Convergent evolution of <i>NFP</i> -facilitated root nodule symbiosis

The origin and phylogenetic distribution of symbiotic associations between nodulating angiosperms and nitrogen-fixing bacteria have long intrigued biologists. Recent comparative evolutionary analyses have yielded alternative hypotheses: a multistep pathway of independent gains and losses of root nodule symbiosis vs. a single gain followed by numerous losses. A detailed reconstruction of the history of genes involved in signaling between nitrogen-fixing bacteria and potential hosts, particularly lipo-chitooligosaccharide (LCO) signaling, is needed to distinguish between these hypotheses. LCO recognition by plants involves the Nod Factor Perception (<i>NFP</i>) gene family; in the legume model <i>Medicago truncatula</i> (Fabales), <i>MtNFP</i> is essential for establishing rhizobial symbiosis. Here, we document convergent evolution of <i>NFP</i>, indicating multiple origins of LCO-driven symbiosis. In contrast to previous models that explain the recruitment of <i>NFP</i> via a single duplication in the ancestor of the nitrogen-fixing clade, our phylogenomic and synteny results suggest this duplication does not span the entire clade. Tandem duplication in a common ancestor of Cucurbitales and Rosales resulted in the <i>NFP1</i> and <i>NFP2</i> groups. In contrast, the phylogenetically closest paralog of <i>MtNFP</i> is <i>MtLYR1</i>, located on a different chromosome within a large syntenic block. All available data indicate that a large-scale duplication resulted in <i>MtNFP</i> and <i>MtLYR1</i>, likely corresponding to a whole-genome duplication in an ancestor of subfamily Papilionoideae of Fabaceae. We show that <i>MtNFP</i> and the <i>NFP2</i>-like group are not orthologous, indicating multiple independent gains of <i>NFP</i>-based LCO signaling. This molecular convergence provides a possible mechanism for multiple gains of root nodule symbiosis across the nitrogen-fixing clade.