The catalytic site of all dihydrofolate reductases contains an invariant carboxylic acid, equivalent to Asp-27 in Escherichia coli dihydrofolate reductase (ecDHFR). It has been found that various kinetic and ligand binding properties of ecDHFR show a pH profile with a pKa of about 6.5. The group responsible for this pKa is often assumed to be carboxyl group of Asp-27. To determine the ionization state of this carboxyl and its pKa, we have employed a novel method, based on Raman difference spectroscopy, to obtain its vibrational spectrum in situ. The method is general for the study of protein carboxyl groups, which are often significantly implicated in protein function and structure; this study establishes the method's limits and problems. The Raman difference spectrum between wild-type ecDHFR and the Asp-27 to serine mutant (D27S) in the pH range 5.6-9.0 has been taken. No protonation of the carboxyl group was detected, implying that its pKa is probably less than 5.0. We did, however, detect a pH dependence in the intensity of Raman bands in the difference spectrum with a pKa of 6.3, indicating that the apo enzyme undergoes a pH-dependent conformational change. Because the carboxyl group of Asp-27 at the active site is the only ionizable group in the binding site, other groups, away from the catalytic site, must be responsible for the pH behavior of ecDHFR.

The NAD(P) dependent dehydrogenases and reductases stereospecifically catalyze the transfer of a hydride ion from C4 of the dihydronicotinamide of the NAD(P) ring to the substrate. We have investigated the vibrational structure of the important C4−H coordinate for NAD(P)H and NADP+ bound to three enzymes in binary and ternary (Michaelis mimics) complexes:  the A-side specific lactate dehydrogenase (LDH) and dihydrofolate reductase (DHFR) and the B-side specific glycerol-3-phosphate dehydrogenase (G3PDH). This is achieved by specifically deuterating the C4 pro-R or pro-S hydrogens of the reduced ring or the C4 hydrogen of the oxidized ring, which results in a vibrational mode localized to the stretching motion of the labeled C4−H bond. We observed relatively minor changes in the stretch frequencies of the C4−H bonds showing that the electronic nature of the bond is not substantially modified by cofactor binding, a mechanism previously proposed to be involved in enzymic “activation” toward catalysis. However, from the observed band narrowing of the C4-D stretch band, it is clear that interactions at the active site in all three proteins greatly reduced the conformational flexibility of either the reduced or oxidized ring as the cofactor moves from solution to the binary complex or ternary complex, guiding the ring structure from the ensemble of structures accessible in solution toward a selected set. Moreover, as NAD(P)H binds to LDH or DHFR forming binary as well as ternary Michaelis mimic complexes, the pro-R hydrogen is brought to a pseudoaxial orientation, which is thought to be the proper geometry for the transition state of hydride transfer. Hence, ground state structural distortions imposed on the cofactor appear to populate preferentially the correct ring geometry for enzymic activity. Surprisingly, the mimics of their Michaelis complexes also contain a substantial second, presumably unproductive, population of the bound cofactor whereby the pro-S hydrogen is pseudoaxial. Unexpectedly, the geometry of NADH bound to G3PDH is nearly planar with the pro-R hydrogen slightly pseudoaxial. This would seem to be a poorly bound cofactor for catalysis although it may well be true that the transition state geometry for G3PDH is not that of LDH. How the results bear on various proposals concerning ground-state regulation of reactivity is discussed.