Symbiont abundance is more important than pre-infection partner choice in a Rhizobium – legume mutualism

Abstract

It is known that the genetic diversity of conspecific rhizobia present in root nodules differs greatly among populations of a legume species, which has led to the suggestion that both dispersal limitation and the local environment affect rhizobial genotypic composition. However, it remains unclear whether rhizobial genotypes residing in root nodules are representative of the entire population of compatible symbiotic rhizobia. Since symbiotic preferences differ among legume populations, the genetic composition of rhizobia found within nodules may reflect the preferences of the local hosts, rather than the full diversity of potential nodulating rhizobia present in the soil. Here, we assessed whether Vicia cracca legume hosts of different provenances select different Rhizobium leguminosarum genotypes than sympatric V. cracca hosts, when presented a natural soil rhizobial population. Through combining V. cracca plants and rhizobia from adjacent and more distant populations, we found that V. cracca hosts are relatively randomly associated with rhizobial genotypes. This indicates that pre-infection partner choice is relatively weak in certain legume hosts when faced with a natural population of rhizobia.

Introduction

Rhizobia are bacteria capable of fixing atmospheric nitrogen and establishing an endosymbiotic mutualistic interaction with legumes (Leguminosae). In return for carbohydrates, they provide their legume hosts with a steady supply of nitrogen by converting atmospheric nitrogen (N₂) into ammonia (NH₃). Rhizobia fix nitrogen in root nodules, which are formed as a result of a molecular dialogue between rhizobia and the host plant [6]. Legume roots exude specific flavonoids which can be recognised by rhizobia using NodD proteins [28]. These receptor proteins activate the transcription of the nodulation (nod) genes, which are located on mobile genetic elements, such as plasmids [9]. Recognition of the Nod factors by plant receptors initiates the formation of root nodules. The backbone of a Nod factor is encoded by the nodABC genes, while other nod genes and the noe and nol genes encode molecular ‘decorations’ that determine host specificity [22]. The recognition of flavonoids by NodD and Nod factors by plant receptors are part of a series of so-called “key and lock” steps. The legume host can use these steps to evaluate potential compatible rhizobial partners [22], and to actively select rhizobial genotypes (i.e. pre-infection partner choice) [14], [26].

A considerable amount of research has demonstrated that the rhizobial genetic composition in roots generally differs significantly among legume populations of the same species [24], [27], [29], [30]. This has led to the conclusion that both dispersal limitation and the local environment affect rhizobial genotypic composition [25], [27], [31]. However, it remains unclear whether the rhizobial genotypes residing in root nodules (in situ) are representative of the entire soil population of rhizobia able to nodulate that legume host species. Rhizobia in nodules may indeed have been selected out of a highly diverse gene pool of rhizobia [3], [34], and thus just represent the preferences of the local legume genotypes, rather than the rhizobial diversity present in the soil [14], [26]. As conspecific legumes from different populations have different rhizobial preferences [16], [17], it could be expected that legumes from allopatric populations select different rhizobial genotypes than sympatric legume genotypes when grown on the same soil.

In a previous experiment we tested whether different pairs of Vicia cracca genotypes and Rhizobium leguminosarum genotypes originating from different regions respond differently to different N levels [32]. We found that different V. cracca genotypes provided significantly different opportunities to nodulate for different R. leguminosarum genotypes and that, when N levels increased, a given V. cracca genotype would reduce nodulation more for certain rhizobia than for others. Furthermore, V. cracca genotypes provided significantly more opportunities to nodulate to sympatric rhizobia than to allopatric rhizobia, indicating discriminative behaviour. Here, our objective was to assess whether conspecific legume hosts of different provenances select different rhizobial genotypes than sympatric hosts, when presented with natural soil populations of rhizobia. Our approach was growing allopatric V. cracca plants on soils obtained from natural populations, from which we previously examined the genetic diversity of R. leguminosarum biovar viciae within the nodules of V. cracca [30], [31]. We included V. cracca plants originating from different regions, of which we previously showed that they provide different nodulation opportunities to different genotypes of R. leguminosarum biovar viciae [32]. To assess the effect of geographic provenance and geographic scale on pre-infection partner choice, we compared the composition of nodulating rhizobia from plants grown on soils from adjacent populations (4.6 km distance on average) and from distantly located regions separated by 100–400 km distance.