The composition of the human gut microbiome and its relationship with diseases has been widely explored (for instance, in the MetaHIT project and the Human Microbiome Project). However, the extent to which gut species interact and shape gut microbiome composition dynamics is unknown.

A new study, led by Dr. Jeroen Raes from the Department of Microbiology and Immunology at KU Leuven (Belgium), has found that human gut bacteria behave differently depending on whether they are grown alone or in a group under in vitro conditions.

The researchers sought to study interactions between human gut commensals under well-defined in vitro conditions and validate a quantitative mechanistic model by predicting gut microbiome dynamics.

First, a synthetic gut community was established that included three human strains that were representative of the adult gut microbiome and which included the human gut commensals Roseburia intestinalis L1-82, Faecalibacterium prausnitzii A2-165 and Blautia hydrogenotrophica S5a33. This defined community was studied with a combination of mono-, bi-, and tri-cultures, mechanistic modeling and sequencing of ribonucleic acid in samples (transcriptomics).

Experiments in laboratory fermenters in batch mode showed that all three strains competed for fructose. Furthermore, formate produced by R. intestinalis L1-82 and F. prausnitzii A2-165 was used by B. hydrogenotrophica S5a22 and acetate generated by B. hydrogenotrophica S5a33 was used by the other two strains. These findings show the bacteria’s team efforts not only to compete against the same substrate, but also to mutually cross-feed at the same time.

The researchers then tried to predict bacterial strain behavior from mono-culture standard properties and growth kinetics using a mechanistic model. It was found that bacterial strains grown together in bi- or tri-cultures behaved differently than when grown alone. Specifically, prediction accuracy of mono-culture dynamics was strain-dependent and relied not only on the presence of other strains, but also on key substrates in the culture medium. After several experiments, the researchers found that bi-culture data was necessary to accurately describe how the three bacterial strains behave together.

For instance, when growing alone in the presence of formate, B. hydrogenotrophica S5a33 produced both hydrogen gas and carbon dioxide, in contrast with its known role as a generator of acetate in the presence of carbon dioxide and hydrogen. On the other hand, when culturing R. intestinalis L1-82 and B. hydrogenotrophica S5a33, fructose and acetate were consumed and both strains generated carbon dioxide, hydrogen gas, butyrate and lactate as by-products.

It was also proven that different bacterial behavior was the result of altered gene expression, due to niche interactions with other strains in mono-cultures compared with tri-cultures. For instance, in tri-cultures, several metabolic pathways involved in vitamin B12 and amino acid and protein biosynthesis were down- and up-regulated, respectively, thus ultimately benefiting the growth of bacteria.

Furthermore, it was observed that the behavior of the three bacteria when grown together relied not only on kinetic parameters, but also on how bacteria were brought together in the beginning (initial conditions and lag phase).

On the whole, these findings confirm bacterial strains behave differently alone than when in a group with other bacteria. Furthermore, bacterial behavior also depends on how bacteria were brought together. This data adds to previous models, such as the generalized Lotka-Volterra model, in the straightforward way they develop and validate predictive models of gut microbial communities to better characterize the dynamics and stability of the gut microbiome.



D’hoe K, Vet S, Faust K, et al. Integrated culturing, modeling and transcriptomics uncovers complex interactions and emergent behavior in a three-species synthetic gut community. eLife. 2018; 7:e37090. doi: 10.7554/eLife.37090.