Microbiome may facilitate host evolution by overcoming nutritional barriers
The changes in microbiome that we identified may affect microbiome function, fish metabolism, and ultimately host fitness. Five of the ten most abundant OTUs in the entire dataset belonged to a single genus, Rhizobiales. Rhizobiales are abundant in the guts of many fish species (Sullam et al. 2012), and decrease as diets become more carnivorous in fish (Sullam et al. 2012) as well as in ants (Russell et al. 2009). Rhizobiales can convert N2to absorbable nutrients, and Russell et al. (2009) suggests hosts could evolve a reliance on these N fixers to supplement a poor quality diet. We found higher Rhizobiales abundance in guts of fish in LP environments compared to those in HP, and the highest abundances in fish translocated 30-60 years ago (old introduction). These populations tended to have diets particularly low in invertebrates (see Fig. S1) but high in detritus (data not shown), which would make them N-limited. In the guppy system, density-dependent factors are thought to underlie predictable evolution of the LP phenotype (Bassar et al. 2010). Adapting to the LP environment involves adapting to higher fish densities and lower per capita resources, which could also potentially result in strong selection for highly efficient gut microbiota. In contrast, HP populations can, and do, select for N-rich invertebrates (Zandonà et al. 2011). Perhaps in LP guts, the abundance of labile C in the form of detritus can support the energy-intensive process of N-fixation from Rhizobiales (Stoll et al. 2007),and satisfy nutrient demand.
It is intriguing that this strategy was more highly adopted in introduced populations, compared to either their sources (HP) or the native (LP) sites that we expected them to converge with (Fig. 5C). The fact that introduced populations had more sources of microbiome transmission (vertical transmission from source population, and horizontal transmission from the new LP environment) may have driven their significantly higher microbiome richness (Fig. 5A). Perhaps this greater gut diversity, representing higher ‘additive variation’ of the hologenome (Zilber-Rosenberg & Rosenberg 2008), was needed to select for nutritionally-beneficial taxa, allowing older translocated populations to take better advantage of this strategy than native fish. This possible mechanism for microbiome-facilitated nutrient acquisition – taken in combination with the observed consistent divergence of microbiome away from HP source populations – points to a nutrition-related role for the microbiome in guppy evolution, even in the absence of a proximal link between gut content and microbiome. Future studies could confirm the causality of this link by manipulating microbiome and documenting changes in host fitness (see Rudman et al. (2019) for work in model organisms). Our study shows that these studies will be much more relevant to wild populations if they consider trait plasticity, realistic dietary patterns, and environmental properties all of which may co-vary with ecotype divergence.