The Phylogenetic Position And Hybrid Origination Of Rosa Praelucens Byhouwer And The Molecular Phylogeny Of Genus Rosa L.

Rosa praelucens Byhouwer,endemic to Yunnan,China,is a famous alpine flower and is highly tolerant to coldness.It is also an important resource plant because its hips are edible and rich in vitamins.Furthermore,Rosa praelucens,so far reported,is the only one decaploid wild rose species in the world.Now,it has been listed as a ‘critically endangered'(CR)plant of China.Thus,Rosa praelucens is an important and singnificant plant material for the study of evolution,biological conservation and plant resource utilization.However,the phylogenetic position of Rosa praelucens is still debatable and its origination is unclear yet.Moreover,more data is needed to reconstruct a robust phylogenetic model of genus Rosa since there exist many inconsistence between the molecular systematics and the traditional classification.In this study,we reconstructed the phylogeny of 50 rose species/varieties based on 5S rDNA,four chloroplast DNA fragments(psbA-atpH,rbcL,rpl16 and trnL-F)and 6 pairs of amplified fragment length polymorphism(AFLP)markers to explore the phylogenetic position of R.praelucens.We postulated the original ancestors of R.praelucens according to the results of sequence alignment and base comparison,and the genetic similarity of AFLP markers among its closely related species.The results were as the follows:(1)The sequence of 5S rDNA ranged from 498 bp to 573 bp in length and varied significantly among different rose species.The combined alignment of 4 cpDNA fragments was 2969 bp in length,153 of which were polymorphic.727 bands were scored by using 6 pairs of AFLP primers,724 of which were polymorphic.(2)Rosa praelucens clustered with species from sect.Cinnamomeae,rather than R.roxburghii of sect.Microphyllae in the phylogenetic trees based on 5S rDNA and cpDNA fragments and in the dendrogram of AFLP markers,suggesting that R.praelucens should be placed in sect.Cinnamomeae rather than in sect.Microphyllae.(3)The origination of R.praelucens was very complex.Rosa caudata,R.murielae,R.davidii,R.gracilifolia and R.moyesii were most closely related to R.praelucens.The common maternal ancestor of R.caudate,R.murielae and R.davidii or R.caudate was most likely to be the maternal ancestor of R.praelucens.Both R.gracilifolia and R.moyesii were most likely to be its paternal ancestors,and R.caudata might also have participated the hybrid origination of R.praelucens as paternal parents.(4)The genus Rosa is monophyletic based on moleculer data.Within the genus,sect.Banksianae,sect.Bracteatae and sect.Laevigatae were closely related to each other in the phylogenetic trees.Sect.Pimpinellifoliae was not monophyletic and closely related to sect.Cinnamomeae.Both in the phylogenetic trees based on 5S rDNA and cpDNA fragments and in the dendrogram of AFLP markers,R.graciliflora of sect.Pimpinellifoliae clustered with species from sect.Cinnamomeae,indicating that R.graciliflora should be placed in sect.Cinnamomeae rather than in sect.Pimpinellifoliae.Also,Rosa pseudobanksiae of sect.Cinnamomeae clustered with species from sect.Banksianae in both the phylogenetic tree of cpDNA sequences and the dendrogram of AFLP data.In the 5S rDNA phylogeny,one clone sequence of R.pseudobanksiae clustered with species from sect.Banksianae,while another clone clustered with R.lichiangensis from sect.Synstylae.This result suggested that R.pseudobanksiae was also a hybrid species.Its maternal ancestor was likely to be R.banksiae var.normalis of sect.Banksianae and its paternal ancestor was likely to be R.lichiangensis of sect.Synstylae-Chinenses,respectively.Thus,R.pseudobanksiae should be removed from sect.Cinnamomeae.Finally,species from sect.Synstylae and sect.Chinenses clustered not according to the sections of the traditional taxonomy but mingled with each other,indicating these two sections in the traditional taxonomy were not monophyly and the Synstylae-Chinenses clade need further studying.