Mendeleev Communications Electronic Version, Issue 1, 1998 (pp. 1–42) Synthesis of spacer-armed glycosides using azidophenylselenylation of allyl glycosides Andrei A. Sherman, Leonid O. Kononov, Alexander S. Shashkov, Georgij V. Zatonsky and Nikolay E. Nifant’ev* 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 Protected 3-aminopropyl spacer-armed glycosides that can be further used for the preparation of neoglycoconjugates have been prepared from allyl glycosides using azidophenylselenylation of the double bond as a key step.Neoglycoconjugates are synthetic compounds that emulate the behaviour of the natural glycoconjugates and are useful tools in glycobiology research.1,2 A prerequisite for the preparation of neoglycoconjugates is the accessibility of a spacer-armed glycoside, i.e.a glycoside with a functional group in the aglycon that can be used for coupling to a carrier. An amino function at the terminal position of an aglycon alkyl chain has been widely used for this purpose.2 For example, 3-aminopropyl glycosides3 have already been used for the preparation of various neoglycoconjugates.However, there still exists a need for the development of new approaches to the preparation of such spacer-armed glycosides from simple glycosides (so called pre-spacer glycosides) under mild conditions. Such an approach would be of special importance for long-chain oligosaccharides. Retrosynthetic analysis shows that 3-aminopropyl glycosides may be obtained by addition of a synthetic equivalent of the amino group to the double bond of allyl glycosides.We anticipated that azidophenylselenylation4 of allyl glycosides followed by subsequent reduction of the azido function and removal of the phenylselenyl moiety (Scheme 1, path A) could constitute a new approach to the preparation of 3-aminopropyl spacer-armed glycosides from allyl glycosides (for other methods of functionalization of allyl glycosides see refs. 5–7). Peracetylated allyl lactoside 1a8,9 was chosen as a model substrate for azidophenylselenylation which was performed under the conditions developed4 for aliphatic alkenes. Treatment of 1a (0.05 mmol) with NaN3 (2.4 equiv.), (PhSe)2 (0.6 equiv.) and PhI(OAc)2 (1.4 equiv.) in CH2Cl2 (0.5 ml) at 20 °C (18 h) PhI(OAc)2 + 2N3 + N3 Path A Path B [PhI(N3)2] [PhI(OAc)(N3)] – PhI – N3 2N3 [PhI(OAc)] RO – PhI RO N3 RO OAc 2 7 RO N3 SePh 3a,b RO N3 10b + CH2N3 RO 11b + (PhSe)2 RO NHX SePh RO OAc SePh 4b X = H 5b X = Boc 8a,b RO NHBoc RO OAc 6b 9b O AcO AcO OAc OAc O O OAc AcO OAc O AcO AcO OAc OAc a R = b R = 1a,b Scheme 1 – (a) (b) H-1 H-1' H-3,3' H-1 –584.0 –582.0 –580.0 –578.0 –576.0 –574.0 4.4 4.0 3.6 3.2 ppm H-1' H-3,3' ppm ppm H-3 H-3' H-1 H-1' H-3,3' 4.8 4.4 4.0 3.6 –590.0 –588.0 –586.0 –584.0 –582.0 –580.0 H-1 H-1' ppm Figure 1 2D 1H–77Se NMR spectra of compounds 3b (a) and 8b (b) (Bruker AM-300, 303 K, C6D6). Numeration of atoms in the aglycon: sugar-1-2-3.Mendeleev Communications Electronic Version, Issue 1, 1998 (pp. 1–42) unexpectedly afforded two adducts 3a (31%, 1.2:1 ratio of diastereoisomers) and 8a (23%, 1.3:1 ratio of diastereoisomers) rather than a single product (cf.ref. 4). Unidentified products possessing neither allyl nor aromatic fragments (1H NMR data) were also isolated in ca. 20% yield. The presence of PhSe groups in both adducts 3a and 8a, seven AcO groups in 3a and eight AcO groups in 8a was evident from 1H, 13C and 77Se NMR data (Table 1).Compound 3a has an absorption band at 2108 cm–1 in the IR spectrum which is characteristic of an azido group whereas in 8a this band was absent. The spectral data obtained allowed us to surmise that 3a and 8a were the products of azidophenylselenylation and acetoxyphenylselenylation of 1a, respectively. The overlap of signals of five aglycon protons and four H-6 protons of glucose and galactose residues in the 1H NMR spectra of 3a and 8a complicated the determination of the exact substitution pattern of the aglycon. In order to simplify interpretation of the spectra, azidophenylselenylation of allyl galactoside 1b6 was performed under the same conditions.The reaction afforded the desired azido adduct 3b (23%, 1.1:1 ratio of diastereoisomers) together with the acetoxy derivative 8b (42%, 1.4:1 ratio of diastereoisomers). The structures of the adducts 3b and 8b were determined similarly by a combination of NMR and IR spectroscopies.However, in this case it was possible to prove unambigously by 2D 1H–77Se NMR spectroscopy the position of the phenylselenyl moiety at C-2 of the aglycon in both 3b and 8b.Thus, the spectra (Figure 1) of both compounds 3b and 8b contained correlation cross-peaks between the selenium signals and the signals of all methylene protons of the aglycon. This is possible only if the PhSe moiety is attached to the C-2 carbon. 2D 1H–77Se HMQC experiments10 were optimized for the observation of couplings with JH–Se 5 Hz, hence the spectra do not contain correlation cross-peaks between selenium and the proton at C-2 of the aglycon since the geminal 2JH–Se is known11 to be ca. 10 Hz. The competitive acetoxyphenylselenylation observed can be rationalised as follows. The azidophenylselenylation reaction (Scheme 1, path A) is thought4 to involve oxidation of two azide anions by PhI(OAc)2 into azide radicals followed by their addition to alkene 1 and subsequent trapping of the resulting carbon-centered azido radical 2 with (PhSe)2.Apparently, oxidation of azide anion by PhI(OAc)2 proceeds (Scheme 1, path A) via an exchange reaction leading to PhI(N3)2, which decomposes rapidly into two azide radicals and PhI. When the concentration of azide anion is low (due to the poor solubility of NaN3 in CH2Cl2), substitution of only one AcO group in PhI(OAc)2 may occur (path B) leading to the mixed species PhI(OAc)(N3), which decomposes into azide and PhI(OAc)• radicals.The latter can react with alkene 1 by transfer of an AcO radical and liberation of PhI. Subsequent trapping of the resulting carbon-centered acetoxy radical 3 with (PhSe)2 completes the acetoxyphenylselenylation. Thus, it is likely that the low concentration of azide anion in the reaction medium is responsible for the formation of the acetoxy adducts 8a,b.In order to increase the effective concentration of azide anion we performed the reaction in other solvents. In MeCN the ratio of the adducts 3b and 8b was similar to that obtained in CH2Cl2, but in pyridine or in tetramethylurea the formation of 3b prevailed over 8b (TLC data).Addition of water did not influence the 3b/8b ratio, but decreased the reaction rate significantly. In N,N-dimethylformamide (DMF) the reaction was slow, however, it resulted in the exclusive formation of 3b. We reasoned that the addition of a crown ether would further increase the effective concentration of azide anion and hence accelerate the reaction. Portion-wise addition of PhI(OAc)2 (2 equiv.in total) to a solution of 1b (0.41 mmol), (PhSe)2 (0.6 equiv.) and NaN3 (3 equiv.) in anhydrous DMF (2 ml) containing 18-crown-6 (1 equiv.) at 20 °C (72 h) yielded 86% of azidophenylselenylation adduct 3b as the only product. Formation of acetoxyphenylselenylation by-product 8b was totally suppressed under high effective concentration of azide anion.Further transformation of the azidophenylselenylation adduct 3b into the target 3-aminopropyl glycoside required removal of the phenylselenyl residue and reduction of the azido moiety, which could be accomplished either simultaneously or in a step-wise manner (in any order). However, attempted reduction of azide and simultaneous deselenation of 3b with Bu3SnH and AIBN in refluxing toluene failed leading to complex mixtures resulting probably from competitive reactions of the amine initially formed from azide (cf.ref. 12). Deselenation of 3b using elimination of PhSeOH from the corresponding selenoxide formed in situ by oxidation (H2O2) of 3b afforded nearly quantitatively the corresponding cis and trans vinyl azides 10b together with cis vinyl ether 11b in a 2:4:1 ratio.Hydrogenation (H2, 10% Pd/C, AcOEt, AcOH, 20 °C) of the mixture of 10b and 11b resulted in decomposition. These results suggested that the phenylselenyl moiety should be cleaved only after reduction of the azide. Thus, 3b was first converted into 5b in 62% overall yield in a one-pot reduction/protection sequence: reaction of azide 3b (0.066 mmol) with Ph3P (1.5 equiv.) in refluxing THF (3 ml) and hydrolysis of the phosphimine thus formed by addition of aNMR spectra were recorded with a Bruker AM-300 instrument at 303 K in CDCl3 unless otherwise stated.Acetone was used as an external standard in 1H (2.225 ppm) and 13C NMR (31.45 ppm) and (PhSe)2 in 77Se NMR (–460 ppm11). In all compounds the chemical shifts of the protons and carbons of the sugar moiety were very close to the published6,8,9 ones and thus are not presented.bNumeration of atoms in the aglycon: sugar-1-2-3. c 1H, 13C and 77Se NMR spectra were recorded in C6D6. d 1H and 77Se NMR spectra were recorded in C6D6. eC6H5Se 6.93–7.47 (5H). fC6H5Se 7.03–7.62 (5H). gC6H5Se 7.29–7.61 (5H); (CH3)3C 1.41 (9H). h(CH3)3C 1.43 (9H). iC6H5Se 6.91–7.53 (5H); CH3CO 1.52–1.96 (24H).jC6H5Se 7.01–7.55 (5H); CH3CO 1.62–1.98 (15H). kCH3CO 1.98–2.16 (15H). Table 1 1H, 13C and 77Se NMR data (d/ppm)a,b for aglycons in compounds 3a,b, 5b, 6b, 8a,b, 9b, 10b and 11b. Compound H-1 H-1' H-2 H-3 H-3' C-1 C-2 C-3 Se Other 3ac 3.63 4.07 3.12 3.27 3.35 70.2 43.3 52.5 –575 e 43.7 –580 3bd 4.09 3.60 3.11 3.35 3.29 69.7 43.4 52.4 –576 f 4.01 3.65 3.14 3.35 3.28 69.9 42.9 52.5 –583 5b 3.65–3.80 4.11–4.17 3.42–3.47 3.34 3.44 70.7 44.2 42.4 g 44.5 6b 3.58 3.94 1.88 3.19 3.19 67.9 29.8 37.8 h 8ac 4.39–4.53 4.39–4.53 3.37 3.58–3.70 4.04–4.16 69.5 42.4 64.3 i 42.7 8bd 4.21 3.64 3.40 4.52 4.38 69.4 41.8 64.1 –588 j 4.12 3.69 3.38 4.37 4.27 69.9 42.3 64.2 –585 9b 3.56 3.96 1.90 4.12 4.12 70.7 28.8 61.1 k 10b (E) 4.15 4.32 5.41 6.15 63.2 131.1 115.2 10b (Z) 4.17 5.03 6.33 66.7 129.1 114.6 11b 6.33 5.01 4.17 101.8 129.1 68.1Mendeleev Communications Electronic Version, Issue 1, 1998 (pp. 1–42) H2O (1.7 ml) to give free amine 4b which was transformed into 5b by treatment with N-(tert-butyloxycarbonyloxy)succinimide (7 equiv.). Reductive deselenation of 5b (0.032 mmol) was effected with Bu3SnH (6 equiv.) and AIBN (0.1 equiv.) in refluxing toluene (1 ml) to give in 15 min the target 3-N-(tertbutyloxycarbonylamino) propyl glycoside 6b {[a]D 29 –15° (c 0.25, CHCl3)} in 93% yield.Similarly, deselenation (Bu3SnH–AIBN, toluene) of 8b afforded 3-acetoxypropyl glycoside 9b {[a]D 30 –5° (c 1, CHCl3)} in 83% yield. The terminal position of the NHBoc and AcO groups in the aglycons of 6b and 9b, respectively, was evident from their NMR spectra (Table 1).This fact served as unambiguous proof of the terminal position of the N3 and AcO moieties in the adducts 3b and 8b, and hence of the penultimate position of the phenylselenyl group in 3b and 8b, thus proving the ascribed regioselectivity of azido- and acetoxy-phenylselenylation. In conclusion, the described sequence of reactions (azidophenylselenylation –reduction of azide–deselenation) is a useful approach for the transformation of allyl glycosides into protected 3-aminopropyl glycosides.This work was supported by the President of the Russian Federation (grant no. 96-15-96991) and the Russian Foundation for Basic Research (grant no. 97-03-33037a). The authors are grateful to Yurii V. Mironov (The Higher Chemical College, Moscow) for technical assistance.References 1 Methods in Enzymology (Neoglycoconjugates, parts A and B), eds. Y. C. Lee and R. T. Lee, Academic Press, San Diego, California, 1994, vols. 242, 247. 2 G. Magnusson, A. Ya. Chernyak, J. Kihlberg and L. O. Kononov, in Neoglycoconjugates: Preparation and Application, eds. Y. C. Lee and R. T. Lee, Academic Press, San Diego, California, 1994, p. 53. 3 N. V. Bovin, E. Yu. Korchagina, T. V. Zemlyanukhina, N. E. Byramova, O. E. Galanina, A. E. Zemlyakov, A. E. Ivanov, V. P. Zubov and L. V. Mochalova, Glycoconj. J., 1993, 10, 142. 4 M. Tingoli, M. Tiecco, D. Chianelli, R. Balduci and A. Temperini, J. Org. Chem., 1991, 56, 6809. 5 M. A. Bernstein and L. D. Hall, Carbohydr. Res., 1980, 78, c1. 6 R. T. Lee and Y. C. Lee, Carbohydr. Res., 1974, 37, 193. 7 M. A. Nashed, Carbohydr. Res., 1983, 123, 241. 8 K. Okamoto, T. Kondo and T. Goto, Tetrahedron, 1987, 43, 5919. 9 L. O. Kononov, A. V. Kornilov, A. A. Sherman, E. V. Zyryanov, A. S. Shashkov, G. V. Zatonsky and N. E. Nifant’ev, Bioorg. Khim., 1998, 24, in press. 10 A. Bax and S. Subramanian, J. Magn. Reson., 1986, 67, 565. 11 W. McFarlane and R. J. Wood, J. Chem. Soc., Dalton Trans., 1972, 1397. 12 T. Ercegovic and G. Magnusson, J. Org. Chem., 1996, 61, 179. Received: Moscow, 15th October 1997 Cambridge, 28th November 1997; Com. 7/07604E