Phylogenomics of Southern European Taxa in the Ranunculus auricomus Species Complex: The Apple Doesn't Fall Far from the Tree

Abstract

The taxonomic status of many Southern European taxa of the Ranunculus auricomus complex remains uncertain despite this region's proximity to the native ranges of the sexual progenitor species of the complex. We investigated whether additional sexual progenitor species are present in the Mediterranean region. Utilizing target enrichment of 736 single-copy nuclear gene regions and flow cytometry, we analyzed phylogenomic relationships, the ploidy level, and the reproductive mode in representatives of 16 populations in Southern Europe, with additional sequence data from herbarium collections. Additionally, phased sequence assemblies from suspected nothotaxa were mapped to previously described sexual progenitor species in order to determine hybrid ancestry. We found the majority of Mediterranean taxa to be tetraploid, with hybrid populations propagating primarily via apomixis. Phylogenomic analysis revealed that except for the progenitor species, the Mediterranean taxa are often polyphyletic. Most apomictic taxa showed evidence of mixed heritage from progenitor species, with certain progenitor genotypes having mapped more to the populations from adjacent geographical regions. Geographical trends were found in phylogenetic distance, roughly following an east-to-west longitudinal demarcation of the complex, with apomicts extending to the southern margins. Additionally, we observed post-hybridization divergence between the western and eastern populations of nothotaxa in Southern Europe. Our results support a classification of apomictic populations as nothotaxa, as previously suggested for Central Europe.

Keywords: Mediterranean; apomixis; hybridization; polyploidy; species complex.

Figure 1

Locations of individuals used in the study. Locations of progenitor taxa are circled and labeled. Triangles = progenitor taxa, circles = presumed nothotaxa. Colors denote ploidy level. Full location data is listed in File S1.

Figure 2

Species tree including all known sexual progenitor species, hybrid taxa, and outgroups (R. sceleratus and R. pygmaeus). The tree was calculated using ASTRAL-III with 100 multi-locus bootstrap repetitions (MLB). High support (≥90 MLB) is found only in basal nodes. Colors denote geographical origin; asterisks indicate sexual progenitor species.

Figure 3

Heatmap of similarity between progenitor genotype reference individuals and suspected hybrid taxa. From left to right: R. cassubicifolius, R. marsicus, R. flabellifolius, R. notabilis, and R. envalirensis. Color bars on the right (Purple = Spain, Light Blue = Italy) indicate geographic origin. The species tree was calculated using ASTRAL-III as in Figure 2 with tips of non-nothotaxa and low branch length values pruned (see Methods). Here, only the tree topology is shown. Mapping performed in HybPhaser, normalized across reference genotypes.

Figure 4

Multilabeled species tree created using phased sequences from suspected hybrid taxa. Legend is shown to the right of box (A). Color corresponds to the progenitor taxon to which the subgenomic contigs comprising the full-phased sequence were mapped to (legend left). (A) Full tree, showing region displayed in (B) (top, non-transparent rectangle). (B) Clade containing Italian subgenomic sequences mapping to R. notabilis, R. marsicus, and R. flabellifolius (in brackets). Additionally, a clade containing all subgenomic sequences from R. ×baldensis is shown (yellow). The tree was computed using ASTRAL-III with 100 multi-locus bootstrap repetitions. Line thickness and node values indicate multilocus bootstrap support.

Figure 5

Multilabeled species tree created using phased sequences from suspected hybrid taxa. Legend is shown to the bottom of box (A). Color corresponds to the progenitor taxon to which the subgenomic contigs comprising the full-phased sequence were mapped to. (A) Full tree, showing region displayed in (B) (upper-middle, non-transparent rectangle). (B) Clades containing subgenomic sequences mapping to R. notabilis (green) and R. flabellifolius (pink), including Spanish taxa and western Italian R. bovioi which is highlighted. Line thickness and node values indicate multilocus bootstrap support. Computed using ASTRAL-III with 100 multi-locus bootstrap repetitions.

Figure 6

Multilabeled species tree created using phased sequences from suspected hybrid taxa. Legend is shown underneath the bottom of box (A). Color corresponds to the progenitor taxon to which the subgenomic contigs comprising the full-phased sequence were mapped to. (A) Full tree, showing region displayed in (B) (lower-middle, non-transparent rectangle). (B) Clades containing subgenomic sequences mapping to R. envalirensis (light blue) and R. marsicus (red), including Spanish taxa and western Italian R. bovioi which is highlighted. Line thickness and node values indicate multilocus bootstrap support. Computed using ASTRAL-III with 100 multi-locus bootstrap repetitions.

Figure 7

Multilabeled species tree created using phased sequences from suspected hybrid taxa. Legend is shown under the bottom of box (A). Color corresponds to the progenitor taxon to which the subgenomic contigs comprising the full-phased sequence were mapped to. (A) Full tree, showing region displayed in (B) (bottom, non-transparent rectangle). (B) Clade containing Italian subgenomic sequences mapping to R. envalirensis (light blue) and R. cassubicifolius (dark blue); Spanish taxa, and R. ×bovioi (yellow); a Western Italian individual is highlighted. Line thickness and node values indicate multilocus bootstrap support. Computed using ASTRAL-III with 100 multi-locus bootstrap repetitions.