COS 31-9
Local variation in fungal endophyte communities among lowland tropical rainforest trees

Tuesday, August 12, 2014: 10:50 AM
Carmel AB, Hyatt Regency Hotel
John B. Vincent, Department of Plant Biology, University of Minnesota, St. Paul, MN
Georgiana May, Ecology, Evolution, and Behavior, University of Minnesota, Saint Paul, MN
George D. Weiblen, Bell Muserum and Department of Plant Biology, University of Minnesota, St. Paul, MN
Background/Question/Methods

A growing body of research has shown that fungal endosymbionts influence plant function and play a role in shaping plant-insect interactions. However, we lack a fundamental understanding of processes influencing the distribution endophyte communities. We tested competing hypotheses of dispersal limitation and host specificity by sampling fungal endophytes from multiple individuals of nine rainforest tree species in a spatially explicit 50 ha forest plot at Wanang, Papua New Guinea. Cultured endophytes were sequenced for ribosomal DNA (ITS) to establish 98% similar molecular operational taxonomic units (MOTUs). Differences in MOTU richness and endophyte abundance among trees were uncovered by analysis of variance (ANOVA). Non-metric multidimensional scaling (NMDS) was performed to visualize variation in fungal endophyte community composition within and between sampled host species. Multivariate statistical methods (PERMANOVA and PERMDISP) were used to model and explore possible sources of variation in endophyte communities. Mantel tests for spatial and phylogenetic structure in fungal endophyte composition were performed. Lastly, bipartite host association graphs were used to explore quantitative distribution of specialist and generalist endophytes.

Results/Conclusions

MOTU richness and abundance did not vary between species. NMDS showed endophyte community compositions to be essentially homogenous, with little differentiation by host species. Results show that some tree species harbor significantly more variable assemblages of endopytes than others. Evidence of host phylogeny, leaf traits, and spatial distributions producing this pattern was equivocal, possibly obscured statistically by large differences in group dispersions. Mantel tests showed no evidence of phylogeny structuring the endophyte community. Likewise, Mantel tests showed no evidence of space structuring the endophyte community in trees in general nor within species or genera. This lack of spatial structure at any level suggests that dispersal limitation does not play a large role in determining endophyte distribution at this scale. Bipartite graphs revealed an endophyte community composed of few abundant generalist MOTUs and many very rare specialist endophytes. Our results suggest that in a homogenous environment at distances of less than one kilometer there appears to be little variation in foliar endophyte communities in rainforest trees despite large differences in host phylogeny, leaf traits, and tree spatial distribution patterns. This conclusion is consistent with the notion that host specificity plays a weak but relatively more important role than dispersal limitation in shaping the distribution of fungal endophyte communities.