UC Berkeley

Hydrogen atom transfer (HAT) is a salient feature of many enzymatic C-H cleavage mechanisms. In systems where kinetic isolation of HAT is achieved, selective labeling of substrate with hydrogen isotopes, such as deuterium, enables the determination of intrinsic kinetic isotope effects (KIEs). While the magnitude of the KIE is itself informative, ultimately the size of the temperature dependence of the KIE, Δ E_a = E_a(D) - E_a(H), serves as a critical, and often misinterpreted (or even ignored) descriptor of the reaction coordinate. As will be highlighted in this Account, Δ E_a is one of the most robust parameters to emerge from studies of enzyme catalyzed hydrogen transfer. Kinetic parameters for C-H reactions via HAT can appear consistent with either classical "over-the-barrier" or "Bell-like tunneling correction" models. However, neither of these models is able to explain the observation of near-zero Δ E_a values with many native enzymes that increase upon extrinsic or intrinsic perturbations to function. Instead, a full tunneling model has been developed that can account for the aggregate trends in the temperature dependence of the KIE. This model is reminiscent of Marcus-like theory for electron tunneling, with the additional incorporation of an H atom donor-acceptor distance (DAD) sampling term for effective wave function overlap; the role of the latter term is manifested in the experimentally determined Δ E_a. Three enzyme systems from this laboratory that illustrate different aspects of HAT are presented: taurine dioxygenase, the dual copper β-monooxygenases, and soybean lipoxygenase (SLO). The latter provides a particularly compelling system for understanding the properties of hydrogen tunneling, showing systematic increases in Δ E_a upon reduction in the size of hydrophobic residues both proximal and distal from the active site iron cofactor. Of note, recent ENDOR-based studies of enzyme-substrate complexes with SLO indicate an increase in DAD for mutants with increased Δ E_a, observations that are inconsistent with "Bell-like correction" models. Overall, the surmounting kinetic and biophysical evidence corroborates a multidimensional approach for understanding HAT, offering a robust mechanistic explanation for the magnitude and trends of the KIE and Δ E_a. Recent DFT and QM/MM computations on SLO are compared to the developed nonadiabatic analytical constructs, providing considerable insight into ground state structures and reactivity. However, QM/MM is unable to readily reproduce the small Δ E_a values characteristic of native enzymes. Future theoretical developments to capture these experimental observations may necessitate a parsing of protein motions for local, substrate deuteration-sensitive modes from isotope-insensitive modes within the larger conformational landscape, in the process providing deeper understanding of how native enzymes have evolved to transiently optimize their active site configurations.