic anhydride, a crystalline product was obtained in 74% yield and assigned as the a-glucosyl mesylate. Indeed the characteristics of this substance were comparable to those of the compound obtained by Schuerch upon treatment of acetobromoglucose with silver methanesulfonate [8,9].

When methanesulfonyl chloride was allowed to react with 2,3,4,6-tetra-O-ben-zylglucopyranose and collidine in dichloromethane, the a-glucopyranosyl chloride was isolated regardless of whether the quaternary ammonium bromide was included. Addition of methanol to the reaction mixture resulted in the formation of an anomeric mixture of methyl glycosides. Similar results were obtained with toluenesulfonyl chloride, although it was noted that the initial sulfonylation was somewhat slower [8,9]. The use of tosyl chloride in the dehydrative coupling of alcohols with pyranoses was later revisited by Szeja and his coworkers, with the difference that aqueous

Scheme 3

phase transfer conditions were used, and glycosyl toluenesulfonates were implied as intermediates [10,11]. Koto and coworkers investigated the coupling of tetra-O-ben-zyl-a-d-glucopyranose and a range of acceptor alcohols with the aid of a mixture of 4-nitrobenzenesulfonyl chloride and silver triflate [12-15]. In the presence of tri-ethylamine, the a-glycoside predominated, whereas the inclusion of jV,V-dimethyl-acetamide resulted in the isolation of the /3 anomer. It was suggested that the anomeric hydroxyl group was sulfonylated with the sulfonyl chloride to a glucopyranosyl 4-nitrobenzenesulfonate and subsequently converted to the active glycosyl donor, the glucopyranosyl triflate, by the action of silver triflate. However, given the relative acidities of triflic and 4-nitromethanesulfonic acid and the related work of Szeja, carried out in the absence of silver triflate, the formation of a covalent glycosyl triflate in this work appears to be somewhat unlikely. The inversion of stereoselectivity on inclusion of the jV,jV-dimethylacetamide was explained by invoking the formation of an a-jV,jV-dimethylacetimidate ester [12,13].

Pavia et al. revisited the reactions of tetra-O-benzylglucopyranose (12) with trifluoromethanesulfonic anhydride, but in the absence of base. They discovered that the corresponding trehalose derivatives (13) were formed in good yield, predominantly as the a,a form (Scheme 4) [16]. Comparable results were obtained in the galacto-, manno-, and arabinopyranose series as with fructofuranose [16]. When tri-fluoromethanesulfonic anhydride was added to a mixture of a perbenzyl-protected glycopyranose and an acceptor alcohol, such as various serine, threonine and hy-droxproline derivatives, coupling was achieved in good yield [17].

Pavia and coworkers carried out a careful study of the mechanism of these reactions using 19F-NMR spectroscopy. They concluded that they were observing simple acid-catalyzed dehydrative couplings in which water was removed from the

0 0

Post a comment