J. Biochem. 82, 1673-1680 (1977) GLC Analysis of the Hydroxy and Oxo Compounds Produced by the Smith Degradation of Reducing Di and Trisaccharides Haruki YAMAGUCHI and Takafumi MUKUMOTO Department of Agricultural Chemistry, College of Agriculture, University of Osaka Prefecture, Sakai, Osaka 591 Received for publication, June 18, 1977 Reducing di and trisaccharides were oxidized with periodate under conditions minimizing overoxidation, and the hydroxy and oxo compounds produced by subsequent reduction and hydrolysis were analyzed by GLC after trimethylsilylation. The present study has demon strated that GLC analysis of the hydroxy and oxo compounds produced by Smith degradation is useful for linkage analysis of reducing di and trisaccharides. In addition, it is suggested that the rate of periodate oxidation of the internal hexopyranose residue of trisaccharides depends markedly upon their structures. Periodate oxidation, followed by borohydride re of the intermediate formyl esters, leading to ad duction and mild hydrolysis with acid (Smith ditional periodate consumption, so-called overoxi degradation), has been generally used for structural dation. It seems likely, however, that the inter- studies of polysaccharides and complex carbohy mediate esters are relatively stable in weak acid drates. The determination of hydroxy and oxo (6). Based on these suggestions, therefore, it is compounds produced by Smith degradation is expected that each of the intermediate products, expected to provide useful information on the upon subsequent reduction followed by hydrolysis structures of carbohydrates. with acid, should give a characteristic set of hy Previously, we have attempted to determine droxy and oxo compounds depending on the link the hydroxy and oxo compounds resulting from ing position of the original disaccharide, as illus the Smith degradation of non-reducing disac trated for glucobioses with different linking charides by GLC (1, 2), and also applied the positions in Table I. Further, reducing trisac technique to a glycopeptide (3) and polysaccharides charides subjected to Smith degradation might also (4, 5) to elucidate their fine structures. However, give hydroxy and oxo compounds which are GLC analysis of the hydroxy and oxo compounds different in types and/or molar ratios depending on produced by Smith degradation of reducing di and the original linking positions. The study reported trisaccharides has never been attempted. here was undertaken to determine the hydroxy and It is known that aldohexoses and reducing oxo compounds resulting from the Smith degrada end-aldohexose residues linked at positions other tion of reducing di and trisaccharides. It has than C-2 are oxidized by periodate in their pyra now been shown that GLC determination of these nose forms to give formyl esters (6-9). Alkaline compounds is useful for linkage analysis of re or strongly acidic conditions cause rapid hydrolysis ducing di and trisaccharides. Vol. 82, No. 6, 1977 1673 1674 H. YAMAGUCHI and T. MUKUMOTO TABLE I. Expected values of hydroxy and oxo compounds produced from D-glucopyranosyl-D-glucopyranoses by normal Malaprade oxidation followed by borohydride reduction and subsequent hydrolysis. Smith degradation of reducing oligosaccharides EXPERIMENTAL and GLC analysis of the products were carried out substantially as reported previously (2), but Materials-All chemicals were of reagent with some improvements. A reducing oligosac grade. Maltose and lactose were purchased from charide (ca. 0.15 ƒÊmol) was dissolved in 0.2 ml Wako Pure Chem. Co. Isomaltose, melibiose, (0.1 ml for (1•¨4)-ƒÀ-n-hexotrioses) of 0.5 M acetate gentiobiose, and maltotriose were obtained from buffer (pH 3.2) containing 2.14 mg of sodium Sigma Chem. Co. 4-O-ƒÀ-D-Mannopyranosyl-D- metaperiodate and the solution was kept at 30•Ž mannopyranose and O-ƒÀ-D-mannopyranosyl-(1•¨ for a certain period (up to 12 h). The reaction 4)- ƒÀ-D-mannopyranosyl-(1•¨4)-D-mannopyranose mixture was subsequently neutralized by the ad were supplied by Dr. T. Mizuno of Shizuoka Uni dition of 0.1 ml of 1 N sodium hydroxide (0.5 N versity and kojibiose was a gift of Dr. S. Chiba of sodium hydroxide for the reaction mixtures of Hokkaido University. Isomaltotriose and panose (1•¨4)-ƒÀ-D-hexotrioses) and immediately treated were gifts from Dr. M. Takagi of this University. with 0.1 ml of 0.01 N sodium hydroxide containing Laminaribiose and O-ƒÀ-D-galactopyranosyl-(1•¨6)- 2.5 mg of sodium borohydride. The additions of ƒ¿ -D- galactopyranosyl-(1•¨6)-D- galactopyranose the aqueous sodium hydroxide and the sodium were prepared by partial acid hydrolysis of a (1 borohydride solution were carried out with external •¨ 3)-ƒÀ-D-glucan (curdran, (10)) and a heterogalactan cooling in an ice-water bath. The slightly alka (5), respectively, and purified by repeated gel-filtra line solution was allowed to stand overnight at tions through a column of Bio-gel P-2 (100-200 4-6•Ž. The excess sodium borohydride was de- mesh). Cellobiose and cellotriose were also ob composed by addition of 0.1 ml of 1 N acetic acid tained by partial acid hydrolysis of cellulose and and the resulting acidic solution was evaporated purified as described above. 3-O-ƒ¿-D-Mannopyra to a syrup. The residue was dissolved in 0.3 ml nosyl-D-mannopyranose was prepared by partial of methanol and allowed to stand at 50•Ž for acetolysis of a glycopeptide obtained from Taka 15 min. The alcoholic solution was evaporated to amylase A (11). A strongly acidic cation-exchange dryness under reduced pressure, and the residue resin (Dowex 50•~8, 200-400 mesh) was reactivated was subjected to two further evaporations from with 2 N hydrochloric acid immediately before use. methanol. The residue was dissolved in 0.2 ml A weakly basic anion-exchange resin (Amberlite of water and immediately passed through ion IR-4B) which had been stored in the chloride form exchange resin. To desalt the reaction mixture, a after activation, was treated successively with 2 N glass column (5.0•~1.0 cm) was packed with ammonium hydroxide and water just before use. Amberlite IR-4B (OH, 0.3 ml), then Dowex General Methods-All evaporations were con 50x8 (H+, 0.3 ml) was layered over the basic ducted under reduced pressure at a bath temper resin. After applying the sample solution to the ature not exceeding 30•Ž. The consumption of column, the resin was quickly washed twice with periodate was estimated by measuring the ab 0.1 ml volumes of water and then with 1.6 ml of sorbance at 290 nm (12). Oligosaccharides were water. The effluent was collected in a glass tube determined by the orcinol-sulfuric acid reaction (5.0•~1.4 cm) with a Teflon-lined screw cap. The (13). desalting procedure was completed within about J. Biochem. SMITH DEGRADATION PRODUCTS OF REDUCING OLIGOSACCHARIDES 1675 5 min. To the neutral effluent was added 0.2 ml of 0.175 N hydrochloric acid containing 5 mg of hydroxylamine hydrochloride. The glass tube was tightly stoppered and maintained at 80•Ž for 90 min. Then the reaction mixture was evaporated to dryness under reduced pressure and the residue was subjected to repeated evaporations from methanol (3 times). The residue was dissolved in 50ƒÊ1 of anhydrous pyridine containing 25 ƒÊg of 2-ethyl-2-hydroxymethyl-1, 3-propanediol as an in ternal standard and treated with 25 pl of hexa methyldisilazane and 15 pl of trimethylchlorosilane for trimethylsilylation. The reaction mixture was Fig. 1. Production of hydroxy and oxo compounds centrifuged and an aliquot (ca. 1 ƒÊ1) of the super- and consumption of periodate in the Smith degradation natant was injected into a gas chromatograph. of maltose. Analyses of the degradation products were GLC was performed with a Shimadzu 4APF gas performed at intervals during periodate oxidation. chromatograph equipped with a hydrogen-flame Reaction conditions were as described in "EXPERI- ionization detector. A steel column (300•~0.3 cm MENTAL." •ü, Glycerol; •œ, meso-erythritol; ?, i.d.) was used, packed with 5% SE-30 on Chro glycolaldehyde; ?, glyceraldehyde;•~, periodate. mosorb W (60-80 mesh), and the column temper ature was programmed from 100•Ž to 190•Ž at a C-1, C-2, and C-3 of the non-reducing end glucose rate of 4•Ž increment per min. unit. The course of this oxidation reaction may be interpreted by conformational analysis. Figure I also shows that the intermediate formyl ester RESULTS AND DISCUSSION derived from maltose was essentially stable at From previous findings (6-9), it is clear that the pH 3.2, and little overoxidation occurred over a pH at which periodate oxidation is performed period of 7 h. significantly affects both the rate of oxidation and 4-0- ƒÀ-D-Glucopyranosyl-D-glucopyranose the kind of oxidation products formed. We there- (cellobiose), 4-0- ƒÀ-D-galactopyranosyl-D-gluco fore checked carefully the effect of pH on the peri pyranose (lactose), and 4-0-ƒÀ-D-mannopyranosyl- odate oxidation of some reducing glucobioses and D-mannopyranose were subjected to periodate glucotrioses in acetate buffer at various pH's. oxidation, and the hydroxy and oxo compounds Our experiments clearly indicated that overoxi produced by subsequent reduction followed by dation has a minimum in the pH 3 to 3.4 region, hydrolysis were analyzed as performed on maltose. in agreement with the results previously reported Cellobiose gave glycolaldehyde more rapidly, and for maltose (6) and cellobiose (8). Therefore, the therefore a smaller amount of glyceraldehyde, than acetate buffer was adjusted to pH 3.2 for periodate maltose, whereas erythritol production was a little oxidation throughout the present study. slower. Clear-cut Malaprade oxidation was ob- GLC determination of the hydroxy and oxo served for lactose. Only the initial 1 h of peri compounds resulting from the Smith degradation odate oxidation led to the production of gly of maltose was carried out with simultaneous ceraldehyde from lactose.