An ancient divide in outer membrane tethering systems in Bacteria
Recent data support the hypothesis that Gram-positive bacteria (monoderms) arose from Gram-negatives (diderms) through loss of the outer membrane (OM). However how this happened remains unknown. Considering that tethering of the OM is essential for cell envelope stability in diderm bacteria we hypothesize that its destabilization may have been involved in OM loss. Here, we present an in-depth analysis of the four main OM tethering systems across all Bacteria. We show that their distribution strikingly follows the bacterial phylogeny with a bimodal distribution matching the deepest phylogenetic cleavage between Terrabacteria (a clade encompassing Cyanobacteria, Deinococcus/Thermus, Firmicutes, etc.) and Gracilicutes (a clade encompassing Proteobacteria, Bacteroidetes, Spirochaetes, etc.). Diderm Terrabacteria display as the main system OmpM, a porin that attaches non-covalently to modified peptidoglycan or to secondary cell wall polymers. In contrast, the lipoprotein Pal is restricted to the Gracilicutes along with a more sporadic occurrence of OmpA. While Brauns lipoprotein Lpp is largely considered as the textbook example of OM attachment, it is actually present only in a subclade of Gammaproteobacteria. We propose an evolutionary scenario whereby the last common bacterial ancestor used a system based on OmpM, which was later replaced by one based on the lipoprotein Pal concomitantly to the emergence of the Lol machinery to address lipoproteins to the OM, with OmpA as a possible transition state. We speculate that the existence of only one main OM tethering system in the Terrabacteria would have allowed the multiple emergences of the monoderm phenotype specifically observed in this clade through OmpM perturbation. We test this hypothesis by inactivating all four ompM gene copies in the genetically tractable diderm Firmicute Veillonella parvula. The resulting mutant is severely affected in growth and displays high sensitivity to OM stress. High resolution imaging and tomogram reconstructions reveal a dramatic - yet non-lethal - phenotype, in which vast portions of the OM detach, producing large vesicles surrounding multiple monoderm-like cells sharing a common periplasm. Complementation by a single OmpM rescues the phenotype to a normal cell envelope. Together, our results highlight an ancient shift in bacterial evolution involving OM tethering systems. They suggest a possible mechanism for OM loss and a high flexibility of the cell envelope in diderm Firmicutes, making them ideal models to further refine our understanding of the mechanisms involved in bacterial OM stability, and opening the way to recapitulate the monoderm/diderm transition in the laboratory.