The isomerization of specifically tritium-labeled [ 1 (R)-3H]dihydroxyacetone phosphate to D-glyceraldehyde 3-phosphate, catalyzed by the enzyme triosephosphate isomerase, has been studied. The distribution of the 3H label amongst the three possible sites (in [ 1 (R)-3Hldihydroxyace-tone phosphate, in ~-[2-~H]glyceraldehyde 3-phosphate, and in the solvent) has been followed as a function of the extent of the reaction. It is shown that the extent of transfer of the 3H label from the substrate dihydroxyacetone phosphate to Dglyceraldehyde 3-phosphate is between 3 and 6% (depending upon the extent of the reaction). The enzymic base responsible Tiosephosphate isomerase catalyzes the interconversion of the two triose phosphates, dihydroxyacetone phosphate and glyceraldehyde 3-phosphate. The crystalline enzyme from chicken muscle is a dimer of identical subunits each of molecular weight about 26 000 (Putman et a]., 1972). It has no cofactor or specific metal ion requirements. As an isomerase, with a single substrate and a single product, it is particularly amenable to mechanistic study, and the equilibrium constant for the reaction is close enough to unity to allow the reaction to be run in either direction, utilizing the appropriate dehydrogenases as coupling enzymes (a-glycerophosphate dehydrogenase to remove dihydroxyacetone phosphate, or glyceraldehyde-phosphate dehydrogenase to remove glyceraldehyde phosphate). In the presence of excess of the appropriate dehydrogenase and cofactor, the reaction can be studied under conditions such that the isomerase step is rate limiting (Plaut and Knowles, 1972).Aside from these attractive features, this system presents a rare opportunity for crystallographic and spectroscopic study. With most enzymes that use two or more substrates in either direction, even if one of these is water, it is currently impossible to study productive enzyme-substrate complexes directly by crystallographic methods. approaches to the problem of the structure of enzyme-substrate complexes have had to rely on the structures of enzyme-inhibitor complexes, or "dead-end" binary complexes or abortive ternary complexes \lor two substrate systems). From such data, one has to try to deduce the structure of enzyme with real substrate bound (see, e.g., Blake et al., 1967; Steitz et al., 1969). For an isomerase, however, it is possible to find the structure both of native enzyme and of an enzyme-substrate complex. Progress with each for proton abstraction from substrate is, therefore, in almost complete isotopic equilibrium with the solvent. The remaining substrate after partial reaction increases in specific radioactivity as the reaction proceeds, showing that the preferential reaction of ' H substrate is more important than the washing out of 3H label at the stage of the exchanging intermediate. Quantitatively, these results provide the data for the first steps in the analysis described in the previous paper (Albery, W. J., and Knowles, J. R. (1976), Biochemistry, preceding paper in this issue). of these probl...