Microbial feedbacks are no match for genotypic control of flowering traits in a model annual plant

Background/Question/Methods

Interactions between genotype and microbiome often shape host traits. In plants, microbial symbioses influence plant morphology, phenology, and growth. The best known of these are plant-soil feedbacks, whereby a plant changes the local soil microbial community in a way that alters the growth of subsequent plants. Although these feedbacks are often important to host growth, we know much less about their effect on plant flowering traits, many of which have a clear genetic basis and whose effects are more proximate to plant fitness. For example, flowering time and flower size shape fitness in many angiosperms and several genes driving these phenotypes have been identified. Here, we sought to test whether plant-soil feedbacks additionally shape flowering traits. We hypothesized that genotypes with contrasting flowering traits would change the soil microbiomes in a way that shifted the subsequent flowering traits of these genotypes. We grew two well-characterized lines of Mimulus guttatus in a naïve soil microbiome. Next, we used these genotype-trained microbiomes as soil inoculum for new plants in a second round, crossing them in a fully factorial way. We then measured biomass, flowering traits, and final microbiomes.

Results/Conclusions

Plant genotype was the dominant force controlling flower width (F1,33 = 25.08, p < 0.001), pistil length (F1,33 = 26.82, p < 0.001), and days to first flower (F1,38 = 27.21, p < 0.001). The genotype which trained soil inoculum did not affect flower width or pistil length. Plants (of both genotypes) grown in soil trained by the late flowering genotype (IM-62) did flower earlier (F1,38 = 6.01, p = 0.019). This genotype grew larger than the rapid flowering IM-767 genotype (F1,75 = 49.45, p < 0.001). Growth was consistent with positive feedbacks for the rapid flowering IM-767 genotype. Ongoing microbiome analyses will test for links between microbial differentiation and these traits. Control plants (sterilized inoculum) did have wider flowers (F1,75 = 6.29, p =0.014) and the IM-767 genotype had longer pistils in controls (F1,75 = 9.25, p =0.003). Our results confirm that genotypic effects dominate flowering traits and growth in this rapidly growing annual. Soil microbes did play a role in flowering, but these effects often depended on live or sterile soil only, rather than compositional changes consistent with plant-soil feedbacks. Multiple plant generations or longer-lived plants may be an important prerequisite for any soil feedbacks to flowering traits.