Identification of Ephedra species by phylogenetic analyses using matK and ITS 1 sequences

In this study, the species identifications of seven Ephedra plants, including three medicinal plants from the Pharmacopoeia of the People’s Republic of China, were conducted using phylogenetic analyses, and the method’s validity was verified. The phylogenetic trees constructed from the maturase-coding gene (matK) and internal transcribed spacer 1 (ITS1) sequences showed that the former could be used for identifying five Ephedra plants, Ephedra intermedia, E. equisetina, E. antisyphilitica, E. major, and E. aphylla, but it had less power to discriminate E. sinica and E. przewalskii, while the latter could distinguish five Ephedra plants, E. przewalskii, E. equisetina, E. antisyphilitica, E. major, and E. aphylla, but it had less power to discriminate E. sinica and E. intermedia. However, when the two genes were combined, the seven species could be completely distinguished from each other, especially the medicinal plants from the others, which is significant in developing their pharmaceutical uses and in performing quality control assessments of herbal medicines. The method presented here could be applied to the analysis of processed Ephedra plants and to the identification of the botanical origins of crude drugs. Additionally, we discovered that E. equisetina and E. major were probably closely related to each other, and that E. sinica, E. intermedia, and E. przewalskii also had a close genetic relationship.


Introduction
The genus Ephedra, which is distributed in the arid and semiarid regions of Asia, Europe, northern Africa, southwestern North America and South America, belongs to the family Ephedraceae and includes ~50 species [1].In China, there are 13 species [2], and three species, Ephedra sinica, E. intermedia, and E. equisetina have long been used in traditional medicines according to the Pharmacopoeia of the People's Republic of China [3].The medicinal Ephedra plants have been used primarily to treat asthma or bronchitis, but they are also prescribed for cold and flu symptoms, including nasal congestion, cough, fever, and chills [4].Now, processed Ephedra herbs with vernacular names are distributed in the markets, making the identification of their species of origin more difficult.Moreover, some adulterants of Ephedra species are confused with the medicinal plants.To maintain quality control, it is essential that the medicinal Ephedra plants are identifiable.Therefore, the identification of the plant sources is critical for their use as herbal medicines.
Many Ephedra plants are morphologically similar, making their identification based on morphology very difficult.For example, E. intermedia and E. przewalskii not only have three-lobed leaves, but also have two-lobed leaves [2], which increases the difficulty to distinguish between them.Moreover, identifying the botanical origin of the processed Ephedra herbs is more difficult because during processing, the natural resource is cut into sections and dried, or broiled with honey.
Recently, molecular systematics in plants, as well as other organisms, has been widely used for species identification and the determination of phylogenetic relationships [5].In plants, chloroplast genes, including the maturase-coding gene (matK), the large subunit of ribulose 1,5-bisphosphate carboxylase-coding gene (rbcL), the non-coding plastid trnH-psbA intergenic spacer region and encoding subunit B of light-independent protochlorophyllide reductase (chlb), are usually used for molecular phylogenetic analyses [6][7][8][9][10].For example, Lahaye et al. [11] used the matK sequences of 1566 orchid specimens representing 1084 species in Costa Rica to identify species and reconstruct a phylogeny.In another study, the use of rbcL gene sequences enabled the majority of the samples (92%) to be identified to the genus level [12].In addition, nuclear ribosomal DNA (nrDNA) containing the internal transcribed spacer (ITS) region is also used in plant species identification.For instance, the ITS/ ITS2 regions could accurately and efficiently distinguish Corni Fructus and its adulterants, and provided a reference for the molecular identification of other Chinese herbal medicines [13].
Similarly, molecular systematics has also been used for Ephedra identification.Peng et al. [14] distinguished the Chinese Ephedra herb from other related species using ITS2 sequences.A novel method to authenticate the Ephedra herb, based on the chloroplast chlB gene and ITS sequence of nrDNA genes was developed and successfully applied to identify the ingredients of crude drugs obtained at a Chinese market [10].The method distinguished medicinal Ephedra plants from E. przewalskii, but their relationships were not recovered.Although the phylogenetic relationships in Ephedra were constructed from the chloroplast matK gene, rbcL gene and nrDNA ITS1 to study the geographic range and morphological diversity of the genus [15], some different Ephedra species were not distinguished, in particular E. intermedia and E. sinica.Recently, the studies on identifying Ephedra species growing in different locations in China based on the phylogenetic analyses of matK and ITS1 sequences have not been reported.
In this study, seven Ephedra species, including three medicinal Ephedra plants, and data deposited in GenBank, were used for species identification based on phylogenetic analyses.We chose the matK and ITS1 sequences to distinguish different species, especially medicinal from non-medicinal plants, and assess the intra-and inter-species relationships of Ephedra.

Plant materials
A total of 45 sequences belonging to seven species of the genus Ephedra and one outroup Gnetum montanum (Tab. 1) were used in this study.Eleven sequences representing four species were from GenBank and the others generated in this study from plants were collected from different locations of China, including Ningxia, Gansu, Inner Mongolia, Shanxi, Xinjiang, and Shaanxi and plant vouchers were deposited in the Ningxia Research Center of Modern Hui Medicine Engineering and Technology, China.
For the amplification of the entire matK or ITS1 regions, PCR was performed in 50 μL reaction mixture consisting of 19 μL sterile water, 25 μL 2× Power Taq PCR Master-Mix (Bio Teke), 2 μL of each primer (10 μM) and 2 μL of template DNA.Amplification conditions for matK consisted of one cycle of an initial denaturation at 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, annealing at 56-58°C for 40 s and extension at 72°C for 1.5 min, with a final amplification of 72°C for 10 min.Amplification program for ITS1 consisted of an initial step at 94°C for 5 min, followed by 35 cycles of 94°C for 30 s, 58-63°C for 40 s, and 72°C for 1.5 min, followed by a final extension at 72°C for 10 min.Amplifications were performed in the T100TM Thermal Cycler (BIO-RAD).PCR products were separated by agarose gel electrophoresis and purified with the Axyprep DNA Gel Extraction Kit (Axygen) according to the manufacturer's instructions.The purified PCR products were sequenced at Beijing Genomics Institution in China.Sequence data were submitted to GenBank and were assigned accession numbers ranging from KT286779 to KT286846.

DNA sequence data analysis
DNA sequences obtained for matK and ITS1, were aligned using Clustal X [17].The maximum likelihood (ML) and Bayesian inference (BI) methods were selected for the construction of phylogenetic trees.The sequence of Gnetum montanum Markgraf was used as the outgroup.Maximum likelihood analyses were processed using the MEGA 4 [18] program with the Kimura-2-Parameter model.The reliability of each branch was tested by a bootstrap analysis with 1000 replications.Bayesian inference analyses were performed using MrBayes 3.0b4 [19].Evolutionary models were chosen with MrModeltest 1.0b [20] in combination with PAUP 4.0 b10 [21].Each analysis consisted of two independent runs with four chains for 2 000 000 generations, sampling one tree every 100 generations.

Results
For the 34 samples containing three Ephedra species, the PCR-amplified fragments of both the matK gene and ITS1 regions were sequenced.The sequenced matK gene was 1408 base pairs (bp), and ITS1 was 918 bp.When the 11 sequences from GenBank were combined, the alignment of the 1141-bp regions of 45 different matK sequences revealed that 1122 bp (98.33%) were conserved, 19 bp (1.67%) were variable, and 14 bp (1.23%) were parsimony informative sites.The estimated transition/transversion ratio was found to be 2.1.The alignment of the 918 bp region of 45 different ITS1 sequences revealed that 848 bp (92.37%) were conserved sites, 70 bp (7.63%) were variable sites, and 66 bp (7.19%) were parsimony informative sites.The estimated transition/transversion ratio was 3.0.
In the BI analyses, the best-fit model (GTR + I) was selected based on the results of the Akaike information criterion or hierarchical likelihood ratio tests.
When the use of the matK gene and the ITS1 region were combined (Fig. 3), the seven Ephedra species could be clearly distinguished from each other.More importantly, the three medicinal plants could be distinguished from the other non-medicinal plants.

Discussion
The phylogenetic tree of the ITS1 sequences indicated that E. sinica and E. intermedia were clustered together, which was consistent with the results of Peng et al. [14], who discovered that the two species belonged to the same clade using the ITS2 sequences.Additionally, Wang et al. [22] found that E. sinica and E. intermedia clustered together using matK + rbcL and 18S + ITS sequences, indicating that the two species were closely related.Ephedra przewalskii formed a sub-cluster in the same clade as E. sinica and E. intermedia.Similarly, the phylogenetic tree of matK sequences showed that E. sinica, E. intermedia, E. przewalskii, and E. equisetina formed a clade, but that E. equisetina had a distant relationship with the other three species.Combining the matK and ITS1 sequence phylogenies, showed that E. sinica, E. intermedia, and E. przewalskii had a closer genetic relationships than E. equisetina.This was in agreement with Guo et al. [10] who showed that E. sinica, E. intermedia, and E. przewalskii were phylogenetically close to each other, while E. equisetina was an outgroup of the three Ephedra species.Similarly, Long et al. [23] also placed the three species (E.sinica, E. intermedia, and E. przewalskii) into one group based on ITS sequences.
Moreover, previous analyses of the ITS1 and ITS2 regions of nrDNA indicated that E. sinica and E. intermedia had identical sequences, while E. przewalskii had several nucleotide sites different from E. sinica and E. intermedia [23,24].Our study also distinguished E. przewalskii from E. sinica and E. intermedia.Yamaji et al. (2001; cited by [10]) reported that E. intermedia and E. przewalskii had the identical chloroplast rbcL sequence, and Guo et al. [10] also found that E. intermedia had the identical chlB sequence as E. przewalskii.In contrast, the present investigation indicates that E. intermedia and E. przewalskii have some different nucleotide sites not only in their matK sequences, but also in their ITS1 sequences, providing a simple method to identify the two species.Previous reports showed that E. sinica and E. equisetina could be identified based on the chlB sequence [10].Similarly, in our report just using the matK or ITS1 sequence, not only E. sinica and E. equisetina were identified, but E. intermedia and E. equisetina also were distinguishable.More importantly, E. intermedia could be separated from other species based on the matK gene, and E. equisetina could also be identified using either the matK or ITS1 sequence, providing a brief and rapid method to identify E. intermedia and E. equisetina.
In addition, as was shown in Fig. 3, E. equisetina and E. major clustered together with strong support, and this result implies that the two species have a close relationship, while E. antisyphilitica and E. aphylla clearly formed two clades, as seen in Fig. 3, indicating that the two species are genetically distant from the other species.
In conclusion, when the matK gene was combined with the ITS1 region, seven Ephedra species could be clearly distinguished from each other, including the medicinal and non-medicinal plants, which is significant for developing their pharmaceutical uses and is also important for the quality control of herbal medicines.

Fig. 2
Fig. 2 Phylogenetic analysis of Ephedra species based on maximum likelihood and Bayesian inference analyses of ITS1 sequence data.Numbers at the nodes indicate bootstrap values (% over 1000 replications ) and Bayesian posterior probability.

Fig. 3
Fig. 3 Phylogenetic analysis of Ephedra species based on maximum likelihood and Bayesian inference analyses of matK + ITS1 sequence data.Numbers at the nodes indicate bootstrap values (% over 1000 replications) and Bayesian posterior probability.