TESTING MODELS OF HOST-SYMBIONT ORGANIZATION IN EVOLUTIONARILY DISTINCT AND FUNCTIONALLY DIVERSE NITROGEN FIXING TREES AND THEIR SYMBIOTIC BACTERIA

Abstract

The legume-rhizobia symbiosis is an association in which legume plants provide carbon (C) in exchange for nitrogen (N) fixed by bacterial rhizobia housed in plant root nodules. Fixed N is critical for legume plant survival and for vegetation growth within entire ecosystems, as N- rich legume tissues fertilize the surrounding soil. Despite the ecological importance of this symbiosis, there is controversy and confusion about host-symbiont structure, function, and evolutionary stability.In this dissertation, I develop predictions for two fundamentally different models of host- symbiont organization: the conserved taxonomic model (CTM) and the biological market model (BMM). I test the two models empirically across four evolutionarily different plant host- symbiont relationships, leveraging a DNA metabarcoding approach. My results show that the CTM, which predicts low N2-fixing symbiont diversity and little to no change in N2-fixing symbiont composition across different environments, best characterizes the host-symbiont relationships between the rapidly fixing and abundant plant species Inga cocleensis and Tachigali versicolor and their respective symbiotic bacteria. I found that the BMM, which predicts greater N2-fixing symbiont diversity and a change in symbiotic bacteria composition across environmental conditions, best describes the relationship between Andira inermis and Lonchocarpus latifolius and their respective N2-fixing bacteria. From these results, I conclude: (i) a single model does not characterize plant-symbiont relationships of N2- fixing plants; and (ii) ecological success in the local plant community appears to be linked to the host-symbiont relationship predicted by the CTM. I subjected my inferences to two more explicit tests, using I. cocleensis and its N2-fixing symbionts. I first examined the symbiont diversity and composition in I. cocleensis across locations in the forest that represent a strong gradient of resource availability. I followed this with an experiment, in which I directly manipulated the amount of N available to plants in replicated .1-hectare plots. Across both the natural gradient and manipulated resource conditions, I. cocleensis maintained the CTM. I conclude that I. cocleensis maintained a limited and unchanging symbiotic association despite dramatic changes in resource availability. The results of my dissertation further our understanding of the principles that characterize host-symbiont organization in legume hosts.