Mild elimination of a glycosidically linked –O CH 2 CH 2 CH 2 NH 2 spacer-arm Ekaterina V. Shipova and Nicolai V. Bovin* M. M. Shemyakin–A. Yu. Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, 117871 Moscow, Russian Federation. E-mail: bovin@carb.siobc.ras.ru DOI: 10.1070/MC2000v010n02ABEH001222 The spacer-arm –OCH 2 CH 2 CH 2 NH 2 of complex oligosaccharides can be removed by oxidative deamination followed by alkaline -elimination. The application of oligosaccharides to various chemical and bio- chemical studies often requires that these were prepared with the free 1-OH group or as spacered glycosides, i.e., as glyco- sides of alcohols whose second function may serve for conjuga- tion with macromolecules or labels. Spacering is one of the key problems of oligosaccharide synthesis strategy. The introduction of a spacer or prespacer group can be performed (i) at the final stage of the synthesis (by glycosidation of the oligosaccharide with a spacer alcohol) or (ii) at the initial stage when the spacer also serves as a 1-O-protecting group. Both approaches have advantages and drawbacks. The former makes it possible to obtain both free and spacered oligosaccharides; however, spac- ering at the final stage usually leads to a loss of yield, especially when a microscale synthesis is performed. The second approach is more economical; however, only the spacered product can be obtained. Here we describe a methodology based on the second ap- proach, namely, the removal of the spacer group –OCH 2 CH 2 – CH 2 NH 2 , which was often employed in our studies. 1,† A simple one-stage removal of the spacer-arm by acid hy- drolysis or acid acetolysis in oligosaccharides 1a and 1b was unsuccessful because of cleavage of the acid-labile Fuc 1-2Gal bond. ‡ In searching for a mild despacering procedure, we examined the applicability of the following two-stage approach. The spacer- arm –OCH 2 CH 2 CH 2 NH 2 in compound 3a or 3b was subjected sequentially to the Corey method 3 and alkaline -elimination by treatment of the compounds with 3,5-di-tert-butyl-1,2-benzo- quinone in methanol and acidification of intermediate azomethines † As a rule, this spacer is used as the trifluoroacetamidopropyl group –OCH 2 CH 2 CH 2 NHCOCF 3 , which can be quantitatively deblocked by an alkali. 2 ‡ No cleavage of the glycosidic bond Gal -sp leading to desired per- acetate 2a or 2b was observed when acid acetolysis of the corresponding trifluoroacetamidopropyl glycosides of trisaccharides 1a or 1b (AcOH/ Ac 2 O/H 2 SO 4 , 100:100:1, 0–20 °C) was carried out. More severe condi- tions (an increase in the amount of sulfuric acid or a higher temperature) led to the cleavage of the Fuc 1-2Gal bond. Analogous results were obtained upon acid acetolysis (AcOH/Ac 2 O/H 2 SO 4 , 100:100:1, 5 °C) of A trisaccharide acetamidopropyl glycoside 4a. However, under the same conditions, the similar B trisaccharide glycoside 4b was converted into peracetate 2b, 90%. 5 with oxalic acid. The final acetylation with acetic anhydride in pyridine simplified the chromatographic purification (see Scheme 1). In this way, the azomethine obtained from the spacered blood group trisaccharide B gave rise to 41% aldehyde 6 § and 39% –OC H 2 CH 2 CH 2 NHAc derivative 4b. The azomethine obtained from spacered blood group trisaccharide A under the same con- ditions gave rise to benzoxazole 7 § (70%). Although according to Corey and Achiva 3 primary amines can be converted into either carbonyl compounds or benzoxazole depending on carbon chain branching, it is surprising that trisaccharides having only minor differences in the sites distant from the reaction centre behaved in such a different fashion under the same conditions ‡, ¶ . The next stage of the proposed method has to be alkaline -elimination of the free trisaccharide from derivatives 6 and 7. However, the treatment with aqueous alkali solutions would result in splitting off the monosaccharide from the position O-3 of the despacered Gal moiety due to additional -elimination, where the deprotected 1-OH group plays a role of the aldehyde group. To avoid the secondary -elimination (so-called peeling), we used the conditions of base-catalysed acetolysis/acetylation (AcOH + AcONa + Ac 2 O) described earlier for the elimination of complex oligosaccharides from the protein core: 4 acetic an- hydride converts 1-OH into 1-OAc and thus blocks it. Thus, the treatment of compounds 6 and 7 with an AcOH–AcONa–A c 2 O mixture at 110 °C for 48 h †† gave despacered derivative 2b or 2a in 68 or 93% yield, respectively. Note that we did not optimise the conditions of either the Corey reaction or -elimination and did not perform a one-pot process, which can increase the yields of final oligosaccharides. Thus, we have demonstrated that 3-aminopropyl glycosides of complex oligosaccharides can be converted into despacered forms in non-acidic conditions. § A solution of amine 3 (0.34 mmol) and 3,5-di-tert-butyl-1,2-benzo- quinone (83 mg, 0.37 mmol) in 30 ml of MeOH was stirred under argon at room temperature. The colour of the reaction mixture changed from dark brown to green in 1 h. The stirring was continued for 24 h; next, oxalic acid dihydrate was added to pH 2, the reaction mixture was evapo- rated to dryness in vacuo, and the solid residue was washed with a 2:1 ethyl acetate–b enzene mixture to remove the remaining benzoquinone. The residue was conventionally acetylated with Ac 2 O in Py for 24 h, and product 6 or 7 was separated by column chromatography. Selected spectral data: 6: 1 H NMR (CDCl 3 ) d: 1.19 (d, 3H, Me'' ), 1.9–2.18 (27H, 9OAc), 3.37 (m, 2H, CH 2 CHO), 3.67 dd (1H, H 2 , J 2,1 7.5 Hz, J 2,3 10 Hz), 3.78 (m, 1H, H 5 ), 3.98, 4.08 (2H, H 6 ), 4.09, 4.30 (2H, H 6' ), 4.10 (m, 1H, OCHCH 2 ), 4.41 (1H, H 5'' ), 4.45 (m, 1H, OCH' CH 2 ), 4.46 (d, 1H, H 1 ), 4.51 (m, 1H, H 5' ), 5.12 (1H, H 3'' ), 5.18 (1H, H 2'' ), 5.23 (1H, H 4'' ), 5.34 (dd, 1H, H 2' , J 2',1' 3 Hz, J 2',3' 10 Hz), 5.37 (d, 1H, H 1' ), 5.39 (d, 1H, H 4 , J 4,3 3.5 Hz), 5.46 (dd, 1H, H 3' ), 5.52 (d, 1H, H 1'' , J 1'' ,2'' 3.5 Hz), 5.61 (dd, 1H, H 4' ). FAB MS, m/z: 922 (M + ). 7: 1 H NMR (CDCl 3 ) d: 1.07 (d, 3H, Me'' ), 1.35 (s, 9H, 3Me-q), 1.5 (s, 9H, 3Me-q), 3.30 (t, 2H, CH 2 C=), 3.81 (dd, 1H, H 2 , J 2,1 7 Hz, J 2,3 9 Hz), 3.89 (dd, 1H, H 3 , J 3,4 3 Hz), 4.13 (m, 1H, OCHCH 2 ), 4.26 (1H, H 5' ), 4.43 (m, 1H, OCH'CH 2 ), 4.5 (1H, H 2' ), 4.51 (d, 1H, H 1 ), 5.02 (dd, 1H, H 3' , J 3',2' 11 Hz, J 3',4' 3 Hz), 5.2 (1H, H 5'' ), 5.24 (d, 1H, H 1' , J 1',2' 3 Hz), 5.25 (1H, H 3'' ), 5.32 (dd, 1H, H 2'' , J 2'' ,3'' 11 Hz, J 2'' ,1'' 3.5 Hz), 5.37 (dd, 1H, H 4 ), 5.46 (dd, 1H, H 4' ), 5.51 (d, 1H, H 1'' ), 6.17 (d, 1H, NHAc, J NH,2' 9 Hz), 7.54 (d, 1H, H-q), 7.55 (d, 1H, H-q). FAB MS, m/z: 1123 (M + ). ¶ Corey and Achiva 3 also described the formation of a benzoxazole product without an explanation of the over-oxidation. O OR O O Y RO OR OR O RO OR OR O OR sp Me 1a sp = OCH 2 CH 2 CH 2 NHCOCF 3 , R = H, Y = NHAc 1b sp = OCH 2 CH 2 CH 2 NHCOCF 3 , R = H, Y = OH 2a sp = OAc, R = Ac, Y = NHAc 2b sp = OAc, R = Ac, Y = OAc 3a sp = OCH 2 CH 2 CH 2 NH 2 , R = H, Y = NHAc 3b sp = OCH 2 CH 2 CH 2 NH 2 , R = H, Y = OH 4a sp = OCH 2 CH 2 CH 2 NHAc, R = Ac, Y = NHAc 4b sp = OCH 2 CH 2 CH 2 NHAc, R = Ac, Y = OAc Mendeleev Commun., 2000, 10(2), 63–64 – 63 –