Vibrational simulation of the biological systems---normal modes analysis and stochastic vibrational self-consistent field calculation

The present research focuses on the vibrational simulation of the biological systems, including the harmonic-level normal modes analysis and the anharmonic-level vibrational self-consistent field (VSCF) calculation. Normal modes analysis was performed on a large biomolecule---ricin-A-chain (RTA)---in both apo- (no substrate) and holo- (adenosine monophosphate (AMP)-bound) state. It revealed that the shearing motion was shared by both apo- and holo-RTAs, whereas the breathing motion, as well as the upward hinge and the a-G bending characteristic motions, was dampened by substrate binding. We hypothesize that the breathing, a-G bending and upward hinging motions play an important role in substrate binding as these motions facilitate the entry of the substrate and provide space for the substrate realignment that is necessary for the depurination. The VSCF calculations, which were typically restricted to small systems, were performed in the present research on a moderate biomolecule---the VA-class dipeptide nanotubes; to the best of our knowledge, this is the first time condensed-phase VSCF calculation of biomolecules that has been preformed. By comparing the calculated Terahertz (THz) spectra against the experimental spectra, we found that in general the VSCF level calculations deomonstrated significant improvement over the harmonic calculations, which was mostly reflected in the overall blueshifts of the VSCF frequencies from the harmonic values. These blueshifts were accounted for the coupling between two similar sidechain-squeezing modes and the subsequent stiffening of the effective potential. Finally, we developed a stochastic-VSCF methodology to enable the anharmonic and coupling-incorporated vibrational simulation for large biomolecules. This method evaluates the inter-mode couplings and performs the vibrational calculations in a stochastic fashion. During this stochastic process, the insignificantly mode pairs are rapidly found and removed from the system without exhaustively exploring the entire PES, and therefore the computational expense is remarkably saved. A validation test was performed for the stochastic-VSCF method on the VA-class dipeptide nanotubes. It was fond that this method remarkably saved the computational expense while yielding the THz spectra very close to the exhaustive VSCF calculations. The stochastic-VSCF calculation was finally performed on a large biomolecule---bacteriorhodopsin.