Complete Chloroplast Genome Sequencing and Phylogenetic Analysis of Two Dracocephalum Plants

Dracocephalum tanguticum and Dracocephalum moldavica are important herbs from Lamiaceae and have great medicinal value. We used the Illumina sequencing technology to sequence the complete chloroplast genome of D. tanguticum and D. moldavica and then conducted de novo assembly. The two chloroplast genomes have a typical quadripartite structure, with the gene's lengths of 82,221 bp and 81,450 bp, large single-copy region's (LSC) lengths of 82,221 bp and 81,450 bp, and small single-copy region's (SSC) lengths of 17,363 bp and 17,066 bp, inverted repeat region's (IR) lengths of 51,370 bp and 51,352 bp, respectively. The GC content of the two chloroplast genomes was 37.80% and 37.83%, respectively. The chloroplast genomes of the two plants encode 133 and 132 genes, respectively, among which there are 88 and 87 protein-coding genes, respectively, as well as 37 tRNA genes and 8 rRNA genes. Among them, the rps2 gene is unique to D. tanguticum, which is not found in D. moldavica. Through SSR analysis, we also found 6 mutation hotspot regions, which can be used as molecular markers for taxonomic studies. Phylogenetic analysis showed that Dracocephalum was more closely related to Mentha.


Introduction
Lamiaceae is the sixth largest family of phanerogam. There are more than 7,000 species in about 220 genera and 10 subfamilies in the world. In China, there are over 800 species in 99 genera. This family is mainly distributed in Asia, Africa, and Europe. There are about 10 genera with more than 100 species, including Salvia, Mesosphaerum, and Scutellaria, among which the number of species accounts for about two-thirds of the species in this family. The plants of this family draw people's attention because of their aroma and medicinal value [1,2], and they have great economic value. Among them, there are more than 60 kinds of plants in the Dracocephalum, which are mainly distributed in northern Asia, and a few extend to northern Europe or central Europe.
Chloroplast is a kind of plastid and one of the most important organelles in plants and algae cells. In addition to photosynthesis, chloroplasts also participate in the   BioMed Research International synthesis of starch, pigments, amino acids, and fatty acids [3][4][5]. Chloroplasts are green and oblate under a high-power microscope [6]. The currently accepted view on the origin of chloroplasts is the endosymbiosis theory. The chloroplasts in eukaryotic cells originated from endosymbiont cyanobacteria, which were engulfed and gradually digested by eukaryotic cells before becoming organelles in host cells through a long period of evolution [7]. As semiautonomous genetic organelles, chloroplasts have their own genetic material-chloroplast genome (cpDNA) [8,9]. The length of the genome is about 120-160 kb, and the difference in the size of the chloroplast genome is mainly caused by the shrinkage and expansion of the IR region [10]. For example, the IR size of the chloroplast genome of Pelargonium xhortorum is 75,741 bp; the total length of its chloroplast genome is 217,942 bp [11]; the IR of Pinus thunbergii is only 495 bp, and its chloroplast genome is only 119,707 bp [12]; the chloroplast genome of Metasequoia glyptostroboides is much shorter due to loss of an IR region [13]. Genome sequencing is often used in the analysis of phylogenetic relationship and research on genetic diversity and evolution [14,15]. The three independent genomes that can provide genetic information are chloroplasts, mitochondria, and cell nucleus. Compared with the complexity of the nuclear genome, the nonconservative nature of the large mitochondrial genome, the chloroplast genome is proper in size and easy for sequencing and acquisition [16][17][18]. Since 1986, the entire chloroplast genome of tobacco was first sequenced successfully [19]. To date, the NCBI GenBank database has collected the complete sequences of chloroplast genomes of more than 1,000 plants. The Lamiaceae family comprises about 200 species. In 2003, Paul Hebert introduced the concept of DNA barcoding, and gene sequencing technology began to be applied to biological species identification systems. Common DNA barcodes used by plants include rbcL, matK, psbK-psbI, trnH-psbA, and ITS [20]. However, using only a single gene for species identification may lead to different results from different genes, which brings about difficulties for species identification [21,22]. Chloroplast genomes offer more variation than a single gene and will be more effective in addressing the evolutionary rules of close species.
In this study, we used Illumina sequencing technology to conduct full genome sequencing to the chloroplast genome of Subunits of ATP synthase atpA, atpB, atpE, atpF, atpH, atpI Large subunit of rubisco rbcL

BioMed Research International
D. tanguticum and D. moldavica and then conducted de novo assembly and analysis. Moreover, based on the chloroplast genome data of D. tanguticum and D. moldavica and the chloroplast genomes of 30 other species, a phylogenetic tree was constructed. These data will help us better understand the evolutionary history of the branch of Lamiaceae and promote the research on the phylogeny, population, and genetic engineering of medicinal plants.

Characterization of Chloroplast Genomes in Labiatae
Species. A total of 12.27 Gb raw data were obtained after sequencing. The chloroplast genome size, reads, and GC content of the two Dracocephalum plants are shown in Table 1. The chloroplast genome size of Lamiaceae plants is basically similar, which is about 150 kb. Through genome sequencing and de novo assembly, we obtained the complete chloroplast genome map of D. tanguticum and D. moldavica (Figure 1).
The total length of the chloroplast genomes of D. tanguticum and D. moldavica was 150,954 bp and 149,868 bp, respectively. Both D. tanguticum and D. moldavica contain a long single-copy region (LSC) with the lengths of 82,221 bp and 81,450 bp and a short single-copy region (SSC) with the lengths of 17,363 bp and 17,066 bp, respectively, which are separated by a pair of inverted repeat regions (IR) with the lengths of 51,370 bp and 51,352 bp, respectively (Table 1, Figure 1). The GC content of the chloroplast genome of these Lamiaceae plants is about 38%, and the GC content of the IR region, LSC region, and SSC region is about 43%, 36%, and 32%, respectively. D. tanguticum and D. moldavica encode 133 (114 unique genes) and 132 (113 unique genes) genes, respectively, including 37 tRNA genes and 8 rRNA genes, as well as 88 and 87 protein-coding genes, respectively. In addition, the rps2 is a gene unique to D. tanguticum, which is not found in D. moldavica (Table 2). Gene deletions in other Lamiaceae species have also been reported before, and the deletion of the trnK-UUU gene in the chloroplast genome of Pogostemon cablin has ever been reported [23]. Generally speaking, the IR region is the most conservative region in the chloroplast genome [24]. The expansion and shrinkage of the IR, LSC, and SSC regions during evolution are the main reason for the difference in the length of the chloroplast genome [15,25].
In addition, 18 genes were found to contain intron sequences, of which 16 genes contain one intron and 2 genes contain two introns. Among them, the intron in trnK-UUU is the largest and contains the matK gene ( Table 3). The size of introns and gene spacers is also a factor that affects the size of the chloroplast genome, and intron loss has been reported in Hordeum vulgare [26,27]. The two pseudogenes, ycf1 and rps19, are located between IR and SSC and between IRa and LSC, respectively, which is consistent with the research in Prunella vulgaris [28]. CodonW software was used to calculate and analyze the codon usage bias (RSCU) of the chloroplast genome. The protein-coding sequences of the chloroplast genomes of D. tanguticum and D. moldavica consist of 26,754 and 26,460 codons, respectively. Among them, leucine (Leu) is used the most, and cysteine (Cys) is used the least. Such bias is considered to be a comprehensive result of natural selection, species mutation, and genetic drift [29,30]. The complete chloroplast genome with gene annotations has been uploaded to the NCBI database, the accession numbers of D. tanguticum and D. moldavica are MT457746 and MT457747, respectively.

Comparative
Analysis of Genome Structure. The shrinkage and expansion of the boundary region between IR, LSC,  Figure 2. The rps19 gene is located at the boundary of LSC/IRa. The size of the gene fragment in the LSC region is 187-281 bp, and the size of the gene fragment in the IRa region is 42-92 bp. The rps19 and ycf1 genes are usually crossborder genes, which is consistent with the reports on Pongamia pinnata [33]. The rpl2 gene is located at the IR/LSC border, which is 1486 bp in both D. tanguticum and D. moldavica, but the distance from rpl2 to the border in the D. tanguticum is 2 bp longer than that in the D. moldavica. Additionally, ndhF is also a gene that crosses the IRb/SSC border, but it was found in the analysis of the six Lamiaceae species that the ndhF gene in Rosmarinus officinalis is located in the SSC region, 213 bp away from the border (Figure 2), which is different from that of the other five Lamiaceae plants. Studies have shown that the small amplification of the IR region border is considered to be an important reason for maintaining the stability of the IR region [34]. Goulding et al. proposed the hypothesis of the evolution of the chloroplast's IR region: there is a boundary gene amplification mechanism in the IR region [35].

Analysis on Simple Repeat Sequences and Mutation
Hotspots. We used vmatch v2.3.0 (http://www.vmatch.de/) software to identify the dispersed duplication sequences (Table 4) The mauve software was used to compare the complete chloroplast genome of the two sequenced species with ten other Lamiaceae plants. Comparing the LCBs (local-collinear blocks) of these species, it was found that most of the genes maintained a consistent position and orientation, and the complete genome of the plastid showed a high degree of consistency, with no gene reversal detected. Through comparison of the complete chloroplast genome of 12 Lamiaceae plants, it was found that the rpl2 gene was in the first location in the LSC region in Lavandula angustifolia, while the other 11 plants were in the last location in the IR region ( Figure 3). This is different from predecessors' studies [36,37]. The two differences can be used as the basis for species identification and molecular markers.

BioMed Research International
D. moldavica were 32 (62.75%) and 35 (64.81%), respectively, with a large proportion. Among SSRs of D. tanguticum and D. moldavica, the four-nucleotide repeat sequences are much more than three-nucleotide repeat sequences, which is consistent with the study of Salvia miltiorrhiza [38,39].
Genes such as rbcL, matK, and atpB are widely applied as molecular markers in general phylogenetic studies [40]. After analysis of the SSRs of D. tanguticum and D. moldavica, we also found six mutation hotspot regions: trnK-UUU-psb1, trnR-UCU-atpI, rpoC2-trnC-GCA, psaA-trnL-UAA, accD-cemA, and clpP. Five of them are located in the intergenic region (IGS), and one is located in the intron region. Through comparison of the two Dracocephalum plants, the order of mutation rate from high to low is the noncoding region, the intron region, the LSC region, the SSC region, and the IR region.

Evolutionary and Phylogenetic Analysis.
In this study, we combined the sequenced chloroplast genomes of two Dracocephalum plants with that of 30 other angiosperms, constructed a phylogenetic tree with RAxML v8.2.10 (https:// cme.h-its.org/exelixis/software.html) software based on the maximum likelihood (ML) method, and selected Cistanche deserticola (NC_021111.1) as the outgroup. Phylogenetic trees based on complete chloroplast genome data have a higher bootstrap value (the branch with a value greater than 75 is considered as a more stable branch). The results showed that D. tanguticum and D. moldavica belong to the same genus. Although the D. tanguticum is more similar to Rosmarinus morphology, the phylogenetic tree shows that the relationship between the two Dracocephalum plants and Mentha is closer than that with Rosmarinus ( Figure 4). This result is consistent with traditional taxonomy. In the seed morphology, Dracocephalum is closer to Mentha [28,41].

Materials and Methods
3.1. Material Acquisition. We used the leaves of D. tanguticum and D. moldavica for chloroplast genome sequencing. The vigorous, healthy, and fresh leaves of D. tanguticum and D. moldavica were obtained from the Science and Technology Park of Xizang Agriculture and Animal Husbandry College, Linzhi City, Tibet Autonomous Region, China. All leaves were immediately frozen in liquid nitrogen and stored at -80°C before analysis and use.   We used REPuter (http://bibiserv.techfak.uni-bielefeld .de/ter/) to look for and analyze the size and position of forward, reverse, palindrome, and complementary repeat sequences. MISA (http://pgrc.ipk.gatersleben.de/misa/) was adopted to identify SSRs, the threshold for single nucleotide SSRs is 10 repeats, the threshold for double-nucleotide SSRs is 5 repeats, the threshold for three-nucleotide SSRs is 4 repeats, and the threshold of four-, five-, and six-nucleotide SSRs is 3 repeats. The mauve software was used to sequence alignment of the chloroplast genome.

Phylogenetic Analysis.
Based on the chloroplast genome data of D. tanguticum and D. moldavica and the chloroplast genomes of 30 other species, a phylogenetic tree was constructed, and Cistanche deserticola (NC_021111.1) chloroplast genome was selected as the outgroup. Before constructing the phylogenetic tree, we used MAFFT software (V7.427, auto mode) for multisequence alignment to obtain neatly arranged chloroplast genomes for phylogenetic analysis. By default, the whole genome is used for evolutionary tree analysis, and the ring sequence is set at the same starting point. MAFFT software was used to perform multiple sequence alignment between species sequences, and tri-mAl (v1. 4   BioMed Research International software was used, its GTRGAMMA model was selected for Bootstrap analysis, and the Bootstrap = 1000 was set to build the maximum likelihood evolutionary tree.

Conclusion
The size, GC content, gene order, and number of chloroplast genomes in D. tanguticum and D. moldavica are highly similar. As D. moldavica lacks the rps2 gene compared to D. tanguticum, which may be one of the reasons that D. moldavica is 1,086 bp shorter than D. tanguticum in the full length of the chloroplast genome. Through SSR analysis, six mutation hotspot regions were found, among which five were located in the intergenic region, and one was located in the intron region. Through phylogenetic tree analysis, it was found that Mentha and Dracocephalum have a closer genetic relationship. In this study, we have provided the complete chloroplast genome of D. tanguticum and D. moldavica, which provides a basis for the identification and overcoming of phylogenetic problems at the species level.

Data Availability
The complete chloroplast genome data used to support the findings of this study have been deposited in the NCBI repository (the accession numbers of D. tanguticum and D. moldavica are MT457746 and MT457747, respectively).

Conflicts of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.