Most insects harbor influential, yet non-essential heritable microbes in their hemocoel. Communities of these symbionts exhibit low diversity. But their frequent multi-species nature raises intriguing questions on roles for symbiont-symbiont synergies in host adaptation, and on the stability of the symbiont communities, themselves. In this study, we build on knowledge of species-defined symbiont community structure across United States populations of the pea aphid, Acyrthosiphon pisum. Through extensive symbiont genotyping, we show that pea aphids’ microbiomes can be more precisely defined at the symbiont strain level, with strain variability shaping ~5 out of 9 previously reported co-infection trends. Field data provide a mixture of evidence for symbiont-symbiont synergies, and symbiont hitchhiking, revealing causes and consequences of these co-infection trends. To test whether within-host metabolic interactions predict common versus rare strain-defined communities, we leveraged the high relatedness of our dominant, community-defined symbiont strains vs. twelve pea aphid-derived Gammaproteobacteria with sequenced genomes. Genomic inference, using metabolic complementarity indices, revealed high potential for cooperation among one pair of symbionts – Serratia symbiotica and Rickettsiella viridis. Applying the expansion network algorithm, through additional use of pea aphid and obligate Buchnera symbiont genomes, Serratia and Rickettsiella emerged as the only symbiont community requiring both parties to expand holobiont metabolism. Through their joint expansion of the biotin biosynthesis pathway, these symbionts may span missing gaps within a multi-party mutualism, within their nutrient-limited phloem-feeding hosts. Recent, complementary gene inactivation, within the biotin pathways of Serratia and Rickettsiella, raise further questions on the origins of mutualisms and host-symbiont interdependencies.

Data were obtained from Sanger sequencing of symbionts harbored by pea aphids, collected from North America. Additional data included published genomes, with various analyses being run on these to assess metabolic complementarity among symbiots.