Systematics and biogeography of the pantropical genus Manilkara Adans. (Sapotaceae)
Mechanisms for the generation of biodiversity in species-rich biomes such as rain forests remain unclear. Molecular phylogenies using DNA sequence data, calibrated with a temporal dimension offer a means of addressing this question, enabling the testing of different hypotheses on biogeographic histories and causes of diversification. Manilkara is a genus of trees in the Sapotaceae consisting of ~79 species distributed throughout the tropics (30 South and Central American, 35 African and 14 Southeast Asian). This species diversity in all major tropical regions of the globe makes it an ideal candidate for in-depth biogeographic studies. Maximum parsimony and Bayesian analyses of nuclear (ITS) and chloroplast (rpl32-trnL, rps16-trnK and trnS-trnFM) sequences were used to reconstruct a species level phylogeny of Manilkara and related genera in the tribe Mimusopeae. Manilkara, as currently defined, is not monophyletic due to the placement of three Asian taxa (M. fasciculata, M. dissecta and M. udoido), which are more closely related to the Madagascan genera Labourdonnaisia and Faucherea than to Manilkara s.s. and need to be re-circumscribed in a new genus. Letestua is nested in Manilkara and the genera Faucherea and Labourdonnaisia are not monophyletic. Nuclear and chloroplast datasets were mostly congruent, however, three instances of hard incongruence were demonstrated, suggesting chloroplast capture events. Bayesian analyses of ITS sequences using a relaxed molecular clock calibrated with fossils, focused on testing biogeographical hypotheses on the origin of Manilkara’s pantropical disjunct distribution and spatio-temporal diversification patterns on each continent. Mimusopeae, originated during the Eocene ~46-57 Ma and fossil evidence supports its existence in the boreotopical region of the northern hemisphere during this time. This suggests that the tribe may have evolved there and found refuge in Africa when Oligocene climatic cooling made higher latitudes uninhabitable for megathermal taxa. The subtribe Manilkarinae was resolved as ~42-36 Myo. These ages fall on the Eocene-Oligocene boundary and the crown node age coincides with the onset of Oligocene cooling and the closing of the boreotropical route. The genus Manilkara is estimated to have evolved ~36-33 Ma. The current distribution of the genus could not, therefore, have been the result of Gondwanan vicariance or migration through the boreotropics, but results instead support long distance dispersal as an important factor influencing the distribution of the group. Resolution along the backbone of the phylogeny is weak and the area of origin is, therefore, difficult to determine. However, all sister taxa to Manilkara are African and this suggests that the most likely explanation is an African origin for the genus with subsequent inter-continental dispersal during the Miocene. Manilkara spread from Africa to the Neotropics and Asia via at least three separate long distance dispersal events. A single lineage dispersed to the Neotropics ~27-21 Ma and spread across the Isthmus of Panama before its closure. Another lineage dispersed to Southeast Asia ~30-25 Ma from mainland Africa and subsequently diversified throughout the region. A third dispersal from Madagascar to the Sahul Shelf, occurred ~31-16 Ma in the M. fasciculata/dissecta/udoido lineage. In South America, diversification is consistent with both aridification and the rearrangement of drainage patterns in the Amazon basin as a result of Andean orogeny. The Atlantic coastal forest clade and the Amazonian clade of Manilkara split from one another ~14 Ma, at approximately the same time as the dry biomes of the Cerrado and Caatinga were forming between them. In Africa diversification coincides with Tertiary cycles of aridification and uplift of the east African plateaux. In Southeast Asia Wallace’s Line did not affect the dispersal of Manilkara. Instead, the limiting factor was the appearance of land in New Guinea ~10 Ma, which coincides with the dispersal and establishment of new taxa east of Wallace’s Line. Spatio-temporal patterns of diversification in Manilkara were compared to those of 34 other wet tropical genera which have intercontinental disjunctions. Ages of disjunctions ranged from the Eocene to the Pliocene, indicating that compilation of the tropical rain forest biome is a dynamic process which has been occurring throughout the Tertiary. Recent migration via long distance dispersal is a significant phenomenon in biome construction. Geo-climatic events have also been shown to be important drivers of diversification in all continental regions.