ИСТИНА

Cyclooxygenases (COXs) and lipoxygenases (LOXs) are the most important oxylipin-forming enzymes in humans and other vertebrates. They have great medical significance as drug targets to treat inflammation, pain and fever. Fungi and plants also have cyclooxygenase homologs that participate in oxylipin biosynthesis and provide immunity signalling, quorum sensing and sexual reproduction. COX- and LOX-like enzymes are also the backbone of the oxylipin biosynthetic pathway in these organisms. COX- and LOX-like enzymes are also described in some bacteria [https://doi.org/10.1134/S0006297920090059, https://dx.doi.org/10.1074/jbc.M114.555904]. Similar functions of COX- and LOX-like enzymes in different kingdoms of life and their evident homology raise a question about their origin and evolution. Recently, we published a paper on LOX evolution [https://doi.org/10.1134/S0006297920090059], where found both a statistical and an evolutionary association between LOXs and multicellularity, as well as a ‘minor’ association between LOXs and cross-kingdom virulence. We assumed that LOXs could have originated in multicellular bacteria and be borrowed by different groups of eukaryotes by horizontal gene transfer to provide cell-to-cell signalling. Here, we present a follow-up of that study aimed to elucidate the second core component of all oxylipin biosynthetic pathways – COX-like enzymes. Such a research has already been published by Gupta and Selinsky [https://doi.org/10.1016/j.bbamem.2014.09.011], but includes a small sequence sample and the resulting phylogeny of bacteria COX-like enzymes seems controversial. We have performed a phylogenetic research for animal COXs, plant alpha-dioxygenases (αDOXs) and fungal Psi-producing oxygenases (PPOs) with a larger amount of bacterial sequence data. We have revealed that plant αDOXs and fungal PPOs have originated independently of animal COXs from bacterial peroxidase-like enzymes with unknown function. Animal COXs appeared to be closely related to some groups of bacterial enzymes, including the single described cyanobacterial COX-like enzyme [https://dx.doi.org/10.1074/jbc.M114.555904]. Combining available PDB data with sequence alignments, we have found that COXs, αDOXs and PPOs share common catalytic residues (that are also shared with peroxidases and catalases), but their substrate-binding residues developed independently. It confirms our assumption that COXs, αDOXs and PPOs originated independently from peroxidase-like enzymes despite very similar biochemical and biological function. Substrate-binding facility appeared to be homologous only in animal COXs and the cluster of bacterial COX-like enzymes closely related to them. A detailed statistical analysis of this cluster revealed the highest abundance of COX-like enzymes in multicellular bacteria (Cyanobacteria and Actinobacteria), similar to observed peaks of LOX abundance in multicellular Myxococcales and Cyanobacteria. It probably reflects that bacterial COX-like enzymes are also associated with cell-to-cell signalling and forming complex structures and raises a possibility that they could also originate and spread across the life to support multicellularity. These findings support the hypothesis that oxylipins have initially originated to provide cell-to-cell signalling even at the ‘bacterial’ stage of their evolution [https://doi.org/10.1134/S0006297920090059].