Person:

Roelofs, Kevin G.

Background: Type VI secretion systems (T6SSs) are contact-dependent antagonistic systems employed by Gram negative bacteria to intoxicate other bacteria or eukaryotic cells. T6SSs were recently discovered in a few Bacteroidetes strains, thereby extending the presence of these systems beyond Proteobacteria. The present study was designed to analyze in a global nature the diversity, abundance, and properties of T6SSs in the Bacteroidales, the most predominant Gram negative bacterial order of the human gut. Results: By performing extensive bioinformatics analyses and creating hidden Markov models for Bacteroidales Tss proteins, we identified 130 T6SS loci in 205 human gut Bacteroidales genomes. Of the 13 core T6SS proteins of Proteobacteria, human gut Bacteroidales T6SS loci encode orthologs of nine, and an additional five other core proteins not present in Proteobacterial T6SSs. The Bacteroidales T6SS loci segregate into three distinct genetic architectures with extensive DNA identity between loci of a given genetic architecture. We found that divergent DNA regions of a genetic architecture encode numerous types of effector and immunity proteins and likely include new classes of these proteins. TheT6SS loci of genetic architecture 1 are contained on highly similar integrative conjugative elements (ICEs), as are the T6SS loci of genetic architecture 2, whereas the T6SS loci of genetic architecture 3 are not and are confined to Bacteroides fragilis. Using collections of co-resident Bacteroidales strains from human subjects, we provide evidence for the transfer of genetic architecture 1 T6SS loci among co-resident Bacteroidales species in the human gut. However, we also found that established ecosystems can harbor strains with distinct T6SS of all genetic architectures. Conclusions: This is the first study to comprehensively analyze of the presence and diversity of T6SS loci within an order of bacteria and to analyze T6SSs of bacteria from a natural community. These studies demonstrate that more than half of our gut Bacteroidales, equivalent to about ¼ of the bacteria of this ecosystem, encode T6SSs. The data reveal several novel properties of these systems and suggest that antagonism between or distributed defense among these abundant intestinal bacteria may be common in established human gut communities. Electronic supplementary material The online version of this article (doi:10.1186/s12864-016-2377-z) contains supplementary material, which is available to authorized users.

ABSTRACT We recently showed that human gut Bacteroidales species secrete antimicrobial proteins (BSAPs), and we characterized in vitro the first such BSAP produced by Bacteroides fragilis. In this study, we identified a second potent BSAP produced by the ubiquitous and abundant human gut species Bacteroides uniformis. The two BSAPs contain a membrane attack complex/perforin (MACPF) domain but share very little sequence similarity. We identified the target molecules of BSAP-sensitive cells and showed that each BSAP targets a different class of surface molecule: BSAP-1 targets an outer membrane protein of sensitive B. fragilis strains, and BSAP-2 targets the O-antigen glycan of lipopolysaccharide (LPS) of sensitive B. uniformis strains. Species-wide genomic and phenotypic analyses of B. fragilis and B. uniformis showed that BSAP-producing strains circumvent killing by synthesizing an orthologous nontargeted surface molecule. The BSAP genes are adjacent to the gene(s) encoding their target replacements, suggesting coacquisition. Using a gnotobiotic mouse competitive-colonization model, we found that the BSAP surface targets are important for colonization of the mammalian gut, thereby explaining why they are maintained in sensitive strains and why they were replaced rather than deleted in BSAP-producing strains. Using isogenic BSAP-producing, -sensitive, and -resistant strains, we show that a BSAP-producing strain outcompetes a sensitive strain but not a resistant strain in the mammalian gut. Human gut metagenomic datasets reveal that BSAP-1-sensitive strains do not cooccur with BSAP-1-producing strains in human gut microbiotas, further supporting the idea that BSAPs are important competitive factors with relevance to the strain-level composition of the human gut microbiota.