Background

Apioideae is the most complicated subfamily of the Apiaceae family in taxonomy, and this subfamily comprises approximately 380 genera and 3,200 species (Angiosperm Phylogeny Website, Stevens, updated 2021, https://www.mobot.org/MOBOT/research/APweb/). Its members are widely distributed all over the world [1]. Its earliest lineage appeared in Southern Africa, but representatives appeared more frequently in the north temperate zone of Eurasia [2, 3]. Apioideae is definitely monophyletic and subdivided into 16 tribes and 14 clades, but many tribes and clades of the subfamily are not monophyletic [4,5,6]. Furthermore, circumscriptions of some genera are poorly delimited, and species of these genera always cluster to more than one tribe or clade based on molecular phylogenetic studies [6]. Therefore, Apioideae is a puzzle in terms of classification and phylogeny.

The Sinodielsia clade was established as a novel clade in Apioideae based on nuclear ribosomal (nr) DNA internal transcribed spacer (ITS) sequence data in 2008 [7] and initially included 14 species from 10 genera of Apioideae (Table 1). Since then, the classification of the Sinodielsia clade has attracted considerable interest, and phylogenetic studies have spurred taxonomic realignments of relative genera so that an increasing number of species from other genera were transferred to this clade in view of nrDNA ITS and chloroplast (cp.) DNA data [8,9,10]. Currently, the Sinodielsia Clade comprises 37 species from 17 genera in total (Table 1), and some of them are important medicinal herbs with great economic value: Angelica sinensis (Oliv.) Diels, Ligusticum sinense Oliv., Cnidium dahuricum (Jacq.) Turcz. ex Fisch. & C.A. Mey., Conioselinum vaginatum (Spreng.) Thell., and so on [5, 7,8,9,10].

Studies on Apioideae systematics have contributed to the circumscription of the Sinodielsia clade in principle, except that several of its members seem to be controversial and indeterminate in systematic positions based on different datasets: Peucedanum delavayi Franch., Ligusticum pteridophyllum Franch. and Seselopsis tianschanica Schischkin [5, 8,9,10,11,12,13]. Therefore, it is necessary and useful to conduct a comprehensive analysis of the phylogeny of the clade and its relatives to confirm the interspecific relationships of the clade and obtain a better understanding of the evolution of the clade and subfamily.

Chloroplast genomes are characterized by a highly conserved structure, fewer gene arrangements, and relatively coincident gene contents among plant species [14]. Furthermore, because developments in next-generation sequencing (NGS) and improvements in algorithms have decreased the cost of data acquisition and complexity of the cp. genome assembly, complete cp. genome sequences have been comprehensively accepted as valuable and informative data sources for comprehensive evolutionary investigation and have become a highly useful tool to assess the phylogenetic relationships of puzzling groups of angiosperms [15,16,17]. More cases also confirmed the advantage of the cp. genome in phylogenetic studies at different taxonomic levels. Huang et al. analyzed the cp. genome of Salix L. and reconstructed a phylogenetic tree based on the whole cp. genome and common protein-coding sequences of the genus, and the results showed that the genus was monophyletic with high support and was subdivided into two subclades [18]. In Urticeae, cp. genome data proved the monophyly of most genera and provided new insights into the phylogenetic relationship and chloroplast structure evolution [19]. Zhou et al. reconstructed the phylogenetic relationship of Bambusoideae based on cp. genome data and further estimated the divergence time and ancestral distribution, which showed that Cephalostachyum Munro and Schizostachyum Nees were homologous, and they originated from the early Miocene Eastern Himalayas to northern Myanmar [20].

The taxonomic history of the Sinodielsia clade is complex, and its members were enrolled from 17 genera. In phylogenetic studies of these 17 genera, some species were far from other members of genera in phylogeny and clustered in the Sinodielsia clade, so these species were adjusted into this clade in the classification system. In this context, we supposed that some members of the Sinodielsia clade were inappropriately transferred to this clade because the sampling coverage was not sufficient to detect their proper positions in relevant research. In this study, we sampled 105 cp. genome sequences of 95 species from the Sinodielsia clade and its relevant genera and reconstructed their phylogeny based on cp. genome sequences to define the circumscription of the Sinodielsia clade and infer the evolutionary history of the clade. This study will promote the comprehension of cp. genome evolution, taxonomy and phylogenetics of Apioideae.

Results

Yong Wang de novo assembled cp. genomes

Clean reads ranging from 0.81 GB (Lig. oliverianum) to 2.75 GB (A. longicaudata) of 39 species were extracted from raw reads obtained by the Illumina HiSeq 4,000 system (Table 2). Complete cp. genomes were de novo assembled successfully from clean reads and further manually verified to prevent potential assembly errors. Complete cp. genomes of 39 species ranged from 145,335 bp (Lig. yushuense) to 165,147 bp (Ple. foetens) in length (Table 2). All of them had a highly conserved typical quadripartite structure with two inverted repeat (IR) regions (17,599–31,486 bp), a large single copy (LSC) region (82,048–94,046 bp) and a small single copy (SSC) region (16,343–17,917 bp) (Fig. 1; Table 2). The total GC contents were between 37.4 and 38.0% (Table 2). The total numbers of genes of these 39 complete cp. genomes ranged from 126 (Ple. hookeri) to 138 (Ple. foetens, T. tianschanicum). In detail, these genes included 83–93 protein–coding DNA sequence (CDS) genes, 35–37 transfer RNA (tRNA) and eight ribosomal RNA (rRNA) genes (Table 2). The organization and CDS order (Table S1) of these cp. genomes were highly identical and similar to those of other plants in Apioideae [21,22,23].

Fig. 1
figure 1

Gene map of cp. genomes of nine Sinodielsia Clade relevant genera (the length of each genus was displayed inside). Transcribed clockwise genes are shown outside, while counterclockwise genes are inside. Different functional groups of genes were identified by different colors. The darker gray represents the GC content and the value was displayed inside

Full size image

We detected 2,734 simple sequence repeats (SSRs) among the 39 cp. genomes. Most were mononucleotide repeats (58%), followed by dinucleotides (25%), trinucleotides (4%), tetranucleotides (10%), pentanucleotides (2%) and hexanucleotides (1%) (Fig. 2A). For each genome, the total numbers of SSRs ranged from 44 (Ple. franchetianum) to 94 (Lig. ajanense) (Fig. 2B). More than half of the SSRs (50.0–69.7%) were mononucleotide repeats in species with newly assembled cp. genomes, except Ses. tianschanica (46.7%). (Fig. 2C, Table S2).

Fig. 2
figure 2

Analysis of simple sequence repeats of 39 species cp. genomes. (A) Proportion of different SSRs types; (B) Total number of SSRs of 39 species; (C) Number of SSRs with different types in 39 species

Full size image
Table 1 Members of the Sinodielsia Clade recorded in references
Full size table

Phylogenetic relationships of 105 genomes related to the Sinodielsia clade

Phylogenetic analyses produced two trees identical in topology based on whole cp. genome and CDS datasets. Our results showed that 100 genomes were clustered into eight clades (Pleurospermeae, East–Asia clade, Komarovieae, Acronema clade, Cachrys clade, Sinodielsia clade, Tordyliinae, and Selineae), while the other five contained two novel clades, the clade of Lig. pteridophyllum and Ses. tianschanica and the clade of Ple. uralense, Lig. discolor and T. tianschanicum, respectively (Fig. 3).

Table 2 Characteristics of the 39 newly assembled cp. genomes
Full size table
Fig. 3
figure 3

Phylogenetic tree inferred from Maximum-Likelihood based on CDSs and whole cp. genomes of 105 sequences. The numbers were listed at each node represent the bootstrap support (BS) values. The red font labels indicate that the species belonged to Sinodielsia Clade in previous studies, and the yellow represents species that were first clustered into Sinodielsia Clade

Full size image

Nineteen species from 10 genera were clustered into the Sinodielsia clade. It was close to Selinese and Tordyliinae (BS = 83/91) in topology, which was in line with previous studies of Apioideae based on both ITS and cp. genome data [5, 6, 10]. The clade was divided into two subclades with strong bootstrap values. One subclade included A. sinensis (8 individuals), Cni. officinale, Lig. sinense, Con. chinense, Lig. chuanxiong, Lig. jeholense, A. omeiensis and Lig. tenuissimum (Fig. 3, subclade I, BS = 99/100), and another subclade comprised 12 other species, namely, A. ternata, A. paeoniifolia, A. multicaulis, (A) sinensis (1 individual), Ple. rivulorum, H. apiolens, L. yushuense J. T. Pan, Ple. hookeri C. (B) Clarke, Meeboldia yunnanensis (H. Wolff) Constance & F. T. Pu, V. thibetica, T. subnudum and Lev. officinale (Fig. 3, subclade II, BS = 93/100).

Comparing the IR boundaries of the Sinodielsia clade

Chloroplast genomes were highly conserved in structure and size, while the change in the location of the IR/SC junction was due to the universally existing expansion and shrinkage of the IR regions [24, 25]. Comparison of IR boundaries among 27 genomes of the Sinodielsia clade displayed diverse expansion and contraction of the IR regions (Fig. 4). The junction site of LSC/IRb (JLB) was located in the ycf2 gene in 18 out of 27 genomes and extended to the rpl22 gene in the Lig. tenuissimum genome. For eight A. sinensis genomes (A. sinensis1–A. sinensis8), it departed 1–44 bp from the trnV gene. Among these 27 genomes, all the junction sites of IRb/SSC (JSB) were close (within 33–162 bp) to or located in the ndhF gene. The junction sites of SSC/IRa (JSA) were located in the ycf1 gene in the 27 genomes. The junction sites of IRa/LSC (JLA) were near the trnH gene (7–1664 bp) in 23 genomes except the cp. genomes from four individuals of A. sinensis (A. sinensis3, A. sinensis4, A. sinensis7, A. sinensis8), whose boundaries extended 0–32 bp into the psbA gene.

Fig. 4
figure 4

Comparison of LSC, SSC, IRs region boundaries of 27 genomes within the Sinodielsia Clade. The figure is not drawn to scale

Full size image

Comparative genomic analysis in the Sinodielsia clade

Using the mVISTA program and referring to A. sinensis8 (MK688991), the sequence identity analysis revealed more sequence mutations in noncoding regions than in coding regions in the Sinodielsia clade (Fig. 5). Furthermore, sliding window analysis showed that the nucleotide diversity (Pi) values of these 27 genomes ranged from 0 to 0.02576 (Fig. 6). The average Pi value was 0.00476 in the LSC regions and 0.00706 in the SSC regions. In contrast, the average Pi values of the IR regions were the lowest (0.00154). We observed six mutation hotspots (highly variable regions) with Pi values over 0.01400, including five noncoding regions and one gene region (Fig. 6). Of them, three noncoding regions were located in the LSC, and they were rbcL–accD (Pi, 0.01456–0.01563), ycf4–cemA (Pi, 0.01570–0.02514) and petA–psbJ (Pi, 0.01538–0.02576). The SSC region contributed the other three mutation hotspots, including two noncoding regions and a gene, which were ycf1–ndhF (Pi, 0.01985–0.02127), ndhF–rpl32 (Pi, 0.01444–0.01632) and ycf1 (Pi, 0.01418), respectively. Furthermore, ndhF–rpl32 and ycf1 were highly variable in all 105 cp. genomes (Fig. S1).

Fig. 5
figure 5

Twenty-seven sequences alignment was performed by mVISTA using Angelica sinensis8 as a reference. The vertical scale represents the percentage of identity, ranging from 50 to 100%

Full size image
Fig. 6
figure 6

The DnaSP graph of nucleotide diversity (Pi) value of the 27 cp. genomes in Sinodielsia Clade

Full size image

Comparative analysis of cp. genome sequences using the Mauve alignment approach showed that the genome structures of the Sinodielsia clade species were conservative, and no potential rearrangement or change was detected in gene order (Fig. 7).

Fig. 7
figure 7

Mauve alignment of the complete cp. genome of 27 Sinodielsia Clade species. The strip structure with the same color in the figure is a local collinear block, representing a set of homologous genes. The strip area below the horizontal line of each genome indicates that inversion has occurred

Full size image

Discussion

Genome features

A total of 105 complete cp. genomes from 16 genera were different in size, ranging from 140,670 bp (A. sinensis) to 165,147 (Ple. foetens) (Table S3), which showed that the cp. genomes in these species had distinct characteristics. Nevertheless, species within the Sinodielsia clade had a moderate-length genome, and their lengths ranged from 145,335 bp to 148,653 bp, except for the minimum length of A. sinensis (8 individuals, 140,670 bp–142,822 bp) and the maximum length of Lig. tenuissimum (158,500 bp). Notably, Pleurospermum had a large average cp. genome size of 154,687 bp, but two species of this genus, Ple. rivulorum and Ple. hookeri, were clustered in the Sinodielsia clade, and the cp. genomes of both species were relatively short in size (146,815 bp and 145,400 bp, respectively).

All 105 cp. genomes had a typical quadripartite structure and were conserved in gene order, similar to other angiosperm cp. genomes [14]. However, many genes of the cp. genome have been lost in different plants [26,27,28,29,30], such as accD, ycf1, infA, clpP, ccsA, rps12, rps16, and rpl23. In this study, the gene number of 105 cp. genomes ranged from 121 to 144, and rps12, rps16, ycf15 and ycf1 were frequently missing from these cp. genomes (Table S1 and Table S3). The phenomenon of these gene losses in Apioideae was probably a result of extensive hybridization and/or cp. genome decay within various lineages [28, 31].

Circumscription and phylogeny of the Sinodielsia clade

In the present study, 105 taxa were sampled from 16 genera related to the Sinodielsia clade, and their phylogeny was constructed based on whole cp. genome sequences and CDSs. We sampled not only species of the Sinodielsia clade but also extensive relevant taxa, which covered eight out of 30 major clades of Apioideae (e.g., Pleurospermeae, East–Asia clade, Komarovieae, Acronema clade, Cachrys clade, Sinodielsia clade, Tordyliinae, and Selineae). The relationships of eight clades inferred based on cp. genome data were consistent with those in previous studies [5, 6, 10, 12], except for two novel clades, the clade of Lig. pteridophyllum and Ses. tianschanica, and the clade of Ple. uralense, Lig. discolor and T. tianschanicum.

Our results confirmed that the Sinodielsia clade was the sister to the cluster of Selineae and Tordyliinae, but the circumscription of the Sinodielsia clade was different from the results of previous studies [5, 7,8,9,10]. Three species of the Sinodielsia clade, Lig. yushuense, Ple. hookeri and M. yunnanensis, were first clustered into this clade, and it seemed that they were new members of the Sinodielsia clade. Nevertheless, we found that seven species formerly accommodated in the Sinodielsia clade were excluded from this clade. Lig. pteridophyllum and Ses. tianschanica formed an independent clade departing from the Sinodielsia clade, and it was the sister to the large cluster of three clades, Selineae, Tordyliinae and Sinodielsia. A. tianmuensis was clustered in the Angelica group. Peu. delavayi was nested in the Acronema clade, and the other three species were in Selineae, Cni. dauricum, Lig. acuminatum and Sil. silaus.

The Sinodielsia clade was established, and its members were enrolled based on ITS, but the positions of the aforementioned seven species were not supported by cp. data. Conflicts between results from cp. DNA and nrDNA data probably reflected complementary processes of speciation in diverse inheritance patterns [32]. Additionally, some genetic events might contribute to these conflicts in phylogeny, including incomplete lineage sorting, hybridization/introgression, paralogy, gene duplication and/or loss, and horizontal gene transfer [33, 34]. Positions of Peu. delavayi, Lig. pteridophyllum and Ses. tianschanica were unstable in different studies when they were investigated together with different species and genera based on ITS [5, 7, 8, 13], while they were not clustered in the Sinodielsia clade on the phylogenetic trees based on cp. genome data. We speculated that the lack of relatives resulted in clustering of these species in the Sinodielsia clade in studies based on ITS, and conflicts between results from cp. DNA and nrDNA could be attributed to low sampling coverage in previous studies. We have no direct evidence to clarify this phenomenon in A. tianmuensis, Cni. dauricum, Lig. acuminatum and Sil. silaus, but we preferred to attribute the conflicts to hybridization or introgression after comparing them with their relatives in terms of morphology and distribution.

Except for cultivated and introduced species, all species in the Sinodielsia clade gathered into two subclades relevant to geographical distributions. Species of subclade I were widely distributed in Asia and Europe, except A. omeiensis in Emei Mountain, Sichuan Province. A. omeiensis is a poorly known specieswith reputed medicinal properties, which is recorded only from a few collections. Recent research suggests that it is conspecific with A. wilsonii and A. sinensis var. wilsonii [35]. Species of subclade II were distributed in Western China except for the introduced Lev. officinale and one individual of cultivated A. sinensis. A. ternata and A. multicaulis were distributed in Xinjiang Province, and Lig. yushuense was in Qinghai Province. The other seven species were endemic to the Himalaya region, A. paeoniifolia, Ple. rivulorum, H. apiolens, Ple. hookeri, M. yunnanensis, V. thibetica and T. subnudum.

Angelica sinensis is a famous and widely used Chinese traditional medicine herb, and it has been cultivated for more than 1,000 years. We sampled nine individuals of A. sinensis from different plant areas, but eight of them were clustered in subclade I and one in subclade II. This result seems provide a hint of its multiple domestication, and this study pointed out further direction on the investigation of the cultivation origin of A. sinensis.

Potential markers for molecular identification and phylogeny of apiaceae at low taxonomic levels

Higher Pi values indicated more mutations and higher evolutionary rates in highly variable regions than in other regions [36]. Multiple variable regions have been identified in angiosperms [37,38,39]. Unfortunately, these regions often lack variations in closely related species in Apiaceae, especially those diverged recently in evolutionary history, so that there are very few choices in chloroplast DNA markers suitable for phylogeny of Apiaceae at low taxonomic levels, and research on molecular systematics of the family relies heavily on the use of ITS [5, 8, 11,12,13]. In the present study, we scanned the whole cp. genomes of the Sinodielsia clade and detected six mutation hotspot regions in both noncoding regions and CDS, rbcL–accD, ycf4–cemA, petA–psbJ, ycf1–ndhF, ndhF–rpl32 and ycf1. Among the six highly variable regions, high variability of ndhF–rpl32 and ycf1 was also observed in sliding window analysis of 105 whole cp. genomes. These six regions, especially ndhF–rpl32 and ycf1, were highly variable and should be the first consideration as screening suitable loci to distinguish closely related species or genera in identification and phylogenetic analyses of the Sinodielsia clade, even Apioideae.

Conclusion

The Sinodielsia clade has been an incomprehensible group of Apiaceae in terms of taxonomy, and its members are tangled with 16 genera from different clades in morphology. In this study, we assembled complete cp. genomes of 39 species relative to the Sinodielsia clade and scanned genome characteristics in terms of genome size, GC content, SC/IR boundaries, gene number, repeat types and distribution. Then, we performed phylogenetic analysis based on 105 cp. genome sequences from 16 genera relative to the Sinodielsia clade. The phylogenetic analysis showed that 19 species were clustered into the Sinodielsia clade, and the clade was subdivided into two subclades relevant to geographical distributions, except cultivated and introduced species. Six mutation hotspot regions were detected from the whole cp. genomes among the Sinodielsia clade; namely, rbcL–accD, ycf4–cemA, petA–psbJ, ycf1–ndhF, ndhF–rpl32 and ycf1, and ndhF–rpl32 and ycf1 were highly variable in the 105 sampled cp. genomes. These mutation hotspot regions could be used as potential DNA markers in identification and phylogenetic analyses of the Sinodielsia clade and Apioideae. Our study provided new insights into the phylogeny of the Sinodielsia clade and provided valuable information on cp. genome evolution in Apioideae.

Methods

Taxon sampling

We sequenced and assembled the cp. genomes of 39 species from nine genera relevant to the Sinodielsia clade. Samples were collected from the National Wild Plant Germplasm Resource Center and Herbarium of Kunming Institute of Botany, and we used them following the prescribed procedures of the Kunming Institute of Botany. Vouchers were preserved in the Herbarium of Kunming Institute of Botany (KUN), and Shaotian Chen reviewed and identified the vouchers. To infer the phylogeny of the Sinodielsia clade, we downloaded published cp. genome sequences of 66 species related to the Sinodielsia clade from GenBank (https://www.ncbi.nlm.nih.gov/genbank/). In total, 105 sequences from 23 species from 12 genera were members of the Sinodielsia clade. Details of 39 newly sequenced samples and 66 published sequences are shown in the supporting information (Table S4 and Table S5).

DNA extraction and cp. genome sequencing

Total genomic DNA was extracted following a modified CTAB protocol [40]. Genomic DNA from each sample was subsequently assessed for quality using both a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, United States) and agarose gel electrophoresis before library preparation. The libraries were generated using the NEBNext Ultra II DNA Library Prep Kit for Illumina following the manufacturer’s instructions. Sequencing was performed on the Illumina HiSeq 4000 platform with 150 bp paired–end reads. The obtained raw reads were adapter–trimmed and quality–filtered by AdapterRemoval v2 (trimwindows = 5 and minlength = 50) [41]. Clean reads were deposited in GSA (Genome Sequence Archive, https://ngdc.cncb.ac.cn/gsa/, Accession No.: CRA007981, CRA006303).

Chloroplast genome assembly and annotation

Clean reads were qualitatively evaluated and assembled using GetOrganelle version 1.7.4 [42]. Assembled circular complete cp. genomes were checked and aligned with the reference to complete cp. genomes of Lig. sinense (MN652884) and A. sinensis (MW820164) using Geneious version 2022.0.1 [43]. The online program GeSeq (https://chlorobox.mpimp–golm.mpg.de/geseq.html) was used to annotate the complete cp. genomes [44]. The annotated genomes were further examined using Geneious version 2022.0.1 to prevent potential annotation errors. The online program OGDRAW (https://chlorobox.mpimp–golm.mpg.de/OGDraw.html) was used to plot the gene maps of complete cp. genomes [45]. Annotated cp. genomes were deposited in GenBank (Accession No.: OP672440–OP672478, Table 2).

Simple sequence repeat analysis

Simple sequence repeats (SSRs) were searched using the online program MISA [46] (https://webblast.ipk–gatersleben.de/misa/). The program parameters were set as follows: minimum numbers of repetitions for mononucleotide SSRs, dinucleotide repeat SSRs, trinucleotide repeat SSRs, tetra–, penta–, and hexanucleotide repeat SSRs were 10, 5, 4, 3, 3 and 3, respectively.

Sequence characteristics analysis

Geneious version 2022.0.1 was used to count genome sizes, GC contents, LSC/SSC/IR lengths and gene numbers [43]. The IR/SC boundaries of the Sinodielsia clade species were compared to describe IR expansion and contraction. Whole cp. genome alignment was performed and visualized by the online tool mVISTA [47] (http://www.gsd.lbl.gov/vista/mvista/). Nucleotide divergence values were computed by DnaSP version 6.12.03 [48]. The parameters of the sliding window method were set to a step size of 200 bp and a window length of 600 bp. Comparative analysis of cp. genome structure and gene rearrangements was performed by Geneious version 2022.0.1 using plague Mauve alignment [49].

Sequence alignment and phylogenetic analysis

Phylogenetic trees were constructed based on two datasets of CDS and whole cp. genome sequences of 105 complete cp. genomes, and two species (Sanicula chinensis Bunge, Sa. odorata (Raf.) K.M. Pryer et L.R. Phillippe) were assigned as outgroups to root trees. To avoid calculating the same information twice, the CDSs in the second inverted repeat region were eliminated from the CDS dataset. Both datasets were aligned by MAFFT version 7.490 [50] and adjusted automatically using TBtools version 1.09876 [51]. Maximum likelihood (ML) trees were constructed for each of the two datasets using IQ–tree version 2.2.0 under a GTR + I + G + F4 model with 1000 bootstrap replicates [52]. The ML tree file was imported into MEGA11 to view and edit the tree, and the vector graph of output trees was saved to the file [53].