Mendeleev Communications Electronic Version, Issue 3, 1999 (pp. 87–128) Effect of the nature of protecting group at O-4 on stereoselectivity of glycosylation by 4-O-substituted 2,3-di-O-benzylfucosyl bromides Alexey G. Gerbst,a Nadezhda E. Ustuzhanina,a Alexey A. Grachev,a Dmitry E. Tsvetkov,b Elena A. Khatuntsevab and Nikolay E. Nifant’ev*b a Higher Chemical College, Russian Academy of Sciences, 125047 Moscow, Russian Federation b N.D. Zelinsky Institute of Organic Chemistry, Russian Academy of Sciences, 117913 Moscow, Russian Federation. Fax: +7 095 135 8784; e-mail: nen@ioc.ac.ru The effect of the nature of the substituent at O-4 on the stereoselectivity of glycosylation by 2,3-di-O-benzylfucosyl bromides was studied by direct chemical experiments and computer modelling. Within the synthesis of fucoidan fragments1 we performed glycosylation of acetonide 1 by 2,3-di-O-benzylated L-fucosyl bromides 2 and 3 with benzyl and benzoyl protecting groups at O-4 (Scheme 1).In case of 4-O-benzoylated bromide 3 glycosylation was more stereoselective than in case of 2 (Table 1). Similar data on the stereoselectivity of fucosylation were reported before,2,3 but the origin of the dependence of the stereoselectivity of fucosylation on the structure of fucosyl donor remains unclear.To explain the predominance of the a-product in case of glycosylation by 3 we supposed the formation of intermediate cation II (Scheme 2), in which the carbonyl group of benzoate provides intramolecular ‘stabilisation’ of the cationic centre. Cation II is hindered from the b-side for a nucleophilic attack leading to the formation of the a-glycoside product.To evaluate the ability of the substituent at O-4 in fucosyl bromide 3 to ‘stabilise’ the cationic centre at C-1, the difference (DE) between the total energy of ‘non-stabilised’ glycosyl cation I and ‘stabilised’ glycosyl cation II was calculated using the MM+ force field.4 Partial charges were calculated on the AM1 level5 of approximation.Both molecular-mechanics and semiempirical calculations were performed using the HyperChem software† (version 5.02). The starting conformations of cations I and II were built using standard geometric parameters and setting the torsion angle H(4)–C(4)–O(4)–C equal to 0° for † HyperChem™, Hypercube, Inc., 1115 NW 4th Street, Gainesville, Florida 32601, USA.cation I and to 180° for cation II. The total geometry optimisation was performed using the Polak–Ribiere conjugate gradient algorithm until the gradient value reached 0.1 kcal mol–1 Å–1. The DE value for 4-O-benzylated compound 2 (Table 1) was close to zero, but it was positive for 4-benzoate 3, thus confirming the stabilisation hypothesis and explaining the difference in the stereoselectivities of fucosylation by bromides 2 and 3.Note that all glycosylation reactions were performed in CH2Cl2,‡ which solvates all mentioned cations in a similar manner. This permitted us to neglect solvation effects in the calculations. To elucidate in more details the stabilising effect of protecting group at O-4, which favours the a-selectivity of fucosylation, we also calculated DE for cations with p-nitrobenzoyl, p-methoxybenzoyl and N-acetylaminoacetyl groups at O-4.The DE values for p-nitrobenzoate 4 and 4-O-(N-acetylaminoacetyl) derivative 6 (Table 1) were lower than that for benzoylated compound 3. On the contrary, the DE value for p-methoxybenzoate 5 was higher than that for benzoate 3. According to these calculation data, we expected that the a-selectivity of fucosylation should decrease in the order 5 > 3 > 4 > 6 > 2.These results were later R Bn Bz p-NO2C6H4CO p-MeOC6H4CO AcNHCH2CO 23456 789 10 11 a-isomer 12 13 14 15 16 b-isomer bromide Scheme 1 O OAll OH Me O O Me Me O OBn Me RO OBn Br O OAll O Me O O Me Me O OBn Me RO OBn O OAll O Me O O Me Me O MeOR OBn BnO 1 2–6 a-isomer 7–11 b-isomer 12–16 All = CH2CH=CH2 O Me O O R OBn OBn O R H a b 3–6 a + b products I O Me O OBn BnO O R H a b II R O a product Scheme 2 aExperimental ratios of a:b isomers were determined by integration of respective 1H signals of Fuc residues at the ‘non-reducing’ end. bDE value for intermediate III.Table 1 The ratios between a- and b-disaccharide products in the glycosylation of acceptor 1 with bromides 2–6 (Scheme 2) and the DE values for corresponding cations of type II.Fucosyl bromide Substituent at O-4 DE/kcal mol–1 a- and b- disaccharide products Ratio between a- and b- products 2 Bn –0.1 7 and 12 1:1a 3 Bz 3.6 8 and 13 3.5:1a 4 p-NO2C6H4CO 2.1 9 and 14 2:1 5 p-MeOC6H4CO 4.7 10 and 15 5:1 6 AcNHCH2CO 1.6 (19.3b) 11 and 16 2:1Mendeleev Communications Electronic Version, Issue 3, 1999 (pp. 87–128) ‡ Preparation of bromides 4–6.Esters 18 {[a]D –134° (c 2, CHCl3)} and 19 {[a]D –162° (c 1, CHCl3)} were prepared in 80–90% yields by acylation of compound 7 (1 mmol) with a corresponding acylchloride (4 mmol) in 40 mmol of pyridine in the presence of a catalytic amount of N,N-dimethylaminopyridine.Amide 20 {[a]D –62° (c 1, CHCl3)} was obtained in 75% yield by reaction of 7 with equimolar amounts of sym anhydride of N-acetylglycine and N,N-dimethylaminopyridine. Deallylation of esters 18–20 in the presence of PdCl2 (0.4 mmol) in methanol gave semiacetals 21–23 with 75–80% yields. The bromination with CBr4 and Ph3P (1.1 mmol each) in 5 ml of boiling methylene chloride resulted in formation of bromides 4–6 in almost quantitative yields.Fucosyl bromides were used directly in glycosylation reactions without special purification. Glycosylation with bromides 2–6 (typical procedure). A solution of 1 mmol of acetonide 1, 1.5 mmol of Hg(CN)2, 10–20 mg of HgBr2 and 1.4 g of molecular sieves 4 Å were stirred for 1 h at room temperature under Ar, and a solution of 1.5 mmol of a corresponding fucosyl bromide was added portionwise within 1 h at room temperature. The mixture was additionally stirred for 24 h at room temperature, then filtered through Celite, diluted with CH2Cl2, washed with saturated aqueous KBr and NaHCO3 solutions, filtered through cotton wool and concentrated in vacuo. The residue was subjected to flash column chromatography to separate a mixed fraction of a- and b-disaccharide products.The ratio between the products was determined from the 1H NMR spectra (Tables 1 and 2). The anomeric configurations of Fuc residues in disaccharides 7–16 at the ‘non-reduced’ end were confirmed by characteristic values of J1,2, which were 3–3.5 Hz for a-anomers and 7–7.5 Hz for b-anomers. proven experimentally by chemical glycosylations (Table 1), except for the coincidence of stereochemical outcomes in glycosylations with compounds 4 and 6.For compound 6 which comprises a carbonyl of the amido group along with the ester carbonyl, the hypothetical intermediate III can also be expected in addition to cation II. The DE value for III is higher (Table 1) than that for ester-stabilised cations of the type II.Taking into account too high DE value for intermediate III, we can expect high a-stereoselectivity of fucosylation with bromide 6. However, the ratio between a- and b-disaccharides in the glycosylation with bromide 6 was as low as 2:1. This result argues that the glycosylation proceeds preferentially via cation II rather than cation III. In conclusion, the data obtained show the mechanism of the influence of the substitutent at O-4 on the a-stereoselectivity of glycosylation by 2,3-di-O-benzylfucosyl donors.Molecularmechanics calculations according to the described procedure can be successfully applied to the estimation of the stereoselectivity of glycosylation. This work was supported by the President of the Russian Federation (grant no. 96-15-96991) and the Russian Foundation for Basic Research (grant nos. 97-03-33037a and 98-03-33025a). References 1 E. A. Khatuntseva, N. E. Ustuzhanina, G. V. Zatonskii, A. S. Shashkov, A. I. Usov and N. E. Nifant’ev, J. Carbohydr. Chem., in press. 2 M. Dejter-Juszynski and H. M. Flowers, Carbohydr. Res., 1973, 28, 61. 3 S. J. Danishefsky, J. Gervay, J. M. Peterson, F. E. McDonald, K.Koseki, D. A. Griffith, T. Oriyama and S. P. Marsden, J. Am. Chem. Soc., 1995, 117, 1940. 4 N. L. Allinger, J. Am. Chem. Soc., 1977, 99, 8127. R p-NO2C6H4CO p-MeOC6H4CO AcNHCH2CO 18 19 20 21 22 23 Scheme 3 O OAll OBn Me HO OBn O OBn Me RO OBn 17 OAll O OBn Me RO OBn OH 4–6 All = CH2CH=CH2 aNMR spectra were recorded on a Bruker AM-300 instrument (300 MHz) in CDCl3 at 303 K. Assignment was performed by 2D 1H–1H correlation spectroscopy.bR is the ‘reducing’ end (i.e., a fucose residue attached to allyl aglycon), N is the ‘non-reducing’ end (i.e., a fucose residue attached to fucose aglycon). cn/d = not determined. dSignals of 6-H in all Fuc residues in the specta of mixtures of disaccharide pairs (9,14), (10,15) and (11,16) were not assigned because of overlapping.eFor all compounds, J4,5 < 1 Hz. Table 2 1H NMR dataa for monosaccharides 17–23 and disaccharides 9–11 and 14–16. Compound Fucose residueb Chemical shifts, d/ppm Coupling constants,e J/Hz 1-H 2-H 3-H 4-H 5-H 6-H J1,2 J2.3 J3,4 J5,6 9 R 4.99 3.87 4.39 4.09 n/dc 1.05–1.50d 3.5 7.9 5.7 n/d N 5.02 3.86 4.28 5.68 4.53 1.05–1.50d 3.5 9.9 3.1 n/d 10 R 4.98 3.86 4.38 4.09 n/d 1.05–1.50d 3.1 7.9 5.5 n/d N 5.01 3.91 4.25 5.66 4.48 1.05–1.50d 3.5 10.0 3.2 n/d 11 R 4.94 3.82 4.34 4.03 n/d 1.05–1.50d 2.9 7.7 5.2 n/d N 5.01 3.71 4.12 5.31 4.39 1.05–1.50d 3.2 10.0 3.0 n/d 14 R 5.09 3.91 4.45 4.09 n/d 1.05–1.50d 3.2 8.0 5.1 n/d N 4.81 3.69–3.74 4.54 3.76 1.05–1.50d 6.5 n/d 2.9 n/d 15 R 5.11 3.90 4.45 4.09 n/d 1.05–1.50d 3.9 7.5 5.1 n/d N 4.75 3.65 3.75 5.58 3.72 1.05–1.50d 7.1 7.1 3.2 n/d 16 R 5.08 3.88 4.40 4.03 n/d 1.05–1.50d 3.1 8.5 4.9 n/d N 4.74 3.60–3.63 5.46 3.62 1.05–1.50d 6.9 n/d 3.5 n/d 17 4.85 3.81 4.05 3.79 n/d 1.20 3.9 9.5 3.1 6.0 18 4.95 3.89 4.13 5.65 4.25 1.19 3.5 9.5 3.5 7.5 19 4.95 3.93 4.12 5.62 4.21 1.20 3.9 9.1 3.0 6.0 20 n/d 3.83 4.07 5.52 4.16 1.20 4.0 10.0 3.9 6.8 21 a 5.35 3.87 4.10 5.65 4.42 1.22 3.2 10.0 3.4 6.9 b n/d 3.65 3.71 5.59 3.82 1.31 7.4 9.8 3.1 6.3 22 a 5.28 3.89 4.01 5.58 4.31 1.18 3.6 10.0 2.9 6.5 b 5.24 4.09 3.63 5.51 3.73 1.21 n/d 7.8 2.8 6.0 23 a 5.23 4.06 3.76 5.48 4.33 1.15 3.5 9.0 4.0 6.7 b n/d 3.48 3.73 5.43 3.84 1.25 8.0 9.5 3.1 6.7 O Me O OBn BnO III O N Me O HMendeleev Communications Electronic Version, Issue 3, 1999 (pp. 87–128) 5 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart, J. Chem. Soc., 1985, 107, 3902. Received: 18th November 1998; Com. 98/1400 (8/09455A)