五味子科Schisandraceae隶属于木兰藤目Austrobaileyales，本文研究所用概念为狭义五味子科，包括五味子属Schisandra和南五味子属Kadsura。五味子科植物具有重要的药用价值，其中五味子（五味子S. chinensis）、南五味子（华中五味子S. sphenanthera）以及滇鸡血藤（内南五味子K. interior）已被收入《中国药典》一部（2015版）。然而，五味子科的系统学研究存在一些问题，例如：五味子属和南五味子属是否为单系类群、属下分类及物种归并是否合理等。这些问题在一定程度上影响到临床用药疗效。因此，本论文开展了五味子科叶绿体基因组学研究、DNA系统发育树的重建和谱系地理学方法的分析，以期通过上述新技术和新方法，对五味子科系统发育和物种鉴定进行探讨。

叶绿体基因组属于母系遗传，具有遗传信息丰富、核苷酸置换率适中、在不同植物类群中有较好的共线性等优点，是植物系统发育研究的理想工具。本研究利用二代测序技术Illumina测序平台，对五味子科12物种16份样本进行测序，摸索出一套叶绿体基因组组装策略，并且通过一代测序的验证，组装正确率100%。通过长度比较发现，五味子科叶绿体基因组长度存在两种明显差异，可分为S型基因组和L型基因组，S型基因组长度在145.4-146.9 kb之间，L型基因组长度在152.7-153.4 kb之间。在五味子科叶绿体基因组中共注释出113个不相同的基因，包括79个蛋白质编码基因、30个tRNA基因和4个rRNA基因，其中有18个内含子基因。重复序列分析显示，五味子科叶绿体基因组平均含有145个SSR，其中包括99个Mono-SSR、29个Di-SSR、6个Tri-SSR、5个Tetra-SSR、2个Penta-SSR和3个Hexa-SSR。五味子科叶绿体基因组平均正向重复（F）、回文重复（P）和串联重复（T）序列分别为24、21和26，根据T、F和P重复序列的比例大致可将五味子科叶绿体基因组分成三种类型，一类为黑老虎K. coccinea，T含量约43-47%，F含量约26-29%，P含量约26-29%；二类为L型基因组物种，其T重复含量为28-39%，F重复含量为29-35%，P重复含量为32-39%；三类为大多数五味子属物种，T重复含量为40-45%，F重复含量为36-40%，P重复含量为16-21%；五味子属物种重瓣五味子S. plena T重复含量为36.36%，F重复含量为33.77%，P重复含量为29.87%，其三者含量与二类较为相似。通过结构比较分析发现，五味子科叶绿体基因组中，S型基因组的IR区发生了收缩，非编码区变异大于编码区，SC区变异大于IR区，合蕊五味子S. propinqua和重瓣五味子与南五味子属物种相似性较高。系统进化分析的结果支持林祁将多蕊五味子群和中华五味子群作为两个组隶属于五味子亚属的观点。

通过比较ITS、ITS2、psbA-trnH、matK和rbcL序列对内南五味子的鉴别能力，结果发现，除rbcL序列外，其余序列都可以通过遗传距离和系统发育树将内南五味子与五味子属和南五味子属黑老虎进行区分，但是只有ITS和psbA-trnH序列能够通过单核苷酸多态性（SNP）分析将内南五味子与异形南五味子K. heteroclita和长梗南五味子K. Longipedunculata进行区分。因此，本研究推荐ITS和psbA-trnH作为鉴定内南五味子的优选条形码。

利用谱系地理学研究方法探讨南五味子物种复合体的物种鉴定，结果发现内南五味子的psbA-trnH序列中存在2个特有单倍型，在ITS序列中，仅与云南屏边县的异形南五味子存在1个共有单倍型，因此通过谱系地理可以达到鉴定内南五味子的目的。通过分析ITS序列单倍型的地理分布情况，结果发现不同单倍型类群的地理分布存在一定规律，如内南五味子所在的类群C_1主要分布于中国西南地区，以日本南五味子为主导的类群C_3主要分布于华东地区。这为物种的界定以及未来药用原料中相似药材的采收提供遗传学依据。

Schisandraceae is a family of Austrobaileyales belonging to basal angiosperms. Schisandraceous plants has important medicinal value. Schisandra chinensis, S. sphenanthera and Kadsura interior has been recorded in Chinese Pharmacopoeia. However, there are a lot of problems in the systematic study of Schisandraceae, such as systematic position of Schisandraceae, monophyletic and species complex problem in Schisandra and kadsura, and species classification. These problems have a great effect on clinical medication. In this study, we hope to promote the development of the classification and identification of Schisandraceae by chloroplast genome, DNA barcoding and phylogeography.

The chloroplast (cp) genome is semiautonomous and maternally inherited. The cp genome has a wealth of genetic variation and moderate rate of nucleotide evolution. It can be used for phylogenetic studies at different taxonomic levels. The total genomes of 16 species in Schisandraceae were completely sequenced by using the Illumina PE150 platform. According to length, these cp genomes were divided into two types, i.e. Long (L) type and Short (S) type. The L type which length range is 152.7-153.4 kb includes K. interior, K. heteroclite, K. longipedunculata, K. japonica and K. oblongifolia. S type which length range is 152.7-153.4 kb includes K. coccinea, S. propinqua, S. plena, S. chinensis, S. sphenanthera, S. rubriflora and S. henryi. 113 unique genes, including 79 protein-coding genes, 30 tRNA genes and four rRNA genes, were annotated in Schisandraceous cp genomes. A total of 145 simple sequence repeats (SSRs), including 99 Mono-SSRs、29 Di-SSRs、6 Tri-SSRs、5 Tetra-SSRs、2 Penta-SSRs and 3 Hexa-SSRs, were detected in Schisandraceous cp genomes. The proportion of T, F and P was classified three types. Type 1 includes K. coccinea with 43-47% of T, 26-29% of F and 26-29% of P. Type 2 includes species of L type with 28-29% of T, 29-35% of F and 32-39% of P. Type 3 includes most species of Schisandra, except S. plena, with 40-45% of T, 36-40% of F and 16-21% of P. By comparing each cp genome with S. chinensis, we found the cp genomes of S. plena and S. propinqua were similar to that of Kadsura species. A large IR expansion of about 8 kb has occurred in L type species. A phylogenetic analysis based on all Schisandraceous cp genome sequences supported that taxa Pleiostema and Sinoschisandra should be set as two sect. which belong to subgen. Schisandra. Species of L type grouped into one clade which involved in all species of subgen. Kadsura distributed in China. Remarkably, four samples of K. heteroclite did not group into one clade.

DNA barcoding was used to identify K. interior. We analyzed five DNA barcodes (ITS, ITS2, psbA-trnH, matK and rbcL) in terms of distance-based, tree-based and character-based identification methods to distinguish K. interior and its adulterants. In distance-based and tree-based identification, K. interior could be distinguished easily from the species of Schisandra and K. coccinea. In character-based identification, there are two single nucleotide polymorphisms (SNPs) in ITS and one SNP in psbA-trnH which can be used to distinguish K. interior from K. heteroclita and K. longipedunculata. The results indicate that DNA barcoding can be used to identify the K. interior. The ITS and psbA-trnH sequence can be the most ideal DNA barcodes for discriminating K. interior and its adulterants by the combination analysis of distance-based, tree-based and character-based identification (SNPs).

Phylogeographical analyses were carried out for species identification in Kadsura. The result shown that K. interior owned two specific haplotypes of psbA-trnH and one shared haplotype of ITS with K. heteroclite which was collected in Pingbian Yunnan. Therefore, by this method we are able to identify K. interior from its affinities. After classifying all haplotypes of ITS into six types (C_1-C_6), we also find out C_1 distributed in southwest China, C_2 in central China, C_3 in east China. It proved genetic basis for species delimitation and application of similar medicinal materials.