Host-specific soil microbes contribute to habitat restriction of closely related oaks (Quercus spp.)

Abstract

Habitat divergence among close relatives is a common phenomenon. Studying the mechanisms behind habitat divergence is fundamental to understanding niche partitioning, species diversification, and other evolutionary processes. Recent studies found that soil microbes regulate the abundance and diversity of plant species. However, it remains unclear whether soil microbes can affect the habitat distributions of plants and drive habitat divergence. To fill in this knowledge gap, we investigated whether soil microbes might restrict habitat distributions of closely related oaks (Quercus spp.) in eastern North America. We performed a soil inoculum experiment using two pairs of sister species (i.e., the most closely related species) that show habitat divergence: Quercus alba (local species) vs. Q. michauxii (foreign), and Q. shumardii (local) vs. Q. acerifolia (foreign). To test whether host-specific soil microbes are responsible for habitat restriction, we investigated the impact of local sister live soil (containing soil microbes associated with local sister species) on the survival and growth of local and foreign species. Second, to test whether habitat-specific soil microbes are responsible for habitat restriction, we examined the effect of local habitat live soil (containing soil microbes within local sister's habitats, but not directly associated with local sister species) on the seedlings of local and foreign species. We found that local sister live soil decreased the survival and biomass of foreign species' seedlings while increasing those of local species, suggesting that host-specific soil microbes could potentially mediate habitat exclusion. In contrast, local habitat live soil did not differentially affect the survival or biomass of the local vs. foreign species. Our study indicates that soil microbes associated with one sister species can suppress the recruitment of the other host species, contributing to the habitat partitioning of close relatives. Considering the complex interactions with soil microbes is essential for understanding the habitat distributions of closely related plants.

Keywords: Quercus; habitat distributions; habitat divergence; host specificity; plant–soil (below‐ground) interactions; soil microbes.

FIGURE 1

Hypothesis of soil microbe‐mediated habitat restriction of sister species. This diagram visualizes the predictions that soil microbes of a local sister constrain habitat distribution of its foreign sister species. (b) Prediction 1—host‐specificity mechanism: local sister live soil collected from adult trees of local sister species (green dashed circles in panel a) increases the fitness of conspecific seedlings due to specialized soil mutualists and tolerance of its own pathogens, while the same soil decreases the fitness of foreign sister's seedlings due to soil pathogens parasitic to the foreign sister and foreign sister's susceptibility. (c) Prediction 2—habitat‐specificity mechanism: local habitat live soil collected from other species co‐occurring within local sister's habitat (brown dashed circles in panel a) differentially affects the fitness of local sister's and foreign sister's seedlings.

FIGURE 2

Bayesian estimates of the effects of soil treatments and host habitat origin (local species vs. foreign species) on oak seedling survival and biomass in a soil inoculum experiment. Sterilized soil is used as a reference level for soil treatment, and foreign species is used as a reference level for host habitat origin. Blue vertical lines represent median estimates of the coefficients derived from the Bayesian models. The truncated distribution outline represents 90% credible intervals (CIs), while the shaded‐light blue region represents 80% CIs. A light‐gray vertical line marks x = 0 in each panel. The tests for Prediction 1 ( P1 ) and Prediction 2 ( P2 ) are highlighted with rectangles. Statistical significance is highlighted with asterisks: ** indicates that 90% CIs of the posterior estimates of the coefficient do not overlap with zero, while * indicates that the 80% CIs do not include zero.

FIGURE 3

Seedling survival probability and aboveground biomass of the local vs. foreign sister species in different soil treatments. Values were derived from the best Bayesian model, using estimated marginal means. Panels (a, c) compare the survival probabilities and aboveground biomass of local sister (green points) when grown in sterilized soil vs. in local sister live soil, and the survival of foreign sister (yellow points) in these two treatments. Panels (b, d) compare the survival probabilities and aboveground biomass of local sister (green points) when grown in sterilized soil vs. in local habitat live soil that does not associate specifically with one host, and the survival of foreign sister (yellow points) in these two treatments. Error bars represent one standard error. Statistical significance, as tested using Bayesian models, is highlighted with asterisks: ** indicates that 90% credible intervals of the posterior distribution of the model coefficient do not overlap with zero. The 90% credible intervals are marked on each panel.