Electron Transfer Pathways and Dynamics in Drosophila Cryptochrome - the Role of Protein Electrostatics

. Dissipative quantum dynamics simulations reveal a branching of charge separation dynamics in Drosophila Cryptochrome due to subtle balanced energetics within the enzyme. In silico mutations of charged amino acids provide control over charge transfer directionality.


Photoreception mechanism of Drosophila cryptochrome
Cryptochromes are highly conserved flavoproteins consisting of an N-terminal photolyase homology region (PHR) that binds a flavin adenine dinucleotide (FAD) cofactor, and a C-terminal α-helical domain with a variable C-terminal tail (CTT) that shows high diversity in sequence among organisms (Fig. 1a). Drosophila cryptochromes (dCRY) function as blue-light photoreceptors by synchronizing the circadian clock to the external stimuli of incident sunlight via conformational changes located in the CTT.
The microscopic details of the photoreception mechanism leading to CTT conformational changes upon light absorption of the FAD cofactor are unknown. It is generally assumed that photoreduction of FAD proceeds via a highly conserved triad of tryptophanes (dCRY Trp-triad: W420-W397-W342, Fig. 1b) where W420 acts as primary electron donor upon photoexcitation, in analogy to tryptophan-triad dependent photoactivation in photolyases. Such Trp-triad functionality has been recently questioned due to observations that dCRY promotes enzymatic activity either in the presence of W → F mutations that abolish in vitro Trp-triad photoreduction [1].
Focussing on the elusive microscopic events initiating signal transduction upon photoexcitation of fully oxidized flavin FAD Ox as plausible resting state [2], state-of-the-art simulations are presented that consider charge separation via the Trp-triad and further take into account intra-FAD charge transfer involving the isoalloxazine (ISO) and adenine (ADE) moieties of the FAD co-factor, as well as charge transfer states involving spatially close W314 and W422 residues (cf. Fig. 1b

MACGIC-QUAPI simulations of electron transfer dynamics
Microscopically derived Hamiltonians were obtained by evaluating QM/MM excitations energies along a ~50 ns segment of a 160 ns dCRY MD trajectory (Fig. 1c-e), inherently accounting for enzyme thermal fluctuations of charge separated states. Additional trajectories conducted in respective charge separated states (e.g. ISO − W420 + , ISO − W314 + etc.) show convergence to Gaussian statistics and allow for the reliable construction of free energy differences G and reorganization energies of participating charge separated states within the enzyme environment ( Fig. 2a-b) [5]. Non-Markovian real-time evolution of charge separation dynamics is simulated with the recently introduced MACGIC-QUAPI method [6] that relies on a non-linear intermediate representation of the influence functional and provides convergence to exact HEOM benchmark results for arbitrary system-bath coupling strength. We find that due to the comparable energetics of charge separated states an assignment to a unique charge separation pathway, either the TRP-triad pathway (ISO − W420 + -ISO − W397 + ) or the alternative pathway towards the CTT (ISO − ADE + -ISO − W314 + ), is precluded. The wild-type enzyme charge separation dynamics confirms the picture that both pathways contribute to the deactivation of the initially excited ππ * state of ISO (Fig. 2c, top) which is depopulated on the tens-of-picosecond timescale, leading to a parallel population of ISO − ADE + and ISO − W420 + charge separated states, followed by depopulation into ISO − W314 + and ISO − W397 + states.
The effect of enzyme environment is investigated by in silico mutations of selected amino acid residues. We find that a small set of charged residues strongly affects the energetics of charge separated states, which in part have destabilizing impact on the TRPtriad pathway and stabilizing effect on the pathway involving TRP314. Accordingly, the dynamics of in silico mutated dCRY (Fig 2c, bottom)