Robust Frankia phylogeny, species delineation and intraspecies diversity based on Multi-Locus Sequence Analysis (MLSA) and Single-Locus Strain Typing (SLST) adapted to a large sample size.

Diazotrophic Actinobacteria of the genus Frankia represent a challenge to classical bacterial taxonomy as they include many unculturable strains. As a consequence, we still have a poor understanding of their diversity, evolution and biogeography. In this study, a Multi-Locus Sequence Analysis (MLSA) using atpD, dnaA, ftsZ, pgk, and rpoB loci was done on a large set of cultured and uncultured strains, compared to 16S rRNA and correlated to Average Nucleotide Identity (ANI) from available Frankia genomes. MLSA provided a robust resolution of Frankia genus phylogeny and clarified the status of unresolved species and complex of species. The robustness of single-gene topologies and their congruence with the MLSA tree were tested. Lateral Gene Transfers (LGT) were few and scattered, suggesting they had no impact on the concatenate topology. The pgk marker - providing the longest sequence, highest mean genetic divergence and least occurrence of LGT - was used to survey an unequalled number of Alnus-infective Frankia - mainly uncultured strains from a broad range of host-species and geographic origins. This marker allowed reliable Single-Locus Strain Typing (SLST) below the species level, revealed an undiscovered taxonomical diversity, and highlighted the effect of cultivation, sporulation phenotype and host plant species on symbiont richness, diversity and phylogeny.