Comparative genomic and extracellular metabolomic profiling of Lactiplantibacillus plantarum LP140 and Bifidobacterium breve BB010 reveals distinct gut–brain-axis-associated metabolite profiles

Background The microbiota-gut-brain axis provides multiple routes through which intestinal bacteria may influence host neurophysiology. Mechanistic evaluation of candidate probiotic strains therefore requires linking genomic metabolic potential with measurable extracellular output. We compared Lactiplantibacillus plantarum LP140 and Bifidobacterium breve BB010 to identify distinct metabolic traits potentially relevant to gut–brain-axis-associated functions. Results Genome-based pathway reconstruction predicted distinct metabolic capacities. LP140 showed stronger support for glutamate-to-γ-aminobutyric acid (GABA) conversion and a broader succinate-associated fermentative framework, whereas BB010 showed stronger de novo B3/NAD-related capacity, PdxST-dependent PLP biosynthesis, and broader Gln/Glu interconversion support. Species-level pangenome analysis further showed that many of the canonical metabolic loci underlying these predictions were conserved across the respective species rather than unique to the focal strains. Replicated targeted LC–MS/MS confirmed major strain-resolved extracellular differences. LP140 showed pronounced accumulation of lactate (mean Δ ± SE = 127.0 ± 8.9 mM, q = 0.0089), succinate (999 ± 49 µM, q = 0.0147), and GABA (60.6 ± 4.2 µM, q = 0.0157). BB010 showed strong accumulation of nicotinic acid (22.8 ± 0.8 µM, q = 0.0009), glutamine (347 ± 38 µM, q = 0.0204), and glycine (5.55 ± 0.64 mM, q = 0.0204). Genome–metabolite relationships were not uniformly direct, indicating that extracellular metabolite pools cannot be predicted from pathway completeness alone. Polyamine-related findings were more consistent with turnover than broad de novo biosynthesis, while histamine and tyramine did not accumulate under the tested conditions. Conclusions LP140 and BB010 exhibit distinct, condition-specific extracellular metabolic profiles that are only partly predictable from genomic pathway content. These strain-resolved genome–metabolome relationships provide a functional basis for further testing in physiologically relevant host-interaction models but do not by themselves demonstrate psychobiotic efficacy.