QM/MM Trajectory Surface Hopping Approach to Photoisomerization of Rhodopsin and Isorhodopsin: The Origin of Faster and More Efficient Isomerization for Rhodopsin

The photoinduced cis-trans isomerization dynamics of rhodopsin and isorhodopsin are studied using a newly developed hybrid QM/MM trajectory surface hopping MD scheme based on the Zhu-Nakamura theory for nonadiabatic transitions. Rhodopsin and isorhodopsin have 11-cis and 9-cis forms of retinal as chromophore and the two proteins are isomerized to bathorhodopsin enclosing the all-trans form. The simulation reproduced faster and more efficient isomerization in rhodopsin than in isorhodopsin. In the excited state, rhodopsin shows a straightforward dynamics, whereas isorhodopsin dynamics is rather complicated and in a back-and-forth manner. The latter complicated dynamics would be mainly due to a narrow space near the active dihedral angle = C8-C9 = C10-C11 =(phi(9)) created by Thr 118 and Tyr 268 in opsin. Rhodopsin gives bathorhodopsin only while isorhodopsin yields a byproduct. The rigorous selectivity in rhodopsin would be another reason why rhodopsin is selected biologically. Comparison with our previous opsin-free investigations reveals that opsin tends to confine the twist of the active dihedral to only one direction and funnels transitions into the vicinity of minimum energy conical intersections (MECI). The twist-confinement totally blocks simultaneous twisting of phi(9) and phi(11) (= C10-C11 = C12-C13 =) and enhances the quantum yields. The opposite rotation of phi(9) and phi(11) (wring-a-wet-towel motion) takes place upon photoexcitation, which also does without opsin. The wring-a-wet-towel motion is dynamically enhanced in comparison with the one expected from locations of the MECI. The present simulation reveals that the Weiss-Warshel model for cis-trans photoisomerization is not applicable for rhodopsin because the branching ratio after transition is crucial.