Exploring the diversity of ascomycete fungi : evolution of mating systems in Pleospora and discovery of new lineages in the Dothideomycetes

This thesis explored aspects of diversity in the ascomycete genus Pleospora and the new family Aliquandostipitaceae. In Pleospora, the main focus was on mating system evolution. I found that homothallism in Pleospora evolved in three different ways from heterothallism. One origin of homothallism resulted from a horizontal transfer across lineages involving a MAT locus. The approach chosen to investigate mating system evolution in Pleospora was based on delimiting the polyphyletic genus Pleospora to Pleospora sensu stricto, inferring a robust species phylogeny of Pleospora sensu stricto and cloning and examining the master regulator locus of sexual development in ascomycetes, the MAT locus. Conclusions were then drawn by integration of the total evidence. Research of mating system evolution in Pleospora was initiated by confirmation of the monophyly of Pleospora isolates with Stemphylium asexual states. This group contained the type of Pleospora, and was thus called Pleospora sensu stricto. Contributing towards a monophyletic genus Pleospora, the marine species P. gaudefroyi collected in British Columbia, and lacking a Stemphylium asexual state, was transferred to the new genus Decorospora. Phylogenetic 18S rDNA analyses and Shimodaira-Hasegawa and Kishino-Hasegawa tests rejected the null hypothesis of monophyly for the two taxa. Instead, Pleospora gaudefroyi grouped distantly from P. herbarum, at the base of the family Pleosporaceae confirming the importance of a Stemphylium asexual state for the definition of Pleospora sensu stricto in this case. Since no existing genus was available to accommodate P. gaudefroyi, the new genus Decorospora had to be erected. To generate a species phylogeny of Pleospora sensu stricto, 114 ingroup taxa with Stemphylium asexual states were used. Phylogenetic analyses based on ITS, GPD, EF-1alpha and vmaA-vpsA DNA sequences with four different algorithms showed that Pleospora sensu stricto contained 22 phylogenetic species. Morphological species generally correlated well with phylogenetic species, except for the type P. herbarum whose phylogenetic species contained the four additional morphological species P. alfalfae, S. vesicarium, P. tomatonis and P. sedicola. Another conflict between morphological and phylogenetic species concept could arise in the phylogenetic species S. xanthosomatis that also contained an isolate of S. lycopersici, which, however, was not certain to represent the type of the species. Three phylogenetic species, one of which was exclusively collected in British Columbia, did not contain any morphological species and might be new to science. The protein-coding gene EF-1 alpha, parts of which where used for phylogenetic inference in Pleospora sensu stricto, contained an unusual intron in the phylogenetic species S. lancipes, S. trifolii and Stemphylium sp. strain P246. The intron was up to 1678 bp long, more than 1400 bp longer than introns in other species of Pleospora, it encoded a protein and was delimited at the 5'-end by the non-canonical splice site GGT, instead of GT. The ORF encoded by the introns of all three Pleospora species was most similar to a hypothetical zinc finger protein from the filamentous ascomycete Gibberella zeae. In case of experimental verification, this would be the first report of a 'parasitic' intron splice site in fungi. To continue investigations of mating system evolution in Pleospora sensu stricto, the MAT locus was PCR amplified using primers targeting the conserved motives alpha box and HMG box on the MAT1-1 and MAT1-2 genes respectively. A chromosome walking approach was then used to recover the MAT flanking regions and neighboring genes. It was shown that Pleospora sensu stricto contained three kinds of MAT regions, comprising either a MAT1- 1 or MAT1-2 idiomorph, or both MAT1-1 and MAT1-2 idiomorphs fused end to end, with the inverted MAT1-1 gene placed between ORF1 and MAT1-2. The genes flanking the MAT regions were ORF1 upstream, and BGL1 downstream of the idiomorphs, as in the close relative Cochliobolus. The idiomorphs of Pleospora sensu stricto were well delimited upstream of the MAT genes, terminating 16 or 17 amino acids inside ORF1 for respectively MAT1-1 and MAT1-2. The downstream boundary of the idiomorphs of Pleospora sensu stricto was poorly defined. There were no important differences between MAT genes from fused and separate MAT regions of Pleospora sensu stricto. MAT1-1 genes from both separate and fused regions were 1193 bp in length and contained one intron of 53 bp inserted at position 218. MAT1-2 genes were all 1093 bp long and comprised one intron of 55 bp inserted at position 491. The fused MAT regions of Pleospora sensu stricto consisted of MAT1-1 and MAT1-2 idiomorphs fused end to end, with MAT1-1 and flanking regions inverted. The gene arrangement found in the fused MAT regions was hypothesized to have evolved from ancestors with separate MAT regions by a crossover following the inversion of MAT1-1. The crossover was possibly facilitated by a short, 4 bp long stretch of DNA sequence similarity between MAT1-1 and MAT1-2 idiomorphs, resulting from the inversion of MAT1-1 plus flanking regions. MAT locus architecture and phylogenetic species correlated well. All phylogenetic species of Pleospora sensu stricto with more than one isolate contained both MAT1-1 and MAT1-2 isolates with separate MAT regions, or only isolates with fused MAT regions. MAT regions also correlated with mating systems in Pleospora sensu stricto. Species with fused MAT regions were homothallic, whereas species with separate MAT regions were heterothallic except one group that was ho mothallic. In all homothallic isolates with separate MAT regions, only MAT1-1 was detected. To evaluate the number of times a switch between mating systems occurred in the evolution of Pleospora sensu stricto, the species phylogeny was compared to the MAT phylogenies, in conjunction with the results from structural analyses of the MAT loci. MAT data of Pleospora sensu stricto suggested a single origin of the fused MAT regions from separate MAT regions. The single origin was supported by the monophyly of the fused MAT regions in MAT phylogenies, as well as the complicated structure of the fused MAT regions that was unlikely to have evolved twice independently. Whereas combined MAT evidence suggested a single origin of the fused MAT regions, the species phylogeny suggested at least two independent origins of the fused MAT regions. The conflicting evidence between MAT data and the Pleospora sensu stricto species phylogeny was consistent with a single origin of the fused MAT regions followed by a horizontal transfer across lineages, by sexual or asexual means. The one time evolution and subsequent horizontal transfer of the fused MAT region constituted two different evolutionary origins of the homothallics with fused MAT regions. A third origin of homothallism in Pleospora sensu stricto was in the group with a separate MAT locus containing a forward-oriented MAT1-1 gene. Homothallism in this case may be do to unknown mutations in other than the MAT genes, as possibly in the homothallic Neurospora africana. The last part of my thesis dealt with two new species in the new family Aliquandostipitaceae. Both Aliquandostipite khaoyaiensis and A. sunyatsenii were collected in Asia, and comprised several features not previously reported in ascomycetes. Novel morphological features were the presence of two types of fruitbodies side by side on the substrate, and the widest hyphae reported in ascomycetes. Overall morphological appearance suggested that species of Aliquandostipite were related to members of the order Pleosporales. However, molecular analyses of the 18S rDNA showed that species of Aliquandostipite did not belong to the Pleosporales, but instead grouped with uncertain affinity in the class Dothideomycetes.Science, Faculty ofBotany, Department ofGraduat