HETEROCYCLES

Communication | Special issue | Vol. 80, No. 1, 2010, pp. 115-123
Received, 15th April, 2009, Accepted, 11th May, 2009, Published online, 12th May, 2009.
DOI: 10.3987/COM-09-S(S)24

■ Mild and Selective O-Glycosylations of Primary Alcohols with the Thioglucosaminide Derivative Promoted by N-Iodosuccinimide and HBF4-Adosrobed on Silica Gel

Michio Kurosu* and Kai Li

Department of Microbiology, Immunology and Pathology, College of Veterinary Medicine and Biomedical Sciences, Colorado State University, 1682 Campus Delivery, Fort Colllins, CO 80523-1682, U.S.A.

Abstract

Selective glycosylations of primary alcohols with the thioglucosaminide 1 are achieved by using NIS and HBF4-SiO2. HBF4-SiO2 is a mild Brønsted acid which requires primary alcohols or phenols to effectively activate 1 with NIS at rt. A wide range of functional groups are tolerated under these conditions.

Drug design in recent years has attempted to explore new chemical spaces resulting in more complex and larger molecular weight molecules. These molecules often possess limited water solubility. In general, these poorly soluble active new chemical entities present problems in drug administration and formulation.1 Prodrug strategy has employed to improve the solubility profile of molecules, resulting in increase of the intestinal absorption of poorly water-soluble hydrophobic compounds.2 Glycosylation is one of the approaches for increasing water solubility and cellular uptake; a number of examples of the enhancement in the solubility and efficacy of natural products via glycosylation have been reported.3 However, rational attempts to improve the oral absorption of poorly soluble active new chemical entities by glycosylation(s) are complicated by the fact that the mechanisms by which new molecules are transported from the gastro intestine to the blood are unknown. In addition, extensive regeneration studies of inactive glycosyl prodrugs by the cellular enzymatic reaction of glycosidase (enzyme specific activation) or hydrolysis in vivo are necessary. To date, glucuronide-, glucosyl- and galactosyl-prodrug approaches have been applied to known chemical entities, and their transport and regeneration mechanisms at specific sites have been well studied.4 Although glucosaminyl or N-acetyl-glucosaminyl prodrugs are known to exhibit beneficial physicochemical properties (i.e. shorter half-live of hydrolysis (at pH 7.4), and susceptibility to β-D-N-acetyl-glucosaminidase),5 applications of these prodrug systems to complex molecules have been limited due in part to lack of robust selective glycosylations with glucosamine derivatives. Herein, we report a mild and selective glycosylation of primary alcohols with thioglucosaminide with N-Troc (2,2,2-trichloroethoxycarbonyl) group using NIS (N-iodosuccinimide) and HBF4-SiO2.
Thioglycosaminide with N-Troc such as 1 has been widely utilized in synthesis of N-acetylglucosamine-containing glycoconjugates.6 The thioglycoside 1 is generally coupled to primary or secondary alcohols in good to excellent yields using a combination of NIS and TfOH7 or other variations of triflate sources.8 The Troc group is reliably deprotected via a wide range of reductive elimination conditions (i.e. Zn, Cd, or In in the presence of a mild Brønsted acid).9 Although reported O-glycosylation conditions for 1 are useful for acyl- or benzyl-protected glycosyl acceptors, application of glycosylation of 1 using NIS and strong Brønsted acids or strong acid salts to glycosyl acceptors possessing acid sensitive groups (i.e. silyl or Boc group) generally requires laborious optimization efforts. In our hands, the NIS/TfOH promoted glycosylation of 2h with 1 gave the desired product 3h in poor yield due to desilylation of 2h or 3h (Scheme 1).10 In addition, no reliable condition for selective glycosylations with 1 against primary alcohols of polyol molecules has been reported to date. In our investigations of an alternative acid which can promote glycosylation of 2h with 1 and NIS, we observed that HBF4-absorbed on silica gel (HBF4-SiO2)11 efficiently catalyzed the reaction to furnish 3h in over 95% yield within 3 h at rt. On the other hand, the glycosylation of the secondary alcohol 2s with 1 did not

provide the disaccharide 3s under the same conditions even after prolonged reaction time. Interestingly, 1 and 2s could be recovered as their intact forms in almost quantitative yields (Scheme 1).12 The glycosyl donor 1 is activated by using a NIS/TfOH condition in the absence of acceptor alcohols to furnish the corresponding anomeric alcohol. Thus, an unusual reaction mechanism, the primary alcohol 2h efficiently participates in the process of activation of 1, observed in Scheme 1 may be attributed to the preferential glycosylation of the primary alcohol and tolerance of the silyl group in these reactions. Indeed, HBF4-SiO2 1) no longer shows a strong Brønsted acid characteristic as observed for HBF4•OEt2 in aprotic solvents, 2) does not catalyze the cleavage of acetal, and acid sensitive ether and Boc groups at ambient temperatures, 3) has a strong affinity for alcohols, but a weak affinity for carbonyls,13 and 4) does not cause HBF4 leaching from silica gel during the reactions or by washing with organic solvents including MeOH.14 Thus, such a strongly bound HBF4 on SiO2 serves as a moderate to weak Brønsted acid.15
In order to generalize the use of NIS/HBF4-SiO2 for selective glycosylations with 1, a wide range of model donor molecules, primary, secondary and phenolic alcohols, were examined. In these experiments, acceptors (3~5 equiv against 1), NIS (2 equiv against 1), 3 Å MS (~two times its weight in 1),16 and HBF4-SiO2 (0.8 mmol/g, 2 equiv against 1) were employed in the reactions in CH2Cl2 at rt. Selected examples are summarized in Table 1 and 2. All primary and phenolic alcohols investigated were glycosylated to furnish the corresponding β-glycosides within 3 h. Reaction rates of the phenols 2b and 2c were faster than those of benzyl alcohol (2a) and alkanols (entries 1–8 in Table 1). Glycosylations of structurally simple alcohols 2a–c resulted in the formation of corresponding β-glycosides 3a–c in greater than 95% yields. The β-hydroxy ester 2d and serine derivatives, 2e and 2f, were coupled with 1, and the desired products 3d, 3e, and 3f were isolated with good to moderate yields (entries 4, 5). The Boc group of 2e was stable under these conditions, however, the isolated yield of 3e was 30% (>95% yield based on recovering 2e) due in part to the attenuated nucleophilicity of the hydoxy group of 2e by forming a hydrogen bond with the neighboring group. Similarly, glycosylation of the Fmoc-protected serine 2f yielded 3f in 50 % yield. The primary alcohols of the pyranose and furanose derivatives 2g, i, j were efficiently coupled with 1 to yield the corresponding β-linked disaccharides 3g, i, j, respectively (entries 6–8). As observed in the reaction of 2s with 1 in Scheme 1, secondary alcohols showed poor reactivity against 1 under the HBF4-SiO2 promoted glycosylation conditions (Table 2). Glycosylation of 2-proponol (2k) with 1 provided the β-linked glycoside 3k in less than 30% yield after 15 h (entry 9); no detectable amount of 3k was identified on TLC within 1 h. However, trace amounts of 3k was formed after 3 h. Less than 5% of the β-cholesteryl glycoside was isolated after 15 h (entry 10). The threonine derivative 2m and secondary alcohols of the pyranose and furanose derivatives 2n–p did not provide detectable amounts of the corresponding products on TLC even after 15 h (entries 11–13).

Having considered the observation in which secondary alcohols are not effectively coupled with 1 under the NIS/HBF4-SiO2 conditions, we explored selectivities (primary vs. secondary alcohol) of glycosylations of a diol and a triol such as 2q and 2r. Glycosylation of 2q furnished a 7:1 mixture of β-glycosides, 3qa and 3qb, in 80% yield (entry 14). Similar selectivity against the primary alcohol was observed in glycosylation of 2r, and the disaccharide 3r was isolated in 72% yield (entry 15). In both cases the over-reaction products were not isolated and the regioisomer is the only byproduct in these reactions. It is worthwhile mentioning that the product selectivities of the reaction with 2q and 2r were not improved by shortening the reaction time (3 h) or lowering the reaction temperatures (-20–0 oC). Less primary/secondary selectivities observed in glycosylations of the glucose and ribose derivatives, 2q and 2r, are not clear. Nonetheless, we observed useful level of selectivity against primary alcohol in glycosylation of a diol and a triol.
The NIS/HBF4-SiO2 promoted selective glycosylations of primary alcohols observed for 1 were applicable to the other tolyl thioglycosides such as 4 and 5. However, the orthoesters 6 and 7 were formed in almost quantitative yields within 10 min,18 and 6 and 7 were not rearranged to the corresponding glycosides 8 and 9 even after 15 h. Isolated 6 and 7 could be completely converted to 8 and 9, respectively, via HBF4•OEt2 in CH2Cl2 (Scheme 2). The experimental data summarized in Scheme 2 confirm that HBF4-SiO2 is a mild Brønsted acid and acid-sensitive orthoesters 6 and 7 are not hydrolyzed by HBF4-SiO2.

In conclusion, we have demonstrated NIS/HBF4-SiO2 promoted selective glycosylations of primary alcohols with thioglucosaminide 1. Interestingly, HBF4-SiO2 requires the thioglycosides (glycosyl acceptors) and primary or phenolic alcohols to efficiently activate NIS.19 As demonstrated in Scheme 1 and 2, a wide range of functional groups are tolerated under NIS/HBF4-SiO2 promoted glycosylation conditions. Although mechanism of the reactions associated with HBF4-SiO2 is far from understood, beneficial features of NIS/HBF4-SiO2 described above can be applied to selective glycosylations of primary alcohols of target molecules, which are difficult to be glycosylated selectively using previously reported conditions.6

ACKNOWLEDGEMENTS
We thank the National Institutes of Health and Colorado State University for generous financial support. Mass spectra were obtained on instruments supported by the NIH Shared Instrumentation Grant GM49631.

References

1. S. S. Davis, Drug Disc. Today, 2005, 10, 249. CrossRef
2. F. Kratz, I. A. Müller, C. Ryppa, and A. Warnecke, Chem. Med. Chem., 2008, 3, 20 and references cited therein. CrossRef
3. (a) M. Hashiba, H. Horabayashi, M. Nishikawa, and Y. Takayama, J. Control. Release., 1997, 46, 129; CrossRef (b) J. J. Lowery, L. Yeomans, C. M. Keyari, P. Davis, F. Porreca, B. I. Knapp, J. M. Bidlack, E. J. Bilsky, and R. Polt, Chem. Biol. Drug Des., 2007, 69, 41; CrossRef (c) F-X. Bernard, S. Sable, B. Cameron, J. Provost, and J.-F. Crouzet, Antimicrob. Agents Chemother., 1997, 41, 992; (d) A. P. Jorge, H. Horst, E. de Sousa, M. G. Pizzolatti, and F. R. M. B. Silva, Chem. Biol. Interact., 2004, 149, 89; CrossRef (e) B. Mishra, K. I. Priyadarsini, M. S. Kumar, M. K. Unnikrishnan, and H. Mohan, Bioorg. Med. Chem., 2003, 11, 2677; CrossRef (f) J. Park, H. Y. Lee, M-H. Cho, and S. B. Park, Angew. Chem. Int. Ed., 2007, 46, 2018; CrossRef (g) H. Y. Lee, J-T. Kwon, M. Koh, M-H. Cho, and S. B. Park, Bioorg. Med. Chem. Lett., 2007, 17, 6335. CrossRef
4. R. G. G. Leendersa, E. W. P. Damena, E. J. A. Bijstervelda, H. W. Scheeren, P. H. J. Houbab, I. H. van der Meulen-Muilemanb, E. Bovenb, and H. J. Haismab, Bioorg. Chem. Lett., 1999, 7, 1597.
5. (a) G. L. Budy, D. T. Pals, J. A. Lawson, S. J. Couch, M. F. Lipton, and M. A. Mauragis, J. Med. Chem., 1990, 33, 2276; CrossRef (b) C. A. Valdez, J. E. Saavedra, B. M. Showalter, K. M. Davies, T. C. Wilde, M. L. Citro, J. J. Barchi, J. R. Deschamps, D. Parrish, S. El-Gayar, U. Schleicher, C. Bogdan, and L. K. Keefer, J. Med. Chem., 2008, 51, 3961. CrossRef
6. (a) M. Imoto, H. Yoshimura, T. Shimamoto, N. Sakaguchi, S. Kusumoto, and T. Shibata, Bull. Chem. Soc. Jpn., 1987, 60, 2205; CrossRef (b) H. Paulsen and C. Krogmann, Liebigs Ann. Chem., 1989, 1203; CrossRef (c) H. Paulsen and B. Helpap, Carbohydr. Res., 1991, 216, 289; CrossRef (d) W. Dullenkopf, J. C. Castro-Palomino, E. I. Ibrahim, E. H. El-Ashry, and R. R. Schmidt, Eur. J. Org. Chem., 1998, 2305; CrossRef (e) G. K. I. Nilsson, Tetrahedron Lett., 1997, 38, 133; CrossRef (f) Z. Zhang, I. Ollmann, X.-S. Ye, R. Wischnat, T. Baasov, and C. H. Wong, J. Am. Chem. Soc., 1999, 121, 734. CrossRef
7. (a) G. H. Veeneman and J. H. van Boon, Tetrahedron Lett., 1990, 31, 275; CrossRef (b) G. H. Veeneman, S. H. van Leeuwen, and J. H. van Boom, Tetrahedron Lett., 1990, 31, 1331; CrossRef (c) P. Konradsson, U. E. Udodong, and B. Fraser-Reid, Tetrahedron Lett., 1990, 31, 4313. CrossRef
8. (a) K. Fukase, A. Hasuoka, I. Kinoshita, Y. Aoki, and S. Kusumoto, Tetrahedron, 1995, 51, 4923; CrossRef (b) N. Tanaka, F. Ohnishi, D. Uchihata, S. Torii, and J. Nokami, Tetrahedron Lett., 2007, 48, 7383 and references cited therein. CrossRef
9. (a) A. J. Pearson and W. R. Roush, Ed.; Handbook of Reagents for Organic Synthesis: Activating and Agents and Protecting Groups. Wiley, New York, 1999; (b) T. W. Greene and P. G. M. Wuts, In Protective Groups in Organic Synthesis. 3rd Ed. John Wiley and Sons: New York, 1999.
10. p-Tolylsulfenyl trifluoromethanesulfonate (p-TolSOTf) was effective in promoting the glycosylation of 2h with 1. However, addition of ScCO3 was indispensable to avoid desialylation of the TBS group, see: M. Kurosu and K. Li, J. Org. Chem., 2008, 73, 9767. CrossRef
11. Anhydrous HBF4 on SiO2 can readily be obtained by a simple procedure of mixing SiO2 with ~48% aq. HBF4 and subsequent removal of H2O, and is not hygroscopic.
Preparation of HBF4-SiO2: To a magnetically stirred suspension of silica gel (20 g, 230–400 mesh) in Et2O (100 mL), HBF4 (48% in water, 2.1 mL, 16 mmol) was added at rt. After 3 h, the mixture was concentrated and the residue was dried under high vacuum (1~5 mmHg) at 110 °C for 72 h to give anhydrous HBF4-SiO2 (HBF4: 0.8 mmol/g). Titration studies of the recovered HBF4-SiO2 and washing solvents supported that negligible amounts of HBF4 from SiO2 are leached during the reactions or washing.
12. In our laboratory, the glycopeptide core of lipid II 3s has been synthesized via NIS/AgBF4 in CH2Cl2. Under these conditions a gram quantity of 3s was synthesized with >80 % yield (based on 2s).
13. Taking advantage of strong affinity of HBF4-SiO2 against primary or phenolic alcohols, we have used HBF4-SiO2 as an affinity matrix for separating primary or phenolic alcohols out of crude materials.
14. Titrations were performed with an analytical 0.1 N NaOH solution and phenolphtalene as an indicator.
15. Recently, HBF4-SiO2 was demonstrated for synthesis of acetals from aldehydes or thiiranes from oxiranes, see: (a) V. T. Kamble, B. P. Bandgar, N. S. Joshi, and V. S. Jamode, Synlett, 2006, 2719; CrossRef (b) B. P. Bandgar, P. V. Abasaheb, V. T. Kamble, and J. V. Totre, J. Mol. Catal. A: Chem., 2007, 273, 114. For a study of glycosylations with thioglycosides promoted by NIS and HClO4-SiO2, see; (c) B. Mukhopadhyay, B. Collet, and R. A. Field, Tetrahedron Lett., 2005, 46, 5923. CrossRef
16. MS 3 Å is not indispensable. In order not to obtain inconsistent results caused by adventitious water, activated MS 3 Å (~two times its weight in 1) was added in all reactions.
17. The following example represents typical experimental procedure: To a stirred suspension of 2i (282 mg, 0.6 mmol), 1 (117 mg, 0.2 mmol) and 3Å MS (550 mg) in dry CH2Cl2 (10 mL) under N2 atmosphere was added NIS (45 mg, 0.2 mmol) and HBF4-SiO2 (250 mg, 0.2 mmol). After 3 h at rt, the reaction mixtures were filtrated and washed with EtOAc (5 mL). The combined organic phase was evaporated in vacuo. Purification by silica gel chromatography (hexanes/EtOAc = 2/1) gave the 3i (178 mg, 96 %). Data for 3i: [α]20D +37° (c 1.0, CHCl3); IR (film): 1746, 1710, 1666, 1070 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.79 (d, J = 8.4 Hz, 1H), 7.43-7.28 (m, 10H), 5.97 (s, 1H), 5.80 (d, J = 8.0 Hz, 1H), 5.48 (s, 2H), 5.17 (m, 2H), 5.09 (m, 1H), 4.88 (d, J = 12.4 Hz, 1H), 4.81 (d, J = 12.4 Hz, 1H), 4.73 (m, 1H), 4.72 (s, 2H), 4.60 (d, J = 8.8 Hz, 1H), 4.55 (d, J = 12.0 Hz, 1H), 4.33-4.26 (m, 3H), 4.10 (dd, J = 2.0, 12.4 Hz, 1H), 4.03 (d, J = 3.6 Hz, 1H), 3.75-3.69 (m, 3H), 3.63 (m, 1H), 3.24 (s, 3H), 2.09 (s, 3H), 2.04 (s, 3H), 2.03 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 171.1, 170.9, 169.5, 163.0, 154.2, 151.0, 139.1, 138.0, 137.5, 128.6, 128.5, 128.2, 127.9, 101.8, 101.1, 95.5, 89.2, 80.5, 78.8, 76.5, 74.6, 72.3, 72.1, 71.9, 70.3, 68.4, 67.5, 61.9, 58.2, 56.6, 20.9, 20.8, 20.7; HRMS (ESI) Calcd. for C40H46Cl3N3NaO16 (M + Na)+: 952.1841, found: 952.1848.
18. The formation of orthoesters has previously been reported with a number of different glycosyl donors which contain a 2-O-acetyl group, see: A. H. Harreus and H. Kunz, Liebigs Ann. Chem., 1986, 717. CrossRef
19. The reaction mechanism of the HBF4-SiO2 promoted glycosylations with 1 and NIS remains far from understood. However, we speculate that primary alcohols interact with HBF4-SiO2 and alcohol protons would activate NIS by coordination with the imido carbonyl of NIS.