HgutMgene-Miner: In silico genome mining tool for deciphering the drug-metabolizing potential of human gut microbiome

Abstract

The biotransformation of drugs by enzymes from the human microbiome can produce active or inactive products, impacting the bioactivity and function of these drugs inside the human host. However, understanding the biotransformation reactions of drug molecules catalyzed by bacterial enzymes in human microbiota is still limited. Hence, to characterize drug utilization capabilities across all the microbial phyla inside the human gut, we have used a knowledge-based approach to develop HgutMgene-Miner software which predicts xenobiotic metabolizing enzymes (XMEs) through genome mining. HgutMgene-Miner derives its predictive power from the MicrobiomeMetDB database, which systematically catalogs all known biotransformation reactions of xenobiotics and primary metabolites mediated by host-associated microbial enzymes. Over 10,000 isolate genomes from 830 different bacterial species found in the Unified Human Gastrointestinal Genome (UHGG) collection have been analyzed by HgutMgene-Miner. This led to the identification of 89,377 xenobiotic metabolizing enzymes (XMEs) across 13 phyla, with the greatest diversity in Bacteroidota, Firmicutes_A, Firmicutes, and Proteobacteria. Bacteroides, Clostridium, and Alitsipes were found to be the richest genera, while Actinomyces were found to encode the fewest XMEs, primarily metabolizing Diclofenac, a nonsteroidal anti-inflammatory drug. Overall, we discovered XMEs in 220 genera, exceeding the number experimentally reported in fewer than 10 genera. Notably, Eggerthella lenta's cgr2 involved in Digoxin inactivation was identified in very distant Holdemania genera, likewise Clostridium leptum's nitroreductase, involved in Nitrazepam metabolism, was found in Fusobacterium. These findings highlight the extensive and diverse distribution of XMEs across microbial taxa.