13C and deuterium isotope effects suggest an aldol cleavage mechanism for L-ribulose-5-phosphate 4-epimerase.

On the basis of (13)C and deuterium isotope effects, L-ribulose-5-phosphate 4-epimerase catalyzes the epimerization of L-ribulose 5-phosphate to D-xylulose 5-phosphate by an aldol cleavage to the enediolate of dihydroxyacetone and glycolaldehyde phosphate, followed by rotation of the aldehyde group and condensation to the epimer at C-4. With the wild-type enzyme, (13)C isotope effects were 1.85% at C-3 and 1.5% at C-4 at pH 7, with the values increasing to 2.53 and 2.05% at pH 5.5, respectively. H97N and Y229F mutants at pH 7 gave values of 3.25 and 2.53% at C-3 and 2. 69 and 1.99% at C-4, respectively. Secondary deuterium isotope effects at C-3 were 2.5% at pH 7 and 3.1% at pH 5.5 with the wild-type enzyme, and 4.1% at pH 7 with H97N. At C-4, the corresponding values were 9.6, 14, and 19%. These data suggest that H97N shows no commitments, while the wild-type enzyme has an external commitment of approximately 1.4 at pH 7 and an internal commitment independent of pH of approximately 0.6. The Y229 mutant shows only the internal commitment of 0.6. The sequence of the epimerase is similar to those of L-fuculose-1-phosphate and L-rhamnulose-1-phosphate aldolases for residues in the active site of L-fuculose-1-phosphate aldolase, suggesting that Asp76, His95, His97, and His171 of the epimerase may be metal ion ligands, and Ser44, Gly45, Ser74, and Ser75 may form a phosphate binding pocket. The pH profile of V/K for L-ribulose 5-phosphate is bell-shaped with pK values of 5.94 and 8.24. The CD spectra of L-ribulose 5-phosphate and D-xylulose 5-phosphate differ sufficiently that the epimerization reaction can be followed at 300 nm.

Role of metal ions in the reaction catalyzed by L-ribulose-5-phosphate 4-epimerase.

H97N, H95N, and Y229F mutants of L-ribulose-5-phosphate 4-epimerase had 10, 1, and 0.1%, respectively, of the activity of the wild-type (WT) enzyme when activated by Zn(2+), the physiological activator. Co(2+) and Mn(2+) replaced Zn(2+) in Y229F and WT enzymes, although less effectively with the His mutants, while Mg(2+) was a poorly bound, weak activator. None of the other eight tyrosines mutated to phenylalanine caused a major loss of activity. The near-UV CD spectra of all enzymes were nearly identical in the absence of metal ions and substrate, and addition of substrate without metal ion showed no effect. When both substrate and Zn(2+) were present, however, the positive band at 266 nm increased while the negative one at 290 nm decreased in ellipticity. The changes for the WT and Y229F enzymes were greater than for the two His mutants. With Co(2+) as the metal ion, the CD and absorption spectra in the visible region were different, showing little ellipticity in the absence of substrate and a weak absorption band at 508 nm. With substrate present, however, an intense absorption band at 555 nm (epsilon = 150-175) with a negative molar ellipticity approaching 2000 deg cm(2) dmol(-1) appears with WT and Y229F enzymes. With the His mutants, the changes induced by substrate were smaller, with negative ellipticity only half as great. The WT, Y229F, H95N, and H97N enzymes all catalyze a slow aldol condensation of dihydroxyacetone and glycolaldehyde phosphate with an initial k(cat) of 1.6 x 10(-3) s(-1). The initial rate slowed most rapidly with WT and H97N enzymes, which have the highest affinity for the ketopentose phosphates formed in the condensation. The EPR spectrum of enzyme with Mn(2+) exhibited a drastic decrease upon substrate addition, and by using H(2)(17)O, it was determined that there were three waters in the coordination sphere of Mn(2+) in the absence of substrate. These data suggest that (1) the substrate coordinates to the enzyme-bound metal ion, (2) His95 and His97 are likely metal ion ligands, and (3) Tyr229 is not a metal ion ligand, but may play another role in catalysis, possibly as an acid-base catalyst.