In vivocommensal control ofClostridioides difficilevirulence

Abstract

We define multiple mechanisms by which commensals protect against or worsen Clostridioides difficile infection.Using a systems-level approach we show how two species of Clostridia with distinct metabolic capabilities modulate the pathogen's virulence to impact host survival. Gnotobiotic mice colonized with the amino acid fermenter Clostridium bifermentans survived infection, while colonization with the butyrate-producer, Clostridium sardiniense, more rapidly succumbed. Systematic in vivo analyses revealed how each commensal altered the pathogen's carbon source metabolism, cellular machinery, stress responses, and toxin production. Protective effects were replicated in infected conventional mice receiving C. bifermentans as an oral bacteriotherapeutic that prevented lethal infection. Leveraging a systematic and organism-level approach to host-commensalpathogen interactions in vivo, we lay the groundwork for mechanistically-informed therapies to treat and prevent this disease.Clostridioides difficile, the etiology of pseudomembranous colitis, causes substantantial morbidity, mortality and >$5 billion/year in US healthcare costs. Infections commonly arise after antibiotic disruption of the microbiota, allowing the pathogen to proliferate and release toxins that ADP ribosylate host rho GTPases (1, 2). In patients with recurrent C. difficile infections, fecal microbiota transplant (FMT) has become standard of care to reconstitute the microbiota and prevent recurrence. While intensive efforts to develop defined microbial replacements for FMT have been undertaken, relatively little is known about the molecular, metabolic, and microbiologic mechanisms by which specific members of the microbiota modulate the pathogen's virulence in vivo, information critical for therapeutics development (3,4). Given deaths in immunocompromised patients from drug-resistant pathogens in FMT preparations (5), therapies informed by molecular mechanisms of action among will enable options with improved safety and efficacy (6, 7).C. difficile's pathogenicity locus (PaLoc) contains the tcdA, tcdB and tcdE genes that encode the A and B toxins, and holin involved in toxin export, respectively. tcdR encodes a sigma factor specific for the toxin gene promoters, and the tcdC gene a TcdR anti-sigma factor (8-10). Multiple metabolic regulators influence PaLoc expression (11,12). In particular, C. difficile elaborates toxin under starvation conditions to extract nutrients from the host and promote the shedding of spores.C. difficile, like other cluster XI Clostridia, possesses diverse genetic machinery to utilize different carbon sources for energy and growth. In addition to carbohydrate fermentation, the pathogen uses Stickland fermentations, and Stickland-independent fermentations of other amino acids including threonine and cysteine (13), to extract energy from amino acids. The pathogen can ferment ethanolamine, extract electrons from primary bile salts, and undergo carbon fixation through the Wood-Ljungdhal pathway to generate acetate for metabolism...

Predictive regulatory and metabolic network models for systems analysis ofClostridioides difficile

“…These findings were confirmed with the high levels of spore release in expanded populations of vegetative C. difficile when co-colonized with CSAR (22). Overall, this analysis suggested that the sporulation pathway is an indicator of C. difficile disease, reinforcing the Spo0A-mediated link between sporulation and toxin production recapitulated by the model (Fig.…”

“…Model predictions also illustrated C. difficile's predicted shift from carbohydrate utilization towards amino acid utilizing pathways in vivo, as shown by the enhanced set of 15 amino acids, including the preferred Stickland donor and acceptor amino acids (leucine and proline) known to support metabolism and growth(43,44,47). Notably, many of these amino acids show high abundance within the gut lumen in gnotobiotic and conventional colonization states that enhance C. difficile's capacity to colonize and expand(22).The Cdiff Web Portal, a resource for the C. difficile communityWe have released a new C. difficile Web Portal (http://networks.systemsbiology.net/cdiff-portal/)to provide a discovery and collaboration gateway for the C. difficile scientific community. The portal aims to accelerate the advancement of the science and understanding of C. difficile biology, gene regulation, and metabolism on its virulence.…”

“…We also curated pathway annotations that were incorrectly designated using default KEGG annotations (61) File S1. Then, the transcriptome of C. difficile profiled from in vivo infections of specifically-colonized gnotobiotic mice (22) was mapped onto the icdf834 model using the GIMME algorithm (46). This resulted in a model with 665 active reactions, with no changes in the number of genes.…”

“…The C. difficile 630 (CD630) genome encodes 4,018 genes, with ~309 putative transcription factors (TFs; including sigma factors), 1,030 metabolic genes, and 1,330 genes (>30%) with unknown function (20,21). The ATCC43255 strain of C. difficile, which is used to generate symptomatic infections in mice, encodes 4,117 genes and ~327 TFs, of which ~97% are significantly orthologous to genes encoded in the CD630 strain (22). To address questions regarding the broader systems-level interplay among genes in colonization and infection, we used computational modeling and network inference algorithms to construct an Environment and Gene In addition to EGRIN, we have advanced a metabolic network model of C. difficile to understand how conditional regulation manifests physiologically, by adding reactions and associated genes supporting the exchange of nutrients required for growth in intestinal environments.…”

“…5C).We next extended and applied the model to predict C. difficile behaviors and gene essentiality in vivo. C. difficile transcriptomes from specifically-colonized gnotobiotic mice(22) were used as input into the GIMME algorithm(46). Analyses of expressed transcripts in vivo identified 665 active reactions(Fig.…”