Photochromism of flavin-based photoreceptors: new perspectives for super-resolution microscopy and optogenetics

Flavin-binding blue-light (BL) sensor proteins (Fl-Blues) of bacterial origin have been investigated in this work with diverse biophysical methods, with the aim to characterize their photophysical and photochemical properties. Investigations have been carried out both in vitro and in vivo, to explore their potential in advance cellular applications and establish novel methods of analysis. Besides the lab work, extensive in silico analysis have been performed, in the attempt to optimize searching in public databanks, identify novel Fl-Blues and gain hints on their evolution patterns. Main object of lab work has been the protein YtvA from Bacillus subtilis, a BL-photoreceptor involved in environmental stress responses. Its photosensing unit is a LOV (Light, Oxygen and Voltage) domain, that binds oxidized flavin mononucleotide (FMN) as a photoactive chromophore, brightly fluorescent in the dark-adapted state. BL illumination triggers a photocycle with formation of a FMN-cysteine covalent adduct (light adapted or lit state), with complete loss of fluorescence and concomitant activation of a biological response (signaling). The adduct thermally recovers to the dark-state with breakage of the covalent bond, on a time scale of hours at room temperature. This protein is thus naturally bi-stable, both for its optical and biological properties. Using steady-state and time resolved optical spectroscopies, we could show here that application of violet or UVA light to the adduct is able to induce a photoequilibrium, with partial recovery of the dark-adapted state and of FMN fluorescence. In other words, YtvA was demostrated to be a photochromic system. It was possible to calculate the quantum yield of the photoinduced light-to-dark reaction as ca. 0.05, that is about 1/10 the efficiency of the forward dark-to-light photochemical step. This newly discovered photochromism of YtvA, allowed us to show its potential in super-resolution microscopy, in particular using Fluorescence Photo-activation Localization Microscopy (FPALM) to visualize YtvA fluorescence inside transformed Escherichia coli cells with nanometer precision. In order to try to optimize and tune the photochromic, photochemical and spectral properties of YtvA, a set of mutagenized variants has been designed and investigated. In particular we were able to identify amino acids that act as major spectral tuners in the UVA region, and other residues able to strongly affects the dynamics of the photocycle. Another notable property of YtvA has emerged by performing epifluorescence microscopy experiments on living cells expressing the proteins. During these studies, still at a preliminary stage, it became clear that the level of hydration is able to affect the thermal recovery rate of the photocycle, a property that is relevant to understand the effects of water on this kind of systems and for applications in vivo. Genome digging has revealed, during the last decade, that LOV domains are widespread among the three life domains and are emerging as BL sensing systems not only in plants, but also in prokaryotes. On full-length proteins LOV modules they are usually linked to diverse effector domains that determine the biological functionality of the protein itself. A second and different type of BL-sensor, also widespread among prokaryotes, is the so-called BLUF (Blue Light sensing Using Flavins), that upon light-activation undergoes a rapidly reversible hydrogen-bond (HB) switch. In this work we performed an extensive database search in order to have a comprehensive scenario of LOV and BLUF proteins in the prokaryotic world, inspect phylogenetic pathways and discover possible novel functionalities for effector domains, a feature connected with optogenetics. From distance-trees obtained by using neighbouring-methods, it was observed that most LOV and BLUF domains from organisms belonging to the same phylum are neighbour, but that in many cases clustering occurs according to effector functions associated to the photosensing domains. Metagenomics and bio-informatic analysis have only recently been initated, but signatures are beginning to emerge that allow definition of a bona fide LOV or BLUF domain, aiming at better selection criteria for novel BL sensors based on sequence logos. Taking advantage of these new criteria it was built, for the first time, the phylogenetic tree for archaeal LOV domains that have reached a statistically significant number, but have not at all been investigated so far.