(169di) Using Enhanced Sampling Methods to Elucidate the Mechanism of Noncanonical Redox Cofactor Dependent Engineered Enzymes

Background: Enzymes are useful proteins that can also produce compounds renewably and under ambient conditions while generating low waste. However, they often require natural cofactors or small, non-protein molecules for optimal activity. These natural cofactors are prohibitively expensive and chemically unstable to be used in specialty chemicals manufacture [1]. Hence, researchers have been engineering enzymes that can utilize smaller natural cofactor mimetics or noncanonical cofactors that are more efficient, easier to synthesize, and are low-cost alternatives.[2]

Dr. Han Li (UC Irvine) recently engineered the enzyme phosphite dehydrogenase (PTDH) to utilize a noncanonical cofactor 1-benzylnicotinamide (BNA+) in biotransformation processes [3]. Dr. Li also produced PTDH mutants that exhibit higher activity compared to the wild type engineered PTDH. However, the reason for that is unknown, and it has been hypothesized that these mutations are space filling that pack against each other where the natural cofactor would have bound, which leads the PTDH mutants to mimic the binding of a natural cofactor and become more active closed proteins. To harness the PTDH mutants fully at industrial scale, we must understand the molecular details and mechanisms of how the mutations lead the enzymes to become more active closed proteins.

Aim: 1) MD Simulations of two PTDH mutants and the native PTDH to investigate the higher activity caused by additional mutations in LY-13-Mut than LY-13.
2) MD Simulations of two PTDH mutants and the native PTDH transitioning from the initial open state to the bound closed state to uncover the closing mechanisms and rates.

Methodology: We outlined the simulations of the two highly active PTDH mutants LY-13 and LY-13-Mut and the wild type PTDH 4EBF, which will provide detailed views into the functional mechanisms of PTDH. Specifically, we used 1 μs long Gaussian accelerated molecular dynamics (GaMD) that adds a harmonic boost potential to the system so that the energy barriers between states are reduced and the energy landscape can be sampled efficiently [4].

MD simulations are run using 10^−15 s time steps, due to being limited by the fastest motions in the system (e.g. vibration of bonds), while protein dynamics range from 10^−6 s to longer. Thus there is a major timescale barrier between MD simulations and protein dynamics. To bridge the timescale barrier, “enhanced sampling methods” for MD simulations were used. Specifically, we used Gaussian accelerated molecular dynamics (GaMD) that adds a harmonic boost potential to the system so that the energy barriers between states are reduced and the energy landscape can be sampled efficiently [4].

Although GaMD is useful for obtaining thermodynamic properties like free energy landscapes, other approaches need to be used to obtain kinetic properties like rate constants and continuous pathways since GaMD alters the real kinetics of the system with its biasing potentials. Hence, we will use the weighted ensemble method (WE), a path sampling method [5] , to fully characterize the closing mechanisms and rates for the PTDH systems starting from their initial open states. Dr. Ahn has successfully applied WE to investigate the activation mechanism of the SARS-CoV-2 spike protein [6] and is confident to introduce and apply the method for systems important for chemical engineering.

Expected Results: We expect to be able to gain atomistic insight into each PTDH system’s structure and dynamics. Specifically, we expect to identify how the two mutations of LY-13-Mut contribute to its higher activity than that of LY-13. We expect that LY-13-Mut will tend more towards the active closed state and/or keep BNA+ bound longer than LY-13, as experiments predict. Moreover, we expect to see the LY-13-Mut show a similar free energy landscape to that of the wild type PTDH, which would indicate that the mutations reshape the conformational landscape of LY-13-Mut to resemble that of the wild type PTDH, similar to the results from Dr. Li’s previous work on the engineered NADH oxidase that can utilize noncanonical cofactors [5].

Implications: PTDHs are highly efficient catalysts in recycling noncanonical redox cofactors due to their superior TTN, ability to maintain turnover at low BNA+ concentrations, and capability to recycle other simpler cofactor biomimetics as established by Dr. Li [3]. However, elucidating the mechanism of redox cofactor dependent enzymes will further our understanding of engineering growth-based selected enzymes for biotechnology applications. This is especially crucial in engineering biomimetics for industrially important reactions that do not exist in nature. In this pursuit, enhanced sampling methods with MD simulations are paramount in helping develop a unified design toolkit for engineered enzymes.

REFERENCES

[1] T. Knaus, C. E. Paul, and et. al, Journal of the American Chemical Society 138, 1033 (2016).
[2] E. King, S. Maxel, and H. Li, (2020), https://doi.org/10.1016/j.copbio.2020.08.005.
[3] L. Zhang, E. King, W. B. Black, C. M. Heckmann, A. Wolder, Y. Cui, F. Nicklen, J. B. Siegel, R. Luo, C. E. Paul, and
H. Li, Nature Communications 13 (2022), 10.1038/s41467-022-32727-w.
[4] Y. Miao, V. A. Feher, and J. A. McCammon, Journal of Chemical Theory and Computation 11, 3584 (2015).
[5] G. Huber and S. Kim, Biophysics Journal (1996), 10.1016/S0006-3495(96)79552-8.
[6] T. Sztain and S.-H. Ahn, (2021), https://doi.org/10.1038/s41557-021-00758-3.
[7] E. King, S. Maxel, Y. Zhang, K. C. Kenney, Y. Cui, E. Luu, J. B. Siegel, G. A. Weiss, R. Luo, and H. Li, Nature
Communications 13 (2022), 10.1038/s41467-022-35021-x.