Big data omics analysis guided discovery of bacteriocins from the human microbiome

Abstract

Bacteriocins are ribosomally synthesized antimicrobial peptides produced by bacteria, exhibiting diverse structures from post-translationally modified to linear forms. Their diversity and potency make them promising alternatives or adjuncts to antibiotics in combating antimicrobial resistance. Beyond mediating microbe-microbe interactions, bacteriocins may engage with the human host, underscoring the importance of exploring human microbiome-derived bacteriocins and their health implications. Although the traditional activity-based screening approach identifies diverse bacteriocins from the human microbiome, they remain largely untapped in the meta-omics era. Therefore, exploring human microbiome-derived bacteriocins can expand the antimicrobial arsenal and deepen insights into their roles in human health. Chapter 1 begins by introducing the bacteriocins’ biology, including their classification, regulation, mechanisms, and immunity. Their potential as promising antibiotic alternatives to contain antibiotic resistance drives ongoing efforts for bacteriocin discovery. Given that the human microbiome is a rich, underexplored reservoir of microbes and that the bacteriocins’ co-evolution with the human host may enable targeting against specific pathogens while contributing to health, the human microbiome stands as a priority source for bacteriocin discovery. A century of bacteriocin discovery has revealed the significant biosynthetic potential of lactic acid bacteria (LAB), which produce a diverse range of bacteriocins that may benefit the human host. However, the biosynthetic potential of LAB-derived bacteriocins remains elusive, particularly in terms of their diversity, abundance, and distribution within the human microbiome. In Chapter 2, we systematically explore LAB-derived bacteriocins, revealing their species-specificity in the LAB species and niche-specificity in the human microbiomes. Notably, unmodified class II bacteriocins are enriched in the vaginal microbiome and may play regulatory roles, as suggested by omics analyses. As proof of principle, we chemically synthesized two novel bacteriocins, confirming the in vitro antibacterial activity of one. In a mouse model of Gardnerella vaginalis-induced bacterial vaginosis, this bacteriocin increases microbial diversity and regulates the vaginal microbiome, alleviating bacterial vaginosis symptoms. While some bioinformatic tools are available for bacteriocin prediction, most focus predominantly on modified bacteriocins and overlook unmodified ones, limiting comprehensive biosynthetic and functional studies for unmodified class II bacteriocins. In Chapter 3, we address this gap by developing IIBacFinder, a tool dedicated to detecting class II bacteriocins. A large-scale bioinformatic analysis reveals their widespread distribution across the bacterial kingdom, supported by the in vitro antibacterial activities of 21 chemically synthesized hypothetical bacteriocins from diverse phyla. To investigate biosynthetic potential in the human gut microbiome, we synthesize 26 hypothetical bacteriocins from gut commensal species, of which 16 exhibit antibacterial activity. These functional bacteriocins represent the first identified from several key human gut commensal species. Further ex vivo tests show minimal impact of these narrow-spectrum bacteriocins on human fecal microbiota. This thesis focuses on discovering human microbiome-derived bacteriocins, particularly class II bacteriocins, through a big data analysis. We highlight the huge biosynthetic potential of untapped bacteriocins in the human microbiome, aiming to identify novel bacteriocins with therapeutic and probiotic potential. Additionally, a comprehensive investigation can advance our understanding of their ecological roles within the human microbiome.