Divergent pattern of nuclear genetic diversity across the range of the Afromontane Prunus africana mirrors variable climate of African highlands.

BACKGROUND AND AIMS: Afromontane forest ecosystems share a high similarity of plant and animal biodiversity, although they occur mainly on isolated mountain massifs throughout the continent. This resemblance has long provoked questions on former wider distribution of Afromontane forests. In this study Prunus africana (one of the character trees of Afromontane forests) is used as a model for understanding the biogeography of this vegetation zone. METHODS: Thirty natural populations from nine African countries covering a large part of Afromontane regions were analysed using six nuclear microsatellites. Standard population genetic analysis as well as Bayesian and maximum likelihood models were used to infer genetic diversity, population differentiation, barriers to gene flow, and recent and all migration among populations. KEY RESULTS: Prunus africana exhibits strong divergence among five main Afromontane regions: West Africa, East Africa west of the Eastern Rift Valley (ERV), East Africa east of the ERV, southern Africa and Madagascar. The strongest divergence was evident between Madagascar and continental Africa. Populations from West Africa showed high similarity with East African populations west of the ERV, whereas populations east of the ERV are closely related to populations of southern Africa, respectively. CONCLUSIONS: The observed patterns indicate divergent population history across the continent most likely associated to Pleistocene changes in climatic conditions. The high genetic similarity between populations of West Africa with population of East Africa west of the ERV is in agreement with faunistic and floristic patterns and provides further evidence for a historical migration route. Contrasting estimates of recent and historical gene flow indicate a shift of the main barrier to gene flow from the Lake Victoria basin to the ERV, highlighting the dynamic environmental and evolutionary history of the region.

Bioactive constituents in Prunus africana: geographical variation throughout Africa and associations with environmental and genetic parameters.

Prunus africana--an evergreen tree found in Afromontane forests--is used in traditional medicine to cure benign prostate hyperplasia. Different bioactive constituents derived from bark extracts from 20 tree populations sampled throughout the species' natural range in Africa were studied by means of GC-MSD. The average concentration [mg/kgw/w] in increasing order was: lauric acid (18), myristic acid (22), n-docosanol (25), ferulic acid (49), ß-sitostenone (198), ß-sitosterol (490), and ursolic acid (743). The concentrations of many bark constituents were significantly correlated and concentration of n-docosanol was highly significantly correlated with all other analytes. Estimates of variance components revealed the highest variation among populations for ursolic acid (66%) and the lowest for ß-sitosterol (20%). In general, environmental parameters recorded (temperature, precipitation, altitude) for the samples sites were not correlated with the concentration of most constituents; however, concentration of ferulic acid was significantly correlated with annual precipitation. Because the concentration of compounds in bark extracts may be affected by tree size, the diameter of sampled plants at 1.3m tree height (as proxy of age) was recorded. The only relationship with tree diameter was a negative correlation with ursolic acid. Under the assumption that genetically less variable populations have less variable concentrations of bark compounds, correlations between variation parameters of the concentration and the respective genetic composition based on chloroplast and nuclear DNA markers were assessed. Only variation of ß-sitosterol concentration was significantly correlated with haplotypic diversity. The fixation index (F(IS)) was positively correlated with the variation in concentration of ferulic acid. Principal Components Analysis (PCA) indicated a weak geographic pattern. Mantel tests, however, revealed associations between the geographic patterns of bioactive constituents and the phylogenetic relationship among the populations sampled. This suggests an independent evolution of bark metabolism within different phylogeographical lineages, and the molecular phylogeographic pattern is partly reflected in the variation in concentration of bark constituents. The results have important implications for the design of strategies for the sustainable use and conservation of this important African tree species.