ИСТИНА

Fluorescent protein KillerRed is a genetically encoded photosensitizer, which produces reactive oxygen species (ROS) under green-orange light irradiation. This feature allows using KillerRed in many areas of biology as a tool for the precise inactivation of cellular proteins and killing target cell populations. Nevertheless, this protein is much less efficient compared to chemical photosensitizers. To date, numerous attempts to increase the phototoxicity of KillerRed by using rational design have not been successful. The application of directed evolution is limited by difficulty in mutant proteins selection: cells expressing more phototoxic protein variants die and can not be sorted by standard methods. In our work, we propose a method that will not only solve the problem of more phototoxic KillerRed mutants selection, but it will be applicable to many similar problems. Our approach is based on high-throughput sequencing of KillerRed mutants library: we create a hundreds of thousands KillerRed mutants, expressed in bacteria. We irradiate some of the bacteria with bright light, and sequence genes, presented in the library before and after irradiation. Irradiation leads to cell death proportionally to the phototoxicity of a KillerRed mutants. Therefore, using next generation sequencing we track the changes in the genes frequencies within the population before and after irradiation and determine the level of phototoxicity for each mutant. Besides to the practical result in the form of obtaining new improved KillerRed mutants, our approach will provide an extensive set of data about impact of numerous mutations on the KillerRed function, and will also contain data about the interaction of mutations in a protein molecule. The obtained results will facilitate understanding the structural basis of fluorescent proteins phototoxicity. So we expect that this information will allow to find out the mechanism of phototoxic fluorescent proteins action and will give an opportunity to rationally design phototoxic proteins with improved efficiency. In addition, our approach will potentially be applicable in the analysis of other toxic molecules libraries, such as antibacterial agents, highly active enzymes (nucleases, proteases, etc.) and guided or interfering RNAs.