• Clematis sibirica;
  • contact zone;
  • phylogeography;
  • separate refugia;
  • Tianshan and Altai Mountains


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results
  5. 3 Discussion
  6. Acknowledgements
  7. References

Abstract  Clematis sibirica, a woody vine occurring primarily under conifer forests, is widespread in northern Eurasia. In this study, we intend to illustrate how the taxon has responded in the area of the Tianshan and Altai Mountains of Central Asia to the Pleistocene climatic fluctuations. The chloroplast intergenic spacer psbA-trnH was sequenced for 125 individuals from 28 populations, and a total of eight chlorotypes were identified. The presence of definite phylogeographic structure was detected for the species (NST > GST, P < 0.001), and phylogenetic analysis indicated that the eight chlorotypes were clustered into two divergent lineages. They split at approximately 550–690 ka BP, according to coalescence analysis, coincident with the Pleistocene maximum glacial stage in these mountains, which suggests the restriction of these lineages to separate refugia at that time. Spatial analysis of molecular variance likewise divided the sampled populations into two associations, an Altai and eastern Tianshan group (populations 1–17), and a western Tianshan group (populations 18–28). Low levels of genetic diversity and unimodal mismatch distributions were obtained for both of these groups, suggesting postglacial range expansions. During the course of these expansions, mountain ranges surrounding the Dzungarian Basin probably served as migration corridors. In addition, a contact zone was identified in the central Tianshan and eastern Altai Mountains between the two phylogeographic lineages.

Lineage divergence and demographic expansion are significant aspects of phylogeographic study. Demographic histories of species are universally associated with the dynamics of paleoclimatic fluctuations (Wang et al., 2009a; Hewitt, 2011), especially the Pleistocene glacial–interglacial cycles. Most parts of China are deemed to have been free of an extensive ice sheet (Shi et al., 2006); however, species still experienced glacial-time retreats and interglacial recolonizations in response to cold–warm climatic cycles. During glacial episodes, separate refugia during glacial advances are hypothesized to have triggered lineage divergence in many species (Gao et al., 2007; Wang et al., 2009a; Guo et al., 2010; Tang et al., 2010; Li et al., 2011; Wang & Guan, 2011). Likewise, evidence from low levels of genetic diversity indicate that demographic expansion, resulting in the founder effect, has frequently occurred following the onset of interglacial warming (Cun & Wang, 2010; Zhang et al., 2010). In Europe, the postglacial migration routes of a number of species have been well inferred from the fossil records and spatial patterns of genetic variation (Hewitt, 2011), and contact and hybrid zones between phylogeographic lineages have been identified that occurred during northward recolonizations (Hewitt, 2011). No similar study has been reported in China so far.

Recently, Qiu et al. (2011) reviewed how plant species have responded to Quaternary climate changes in East Asia, and summarized the locations of putative refugia and putative colonization routes. For example, in the Qinghai–Tibet Plateau, the northeastern or southern margins have served as refugial regions for numerous species (e.g. Zhang et al., 2005; Cun & Wang, 2010; Zhang et al., 2010). These species recolonized the plateau platform or high-altitude regions with the decreased genetic diversity from the assumed refugia. However, some species may have existed in the separate refugia and the Quaternary climatic changes accelerated intraspecific divergences (e.g. Wang et al., 2009a; Jia et al., 2011; Li et al., 2011). In addition, these deep lineages may have been further mixed in the contacting zones during the postglacial expansions (Wang et al., 2009a).

The Tianshan and Altai ranges, located on the peripheries of the Dzungarian Basin in arid northwestern China, are parts of the mountain system of the Central and High Asia regions. The uplift of these mountains was a result of the Cenozoic India–Asia collision, along with the raising of the Qinghai–Tibet Plateau (Jiang et al., 2006). The forest flora of these mountains forms the “rear edge” of the Eurasian forest (Wu & Wang, 1983), which is profoundly susceptible to the influence of variations in the local arid climate. In this area, alpine glaciers are understood to have been well developed during the Pleistocene glacial ages (Shi et al., 2006). Evidenced from multiple sets of glacial tills, at least four Pleistocene glacial advances were experienced in these mountain ranges (Xu et al., 2010). Wen & Shi (1993) showed that the vegetation of the Tianshan Mountains has been responsive to Pleistocene glacial–interglacial cycles. By means of molecular phylogeography, Guo et al. (2010) showed that divergence occurring in Juniperus sabina L. between lineages of the Altai Mountains and those of the Ili Valley, in the western Tianshan Mountains, was due to their restriction to separate glacial refugia during the Pleistocene. These Pleistocene glacial–interglacial cycles are supposed to have affected forest plants in the Tianshan and Altai Mountains.

Clematis sibirica (L.) Mill., a widespread species in the Eurasian forest area, has a continuous distribution covering almost the whole of the forested area in the Tianshan and Altai ranges (Wang & Bartholomew, 2001). It prefers to occur under coniferous forests. The widespread distribution of this species and its particular ecological niche are examined to elucidate possible routes of recolonization along the mountain ranges. In addition, maternally inherited chloroplast DNA (cpDNA), owing to a low frequency of genetic recombination, shows more geographic structure than the nuclear genome in most angiosperms (Avise, 2009). It has been shown that cpDNA markers are of primary usefulness for phylogeography. In this study, we investigate the phylogeographic structure of C. sibirica in the Tianshan and Altai Mountains based on a cpDNA fragment, and focus on the following two issues: (i) inference of the spatial phylogeographic pattern of C. sibirica and the locations of past glacial refugia; and (ii) reconstruction of its postglacial demographic history and possible recolonization routes. This study, together with several previous reports, highlights the importance of the maximum glaciation occurring during the middle Quaternary to drive the deep intraspecific divergences.

1 Material and methods

  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results
  5. 3 Discussion
  6. Acknowledgements
  7. References

1.1 Population sampling

A total of 125 individuals of Clematis sibirica were collected from 28 populations in the Tianshan and Altai Mountains, which cover all the mountain ranges around the Dzungarian Basin (Fig. 1). Detailed information on these populations is shown in Table 1. Fresh leaves were collected then dried and stored in silica gel. Voucher specimens for all populations were deposited in the Herbarium of Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences (XJBI).


Figure 1. Sampling locations and lineage distribution of Clematis sibirica in the Tianshan and Altai Mountains, northwestern China. Black circles, population type A; half black/half white circles, type A+B; white circles, type B. The median-joining network for the eight recovered chloroplast DNA haplotypes (H1–H8) is shown in the bottom right corner. Population codes: 1, BL; 2, TL; 3, HF; 4, JMN; 5, HBH; 6, BEJA; 7, BEJB; 8, ALT; 9, QH; 10, YW; 11, BLK; 12, QT; 13, FH; 14, FY; 15, HM; 16, FK; 17, WLMQ; 18, SW; 19, CJ; 20, MNS; 21, WUS; 22, XY; 23, JH; 24, HC; 25, CBCE; 26, TKS; 27, ZS; 28, WENS.

Download figure to PowerPoint

Table 1.  Details of sample localities in the Tianshan and Altai Mountains for the 28 Clematis sibirica populations studied
CodePopulationLatitude/longitude N ind h (±SD)π (±SD)HaplotypeType
  1. Coordinate, number of individuals sampling (Nind), haplotypes and lineages distribution, and estimation of haplotype diversity (h) and nucleotide diversity (π) are shown for each population.

Overall  1250.6648 ± 0.02870.0166 ± 0.0089  
Altai and eastern Tianshan  760.4060 ± 0.06770.0058 ± 0.0037  
1BL45°12′N/81°46′E50.6000 f± 0.17530.0019 ± 0.0021H1, H2A
3HF46°59′N/85°50′E50.4000 ± 0.23730.0013 ± 0.0016H1, H2A
5HBH48°19′N/86°45′E30.6667 ± 0.31430.0021 ± 0.0026H1, H2A
6BEJA48°25′N/87°12′E50.6000 ± 0.17530.0019 ± 0.0021H1, H2A
7BEJB48°30′N/87°8′E50.6000 ± 0.17530.0019 ± 0.0021H1, H2A
10YW43°24′N/93°59′E50.4000 ± 0.23730.0025 ± 0.0024H1, H5A
12QT43°33′N/89°44′E40.5000 ± 0.26520.0016 ± 0.0019H1, H2A
13FH47°42′N/89°1′E21.0000 ± 0.50000.0162 ± 0.0171H1, H3A+B
14FY47°11′N/89°52′E50.6000 ± 0.17530.0094 ± 0.0068H1, H3A+B
15HM43°17′N/93°48′E50.8000 ± 0.16400.0211 ± 0.0139H1, H3, H4A+B
16FK43°54′N/88°7′E50.4000 ± 0.23730.0125 ± 0.0087H1, H6A+B
17WLMQ43°14′N/87°9′E50.7000 ± 0.21840.0161 ± 0.0112H1, H3, H6A+B
Western Tianshan  490.3461 ± 0.07690.0023 ± 0.0019  
18SW43°53′N/85°24′E50.4000 ± 0.23730.013 ± 0.0088H1, H6A+B
20MNS43°49′N/86°12′E50.6000 ± 0.17530.0019 ± 0.0021H6, H7B
22XY43°14′N/84°38′E50.4000 ± 0.23730.0013 ± 0.0016H6, H7B
24HC44°27′N/81°8′E30.6667 ± 0.31430.0021 ± 0.0026H6, H7B
26TKS42°56′N/81°46′E40.6667 ± 0.20410.0021 ± 0.0023H6, H7B
27ZS42°41′N/80°46′E50.4000 ± 0.23730.0013 ± 0.0016H6, H7B
28WENS41°49′N/80°41′E50.7000 ± 0.21840.0026 ± 0.0025H6, H7, H8B

1.2 DNA extraction, amplification, and sequencing

Total genomic DNA was extracted from approximately 50 mg silica gel-dried leaf material per sample, following a CTAB protocol modified from Doyle & Doyle (1987). One chloroplast intergenic spacer region, psbA-trnH, was amplified in the present study using the primers described in Sang et al. (1997). The polymerase chain reaction (PCR) mixture and amplification program followed the protocols of Wen et al. (2010), except an annealing temperature of 52 °C was used. The PCR products were purified from an agarose gel using the PCR product purification kit following the recommended protocol (Sangon Biotech, Shanghai, China). Sequencing reactions were carried out, with the primers described above, in both directions by standard methods, on an ABI 3730 automated sequencer at the laboratories of Sangon Biotech (Shanghai, China). Sequences were aligned using ClustalX (Thompson et al., 1997) then checked manually. The cpDNA haplotypes were determined based on nucleotide substitutions and indels. All intraspecies haplotype sequences were deposited in GenBank databases under accession numbers JX026666–JX026673.

1.3 Data analysis

To examine the phylogenetic relationships among the observed cpDNA haplotypes, a median-joining network was constructed using the program Network version 4.6 (Bandelt et al., 1999). A maximum likelihood (ML) topology of cpDNA haplotypes was also retrieved using the program PhyML 3.0 (Guindon & Gascuel, 2003), under the HKY substitution model, which was suggested by jModelTest 0.1.1 (Posada, 2008). Based on previous morphological (Wang, 1980) and molecular (Wang et al., 2009b) studies, Anemone acutiloba (DC.) G. Lawson (GenBank accession number HQ596596), a genus closely related to Clematis, was treated as the outgroup for phylogenetic analysis. In addition, a nested clade phylogeographic analysis was carried out following the approach of Templeton et al. (2005), using the program ANeCA (Panchal, 2007). In this analysis and the median-joining network analysis, each indel was treated as a single mutation event.

To determine the population structure of the 28 sampled populations, spatial amova of chlorotypes was carried out in Samova 1.0 (Dupanloup et al., 2002). This program implements a simulated annealing approach to gather geographically homogenous populations and maximally differentiate them from each other within defined groups of populations (K). The simulated annealing process was repeated for 1000 replications. An FCT value was given for every calculation. We carried out these analyses for the range of 2 ≤K≤ 8.

Genetic diversity, nucleotide diversity (π) (Nei, 1987), and haplotype diversity (h) (Nei, 1987) were calculated using Arlequin version 3.11 (Excoffier et al., 2005) for the species, each population group, and for each population. Two parameters for population diferentiation (GST, NST) and two for genetic diversity (HS, HT) were analyzed using the program Permut version 1.0 (available at, with 10 000 permutations. We also tested whether NST was significantly larger than GST, which would indicate the presence of phylogeographic structure. To estimate genetic variation within populations, among populations within groups (as identified by Samova), and between groups, amova was carried out in the program Arlequin version 3.11 (Excoffier et al., 2005), with significance tests based on 1000 permutations.

To estimate the divergence times of phylogenetic lineages, we used a Bayesian strict clock using the program Beast version 1.6.1 (Drummond & Rambaut, 2007). Based on the substitution rate of psbA-trnH proposed by Cosacov et al. (2010), we used a rate of 8–10 × 10−3 substitutions per site per million years to estimate divergence times among haplotypes. Beast uses a Bayesian Markov chain Monte Carlo (MCMC) approach with a coalescent tree prior and HKY substitution model. The MCMC chains were run for 10 000 000 generations, sampling every 1000 generations. The combined parameters were checked in Tracer version 1.4 (Drummond & Rambaut, 2007). Effective sample sizes for the relevant estimated parameters were well above 200. Finally, trees were edited in FigTree version 1.3.1 (

Pairwise mismatch distribution analysis was carried out for the defined population groups using Arlequin version 3.11 (Excoffier et al., 2005), to test historical demographic expansion. Populations that have experienced expansion are expected to have a unimodal shape in the mismatch distribution, whereas stable populations should have a bi- or multimodal distribution. The Harpending's Raggedness index (r) (Harpending, 1994) and their P-values were computed to test the significance of this population expansion model. If the sudden expansion model was not rejected, the expansion time (t) was calculated using the relationship τ= 2ut, where u is the mutation rate per generation for the whole length of analyzed sequence (Rogers & Harpending, 1992). Values of u were estimated as u= 2 μkg, where μ is the mutation rate, k is the length of the cpDNA fragment, and g is the generation time in years. According to previous results that the first flower of Clematis is produced 4 or 5 years following germination (Erickson, 1945), we assumed the generation time to be 5 years.

2 Results

  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results
  5. 3 Discussion
  6. Acknowledgements
  7. References

2.1 Chloroplast variation and haplotype distribution

The aligned cpDNA psbA-trnH data matrix was 372 bp in length. A total of eight chlorotypes were identified in the 28 sampled populations, based on four nucleotide substitutions and five indels in this cpDNA region (Table 2). Chlorotypes H1 and H6 showed widespread distributions. H1 was present in populations 1–18 (Table 1), and H6 was shown in populations 16–28 (Table 1). H2 occurred in populations BL (1), HF (3), HBH (5), BEJA (6), BEJB (7), and QT (12) (Table 1); H3 in populations FH (13), FY (14), HM (15), and WLMQ (17); and H7 was scattered in populations MNS (20), XY (22), HC (24), TKS (26), and ZS (27). The other three haplotypes (H4, H5, and H8) only occurred in populations HM (15), YW (10), and WENS (28), respectively (Table 1).

Table 2.  Variable sites of the aligned sequences of chloroplast DNA fragment (psbA-trnH) in eight haplotypes of Clematis sibirca
HaplotypeVariable sites
  1. —, indel.

 11222  2222  223
 19011  1111  264
 69001  4567  009

2.2 Phylogenetic analysis of chlorotypes and patterns of phylogeographic lineages

An ML tree of the eight observed chlorotypes was constructed using Anemone acutiloba as an outgroup. The ML topology was the same as that obtained from coalescence analysis (Fig. 2). Two independent lineages, A and B, were supported with a high bootstrap value (100%). Lineage A includes haplotypes H1, H2, and H5, and the other five haplotypes (H3, H4, H6, H7, and H8) formed lineage B. A median-joining network of the eight haplotypes was also constructed, and showed two separate lineages with the same phylogentic relationships for the eight chlorotypes (Fig. 1). Based on these lineages, we divided the 28 sampled populations into three types (Fig. 1, Table 1): type A, populations 1–12; type B, populations 19–28; and type A+B, populations 13–18. Type A was mainly distributed in the Altai Mountains and the eastern Tianshan; Type B covered populations in the western Tianshan. Populations of type A+B are mostly located in the contact zone between type A and type B. Nested clade phylogeographic analysis (Fig. 3) showed similar relationships to the median-joining network (Fig. 1). Clade 2–1 included H1, H2, H5; Clade 2–2 included H3; Clade 2–3 included H4; and Clade 2–4 included H6, H7, H8.


Figure 2. Divergence dating of the two phylogeographic lineages of Clematis sibirica based on coalescence analysis.

Download figure to PowerPoint


Figure 3. Haplotype network and nested clades of the eight haplotypes (H1–H8) in Clematis sibirica. Blank dots indicate hypothetical haplotypes. Black pie charts, populations belonging to type A, mainly distributed in the Altai Mountains and the eastern Tianshan. Gray pie charts, type A+B, mostly located in the contact zone between types A and B. White pie charts, type B, mainly in the western Tianshan.

Download figure to PowerPoint

2.3 Spatial genetic structure and genetic diversity

Genetic diversity analysis of C. sibirica revealed that total genetic diversity (HT= 0.682) across all populations was much higher than the average intrapopulation diversity (HS= 0.396). Significant phylogeographic structure (NST > GST, P < 0.001; Table 3) was detected for this species. To assess hierarchical genetic variation and make lineage demographic inferences, Samova was applied for the 28 sampled populations. The highest FCT value (0.595) was obtained when we defined K= 2, which suggested dividing the 28 populations into two groups: Altai and eastern Tianshan group (populations 1–17); and western Tianshan group (populations 18–28). The amova showed that a larger proportion of the variation was distributed among populations than within populations (76% vs. 24%; Table 4). When the partitioned groups were taken into account, a higher ratio of total variation (85%; Table 4) occurred between the two defined groups, indicating very strong genetic differentiation between them as identified by Samova. Two genetic diversity indexes, haplotype frequency (h) and nucleotide diversity (π), for each population are presented in Table 1. At the species level, haplotype and nucleotide diversities were 0.6648 and 0.0166, respectively. Nonetheless, populations from the Altai and eastern Tianshan group had on average higher levels of diversity (h= 0.4060; π= 0.0058; see Table 1) than those from the western Tianshan region (h= 0.3461; π= 0.0023; see Table 1).

Table 3.  Estimates of average gene diversity and demographic expansion results for the total distribution of Clematis sibirica
Region H S H T G ST N ST τ t
  1. G ST, interpopulation differentiation; HS, average gene diversity within populations; HT, total gene diversity; NST, number of substitution types (mean ± SE in parentheses); τ, expansion parameter; t, expansion time.

Total distributional range0.396 (0.0582)0.682 (0.0315)0.419 (0.0708)0.768 (0.0656)10.7151.1–188.9 ka
Table 4.  Results of analysis of molecular variance (amova) of chlorotype frequencies for populations and population groups of Clematis sibirica
Source of variation d.f. SS VC PV (%)Fixation index
  1. Population numbers in parentheses represent the grouping pattern for each group. d.f., degrees of freedom; FCT, correlation of chlorotypes within groups relative to total; FSC, correlation within populations relative to groups; FST, correlation within populations relative to total; PV, percentage of variation; SS, sum of squares; VC, variance component. **P < 0.001, 1000 permutations.

Total populations
 Among populations27241.7682.09732 Va76.02 F ST= 0.76018**
 Within populations8354.9170.66165 Vb23.98 
Altai and eastern Tianshan group (1–17) vs. western Tianshan group (18–28)
 Among groups1244.9824.09418 Va85.01 F CT= 0.85013**
 Among populations within groups2626.3700.08485 Vb 1.76 F SC= 0.11755**
 Within populations9761.7830.63694 Vc13.23 F ST= 0.86774**

2.4 Lineage divergence time and population demography

In this study, coalescence analysis estimated the divergence of the two lineages of C. sibirica at approximated 550–690 ka BP. The mismatch distributions of populations in both of these groups were unimodal (Fig. 4), indicating a sudden expansion model. Similarly, raggedness index (HRag) values under the sudden expansion model (Altai and eastern Tianshan group: HRag = 0.227, P= 0.538; western Tianshan group: HRag = 0.213, P= 0.623) did not reject an expansion event. Based on cpDNA substitution rate ranges, the expansion age for C. sibirica was estimated as 151.1–188.9 ka BP (Table 3).


Figure 4. Pairwise mismatch distributions of chloroplast DNA haplotypes for the two phylogeographic groups of Clematis sibirica in the Tianshan and Altai Mountains.

Download figure to PowerPoint

3 Discussion

  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results
  5. 3 Discussion
  6. Acknowledgements
  7. References

3.1 Lineage divergence and separate glacial refugia

Phylogeographic studies of many north temperate plant species have revealed lineage divergences that were likely related to the existence of two or more separate refugia during glacial times (Chen et al., 2008; Wang et al., 2009a). In the present study, significant phylogeographic structure (NST > GST, P < 0.001) and two independent lineages are displayed among the 28 sampled populations of C. sibirica from the Tianshan and Altai Mountains (Fig. 2). Lineage A, including haplotypes H1, H2, and H5, is widespread in the Altai and eastern Tianshan region, whereas the other lineage is mainly distributed in the western Tianshan region (Fig. 1). A similar spatial genetic pattern was shown by Samova, which likewise defined two groups for the 28 populations, Altai and eastern Tianshan group and western Tianshan group. The amova provided evidence that 85% of the genetic variation is distributed between these groups (Table 4), indicating a significant genetic differentiation. The population structure of C. sibirica seen in the present study is similar to that shown by Juniperus sabina (Guo et al., 2010), a species occurring in arid rocky habitats, with separate lineages between the Altai and western Tianshan (Ili Valley) areas.

Our coalescence dating, at approximately 550–690 ka BP, indicates that the lineage divergence of C. sibirica occurred during the mid-Pleistocene. In the Tianshan and Altai Mountains, at least four Pleistocene glacial–interglacial cycles were experienced (Shi et al., 2006; Xu et al., 2010). Glaciological studies have estimated maximum glaciation to have occurred at marine isotope stage 12 (approximately 500 ka BP) in this area (Shi et al., 2006; Xu et al., 2010). The consistency of times between molecular dating and the glacial chronology suggests that the lineage divergence of the species was related to the existence of refugia during the maximum glacial interval. Refugia are usually correlated with high levels of genetic diversity (Stewart et al., 2010), although this is not the only criterion (Petit et al., 2003). As phylogeographic studies in the Tianshan and Altai Mountains (Qiu et al., 2011) are generally lacking, refugial locations for plant species in this region remain rather cryptic. In the absence of an ice sheet, the flora of this area was deemed to have been influenced primarily by Pleistocene glacial-dry to interglacial-humid changes (Wen & Shi, 1993). Under the pressure of arid local climates, forest species are expected to migrate to more humid locations. Zhang et al. (2008) suggested that humid valleys (such as the Gongnaisi valley in the western Tianshan and the Kanas valley in the Altai Mountains) may have provided refugial habitats for green toads, a widespread terrestrial vertebrate. Because they are less influenced by the high barrier of the Qinghai–Tibet Plateau and surrounding mountains, the western areas are more humid than those in the eastern Tianshan and Altai Mountains. Thus, these western locations were likely to have served as refugia for C. sibirica persistence during the arid-glacial climate. In consideration of their humid habitats and high levels of genetic diversity, the sites of BL (1), HF (3), HBH (5), BEJA (6), and BEJB (7) (Fig. 1, Table 1) are speculated to be the possible refugial locations for these populations in the Altai and eastern Tianshan group. Similarly, populations of XY (22), HC (24), ZS (27), TKS (26), and WENS (28) (Fig. 1, Table 1) are other possible refugia within the western Tianshan group.

3.2 Postglacial expansions and a contact zone between two phylogeographic lineages

Two prevalent haplotypes (H1 and H6) dominate the lineages of C. sibirica (Table 1). This result suggests that the species could have experienced extensive postglacial colonization from two separate refugia. Historical demographic expansion of these population groups was also tested and supported by mismatch distribution analysis (Fig. 4). During the course of demographic expansions, the founder effect can decrease the levels of genetic diversity (Cun & Wang, 2010) and produce a pattern of one prevailing haplotype in a lineage (Zhang et al., 2010). Within the two partitioned groups of C. sibirica, low levels of genetic diversity were evident (Altai and eastern Tianshan group: h= 0.4060, π= 0.0058; western Tianshan group: h= 0.3461, π= 0.0023). These results support the idea that the two lineages of C. sibirica experienced postglacial range expansions. The time of this expansion was estimated at 151.1–188.9 ka, almost coincident with the interglacial period following the Pleistocene glacial maximum in the Tianshan and Altai Mountains (Shi et al., 2006). The interglacial climate existing at this time should have provided suitable conditions for the historical demographic expansions of C. sibirica.

Clematis sibirica occurs in conifer forests and has continuous distributions in the Tianshan and Altai Mountains surrounding the Dzungarian Basin (Fig. 1). As evidenced by many plant species (Cun & Wang, 2010; Gugger et al., 2010), the mountain ranges are supposed to have served as migration corridors during postglacial colonizations. Under pressure from the cold glacial climate, many species in the Northern Hemisphere retreated to southern warm refugia, and experienced northward colonization after the interglacial warming (Naydenov et al., 2007; Bai et al., 2010; Hewitt, 2011). However, in the present study, migration routes of C. sibirica under the influence of an arid glacial climate are different from the traditional northward expansions. Forest species most likely expanded their ranges into arid areas during the humid-interglacial periods. Two prevalent haplotypes dominate the two lineages of C. sibirica (Table 1), suggesting the ancestor haplotypes for the two lineages, which indicate their interglacial expansions. For the Altai and eastern Tianshan group (populations 1–17), lineage A with prevalent haplotype H1 experienced demographic expansion from western refugial locations in the Altai Mountains into the eastern Tianshan (Fig. 5). At the same time, lineage B with prevalent haplotype H6 in the western Tianshan Mountains also experienced eastward colonization and occupied areas of the eastern Tianshan and eastern Altai Mountains (Fig. 5). The mountain ranges surrounding the Dzungarian Basin were hypothesized to have played important roles in the postglacial expansion process of these two lineages, as migration corridors for C. sibirica. Between the two lineages, we identified a contact zone in the central Tianshan and eastern Altai Mountains showing a mixed ancestry (Fig. 5). Populations of this contact zone possessed haplotypes from both lineage A and lineage B (Fig. 1, Table 1). These populations of the contact zone (type A+B) are connected with the populations of both type A and type B in the nested clade analysis (Fig. 3). Because of their mixed ancestry, populations of the contact zone display high levels of genetic diversity (Table 1). High levels of genetic diversity are commonly observed in contact zones (Petit et al., 2003; Hewitt, 2011). In addition, as C. sibirica primarily occurs under conifer forests, which indicates this species should always follow forests, the demographic history of this species can provide an insight into the forest dynamic changes during the Pleistocene. The dynamic changes of glacial retreat and interglacial expansion for the conifer forests should have been experienced in the Tianshan and Altai Mountains.


Figure 5. Map showing the possible postglacial colonization routes of the two phylogeographic lineages of Clematis sibirica in the Tianshan and Altai Mountains and a contact zone between these two lineages in the central Tianshan Mountains and the eastern Altai Mountains. Circles, populations of type A; squares, populations of type B; triangles, populations of type A+B. Dotted line shows the range of phylogenetic lineage A, and dashed line shows that of lineage B.

Download figure to PowerPoint

In conclusion, our results suggest that Pleistocene climatic oscillations have significantly affected the current spatial genetic structure of C. sibirica in the Tianshan and Altai Mountains. Because of separate glacial refugia, two phylogeographic lineages of this forest species diverged during the Pleistocene maximum glacial stage. Following interglacial warming, demographic expansions were also identified for these phylogeographic lineages, and the mountain ranges surrounding the Dzungarian Basin are hypothesized to have served as migration corridors for the species. In addition, a contact zone between the two lineages was identified in the central Tianshan and eastern Altai Mountains. In fact, in China, the maximum glaciation of the Quaternary was suggested to have occurred between 0.8 and 0.6 Mya; during this stage, ice sheets covered most high mountains, especially in western China (Zheng et al., 2002). The deep lineages within the other several species were also dated to have corresponded to this stage (Wang et al., 2009a; Jia et al., 2011, 2012; Li et al., 2011). Therefore, our results, together with previous reports, suggest that the largest glaciation during the middle Quaternary may have played an important role in triggering the development of the divergent lineages within a few species and promoted allopatric speciation in western China.


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results
  5. 3 Discussion
  6. Acknowledgements
  7. References

We thank Prof. Jian-Quan LIU in Lanzhou University for his helpful suggestions on this manuscript. Dr. Stewart C. SANDERSON in the Shrub Sciences Laboratory, Rocky Mountain Research Station, US Department of Agriculture, Utah, USA, is acknowledged for his English improvement to the manuscript. We also thank the editor Prof. Xue-Jun GE and the anonymous referees for their valuable comments which improved the earlier version of this paper. This research was financially supported by the Important Direction for Knowledge Innovation Project of the Chinese Academy of Sciences (Grant No. KZCX2-EW-305), and the Xinjiang Institute of Ecology and Geogeraphy, Chinese Academy of Sciences (Grant No.Y276031).


  1. Top of page
  2. Abstract
  3. 1 Material and methods
  4. 2 Results
  5. 3 Discussion
  6. Acknowledgements
  7. References
  • Avise JC. 2009. Phylogeography: Retrospect and prospect. Journal of Biogeography 36: 315.
  • Bai WN, Liao WJ, Zhang DY. 2010. Nuclear and chloroplast DNA phylogeography reveal two refuge areas with asymmetrical gene flow in a temperate walnut tree from East Asia. New Phytologist 188: 892901.
  • Bandelt HJ, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution 16: 3748.
  • Chen KM, Abbott RJ, Milne RI, Tian XM, Liu JQ. 2008. Phylogeography of Pinus tabulaeformis Carr. (Pinaceae), a dominant species of coniferous forest in northern China. Molecular Ecology 17: 42764288.
  • Cosacov A, Sérsic AN, Sosa V, Johnson LA, Cocucci AA. 2010. Multiple periglacial refugia in the Patagonian steppe and post-glacial colonization of the Andes: The phylogeography of Calceolaria polyrhiza. Journal of Biogeography 37: 14631477.
  • Cun YZ, Wang XQ. 2010. Plant recolonization in the Himalaya from the southeastern Qinghai–Tibetan Plateau: Geographical isolation contributed to high population differentiation. Molecular Phylogenetics and Evolution 56: 972982.
  • Doyle JJ, Doyle JL. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf material. Phytochemical Bulletin 19: 1115.
  • Drummond AJ, Rambaut A. 2007. Beast: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7: 214.
  • Dupanloup I, Schneider S, Excoffier L. 2002. A simulated annealing approach to define the genetic structure of populations. Molecular Ecology 11: 25712581.
  • Excoffier L, Laval G, Schneider S. 2005. Arlequin (version 3.0): An integrated software package for population genetics data analysis. Evolutionary Bioinformatics Online 1: 4750.
  • Erickson RO. 1945. The Clematis fremontii var. riehlii population in the Ozarks. Annals of the Missouri Botanical Garden 32: 413460.
  • Gao LM, Möller M, Zhang XM, Hollingsworth M, Liu J, Mill RR, Gibby M, Li DZ. 2007. High variation and strong phylogeographic pattern among cpDNA haplotypes in Taxus wallichiana (Taxaceae) in China and North Vietnam. Molecular Ecology 16: 46844698.
  • Gugger PF, González-Rodríguez A, Rodríguez-Correa H, Sugita S, Cavender-Bares J. 2010. Southward Pleistocene migration of Douglas-fir into Mexico: Phylogeography, ecological niche modeling, and conservation of ‘rear edge’populations. New Phytologist 189: 11851199.
  • Guindon S, Gascuel O. 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Systematic Biology 52: 696704.
  • Guo YP, Zhang R, Chen CY, Zhou DW, Liu JQ. 2010. Allopatric divergence and regional range expansion of Juniperus sabina in China. Journal of Systematics and Evolution 48: 153160.
  • Harpending HC. 1994. Signature of ancient population growth in a low-resolution mitochondrial DNA mismatch distribution. Human Biology 66: 591600.
  • Hewitt GM. 2011. Quaternary phylogeography: The roots of hybrid zones. Genetica 139: 617638.
  • Jia DR, Liu RL, Wang LY, Zhou DW, Liu JQ. 2011. Evolutionary history of an alpine shrub Hippophae tibetana (Elaeagnaceae): Allopatric divergence and regional expansion. Biological Journal of the Linnean Society 102: 3750.
  • Jia DR, Abbott RJ, Liu TL, Mao KS, Bartish IV, Liu JQ. 2012. Out of the Qinghai–Tibet Plateau: Evidence for the origin and dispersal of Eurasian temperate plants from a phylogeographic study of Hippophaë rhamnoides (Elaeagnaceae). New Phytologist 194: 11231133.
  • Jiang M, Li HO, Wang YJ, Zhang LS, Li QQ. 2006. The effect of Qinghai–Tibet plateau uplift on the structure of crust and upper mantle beneath the Tianshan mountains in Xinjiang. Earth Science Frontiers 13: 401407.
  • Li Y, Zhai SN, Qiu YX, Guo YP, Ge XJ, Comes HP. 2011. Glacial survival east and west of the ‘Mekong-Salween-Divide’ in the Himalaya–Hengduan Mountains Region as revealed by AFLP and cpDNA sequence variation in Sinopodophyllum hexandrum (Berberidaceae). Molecular Phylogenetics and Evolution 59: 412424.
  • Naydenov K, Senneville S, Beaulieu J, Tremblay F, Bousquet J. 2007. Glacial vicariance in Eurasia: Mitochondrial DNA evidence from Scots pine for a complex heritage involving genetically distinct refugia at mid-northern latitudes and in Asia Minor. BMC Evolutionary Biology 7: 233.
  • Nei M. 1987. Molecular evolutionary genetics. New York : Columbia University Press.
  • Panchal M. 2007. The automation of nested clade phylogeographic analysis. Bioinformatics 23: 509510.
  • Petit RJ, Aguinagalde I, de Beaulieu JL, Bittkau C, Brewer S, Cheddadi R, Ennos R, Fineschi S, Grivet D, Lascoux M. 2003. Glacial refugia: Hotspots but not melting pots of genetic diversity. Science 300: 15631565.
  • Posada D. 2008. jModelTest: Phylogenetic model averaging. Molecular Biology and Evolution 25: 12531256.
  • Qiu YX, Fu CX, Comes HP. 2011. Plant molecular phylogeography in China and adjacent regions: Tracing the genetic imprints of Quaternary climate and environmental change in the world's most diverse temperate flora. Molecular Phylogenetics and Evolution 59: 225244.
  • Rogers AR, Harpending H. 1992. Population growth makes waves in the distribution of pairwise genetic differences. Molecular Biology and Evolution 9: 552569.
  • Sang T, Crawford DJ, Stuessy TF. 1997. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). American Journal of Botany 84: 11201136.
  • Shi YF, Cui ZJ, Su Z. 2006. The Quaternary glaciations and environmental variations in China. Shijiazhuang : Hebei Science and Technology Press.
  • Stewart JR, Lister AM, Barnes I, Dalén L. 2010. Refugia revisited: Individualistic responses of species in space and time. Proceedings of the Royal Society B: Biological Sciences 277: 661.
  • Tang LZ, Wang LY, Cai ZY, Zhang TZ, Ci HX, Lin GH, Su JP, Liu JQ. 2010. Allopatric divergence and phylogeographic structure of the plateau zokor (Eospalax baileyi), a fossorial rodent endemic to the Qinghai–Tibetan Plateau. Journal of Biogeography 37: 657668.
  • Templeton AR, Maxwell T, Posada D, Stengard JH, Boerwinkle E, Sing CF. 2005. Tree scanning: A method for using haplotype trees in phenotype/genotype association studies. Genetics 169: 441453.
  • Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG. 1997. The Clustal_X windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Research 25: 48764882.
  • Wang LY, Abbott RJ, Zheng W, Chen P, Wang YJ, Liu JQ. 2009a. History and evolution of alpine plants endemic to the Qinghai–Tibetan Plateau: Aconitum gymnandrum (Ranunculaceae). Molecular Ecology 18: 709721.
  • Wang W, Lu AM, Ren Y, Endress ME, Chen ZD. 2009b. Phylogeny and classification of Ranunculales: Evidence from four molecular loci and morphological data. Perspectives in Plant Ecology, Evolution and Systematics 11: 81110.
  • Wang WT. 1980. Ranunculaceae. In: Flora Reipublicae Popularis Sinicae. Beijing : Science Press. 27: 3336.
  • Wang WT, Bartholomew B. 2001. Clematis. In: Wu ZY, Raven PH eds. Flora of China. Beijing : Science Press; St. Louis : Missouri Botanical Garden Press. 6: 333386.
  • Wang ZJ, Guan KY. 2011. Genetic structure and phylogeography of a relict tree fern, Sphaeropteris brunoniana (Cyatheaceae) from China and Laos inferred from cpDNA sequence variations: Implications for conservation. Journal of Systematics and Evolution 49: 7279.
  • Wen QZ, Shi YF. 1993. The Quaternary climo-environment changes in Chaiwopu Basin of Xinjiang region. Chinese Geographical Science 3: 147158.
  • Wen ZB, Zhang ML, Zhu GL, Sanderson SC. 2010. Phylogeny of Salsoleae s.l. (Chenopodiaceae) based on DNA sequence data from ITS, psbB-psbH, and rbcL, with emphasis on taxa of northwestern China. Plant Systematics and Evolution 288: 2542.
  • Wu ZY, Wang HS. 1983. Chinese physical geography – phytogeography (I). Beijing : Science Press.
  • Xu XK, Kleidon A, Miller L, Wang SQ, Wang LQ, Dong GC. 2010. Late Quaternary glaciation in the Tianshan and implications for palaeoclimatic change: A review. Boreas 39: 215232.
  • Zhang Q, Chiang TY, George M, Liu JQ, Abbott RJ. 2005. Phylogeography of the Qinghai–Tibetan Plateau endemic Juniperus przewalskii (Cupressaceae) inferred from chloroplast DNA sequence variation. Molecular Ecology 14: 35133524.
  • Zhang YH, Volis S, Sun H. 2010. Chloroplast phylogeny and phylogeography of Stellera chamaejasme on the Qinghai–Tibet Plateau and in adjacent regions. Molecular Phylogenetics and Evolution 57: 11621172.
  • Zhang YJ, Stöck M, Zhang P, Wang XL, Zhou H, Qu LH. 2008. Phylogeography of a widespread terrestrial vertebrate in a barely-studied Palearctic region: Green toads (Bufo viridis subgroup) indicate glacial refugia in Eastern Central Asia. Genetica 134: 353365.
  • Zheng BX, Xu QQ, Shen YP. 2002. The relationship between climate change and Quaternary glacial cycles on the Qinghai–Tibetan Plateau: Review and speculation. Quaternary International 97–98: 93101.