Activation Heat Capacities in Pyridoxal Phosphate Enzymes

Abstract

Much attention has been given to the temperature dependence of enzyme catalyzed reaction rates in the last ∼100 years. Over the last couple of decades, it has become apparent that activity decreases before enzymes lose their structural integrity. Two viable models have been proposed to account for this behavior. In one, an inactive conformation(s) becomes more populated at higher temperatures. In the other, a difference in heat capacity between ground and transition state conformational ensembles is invoked. Here, the temperature dependence of the activity of 16 different combinations of pyridoxal phosphate enzymes and substrates has been measured. All show non-Arrhenius activity vs temperature profiles. These are generally best accommodated by the second model where the ΔG ‡ term in the Eyring equation is expanded in terms of ΔH ‡ , ΔS ‡ , and ΔC p ‡ (macromolecular rate theory (MMRT) equation). The values of ΔC p ‡ extracted by curve fitting are negative and in the range of −1 to −2 kJ/(mol K). Several data sets fit better to a modified MMRT equation that additionally includes the temperature dependence of ΔC p ‡ (dΔC p ‡ /dT). The values of dΔC p ‡ /dT from curve fitting are also negative. The additional dΔC p ‡ /dT term has major effects on temperature profiles at the low and high ends of the biological temperature range. Molecular dynamics simulations using the AMBER-FB15 force field were performed on a set of six model proteins and two of the enzymes studied experimentally. Convex energy vs temperature relationships are observed, which require a positive dC p /dT term to fit the data well. The simulations for ground and transition state structures for aspartate aminotransferase and D-serine deaminase allow calculation of ΔC p ‡ and dΔC p ‡ / dT for comparison to experiments. Negative values for ΔC p ‡ and dΔC p ‡ /dT are obtained for both enzymes. The simulations generally reproduce well the experimental values of ΔC p ‡ but the dΔC p ‡/dT values are overestimated. The results support a model in which a large, looser ensemble of ground state-compatible conformations is in equilibrium with a very small, tighter transition stateoptimized ensemble that allows bond making and breaking to occur with a low energetic barrier.