Phylogenetic Study Of Epimedium Genus Based On Chloroplast Genome And RDNA Sequences

"Herba Epimedii" was firstly recorded in Shennong's Materia Medica Classic(Shennong Bencao Jing)",and has been used as a kidney tonic and antirheumatic herb"Herba Epimedii" in traditional Chinese medicine for thousands of years.Modern research has proven that it has many remarkable effects,such as anti-tumor,etc.Herba Epimedii is derived from the dried leaves of several plant species that belongs to the genus Epimedium of the family Berberidaceae.However,their similar morphology has sowed serious confusion.Therefore,the systematic taxonomic research of genus Epimedium is essential for accurate identification of medicinal materials.There are approximately 60 Epimedium species in the world,and of which about 50 species are endemic to China.In 2002,based on the geographical distribution,C-banding of chromosomes and flower and leaf morphology,Steam(a British taxonomist)proposed the most acknowledged classification system,which firstly divided Epimedium genus into two subgenera(Subgenus Rhizophyllum and Subgenus Epimedium).Subgenus Epimedium is furhter subdivided into four sections,namely Section Epimedium,Section Macroceras,Section Polyphyllon,and Section Diphyllon.Among them,Section Diphyllon,the largest section,which entirely comprises of Chinese species,is further divided into four series(Series Campanulatae,Series Davidianae,Series Dolichocerae,and Series Brachycerae).Although various methods had been applied by many scholars to resolve the phylogenetic relationships,great controversies still exist in these research results.Due to the limitation of sampling and insufficient resolution of molecular markers,the phylogenetic relationship of Epimedium still remains controversial and the most serious chaos is within section Diphyllon.In this study,the samples of 56 Epimedium species(58 populations totally)which represent all subgenera,sections,and series of Epimedium genus,and as well as Vancouveria chrysantha from sister genus Vancouveria were obtained for re-sequencing with a sequencing depth of more than 10×.Consequently,59 chloroplast genomes and 53 nuclear rDNA sequences was successfully assembled.Through comparative analysis,the characteristics of Epimedium chloroplast genomes were revealed.The phylogenetic relationship within Epimedium was reconstructed based on the maternally inherited chloroplast genomes and the biparentally inherited rDNA sequences.Furthermore,the chloroplast genome and rDNA sequence of two natural hybrids,and together with their parents,were analyzed.Conclusions of this study were drawn as follows:1.Phylogenetic analysis of Epimedium chloroplast genomes:The full length of 58 chloroplast genomes from 56 Epimedium species ranged from 156,078 bp to 159,716 bp.They all possess a typical quadripartite structure with high similarity in gene order and the number of introns.Among them,54 Epimedium chloroplast genomes contain 112 unique genes,including 78 CDS genes,30 tRNA genes and 4 rRNA genes.Four species(E.flavum,E.latisepalum,E.pinnatum and E.perraladerianum)have 111 unique genes due to the loss of rpl32 gene,and the length and location of gene loss varied,which is inferred to be caused by three independent gene deletion events.The gene trnQ-UUG duplicated in the LSC region occurred for all the 59 Epimedium species and Vancouveria chrysantha,which did not occur for other genus the family Berberidaceae.Several expansion and contraction types were detected at the IR boundaries and there were 16-27 genes duplicated in the IR regions.The genus Epimedium was firstly divided into two major clades in the phylogenetic tree based on the protein-coding regions(CDS)of chloroplast genomes:section Macroceras firstly clustered with section Epimedium(BS=100,PP=1),and then clustered with Subgenus Rhizophyllum(BS=91,PP=1);and for the other clade,section Diphyllon formed a sister relationship with section Polyphyllon(BS=88%,PP=1),indicating that subgenus Rhizophyllum and subgenus Epimedium are not monophyletic groups.In the largest section which was entirely formed by Chinese Epimedium species,phylogenetic relationships are closely related to geographical distribution.The Chinese Epimedium species(Section Diphyllon)can be separated into two major subclades along the central stretch of Sichuan Basin,namely the western and central taxa,which corresponds to the two major diversity centers of Epimedium in China.The result was also supported by the distinct differences about the size of the chloroplast genomes and the gene types at the expanded IR region boundary.The fact that chloroplast genome phylogeny was strongly related to geographical distribution rather than morphology indicated that Epimedium species has experienced complicated historical and modern interspecific hybridization events,leading to frequent occurrence of chloroplast capture events.This is an important reason for the intractable interspecific relationship of Epimedium species,especially within Section Diphyllon,and shows the limitation of chloroplast genome-based phylogeny.2.Phylogenetic analysis of Epimedium rDNA sequencesThe rDNA sequences assembled in this study included 18S,ITS1,5.8s,ITS2,26S and IGS regions,and a total of 52 rDNA sequences of Epimedium and outgroup specied V.Chrysantha.The full length of these rDNA sequences 8,467-10,464 bp.Similary,the phylogenetic tree based on rDNA sequences disagreed with the monophyly of two subgenera which was proposed by Steam in 2002.Therefore,it is suggested to cancelled the classification of two subgenera,and straightly classified Epimedium genus into five sections,namely section Polyphyllon,section Rhizophyllum,section Epimedium,section Macroceras and section Diphyllon;The rDNA sequence tree also disagreed with the classification of four series within section Diphyllon.Morphologically,there are transitional types between series campanulatum and Davidianae,so it is suggested that the four series should be cancelled.The rDNA sequence tree was significantly inconsistent with the chloroplast genome tree(nucleo-cytoplasmic conflict).In the rDNA sequence tree,E.elatum,the only species in the section Polyphyllon,was located at the base of genus Epimedium,while in the chloroplast genome tree,E.elatum was sister to section Diphyllon.It is speculated that ancient hybridization events may have occurred for the ancestor species of E.elatum and the ancestor species of section Diphyllon,resulting in chloroplast capture.There are more nucleo-cytoplasmic conflicts in the largest section n Diphyllon,and the short branches of the phylogenetic trees indicated that the complicated nucleo-cytoplasmic conflicts may result from hybridization and incomplete lineage sorting(ILS).3.Analysis of two natural hybridization casesThrough field investigation,a possible natural hybrid E.stellulatum× E.franchetii was discovered in the Wudang Mountains.The analysis of chloroplast genome and rDNA sequence of E.stellulatum× E.franchetii and its parental species revealed that the chloroplast genome of E.stellulatum_WDS was captured by E.franchetii_WDS,and 15 alleles of the rDNA sequence of E.franchetii_WDS have introgressed into E.stellulatum_WDS.The chloroplast genome and rDNA sequence of the natural hybrid E.stellulatum × E.franchetii were almost identical with the parental species E.stellulatum_WDS.Analysis of rDNA sequence of a known natural hybrid E.× omeiense and its possible parental species(E.fangii and E.acuminatum)revealed that there were 56 and 39 variable loci of E.× omeiense were identical with E.fangii and E.acuminatum,respectively.E.× omeiense have also accumulated 18 unique variable loci itself,which should be the reason why it did not cluster with any of its parents.Therefore,interspecific hybridization and introgression have resulted in a variety of morphological and molecular manifestations in hybrid species.Therefore,extensive and detailed work remains to be done to clarify the complex interspecific relationships within section Diphyllon.