Pre-Pleistocene origin of phylogeographical breaks in African rain forest trees: New insights from Greenwayodendron (Annonaceae) phylogenomics

Aim Palaeoecological records indicate that Pleistocene glaciations affected the African rain forest, probably causing its fragmentation, which could explain phylogeographical breaks documented in many tree species. This refuge hypothesis was further tested through species distribution models, hindcasting persistence during the Last Glacial Maximum. However, previous studies failed to estimate with sufficient precision the divergence time between phylogeographical entities to confirm their Pleistocene origin. Developing genomic tools on a representative tree of mature rain forests, we test if parapatric genetic clusters documented in widespread tree species can be interpreted as the legacy of past population fragmentation during the last glacial period(s). Location Tropical Africa, Guineo-Congolian forests. Taxon Greenwayodendron (Annonaceae). Methods To further test the Pleistocene refuge hypothesis by molecular dating, we sequenced the plastome of 145 individuals of the shade-tolerant rain forest tree Greenwayodendron suaveolens and congeneric species, and genotyped the same samples using nuclear microsatellites to identify genetic clusters. Results Five plastid phylogroups of G. suaveolens occur in parapatry throughout Central Africa, following a spatial pattern generally congruent with genetic clusters. Four of them diverged 3.5-4.5 Ma, whereas the fifth one, located in the Cameroon volcanic line (CVL), diverged 8.3 Ma, in the range of divergence times between Greenwayodendron species, highlighting the key role of the CVL in hosting ancient lineages. Within phylogroups, most nodes were dated from 0.9 to 3.2 Myr and a correlation between haplotype divergence and spatial distance was still perceptible, indicating a slow population dynamic. Main conclusions The phylogeographical structures of Central African trees probably established during the Pliocene or early Pleistocene, and while they might have been reinforced during subsequent glacial-interglacial cycles, interglacial phases did not lead to genetic homogenization. Therefore, interpreting phylogeographical patterns of African trees must account for a much deeper past than previously assumed, and cannot be limited to the last glacial period.