Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx http://www.paper.edu.cn

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier.com/locate/ympev

Phylogeography of a marsh herb trifolia () in inferred from cpDNA atpB–rbcL intergenic spacers

Jin-Ming Chen a, Fan Liu a, Qing-Feng Wang a,b,*, Timothy J. Motley b a Laboratory of Systematics and Evolutionary Biology, College of Life Sciences, Wuhan University, Wuhan 430072, Hubei, PR China b Department of Biological Sciences, Old Dominion University, Norfolk, VA 23529-0266, USA article info abstract

Article history: Sagittaria trifolia L. is a perennial, erect herb that is confined to ponds, rice fields, ditches, and freshwater Received 27 October 2007 wetlands. Using chloroplast DNA (cpDNA) atpB–rbcL intergenic spacer sequences, we studied the phylog- Revised 2 March 2008 eographic pattern and demographic history of S. trifolia with 108 samples from 42 populations represent- Accepted 4 March 2008 ing the entire geographic range in China. Twenty-seven haplotypes were characterized and two of them Available online xxxx were widely distributed in the populations. In the minimum-spanning network, all tip haplotypes were unique to a particular population, while the interior nodes represented widespread haplotypes. Nested Keywords: clade analysis (NCA) of cpDNA haplotypes indicated that long distance dispersal characterized the cpDNA atpB–rbcL noncoding spacer post-glacial recolonization of S. trifolia in China. No specific refugia areas were suggested because genetic sequence Marsh herb differentiation was low among the sampled regions and among populations within regions although a Nested clade analysis large number of the haplotypes were unique to a single population. The present data support that the Phylogeography unique haplotypes in individual population most likely represent recent mutational derivatives after long Sagittaria trifolia distance dispersal rather than the relics in refugia. These results for S. trifolia represent the first phylog- eographic analysis of a widespread marsh herb in China and support the importance of long distance dis- persal events in the post-glacial migrations of . Ó 2008 Published by Elsevier Inc.

1. Introduction (Cox and Moore, 2000; Hewitt, 2000, 2004). Most of these studies have been concerned with the patterns of distributional changes In the evolutionary history of a species, the distribution of pop- following climatic oscillations in Europe and the North America ulations could be constricted or expanded due to the environmen- (Soltis et al., 1997; Taberlet et al., 1998; Griffin and Barrett, tal changes. The climatic oscillations during the Quaternary ice age 2004; Schonswetter et al., 2005). have been considered to play a major role in changing the geo- China has some distinctive biogeographical features. The Qing- graphical distribution of plant and animal species (Hewitt, 1996, hai–Tibet (Q–T) Plateau located in the western China is the highest 2000; Comes and Kadereit, 1998). The climatic oscillations of the and the largest plateau in the world. The Q–T Plateau was a part of Quaternary, 2 million years ago, resulted in several glacial and the Ancient Tethys Sea and began to uplift about 50 million years interglacial cycles (Shackleton and Opdyke, 1973). In general, the ago in the Eocene of the Lower Tertiary (Harrison et al., 1992; An distribution of organisms shifted across latitudes and elevations et al., 2001). With the large-scale intense uplift of the Q–T Plateau, in response to these paleoclimatic cycles, contributing to species the Chinese three-step land features were formed (Zhang et al., distributions being restricted to refugia during cold, glacial periods 2000). The formation of the Q–T Plateau has dramatically influ- and rapidly range expansions with the onset of interglacial warm- enced the natural environments of the East and the East Asia ing (Hewitt, 1996; Taberlet et al., 1998). flora (Coleman and Hodges, 1995; Zheng and Li, 1999; An et al., Although the effects of glacial–interglacial cycling of the Qua- 2001; Sun, 2002; Shi et al., 1998). Although no massive ice sheet ternary were global, the biological response appears to vary with developed in most parts of China during glacial periods, the tre- locality and taxa (Cox and Moore, 2000; Hewitt, 1996; Taberlet mendous global climatic changes combined with the local climatic et al., 1998). Recent progress in the use of molecular markers has changes caused by the Q–T Plateau uplift, particularly during Qua- facilitated many phylogeographic studies in northern hemisphere ternary glaciations, have influenced the distribution and evolution of many plant species in China and its neighboring areas (Zhang et al., 2005; Wang and Ge, 2006). In contrast to the numerous * Corresponding author. Address: Laboratory of Plant Systematics and Evolution- investigations on plants of Europe and North America, understand- ary Biology, College of Life Sciences, Wuhan University, Wuhan 430072, Hubei, PR China. Fax: +86 27 68752560. ing of the effects of past climate events on the current distribution E-mail address: [email protected] (Q.-F. Wang). patterns of species in China is relatively limited.

1055-7903/$ - see front matter Ó 2008 Published by Elsevier Inc. doi:10.1016/j.ympev.2008.03.008 转载

Please cite this article in press as: Chen, J.-M. et al., Phylogeography of a marsh herb Sagittaria trifolia (Alismataceae) in China ..., Mol. Phylogenet. Evol. (2008), doi:10.1016/j.ympev.2008.03.008 中国科技论文在线 http://www.paper.edu.cn

2 J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx

Recent phylogeographical studies of Chinese plant species seven regions, i.e., Northeast China (JL-1–3, HLJ-1–8), Northwest have focused on the endangered or endemic species that with China (XJ-1 and GS-1), North China (HEB-1), Central China (HUB- narrow distribution ranges in China. Most phylogeographical 1–4, HEN-1, HUN-1), East China (JS-1 and 2, and JX-1–3), South studies of plant migration in China have dealt with woody, China (GX-1 and 2, and HAIN-1), and Southwest China (YN-1–10 long-lived tree and shrub species, e.g., Cathaya argyrophylla,an and GZH-1–4). The definition of regions in this study was followed endangered conifer restricted to subtropical mountains of China the criteria of Ren and Yang (1961), which was based on the differ- (Wang and Ge, 2006); Juniperus przewalskii, a tree endemic to ences of the altitude and latitude, the distribution of the flood and the Q–T Plateau region (Zhang et al., 2005); Alsophila spinulosa, field, the topography, the climate, and the distribution of vegeta- a relictual tree fern distributed in southern China (Su et al., tion. A total of 108 individuals from the 42 populations were in- 2005a,b) ; and Dunnia sinensis, an endangered, endemic shrub re- cluded in the study. Details on materials are given in Table 1. stricted to the southern part of Guangdong Province (Ge et al., About 5 g of fresh leaves per plant was collected and immediately 2002). Although these studies have enriched the knowledge of dried with silica gel. the effects of historical events on the current geographical distri- butions of several species in China, due to large gaps in taxo- 2.2. DNA extraction, PCR amplification, and sequencing nomic and habitat sampling, for example, as yet there are few published data examining phylogeographical patterns across a Total genomic DNA was isolated from 0.5 g of silica-dried leaf wide geographic range (Ran et al., 2006) and relating to herba- tissue following the procedure described by Fu et al. (2003). PCR ceous plant phylogeography (Huang et al., 2005), our understand- was performed in a reaction volume of 50 ll containing 0.25 mM ing of the historical biogeography events in China still remains of each dNTP, 2.5 llof10Taq buffer [10 mM Tris–HCl (pH 8.3),

incomplete. 1.5 mM MgCl2 and 50 mM KCl], 1 mM of each primer, 2 U Taq Poly- Sagittaria trifolia L. is a perennial, erect marsh herb that be- merase (Tian Yuan Biotech) and 60 ng of DNA template. Primers of longs to the family Alismataceae. It is a self-compatible species, Chiang et al. (1998) were used to amplify the cpDNA atpB–rbcL which can reproduce both sexually by selfed and out-crossed noncoding spacer. Amplification of genomic DNA was made on a seeds and vegetatively through corms (Chen, 1989). The species PTC-100TM thermocycler (MJ Research, Inc.), and commenced with is insect-pollinated by bees (Hymenoptera). Seeds are dispersed 4 min at 94 °C, followed by 35 cycles of 1 min at 94 °C, 1 min by water or animals (Cook, 1990). Most of the seeds drop and annealing at 50 °C and 2 min extension at 72 °C, and a final exten- germinate within populations (pers. obs. Chen J.M.). S. trifolia sion cycle of 7 min at 72 °C. The size of PCR products was deter- is one of the most widespread species in the genus Sagittaria mined by agarose electrophoresis. All PCR products were purified ranging from north Beikal in Russia to the Southeastern part of from an agarose gel using the PCR product purification kit (Shang- Asia and Europe. It is confined to ponds, rice fields, ditches, hai SBS, Biotech Ltd., China). The purified PCR products were se- and freshwater wetlands (Chen, 1989). In China, the distribution quenced in both directions by standard methods on an ABI 377 range of this species extends to most parts of the country except automated sequencer in Beijing Genomics Institute, Chinese Acad- for the Q–T Plateau and the western desert area. The current emy of Sciences. widespread distribution range probably spans the former glaci- ated (e.g., the Northeast and the Northeastern regions of China) 2.3. Data analysis and ice-free (e.g., the central China and the Southeastern region of China) portions of China (Wu, 1979; Chen, 1989). Therefore, Sequences were aligned using CLUSTAL W program (Thompson S. trifolia appears to be an ideal candidate for inferring the pat- et al., 1994). The alignments were then adjusted manually. Inser- terns of plant migration following the historical climatic oscilla- tions/deletions (indels) were generally placed so as to increase tions in China. the number of matching nucleotides in a sequence position. We The Chloroplast DNA (cpDNA) noncoding spacers have been uti- determined cpDNA haplotypes based on nucleotide substitutions lized frequently in recent years to survey population variation and and indels. Tests of neutrality including Tajima (1989) D, Fu and phylogeography of plants (Cannon and Manos, 2003; Honjo et al., Li’s (1993); D* and F, and the determination of their associated sig- 2004; Ikeda and Setoguchi, 2007). Their uniparental inheritance, nificance were performed using the DnaSP 4.0 program (Rozas nearly neutral, slow rate of evolution, which allows the preserva- et al., 2003). tion of haplotypes over generations with few changes among sites, In order to describe genetic structure and variability among were well suited for locating refugia and post-glacial recoloniza- populations, the non-parametric Analysis of Molecular Variance tion routes (Petit et al., 2003; Heuertz et al., 2004; Grivet and Petit, (AMOVA) was performed using squared Euclidean distances 2003; Lascoux et al., 2003). In this study we use a broad geograph- (Excoffier et al., 1992). Variance was partitioned to the following ical sampling of populations of S. trifolia in China and phylogeo- components: among individuals within populations, among popu- graphical methods, e.g., nested clade analysis (NCA), to analyze lations within regions and among regions. Genetic analyses were the genetic structure of the cpDNA atpB–rbcL intergenic spacer performed with ARLEQUIN ver. 2.001 (Schneider et al., 2000). haplotypes. We were particularly interested in addressing the fol- Pairwise differences between DNA haplotypes were also cal- lowing specific questions: (1) what is the genetic structure of S. tri- culated using ARLEQUIN ver. 2.001 (Schneider et al., 2000). folia populations in China as revealed by cpDNA variation? And (2) These were used to construct a parsimony network with the what does this structure indicate about the post-glacial history of aid of the TCS 1.06 (Clement et al., 2000). We further subjected S. trifolia in China? data to nested clade analysis (NCA). Clade distances (Dc), a mea- sure of the geographical spread of a clade, and nested clade dis- tances (Dn), a measure of the degree of a particular clade 2. Materials and methods relative to its closest sister clades, were defined based on geo- graphical locations of samples in the nesting cladogram, and 2.1. Plant materials were estimated as described in Templeton et al. (1995).Differ- ences between interior (ancestral) and tip (recent) clades, Dc The plant materials used in this investigation were from 42 and Dn distances, were calculated, where I and T were interior populations representing almost all the natural distribution area and tip clades, respectively. The null hypothesis of no geograph- of S. trifolia in China; these populations could be grouped into ical associations for tip clades and interior clades was tested by 中国科技论文在线 http://www.paper.edu.cn

J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx 3

Table 1 Names, locations, sample size, and the distribution of cpDNA haplotypes of each of the 42 investigated populations of Sagittaria trifolia

Populations Locations Coordinates Regions Sample size Haplotype number Haplotypes JL-1 Dunhua, Jilin 128°130E/43°190N Northeast 4 2 8, 18 JL-2 Dunhua, Jilin 128°150E/43°230N Northeast 3 2 5, 8 JL-3 Antu, Jilin 128°280E/42°570N Northeast 2 1 5 HLJ-1 Mishan, Heilongjiang 131°510E/45°300N Northeast 2 2 21, 25 HLJ-2 Majiagang, Heilongjiang 132°160E/45°270N Northeast 1 1 8 HLJ-3 Xingkaihu, Heilongjiang 132°160E/45°270N Northeast 2 2 20, 22 HLJ-4 Dongcheng, Heilongjiang 129°110E/44°060N Northeast 3 3 3, 5, 8 HLJ-5 Raohe, Heilongjiang 134°010E/46°440N Northeast 2 2 5, 8 HLJ-6 Tongjiang, Heilongjiang 134°240E/48°200N Northeast 2 1 5 HLJ-7 Bacha, Heilongjiang 133°180E/47°250N Northeast 3 2 5, 7 HLJ-8 Hulin, Heilongjiang 133°350E/45°470N Northeast 3 3 13, 15, 23 XJ-1 Kekesu, Xinjiang 87°190E/47°370N Northwest 1 1 27 GS-1 Zhangye, Gansu 100°270E/39°030N Northwest 2 1 8 HEB-1 Baiyangdian, Hebei 125°560E/38°540N North 2 1 8 YN-1 Lashihai, Yunnan 100°110E/26°140N Southwest 3 3 3, 5, 8 YN-2 Heqing, Yunnan 100°100E/26°360N Southwest 3 1 8 YN-3 Zizhi, Yunnan 98°330E/25°430N Southwest 2 2 8, 19 YN-4 Lijiang, Yunnan 100°150E/26°340N Southwest 2 2 5, 8 YN-5 Simao, Yunnan 100°560E/22°400N Southwest 3 2 4, 5 YN-6 Tengchong, Yunnan 98°330E/25°070N Southwest 3 3 5, 8, 14 YN-7 Simao, Yunnan 101°010E/22°450N Southwest 4 3 1, 2, 26 YN-8 Cibihu, Yunnan 99°560E/26°060N Southwest 3 2 5, 8 YN-9 Erhai, Yunnan 100°60E/25°420N Southwest 3 1 5 YN-10 Caohai, Yunnan 100°100E/26°360N Southwest 3 2 5, 8 GZH-1 Huaxi, Guizhou 106°400E/26°240N Southwest 3 2 1, 5 GZH-2 Guiyang, Guizhou 106°200E/26°250N Southwest 2 1 8 GZH-3 Guiyang, Guizhou 106°270E/26°290N Southwest 3 2 5, 8 GZH-4 Pingba, Guizhou 106°160E/26°250N Southwest 2 2 5, 8 HUB-1 Ezhou, Hubei 114°370E/30°240N Central 3 2 5, 8 HUB-2 Ezhou, Hubei 114°410E/30°200N Central 3 2 5, 8 HUB-3 Zhijiang, Hubei 111°350E/30°280N Central 3 2 5, 8 HUB-4 Shayang, Hubei 112°220E/30°310N Central 2 2 5, 12 HUN-1 Chaling, Hunnan 113°400E/26°160N Central 3 2 5, 8 HEN-1 Huangchuan, Henan 115°240E/32°70N Central 3 2 5, 10 JS-1 Baoying, Jiangsu 119°290E/33°240N East 2 2 16, 17 JS-2 Baoying, Jiangsu 119°290E/33°190N East 2 2 8, 11 JX-1 Dongxiang, Jiangxi 116°370E/28°140N East 3 2 5, 24 JX-2 Changbei, Jiangxi 115°510E/28°450N East 3 1 5 JX-3 Dongxiang, Jiangxi 116°310E/28°050N East 2 1 8 HAIN-1 Hainan 110°340E/19°210N South 4 3 5, 6, 9 GX-1 Lingchuan, Guangxi 110°180E/25°230N South 2 1 5 GX-2 Guilin, Guangxi 110°160E/25°050N South 2 2 5, 8

considering that the dispersion distance of clades was not great- The geographical distributions of the 27 haplotypes are shown er or less than expected by chance, and by comparing observed in Table 1. The most widespread haplotypes were haplotype 5 (in DcandDn values with a distribution of such values, calculated 27 of 42 populations) and 8 (in 24 of 42 populations). Haplotype for each 10,000 random permutations of clades against sampling 1 and 3 were all found in two populations (Haplotype 1: YN-7 locations at 0.05 level (Templeton et al. 1995). Permutation tests and GZH-1; Haplotype 3: YN-1 and HLJ-4), each of the remainders were conducted separately for each level of the nested clado- of the haplotypes was unique to a particular population (Table 1 gram using GeoDis version 2.5 (Posada et al., 2000). After signif- and Fig. 1). icance levels for DcandDn were determined, inferences about In the haplotype network, 28 one-step clades, 11 two-step the historical processes that were likely to be responsible for ob- clades, five three-step clades and two four-step clades were iden- served patterns of clade structure were made following methods tified from the minimum-spanning network (Fig. 3). Two groups of Templeton et al. (1995). were detected among the 27 haplotypes in the network. One group consists of haplotype 5 and its closely related haplotypes 3. Results and the other group was haplotype 8 and its diverged haplotypes (Fig. 2). The distribution ranges of the two groups overlap Sequences of the cpDNA atpB–rbcL intergenic spacer varied geographically. from 947 to 955 bp in length. The length after alignment of Hierarchical analyses of molecular variance (AMOVA) showed the sequences was 984 bp. There were 12 nucleotide substitu- little differentiation among the sampled populations or among tions and 27 indels. From these nucleotide substitutions and in- the populations within the regions (P < 0.001). An examination dels, a total of 27 haplotypes were identified (Table 1). The of the diversity among regions, among populations within a re- sequences of these haplotypes are deposited in the GenBank gion, and within populations indicated that a great amount of (Accession Nos. EU247764 to EU247790). Sequence variation variation (80.2%) occurred within populations. Variation among was determined as a non-significant deviation from expectations regions and between populations within a region was 3.8% and of neutrality, both by Tajima’s criterion (D = 1.845, 16.0%, respectively (Table 2). 0.1 > P > 0.05) and Fu and Li’s tests (D* = 2.370, 0.1 > P >0.05; All tip haplotypes (2, 4, 6–25, and 27) in the minimum-span- F* = 2.578, 0.1 > P > 0.05). ning network were unique to a particular population, while the 中国科技论文在线 http://www.paper.edu.cn

4 J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx

Fig. 1. The localities of populations sampled in this study and geographical distribution of the cpDNA haplotypes in each population of Sagittaria trifolia in China. Populations were assigned to seven geographical regions as described in the text.

haplotypes 5 and 8 of the interior nodes were widespread population JL-4, HLJ-6, YN-9, JX-1, and GX-1; haplotype 8: pop- (Table 1). Geographic structures were significant for the 2-step ulation HLJ-2, YN-2, GZH-2, GS-1, JX-3, and HEB-1). The most clade 2–6, 3-step clade 3–4, and the total clade (Table 3). parsimonious interpretation of these widespread distributions From the NCA analysis, the evolutionary events, such as long was that S. trifolia had spread into most of its extant distribu- distance colonization at the 3–5 clade (inference chain: 1–2– tions through colonization by a small number of founders. S. 11–12–13—YES) and the 4–2 clade levels (inference chain: 1– trifolia is a clonal aquatic herb. Vegetative reproduction of this 2–3–5–15—NO) and restricted gene flow/dispersal but species is by corms, which is potentially a mode of long dis- with some long distance dispersal at the 4–1 clade level (infer- tance dispersal. The simple haplotype composition (several pop- ence chain: 1–2–3–5–6–7—YES), were inferred to be responsi- ulations were composed of a single haplotype) might be ble or the geographic distribution of cpDNA haplotypes in attributed to founder effects or genetic drift in small popula- S. trifolia. tion. In the minimum-spanning network, interior positions cou- pled with the high frequency occurrence indicated haplotype 5 4. Discussion and 8 were the ancestral haplotypes, each of the haplotypes at tip positions were unique to their particular population (Fig. 3). Among the 27 cpDNA haplotypes revealed in S. trifolia in this According to the coalescent theory, haplotype 5 and 8 may rep- study, haplotype 5 and 8 were found to be most widespread. resent the ancestral genotypes and probably produce other Twenty-seven of the 42 populations investigated harbored hap- mutational derivatives following long distance dispersal and lotype 5 and 24 populations harbored haplotype 8. Of these, 12 colonization (Donnelly and Tavare, 1986; Crandall and Temple- populations were composed of a single haplotype (haplotype 5: ton, 1993). 中国科技论文在线 http://www.paper.edu.cn

J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx 5

5

4 1 3 12 24 8

17 11 16 26 7 19

2 23 15 13

9 20 22 21 18 14

10

25

6 27

Fig. 2. Minimum-spanning network of 27 cpDNA haplotypes found in Sagittaria trifolia. Solid bars represent single base mutations, ‘‘0” represents hypothetical ancestral nodes that were not detected or became extinct in populations.

Table 2 Analysis of molecular variance (AMOVA) for populations of Sagittaria trifolia based on cpDNA atpB–rbcL noncoding spacer sequences data (d.f., degree of freedom; SSD, sum of squares)

Source of variation d.f. SSD Variance component Percentage of variation Among regions 6 9.264 0.03429 3.83 Among populations within regions 35 38.051 0.14333 16.02 Within populations 66 47.333 0.71717 80.15

The geographic history of South China (including southeastern, glacial refuge areas harbor a large fraction of the intraspecific southern and southwestern China) has close similarity to that of diversity. The plant populations in refuge areas have high genet- the Atlantic coastal zone of North America and both regions were ic divergence and uniqueness, rather than a high number of proposed to be refugia area during the last glaciation (Chen, haplotypes. In this study, low levels of genetic divergence among 1989). Since the Cretaceous period the climatic conditions in Asia the sampled regions and among populations within regions were (especially in the region between latitude 20°(N) and latitude caused by widespread common haplotypes. The patterns of ge- 40°(N)) have been considerably stable, thus providing a suitable netic structure found in S. trifolia do not seem to fit the criterion environment for speciation (Chen, 1989). Although relatively few for locating regions of glacial refugia (Petit et al., 2003), although studies have been conducted on the post-glacial history of Chinese a large number of the haplotypes are unique to a single popula- plants, phylogeographical studies on a number of endangered or tion. Each region, except for the Northern region (only one pop- endemic species in China have revealed several refugia and post- ulation with small sample size was included in this study), glacial population growth. Based on a phylogeographical analysis harbors a number of the unique haplotypes. It was not likely of a conifer (Cunninghaia konishii), Lu et al. (2001) and Hwang that all the sampled regions in this study served as the refugia. et al. (2003) inferred several refugia in southern China and de- As stated above, the unique haplotypes in each population may tected post-glacial population growth after a glacial induced genet- be mutational derivatives from the ancestral haplotypes after ic bottleneck. Shen et al. (2005) also found possible refugia area of settlement by long distance dispersal rather than the relics in Ginkgo biloba in southern China. Wang and Ge (2006) identified refugia. four possible refugia for Cathaya argyrophylla, although they did The phylogeographical pattern of cpDNA haplotypes demon- not find evidence of long distance dispersal and population strates two ‘‘star-like” clusters that were the result of haplo- expansion. types being linked to a central haplotype (Figs. 2 and 3). This Eight species of Sagittaria including four endemics occur in relatively simple pattern could be explained by populations China and three of the endemics are in southern China (Chen, preserved in refugia that have experienced population expan- 1989). Based on the high proportion of the endemic Sagittaria sion after glaciations, where, there has been insufficient time species in southern China, Chen (1989) proposed that the south- to form a more complicated population structure (Dynesius ern region was the likely refugia area and the center of specia- and Jansson, 2000). However, this explanation seems unlikely tion for species of Sagittaria. Petit et al. (2003) suggested that for S. trifolia because in the minimum-spanning network the 中国科技论文在线 http://www.paper.edu.cn

6 J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx

Table 3 Table 3 (continued) Results of nested clade analyses of the geographical distance of cpDNA haplotypes of Nested clades Type of geographical distance Permutational Probability Sagittaria trifolia and haplotypes v2 statistic (*: <0.05) Within Clade Nested Clade Nested clades and Type of geographical Permutational Probability * (Dc) (Dn) haplotypes distance v2 statistic ( : <0.05) Clade 4–2 28.0889 0.5080 Within Nested 3–1 (Tip) 47.3390S 2040.8338L Clade (Dc) Clade (Dn) 3–5 (Interior) 1317.6467S 1351.4614S Clade 1–10 18.9868 0.8480 I–T 1270.3076L 689.3724S 8 (Interior) 1199.1754 1192.4084 Total clade 66.3940 0.0020* 19 (Tip) 0.0000 1475.0480 4–1 (Tip) 1398.8002 1387.0088 I–T 1199.1754 282.6396 4–2 (Tip) 1253.7714 1268.5667 Clade 1–18 3.0000 1.0000 I–T 145.0288 118.4421 3 (Interior) 1617.1984 1737.5665 7 (Tip) 0.0000 1429.8222 Note.S=DcorDn values that are significantly smaller than expected at the 5% level I–T 1617.1984 307.7444 based on 1000 permutations. L = DcorDn values that are significantly larger than expected at the 5% level based on 1000 permutations. Clade 1–19 50.5357 0.9780 4 (Tip) 0.0000 1631.3867 5 (Interior) 1214.3283 1202.9586 12 (Tip) 0.0000 200.4161 24 (Tip) 0.0000 438.0402 I–T 1214.3283 446.3442 central haplotypes were widely distributed in the sampled pop- Clade 1–24 0.4444 1.0000 ulations. The two clades of cpDNA haplotypes are not associ- 1 (Interior) 336.9476 323.9600 ated with two distinct geographical regions. The results of 26 (Interior) 0.0000 217.2255 NCA analysis indicate that long distance colonization has likely I–T resulted in the cpDNA haplotype structure of S. trifolia.Someof Clade 2–1 2.0000 1.0000 the extant populations may have served as refugia during the 1–3 (Tip) 0.0000 43.7598 interglacial period, although no evidence of refugia was found. 1–4 (Interior) 0.0000 65.4639 The extent of refugia populations may have been so small that I–T 0.0000 21.7041 the subsequent long distance dispersal might have replaced the Clade 2–2 2.0000 1.0000 pre-existing genetic structure and swamped evidence of 1–6 (Interior) 0.0000 214.7819 1–5 (Tip) 0.0000 290.6920 refugia. I–T 0.0000 75.9100 It should be mentioned that the sample sizes per population in Clade 2–3 106.9625 0.1410 the present study were relatively small. The low sampling within 1–7 (Tip) 0.0000 1556.4465 populations may be partially responsible for the sparse network 1–8 (Tip) 0.0000 2228.6088 phylogeny with many possible unsampled haplotypes and these 1–9 (Tip) 0.0000 2358.7376 small sample sizes may greatly reduce power for identification 1–10 (Interior) 1202.1078 1221.1209S S of refugia (because not enough were sampled to see differences I–T 1202.1078 826.8100 in haplotype diversity among populations). It will also reduce Clade 2–6 0.0000 0.0000* 1–15 (Tip) 0.0000 418.8688 the robustness of statistical tests in the NCA at lower clade levels. 1–16 (Tip) 0.0000 422.2226 However, the inferences of the phylogeography pattern of the S. 1–17 (Tip) 0.0000 422.2226 trifolia in China seems not be greatly limited by the low sample 1–18 (Tip) 1234.0922 1376.6104 sizes per population. Because the tip haplotypes from the mini- 1–19 (Interior) 1165.7200 1194.7673 mum-spanning network are generally restricted to a single popu- I–T 460.5245 227.6593 lation, the large number of the possible missing haplotypes in Clade 2–8 6.3750 0.6010 each population were unlikely the widespread haplotypes, thus, 1–22 (Interior) 0.0000 2583.3121 1–23 (Tip) 0.0000 1218.6635 the low samples might have little effect on the sparse network 1–24 (Interior) 299.3290 1004.1707 phylogeny. Although there were low sample sizes at lower clade I–T 239.4632 101.3355 levels in the NCA, the samples are enough at higher clade levels; Clade 3–1 3.0000 1.0000 thus, the demographic inferences from the nested clade distance 2–1 (Interior) 52.4415 47.8531 analysis in S. trifolia should be reasonable. In addition, during 2–11 (Tip) 0.0000 46.4823 the high rates of colonization of S. trifolia in China, it was possible I–T 52.4415 1.3708 that there were only very few haplotypes as the initial founders Clade 3–2 7.0000 0.4320 in each population. 2–7 (Tip) 0.0000 1791.0265 2–8 (Interior) 1464.0947 1749.9142 This study initiates a research on the post-glacial colonization I–T 1464.0947 41.1123 of a marsh herb in China. From the phylogeographic analyses, Clade 3–4 0.0000 0.0000* the evolutionary events, such as long-distance dispersal, have 2–4 (Tip) 0.0000 1481.6463 been important factors in the post-glacial history of S. trifolia 2–5 (Tip) 0.0000 25.2349 and the present-day distribution of genetic variation. S. trifolia 2–6 (Interior) 1148.0664 1144.3358 is widely distributed in Asia and Southeast Europe, it is expected I–T 1148.0664 390.8952 that more samples being collected outside China to construct the Clade 3–5 15.5669 0.9790 haplotype structure, and to infer the migratory routes and diver- L 2–2 (Tip) 247.3148 2070.2053 gence times, which will help us to explore the evolutionary his- 2–3 (Interior) 1277.9623 1292.5614 I–T 1030.6475 777.6439S tory of the species in more detail. In addition, more species with widely distribution ranges should be included in future work in Clade 4–1 68.8100 0.2450 3–2 (Tip) 1764.4245L 2047.3478L order to investigate phylogeographic patterns in herbs in China, 3–3 (Tip) 0.0000 1477.4566 this would be valuable to determine if the other herbaceous spe- 3–4 (Interior) 1129.8401S 1138.0700S cies also display similar migratory histories as revealed in S. S S I–T 414.0313 838.0414 trifolia. 中国科技论文在线 http://www.paper.edu.cn

J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx 7

25 1-20 10 O 27 9 1-1

2-7 1-11 1-13 3-1 O 1-21 O O O O 2-11

2-4 2-5 1-3 3-2 O O 22 O O O 1-2 2-8 3-4 1-14 2-1 1-12 O O 1-22 20 O O O 1-4 O 23 11 1-16 O 16 2 26 1-15 24 1-23 1-24 O 12 15 18 5 1-10 O 1 8 1-6 2-2 1-5 2-6 1-19 4 O O O O O 3 19 6 O O O 1-9 1-17 1-18 2-3 1-7 1-25 1-28 3-5 13 1-8 O 21 17 7 14 O 3-3 O

2-9 2-10

O O 4-1 1-26 1-27 4-2

O O

Fig. 3. Nested clades of 27 haplotypes of cpDNA of Sagittaria trifolia. One- (1–1 to 1–28), two- (2–1 to 2–11), three- (3–1 to 3–5) and four- step (4–1 and 4–2) clades are indicated (one-step clade: dotted line; two-step clade: thin solid line; there-step clade: bold line; four-step clade: separated by vertical line).

Acknowledgments Donnelly, P., Tavare, S., 1986. The ages of alleles and a coalescent. Adv. Appl. Prob. 18, 1–19. Dynesius, M., Jansson, R., 2000. Evolutionary consequences of changes in species’ We sincerely thank Chun-Feng Yang, Yi-Hong Zheng, Zhi-Yuan geographical distributions driven by Milankovitch climate oscillations. Proc. Du, Jun Li, and Xiao-Lin Zhang for their help in fieldwork and Natl. Acad. Sci. USA 97, 9115–9120. Xiao-Li Yue and Shu-Ying Zhao for their assistance in the labora- Excoffier, L., Smouse, P.E., Quattro, J.M., 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: application to human tory. This study was supported by Grants from the National Natu- mitochondrial DNA restriction data. Genetics 131, 479–491. ral Science Foundation of China (No.: 30570111) and the Program Fu, C.X., Qiu, Y.X., Kong, H.H., 2003. RAPD analysis for genetic diversity in Changium for New Century Excellent Talents in University (from Ministry of smyrnioides(Apiaceae) an endangered plant. Bot. Bull. Acad. Sin. 44, 13–18. Fu, Y.X., Li, W.H., 1993. Statistical tests of neutrality of mutations. Genetics 133, Education, People’s Republic of China) granted to W.Q.F. (NCET- 693–709. 05-0619). Ge, X.J., Chiang, Y.C., Chou, C.H., Chiang, T.Y., 2002. Nested clade analysis of Dunnia sinensis(Rubiaceae), a monotypic genus from China based on organelle DNA sequences. Conserv. Genet. 3, 351–362. References Griffin, S.R., Barrett, S.C.H., 2004. Post-glacial history of Trillium grandiflorum (Melanthiaceae) in eastern North America: inferences from phylogeography. An, Z.S., Kutzbch, J.E., Prell, W.L., Porter, S.C., 2001. Evolution of Asian monsoons and Am. J. Bot. 91, 465–473. phased uplift of the Himalaya–Tibetan plateau since late miocene times. Nature Grivet, D., Petit, R.J., 2003. Chloroplast DNA phylogeography of the hornbeam in 411, 62–66. Europe: evidence for a bottleneck at the outset of postglacial colonization. Cannon, H.C., Manos, P.S., 2003. Phylogeography of the Southeast Asian stone oaks Conserv. Genet. 4, 47–56. (Lithocarpus). J. Biogeogr. 30, 211–226. Harrison, T.M., Copeland, P., Kidd, W.S.F., Yin, A., 1992. Raising Tibet. Science 255, Chen, J.K., 1989. Systematic and Evolutionary Biology Studies on Chinese Sagittaria. 1663–1670. Wuhan University Press, Wuhan. Heuertz, M., Fineschi, S., Anzidei, M., Pastorelli, R., Salvini, D., Paule, L., Frascaria- Chiang, T.Y., Schaal, B.A., Peng, T., 1998. Universal primers for amplification and Lacoste, N., Hardy, O.J., Vekemans, X., Vendramin, G.G., 2004. Chloroplast DNA sequencing a noncoding spacer between atpB and rbcL genes of chloroplast variation and postglacial recolonization of common ash (Fraxinus excelsior L.) in DNA. Bot. Bull. Acad. Sin. 39, 245–250. Europe. Mol. Ecol. 13, 3437–3452. Clement, M., Posada, D., Crandall, K., 2000. TCS: a computer program to estimate Hewitt, G.M., 1996. Some genetic consequences of ice ages, and their role in gene genealogies. Mol. Ecol. 9, 1657–1660. divergence and speciation. Biol. J. Linn. Soc. 58, 247–276. Coleman, M., Hodges, K., 1995. Evidence for Tibetan plateau uplift before 14 Myr Hewitt, G.M., 2000. The genetic legacy of the Quaternary ice ages. Nature 405, 907– ago from a new minimum age for east–west extension. Nature 374, 49–52. 913. Comes, H.P., Kadereit, J.W., 1998. The effect of Quaternary climatic changes on plant Hewitt, G.M., 2004. Genetic consequences of climatic oscillations in the Quaternary. distribution and evolution. Trends Plant Sci. 3, 432–438. Philos. Trans. R. Soc. Lond. B Biol. Sci. 359, 183–195. Cook, C.D.K., 1990. Aquatic Plant Book. SPB Academic Publishing, The Hague, The Honjo, M., Ueno, S., Tsumura, Y., Washitani, I., Ohsawa, R., 2004. Phylogeographic Netherlands. study based on intraspecific sequence variation of chloroplast DNA for the Cox, C.B., Moore, P.D., 2000. Biogeography: An Ecological and Evolutionary conservation of genetic diversity in the Japanese endangered species Primula Approach. Blackwell Science, Oxford. sieboldii. Biol. Conserv. 120, 211–220. Crandall, K.A., Templeton, A.R., 1993. Empirical tests of some predictions from Huang, J.C., Wang, W.K., Peng, C.I., Chiang, T.Y., 2005. Phylogeography and coalescent theory with applications to intraspecific phylogeny reconstructure. conservation genetics of Hygrophila pogonocalyx (Acanthaceae) based on atpB- Genetics 134, 959–969. rbcL noncoding spacer cpDNA. J. Plant Res. 118, 1–11. 中国科技论文在线 http://www.paper.edu.cn

8 J.-M. Chen et al. / Molecular Phylogenetics and Evolution xxx (2008) xxx–xxx

Hwang, S.Y., Lin, T.P., Ma, C.S., Lin, C.L., Chung, J.D., Yang, J.C., 2003. Postglacial Shi, Y.F., Li, J.J., Li, B.Y., 1998. Uplift and Environmental Changes of Qinghai–Tibetan population growth of Cunninghamia konishii (Cupressaceae) inferred from Plateau in the Late Cenozoic. Guangdong Science and Technology Press, phylogeographical and mismatch analysis of chloroplast DNA variation. Mol. Guangzhou, China. Ecol. 12, 2689–2695. Soltis, D.E., Gitzendanner, M.A., Strenge, D.D., Solits, P.S., 1997. Chloroplast DNA Ikeda, H., Setoguchi, H., 2007. Phylogeography and refugia of the Japanese intraspecific phylogeography of plants from the Pacific Northwest of North endemic alpine plant, Phyllodoce nipponica Makino (Ericaceae). J. Biogeogr. 34, America. Plant Syst. Evol. 206, 353–373. 169–176. Su, Y.J., Wang, T., Zheng, B., Jiang, Y., Chen, G.P., Ouyang, P.Y., Sun, Y.F., 2005a. Lascoux, M., Palmem, A.E., Cheddadi, R., Latta, R.G., 2003. Impact of ice ages on the Genetic differentiation of relictual populations of Alsophila spinulosa in southern genetic structure of trees and shrubs. Trans. R. Soc. Lond. B Biol. Sci. 359, 197– China inferred from cpDNA trnL–F noncoding sequences. Mol. Phylogenet. Evol. 207. 34, 323–333. Lu, S.Y., Peng, C.I., Cheng, Y.P., Hong, K.H., Chiang, T.Y., 2001. Chloroplast DNA Su, Y.J., Wang, T., Zheng, B., Jiang, Y., Ouyang, P.Y., Chen, G.P., 2005b. Genetic phylogeography of Cunninghamia konishii (Cupressaceae), an endemic conifer of variation and phylogeographical patterns in Alsophila podophylla from Southern Taiwan. Genome 44, 797–807. China based on cpDNA atpB-rbcL sequence data. Am. Fern. J. 95, 68–79. Petit, R.J., Aguinagalde, I., de Beaulieu, J.L., Bittkau, C., Brewer, S., Cheddadi, R., Ennos, Sun, H., 2002. Tethys retreat and Himalaya–Hengduanshan mountains uplift and R.A., Fineschi, S., Grivet, D., Lascoux, M., Mohanty, A., Muller-Starck, G., their significance on the origin and development of the Sino–Himalayan Demesure-Musch, B., Palme, A., Martin, J.P., Rendell, S., Vendramin, G.G., elements and alpine flora. Acta Bot. Yunnanica 24, 273–288. 2003. Glacial refugia: hotspot but not melting pots of genetic diversity. Science Taberlet, P., Fumagalli, L., Wust-Saucy, A.G., Cosson, J.F., 1998. Comparative 300, 1563–1565. phylogeography and postglacial colonization routes in Europe. Mol. Ecol. 7, Posada, D., Crandall, K.A., Templeton, A.R., 2000. GeoDis, a program for the cladistic 453–464. nested analysis of the geographical distribution of genetic haplotypes. Mol. Ecol. Tajima, F., 1989. Statistical method for testing the neutral mutation hypothesis by 9, 487–488. DNA polymorphism. Genetics 123, 585–595. Ran, J.H., Wei, X.X., Wang, X.Q., 2006. Molecular phylogeny and biogeography of Templeton, A.R., Routman, E., Philips, C.A., 1995. Separating population structure Picea (Pinaceae) implications for phylogeographical studies using cytoplasmic from population history: a cladistic analysis of the geographical distribution of haplotypes. Mol. Phylogenet. Evol. 41, 405–419. mitochondria DNA haplotypes in the tiger salamander, Ambystoma tigrinum. Ren, M.E., Yang, Z.Z., 1961. Problems on the physio-geographic regionalization of Genetics 140, 767–782. China. Acta Geogr. Sin. 27, 66–74. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: improving the Rozas, J., Sanchez-DeIBarrio, J.C., Messeguer, X., Rozas, R., 2003. DNASP, DNA sensitivity of progressive multiple sequence alignment through sequence polymorphism analyses by the coalescent and other methods. Bioinformatics weighting, position-specific gap penalties and weight matrix choice. Nucleic 19, 2496–2497. Acids Res. 22, 4673–4680. Schneider, S., Roessli, D., Excoffier, L., 2000. ARLEQUIN version 2.000: a software for Wang, H.W., Ge, S., 2006. Phylogeography of the endangered Cathaya population genetic data analysis. Genetics and Biometry Laboratory, argyrophylla(Pinaceae) inferred from sequence variation of mitochondrial and Department of Anthropology, University of Geneva, Switzerland. nuclear DNA. Mol. Ecol. 15, 4109–4122. Schonswetter, P., Stehlik, I., Holderegger, R., Tribsce, A., 2005. Molecular evidence Wu, Z.Y., 1979. On zonal problem of flora in China. Acta Bot. Yunnanica 1, 1–22. for glacial refugia of mountain plants in the European alps. Mol. Ecol. 14, 3547– Zhang, D.F., Fengquan, L., Jianmin, B., 2000. Eco-environmental effects of the 3555. Qinghai–Tibet plateau uplift during the Quaternary in China. Environ. Geol. 39, Shackleton, N.J., Opdyke, N.D., 1973. Oxygen isotope and palaeomagnetic 1352–1358. stratigraphy of equatorail Pacific core v28–238 isotope temperatures and ice Zhang, Q., Chiang, T.Y., George, M., Liu, J.Q., Abbott, R.J., 2005. Phylogeography of volumes on a 105 year and 106 year scale. Quat. Res. 3, 39–55. the Qinghai–Tibetan plateau endemic Juniperus przewalskii (Cupressaceae) Shen, L., Chen, X.Y., Zhang, X., Li, Y.Y., Fu, C.X., Qiu, Y.X., 2005. Genetic variation of inferred from chloroplast DNA sequence variation. Mol. Ecol. 14, 3513–3524. Ginkgo biloba L. (Ginkgoaceae) based on cpDNA PCR-RFLPs: inference of glacial Zheng, L., Li, B.Y., 1999. Progress in studies on geographical environments of the refugia. Heredity 94, 396–401. Qinghai–Xizang plateau. Chin. Geogr. Sci. 9, 289–296.