Phylogeny And Phenotypic Evolution Of Epiphyllous Liverwort Genus Cololejeunea (Lejeuneaceae, Hepaticae)

Despite substantial evidence based on molecular data greatly improved our understanding of plant taxonomy and systematics, the species of bryophyte were defined mainly on morphological evidence in current bryophyte taxonomy. Recent molecular studies have indicated that the phylogenetic relationship of derived liverworts lineages can’t be recovered based only on morphology as well as species delimitation. Some hypotheses have been suggested, e.g. the conflict between morphological and molecular evidence might result from parallel evolution and/or convergence and occurrence of cryptic species. Particularly for plant lineages growing in extreme habitats, rampant homoplasy of morphological characters have been frequently recovered in context of the stabilizing selection imposing by extreme habitats. Cololejeunea is one of the largest genera in derived liverworts, with preferrablely growing in epiphyllous habitats. Thus, we hypothesized that morphological-typological species (MTSs) concept could not be used to identify the real extent species in context of limited mophorspace and extreme habitat. Here we used three approaches to explore evidence of phenotypic evolution of Cololejeunea. Firstly, we explored evidences of rampnt homoplasy of morphological characters commonly used in Cololejeunea taxonomy using phylogenetic trees reconstructed based on DNA sequences. Secondly, we inferred the correlation of genetic diversity and morphological disparity. Thirdly, we used DNA taxonomy to compare the number of species defined either by morphology or by genetic distinctiveness based on chloroplast and nuclear markers. The results are as follow:(1) We firstly constructed a comprehensive phylogeny of Cololejeunea and relative genera based on132accessions and three gene regions (rbcL, trnL-F, and nrITS). The molecular data support three major lineages: Colura, Myriocoleopsis, and Cololejeunea. Myriocoleopsis species were resolved in a clade with Cololejeunea minutissima. Aphanolejeunea and Chondriolejeunea were both resolved as putative monophyletic groups, and nested in the Cololejeunea clade. The clade comprising of Cololejeunea angustiflora and Cololejeunea magnilobula was found to be sister to the clade comprising the remaining Cololejeunea species. Currently accepted subgenera of Cololejeunea were paraphyletic or polyphyletic. Seven out of31morphological-based species with two or multiple accessions were not supported as natural taxa. The results found evidence to support the rampant homoplasy of morphological characters used in Cololejeunea taxonomy, which could be used to explain a sum of arguments on Cololejeunea classification based on morphology only. Our data did not provide evidence for ancient vicariance events where major clades and continents were associated, instead some evidence was recovered for recent intercontinental species range expansion.(2) To infer the correlation between phylogenetic and phenotypic distance matrix, the morphological characters/traits distance matrix was generated based on70Cololejeunea MTSs and13gametophyte and life history characters/traits. Whereas, the phylogenetic relatedness were qualified by Patristic and Hamming distance based on three gene markers (rbcL, trnL-F, and nrITS). The result of Mantel test indicated that there was strong correlation between9gametophyte characters and phylogenetic distance matrices, while no phylogenetic signals were discovered in life history relevant traits might resulting from rampant convergence of this characters morphological characters or random phenotypic drift. It is assumed that the degree of phenotypic disparity of characters varies within characters as well as clades. Thus special attentions should be paid on the selection of morpholoigal characters for taxonomy.(3) We studied the occurrence of cryptic species in Cololejeunea lanciloba complex. The DNA taxonomy was applied to calculate the genetic distinctness using the statistic measures of species delimitation. We used two variable region markers(trnL-F and nrITS) and62accessions.We found evidence of13molecular operational taxonomy units with genetic differentiation rather than four MTSs in C. lanciloba complex of interest, suggesting the occurrence of cryptic species in this complex. The accessions of one MTS were nested in more than one clade, while some clades included multiple phenotypes, supporting the limited morphological disparity in some clades whereas occurrence of plasticity in other clades. In this study we compared the approaches of species delimitation based on long-stand monophly and paraphyly and P (AB) and P (BD). The P (AB) and P (BD) measures were proved to be more sensitive to genetic differentiation. As we knew, this was the first time to apply statistic measures in species delimitation. Furthermore, the feasibility of Boykin "tip-to-root" iterative methods in species identification was proved.In total, we constructed a comprehesively phylogeny of Cololejeunea and its closely relatives, outlined the taxonomic frame of this genus, solved several problems of genera and subgenera classification. In the study of phenotypic evolution of morphological characters incorporated with phylogenetic trees, we found the evidence of rampant homoplasy might as a result of stabilizing selection exporting by epiphyllous habitat and limited development options. Rampant homoplasy of morphological characters could be used to explain the occurance of observed correlation between phenotypic and phylogenetic divergences in several clades of Cololejeunea. Cryptic species is one of important challenges for Cololejeunea taxonomy. We used P (AB) and P (RD) measures successfully worked in the species delineation of other organism lineages to delimitate species with genetic distinctiveness and morphological uniformity in Cololejeunea. This study of phenotypic evolution and phylogeny of Cololejeunea was a good model to evoke the future taxonomic work on other derived liverworts particularly for lineages in extreme habitats.