When a stool sample is analysed for the compounds gut bacteria make, the usual method breaks every cell open first and reports one total. That total mixes two very different things: the compounds floating free between the bacteria, which are available to other microbes and to the person, and the compounds still sitting inside bacterial cells, which are not. This study measured the same sample twice, once gently and once after breaking the cells, and treated the difference as the inside pool.
Drawn from the paper's introduction.
The authors write that because fecal concentrations are simultaneously influenced by microbial production, intracellular retention, consumption by other species and host absorption, it is fundamentally difficult to distinguish producers from consumers based on a single total measurement. They describe this as an interpretive bottleneck, where changes suggested by functional gene analysis do not necessarily align with changes in measured metabolite levels.
Drawn from the paper's introduction.
Stool from 63 healthy Japanese adults was split into paired fractions from the same specimen: an undisrupted fraction standing for the extracellular pool, and a strongly bead beaten fraction standing for the total. The intracellular pool was defined as the difference. Short chain fatty acids, polyamines and water soluble vitamins were quantified in both, then integrated with shotgun metagenomic profiles of species composition and functional genes.
Vitamins split into two clear behaviours. Adenosylcobalamin, a coenzyme form of B12, was detectable only after disruption, and together with thiamine and niacin was classified as intracellular retained. Methylcobalamin, thiamine, nicotinate and pyridoxamine all showed a gentle fraction below 10 percent of the total. Biotin and pantothenate behaved differently, with gentle fractions around 40 percent, and riboflavin sat with them.
The ecological associations followed the split. Biotin correlated with Alistipes abundance at 0.56 in both fractions, while thiamine carried a negative correlation of minus 0.31, consistent with intensive microbial utilisation. For the locked inside vitamins, no bacterial genus reached a correlation above 0.3 in the gentle fraction, meaning the associations that a conventional total measurement would report for those compounds are being driven by the pool nothing else can reach.
Species within the same genus excluded one another far more often than species across genera, at 40.0 percent of 20 pairs and 44.4 percent of 72 pairs against 17.8 percent of 90 pairs.
| Cohort | 63 healthy adults, 31 male and 32 female, mean age 66.9±9.8 y, mean BMI 22.6±3.1 |
| Adenosylcobalamin, a B12 coenzyme form | undetectable in the gentle fraction; identified only after cells were broken open |
| Methylcobalamin, thiamine, nicotinate, pyridoxamine | gentle fraction below 10% of total |
| Pantothenate, riboflavin, biotin | gentle fraction roughly 40% of total |
| Biotin against Alistipes abundance | correlation 0.56 in both fractions |
| Thiamine, negative correlation | minus 0.31 |
| Locked inside vitamins, gentle fraction | no bacterial genus reached a correlation above 0.3 |
| Species level mutual exclusivity within vs between genera | 40.0% of 20 pairs and 44.4% of 72 pairs, vs 17.8% of 90 pairs |
Proposed by the authors This is the explanation the authors offer in their discussion. This study did not test it.
For the B12 result the authors propose that the failure of adenosylcobalamin to accumulate outside the cells reflects the combined consequences of its scarcity and its essentiality, which drive highly efficient intracellular sequestration and competitive community wide reclamation. They note that making it from scratch takes roughly thirty enzymatic steps, and that any released is likely to be rapidly scavenged by surrounding microbes through high affinity transporters.
For thiamine they offer two explicitly non exclusive routes: an ecological one in which dominance of non producers reduces community level production capacity, and a physiological one in which increased demand during vigorous fermentation outpaces supply.
They then state the limit directly, writing that correlation based analyses alone are insufficient to fully capture the underlying biology, and that establishing causal mechanisms will require culture based validation, metatranscriptomic profiling and stable isotope tracing.
Drawn from Discussion, PMC13505459.
The single number that fecal metabolite studies normally report turns out to be two different quantities added together, and for several vitamins almost all of it is the pool that is not actually available.
Healthy Japanese adults with an average age in the sixties, one stool sample each.
Isotope tracing or controlled co-culture, which the authors name themselves, is what would turn these correlations into a map of who produces and who consumes.
Newsletter
Each issue by email, when the newsletter launches. Leaving your address puts you on the list, nothing is sent yet.