Cryo‐kinetics Reveal Dynamic Effects on the Chemistry of Human Dihydrofolate Reductase

Abstract Effects of isotopic substitution on the rate constants of human dihydrofolate reductase (HsDHFR), an important target for anti‐cancer drugs, have not previously been characterized due to its complex fast kinetics. Here, we report the results of cryo‐measurements of the kinetics of the HsDHFR catalyzed reaction and the effects of protein motion on catalysis. Isotopic enzyme labeling revealed an enzyme KIE (k H LE/k H HE) close to unity above 0 °C; however, the enzyme KIE was increased to 1.72±0.15 at −20 °C, indicating that the coupling of protein motions to the chemical step is minimized under optimal conditions but enhanced at non‐physiological temperatures. The presented cryogenic approach provides an opportunity to probe the kinetics of mammalian DHFRs, thereby laying the foundation for characterizing their transition state structure.


Protein Production and Purification
The gene coding for HsDHFR (UniProt Accession ID: P00374) was expressed in E. coli BL21 (DE3) cells. Cells were grown in M9 media and supplemented with 50 µg/mL kanamycin, vitamins and either natural abundance or isotopically labeled ammonium chloride and glucose (that is, 14 NH4Cl/ 15 NH4Cl and 12 C-glucose/ 13 C-glucose) for light and heavy enzyme production, respectively. Gene expression was induced at O.D. 600 nm of 0.9 by the addition of 1 mM IPTG. The culture was further grown at 25 °C overnight. Cells were harvested by centrifugation. HsDHFR was purified using a modified protocol involving Q-Sepharose and size exclusion chromatography. The purification protocol, to be published separately by the authors, neither employ folic acid to enhance the enzyme's stability nor require a refolding step to remove protein and ligand impurities as recommended in a previous protocol. 1 Protein concentration was determined using a bicinchoninic acid assay. 2 Enzyme purity was assessed by LC/MS and circular dichroism employed to ascertain the secondary structure of the enzymes in different buffer systems.

Pre-Steady State kinetic measurements
TgK Scientific stopped-flow equipment was modified for cryogenic measurements by diverting the water bath cycler from the sample handling unit to the T-pod to prevent cryo-damage of the unit. A constant flow of nitrogen ensured the optics are free from condensation. Injection volume was reduced to 80 μL and a 60 sec delay between each acquirement was used to ensure that the reacting solutions are equilibrated to the temperature within the cell block. Hydride transfer reaction by HsDHFR is fast and often complete within 0.1 sec. 20 µM enzyme was pre-incubated with 10 µM NADPH/D and the reaction initiated by the addition of 200 µM DHF. Following excitation at 297 nm, emission was measured using an output filter with a 400 nm cut-off.
Hydride transfer rates were determined by fitting the relaxation of the fluorescence energy transfer from the enzyme to the reduced cofactor to a first-order exponential (Section 1.3.1). Final assay conditions were 10 µM HsDHFR, 5 µM NADPH/D and 100 µM DHF. The buffer was 50 mM potassium phosphate, 20 mM boric acid (pH, 8.5), 150 mM KCl, 5 mM DTT in the presence of 30% methanol. Pre-steady state measurements were also carried out in the absence of methanol between 0 and +20 °C using NADPD as the cofactor. NADPH/D exhibit biphasic binding to HsDHFR, especially at non-saturating concentrations. 5,6 Hence, a change in time course during fluorescence relaxation that became more prominent at lower temperatures was observed after 100 msec.

Hydride transfer rate constant measured under pre-steady-state conditions
The relaxation of the fluorescence energy transfer from the enzyme to the cofactor was fitted to a single exponential with a slope function. The definition (Kinetic Studio 4.0) of the model is A exp (-R*X) + (M*X) + C, with R representing the rate constant of the best fit of the curve.   (Table S5).
However, minor structural effects were observed when the CD spectra were analyzed (Table S6), which might be due to the perturbation of water-molecules that had been proposed to stabilize the structure of HsDHFR. [20] In addition, the steady-state (kcat) at pH 7.0 (Table S5) and pre-steady state rate constants above 10 °C were reduced due to the addition of methanol (data not shown).   Figure S4. CD spectra at 20 ºC for HsDHFR (10 µM     Ea (kcal·mol -1 ) 7.9 ± 0.2 11.6 ± 0.7 Figure S6. Temperature-dependence of enzyme KIE for the transfer of hydride ( ) and deuteride in buffers with ( ) and without 30% methanol ( ).