Multilocus analysis of the Exophiala jeanselmei clade containing black yeasts involved in opportunistic disease in humans

To confirm species delimitations in the 'jeanselmei-clade' in the Chaetothyriales, four independent markers were analysed, and phylogenetic trees were reconstructed using different algorithms. Reproductive isolation within the complex and reproductive modes in the species involved were determined, in order to explore specific borderlines using Genealogical Concordance Phylogenetic Species Recognition (GCPSR). Sequences of the Internal Transcribed Spacer (ITS), partial Translation Elongation Factor 1-alpha (TEF1), beta-Tubulin (BT2), and Actin (ACT1) genes were analysed for a set of 81 strains. Phylogenetic reconstruction was performed using neighbour-joining (NJ), maximum parsimony (MP), maximum likelihood (ML) and Bayesian analysis (BA) to evaluate the concordance of topologies obtained under different optimization criteria. Incongruence among lineages was detected by Partition Homogeneity Test (PHT). Most Exophiala species analysed in this study seemed to be preponderantly clonal. A group of species around Exophiala spinifera showed a tendency toward human pathogenicity, whereas species around E. bergeri were often associated with environments rich in hydrocarbons. Though the tree topologies were not completely identical when different algorithms were used, the four gene lineages were nearly congruent and the trees of the combined multilocus data set did not provide more phylogenetic information than those of separate data sets. ITS data are sufficient for identification of species in the 'jeanselmei-clade'.