With the rise of antibiotic resistance, there is a need for alternatives to antibiotics. Predatory bacteria attack and kill animal and plant pathogens, making them potential biological alternatives to antibiotics. To develop effective clinical applications of predatory bacteria, understanding the genomic basis of the predatory lifestyle is crucial. Comparative genomics studies are needed to define genome differences between predatory and non-predatory bacteria and to identify gene families unique to predatory bacteria. Further, comparative genomics of the well-studied predatory genus Bdellovibrio are needed to understand variation in predatory phenotypes.
To this end, we compiled a dataset of 29 complete genomes. This includes all 15 publicly available complete predatory bacteria genomes belonging to Bdellovibrionota, a newly defined phylum, and comprises 10 Bdellovibrio, 3 Halobacteriovorax, and 2 Bacteriovorax. The Bdellovibrio strains include 2 epibiotic and 8 intraperiplasmic predators, representing diversity in predatory strategy. Epibiotic predators attach to the outer surface of prey bacteria but do not invade, while intraperiplasmic predators invade the prey’s periplasm using lytic enzymes. To complete the dataset, we chose bacteria that belong to Bdellovibrionota and a closely related phylum and are not predicted to have a predatory component to their lifestyle (Waite et al., 2020).
We then used anvi’o, a pangenome analysis software package, to build clusters of protein-coding genes and analyze differences in gene cluster distribution. We tested multiple settings for the mcl inflation parameter, which defines granularity of gene clusters. Due to phylogenetic diversity among the genomes, we selected a setting of 1.4 to define coarse gene clusters. We generated three pangenomes: a full pangenome of predatory and non-predatory bacteria, a predatory pangenome of only predatory bacteria, and a Bdellovibrio pangenome of only Bdellovibrio strains.
We identified single copy core gene clusters, which include one protein-coding gene from each genome, and singleton gene clusters, which include protein-coding genes from only one genome. For each of the pangenomes, singleton gene clusters greatly outnumbered single copy core gene clusters, indicating few conserved genes across the genomes. Preliminary analysis of gene clusters identified some genes that may be enriched in predatory bacteria and contribute to their lifestyle, such as a murein L,D-transpeptidase catalytic domain family protein, a penicillin-insensitive murein endopeptidase, and a murein peptide amidase. Functional annotation of these genes indicate involvement in peptidoglycan metabolism, which is a significant aspect of predation. Across all 29 genomes, functional annotation combined from multiple sources, including PGAP and InterProScan, associated 3879 gene clusters with “hydrolase,” 1454 gene clusters with “peptidase,” and 988 gene clusters with “nuclease.” Moving forward, we are constructing a catalog of these lytic enzyme gene clusters to determine their distribution across the genomes and identify those unique to predatory bacteria. Results from this project will further our understanding of the importance of lytic enzyme gene families to the predatory lifestyle.
SD was supported by a Walsh Fellowship and a Summer Undergraduate Research Grant from Providence College.
