The nucleotide 2'-deoxythymidine 5'-monophosphate (thymidylate, dTMP) is phosphorylated twice to become a substrate for DNA polymerases, which copy a cell’s genetic information in advance of cell division. The main route to dTMP is mediated by the enzyme thymidylate synthase (TSase) and goes through 2'-deoxyuridine 5'-monophosphate (dUMP); dUMP’s heterocyclic aromatic pyrimidine ring loses a proton from its C5 position and gains a methylene and a hydride from the other reactant, methylene tetrahydrofolate (MTHF). In general, intricate knowledge of an enzyme’s mechanism can yield insight that leads to the development of precision-targeted inhibitors tailored exactly to thymidylate synthase. In fact, even more careful targeting could be achievable: Although E. coli TSase has served as a model system, investigators have increasingly been directing their lines of inquiry toward human TSase.
A general enzymatic catalytic cascade is complex, comprising substrate binding, the chemical steps and product release; typically, the product release step is rate-limiting. TSase, however, is partially rate-limited by the chemistry portion of the process. The enzymatic mechanism has been considered for decades, yet recently has undergone a reassessment. After substrate binding – for which there is strong evidence for preference to dUMP as the first ligand in the wild-type E. coli enzyme – the important events are methylene transfer from MTHF to dUMP, proton abstraction and hydride transfer. The last of these – hydride transfer – is irreversible and rate-limiting (to a large degree without Mg2+, and to a small but noticeable degree with Mg2+). The studies described here are aimed at three therapeutically relevant questions: (a) determining the extent of negative charge accumulation at the O4 position of the hydride transfer acceptor; (b) expanding knowledge of the differential properties of E. coli and human TSase; and (c) gaining insight into the molecular origin of the drug resistance seen in a clinically relevant human TSase mutant.
The properties touched on in this work include steady-state kinetics; inhibition constants toward 5-fluoro dUMP, substrate binding sequence and the temperature dependency of intrinsic hydride transfer kinetic isotope effects (KIEs). Intrinsic KIEs are a specialized measurement that permits the investigator to examine a particular hydrogen transfer step in isolation; it is achieved by labeling the bond to hydrogen broken in the reaction with protium (1H, also written as H), deuterium (2H, also written as D) or tritium (3H, also written as T). The latter is radioactive. The reaction is conducted with a mixture of two hydrogen isotopes at a time, and the extent to which the heavier isotope is disfavored against reaction is assessed; this covers multiple steps. Heavier isotopes directly participating in a chemical step react slower both because of zero-point vibrational energies if a semi-classical view is taken and because of the mass-dependence of tunneling probabilities if a quantum-mechanical view is taken. Each of the two-way isotopic comparisons mentioned above furnishes an observed KIE for that competition between two isotopes. Mathematical combination of two isotopic comparisons cancels out the effect of isotopically insensitive steps and provides rich insight into the hydride transfer alone. The ultimate result is the ratio of rate constants for the isotopologues; this ratio’s magnitude and variation with temperature report on the compactness of the active site and its resistance to thermal fluctuation, respectively.
Our results reveal a possible role for E. coli asparagine 177 (N177) in the hydride transfer transition state (TS) stabilization, as revealed by its disruption in the aspartate mutant, N177D. This disruption was found to be alleviated to a high extent when the substrate was changed to dCMP, consistent with the N177 stabilizing partial negative charge at the TS for hydride transfer. This has drug design implications. Our work on human TSase underscores slightly weaker substrate binding preference, insensitivity to Mg2+ and mild alteration of hydride transfer TS when compared with E. coli TSase. Finally, analysis of the Y33H mutant of human TSase – the affected residue being remote from the active site – indicated the drug resistance was because of a higher inhibition constant for 5F-dUMP and that the hydride transfer step is disrupted, with a wider variation among donor-acceptor distances (between the two carbons involved in the hydride transfer at the TS for that step). Other researchers’ crystallographic evidence reveals greater positional uncertainty for a set of active-site side chains in the E. coli equivalent mutant. In totality, the data available implicate enzyme motions as relevant to drug binding and to catalysis for human TSase.
In summary, the research described herein enriches the understanding of several aspects of the behavior of multiple TSase variants – the overall performance as seen via steady-state kinetics; the pattern of substrate binding as seen with observed KIEs for the proton abstraction step; and the efficiency of active site preparation for hydride transfer as evidenced in the temperature dependency of intrinsic hydride transfer KIEs.