Autoinducer-2 (AI-2) is a broadly studied quorum sensing molecule derived from the spontaneous cyclization of DPD, a product released during the conversion of S-adenosyl-homocysteine to homocysteine by the MtnN–LuxS pathway. An alternative enzyme, SahH, catalyzes the direct conversion of S-adenosyl-homocysteine to homocysteine without producing DPD/AI-2. The authors performed a comprehensive genomic survey across members of the phylum Bacteroidota to determine the distribution of luxS (associated with mtnN-luxS) versus sahH.
Their analysis indicates that the ancestral Bacteroidota genome likely contained sahH and that multiple independent replacement events introduced mtnN-luxS into diverse lineages during evolutionary diversification. This pattern suggests repeated gains of the AI-2–producing pathway rather than a single conserved acquisition across the phylum.
Within several gut-associated groups—specifically the Bacteroidaceae, Parabacteroides, and many members of Prevotellaceae—the authors observed that either the mtnN-luxS module or sahH is located in the same genomic region adjacent to the gene yfhO. The conserved neighborhood implies a locus prone to replacement between the two biochemical options (AI-2–producing versus non–AI-2-producing) in these lineages. The source reports this conserved adjacency but does not provide further sequence details or mechanistic explanation in the abstract.
To test whether genomic presence of luxS translates into AI-2 production, the study examined representative gut species with contrasting gene content. The authors used Bacteroides fragilis, which encodes sahH, and two species that encode mtnN-luxS, Bacteroides uniformis and Phocaeicola vulgatus. According to the reported results, strains that contain luxS produced detectable AI-2, demonstrating that the mtnN-luxS pathway is functional in gut Bacteroidales and yields the quorum sensing molecule.
The abstract does not report the exact assay types, quantitative AI-2 levels, or experimental controls; those methodological specifics and data values are not included in the provided source text.
The authors carried out transcriptomic analyses and gnotobiotic mouse experiments using wild-type strains and engineered genetic swaps that exchanged mtnN-luxS and sahH alleles between strains. These experiments were designed to test whether Bacteroidaceae species both produce and respond to AI-2.
Under the conditions tested, the data showed that Bacteroidaceae species did not exhibit a detectable response to AI-2. This result applied to the transcriptomic readouts and to the gnotobiotic in vivo experiments reported in the abstract. The source does not provide additional details on the environmental conditions, time points, or the criteria used to define responsiveness.
To assess the potential ecological significance of AI-2 production by Bacteroidales, the authors analyzed 15 human gut metagenomic datasets. They report that species containing mtnN-luxS account for a substantial fraction of many gut microbial communities, averaging between about 5.1% and 33% of the total bacteria across the datasets examined. This prevalence indicates that luxS-containing Bacteroidales could be a major source of community AI-2 in the human gut environment.
The abstract does not list the datasets, sampling populations, or the computational thresholds used to assign species or pathways in metagenomes; those specifics were not reported in the provided source excerpt.
Taken together, the genomic, experimental, and metagenomic observations support a model in which the ancestral Bacteroidota enzyme repertoire lacked AI-2 production, but multiple independent acquisitions of mtnN-luxS have produced numerous modern gut lineages capable of synthesizing AI-2. Although luxS-positive Bacteroidales produce AI-2, the tested Bacteroidaceae strains did not show intrinsic responses to AI-2 under the experimental conditions reported, suggesting that these taxa may primarily serve as AI-2 producers rather than responders in the gut ecosystem.
Because mtnN-luxS-containing species can represent up to a third of gut microbial communities in some datasets, the community-level production of AI-2 may influence other AI-2–responsive symbionts or pathogens in the human gut.
Limitations and caveats: the source is a preprint and has not been peer reviewed. The abstract does not include methodological details such as assay types, quantitative AI-2 measurements, statistical analyses, or full experimental conditions for transcriptomics and gnotobiotic studies. Those details are required to assess reproducibility and interpretation but were not reported in the provided source material.