Gene Co-Expression Highlights Coordination of Microbial Community Behavior

Association networks built from parallel transcriptomic datasets can answer fundamental questions regarding gene control and species interactions in microbial systems.  Here, we combine several RNA-seq datasets examining a model consortium comprised of a cyanobacterial phototroph (Thermosynechococcus elongatus BP1) and a heterotrophic partner (Meiothermus ruber strain A.  During coculture, M. ruber is completely dependent upon T. elongatus for sources of energy, carbohydrates and organic nitrogen sources.  Concurrently, M. ruber removes potentially toxic molecular oxygen from the system and provides key micro/macronutrients to T. elongatus.  Because of these dependencies there is a large amount of interaction, metabolite exchange and crosstalk between the two organisms.  To understand and identify these interactions at the systems level we utilized RNA-seq in conjunction with the Context Likelihood of Relatedness (CLR) program to build a network of genes (nodes) from both species and identify correlations of expression (edges) both within and between species of this consortium.  Resulting networks were then parsed to identify possible points of interaction and the genes and processes that show the greatest number of links between species.  Networks comprised of each species individually showed structures similar to what has been previously observed with genes involved in central functions included photosynthesis, oxidative phosphorylation, cell motility, nitrogen and amino acid metabolism, carbohydrate metabolism and translation grouped separately into tight modules of highly co-expressed genes.  When examining both species networks became much larger (~10x more edges with only ~2x more genes).  When focusing on edges between species, M. ruber genes with the highest number of interspecies links include those involved in amino acid metabolism, cell adhesion and B12 metabolism.  T. elongatus genes with multiple interspecies links include those involved with photosynthesis, carbohydrate metabolism, nitrogen metabolism and terpenoid metabolism.  Through a view of species interactions at the network level these studies identify putative points of interaction in phototroph-heterotroph communities.  These include both central and expected processes (carbon, nitrogen and energy metabolism) as well as additional areas of interaction (adhesion, terpenoid metabolism and B12 metabolism).  Both the developed methodology and conclusions derived from this work are widely applicable to microbial communities for identification of interactions between species and characterization of community functioning as a whole.  Future efforts will be focused on examining these putative interactions further and in constructing biological networks to examine interactions of microbial communities comprised of several heterotrophs supported by a single photoautotroph.