Phylogenomic analysis of the genus Leuconostoc

Abstract

Leuconostoc is a genus of saccharolytic heterofermentative lactic acid bacteria that inhabit plant-derived matrices and a variety of fermented foods (dairy products, dough, milk, vegetables, and meats), contributing to desired fermentation processes or playing a role in food spoilage. At present, the genus encompasses 17 recognized species. In total, 216 deposited genome sequences of Leuconostoc were analyzed, to check the delineation of species and to infer their evolutive genealogy utilizing a minimum evolution tree of Average Nucleotide Identity (ANI) and the core genome alignment. Phylogenomic relationships were compared to those obtained from the analysis of 16S rRNA, pheS, and rpoA genes. All the phylograms were subjected to split decomposition analysis and their topologies were compared to check the ambiguities in the inferred phylogenesis. The minimum evolution ANI tree exhibited the most similar topology with the core genome tree, while single gene trees were less adherent and provided a weaker phylogenetic signal. In particular, the 16S rRNA gene failed to resolve several bifurcations and Leuconostoc species. Based on an ANI threshold of 95%, the organization of the genus Leuconostoc could be amended, redefining the boundaries of the species L. inhae, L. falkenbergense, L. gelidum, L. lactis, L. mesenteroides, and L. pseudomesenteroides. Two strains currently recognized as L. mesenteroides were split into a separate lineage representing a putative species (G16), phylogenetically related to both L. mesenteroides (G18) and L. suionicum (G17). Differences among the four subspecies of L. mesenteroides were not pinpointed by ANI or by the conserved genes. The strains of L. pseudomesenteroides were ascribed to two putative species, G13 and G14, the former including also all the strains presently belonging to L. falkenbergense. L. lactis was split into two phylogenetically related lineages, G9 and G10, putatively corresponding to separate species and both including subgroups that may correspond to subspecies. The species L. gelidum and L. gasicomitatum were closely related but separated into different species, the latter including also L. inhae strains. These results, integrating information of ANI, core genome, and housekeeping genes, complemented the taxonomic delineation with solid information on the phylogenetic lineages evolved within the genus Leuconostoc.

Keywords: 16S rRNA gene; Leuconostoc; average nucletide identity (ANI); biopreservatives; cosmeceutics; phylogenomics.

Figure 1

Heatmap of the ANI similarities matrix reporting pairwise ANI values of 216 Leuconostoc genomes. Blue, ANI > 98%; green, 95% < ANI <98%; gray shades, ANI <95% (77–94.9%). Strain labels are colored according to the groups.

Figure 2

Phylogenomic analysis of Leuconostoc species: (A) Minimum evolution tree of ANI, reporting the 18 proposed clades with values >95%; T, type strain; strains currently ascribed to the species of L. falkenbergense, L. garlicum, and L. inhae are indicated in green, cyan, and yellow, respectively. (B) Heatmap of the mean ANI values between the strains of the groups.

Figure 3

Split decomposition analysis of the phylogenetic trees of ANI, core genome, AAI, rpoA, pheS, and 16S rRNA genes. The analysis was performed utilizing gene alignments for the core genome, rpoA, pheS, and 16S rRNA genes, and distance matrix for ANI. Blue circles indicate the position of groups on split-decomposed trees. *For clarity of representation, the strain L. lactis AV1N that harbored a highly divergent pheS sequence, was not included.

Figure 4

(A) Phylogenetic trees representing ANI distances and the alignment of the core genome, rpoA, pheS, and the 16S rRNA gene. For core genome, rpoA, pheS, and 16S rRNA genes trees, an alignment was produced with Clustal Omega, respectively, and it was used to infer a tree using RAxML. ANI tree was inferred from the ANI distance matrix with R package APE. AAI distance matrix was used to compute the UPGMA tree. Strains are collapsed into their corresponding group. The percentage bootstrap value of each clade is reported in red brackets. *For clarity of representation, the strain L. lactis AV1N that harbored a highly divergent pheS sequence, was not included. (B) Comparison of the ANI tree with the other phylogenetic trees. Normalized similarity score obtained with the generalized Robinson–Foulds metrics Jaccard–Robinson–Foulds (JRF), Mutual Clustering Information (MCI), and Shared Phylogenetic Information (SPI) are reported. (C) JRF scores of the 15 groups identified in the ANI tree.