Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

DNA Barcoding of Rhodiola (Crassulaceae): A Case Study on a Group of Recently Diversified Medicinal Plants from the Qinghai-Tibetan Plateau

DNA Barcoding of Rhodiola (Crassulaceae): A Case Study on a Group of Recently Diversified Medicinal Plants from the Qinghai-Tibetan Plateau

  • Jian-Qiang Zhang, 
  • Shi-Yong Meng, 
  • Jun Wen, 
  • Guang-Yuan Rao
PLOS
x

Abstract

DNA barcoding, the identification of species using one or a few short standardized DNA sequences, is an important complement to traditional taxonomy. However, there are particular challenges for barcoding plants, especially for species with complex evolutionary histories. We herein evaluated the utility of five candidate sequences — rbcL, matK, trnH-psbA, trnL-F and the internal transcribed spacer (ITS) — for barcoding Rhodiola species, a group of high-altitude plants frequently used as adaptogens, hemostatics and tonics in traditional Tibetan medicine. Rhodiola was suggested to have diversified rapidly recently. The genus is thus a good model for testing DNA barcoding strategies for recently diversified medicinal plants. This study analyzed 189 accessions, representing 47 of the 55 recognized Rhodiola species in the Flora of China treatment. Based on intraspecific and interspecific divergence and degree of monophyly statistics, ITS was the best single-locus barcode, resolving 66% of the Rhodiola species. The core combination rbcL+matK resolved only 40.4% of them. Unsurprisingly, the combined use of all five loci provided the highest discrimination power, resolving 80.9% of the species. However, this is weaker than the discrimination power generally reported in barcoding studies of other plant taxa. The observed complications may be due to the recent diversification, incomplete lineage sorting and reticulate evolution of the genus. These processes are common features of numerous plant groups in the high-altitude regions of the Qinghai-Tibetan Plateau.

Introduction

DNA barcoding refers to rapid, accurate taxon identification using one or a few short, standardized DNA region(s) [1,2]. Through large-scale standardized sequencing of the mitochondrial gene CO1, it has become an efficient tool for identifying species in many animal groups [2]. However, three obstacles still hinder its extensive application in plants, despite strenuous efforts [35]. Firstly, designing universal primers for targeted markers in all plants is problematic. Secondly, rates of successful amplification and sequencing of candidate DNA markers widely vary amongst plant groups. Thirdly, between-species differences in candidate barcodes are also highly variable and even non-existent in some taxa. Nevertheless, DNA barcoding can often be applied in plants by using two or three DNA markers [4,6,7].

An effective barcoding marker must be easy to amplify and sequence using a universal pair of primers, suitably long, usually 500–800 bp [5], and sufficiently variable between species but homogeneous within a species to distinguish closely related species robustly [3]. Several DNA regions have been identified as potentially suitable barcodes. The plastid gene rbcL and the nuclear ribosomal internal transcribed spacer (ITS) can reportedly distinguish species of Moraea Mill. and Protea L., according to BLAST tests [8]. Kress and colleagues have recommended the use of ITS and plastid trnH-psbA sequences, and subsequently trnH-psbA and rbcL genes, as two-locus universal barcodes for land plants [4]. The CBOL Plant Working Group recommended the combination rbcL + matK as a core plant barcode, and also advocated the plastid trnH-psbA and ITS [9] as complementary markers.

Rhodiola L. (Crassulaceae) consists of about 70 species mainly distributed in high altitudes and cold regions of the Northern Hemisphere [10,11]. There are about 55 Rhodiola species recognized in China (16 species endemic), especially in the western alpine regions (i.e., the Hengduan Mountains and the Qinghai-Tibetan Plateau) [10]. Species of this genus are herbaceous perennials that often grow on gravel-covered slopes or in cracks of exposed rocks at ca. 3500–5000 m elevations, thus collecting and studying them have been notoriously difficult [10]. Rhodiola species, historically used as adaptogens in Russia and northern Europe, have been widely recognized for enhancing human resistance to stress or fatigue and promoting longevity [1214]. In China, the Rhodiola species known as Hongjingtian have been frequently used as adaptogens, hemostatics, and tonics in traditional Tibetan medicines for thousands of years [14]. The type species R. rosea L. is a popular traditional medicinal plant in east Europe and Asia, with a reputation for stimulating the nervous system, decreasing depression, enhancing work performance, eliminating fatigue, and preventing high-altitude sickness [15]. The roots and rhizomes of R. crenulata (J. D. Hooker & Thomson) H. Ohba have been included in the Pharmacopoeia of China [16]. Furthermore, several other Rhodiola species, such as R. sachalinensis A. Bor., R. himalensis (D. Dons) S. H. Fu, R. serrata H. Ohba, and R. fastigiata (Hook. f. et Thomson) S. H. Fu are also used as medicines in China. Consequently, many species of this genus are severely endangered in Asia due to excessive and indiscriminate exploitation [17,18]. In spite of their wide use and medicinal importance, the identification of closely related species of Rhodiola is often difficult due to their morphological similarity.

A recent phylogenetic study of Rhodiola revealed significant convergent evolution of important morphological characters, such as dioecy and marcescent flowering stems [19]. Consequently, many of the previously defined infrageneric taxa are not monophyletic. Historical biogeographic studies have suggested that rapid radiations occurred in the evolution of this genus [20]. Rhodiola is thus an excellent model for evaluating the effectiveness and universality of rbcL + matK as a core plant barcode, and the plastid trnH-psbA and ITS as complementary markers in a group of closely related, recently diversified plant species. Furthermore, although several studies have assessed the suitability of barcoding markers for identifying important medicinal plants [21,22], they have generally focused on a small fraction of species in a genus, or single targeted species. Few studies have tested the suitability of DNA barcodes using extensive samples covering most species of a medicinal plant genus of moderate size. In the present study, we analyzed a broad set of samples of Rhodiola as a model group of recently diversified medicinal plants, to evaluate the proposed core and complementary DNA barcodes, as well as trnL-F. We also specifically assessed the power of the barcodes to discriminate six widely used medicinal species of Rhodiola.

Materials and Methods

Ethics statement

No specific permits were required for the described locations in China because all researchers collecting the samples had introduction letters from College of Life Sciences, Peking University, Beijing. The field studies did not involve protected species. The localities of all accessions sampled were shown in S1 Table.

Plant materials

In total, 189 accessions representing 47 Rhodiola species (including the widely used medicinal species) were collected from sites in Xizang, Qinghai, Gansu, Sichuan, Yunnan, and Xinjiang provinces of China, and north America from 2009 to 2012 (see S1 Table for collection information and NCBI accession numbers). Fresh leaves were dried in silica gel upon collection. Three to six accessions per taxon were sampled to cover the diversity within each taxon and most of their respective geographical ranges. Voucher specimens of the collected taxa were deposited in the Herbarium of Peking University (PEY) and the National Herbarium of the United States of America (US).

DNA extraction, amplification, and sequencing

Genomic DNA was isolated from ca. 15 mg of each leaf sample following the CTAB protocol [23]. The primers for amplification and sequencing were: ITS-1 and ITS-4 for ITS [11], psbAF and trnHR for trnH-psbA [24], c and f for trnL-F [25], rbcL-1F and rbcL-R for the rbcL gene [26], KIM 3-F and KIM 1-R for matK [27] (Table 1). The four candidate DNA barcodes were amplified by the polymerase chain reaction (PCR) in 20 μL mixtures containing 2 μL of 10 × buffer, 0.5 μM of each primer, 200 μM of each dNTP, 1 U of Taq polymerase (TianGen Biotech, Beijing, China), and 1 μL template genomic DNA. The temperature program consisted of 5 min at 95°C, 35 cycles of 1 min at 95°C, 1 min at 56°C, and 1 min at 72°C, with a final extension of 5 min at 72°C. PCR products were purified by polyethylene glycol (PEG) precipitation [28], then sequenced using BigDye 3.1 reagents with an ABI 3730 automated sequencer (Applied Biosystems, Foster City, California) at the Biomed Corporation (Beijing, China).

Data analyses

Contigs were assembled and edited using the ContigExpress module of Vector NTI Suite 6.0 (InforMax). Sequences were aligned using MUSCLE 3.8.31 [29], followed by manual adjustments in Geneious 7.1.7 [30]. We calculated the Kimura 2-parameter (K2P) distance for all five DNA regions in MEGA v. 6.0 [31] to estimate intra- and inter-specific divergence, and assessed its significance using the median and Wilcox two-sample tests (http://www.fon.hum.uva.nl/Service/Statistics.html). We also graphed the distribution of intra- and inter-specific divergence (i.e., K2P distances) of each candidate barcode. Here intra-specific distances include all possible intra-specific comparisons and inter-specific distances represent all possible inter-specific comparisons. We also used the BLAST procedure to evaluate the generic-level identification power of the five tested markers, using every sequence generated for the five candidate barcodes as a seed sequence to check whether best matches in the National Centre for Biotechnology Information nucleotide database (http://www.ncbi.nlm.nih.gov/BLAST/) were from the same genus.

To evaluate the monophyly of the individuals representing the same species based on morphological assessment, tree-based methods were used to display the molecular identification results. The species identification rates of the barcodes (either singly or in all combinations) were determined by evaluating the percentages of their assignments for each species that were monophyletic according to the UPGMA, NJ, MP, and ML analyses. The neighbor-joining (NJ) and the unweighted pair group method with arithmetic mean (UPGMA) trees were reconstructed using MEGA v. 6.0 with the K2P model [31]. The program PAUP v. 4.0b10 [32] was used to generate the maximum parsimony (MP) tree with a heuristic search strategy followed by random addition starting trees with tree-bisection-reconnection (TBR) branch swapping and MulTrees selected. Indels were treated as missing data and all characters were weighted equally. Support for individual nodes was assessed by calculating bootstrap values [33]. Parsimony bootstrap (PB) values were obtained from 1,000 replicates of heuristic searches as described above (TBR branch swapping and MulTrees selected), but with branch swapping limited to 10 million rearrangements per replicate due to memory constraints. Nucleotide substitution model parameters were determined using the Akaike Information Criterion (AIC) in Modeltest version 3.7 [34,35]. Maximum likelihood (ML) analyses were performed using RAxML 8.0.0 with 1000 bootstraps under the GTRIG model [36]. We used the following code to set the parameters: –b 1000 –m GTR –v e –f e –t e –a e –o tlr (see the RAxML manual for details).

We also used the genetic distance-based program TaxonDNA to analyze the species identification rates of the DNA barcodes using three criteria: “Best Match” (which assigns queries to species with the best-matching sequences, regardless of their similarity); “Best Close Match” (which assigns queries to species if a threshold similarity is met); and “All Species Barcodes” (which assigns queries to species if they match all known barcodes for the species and there are at least two conspecific matches) [37].

Results

Amplification and sequencing

The primers for all five selected DNA regions were found to be applicable for all the 47 species of Rhodiola, and 933 barcode sequences were generated from the study. Newly generated sequences were deposited in the GenBank database with accession numbers shown in S1 Table. The amplification and bidirectional sequencing success rates consistently exceeded 95% for all the markers except matK (Table 2). We tested the amplification efficiency of several other published primers for amplifying matK, including a pair designed for amplifying Saxifragaceae sequences [38], and a recently designed pair of “universal” primers for an angiosperm barcode [39], but the success rate was low. The primer pair used in the present study (KIM 3-F and KIM 1-R [27]) worked the best, but the success rate for matK was still lower than that of the other markers (Table 2).

thumbnail
Table 2. Information used to evaluate the utility of the five DNA barcoding loci.

https://doi.org/10.1371/journal.pone.0119921.t002

Alignment, variability, and BLAST procedure

The aligned lengths of the rbcL, matK, trnH-psbA, trnL-F and ITS barcode data set were 1415, 847, 423, 927, and 668 bp, respectively. Length variation exists in each region, which is 1259–1272 bp for rbcL, 762–786 bp for matK, 277–347 bp for trnH-psbA, and 600–642 bp for ITS. The numbers of informative sites and variable sites (Table 2) were the highest for ITS (204 and 254, respectively), and the lowest for matK (62 and 101, respectively). The best matches to sequences of the five candidate barcodes from the 47 investigated species, identified by BLAST searches of the NCBI database, were all from species of Rhodiola.

Monophyly tests of species based on phylogenetic trees

In the monophyly tests based on phylogenetic trees, UPGMA analyses provided the highest indications of discriminatory power, followed by NJ, ML and MP analyses. As single barcodes, ITS provided the highest species identification rate (66.0%), followed by matK, trnH-psbA, trnL-F and rbcL (36.2, 36.2, 29.8 and 19.1%, respectively). Furthermore, even all four plastid-derived barcodes provided a lower identification rate (61.7%) than ITS, demonstrating the discriminatory power of ITS as a DNA barcode. However, the highest percentage (80.9%) was achieved using all five candidate barcodes, followed by combinations including ITS (e.g., 76.6% using both rbcL + matK + trnL-F + ITS and rbcL + matK + trnH-psbA + ITS). The “core” barcode combination rbcL and matK yielded a species identification rate of just 40.4% (Table 3).

thumbnail
Table 3. Percentage of Rhodiola species recovered as monophyletic based on phylogenetic trees for each barcode.

https://doi.org/10.1371/journal.pone.0119921.t003

Barcoding gap test

The barcoding gaps between intra- and inter-specific distances assessed by graphing the distribution of variation in K2P genetic distance for rbcL, matK, trnH-psbA, trnL-F and ITS were shown in S1 Fig. For each barcode candidate, both the median and Wilcoxon two-sample tests showed that the intra-specific distance was always significantly lower than the inter-specific distance (Table 4).

thumbnail
Table 4. Results of the median and the Wilcoxon two-sample tests based on interspecific versus intraspecific Kimura 2-parameter distances for each barcode.

https://doi.org/10.1371/journal.pone.0119921.t004

TaxonDNA analysis

The results of TaxonDNA analysis using three criteria was shown in Table 5. According to the “Best Match” and “Best Close Match” criteria, ITS, trnH-psbA, trnL-F, matK and rbcL provided species identifications for 65.4, 38.9, 34.9, 34.5 and 22.3% of the samples, respectively. Slightly lower percentages, but similar patterns, were obtained using the “All species Barcodes” strategy. ITS alone and combinations including ITS provided higher success rates than other markers (and combinations of markers) and generally combinations provided higher success rates than single markers (Table 5) based on all three criteria.

thumbnail
Table 5. Success rates of species identification based on TaxonDNA analysis.

https://doi.org/10.1371/journal.pone.0119921.t005

Discussion

New techniques are needed to improve descriptive taxonomy and to ensure that organisms used for various scientific and applied purposes are correctly identified [37]. DNA barcoding is still a relatively new technique, and it has been extensively applied for rapidly identifying diverse taxa [21,27]. However, substantial advances are still needed to reliably apply DNA barcoding in plants. There is no consensus on standard markers, procedures and strategies for barcode development, although a “core barcode” and several other combinations of candidate barcode regions have been proposed [5,27]. A standard DNA barcode for plants must be able to differentiate challenging plant species, such as recently diversified taxa with complex evolutionary histories.

Identification power of barcode loci in Rhodiola

As reported in other plant barcoding studies [22,4042], we also found the ITS region to be the most powerful and the most useful one of the five tested barcodes in Rhodiola. ITS, alone or in combination with plastid markers, can discriminate more than 70% of the species in this genus, confirming its utility as a core barcode marker [9]. One of the main constraints of using ITS as a standard barcode is that it is difficult to amplify and sequence in some taxa [40,42], due to incomplete concerted evolution [9]. Although the success rates for ITS amplification and sequencing were high (Table 2), attempts to acquire ITS sequences for some individuals of R. discolor (Franch.) S. H. Fu and R. bupleuroides (Wall. ex Hook. f. & Thomson) S. H. Fu failed or were difficult, possibly due to multiple polyploidy events, at least for the R. bupleuroides samples, with the reported chromosome numbers of this species ranging from 20 to 110 [43].

The plastid markers showed significantly lower discriminatory power than the ITS region, although the CBOL Plant Working Group has recommended rbcL and matK genes as core plant barcodes [27]. Among the four plastid markers, rbcL, which is a coding region, showed the least variability in Rhodiola (Tables 2 & 3). Low variability of rbcL has also been found in other barcoding studies [4042]. Thus, this conservative marker is often used to determine the taxonomic affinity of unknown samples. Our BLAST analyses of the rbcL sequences all provided best matches with a congeneric species, confirming its utility for placing the plants into the correct genus. Another coding region recommended by the CBOL Plant Working Group is matK, which showed lower variability and lower discrimination power than the ITS region, but greater utility than rbcL in Rhodiola (Tables 2 & 3). However, developing universal primers for matK has been reported to be problematic [4,40,44], and we encountered the same obstacle in the present study. Several previously designed primers, including a pair of recently designed “universal primers” for angiosperm barcoding for amplifying partial matK regions [39] also failed. The primer pair used in the present study (Table 1) worked relatively well, but the success rate was still lower than that of other barcodes (Table 2).

Compared to the two coding regions (rbcL, matK), which tend to be more conserved, the two non-coding regions, trnH-psbA and trnL-F, were more useful for distinguishing similar species. The trnH-psbA spacer has been suggested as a robust DNA barcode for various plants, including Ligustrum L. [41], orchids [45], and Tetrastigma (Miq.) Planch. [40]. Our results show that it has more variable sites, and provided higher species identification rates, than both rbcL and matK (Tables 2 & 3), even though it is much shorter (423 vs 1415 and 847 bp, Table 2). However, its length may be highly variable, ranging from > 300 bp in some groups [5], such as Solidago L., to > 1000 bp in ferns [45], some monocots [46] and conifers [27]. Such length variation may make it difficult for bidirectional sequencing using universal primers as well as for accurate alignment. Our data indicate that the average length of trnH-psbA in Rhodiola is appropriate for a barcode. We detected five large indels (≥10 bp) in the alignment for this marker, but their occurrence does not seem to correlate with species; for example, the same 14 bp insertion was detected in one R. heterodonta (Hook. f. & Thomson) Boriss. individual and one R. macrocarpa (Praeger) S. H. Fu individual. Thus, as pointed out by Kress et al. [5] and Fu et al. [40], indel variations in trnH-psbA may not be suitable for distinguishing species.

Few studies have discussed the suitability of trnL-F as a potential DNA barcode in angiosperms. It was suggested to be a powerful barcode marker in ferns, because the region shows sufficient variations among species, and the universal primers show high success rates for amplification and sequencing [47,48]. We confirmed the amenability of the trnL-F region for amplification and sequencing. Furthermore, although it did not have higher variation (i.e., discrimination power) than matK or trnH-psbA, it was slightly more variable than rbcL (Tables 2 & 3). Thus, given the difficulties of amplifying and sequencing matK and the potential “indel” problem of the trnH-psbA spacer, the trnL-F regions seems a promising alternative marker for barcoding Rhodiola and other plants.

Multi-locus barcodes have consistently provided stronger identification power than single regions [27]. Accordingly, the combinations of markers generally showed higher discriminatory power than single markers in the present study (Tables 3 & 5). For example, the two core barcodes (rbcL and matK) in combination distinguished 40.4% of the included Rhodiola species, while separately they distinguished 19.1 and 36.2%, respectively. The TaxonDNA analysis based on different criteria showed the same pattern (Table 5). We also tested the robustness of rbcL + matK + X (ITS, trnH-psbA, trnL-F, or combinations thereof), as advocated by representatives of the Chinese Plants Barcoding program. The results showed that using more than two cpDNA barcodes increased identification rates, although the ITS region alone seems to discriminate species equally well.

DNA barcoding evaluation in a well sampled data set

Extensive sampling and analysis of taxonomically well understood groups are needed to thoroughly validate and standardize markers and procedures for DNA barcoding [49]. A number of studies have tested the application of DNA barcoding in various plant groups from familial to generic levels [4042,5056]. Familial-level studies often included representatives of each genus [53,54,57]. In contrast, most generic level studies only examined a small fraction of species [4042,51], or focused on a single species [55], with some exceptions [50]. Thus, a major objective of the present study was to test the candidate DNA barcodes at the generic level using an extensive taxon sampling scheme [5862]. Both morphologically divergent (Fig. 1, A-F) and highly similar species (Fig. 1, I, K) were included in the data set. Not surprisingly, morphologically divergent species were easily identified by DNA barcodes, whereas species with highly similar morphological characters remained largely unresolved (Fig. 1). For example, R. smithii (Raym.-Hamet) S. H. Fu (Fig. 1A) is easily distinguished from other species of Rhodiola by its radical leaves with appendages on them, and accessions of the species were found to be monophyletic with high bootstrap support (Fig. 2). This is also the case for three other morphologically distinct species, R. humilis (Hook. f. & Thomson) S. H. Fu, R. stapfii (Raym.-Hamet) S. H. Fu and R. prainii (Raym.-Hamet) H. Ohba (Fig. 1, D). On the other hand, accessions of two morphologically similar species, R. fastigiata (Fig. 1, I) and R. tibetica (Hook. f. & Thomson) S. H. Fu were scattered in one of the Rhodiola clades (Fig. 2, Clade A). Furthermore, accessions of the species of R. sect. Trifida all mingled together on clade B (Fig. 2). Species of this section are highly similar morphologically and only show slight differences in leaf morphology [19].

thumbnail
Fig 1. Representatives of species illustrating the morphological variation in Rhodiola.

(A. Rhodiola smithii; B. R. yunnanensis; C. R. chrysanthemifolia; D. R. prainii; E. R. dumulosa; F. R. hobsonii; G. R. rosea; H. R. kirilowii; I. R. fastigiata; J. R. crenulata; K. R. alsia; L. R. bupleuroides).

https://doi.org/10.1371/journal.pone.0119921.g001

thumbnail
Fig 2. Neighbor joining tree using the Kimula 2-parameter distances based on all five barcoding markers for Rhodiola species.

Numbers at nodes represent bootstrap values with 1000 replicates (only values > 50 were shown).

https://doi.org/10.1371/journal.pone.0119921.g002

DNA barcoding in a recently diversified plant group

With the potential of DNA barcoding to facilitate species identification and discovery [49,63], its applicability in various plant groups has been evaluated recently [22,40,51]. Problems have been encountered in some cases due to complexities arising from the reproductive behavior and evolutionary history, such as interspecific hybridization, introgression, allopolyploidy, mixtures of sexual and asexual reproduction, and recent divergences [22,64]. Such biological factors may have blurred species boundaries in Rhodiola, which was dated to have diversified recently, with its crown group diverged ca. 6.32 Ma [20]. Furthermore, hybridization and introgression may have played important roles in its evolutionary history [20].

The complexities seen in Rhodiola are common in many plant groups, but their effects on DNA barcoding have not been rigorously assessed. The rate of successful species identification was lower than in other plant barcoding studies [4042,5056]. As shown in Table 3, at most 80.9% of the species were successfully recovered as monophyletic groups, even when using all of the markers (rbcL + matK + trnH-psbA + trnL-F + ITS), and at most 72% of the species were successfully identified using the TaxonDNA test (Table 5). We attribute the low discrimination power of the tested barcodes to two main factors. Firstly, the rapid recent species radiations of Rhodiola may have resulted in polytomies of the gene trees, preventing most markers from accumulating sufficient variation to distinguish different species reliably, even if they can be distinguished morphologically (Fig. 1). Secondly, incomplete lineage sorting (ILS) and reticulate evolution, which may occur alone or together, may have blurred species boundaries, impeding clear barcoding. ILS, caused by the retention of ancestral polymorphisms [6567], is likely to lead to discordant and unpredictable associations between accessions of different species due to its stochastic nature. In contrast to ILS, reticulate evolution resulting from post-speciation hybridization and organelle capture among pairs of taxa may show systematic associations between species. Both stochastic and systematic associations between different accessions of species were observed (Fig. 2), indicating that both processes may have played a role in the evolutionary history of Rhodiola, leading to the complications to barcode species of this genus.

In contrast to animals, many plant species are likely to have paraphyletic or polyphyletic origins due to the higher frequency of reticulate evolution in plants, as facilitated by hybridization and polyploidization [68]. Under these circumstances, barcoding based solely on plastid markers may not reliably distinguish species [69]. However, nuclear DNA sequences, e.g., the internal transcribed spacer region (ITS), may improve the resolution among plant species due to its generally higher synonymous substitution rates [70] and less sensitivity to problems caused by hybridization [8]. Our results show that ITS can distinguish more species than the combination of four plastid markers (66.0% vs 61.7%, Table 3). Thus, our DNA barcoding analysis confirms that ITS, and probably other nuclear genes, are powerful tools for identifying plant species with complex evolutionary histories.

Identification of medicinal plants using DNA barcodes

The present study included six species of Rhodiola reported with medicinal properties (R. rosea, R. crenulata, R. sachalinensis, R. himalensis, R. serreta, and R. fastigiata). As one of the most important traditional herbal remedies, R. crenulata has been used in treating long-term illnesses and weaknesses caused by infection in Tibet and other regions for more than 1000 years [13,17]. Four individuals of the species from four different geographic areas form a monophyletic group based on the five-marker barcodes (Fig. 2), showing the feasibility of using DNA barcoding to distinguish this species from other species of the genus or other adulterants. Rhodiola rosea has also been widely used in East Europe and Asia for stimulating the nervous system and decreasing depression [15]. In our analysis, five accessions of R. rosea from different localities formed a clade with the only accession of R. sachalinensis, and one accession of R. tangutica. Rhodiola rosea and R. sachalinensis have strong morphological similarity [10]. Two other species (R. himalensis and R. serreta) could also be correctly identified (Fig. 2). However, R. fastigiata can not be identified successfully using the five-marker barcode because of its high intraspecific variation, probably due to incomplete lineage sorting. In summary, five of the six plant species with medicinal properties are each identifiable using barcodes tested in the present study. Thus, the results indicate not only the potential utility of barcoding for these plants, but also the need to validate the barcodes and interpret the results carefully.

Supporting Information

S1 Fig. Distributions of intra- and inter-specific Kimura 2-parameter (K2P) distances for five candidate barcodes in Rhodiola.

https://doi.org/10.1371/journal.pone.0119921.s001

(PPTX)

S1 Table. Localities, voucher information and GenBank accessions numbers for sequenced taxa.

https://doi.org/10.1371/journal.pone.0119921.s002

(DOC)

Acknowledgments

The authors thank two anonymous reviewers and the academic editor whose suggestions greatly improved the original manuscript. We deliver our gratitude to Yiming An and Dongqing Zhang for field assistance. The study represents part of Jianqiang Zhang’s dissertation research. JQZ also would like to dedicate this paper to Miss Dong-Qing Zhang for their forthcoming wedding ceremony.

Author Contributions

Conceived and designed the experiments: JQZ GYR JW. Performed the experiments: JQZ. Analyzed the data: JQZ. Contributed reagents/materials/analysis tools: JQZ SYM. Wrote the paper: JQZ GYR JW.

References

  1. 1. Floyd R, Abebe E, Papert A, Blaxter M. Molecular barcodes for soil nematode identification. Mol Ecol. 2002; 11: 839–850. pmid:11972769
  2. 2. Hebert PDN, Cywinska A, Ball SL, DeWaard JR. Biological identifications through DNA barcodes. Proc R Soc B. 2003; 270: 313–321. pmid:12614582
  3. 3. Cowan RS, Chase MW, Kress WJ, Savolainen V. 300,000 species to identify: problems, progress, and prospects in DNA barcoding of land plants. Taxon. 2006; 55: 611–616.
  4. 4. Kress WJ, Erickson DL. A two-locus global DNA barcode for land plants: the coding rbcL gene complements the non-coding trnH-psbA spacer region. PLoS One. 2007; 2.
  5. 5. Kress WJ, Wurdack KJ, Zimmer EA, Weigt LA, Janzen DH. Use of DNA barcodes to identify flowering plants. Proc Natl Acad Sci USA. 2005; 102: 8369–8374. pmid:15928076
  6. 6. Fazekas AJ, Burgess KS, Kesanakurti PR, Graham SW, Newmaster SG, Husband BC, et al. Multiple multilocus DNA barcodes from the plastid genome discriminate plant species equally well. PLoS One. 2008; 3.
  7. 7. Hollingsworth ML, Clark AA, Forrest LL, Richardson J, Pennington RT, Long DG, et al. Selecting barcoding loci for plants: evaluation of seven candidate loci with species-level sampling in three divergent groups of land plants. Mol Ecol Resour. 2009; 9: 439–457. pmid:21564673
  8. 8. Chase MW, Salamin N, Wilkinson M, Dunwell JM, Kesanakurthi RP, Haidar N, et al. Land plants and DNA barcodes: short-term and long-term goals. Phil Trans R Soc B. 2005; 360: 1889–1895. pmid:16214746
  9. 9. Li DZ, Gao LM, Li HT, Wang H, Ge XJ, Liu JQ, et al. Comparative analysis of a large dataset indicates that internal transcribed spacer (ITS) should be incorporated into the core barcode for seed plants. Proc Natl Acad Sci USA. 2011; 108: 19641–19646. pmid:22100737
  10. 10. Fu KT, Ohba H. Crassulaceae. In: Wu ZY, Raven PH, editors. Flora of China, vol. 8. Beijing: Science Press; 2001. pp. 202–268.
  11. 11. Mayuzumi S, Ohba H. The phylogenetic position of eastern Asian Sedoideae (Crassulaceae) inferred from chloroplast and nuclear DNA sequences. Syst Bot. 2004; 29: 587–598.
  12. 12. Mattioli L, Titomanlio F, Perfumi M. Effects of a Rhodiola rosea L. extract on the acquisition, expression, extinction, and reinstatement of morphine-induced conditioned place preference in mice. Psychopharmacology. 2012; 221: 183–193. pmid:22421739
  13. 13. Rohloff J. Volatiles from rhizomes of Rhodiola rosea L. Phytochemistry. 2002; 59: 655–661. pmid:11867098
  14. 14. Yang YC, He TN, Lu SL, Hung RF, Wang ZX. Tibet Medicine. Xining: Qinghai People’s Publishing House; 1991.
  15. 15. Gregory S, Kelly ND. Rhodiola rosea: a possible plant adaptogen. Altern Med Rev 2001; 6: 293–302 pmid:11410073
  16. 16. Pharmacopoeia. Pharmacopoeia of the People’s Republic of China. Beijing: Chemical Industry Press; 2010.
  17. 17. Lei YD, Gao H, Tsering T, Shi SH, Zhong Y. Determination of genetic variation in Rhodiola crenulata from the Hengduan Mountains region, China using inter-simple sequence repeats. Genet Mol Biol. 2006; 29: 339–344.
  18. 18. Yan TF, Zu YG, Yan XF, Zhou FJ. Genetic structure of endangered Rhodiola sachalinensis. Conserv Genet. 2003; 4: 213–218.
  19. 19. Zhang JQ, Meng SY, Wen J, Rao GY. Phylogenetic relationships and character evolution of Rhodiola (Crassulaceae) based on nuclear ribosomal ITS and plastid trnL-F and psbA-trnH sequences. Syst Bot. 2014a; 39: 441–451.
  20. 20. Zhang JQ, Meng SY, Allen GA, Wen J, Rao GY. Rapid radiation and dispersal out of the Qinghai-Tibetan Plateau of an alpine plant lineage Rhodiola (Crassulaceae). Mol Phylogenet Evol. 2014b; 77: 147–158. pmid:24780751
  21. 21. Li DZ, Liu JQ, Chen ZD, Wang H, Ge XJ, Zhou SL, et al. Plant DNA barcoding in China. J Syst Evol. 2011; 49: 165–168.
  22. 22. Zuo YJ, Chen ZJ, Kondo K, Funamoto T, Wen J, Zhou SL, et al. DNA barcoding of Panax species. Planta Med. 2011; 77: 182–187. pmid:20803416
  23. 23. Doyle J. DNA protocols for plants-CTAB total DNA isolation. In: Hewitt GM, Johnston A, editors. Molecular Techniques in Taxonomy. Berlin: Springer; 1991. pp. 283–293;
  24. 24. Sang T, Crawford DJ, Stuessy TF. Chloroplast DNA phylogeny, reticulate evolution, and biogeography of Paeonia (Paeoniaceae). Am J Bot. 1997; 84: 1120–1136. pmid:21708667
  25. 25. Taberlet P, Gielly L, Pautou G, Bouvet J. Universal primers for amplification of 3 noncoding regions of chloroplast DNA. Plant Mol Biol. 1991; 17: 1105–1109. pmid:1932684
  26. 26. Asmussen CB, Chase MW. Coding and noncoding plastid DNA in palm systematics. Am J Bot. 2001; 88: 1103–1117. pmid:11410476
  27. 27. Hollingsworth PM, Forrest LL, Spouge JL, Hajibabaei M, Ratnasingham S, et al. A DNA barcode for land plants. Proc Natl Acad Sci USA. 2009; 106: 12794–12797. pmid:19666622
  28. 28. Wen J, Nie ZL, Soejima A, Meng Y. Phylogeny of Vitaceae based on the nuclear GAI1 gene sequences. Botany. 2007; 85: 731–745.
  29. 29. Edgar RC. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 2004; 32: 1792–1797. pmid:15034147
  30. 30. Drummond A, Ashton B, Buxton S, Cheung M, Cooper A, Duran C, et al. Geneious version 7.1.7 created by Biomatters. 2011. Available from http://www.geneious.com/.
  31. 31. Tamura K, Stecher G, Peterson D, Filipski A, Kumar S. MEGA6: Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol Evol. 2013; 30: 2725–2729. pmid:24132122
  32. 32. Swofford DL. PAUP*. Phylogenetic Analysis Using Parsimony (* and Other Methods). Version 4. 2003.
  33. 33. Felsenstein J. Confidence-limits on phylogenies—an approach using the bootstrap. Evolution. 1985; 39: 783–791.
  34. 34. Posada D. ModelTest Server: a web-based tool for the statistical selection of models of nucleotide substitution online. Nucleic Acids Res. 2006; 34: W700–W703. pmid:16845102
  35. 35. Posada D, Buckley TR. Model selection and model averaging in phylogenetics: advantages of akaike information criterion and Bayesian approaches over likelihood ratio tests. Syst Biol. 2004; 53: 793–808. pmid:15545256
  36. 36. Guindon S, Gascuel O. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol. 2003; 52: 696–704. pmid:14530136
  37. 37. Meier R, Shiyang K, Vaidya G, Ng PK. DNA barcoding and taxonomy in Diptera: a tale of high intraspecific variability and low identification success. Syst Biol. 2006; 55: 715–728. pmid:17060194
  38. 38. Johnson LA, Soltis DE. matK DNA sequences and phylogenetic reconstruction in Saxifragaceae s. str. Syst Bot. 1994; 19: 143–156.
  39. 39. Yu J, Xue JH, Zhou SL. New universal matK primers for DNA barcoding angiosperms. J Syst Evol. 2011; 49: 176–181.
  40. 40. Fu YM, Jiang WM, Fu CX. Identification of species within Tetrastigma (Miq.) Planch.(Vitaceae) based on DNA barcoding techniques. J Syst Evol. 2011; 49: 237–245.
  41. 41. Gu J, Su JX, Lin RZ, Li RQ, Xiao PG. Testing four proposed barcoding markers for the identification of species within Ligustrum L. (Oleaceae). J Syst Evol. 2001; 49: 213–224.
  42. 42. Zhang D, Duan L, Zhou N. Application of DNA barcoding in Roscoea (Zingiberaceae) and a primary discussion on taxonomic status of Roscoea cautleoides var. pubescens. Biochem Syst Ecol. 2014; 52: 14–19.
  43. 43. Ohba H, Wakabayashi M. Cytotaxonomic study of Rhodiola bupleuroides in the Himalaya and Yunnan, China (Crassulaceae). XV International Botanical Congress; 1993.
  44. 44. Kress WJ, Erickson DL. DNA barcodes: genes, genomics, and bioinformatics. Proc Natl Acad Sci USA. 2008; 105: 2761–2762. pmid:18287050
  45. 45. Nitta JH. Exploring the utility of three plastid loci for biocoding the filmy ferns (Hymenophyllaceae) of Moorea. Taxon. 2008; 57: 725.
  46. 46. Chase MW, Cowan RS, Hollingsworth PM, Van Den Berg C, Madriñán S, Petersen G, et al. A proposal for a standardised protocol to barcode all land plants. Taxon. 2007; 56: 295–299. pmid:17464884
  47. 47. Chen CW, Huang YM, Kuo LY, Nguyen QD, Luu HT, Callado JR, et al. (2013) trnL-F is a powerful marker for DNA identification of field vittarioid gametophytes (Pteridaceae). Ann Bot. 2013; 111: 663–673. pmid:23380240
  48. 48. de Groot GA, During HJ, Maas JW, Schneider H, Vogel JC, Erkens RH. Use of rbcL and trnL-F as a two-locus DNA barcode for identification of NW-European ferns: an ecological perspective. PLoS One. 2011; 6: e16371. pmid:21298108
  49. 49. Meyer CP, Paulay G. DNA barcoding: error rates based on comprehensive sampling. PLoS Biol. 2005; 3: e422. pmid:16336051
  50. 50. Dong LN, Wortley AH, Wang H, Li DZ, Lu L. Efficiency of DNA barcodes for species delimitation: A case in Pterygiella Oliv.(Orobanchaceae). J Syst Evol. 2011; 49: 189–202.
  51. 51. Guo X, Simmons MP, But PPH, Shwa PC, Wang RJ. Application of DNA barcodes in Hedyotis L. (Spermacoceae, Rubiaceae). J Syst Evol. 2011; 49: 203–212.
  52. 52. İpek M, İpek A, Simon PW. Testing the utility of matK and ITS DNA regions for discrimination of Allium species. Turk J Bot. 2014; 38: 203–212.
  53. 53. Shi LC, Zhang J, Han JP, Song JY, Yao H, Zhu YJ, et al. Testing the potential of proposed DNA barcodes for species identification of Zingiberaceae. J Syst Evol. 2011; 49: 261–266.
  54. 54. Xiang XG, Zhang JB, Lu AM, Li RQ. Molecular identification of species in Juglandaceae: A tiered method. J Syst Evol. 2011; 49: 252–260.
  55. 55. Xue CY, Li DZ. Use of DNA barcode sensu lato to identify traditional Tibetan medicinal plant Gentianopsis paludosa (Gentianaceae). J Syst Evol. 2011; 49: 267–270.
  56. 56. Yan HF, Hao G, Hu CM, Ge XJ. DNA barcoding in closely related species: A case study of Primula L. sect. Proliferae Pax (Primulaceae) in China. J Syst Evol. 2011; 49: 225–236.
  57. 57. Du ZY, Qimike A, Yang CF, Chen JM, Wang QF. Testing four barcoding markers for species identification of Potamogetonaceae. J Syst Evol. 2011; 49: 246–251.
  58. 58. Fu SH, Fu KT. Crassulaceae. In: Chen WQ, Ruan YZ, editors. Flora Reipublicae Popularis Sinicae, vol 34. Beijing: Science Press; 1984. pp. 33–216.
  59. 59. Ohba H. Generic and infrageneric classfication of the Old World Sedoideae (Crassulaceae). J Fac Sci Univ Tokyo. 1978; 3: 139–138.
  60. 60. Ohba H. A revision of the Asiatic species of Sedoideae (Crassulaceae). Part 1. Rosularia and Rhodiola (subgen. Primuloides and Crassipedes). J Fac Sci Univ Tokyo. 1981a; 3: 337–405.
  61. 61. Ohba H. A revision of the Asiatic species of Sedoideae (Crassulaceae). Part 2. Rhodiola (subgen. Rhodiola sect. Rhodiola). J Fac Sci Univ Tokyo. 1981b; 3: 65–119.
  62. 62. Ohba H. A revision of the Asiatic species of Sedoideae (Crassulaceae). Part 3. Rhodiola (subgen. Rhodiola sect. Pseudorhodiola, Prainia & Chamaerhodiola). J Fac Sci Univ Tokyo. 1982; 3: 121–174.
  63. 63. Miller SE. DNA barcoding and the renaissance of taxonomy. Proc Natl Acad Sci USA. 2007; 104: 4775–4776. pmid:17363473
  64. 64. Spooner DM. DNA barcoding will frequently fail in complicated groups: an example in wild potatoes. Am J Bot. 2009; 96: 1177–1189. pmid:21628268
  65. 65. Bock DG, Kane NC, Ebert DP, Rieseberg LH. Genome skimming reveals the origin of the Jerusalem Artichoke tuber crop species: neither from Jerusalem nor an Artichoke. New Phytol. 2013; 201: 1021–1030. pmid:24245977
  66. 66. Joly S, McLenachan PA, Lockhart PJ. A statistical approach for distinguishing hybridization and incomplete lineage sorting. Am Nat. 2009; 174: E54–E70. pmid:19519219
  67. 67. Maddison WP, Knowles LL. Inferring phylogeny despite incomplete lineage sorting. Syst Biol. 2006; 55: 21–30. pmid:16507521
  68. 68. Rieseberg LH, Brouillet L. Are many plant species paraphyletic? Taxon. 1994; 43: 21–32.
  69. 69. Fazekas AJ, Kesanakurti PR, Burgess KS, Percy DM, Graham SW, Barrett SCH, et al. Are plant species inherently harder to discriminate than animal species using DNA barcoding markers? Mol Ecol Resour. 2009; 9: 130–139. pmid:21564972
  70. 70. Hajibabaei M, Janzen DH, Burns JM, Hallwachs W, Hebert PD. DNA barcodes distinguish species of tropical Lepidoptera. Proc Natl Acad Sci USA. 2006; 103: 968–971. pmid:16418261