Modifying a Mini Protein with Two Conformational States to Instead Adopt Only One Conformation

Oliver Cho
Oliver Cho

Oliver Cho is a rising Senior (’22). He graduated from the Collegiate School in New York, NY. His interests are rather sporadic, but he has particularly enjoyed running, drumming, reading fantasy novels, and annoying his current roommates with bad jokes he thinks of when he is trying to fall asleep. He is a Chemistry major (A-Track). He would like to perhaps conduct research in agriculture or meat-alternative foods.


Our project is focused on the computationally-designed mini protein EHEE_rd2_0005. While the structure of this protein was previously determined in molecular dynamics simulations, structural knowledge derived from simulations are not as precise as real-world analysis. Members of our lab have previously attempted to obtain more accurate data with NMR spectroscopy but were unable to assign several peaks because of extensive peak broadening. Because this protein’s TRP-56 residue switches between two equally-favored 𝞆 dihedral angles, we believe that it is responsible for an intermediate-rate dynamic exchange that is the source of the NMR peak broadening. The motion of the tryptophan sidechain interferes heavily with NMR analysis because of its aromatic sidechain’s strong electromagnetic ring current effects on surrounding nuclei. Our goal was to find a single-point mutation that would not drastically affect the secondary structure of our protein. In this attempt to modify a protein to adopt only a single conformational state, we chose and tested multiple single-point mutations in silico. By mutating the wild type protein and measuring the relative free energy values of the two conformational states through “alchemical” fast-morph simulations, we can predict which mutations are most likely to slow the rate of TRP-35’s conformational isomerism. The bulk of our work was done with molecular dynamics simulations in Gromacs. While we selected many mutations ourselves, we also utilized rigid-backbone Rosetta simulations to help us decide which mutations to run. Rosetta is very imprecise in comparison to Gromacs, but it is much faster. We were able to perform an exhaustive scan of all possible single-point mutations in our 40-residue protein. Although Rosetta was inaccurate in identifying specific mutations that would meet our goal, its data revealed particular positions in the protein that would greatly influence TRP-35 to favor a single conformation. With this hybrid approach, we were able to better identify which mutations to make and analyze with NMR in the wet lab.

Your Name (Your Department)

Live Poster Session:
Thursday, July 29th 3:03pm EDT