Polymer-assisted solution-phase synthesis of glycosyl chlorides and bromides using a supported dialkylformamide as catalyst.

Glycosyl halides are a very valuable and versatile group of synthetic intermediates widely used in carbohydrate chemistry for the ready functionalization at the anomeric carbon. Thus, glycosyl halides can be easily converted into O-glycosides, C-glycosides, or glycals via the generation of intermediate anomeric carbocations, radicals, or cabanions. Because of their moderate stability, glycosyl halides possessing elaborate protecting group patterns are best prepared by introduction of the halide in the last step. A number of methods have been developed for this transformation starting from glycosyl esters, glycosyl hemiacetals, alkyl glycosides, glycosyl orthoesters, glycosyl acetals, glycals, and thio-, seleno-, or telluro-glycosides. However, the relatively harsh reaction conditions employed in many of these procedures often limit the use of acid-sensitive protecting groups in the starting carbohydrate substrate. With a combination of high reactivity with reasonable stability, glycosyl chlorides and bromides were the first, and for a long time practically the only, glycosyl donors used for the synthesis of complex O-glycosides in the presence of heavy metal salts (the Koenigs-Knorr reaction) or halide ions as promoters. Although important developments have been achieved in the preparation and activation of the highly stable glycosyl fluorides, and the highly reactive glycosyl iodides, the corresponding chlorides and bromides are still the most widely used glycosyl halides. Generation of these glycosyl donors under mild neutral conditions is possible using Ph3P/CX4 (X ) Cl, Br), p-TsCl/DMAP, haloenamines, n-BuLi/ClPO(OPh)2, or triphosgene-pyridine. In some cases, the synthesis of these glycosyl donors and the subsequent glycosylation reaction can be accomplished in a sequential one-pot manner, avoiding the isolation of the intermediate halide. However, purification of the glycosyl halide is usually necessary and frequently problematic. The use of solid-supported reagents opens an interesting perspective for the efficient preparation of these delicate compounds by greatly simplifying their isolation. Successful implementations of this kind of strategy have been recently achieved through the use of a polymer-supported triarylphosphine in the presence of CBr4 or I2 and by means of a polymer-supported haloenamine. In the context of a project aimed at the development of new methods for the efficient and high-throughput synthesis of oligosaccharides, we decided to develop an alternative polymer-assisted solution-phase synthesis of glycosyl chlorides and bromides based on the widely used and high yielding anomeric halogenation of carbohydrate hemiacetals with oxalyl chloride or bromide catalyzed by N,N-dimethylformamide. In this well-known procedure, the N,Ndimethylformide is transformed in situ into the corresponding Vilsmeier salt, which is the actual halogenating reagent. The use of a polymer supported N,N-dialkylformamide as catalyst could streamline the isolation of the glycosyl halide product through simple filtration of the crude reaction mixture, followed by evaporation of the solvent and excess oxalyl reagent at reduced pressure. Two known polystyrenesupported dialkylformamides (2 and 5) were examined for this purpose (Scheme 1). Polymer-supported N-formylpiperazine 2 was prepared using a modification of the described procedure, by N-acylation of commercially available piperazine resin 1 with ethyl formate under DMAP catalysis. Supported N,N-dialkylformamide 5 was obtained from crosslinked chloromethylated polystyrene (Merrifield resin, 3) by successive reaction with methylamine followed by Nacylation of the resultant methylamine resin 4 with ethyl formate as above. The progress of these solid-phase reactions was monitored by FT-IR spectroscopy, and the final resin loadings were determined by nitrogen elemental analysis. Initial experiments were performed using 2,3,4,6-tetra-Obenzyl-D-glucose (6a, Chart 1) as a test substrate and dichloromethane as solvent in the presence of the resin (2 or 5; 0.1–1.0 mol equiv) by dropwise addition of oxalyl chloride (3 mol equiv.) under mild magnetic stirring. When the reaction was judged complete by TLC monitoring, the mixture was filtered; the resin was thoroughly rinsed with CH2Cl2, and the solvent and excess oxalyl chloride were removed at reduced pressure. The crude product was analyzed by H NMR, which showed that the corresponding glycosyl chloride was formed in very high yield and >95% purity. In general, the crude product was somewhat cleaner with resin 5, and this resin was finally selected for further experiments. The reaction was reasonably fast (2 h) when a molar equivalent of 5 was used under room temperature conditions (Table 1, entry 1), although substoichiometric amounts of 5 (0.1–0.5 mol equiv; entries 3 and 4) could be employed without compromising the yields at the expense of increased reaction times (4 h). No reaction was observed in the absence of the resin. Compared to the homogeneous reaction using DMF as catalyst, which produces the thermodynamic product exclusively, the R-glycosyl chloride, 36 its supported analog 5 affords an R/ mixture of glycosyl chlorides 7a. As expected, lower temperatures produced increased amounts of the kinetic product, the -glycosyl chloride (entry 2). Reduced amounts of resin have the same effect (entries 3 and 4), which supports the role of the dialkylformamide as a catalyst of the R/ -isomerization process of the glycosyl halide, as recently reported. Several differently protected mono and disaccharide hemiacetals (6b, * To whom correspondence should be addressed. E-mail: jl.chiara@ iqog.csic.es. J. Comb. Chem. 2008, 10, 361–363 361