Electron transfer is a fundamental process in numerous biological systems, playing a crucial role in photosynthesis, cellular respiration, and DNA repair. Photolyases and cryptochromes, a family of flavoproteins, exemplify the importance of electron transfer in biological function. These proteins utilize light energy to repair DNA damage (photolyases) or regulate biological processes like circadian rhythms (cryptochromes). Despite their functional diversity, photolyases and cryptochromes share a conserved electron transfer chain involving a flavin adenine dinucleotide (FAD) cofactor and a series of tryptophan residues. Understanding the mechanism of this electron transfer is essential to comprehend the function of these vital proteins. Previous studies have used spectroscopy and simulations to investigate the kinetics of electron transfer in photolyases and cryptochromes, yielding estimates of free-energy gains and reorganization energies for each transfer step. However, these studies often struggle to resolve the rates of transfer between spectroscopically similar tryptophan residues and the validity of Marcus theory in these systems remains unclear. A key limitation of past approaches has been the inability to directly observe the protein structural changes that accompany electron transfer. This study aims to address this gap by employing femtosecond time-resolved serial crystallography (SX) to capture the structural dynamics of electron transfer in *Drosophila melanogaster* (6-4) photolyase.

The electron transfer mechanism in photolyases and cryptochromes has been extensively studied using various techniques, including spectroscopy and molecular dynamics simulations. Spectroscopic studies have established the sequence of electron transfer events, demonstrating that the photoexcited FAD extracts an electron from a nearby tryptophan within picoseconds. This initiates a cascade of electron transfer reactions along a chain of tryptophan residues, ultimately establishing a long-range radical pair. While these studies have provided valuable insights into the kinetics of electron transfer, they have not directly visualized the associated protein structural changes. Molecular dynamics simulations have also been employed to study the electron transfer process in photolyases, providing information on the free-energy landscapes and reorganization energies. However, these simulations are often challenged by limitations in sampling protein conformations and accurately representing the interactions between the protein and the surrounding solvent. The linear solvent response approximation in Marcus theory, often used to model electron transfer, may not be valid in proteins where environmental relaxation timescales are comparable to electron transfer rates. Therefore, understanding the role of protein dynamics in guiding electron transfer is essential to accurately modeling the system.

This research utilizes femtosecond time-resolved serial crystallography (SX) to directly visualize the structural dynamics of electron transfer in *Drosophila melanogaster* (6-4) photolyase. Microcrystals of the photolyase were grown, crushed, and embedded in a hydroxyethyl cellulose (HEC) matrix. This matrix allowed for the continuous flow of crystals through an X-ray beam at the SwissFEL Alvra instrument. The crystals were photoexcited using a 150 fs laser pulse at 474 nm, and time-dependent crystallographic data were collected at various time delays (400 fs, 1 ps, 2 ps, 20 ps, 300 ps, and 100 µs) after photoexcitation. To ensure that the excitation fluence was in the linear one-photon absorption regime, control experiments were performed at lower fluences. Difference electron density (DED) maps were calculated by comparing diffraction patterns from photoexcited and dark crystals. The DED maps were analyzed to identify changes in protein structure occurring near the electron-transfer sites. In parallel, femtosecond transient absorption spectroscopy was conducted to obtain independent kinetic data on the charge transfer process and for comparison with the SX data. Data processing included steps such as peak finding, indexing, scaling, and refinement to generate high-quality structural models. The photoinduced structural changes were refined using real-space refinement against extrapolated maps, and the accuracy of the structural models was assessed by comparing observed and calculated DED maps.

The study reveals a series of precisely timed and spatially localized structural changes accompanying electron transfer in the photolyase. Immediately following photoreduction of FAD by the first tryptophan (Trp407), changes were observed in the side chain of Asn403, a residue crucial for charge stabilization. Asn403's carbonyl group moves away from the N5 of FAD, and nearby water molecules reorient to establish a new hydrogen bonding network, stabilizing the negative charge on the reduced FAD. These changes occur within a few picoseconds. The conserved Asp397-Arg368 salt bridge also exhibits a rapid response, with a transient distance increase between the two residues occurring within 400 fs, which decays after 2 ps. Around the second tryptophan (Trp384), structural changes are detected from 1 ps to 20 ps. Interestingly, a nearby methionine (Met408) shows a positive DED signal, suggesting its involvement in the electron transfer chain. Finally, prominent DED features around the fourth tryptophan (Trp381) are observed from 20 ps to 100 µs, consistent with the expected arrival time of the charge at the end of the electron-transfer chain. The observations indicate that the protein undergoes highly directed and carefully timed structural adaptations to facilitate electron transfer. The timescale of the various events identified were: (1) Asp397-Arg368 salt bridge response (400 fs), (2) Asn403 reorientation (1-20 ps), (3) Water molecule reorientation (2-300 ps), (4) Met408 structural change (1-20ps), and (5) Trp381 structural change (20ps-100µs). The overall process is faster than previous estimates based on spectroscopy and other methods.

The findings challenge the simple application of the linear solvent response approximation within Marcus theory to electron transfer in proteins. The precisely timed and spatially directed structural changes observed demonstrate that protein dynamics play a much more active and specific role in modulating electron transfer rates than previously appreciated. The ultrafast responses of the protein environment, especially the reorientation of water molecules, suggest a level of pre-organization and dynamic adaptation that is finely tuned for optimal electron transfer. The observation of Met408’s involvement suggests the protein redox pathway might be more complex than previously thought. This highlights the importance of considering the interplay between protein dynamics and electronic coupling in electron transfer reactions. The fact that the protein does not exhibit large-scale conformational changes suggests that the protein is engineered to avoid trapping charges in local energy minima, a crucial balance between efficient charge separation and charge availability for downstream DNA repair.

This study provides the first direct visualization of ultrafast structural dynamics accompanying electron transfer in a photolyase, using femtosecond time-resolved serial crystallography. The highly directed and carefully timed conformational changes observed highlight the essential role of protein dynamics in facilitating efficient and controlled charge transfer. The findings challenge current theoretical models of electron transfer in proteins and suggest that evolution has optimized protein fluctuations for optimal charge transfer. Further research could focus on investigating the detailed mechanisms of water reorganization and exploring the generality of these findings in other electron-transfer proteins.

While this study provides valuable insights into the structural dynamics of electron transfer, some limitations should be acknowledged. The time resolution of the experiment is limited by the X-ray pulse duration and the experimental setup. While the results are consistent with spectroscopy data, further improvements in time resolution could be desirable for a complete and detailed picture of the process. Furthermore, the study focuses on a single photolyase, and the generalizability of the findings to other photolyases and cryptochromes requires further investigation.