Introduction

Vicatia thibetica is a medicinal and edible plant species of the Apiaceae family, which belongs to the Apioideae subfamily (Pu et al. 2005). V. thibetica is distributed mainly in Tibet, Yunnan, and Sichuan provinces of China, growing on hillsides, grasslands, forests, river beaches, at an altitude of 2700–4000 m (She et al. 1979), and have been artificially cultivated in Dali city (Zhou et al. 2007). The dry root of V. thibetica was called Xigui (Zhou et al. 2007), a Bai ethnic medicine with a dual purpose of medicinal and edible use (Zhou et al. 2007; Jiang et al. 2016). Studies have shown that Xigui contained umbelliferone, bergapten, ferulic acid, apigenin, and other chemical components (Zhang et al. 2004). It has different pharmacological effects such as anti-aging, anti-fatigue, anti-dysmenorrhea, anti-oxidation, promoting intelligence, and improving immunity (Dong et al. 2018).

Xigui has been used by Bai folk in Dali City for more than a hundred years and has the efficacy of supplementing blood, invigorating qi, and regulating menstruation (Zhou et al. 2007; Jiang et al. 2016). Interestingly, Xigui was often used to substitute for Chinese herbal medicine Angelica sinensis in Northwest Yunnan and West Sichuan (Zhou et al. 2007). However, previous studies through differences in chemical composition revealed that Xigui could not be substituted for A. sinensis (Zhang et al. 2004). Since the morphological similarity of V. thibetica and A. sinensis makes it challenging to make an accurate distinction in the appearance, a molecular method is urgently needed to distinguish them.

Chloroplast is the site of photosynthesis in green plants and a vital organelle involved in the synthesis of pigments, lipids, hormones, and ribosomes (Raven and Allen 2003). Studying chloroplast genomes is essential for exploring plant molecular markers, the structure of chloroplast DNA, and species relationships (Tang et al. 2011). In addition, the plant cp genomes were characterized by a conserved structure and a high substitution rate and were considered a valuable source for plant molecular identification, genetic diversity assessment, and phylogenetic analysis (Dong et al. 2012, 2014). Park et al. (2019) found that the ycf4-cemA fragment could distinguish the herbal medicine Ligusticum officinale and Angelica polymorpha based on divergent region analysis of chloroplast genome. Zhang et al. (2019a, b) indicated that the chloroplast genome was used as a super-barcode to accurately discriminate various Dracaena species, successfully solving the problem of identifying Dracaena plants.

So far, the NCBI database has included above 5488 chloroplast genome sequences. However, there have been no reports on chloroplast genomes of  related species of Vicatia. Therefore, the entire V. thibetica cp genome sequence was obtained by Illumina NovaSeq sequencing in this study. Meanwhile, comparative analysis with the cp genome of other Apiaceae species, including A. sinensis, facilitates this genus’s phylogenetic and molecular identification studies. Moreover, this study provides a basis for the classification, identification, conservation genetics, and resource exploitation of Vicatia plants.

Materials and methods

Plant materials, DNA extraction, and illumina sequencing

Clean, fresh leaves of V. thibetica were obtained from Machang Town, Heqing County, Dali, Yunnan Province, China (100° 20′ 4″ E, 26° 3′ 57″ N; elevation 3010 m) and identified by Professor Cong-long Xia (College of Pharmacy, Dali University). Total DNA was extracted using the E.Z.N.A® Plant DNA kit (OMEGA). Extracted DNA was checked for quality and integrity by 1% agarose gel electrophoresis, followed by concentration and content using TBS380 Picogreen (Invitrogen). Then, DNA extracts were fragmented to 300–500 bp using Covaris M220 sonication. These fragments were then purified using TruSeq™ Nano DNA Sample Prep Kit for the trimming, 3’-end adenylation, and ligation index adaptor. The sequencing library was created by PCR amplification of appropriate size fragments, whose library was sequenced with paired-end (2 × 150 bp) using Illumina NovaSeq 6000 platform (Shanghai Biozeron Biotech Co, Ltd).

Genome assembly and annotation

Raw data for V. thibetica were created with 150 bp paired-end read lengths. Since there will be some data with lower quality in the raw sequencing data, to make the subsequent assembly more accurate, it will be quality sheared using the software Trimmomatic (Bolger et al. 2014).

After that, the use of NOVOPlasty software assembled high-quality reads into contigs (Dierckxsens et al. 2016). Next, clean reads were aligned back onto the scaffold obtained from the assembly, and the assembly results were locally assembled and optimized according to the paired-end and overlap relationship of the reads. Then, the inner holes of the assembly result were repaired using GapCloser v1.12 software.

Finally, the start position of assembled chloroplast sequence was corrected using the reference sequence, and the positions and orientations of the four chloroplast partitions (LSC/IRA/SSC/IRB) were determined, resulting in the ultimate genome sequence. Download the species closest to this species from NCBI as a reference sequence, and the assembly results were annotated using DOGMA (Dual Organellar Genome Annotator) (Wyman et al. 2004). A physical map of the cp genome was mapped in the OGDRAW (Organellar Genome Draw) program (Lohse et al. 2007). Finally, the sequence was submitted to NCBI with accession number MZ189732.

Codon preference analysis

Statistical and preference analyses of amino acid usage frequency and the relative synonymous codon usage (RSCU) of the 84 CDS sequences in the chloroplast genome of V. thibetica were performed using the codonw1.4.2 program (Sharp et al. 1986).

Repeat analysis in V. thibetica chloroplast genome

The SSRs were detected using MISA (Thiel et al. 2003) with the following thresholds: Ten repeat units for mononucleotide SSRs, five units for dinucleotide SSRs, four repeat units for trinucleotide repeat SSRs, and three repeat units for tetranucleotide, pentanucleotide, and hexanucleotide repeat SSRs. The online Tandem Repeats Finder (TRF) v4.04 was used to find tandem repeats (Benson 1999). Using the software REPuter (http://bibiserv.techfak.uni-bielefeld.de/reputer/) for long repeat analysis with parameters set to minimal repeat size of 30 bp, Hamming distance of 3 (Kurtz et al. 2001). Four repeat types, forward repeat (F), palindromic repeat (P), reverse repeat (R), and complement repeat (C), were detected with sequence similarity ≥ 90%.

Comparative genomic analysis of the V. thibetica chloroplast genome

Sequence similarity alignment analysis was performed using the Shuffle-LAGAN model (Brudno et al. 2003) of the mVISTA online program (http://genome.lbl.gov/vista/mvista/submit.shtml) (Frazer et al. 2004). Among them, five published chloroplast genomes (C. paradoxum [MK780227.1], B. chinense [NC_046774.1], L. sinense [NC_038088.1], A. sinensis [MH430891.1], and P. praeruptorumin [MN016968.1]) were selected for pairwise comparison using the plastome of V. thibetica as the reference genome. In addition, the boundaries of chloroplast genomes from nine species of the family Apiaceae were mapped using the IRscope online program (https://irscope.shinyapps.io/irapp/) (Amiryousefi et al. 2018).

Analysis of chloroplast genome by sliding window

V. thibetica was the species of this study, and the remaining eight complete chloroplast genome sequences were downloaded from NCBI. Chloroplast genomes were aligned using MAFFT v.7.129 (Katoh and Standley 2013) and BioEdit software (Hall. 1999). DanSP v 5.1 software was employed to perform sliding window analysis and computed nucleotide diversity index Pi. The step length and window length were set to 200 bp and 600 bp, respectively (Rozas et al. 2017).

Phylogenetic analysis

Phylogenetic inference was performed based on plastid genomes, and the complete chloroplast genome sequences of 37 Apiaceae species were used for phylogenetic tree construction (Table S5). The cp genomes of species E. trifoliatus and A. cordata of the Araliaceae family were set as outgroup. The selected sequences were first aligned using MAFFT v.7.129 (Katoh and Standley 2013) and then manually adjusted with BioEdit (Hall 1999). After that, the aligned sequences were analyzed for phylogenetic reconstruction based on the maximum likelihood method using the IQTree software with 1000 bootstrap replicates (Nguyen et al. 2015). In the Partition Finder V2.1.1 software, the most suitable model of nucleotide substitution was selected using the Bayesian information criterion (BIC) (Lanfear et al. 2012). Reconstruction of the ML tree with the best-fit model TVM + F + R7. Bootstrap values were computed using UFBoo built into IQTree, computed by fast bootstrapping to avoid low support values (Minh et al. 2013). In addition, we also constructed a Neighbor-Joining (NJ) phylogenetic tree by MEGA X software with kimura-2 as the parameter and 1000 bootstrap values.

Results

Composition and characteristics of the chloroplast genome

The complete circular four-segment structure of the V. thibetica chloroplast genome was obtained, with a total length of 145,796 bp, consisting of a large single-copy region (LSC, 92,186 bp), a small single-copy region (SSC, 17,452 bp), and a pair of inverted repeat regions (18,079 bp) (Fig. 1). Annotation of the whole genome sequence of V. thibetica resulted in 128 genes, including 37 tRNA genes, eight rRNA genes, 84 protein-coding genes, and one pseudogene ycf1. We observed 14 genes with duplications in the IRs region, containing four PCGs (one gene containing introns), four rRNAs, and six tRNAs (two genes containing introns) (Tables 1, 2). Meanwhile, these genes contained 19 identified introns, of which 11 were PCGs and eight were tRNAs. The ycf3 and clpP were genes that contained two introns, whereas all others contained only a single intron (Table S1). Moreover, total GC content was 37.7% in the V. thibetica plastome, whereas GC content in IR region (44.8%) exceeded that in LSC (36.1%) and SSC (31.1%) regions (Table 1). In addition, the genome was composed of approximately 56.4% of the coding region (82,220 bp) and 43.6% of the non-coding region (63,576 bp). The frequencies of adenine (A), thymine (T), cytosine (C), and guanine (G) in the cp genome of V. thibetica were 44,785, 46,091, 28,055, and 26,864 bp, accounting for 30.7%, 31.6%, 19.2%, and 18.4% of the genome, respectively (Table 1).

Fig. 1
figure 1

Physical map of V. thibetica cp genome. Genes shown inside the circle are transcribed clockwise, and genes outside the circle are transcribed counter-clockwise. Thick black lines represent inverted repeats

Full size image
Table 1 Summary of chloroplast genome characteristics of V. thibetica
Full size table
Table 2 List of all genes present in the V. thibetica chloroplast genome
Full size table

In V. thibetica, a total of 84 protein-coding genes encoding 23,508 codons were observed. These included 61 unique codons encoding 20 amino acids and three stop codons (Fig. 2). Of these, 2506 (10.66%) codons encoded leucine and 252 (1.36%) codons encoded cysteine, which were the most and least encoded amino acids, respectively, in the V. thibetica chloroplast genome. Thirty codons with an RSCU value greater than 1 indicated a high preference for codon usage, whereas 32 codons with an RSCU value less than 1 indicated a low preference for usages (Table S2).

Fig. 2
figure 2

Analysis of relative synonymous codon usage (RSCU) in the chloroplast genome of V. thibetica

Full size image

Comparison with chloroplast genomes of other apiaceae

Comparative analysis based on the nine genera of chloroplast genomes of family Apiaceae (Vicatia, Angelica, Peucedanum, Ligusticum, Semenovia, Heracleum, Bupleurum, Chamaesium, and Hydrocotyle) has never been reported earlier. The overall GC content of these species ranged from 36.7% (S. gyirongensis) to 38.6% (L. sinensis). The size of the V. thibetica cp genome was slightly similar to that of the L. sinense and P. praeruptorum (Wu et al. 2020; Li et al. 2019). In addition, the plastid genome was largest for B. chinense (155,545 bp) and smallest for A. sinensis (142,822 bp). In all species, the length of the SSC region was conservative, ranging from 17,505 bp (B. Chinese) to 18,690 bp (H. sibthorpioides). IR regions varied markedly in length, from 25,108 (A. sinensis) to 52,610 bp (B. chinense). The IR were the smallest in A. sinensis (25,108 bp). The length of the LSC region is highest for A. sinensis (99,964 bp) and shortest for C. paradoxum (84,162 bp).

Except for 134 genes observed in species belonging to the Hydrocotyloideae subfamily (H. sibthorpioide), the total gene numbers in all species were conservative (Table 3). In most species, 129 genes were identified, and eight rRNAs (four rRNA duplications) were detected in each species. V. thibetica, C. paradoxum, B. chinense, and P. praeruptorum contained 84 protein-coding genes. S. gyirongensis, H. sibthorpioides, and H. yungningense contained 85 protein-coding genes, whereas L. sinense contained 87 and A. sinensis contained 83. The V. thibetica, A. sinensis, and P. praeruptorum chloroplast genomes possessed 35 tRNA genes, C. paradoxum, B. chinense, and H. sibthorpioides possessed 37 tRNA genes, whereas the L. sinense, S. gyirongensis, and H. yungningense chloroplast genome possessed 36 tRNA genes (Table 3).

Table 3 Comparison of characteristics of the Apiaceae species cp genomes
Full size table

Repeats analysis

Simple sequence repeats (SSRs) are extensively used in species evolution and population genetics studies (Powell et al. 1995; Roullier et al. 2011). We determined the type, distribution, and frequency of simple sequence repeats in the V. thibetica plastome. A total of 75 perfect SSRs were identified by MISA analysis, which included 40 mono-, 22 di-, one tri-, eight tetra-, two penta- and two hexanucleotide repeats (Fig. 3A). In the V. thibetica whole chloroplast genome, 10, 53, and 12 SSRs were present in the IR, LSC, and SSC regions (Fig. 3B). Thus, simple sequence repeats were primarily distributed in the LSC region, which accounted for 71% of the total, whereas IR regions (13%) were least distributed (Fig. 4A). The repeat type that occurred three times was the most frequent in the LSC region, including tetranucleotide, pentanucleotide, hexanucleotide, whereas other frequencies of classified repeat types appeared either once or did not appear  in LSC, SSC, and IR regions (Fig. 4B–D). The results showed that mononucleotide motifs accounted for approximately 53% of SSRs. Out of these, mononucleotide A and T repeat units were the most dominant parts (Table S3), accounting for 46.7%.

Fig. 3
figure 3

Analysis of SSR repeat types and numbers in V. thibetica chloroplast genome. A SSRs in the complete cp genome. B SSRs in the LSC, SSC, and IR regions

Full size image
Fig. 4
figure 4

The type, distribution, and frequency of SSRs in the V. thibetica chloroplast genome. A Distribution of SSRs in the LSC, SSC, and IR regions. B–D Frequencies of repeat motifs in the LSC, SSC, and IR regions

Full size image

Repetitive sequence utilized illegitimate recombination and slipped strand mismatches play a significant role in genomic rearrangements and mutation (Bausher et al. 2006; Jansen et al. 2007). Twenty-nine tandem repeats were determined in the plastome of V. thibetica (Table S4), which was higher compared to Chamaesium paradoxum (15), Bupleurum chinense (26), Ligusticum sinense (24), and smaller compared to Angelica sinensis (32), and Peucedanum praeruptorum (38). Detected repeats were distributed in intergenic spacer (IGS), protein-coding regions (CDS), and intronic regions, while most tandem repeats were in IGS and intronic regions. Of the tandem repeats detected in the V. thibetica chloroplast genome, 23 were situated in the IGS regions, and two were distributed in the intronic region of trnI-GAU. However, just three and one repeats were presented in CDS regions of ycf2 and ycf1, respectively. The size of these repeats varied from 25 to 98 bp, which was the longest repeat present in the IGS region of rps16/trnQ-UUG.

Long repeats play an essential role in the variation, expansion, and rearrangement of the complete chloroplast genome (Asaf et al. 2018). The following four repeat types were observed by the software REPuter: forward (F), reverse (R), complement (C), palindromic (P). The minimal repeat size is 30 bp for all repeat types. There were 19 F repeats, 16 P repeats, one R repeats, and one C repeats in the cp genome of V. thibetica (Table 4). In all, there were 37 long repeats in V. thibetica plastome. Most of the repeat sizes were between 30 and 39 bp (81.08%), followed by 40–49 bp (16.21%), whereas 50–59 bp were the least (0.03%). Meanwhile, the R and C repeat of the V. thibetica cp genome only contained 30–39 bp (Fig. 5). At the first long repeat position, 83.8% of repeat sequences were observed in non-coding regions. Two repeats were situated in the tRNAs (0.06%), the other five repeats (13.51%) were situated in the PCGs, particularly psaA, psaB, and ycf2 (Table 4).

Table 4 The distribution of long repeats sequences in V. thibetica chloroplast genome
Full size table
Fig. 5
figure 5

Long repeat sequences in cp genome of V. thibetica. REPuter was used to identify repeat sequences with a length ≥ 30 bp, and sequences identified ≥ 90% in the chloroplast genomes. F, P, R, and C indicate the repeat types F (forward), P (palindrome), R (reverse), and C (complement). Repeats with different lengths are indicated in different colors

Full size image

Comparative genomics analysis

This study performed a global alignment of published whole chloroplast genome sequences from five Apiaceae species using the online genome alignment tool mVISTA, while V. thibetica was set as reference sequence (Fig. 6). This result confirmed that the variation in the non-coding region of the six cp genomic sequences was higher compared to that in the conserved protein-coding region. In addition, the LSC region, and the SSC region had significantly higher number of variation compared to the IR regions, whereas the rRNA gene region was highly conservative, with few variants detected. The genes which showed more variations were matK, rpoC2, rpoC1, rpoB, ycf1, ycf2, and ndhF, while others were very highly conserved. Variants in IGS were exceeded in gene regions such as rps16-trnQ-UUG, atpH-atpI, rpoB-trnC-GCA, trnE-UUC-psbD, ndhC-trnV-UAC, rpl16-rps3, ycf4-cemA, and petA-psbJ. As illustrated in Fig. 6, the plastid genome of L. sinensis, A. sinensis, and P. praeruptorum had high similarity with V. thibetica, indicating that they are closely related species. However, the genomes of C. paradoxum and B. chinense presented more variations compared with the reference genome.

Fig. 6
figure 6

Comparing the chloroplast genome sequences of V. thibetica, C. paradoxum, B. chinense, L.sinense, A. sinensis, P. praeruptorum generated with mVISTA. Grey arrows indicate the position and direction of each gene. Red and blue areas show the intergenic and genic regions, respectively. The vertical scale indicates the percentage of identity, ranging from 50 to 100%

Full size image

We further investigated the LSC/IRb/SSC/IRa borders in the chloroplast genomes of V. thibetica and eight Apiaceae, including C. paradoxum, B. chinense, L. sinense, A.sinensis, S. gyirongensis, P. praeruptorum, H. sibthorpioides, and H. yungningense (Fig. 7). As shown in Fig. 7, ycf1 was situated at the SSC/IRa boundary of all nine chloroplast genomes, indicating that it is a universal characteristic of the Apiaceae cp genome. At the IRb/SSC junction, ndhF and the duplicated pseudogene ycf1 had 3 and 37 bp overlapping in the V. thibetica and H. sibthorpioides cp genomes, and the duplicated ycf1 gene was already absent in C. paradoxum, B chinense, and S. gyirongensis. Except for V. thibetica and B. chinense, the other cp genomes contained trnH genes.

Fig. 7
figure 7

Comparison of the borders of the large single-copy (LSC), small single-copy (SSC), and inverted repeat (IR) regions among nine Apiaceae chloroplast genomes. JLA, junction between LSC and IRa; JLB, junction between LSC and IRb; JSA, junction between SSC and IRa; JSB, junction between SSC and IRb

Full size image

Furthermore, no novel genes were observed in the cp genome of V. thibetica compared with the gene composition of other Apiaceae family species, yet two genes (psbA and trnH) were lost. The trnL gene was only observed in C. paradoxum and V. thibetica. Finally, it was worth mentioning that the novel gene ORF8 was observed in the A. sinensis cp genome. The above results demonstrated that the plastid genomes of these nine Apiaceae species were relatively divergent.

Highly variable hotspot analysis of chloroplast genomes

Nucleotide diversity (Pi) values of plastomes among nine species were calculated using a sliding window to evaluate the level of diversity in different regions with species including V. thibetica, C. paradoxum, B. chinense, L. sinense, A. sinensis, S. gyirongensis, P. praeruptorum, H. sibthorpioides, and H. yungningense (Fig. 8). Highly divergent sites among nine Apiaceae species of cp genomes were detected using DNA polymorphism analysis. There were 17,234 variable sites, 3483 parsimony informative sites, and 1459 indels detected in nine plastid genomes, causing Pi values ranging from 0 to 0.1313. Meanwhile, we observed low variability in IRs compared to LSC and SSC partitions. Overall, the average nucleotide diversity within the nine Apiaceae cp genomes was 0.03650, representing a high divergence level among those species. Afterwards, it was further confirmed that the regions showing the higher Pi value peaks (> 0.09) were rps16 (pi = 0.11), ndhc-trnV-UAC (pi = 0.09), clpP (pi = 0.10), ycf1 (pi = 0.13), ndhB (pi = 0.11), including four gene regions and one intergenic region. These variable loci may provide valuable information for resolving species identification, phylogenetic relationships, and genetic diversity.

Fig. 8
figure 8

Analysis of sliding windows across chloroplast genomes in nine Apiaceae. X-axis: position of the window's midpoint; Y-axis: nucleotide diversity (Pi) per window. The red line represents the threshold value for variant loci (Pi threshold = 0.09). The high variation loci of the Apiaceae genome were marked in the figure. LSC, large single-copy; IRb, inverted repeat; SSC, small single-copy; IRa, inverted repeat

Full size image

Phylogenetic analysis of the cp genomes

We constructed a maximum likelihood (ML) tree with 37 entire plastid genomes to determine the phylogenetic position of V. thibetica (Fig. 9). Of these, we downloaded 36 already published complete chloroplast genomes containing ten genera from NCBI (Table S5). The Eleutherococcus trifoliatus and Aralia cordata genome sequences of the Araliaceae family were used as outgroup. In parallel, entire topology of the ML tree and the node support for the principal clades were consistent with the NJ (Neighbor-Joining) tree (Fig. S1). As shown in Fig. 9, all species submitted for analysis were divided into three clades, corresponding to the Apioideae, Hydrocotyloideae subfamily, and outgroup, respectively. Here, V. thibetica was gathered into a single branch in the subfamily Apioideae, which verified the classification research of the previous scientists on the Apiaceae (Apioideae) (Watson 1998). Furthermore, V. thibetica was clustered together with genera Angelica, Peucedanum, and Ligusticum. A close relationship among them was also uncovered. Overall, Vicatia should be considered a separate genus in the Apiaceae family.

Fig. 9
figure 9

Plastome-based phylogenetic relationship among  37 Apiaceae species. The phylogenetic tree was constructed with the maximum likelihood (ML) method. The plastome sequences that E. trifoliatus and A. cordata of Araliaceae were used as outgroup. Values beside branch nodes denote support values for the bootstrap

Full size image

Discussion

Genomic comparison

The chloroplast genome of V. thibetica reported in this study exhibited a circular quadripartite structure and was 145,796 bp in length, similar to the sequence of L. sinensis (Table 3). Thus, it indicated that V. thibetica and L. sinense were related species compared with other Apiaceae species. However, we have also noted that the cp genomes of the nine Apiaceae species, including V. thibetica, were varied in length (145,796–155,545) (Table3) (Wu et al. 2020; Li et al. 2019; Zheng et al. 2019, 2020; Zhang et al. 2019a, b; Tian et al. 2019; Xiao et al. 2019; Ge et al. 2017). Generally, the expanded IR regions and variable SSC regions were considered the main factors which were contributing to the chloroplast genome length variation in angiosperms (Kim and Lee 2004).

Furthermore, the difference was relatively significant in the comparative analysis of V. thibetica and A. sinense. The novel gene ORF8 was observed in the chloroplast genome of A. sinensis, but not in the V. thibetica. Insertions and deletions of genes were observed in the cp genomes of V. thibetica and A. sinense, which could be molecular markers to identify the two species (Fig. 7). At the same time, the variable fragments obtained from the whole genome alignment of V. thibetica and A. sinense also provided important information for their plant identification (Fig. 6). Despite their similarity in efficacy and morphology, they can be distinguished based on chloroplast genome super-barcode.

Variation of apiaceae cp genome

A common feature of Apiaceae chloroplast genomes revealed by genomic variation analysis is that the regions of IR were more conservative compared to LSC and SSC. Moreover, the variation level was significantly greater in non-coding compared to coding regions. In general, non-coding regions could be divided into intronic and intergenic regions, whereas intergenic spacers were highly variable among Apiaceae species in this study (Shaw et al. 2007). Previously, highly variable hotspots in the cp genome have been demonstrated to be useful for species identification and analysis of phylogeny while providing critical information at the population level to explore species differences and population structure (Bi et al. 2018; Du et al. 2017). Therefore, the rps16, ndhc-trnV-UAC, clpP, ycf1, and ndhB variable hotspots obtained by sliding window analysis (Fig. 8) could be used not only as molecular markers for the identification of V. thibetica and A. sinense, but also for the phylogenetic investigation of Apiaceae species. Similar results associated with these variant regions were also reported in recent researches of Fritillaria (Bi et al. 2018; Chen et al. 2019) and Dioscoreales (Biju et al. 2019).

Repeat analysis

Thirty-seven long repeats of 30–59 bp were identified in the cp genome of V. thibetica. However, in the reported species of the genus Ligusticum, 308 repeats of 30–82 bp were identified (Ren et al. 2020), indicating that long repeats are variable between related species and may serve as molecular markers for species identification (Nie et al. 2012). We observed highest number of SSRs in the LSC region, and the reasonable explanation is that the LSC region is more extended compared to SSC and IR (Ren et al. 2020). Among these SSRs, the A/T units of mononucleotide repeats are the most abundant, which may be explained by the higher proportion of  polyA and polyT in the cp genome (Zhu et al. 2020). The identified SSRs may provide candidate SSR markers for molecular genetic correlation studies of medicinal plants in the genus Vicatia.

Phylogenetic analysis

We reconstructed ML and NJ trees using the whole cp genomes of 37 species, including V. thibetica. First, the Apioideae and Hydrocotyloideae subfamilies were clustered separately in two large clades of the phylogenetic tree, and these two large clades were further divided into distinct clades. Then, Vicatia, Angelica, Peucedanum, Semenovia, Ligusticum, Heracleum, Bupleurum, and Chamaesium clustered in Apioideae. And then, Angelica, Vicatia, Peucedanum, Ligusticum clustered in one smaller branch. On this smaller branch, when assigned, the genera Angelica, Peucedanum, and Ligusticum are sister groups, whereas Vicatia forms a separate branch independently. These observations indicated that the genera Vicatia, Angelica, Peucedanum, Ligusticum have been closely related, whereas V. thibetica was retained in the genus Vicatia. Our results have been consistent with previous studies (Pu 2005). However, the phylogenetic position of V. thibetica is still not sufficiently clear due to the lack of chloroplast genomic information for other Vicatia species. Therefore, it is indispensable to increase the cp genome of other Vicatia species.

Conclusions

This study reported the first chloroplast genome of V. thibetica and compared it with other Apiaceae species. The chloroplast genome of V. thibetica was similar in size, number of genes, genomic structure, and gene order to other angiosperm chloroplast genomes. We detected 75 SSRs, 29 tandem repeats, and 37 long repeats useful for genetic breeding and population genetics studies within Vicatia. In addition, the highly variable sites and divergent regions of nine Apiaceae species were identified as possible pathways for further use in studying genetic markers for population genetics. Comparative analysis with other species of the family Apiaceae showed that the IR regions were more conservative compared to the SSC and LSC, suggesting that more DNA barcodes could be developed from these regions to identify species. Meanwhile, the chloroplast genome could be used to distinguish V. thibetica from A. sinensis. Therefore, it may be unreasonable to use V. thibetica instead of traditional Chinese medicine A. sinensis. The phylogenetic study based on 37 complete chloroplast genomes showed that V. thibetica formed a single clade within the Apioideae subfamily, supporting the earlier view that Vicatia is an independent genus of the family Apiaceae (Watson 1998; Pu 2005).