Pedicularis section Cyathophora is a monophyletic group characterized by perfoliate leaf and/or bract bases at each node. This section comprises four series, corresponding to four general corolla types of Pedicularis, i.e. toothless, toothed, beaked and long-tubed corollas. In this study, we aim to reconstruct a comprehensive phylogeny of section Cyathophora, and compare phylogenetic incongruence between nuclear and chloroplast datasets. Sixty-seven accessions belonging to section Cyathophora and 9 species for other Pedicularis were sampled, and one nuclear gene (nrITS) and four chloroplast genes (matK, rbcL, trnH-psbA and trnL-F) were sequenced. Phylogenetic analyses show that the topologies and networks inferred from nrITS and the concatenated chloroplast datasets were incongruent, and the nrITS phylogenies and network agreed with the morphology-based taxonomy to some degree. The chloroplast genome of two Sichuan samples of P. cyathophylloides (E4 and E5) may show introgression from an ancestor of P. cyathophylla. Neither the nrITS dataset nor the concatenated chloroplast dataset were able to adequately resolve relationships among species in the series Reges; this is most likely due to incomplete lineage sorting and/or introgression/hybridization. The nrITS phylogeny indicates the beakless (toothed and toothless) and beaked galeas may have evolved independently within section Cyathophora, and the chloroplast phylogeny reveals that the long corolla tube with beaked galea is derived from the short one.
Citation: Yu W-B, Huang P-H, Li D-Z, Wang H (2013) Incongruence between Nuclear and Chloroplast DNA Phylogenies in Pedicularis Section Cyathophora (Orobanchaceae). PLoS ONE 8(9): e74828. https://doi.org/10.1371/journal.pone.0074828
Editor: Jonathan H. Badger, J. Craig Venter Institute, United States of America
Received: May 20, 2013; Accepted: August 6, 2013; Published: September 19, 2013
Copyright: © 2013 Yu et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This study is supported by grants from the National Natural Science Foundation of China (31200185), the West Light Foundation of the Chinese Academy of Sciences, the National High-tech Research & Development Program of China (2012AA021801), the National Science and Technology on Basic Research Program (2013FY112600), the Key Research Program of the Chinese Academy of Sciences (KJZD-EW-L07), the Large-Scale Scientific Facilities of the Chinese Academy of Sciences (2009-LSF-GBOWS-01), and the National Key Basic Research Program of China (2014CB954100). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Pedicularis (Orobanchaceae) comprises approximately 600-800 species, one of the largest genera of flowering plants in the north temperate zone [1,2]. Flower morphology of Pedicularis exhibits great variation among species in the shape of upper corolla lip (galea) and the length of corolla tube. Species may be described as having one of four different corolla morphologies [1,3,4]: (A) short tubular corolla with beakless and a toothless galea, (B) short tubular corolla with a toothed galea, (C) short tubular corolla with a beaked galea, and (D) long tubular corolla with a beaked galea (see Figure 1). The high degree of parallel evolution in floral morphology of Pedicularis makes traditional infrageneric classifications controversial . Noteworthily, section Cyathophora H. L. Li  is one of few monophyletic subgeneric taxa supported by molecular phylogenies [3,7,8]. This section is characterized by the bases of whorled leaves and/or bracts that are dilated and fused together to form a cup-like structure around the stem at each node (Figure 2). Despite consisting of a few species and restricted to the eastern Himalaya-Hengduan Mountains, section Cyathophora is highly diverse in floral shape, including all four general corolla types (Figure 1) [1,6,9].
(A) short tubular corolla with a beakless and toothless galea (P. cyathophylloides); (B) short tubular corolla with a toothed galea (P. rex subsp. rex); (C) short tubular corolla with a beaked galea (P. superba); and (D) long tubular corolla with a beaked galea (P. cyathophylla).
Topology shows the majority rule consensus of the Bayesian inference tree, with thicker lines for the posterior support ≥ 0.95. Bootstrap values (≥50) of maximum likelihood/parsimony are indicated above branches. Sample code indicated in the parenthesis after species name and on the species photo is summarized in Table S1.
Before the system of Li , authors placed the species of section Cyathophora in the series Superbae Maxim. as a cohesive unit [10-13]. Li  categorized eight species of section Cyathophora into four series in accordance with four general corolla types (Table 1). Tsoong  adopted the four-series classification of section Cyathophora by Li , with a revised circumscription of these series (Table 1). In addition, Tsoong  disagreed with Li  on the likely phylogenetic position of the monotypic series Cyathophylloides H. L. Li, which Tsoong argued was the most basal of the section He classified P. cyathophylloides as the beaked type, given its long corolla tube, swollen galea, obscure beak, and spreading lower lip. Pedicularis xiangchengensis H. P. Yang was initially placed in series Cyathophyllae , but is now considered as a synonym of P. cyathophylla Franch. .
|Sect. Cyathophora H. L. Li||System of Li (1948)||System of Tsoong (1963)||In this study||Corolla types|
|Ser. Reges H.L. Li||P. rex Maxim. var. rex (syn.: P. lopingensis Hand.-Mazz., P. mahoangensis Bonati)||P. rex subsp. Rex (syn.: P. lopingensis Hand.-Mazz., P. mahoangensis Bonati)||P. rex subsp. Rex (syn.: P. lopingensis, P. mahoangensis, P. rex var. parva, P. rex var. pseudocyathus)||B|
|P. rex var. rockii (Bonati) H.L. Li (≡P. rockii Bonati)||P. rex subsp. rex var. rockii (Bonati) H.L. Li (≡ P. rockii)||P. rex var. Rockii (≡ P. rockii Bonati)||A|
|P. rex var. parva Bonati||P. rex subsp. parva (Bonati) P.C. Tsoong (≡ P. rex var. parva)||B|
|P. rex var. pseudocyathus Bonati||P. rex subsp. pseudocyathus (Bonati) P.C. Tsoong (≡ P. rex var. pseudocyathus)||B|
|P. lipskyana Bonati (syn.: P. lamarum Limpr.)||P. rex subsp. lipskyana (Bonati) P. C. Tsoong (≡ P. lipskyana)||P. rex subsp. Lipskyana (≡ P. lipskyana; syn.: P. lamarum, P. rex subsp. zayuensis)||B|
|P. rex subsp. zayuensis H.P. Yang (1990)||B|
|P. thamnophila (Hand. -Mazz.) H. L. Li (≡ P. rex var. thamnophila Hand. -Mazz.)||P. thamnophila (≡ P. rex var. thamnophila)||P. thamnophila (≡ P. rex var. thamnophila; syn.: P. cupuliformis)||B|
|P. cupuliformis H.L. Li||P. thamnophila subsp. cupulifomis (H.L. Li) P. C. Tsoong (≡P. cupuliformis)||B|
|Ser. Cyathophylloides H. L. Li||P. cyathophylloides Limpr.||P. cyathophylloides||P. cyathophylloides||A or C|
|Ser. Superbae Maxim.||P. superba Franch. ex Maxim.||P. superba||P. superba||C|
|P. connata||P. connata (?)||C|
|Ser. Cyathophyllae H. L. Li||P. connata H.L. Li|
|P. cyathophylla Franch.||P. cyathophylla Franch.||
|P. xiangchengensis H.P. Yang (1990)||D|
Li  hypothesized that the ancestral condition in section Cyathophora was toothless (series Cyathophylloides), and subsequently transformed to toothed (series Reges) and to beaked (series Superbae), culminating in the long tubular type (series Cyathophyllae), which is the general evolution model of floral characters in Pedicularis [1,6]. Tsoong  concurred Li’s hypothesis, whereas he placed the obscure-beaked series Cyathophylloides at the middle stage between series Reges and series Superbae. In addition, Tsoong  agreed that floral evolution of section Cyathophora may have begun with the toothless type, because P. rex var. rockii is toothless in series Reges. However, Macior  suspected that the extremely long-tubed and beaked P. cyathophylla was derived from the short-tubed and toothed P. rex. He suggested that such a saltatory evolution of floral form is possible and the gradual sequence may not represent a historical series of events. Ree  demonstrated that the long-tubed corolla appears to derive from a short-tubed corolla, and suggested that the pollination advantage of the beakless galea may facilitate to change from beaked to beakless in section Cyathophora.
Despite a long history of uncertainty regarding the evolution of floral form in section Cyathophora, few individuals have ever been included in previous phylogenetic reconstructions [3,7,8]. All phylogenies strongly support monophyly of section Cyathophora, however, relationships among its members are poorly understood. Phylogenetic incongruence was found among nuclear ribosomal internal transcribed spacer (nrITS), and chloroplast matK and trnT-trnF datasets. For example, the nrITS and matK datasets strongly support P. rex + P. thamnophila (i.e. series Reges) together [7,18], while the trnT-trnF dataset places P. rex sister to a clade P. thamnophila – P. cyathophylla ; the nrITS dataset strongly supports P. cyathophylla + P. superba together, while the matK dataset places P. cyathophylla sister to P. rex + P. thamnophila . In this study, we aim to reconstruct a comprehensive phylogeny of section Cyathophora on the basis of extensive sampling (multiple samples for all taxa) and more genetic coverage (nrITS, matK, rbcL, trnH-psbA and trnL-F) with a specific emphasis on series Reges. The main goals of this analysis are to: (i) explore patterns and causes of phylogenetic incongruence between nuclear and chloroplast datasets; (ii) compare phylogenetic trees of series Reges with morphology-based taxonomy; and (iii) discuss evolution of floral characters in this section based on its reconstructed phylogeny.
Materials and Methods
Taxon sampling and ethics statement
A total of 76 accessions, 67 belonging to section Cyathophora and 9 outgroups, were sampled in this study (Table S1). Pedicularis connata has only been collected from the type locality near Muli in southwestern China, and was not included here. No samples of endangered or protected species were included . Each population of the studied species had more than several hundreds of individuals, and around three plants per population were collected as the vouchers, which were deposited in the herbarium of the Kunming Institute of Botany (KUN), Chinese Academy of Sciences. Samples were collected from public land instead of protected areas in the southwest China; therefore, field permits were not required. The morphology of vouchered specimens was compared with the type specimens as an aid in identification. Pedicularis cyathophylla, P. cyathophylloides and P. superba are easy to identify. In series Reges, P. thamnophila is distinguished from P. rex by having pilose stems, leaves and cup-like bracts, long and persistent basal leaves, a corolla lower lip spreading at right-angle to the corolla tube, and a small flower. Comparisons of the type specimens and field observations indicated that the type of P. cupuliformis consists of some short plants of P. thamnophila. In this study, therefore, we treat P. cupuliformis is conspecific with P. thamnophila. Noteworthily, we collected one unknown taxon, whose floral structure and size resemble those of P. thamnophila, whereas its stems, leaves and bracts are glabrous or sparely pubescent, and basal leaves are short or absent. For infraspecies of P. rex, purple/pink corollas and toothless corolla galeas were use to identify subspecies lipskyana and variety rockii, respectively. We did not accept the treatment of subspecies parva, pseudocyathus and zayuensis in P. rex, because the type specimens of the first two subspecies cannot be distinguished from specimens of the type of subspecies rex, and those of the third subspecies were identical to those of subspecies lipskyana. Herein, five species (including three infraspecific taxa for P. rex) (Table S1) and one unknown taxon of section Cyathophora were included. Nine species of Pedicularis were selected as outgroups. Of the list of sampled taxa is presented in Table S1, including series position, voucher information, population code and GenBank accession number. Geographical information of samples from section Cyathophora is indicated in Figure 3.
Detailed collection of sample code is presented in Table S1. The black, green and light blue dash lines separates sample of P. rex subsp. rex, P. thamnophila and the unknown taxon in accordance with the nrITS phylogeny, respectively; and the blue, red and orange dash lines separates samples of P. cyathophylla, P. cyathophylloides and P. superba in accordance with the chloroplast phylogeny, respectively.
Total genomic DNA was extracted using a modified CTAB method  from silica gel dried leaves obtained from fresh field-collected specimens. One nuclear gene (nrITS) and four chloroplast genes (matK, rbcL, trnH-psbA and trnL-F) were sequenced in this study (Table 2). Protocols of PCR ampliﬁcation and sequencing were detailed in Yu et al. . All new raw sequences were assembled and edited using SeqMan (DNAstar packages). Preliminary alignments were produced using MAFFT version 6.0 , then adjusted manually in BioEdit version 7.0 . Sequence data matrixes were concatenated using SequenceMatrix version 1.7 .
|matK||rbcL||trnH-psbA||trnL-F||Combined chloroplast dataset|
|Aligned length (bp)||629||1081||645||725||912||3363||3992|
|Variable sites/Parsimony informative sites|
|Sect. Cyathophora + Outgroups||76||161/92||139/50||37/16||149/60||107/38||432/164||593/256|
|P. sp. (unknown taxa)||7||5/3||2/0||1/1||13/11||0/0||16/12||21/15|
Reconstruction of phylogenetic trees
Bayesian inference (BI), maximum likelihood (ML) and maximum parsimony (MP) methods were used to reconstruct phylogenetic trees of Pedicularis section Cyathophora. Phylogenetic resolution of the single chloroplast datasets was low, and no or low phylogenetic conflict was found among four chloroplastic genes for all three phylogenetic methods. Therefore, chloroplast loci were concatenated into a single chloroplast dataset. Chloroplast and nrITS datasets were analyzed separately. All characters were equally weighted and treated as unordered. Gaps were not coded. For BI, model selection was based on the Akaike information criterion (AIC) estimated by jModelTest [25,26]. The BI analyses were performed using MrBayes version 3.1 , with 10,000,000 Markov chain Monte Carlo (MCMC) generations and four incrementally heated chains. MCMC started from a random tree and sampling one of every 1000 generations, with the first 10% of the trees discarded as burn-in. The remaining trees were used to generate a majority-rule consensus tree. Consensus clades with posterior probability (PP) ≥ 0.95 are considered as strongly supported [28-32]. ML tree searches and bootstrap estimation of clade support were conducted using RAxML  on the CIPRES Science Gateway (http://www.phylo.org). Rate heterogeneity and proportion of invariant sites were included as model parameters, with other options left in the default settings. Support values for the node and clade were estimated from 100 bootstrap replicates. The MP analysis was conducted using PAUP* version 4.0b10 . Heuristic searches were implemented with 1000 random addition sequence replicates, and tree bisection-reconnection (TBR) branch swapping with the MULTREE option. Support for the node and clade was evaluated using bootstrap analysis  of 1000 replicates, each with 500 random addition sequence replicates and TBR branch swapping.
Phylogenetic network analysis
Phylogenetic networks were constructed to construct the evolutionary history of section Cyathophora using the SplitsTree version 4.12 . The Neighbor-net model was performed using the Kimura 2-parameter (K2P) distances and Ordinary Least Square Method. The nrITS and the concatenated chloroplast datasets were analyzed separately. Bootstrap values for the respective splits were estimated from 1000 bootstrap replicates.
Of the 380 sequences from 76 accessions analyzed in this study (Table S1), sequences of three bar coding loci (nrITS, rbcL and trnH-psbA) for 32 accessions were previously published . The remaining 284 sequences (accession numbers: KC733277-KC733352, KF011707-KF011914) were generated for this study (Table S1). Sequence characteristics of nrITS and single and concatenated chloroplast datasets are summarized in Table 2.
In the raw sequences of nrITS, ambiguous basecalls were the result of multiple superimposed peaks in chromatograms. One such base was found in five samples (E4, L2, R23, S2 and U5), two were found in three samples (R6, S5 and T7), and three were found in two sample (C1 and R5). The ambiguous site was assigned using IUPAC ambiguity characters. At the level of section and series, sequences of nrITS are more variable and informative than those of any single chloroplast gene in section Cyathophora and series Reges. Nevertheless, sequences of some chloroplast genes are more variable and informative than those of nrITS at the species level for some species. For example, psbA-trnH provides the most variation and parsimony informative characters in Reges. Similarly, matK provides the most variation and parsimony informative characters in P. cyathophylla and P. superba. Further, sequences of all chloroplast genes were more variable and informative than those of nrITS in P. cyathophylloides.
Phylogenetic analysis of nrITS sequences
For nrITS dataset, monophyly of section Cyathophora is strongly supported using BI analysis (Figure 2A). Subsequently, section Cyathophora is divided into two well-supported clades. Clade II corresponds to series Reges, while Clade I includes the other three series, and Clade II corresponds to series Reges. In clade I, monophyly of P. cyathophylloides is strongly supported by all analyses, but monophyly of P. cyathophylla and P. superba is strongly-supported in BI analysis alone (Figure 2A). In addition, samples E4 and E5 form a Sichuan clade in P. cyathophylloides, and samples S1 and S2 from Lijiang in southwestern Yunnan group together in P. superba. Within the clade of series Reges, four subclades are identified (Figure 2A). Subclade IIa contains two samples of P. rex var. rockii from Nixi (K1) of Shageri-La in Northwestern Yunnan. Subclade IIb corresponds to P. thamnophila; within this clade, five samples (T3-T7) from northwestern Muli of southwestern Sichuan are grouped together, and the other two samples from southeastern Muli (Changhaizi) are placed at the base of this subclade (Figure 3). Subclade IIc comprises of 16 samples of P. rex subsp. rex from the southern margin of distribution area, i.e., western Guizhou, western and central Yunnan, and Southestern Xizang (see Figure 3). Subclade IInd includes all selected infraspecific taxa of P. rex and all seven samples of the unknown taxon. In this subclade, two samples of P. rex var. rockii from Baishuitai (K2) of Shageri-La and all five samples of P. rex subsp. lipskyana are monophyletic, respectively. Monophyly of the unknown taxon is not supported, the seven samples occurring in four groups. With the exception of the sample ‘LIDZ0990’ from Yongnian (U7) of Ninglang is clustered with sample R23 of P. rex subsp. rex in a separate clade, the other six samples s occur in a poorly supported clade along with six specimens of P. rex ssp. rex, forming groups U1 + U2 + U3 and U4 + U6, and sample U5 alone.
Phylogenetic analysis of chloroplast sequences
The monophyly of section Cyathophora is strongly supported by all analyses using the concatenated chloroplast dataset (Figure 2B). Within section Cyathophora, series Reges (Clade 4) and P. superba (Clade 1) are resolved as monophyletic with high support values; and five samples of P. cyathophylloides separate into two groups: three Xizang samples (E1, E2 and E3) form an independent clade, and two Sichuan samples (E4 and E5) are clustered in a clade including all accessions of P. cyathophylla (Figure 2B). Pedicularis superba (Clade 1) is sister to the remainder of section Cyathophora, followed by three Xizang samples of P. cyathophylloides (Clade 2). A clade including the paraphyletic clade of P. cyathophylla (Clade 3), and two (E4 and E5) of the five accessions of P. cyathophylloides is strongly supported as the sister group of series Reges (Clade 4) In the clade of series Reges, no taxon is resolved as monophyletic (Figure 2B). Samples R8 and R10 of P. rex spp. rex are placed at the base of series Reges, followed by K1A of P. rex var. rockii and sample T5 of P. thamnophila. The remaining samples belong to a well-supported clade; however, relationships within this clade correspond poorly with named taxa.
Phylogenetic networks of section Cyathophora
Both nrITS and the concatenated chloroplast datasets split section Cyathophora into two major groups, series Reges and P. cyathophylla + P. cyathophylloides + P. superba (Figure 4). Incongruence between nrITS and the chloroplast networks was evident. The nrITS placed P. cyathophylloides close to P. cyathophylla + P. superba, whereas the chloroplast dataset split cyathophylloides into two groups, two Sichuan samples (E4 and E5) grouped with P. cyathophylla, and three Xizang samples (E1, E2 and E3) close to P. superba. Within the series Reges subnetwork of the nrITS analysis, monophyletic groups of P. rex subsp. lipskyana, subsp. rex and P. thamnophila form subnetworks of their own, while none of these groups occur in the chloroplast network.
Bootstrap support values for clusters are indicated next to the respective cluster delimitation (dashed lines); monophyly of accepted taxa is annotated in gray, except P. cyathophylloides in the chloroplast network. Scale bar indicates changes.
Informativeness of DNA markers
The nrITS region has been screened multiple times in previous studies of Pedicularis, and it is in general the most informative DNA marker for phylogeny [3,7,8] and species discrimination  in Pedicularis to date. Within section Cyathophora, the nrITS dataset also shows greater polymorphism than the other four screened chloroplast markers. Curiously, intraspecific sequences variation of trnH-psbA are higher than those of nrITS in all species, and those of all four chloroplast markers are higher than those of nrITS in P. cyathophylloides. Our results show that trnH-psbA sequences have high rates of intra-specific insertion/deletion , which may make difficult to align for large scale phylogeny of Pedicularis. The high amount of polymorphisms in the chloroplast markers of P. cyathophylloides results from polyphyletic relationships among five samples. The rbcL gene shows low variation within the genus Pedicularis , however, it does provide more informative sites than trnH-psbA and trnL-F in P. cyathophylloides, which indicating that the chloroplast genomes of two subclades in P. cyathophylloides may be heterogenous.
Incongruence between nuclear and chloroplast DNA phylogenies
Monophyly of section Cyathophora is well supported by all phylogenetic analyses in this study, as well as in previous studies [3,7,8]. Furthermore, monophyly of series Reges is strongly supported by all analyses. We did not concatenate nrITS and chloroplast datasets to infer phylogenies, because phylogenetic topologies between nrITS and the concatenated chloroplast datasets are heavily incongruent (see Figure 2). The only strongly congruent results between separate analysis of plastid and nuclear DNA are: (i) series Reges is monophyletic in both analyses (although no taxon is supported as monophyletic in either analysis), and (ii) all samples of P. superba form clades in both analyses. The chloroplast phylogeny was incongruent with morphology-based taxonomy, except P. superba, whereas the nrITS phylogenies agreed with the morphology-based taxonomy to some degree. The nrITS phylogeny successfully delimits four species and one infraspecific taxon as monophyletic: P. cyathophylloides, P. cyathophylla, P. superba and P. thamnophila, and P. rex spp. lipskyana. Furthermore, the nrITS phylogeny supports the contentions of Tsoong  that P. cyathophylloides is close to P. cyathophylla and P. superba, a result that has also been supported by pollen morphology . In addition, nrITS sequences of the sample ‘HW10215’ of P. cyathophylla (C3) collected from the type locality of ‘P. xiangchengensis’  is indistinct from the other six samples of P. cyathophylla in agreement with previous morphological comparisons . Our results support the prediction of Petit and Excoffier  that biparentally inherited nuclear loci experiencing high rates of intraspecific gene flow should enhance species delimitation, and that maternally inherited chloroplast loci should be more frequently introgressed and hence of more limited use in species delimitation than nuclear loci. In addition, incomplete lineage sorting in the slower evolving chloroplast genes may cause phylogeny of chloroplastic gene incongruent with that of nuclear gene and the morphology-based taxonomy in section Cyathophora [38-40].
Chloroplast capture in P. cyathophylloides
Incongruence between nuclear and chloroplast phylogenies might be explained as convergent evolution, lineage sorting, or reticulate evolution [39,41-43]. However, such processes cannot explain the incongruence in the placement of P. cyathophylloides. Chloroplast phylogenies separate five samples of P. cyathophylloides into two groups, two Sichuan samples and three Xiang samples, and the Sichuan group falls into the clade P. cyathophylla. Haplotypes did not vary within sites, evidence of strong geographical partitioning. Herein, we suggest that chloroplast capture, the introgression of a chloroplast from one species into another following a hybridization event followed by backcrossing of F1s with parental types, may be the most likely explanation for the pattern observed here [44-46]. The geographical pattern indicates that the chloroplast of the Sichuan group of P. cyathophylloides may be introgressed from that of P. cyathophylla. Collection records show that the distributions of P. cyathophylloides and P. cyathophylla overlap in western Sichuan. Hybridization may be mediated by bumblebee pollinators. However, the nrITS phylogenies clearly reject the recent hybridization between two species. Therefore, chloroplast capture in P. cyathophylloides is likely the result of an ancient hybridization/introgression event; and the current overlapping distribution may be a region the secondary contact region, in which evolution of floral character displacement effectively mediates reproductive isolation between two species.
Phylogenetic implications in series Reges
Series Reges is a monophyletic group supported by all analyses in this study and it is characterized by a yellow corolla (except P. rex subsp. lipskyana), short corolla tube, toothed galea (except P. rex var. rockii) and small lower lip. The nrITS phylogenies partially support the morphology-based taxonomic treatment in series Reges. While the number of phylogenetically informative sites and clade resolution of the concatenated chloroplast dataset are higher than those of nrITS dataset in series Reges, none of the morphology-based taxa is resolved in the chloroplast phylogeny. Even two individuals from the same population of P. rex subsp. rex (R9 and R17) and of P. rex var. rockii (K1 and K2) occur in different subclades of Clade 4. Phylogenetic incongruence with morphology-based taxonomy in series Reges may be caused by incomplete lineage sorting  and/or hybridization/introgression [37,42,47-49], and/or other biological factors [39,44,46,50-52].
In the nrITS phylogenies, taxonomical treatments for P. thamnophila and P. rex subsp. lipskyana are supported, while P. rex spp. rex and P. rex var. rockii are polyphyletic. Noteworthily, the subclade IInd of the nrITS phylogeny contains all recognized infraspecific taxa of P. rex and the unknown taxon, which are overlapped in the Hengduan Mountains (see Figure 3). By contrary, samples of P. rex subsp. rex from south margins of distribution range are fallen into the subclade IIc. The phylogenetic network shows that evolutionary history of subclade IInd was more complicated than that of other subclades. We suggested that paraphily of P. rex spp. rex may be explained by incomplete lineage sorting during population expansion, and polyphyly of P. rex var. rockii may be caused by ancient hybridization/introgression or convergent evolution . The nrITS dataset shows that the unknown taxon is close to P. rex, than to P. thamnophila, and the sample ‘LIDZ0990’ (U7) is distinct from its other six samples, most closely related to the sample ‘LIDZ1011’ of P. rex subsp. rex (R23). Morphologically, specimens of the sample ‘LIDZ0990’ have a spreading lower lip that resembles closely that of the unknown taxon, however, their dense inflorescences and large corollas  resemble those of P. rex subsp. rex. Both morphology and nrITS phylogeny imply that the sample ‘LIDZ0990’ may be an evolutionary intermediate or a hybrid between P. rex subsp. rex and the unknown taxon.
Phylogeographic patterns in section Cyathophora
Phylogenetic relationships are correlated with geographic distribution in P. rex subsp. rex and var. rockii and P. thamnophila using nrITS dataset, and in P. cyathophylla, P. cyathophylloidesa and P. superba using chloroplast dataset. In the nrITS phylogeny, samples of P. rex subsp. rex from the south margin of the distribution occur in subclade IIc, while the remaining samples are included in subclade IInd. This pattern indicates that southern samples of P. rex subsp. rex may still be subject to gene flow, or may have rapidly expanded in recent years. Meanwhile, the most recent common ancestor (MRCA) of the northern samples in subclade IInd may have experienced rapid species divergence, with four other taxa derived from this MRCA. Four samples of P. rex var. rockii occur in two widely separated subclades corresponding with geographic localities. This pattern indicates that gene flow between two populations in this taxon is lower, and samples of K2 may have gene exchange with the other taxa in the subclade IInd. For P. thamnophila, five samples from northwestern Muli (T3-T7) are separated from samples T1 and T2 in southeastern Muli by a long branch with strong support, indicating gene flow among the five northwestern samples is stronger than that between the two groups, or the five northwestern samples may be derived from a MRCA that separated from the MRCA of samples T1/T2 and rapidly expanded in recent years.
In the chloroplast phylogenies, the phylogeography of P. cyathophylloides is likely caused by chloroplast capture in two Sichuan samples (E4 and E5) (see above). Incidentally, the nrITS phylogeny also supports samples E4 and E5 together. For P. cyathophylla, three southwestern samples in Daxueshan are divided into two groups: two samples (C2 and C3) group together, while sample C1 is clustered with the other four northeastern samples (C4 ‒ C7). The geographic pattern of P. superba is close to that of P. cyathophylla. Two southwestern samples (S1 and S2) from Lijiang, Yunnan are isolated from the other three northeastern samples collected in Muli and Jiulong of Sichuan.
The geographic distribution of maternally inherited chloroplast haplotypes provides key evidence for investigating history of migration and population expansion at species level. Several phylogeographic studies have documented that the south margin of the Hengduan Mountains was one of the major glacial refugia for high altitude adapted during ice-age cycles of the Quaternary period [55-58]. In this study, phylogenetic networks inferred from both plastid and nuclear datasets indicate that evolutionary lineages of samples C2 and C3 of P. cyathophylla and samples S1 and S2 of P. superba may be the earliest-derived in their respective lineages. Therefore, Yulongshan in Lijiang and Daxueshan in the boundary of Yunnan and Sichuan may be or be close to the origin/refuge for P. cyathophylla and P. superba, respectively. For series Reges, phylogenetic topology and network do not support any morphology-based taxa and geographical pattern. Reticulate evolution of this group suggests that incomplete lineage sorting and/or hybridization/introgression may affect all taxa of series Reges.
Evolution of floral characters in section Cyathophora
The nrITS phylogeny shows that evolution of the beakless (toothed and toothless) and beaked galeas are independent within section Cyathophora, and the long corolla tube is associated with the beaked galea. The concatenated chloroplast phylogeny reveals that the long corolla tube is derived from a short corolla tube, and the beakless galea in series Reges may be inherited from that of the common ancestor or be reversed from the beaked one. Ree  suggested that pollination advantage of the beakless galea in series Reges may have facilitated its geographic expansion in the Hengduan Mountains. Field observations show that P. rex and P. thamnophila reward bumblebees and Amegilla sp. with both nectar and pollen [54,59,60]. On the contrary, bumblebees forage P. cyathophylla and P. superba for pollen alone [61; W.-B. Yu unpublished data]. Morphologically, the nectary of both P. rex and P. thamnophila is virgate at the base of the ovary, whereas that of P. cyathophylla and P. superba is absent or indistinct (H. Wang, unpublished data). Loss of nectar production may be associated with evolution of a long corolla tube and beaked galea . Pollen morphology suggests that the bisyncolpate aperture in series Superbae and Cyathophyllae may be derived from the trisyncolpate aperture in P. cyathophlloides and/or series Reges [9,62]. Both nectar production and trisyncolpate aperture may be plesiomorphic in Pedicularis, indicating that the beakless galea in series Reges may also be plesiomorphic.
Pedicularis rex includes both toothed and toothless varieties: subspecies rex and lipskyana are toothed, and variety rockii is toothless. Due to the smaller corollas of variety rockii, Li  suggested that the changed from a toothed galea to a toothless galea might have resulted from diminishing the galea size. Nevertheless, Tsoong  considered that the the toothless galea of variety rockii may be a reversal. Recently, stochastic mapping has shown that the derived toothed galea is rarely reversed to toothless galea at the genus level . In this study, the nrITS phylogeny indicates that the toothless variety rockii may be independent from other of taxa in series Reges, whereas the concatenated chloroplast phylogeny suggests that the toothless variety rockii may be derived from the toothed subspecies rex. However, phylogenetic incongruence does not allow a straightforward interpretation. Noteworthily, morphometric analyses in series Reges show that galea size of P. rex var rockii is the smallest, followed by that of P. thamnophila and the unknown taxon (excluding LIDZ1990), and that the galea of subspecies rex and lipskyana of P. rex is the largest . Morphologically, the galea teeth of P. thamnophila and the unknown taxon (excluding LIDZ1990) are smaller than those of subspecies rex and lipskyana of P. rex (see Figure 2). If the galea teeth of variety rockii have been reduced , the evidence of galea teeth can be revealed by investigating the galea development [e.g. 63] in a further study.
We are grateful to Lian-Ming Gao, Wei Jiang, Yang Luo, Hang Sun, Ding Wu, Ji-Dong Ya, and Qiu-Lin Yang for their kind help in the field and/or providing samples for this study; to Jun-Bo Yang, Jing Yang, Chun-Xia Zeng, and Yu-Xiao Zhang for their assistance in molecular experiments and/or data analysis; and to Michael Möller, Christopher Randle, Renate Wesselingh and an anonymous reviewer providing valuable comments and suggestions on the manuscript.
Conceived and designed the experiments: DZL HW. Performed the experiments: WBY PHH. Analyzed the data: WBY PHH. Contributed reagents/materials/analysis tools: WBY PHH. Wrote the manuscript: WBY DZL HW.
- 1. Li H-L (1951) Evolution in the flowers of Pedicularis. Evolution 5: 158-164. doi:https://doi.org/10.2307/2405766.
- 2. Wu Z-Y, Lu A-M, Tang Y-C, Chen Z-D, Li D-Z (2003) The families and genera of angiosperms in China: a comprehensive analysis. Beijing: Science Press.
- 3. Ree RH (2005) Phylogeny and the evolution of floral diversity in Pedicularis (Orobanchaceae). Int J Plant Sci 166: 595-613. doi:https://doi.org/10.1086/430191.
- 4. Wang H, Yu W-B, Chen J-Q, Blackmore S (2009) Pollen morphology in relation to floral types and pollination syndromes in Pedicularis (Orobanchaceae). Plant Syst Evol 277: 153-162. doi:https://doi.org/10.1007/s00606-008-0112-1.
- 5. Tsoong P-C (1955) A new system for the genus Pedicularis. Acta Phytotaxon Sin 4: 71-147.
- 6. Li H-L (1948) A revision of the genus Pedicularis in China. part I. Proc Acad Nat Sci Phila 100: 205-378.
- 7. Yang F-S, Wang X-Q (2007) Extensive length variation in the cpDNA trnT-trnF region of hemiparasitic Pedicularis and its phylogenetic implications. Plant Syst Evol 264: 251-264. doi:https://doi.org/10.1007/s00606-006-0510-1.
- 8. Yang F-S, Wang X-Q, Hong D-Y (2003) Unexpected high divergence in nrDNA ITS and extensive parallelism in floral morphology of Pedicularis (Orobanchaceae). Plant Syst Evol 240: 91-105. doi:https://doi.org/10.1007/s00606-003-0005-2.
- 9. Yu W-B, Wang H (2008) Pollen morphology of Pedicularis sect. Cyathophora, a group endemic to the eastern Himalaya-Hengduan Mountains region. J Integr Plant Biol 50: 244-252. doi:https://doi.org/10.1111/j.1744-7909.2007.00364.x. PubMed: 18713447.
- 10. Limpricht W (1924) Studien über die Gattung Pedicularis. Repert Spec Nov Regni Veg 20: 161-265.
- 11. Prain D (1890) The species of Pedicularis of the Indian Empaire and its frontiers. Ann Roy Bot Gard Clac 3: 1-196.
- 12. Bonati G (1910) Contribution à l’étude du genre Pedicularis. Bull Soc Bot France 57(18): 1-35.
- 13. Maximovicz CJ (1888) Diagnoses des plantes nouvelles asiatiques. VII 32. Bull Acad Sci St Petersbourg. pp. 477-629.
- 14. Tsoong P-C (1963) Scrophulariaceae (Pars II). In: S-S ChienW-Y Chun. Flora Reipublicae Popularis Sinacae, Vol 68. Beijing: Science Press. pp. 1-378.
- 15. Yang H-P (1990) New taxa of Pedicularis from the Hengduan Mountains. Acta Phytotaxon Sin 28: 136-144.
- 16. Yu W-B, Huang P-H, Wang H (2011) A new synonym of Pedicularis cyathophylla (Orobanchaceae). J Sys Evol 49: 507. doi:https://doi.org/10.1111/j.1759-6831.2011.00159_3.x.
- 17. Macior LW (1982) Plant community and pollinator dynamics in the evolution of pollination mechanisms in Pedicularis (Scrophulariaceae). In: JA ArmstrongJM PowellAJ Richards. Pollination and Evolution. Sydney: Royal Botanic Gardens. pp. 29–45.
- 18. Ree RH (2001) Homoplasy and phylogeny of Pedicularis. PH. D. Dissertation Cambridge, MA: Havard University.
- 19. Wang S, Xie Y (2004) China species Red List (Red List), Volume 1. Beijing: Higher Education Press. 324pp.
- 20. Doyle JJ, Doyle JL (1987) A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemistry 19: 11-15.
- 21. Yu W-B, Huang P-H, Ree RH, Liu M-L, Li D-Z et al. (2011) DNA bar coding of Pedicularis Linn. (Orobanchaceae): testing four candidate DNA bar coding loci in a large and hemiparasitic genus. J Sys Evol 49: 425-437. doi:https://doi.org/10.1111/j.1759-6831.2011.00154.x.
- 22. Katoh K, Toh H (2010) Parallelization of the MAFFT multiple sequence alignment program. Bioinformatics 26: 1899-1900. doi:https://doi.org/10.1093/bioinformatics/btq224. PubMed: 20427515.
- 23. Hall TA (1999) BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucleic Acids Symp Ser 41: 95-98.
- 24. Vaidya G, Lohman DJ, Meier R (2011) SequenceMatrix: concatenation software for the fast assembly of multi-gene datasets with character set and codon information. Cladistics 27: 171-180. doi:https://doi.org/10.1111/j.1096-0031.2010.00329.x.
- 25. Posada D (2008) jModelTest: Phylogenetic model averaging. Mol Biol Evol 25: 1253-1256. doi:https://doi.org/10.1093/molbev/msn083. PubMed: 18397919.
- 26. Guindon S, Gascuel O (2003) A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52: 696-704. doi:https://doi.org/10.1080/10635150390235520. PubMed: 14530136.
- 27. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572-1574. doi:https://doi.org/10.1093/bioinformatics/btg180. PubMed: 12912839.
- 28. Alfaro ME, Zoller S, Lutzoni F (2003) Bayes or bootstrap? A simulation study comparing the performance of Bayesian Markov chain Monte Carlo sampling and bootstrapping in assessing phylogenetic confidence. Mol Biol Evol 20: 255-266. doi:https://doi.org/10.1093/molbev/msg028. PubMed: 12598693.
- 29. Huelsenbeck JP, Rannala B (2004) Frequentist properties of Bayesian posterior probabilities of phylogenetic trees under simple and complex substitution models. Syst Biol 53: 904-913. doi:https://doi.org/10.1080/10635150490522629. PubMed: 15764559.
- 30. Kolaczkowski B, Thornton JW (2007) Effects of branch length uncertainty on Bayesian posterior probabilities for phylogenetic hypotheses. Mol Biol Evol 24: 2108-2118. doi:https://doi.org/10.1093/molbev/msm141. PubMed: 17636043.
- 31. Wilcox TP, Zwickl DJ, Heath TA, Hillis DM (2002) Phylogenetic relationships of the dwarf boas and a comparison of Bayesian and bootstrap measures of phylogenetic support. Mol Phylogenet Evol 25: 361-371. doi:https://doi.org/10.1016/S1055-7903(02)00244-0. PubMed: 12414316.
- 32. Erixon P, Svennblad B, Britton T, Oxelman B (2003) Reliability of Bayesian posterior probabilities and bootstrap frequencies in phylogenetics. Syst Biol 52: 665-673. doi:https://doi.org/10.1080/10635150390235485. PubMed: 14530133.
- 33. Stamatakis A, Hoover P, Rougemont J (2008) A rapid bootstrap algorithm for the RAxML web servers. Syst Biol 57: 758-771. doi:https://doi.org/10.1080/10635150802429642. PubMed: 18853362.
- 34. Swofford D (2003) PAUP*. Phylogenetic analysis using parsimony (* and other methods), version 4. Sunderland, MA, USA: Sinauer Associates.
- 35. Felsenstein J (1985) Confidence limits on phylogenies: an approach using the bootstrap. Evolution 39: 783-791. doi:https://doi.org/10.2307/2408678.
- 36. Huson DH, Bryant D (2006) Application of phylogenetic networks in evolutionary studies. Mol Biol Evol 23: 254-267. PubMed: 16221896.
- 37. Petit RJ, Excoffier L (2009) Gene flow and species delimitation. Trends Ecol Evol 24: 386-393. doi:https://doi.org/10.1016/j.tree.2009.02.011. PubMed: 19409650.
- 38. Maddison WP, Knowles LL (2006) Inferring phylogeny despite incomplete lineage sorting. Syst Biol 55: 21-30. doi:https://doi.org/10.1080/10635150500354928. PubMed: 16507521.
- 39. Degnan JH, Rosenberg NA (2009) Gene tree discordance, phylogenetic inference and the multispecies coalescent. Trends Ecol Evol 24: 332-340. doi:https://doi.org/10.1016/j.tree.2009.01.009. PubMed: 19307040.
- 40. Maddison WP (1997) Gene trees in species trees. Syst Biol 46: 523-536. doi:https://doi.org/10.1093/sysbio/46.3.523.
- 41. Nishimoto Y, Ohnishi O, Hasegawa M (2003) Topological incongruence between nuclear and chloroplast DNA trees suggesting hybridization in the Urophyllum group of the genus Fagopyrum (Polygonaceae). Genes Genet Syst 78: 139-153. doi:https://doi.org/10.1266/ggs.78.139. PubMed: 12773814.
- 42. Pelser PB, Kennedy AH, Tepe EJ, Shidler JB, Nordenstam B et al. (2010) Patterns and causes of incongruence between plastid and nuclear Senecioneae (Asteraceae) phylogenies. Am J Bot 97: 856-873. doi:https://doi.org/10.3732/ajb.0900287. PubMed: 21622451.
- 43. Raamsdonk LWDV, Smiech MP, Sandbrink JM (1997) Introgression explains incongruence between nuclear and chloroplast DNA-based phylogenies in Allium section Cepa. Bot J Linn Soc 123: 91-108. doi:https://doi.org/10.1006/bojl.1996.0075.
- 44. Acosta MC, Premoli AC (2010) Evidence of chloroplast capture in South American Nothofagus (subgenus Nothofagus, Nothofagaceae). Mol Phylogenet Evol 54: 235-242. doi:https://doi.org/10.1016/j.ympev.2009.08.008. PubMed: 19683588.
- 45. Okuyama Y, Fujii N, Wakabayashi M, Kawakita A, Ito M et al. (2005) Nonuniform concerted evolution and chloroplast capture: Heterogeneity of observed introgression patterns in three molecular data partition phylogenies of Asian Mitella (Saxifragaceae). Mol Biol Evol 22: 285-296. PubMed: 15483320.
- 46. Tsitrone A, Kirkpatrick M, Levin DA (2003) A model for chloroplast capture. Evolution 57: 1776-1782. doi:https://doi.org/10.1554/02-746. PubMed: 14503619.
- 47. Zhang YX, Zeng CX, Li DZ (2012) Complex evolution in Arundinarieae (Poaceae: Bambusoideae): Incongruence between plastid and nuclear GBSSI gene phylogenies. Mol Phylogenet Evol 63: 777-797. doi:https://doi.org/10.1016/j.ympev.2012.02.023. PubMed: 22415014.
- 48. Yoo K-O, Lowry PP, Wen J (2002) Discordance of chloroplast and nuclear ribosomal DNA data in Osmorhiza (Apiaceae). Am J Bot 89: 966-971. doi:https://doi.org/10.3732/ajb.89.6.966. PubMed: 21665695.
- 49. Lu L, Fritsch PW, Cruz BC, Wang H, Li DZ (2010) Reticulate evolution, cryptic species, and character convergence in the core East Asian clade of Gaultheria (Ericaceae). Mol Phylogenet Evol 57: 364-379. doi:https://doi.org/10.1016/j.ympev.2010.06.002. PubMed: 20558306.
- 50. Zou X-H, Ge S (2008) Conflicting gene trees and phylogenomics. J Syst Evol 46: 795-807.
- 51. Schmidt-Lebuhn AN, de Vos JM, Keller B, Conti E (2012) Phylogenetic analysis of Primula section Primula reveals rampant non-monophyly among morphologically distinct species. Mol Phylogenet Evol 65: 23-34. doi:https://doi.org/10.1016/j.ympev.2012.05.015. PubMed: 22643289.
- 52. McKinnon GE, Smith JJ, Potts BM (2010) Recurrent nuclear DNA introgression accompanies chloroplast DNA exchange between two eucalypt species. Mol Ecol 19: 1367-1380. doi:https://doi.org/10.1111/j.1365-294X.2010.04579.x. PubMed: 20298471.
- 53. Eaton DAR, Ree RH (2013) Inferring phylogeny and introgression using RADseq data: an example from flowering plants (Pedicularis: Orobanchaceae). Syst Biol 62: 689-706. doi:https://doi.org/10.1093/sysbio/syt032. PubMed: 23652346.
- 54. Yu W-B (2011) Species divergence and molecular identification of Pedicularis L. (Orobanchaceae) in the Himalaya-Hengduan Mountains region. Ph.D. Dissertation, Kunming Institute of Botany, Chinese Academy of Sciences.
- 55. 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. Mol Phylogenet Evol 59: 225-244. doi:https://doi.org/10.1016/j.ympev.2011.01.012. PubMed: 21292014.
- 56. Yang FS, Li YF, Ding X, Wang XQ (2008) Extensive population expansion of Pedicularis longiflora (Orobanchaceae) on the Qinghai-Tibetan Plateau and its correlation with the Quaternary climate change. Mol Ecol 17: 5135-5145. doi:https://doi.org/10.1111/j.1365-294X.2008.03976.x. PubMed: 19017264.
- 57. Yang Z-Y, Yi T-S, Pan Y-Z, Gong X (2012) Phylogeography of an alpine plant Ligularia vellerea (Asteraceae) in the Hengduan Mountains. J Sys Evol 50: 316-324. doi:https://doi.org/10.1111/j.1759-6831.2012.00199.x.
- 58. Liu J-Q, Sun Y-S, Ge X-J, Gao L-M, Qiu Y-X (2012) Phylogeographic studies of plants in China: Advances in the past and directions in the future. J Syst Evol 50: 267-275. doi:https://doi.org/10.1111/j.1759-6831.2012.00214.x.
- 59. Wang H (1998) The pollination syndrome of Pedicularis rex (Scrophulariaceae) and its biogeographic significance. Acta Bot Sin 40: 781-785.
- 60. Tang Y, Xie H-S, Sun H (2007) The pollination ecology of Pedicularis rex subsp lipkyana and P. rex subsp rex (Orobanchaceae) from Sichuan, southwestern China. Flora 202: 209-217. doi:https://doi.org/10.1016/j.flora.2006.09.001.
- 61. Tang Y, Xie H-S (2006) A pollination ecology study of Pedicularis Linnaeus (Orobanchaceae) in a subalpine to alpine area of Northwest Sichuan, China. Arct Antarct Alp Res 38: 446-453. doi:https://doi.org/10.1657/1523-0430(2006)38[446:APESOP]2.0.CO;2.
- 62. Tsoong P-C, Chang K-T (1965) Palynology study of Pedicularis and its relation with the taxonomic systems of the genus II. Acta Phytotaxon Sin 10: 357-374.
- 63. Yu W-B, Cai J, Li D-Z, Mill RR, Wang H (2013) Floral ontogeny of Pedicularis (Orobanchaceae), with an emphasis on the corolla upper lip. J Sys Evol 51: 435-450. doi:https://doi.org/10.1111/jse.12018.