Non-additive interactions among symbionts are context-dependent

Background/Question/Methods

Multiple symbionts inhabiting the same host can interact in ways that influence whether the symbiosis is mutualistic or antagonistic. In the absence of interactions, we expect the effects of each symbiont on the plant to be additive. However, we expect that interspecific interactions between symbionts can be non-additive when competition or complementarity occurs between ecologically similar or dissimilar species, respectively. To test these ideas, we examined how plants were affected by individuals and pairs of horizontally transmitted foliar fungal endophytes selected to represent different degrees of fungal trait similarity based on their host effects. Specifically, we inoculated the grass Panicum virgatum with six two-species endophyte mixtures composed of fungi ranging from high to low trait similarity, monocultures of each fungus, or a fungus-free inoculum. In addition, we tested the generality of these interaction outcomes in different environments by imposing either drought or well-watered conditions. Plants were grown in the greenhouse for one month and plant growth (height, tillers) and days to first wilt were measured weekly. Non-additive effects were examined via overyielding. Trait similarity was analyzed with linear regression. 

Results/Conclusions

We found predominantly non-additive effects in the six-endophyte mixtures. Under drought conditions, three pairs increased and one pair decreased plant growth beyond what was expected, whereas only two mixtures had additive effects on the plant. The same patterns were observed under well-watered conditions, but the mixtures that were synergistic and antagonistic differed. Competition likely occurred in drought conditions where increasing trait similarity between endophytes was associated with decreasing plant growth. Surprisingly, however, the opposite was observed under well-watered conditions, where increasing trait similarity between fungi resulted in greater plant growth. It is possible that our measurement of trait similarity did not accurately reflect functional similarity in symbiosis. For example, stress responses may depend exclusively on traits related to fungal stress tolerance such as production of osmolytes or secondary metabolites. Alternatively, there may be trait tradeoffs that we did not capture with our approach. Nevertheless, interactions between endophytic fungi clearly determine the outcome of symbiosis and depend on the level of environmental stress.  Because the majority of plant symbioses involve one host colonized by multiple symbionts, it is important to move beyond single symbiont studies to develop a predictive framework for symbiont interactions.