Phylogeography And Molecular Identification Of Tetrastigma Hemsleyanum, A Medicinal Plant In Vitaceae

Tetrastigma hemsleyanum Diels et Gilg belongs to the Vitaceae family, and is known as an important medicinal herb endemic to China. This species is widely distributed throughout subtropical China, and occurs rarely farther south on Hainan and Taiwan, which represents a widespread component of China’s subtropical broad-leaved evergreen forest (BLEF). BLEF makes up much of the primary vegetation of subtropical China in lower elevations, and experienced complex vegetation dynamics since late Tertiary. Palaeobiome reconstructions for the Last Glacial Maximum (LGM) suggested a complete southward shifts of BLEF till the edge of the current tropical rain forest (below 25°N). On the contrary, phylogeography studies focused on alleged constituents of BLEF in subtropical China, consistently revealed limited post-glacial expansion from localized refugia in situ. Thus, how BLEF and its component species responded to past environmental change remains largely unknown. Studies on other BLEF species with wide distributions are required to more rigorously assess the spatial and temporal dynamics over geological scales. In this study, we focused on T. hemsleyanum, and conducted a multidisciplinary survey using principles and methods from phylogeography, phylogeny, biogeography, and ecological niche modelling to gain further insights into the intraspecific divergence and spatiotemporal population dynamics of BLEF species in subtropical China under conditions of late Tertiary/Quaternary climate change. Despites, T. hemsleyanum is a traditional Chinese medicine commonly used for inflammation, fever, asthma etc., but its medicinal parts (roots) are often added with adulterants in herbal medicine market, hence we designed sequence characterized amplified region (SCAR) markers to distinguish T. hemsleyanum from its congenric relatives and imitations. The results of our study are as follows:(1) Intraspecific lineage divergence and population dynamics based on cpDNABased on the polymorphisms of three chloroplast regions (petL-psbE, trnK-matK, rbcL), we identified 20 haplotypes across the 20 surveyed populations of T. hemsleyanum. At the species level, the cpDNA data revealed high estimates of genetic diversity (hT= 0.918; πT= 2.110×10-3), and significant phylogeographic structures at the range-wide scale (NST= 0.851, GST= 0.795, P< 0.05). The phylogenetic analysis and parsimony network result consistently recovered two major lineages within T. hemsleyanum:Southwest (SW) lineage and Central-South-East (CSE) lineage. The CSE lineage can be further divided into three major clades, which is Hainan, Central and South/East clade. Hierarchical AMOVA revealed that 67.9% of the species’total variation was distributed among the four regions (Southwest, Central, South/East China, Hainan), with 19.5% explained by variation among populations within regions. Based on four calibration points, BEAST dated the split between the SW lineage and the CSE lineage at about the early Pliocene (c.5.07 Ma). Within the CSE lineage, Hainan diverged early at c.2.78 Ma, followed by a split of the remainder into a Central and South/East clade at c.2.12 Ma. The ancestral area reconstructions based on cpDNA data supported a probably ancient (pre-Quaternary) distribution of T. hemsleyanum in Southwest and South China. A vicariant event, potentially trigged by the climate and geomorphological changes in the early Pliocene, resulted in the early formation of cpDNA lineages, and was likely followed by two independent colonizations from South China to, respectively, Central and East China, possibly during the late Pliocene to early Pleistocene. MDA analysis as well confirmed the long-term stability of SW lineage, and only detected the existence of spatial expansions in CSE lineage and the Central clade, which were both dated to glacial cycles predating the LGM (CSE lineage:0.81 Ma; Central clade:0.71 Ma). These finding illustrates that pre-Quaternary environmental changes have left strong footprints on the modern genetic structure of the BLEF species.(2) Population structure and inter-region gene flows based on nSSR dataWe developed 8 polymorphic nuclear microsatellite markers for T. hemsleyanum using dual-suppression-PCR based approach. In the Bayesian analysis of population structure based on nSSR data, T. hemsleyanum individuals were clustered into four groups, with geographic distributions corresponding to Southwest China, Central China, South China and East China, respectively. This pattern suggested that T. hemsleyanum persisted through the LGM in at least four refugia. We estimated the pairwise postglacial gene flow between the four genealogically distinct units of chloroplast DNA using MIGRATE, and discovered only low to medium levels of inter-regional gene flow, despite some exceptions of (asymmetrical) gene flow detected between adjacent areas (i.e. SW vs Central China; South/East China vs Hainan). The ENM results predicted that at the LGM, the species’potential range showed a southwestward range shift in general, but did not retreat entirely to the tropical south. The suitable habitat remained more or less stable in the Southwest and South, and expanded on the exposed South China Sea Shelf (including present-day Hainan Island), which might provide migration route between mainland and Hainan populations. These results indicated that this species exhibited characteristic of long-term separation in multiple refugia, limited demographic fluctuations, and little evidence of gene flow impeding interregional divergence over the last glacial-interglacial cycle, in contrast with palaeobiome reconstructions showing that this forest biome retreated to areas of today’s tropical South China during the LGM.(3) Comparative transcriptomes and the development of single-copy nuclear genesWe sequenced transcriptomes of T. hemsleyanum individuals from CSE lineage and SW lineage, respectively, using Hiseq 2000 sequencing platform. By de novo assembly, 52,838 (CSE) and 65,197 (SW) unigenes were obtained, of which more than 89% have function annotations in public databases. Based on sequence similarity, we discovered 6692 pairs of orthologs between lineages, of which ten pairs of orthologous genes showed Ka/Ks ratios significantly greater than 0.5 (P< 0.05), while 3734 pairs had Ka/Ks ratios significantly smaller than 0.5 (P< 0.05), indicating that these gene pairs are likely under purifying selection. We found 286 SSRs among the identified pairs of orthologs, of which 78 probably had variation between lineages. We integrated three different bioinformatic approaches for the discovery of single-copy nuclear genes (SCN), and altogether identified 1018 SCNs. Nearly 78% of the SCN loci exhibited low to medium levels of genetic divergence between lineages (Pairwise identity≤ 99.5%). Using the Vitaceae genome as reference, we predicted the putative intron-exon boundaries and approximate intron size of each SCN loci, which further facilitated the successfully design of 12 exon-anchored SCN primer pairs. Among the 12 developed SCN primers,8 showed high levels of polymorphisms in the cpDNA haplotype individuals (hT:0.909-0.993; πT:7.32-16.01×10-3). This study provided a wealth of molecular resources potentially useful for future evolutionary studies in T. hemsleyanum and closely related species.(4) Molecular identification of T. hemsleyanumOut of 100 ISSR universal primers,12 produced strong, clear and reproducible bands, and were selected for further experiments. Among them, primer UBC843 amplified an approximately 1800 bp band unique to samples of T. hemsleyanum. After cloning and sequencing, this DNA fragment turned out to be 1773 bp, and was used for the development of SCAR markers (ThkF and ThkR). PCR tests using SCAR primers showed that only T. hemsleyanum samples produced a single band with the expected fragment size. Therefore, the SCAR primers designed in this study were proven to be diagnostic probe markers for identifying T. hemsleyanum from its congenric relatives and imitations.