| Resumo: |
Enzymes are the most powerful catalysts created by natural evolution. They are also the target of ~50% of all drugs in the market. Despite their enormous importance, the molecular source for their catalytic power is still unclear, and their modus operandi at the atomic level is still poorly understood. This project is devoted to understand how enzyme flexibility controls the reaction mechanism and reaction rate. We have been focusing in a fundamental aspect of enzymatic catalysis - the influence of enzyme conformational diversity on the reaction mechanism and reaction rate. Single molecule experiments confirmed the uneven landscape of catalysis, with different reaction rates at different times for the same enzyme molecule [1], the macroscopic rate constant being an average of distinct instantaneous rate constants over time. However, experiments address microscopic amounts of matter over macroscopic times. The only way to go down to the study of single molecules in microscopic times is through first-principles- physics computer simulations. The project proposes to identify local and collective structural motifs along enzyme dynamics that induce a given reaction mechanism and a high reaction rate, which ultimately, renders enzymes as the supreme catalysts. The computer simulation of this phenomenon is so much demanding that current state-of-the-art studies either use a high-level theoretical level and a reduced exploration of the enzyme conformational diversity or a more extended exploration of the conformational space at the cost of sacrificing the accuracy of the theoretical level[2]. The conciliation between Hamiltonian accuracy and phase space exhaustiveness, needed to get a firm answer to the problem, can only be obtained with top HPC resources. We want to make a proof-of-concept that HPC can solve the problem of understanding the intricacies of enzyme catalysis. To do so, we focus on a specific enzyme (HIV-1 protease, PR) which we have been studying recently  |
Resumo
Enzymes are the most powerful catalysts created by natural evolution. They are also the target of ~50% of all drugs in the market. Despite their enormous importance, the molecular source for their catalytic power is still unclear, and their modus operandi at the atomic level is still poorly understood. This project is devoted to understand how enzyme flexibility controls the reaction mechanism and reaction rate. We have been focusing in a fundamental aspect of enzymatic catalysis - the influence of enzyme conformational diversity on the reaction mechanism and reaction rate. Single molecule experiments confirmed the uneven landscape of catalysis, with different reaction rates at different times for the same enzyme molecule [1], the macroscopic rate constant being an average of distinct instantaneous rate constants over time. However, experiments address microscopic amounts of matter over macroscopic times. The only way to go down to the study of single molecules in microscopic times is through first-principles- physics computer simulations. The project proposes to identify local and collective structural motifs along enzyme dynamics that induce a given reaction mechanism and a high reaction rate, which ultimately, renders enzymes as the supreme catalysts. The computer simulation of this phenomenon is so much demanding that current state-of-the-art studies either use a high-level theoretical level and a reduced exploration of the enzyme conformational diversity or a more extended exploration of the conformational space at the cost of sacrificing the accuracy of the theoretical level[2]. The conciliation between Hamiltonian accuracy and phase space exhaustiveness, needed to get a firm answer to the problem, can only be obtained with top HPC resources. We want to make a proof-of-concept that HPC can solve the problem of understanding the intricacies of enzyme catalysis. To do so, we focus on a specific enzyme (HIV-1 protease, PR) which we have been studying recently [3]. As PR is essential for the life-cycle of HIV, its detailed understanding will not only materialize the proposed proof-of-concept but will also help in the development of new anti-HIV drugs. The reaction mechanism of HIV-1 protease will be addressed through QM/MM MD simulations with different starting conformations of the enzyme: substrate complex. The study will comprehend an accurate characterization of the minimum free energy path (MFEP) for the cleavage of the peptide bond by HIV-1 protease with the string method approach. Then, and an efficient sampling of the resulting MFEP through umbrella sampling QM/MM MD simulations will be conducted along the converged string. Ultimately, through extensive sampling, the underlying features of HIV-1 protease's catalysis should be determined, shedding light on the nature of the catalysis by the enzyme. This project will provide also important benchmarking data to prepare a future application for an INCITE application.
This grant has provided 10 million CPU core-hours in Cray Titan and IBM Summit (the fastest supercomputer in the world), funded by the Oak Rigde National Laboratory, U.S.A. The approximate cost of these CPU hours is 1 million ¤. |