3l6 SHLOMO HESTRIN 



can be expected to exert a more profound effect on the reaction rate than 

 a similar change effected at a relatively remote atom position, it might be 

 possible on the basis of enzyme specificity studies in a family of substrate 

 analogues to infer the position of the bond at which the enzymic cleavage 

 occurs [3]. However, an attempt to derive this inference from a considera- 

 tion of the substrate range of the polymer-synthesizing carbohydrases — 

 levansucrase and dextransucrase — encounters the difficulty that in both 

 these systems minor changes in glycose structure at atom positions on 

 both sides of glycosidic oxygen in sucrose result in a complete suppression 

 of the reactivity [4]. The question therefore arose whether these poly- 

 merizing enzymes are glycosylases, as are the common hydrolases, or 

 whether they are glycosidases and thus perhaps different in a salient aspect 

 of structure from the common hydrolase group. An investigation which 

 made it possible to select between these concepts was carried out in 

 collaboration with Dr. Frank Eisenberg during a visit to the National 

 Institute of Health at Bethesda. 



Sucrose was synthesized enzymically by levansucrase-catalyzed trans- 

 fer of fructose from raffinose (melibiosyl fructoside) to [i-^^0]-glucose. 

 Sucrose formed was converted chemically first into its octoacetate and 

 then, by transacetylation to methanol, was reconstituted in crystalline 

 form. If a fructosido radical is transferred by the action of levansucrase, 

 synthesized sucrose should have been devoid of any ^^O. If fructosyl rather 

 than fructosido was the group transferred from raffinose to the added 

 glucose acceptor, the atom excess of ^*^0 in the synthesized sucrose was 

 expected, on the basis of the atom excess of ^^O in the glucose used, to be 

 o • 65 %. Experimentally the atom excess of ^^O in the recovered crystallized 

 sucrose was found to be o-6\%, in close agreement with the calculated 

 value. Hence it could be concluded that levansucrase, like the common 

 hydrolyzing /S-fructofuranosidase, is a glycosylase and not a glycosidase. 



P^O]-Sucrose which had been synthesized in the above manner was 

 then incubated with dextransucrase of Leiiconostoc mesenteroides. Dextran 

 formed in this system and isolated from aqueous solution by repeated 

 precipitation with ethanol proved to be devoid of ^^O. Hence we are able 

 to draw the further conclusion that dextransucrase, like levansucrase, is a 

 glycosylase and not a glycosidase. 



It follows that levansucrase and dextransucrase attack sucrose each on 

 a different side of the oxygen bridge. Should we be inclined, accordingly, 

 to write the structure of sucrose with a Lipman wiggle ( ^ ) to indicate the 

 site of the "high-energy-bond", we would be at a loss to decide the side 

 of the oxygen bridge in which the symbol could properly be placed. 



The results with dextransucrase further afford conclusive proof that 

 the only site in enzymically synthesized sucrose at which ^^O occurred 

 was the glycosidic oxygen, P^O]- Sucrose synthesized as described by the 



