1-Deoxy-d-xylulose-5-phosphate reductoisomerase (DXR),
which catalyzes the first committed step in the 2-C-methyl-d-erythritol 4-phosphate pathway of isoprenoid biosynthesis
used by Mycobacterium tuberculosis and other infectious
microorganisms, is absent in humans and therefore an attractive drug
target. Fosmidomycin is a nanomolar inhibitor of DXR, but despite
great efforts, few analogues with comparable potency have been developed.
DXR contains a strictly conserved residue, Trp203, within a flexible
loop that closes over and interacts with the bound inhibitor. We report
that while mutation to Ala or Gly abolishes activity, mutation to
Phe and Tyr only modestly impacts kcat and Km. Moreover, pre-steady-state kinetics
and primary deuterium kinetic isotope effects indicate that while
turnover is largely limited by product release for the wild-type enzyme,
chemistry is significantly more rate-limiting for W203F and W203Y.
Surprisingly, these mutants are more sensitive to inhibition by fosmidomycin,
resulting in Km/Ki ratios up to 19-fold higher than that of wild-type DXR. In
agreement, isothermal titration calorimetry revealed that fosmidomycin
binds up to 11-fold more tightly to these mutants. Most strikingly,
mutation strongly tips the entropy–enthalpy balance of total
binding energy from 50% to 75% and 91% enthalpy in W203F and W203Y,
respectively. X-ray crystal structures suggest that these enthalpy
differences may be linked to differences in hydrogen bond interactions
involving a water network connecting fosmidomycin’s phosphonate
group to the protein. These results confirm the importance of the
flexible loop, in particular Trp203, in ligand binding and suggest
that improved inhibitor affinity may be obtained against the wild-type
protein by introducing interactions with this loop and/or the surrounding
structured water network.