Tracking the primary photoconversion events in rhodopsins by ultrafast optical spectroscopy

Abstract

Opsins are a broad class of photoactive proteins, found in all classes of living beings from bacteria to higher animals, which work either as light-driven ion pumps or as visual pigments. The photoactive function in opsins is triggered by the ultrafast isomerization of the retinal chromophore around a specific carbon double bond, leading to a highly distorted, spectrally red-shifted photoproduct. Understanding, by either experimental or computational methods, the time course of this photoisomerization process is of utmost importance, both for its biological significance and because opsin proteins are the blueprint for molecular photoswitches. This paper focuses on the ultrafast 11-cis to all-trans isomerization in visual rhodopsins, and has a twofold goal: (i) to review the most recent experimental and computational efforts aimed at exposing the very early phases of photoconversion; and (ii) discuss future advanced experiments and calculations that will allow an even deeper understanding of the process. We present high time resolution pump-probe data, enabling us to follow the wavepacket motion through the conical intersection connecting excited and ground states, as well as femtosecond stimulated Raman scattering data allowing us to track the subsequent structural evolution until the first stable all-trans photoproduct is reached. We conclude by introducing computational results for two-dimensional electronic spectroscopy, which has the potential to provide even greater detail on the evolution of the electronic structure of retinal during the photoisomerization process.

QM/MM Photodynamics of Retinal in the Channelrhodopsin Chimera C1C2 with OM3/MRCI

“…To generate initial geometries and velocities for MD trajectories we performed a zero‐point energy (ZPE) sampling on the retinal chromophore using Gaussian09, thus simulating a vibrational distribution at 0 K. The OM3‐optimized ground state geometry including retinal and the lysine residue was used to obtain the required displacements via vibrational frequency calculations. High‐frequency C−H and N−H modes were excluded from the sampling procedure as they may lead to total energy fluctuations and do not influence the reaction …”

“…High-frequency CÀH and NÀH modes were excluded from the sampling procedure as they may lead to total energy fluctuations and do not influence the reaction. [38][39][40][41] Excited-state dynamics. Excited-state dynamics calculations were performed using a velocity-Verlet integrator scheme with an adapted time step [42] starting at 0.1 fs.…”

“…State populations were computed by propagating the time-dependent wavefunction in each individual trajectory with a step size of 1/100 of the MD step. Tully's fewest-switches surface-hopping algorithm [43,44] with decoherence correction [45,46] was applied to provide hopping probabilities. State populations and state occupations resulting from trajectory surface hopping were collected each full fs and averaged over the whole ensemble.…”

Recent Advances and Perspectives on Nonadiabatic Mixed Quantum–Classical Dynamics

“…Alternatively, the ensemble can be built from NA-MQC trajectories. 415 For a steady-state spectrum simulation, the initial transient dynamics should be discarded, and only points obtained after equilibration in the minimum of the excited states should be considered. 67…”

“…The SOS result has been recently explored by Garavelli's group to simulate 2DES based on QM/MM TSH dynamics. 306,415,425 Although it has delivered promising results, the drawback of this approach is the enormous number of electronic states required in the simulation;…”