Cooperative Metabolism in a Three-Partner Insect-Bacterial Symbiosis Revealed by Metabolic Modeling

An important factor determining the impact of microbial symbionts on their animal hosts is the balance between the cost of nutrients consumed by the symbionts and the benefit of nutrients released back to the host, but the quantitative significance of nutrient exchange in symbioses involving multiple microbial partners has rarely been addressed. In this study on the association between two intracellular bacterial symbionts, Candidatus Portiera aleyrodidarum and Candidatus Hamiltonella defensa, and their animal host, the whitefly Bemisia tabaci, we apply metabolic modeling to investigate host- symbiont nutrient exchange. Our in silico analysis revealed that > 60% of the essential amino acids and related metabolites synthesized by Candidatus Portiera aleyrodidarumare utilized by the host, including a substantial contribution of nitrogen recycled from host nitrogenous waste, and that these interactions are required for host growth. In contrast, Candidatus Hamiltonella defensa retains most or all of the essential amino acids and B vitamins that it is capable of synthesizing. Furthermore, Candidatus Hamiltonella defensa suppresses host growth in silico by competition with Candidatus Portiera aleyrodidarumfor multiple host nutrients, by suppressing Candidatus Portiera aleyrodidarum growth and metabolic function, and also by consumption of host nutrients that would otherwise be allocated to host growth. The interpretation from these modeling outputs that Candidatus Hamiltonella defensa is a nutritional parasite could not be inferred reliably from gene content alone but requires consideration of constraints imposed by the structure of the metabolic network. Furthermore, these quantitative models offer precise predictions for future experimental study and the opportunity to compare the functional organization of metabolic networks in different symbioses. IMPORTANCE The metabolic functions of unculturable intracellular bacteria with much reduced genomes are traditionally inferred from gene content without consideration of how the structure of the metabolic network may influence flux through metabolic reactions. The three-compartment model of metabolic flux between two bacterial symbionts and their insect host constructed in this study revealed that one symbiont is structured to overproduce essential amino acids for the benefit of the host, but the essential amino acid production in the second symbiont is quantitatively constrained by the structure of its network, rendering it selfish with respect to these nutrients. This study demonstrates the importance of quantitative flux data for elucidation of the metabolic function of symbionts. The in silico methodology can be applied to other symbioses with intracellular bacteria.