Estimating Rhizobial Fitness During Legume Symbiosis: Enriching Viable Undifferentiated Bacteria from Root Nodules

Advances in understanding the evolutionary ecology of the rhizobia-legume mutualism have been constrained by methodological limitations in efficiently measuring relative strain frequencies alongside measurements of absolute population sizes of rhizobia living in nodules. To examine strain competition in natural and agricultural ecosystems that harbor multiple strains of rhizobia, an increasing number of manipulative and observational studies have recently begun to examine dozens or hundreds of strains simultaneously. Assessing the competitive fitness of multiple strains in legume nodules requires, first, processing pools of dozens to hundreds of nodules to overcome the stochasticity of nodule formation; second, focusing on the reproductively viable rhizobial population, since this trait represents rhizobia's reproductive success in nodules and is pivotal for evolutionary interpretations. Our approach has been optimized in the Medicago truncatula-Sinorhizobium meliloti system, where rhizobia induce the formation of indeterminate nodules that harbor two subpopulations: terminally-differentiated bacteroids and undifferentiated rhizobia that retain reproductive viability. This protocol has also been used for other legumes with terminally-differentiated bacteroids, such as pea and vetch, as well as for those with non-terminally differentiated bacteroids, such as soybean and cowpeas. The protocol we present enables rapid and reproducible homogenization of pools containing hundreds of nodules using a tissue homogenizer. We also enrich for undifferentiated rhizobia using two centrifugation steps: first, a low-speed centrifugation to deplete nodule debris and large, endoreduplicated, terminally-differentiated bacteroids, followed by a high-speed centrifugation to pellet the remaining undifferentiated rhizobia. The pellet can later be used for DNA extraction, followed by whole-genome or amplicon sequencing, and then downstream analysis to estimate strain fitness. Finally, we include an optional step for a reliable, reproducible system for nodule imaging, which is especially useful for quantifying nodule abundance and studying morphological variation. ​.