Introduction

Sanicula L. (Apiaceae-Saniculoideae), consists of ca. 45 taxa, is widely distributed from East Asia to North America [1, 2]. China is one of the most important distributed centers, with ca. 18 species and two varieties [3,4,5]. It was known as the considerably complex taxonomic genus, with its varied morphological characters in rhizomes, leaves, inflorescences and fruits, placed comparatively primitive within Subfam. Saniculoideae Burnett in primitive of Apiaceae [6,7,8,9,10]. Traditionally, based on the features of leaves, flower and fruits, Shan [6] divided the species of world Sanicula into five sections, i.e. Tuberculatae, Pseudopetagnia, Sanicla, Sandwicenses and Sanicoria, and demonstrated that the Chinese Sanicula taxa belonged to the former three sections. A classification was accepted by many later authors [3, 4, 11].

Sanicula had consistently been viewed as a relatively natural genus within the family Apiaceae [12,13,14]. Molecular phylogenetic analyses had also suggested that the genus was a monophyletic group yet based on few Sanicula samples by using the nuclear ribosomal internal transcribed spacer (ITS) region and chloroplast DNA (cpDNA) rpl16 intron, rpoC1 intron, trnQ-rps16 and rps16-trnK intergenic spacers[8, 13,14,15,16,17]. Then, a revised phylogeny of Apioideae and Saniculoideae in Apiaceae based on the 90 whole plastome sequences, including only four Chinese Sanicula species, suggested that sectional relationships in Sanicula were distinct from the traditional classification system [2].

Furthermore, based on recent wild observations in eastern, southern, and western China, the interspecific relationships of some groups in the genus were extremely perplexing [9, 10]. It was also mentioned by many authors, including Chen et al. [14, 17], Shan & Constance [6], and Yang et al. [2]. In addition, numerous species and varieties in Sanicula were poorly defined due to a lack of field studies and consistent characteristics for diagnostic methods in the literature [2, 6, 14, 17]. Therefore, further exploration into more stable genetic variations and effective markers is critical for utilizing and protecting the Sanicula plants.

Chloroplast (cp) genomes were highly conserved in terms of the genetic replication mechanisms in uniparental inheritance and possess the relatively high level of genetic variation resulting from the low selective pressure [18]. In addition to its low sequencing cost caused by rapid development of illumine and assembly technologies, the cp genome had been relatively more successful than fragments in resolving the relationship between many species at different taxonomic levels in many species [3, 19].

The plastomes of six Chinese Sanicula species were reported previously, including S. astrantiifolia H. Wolff ex Kretschmer, S. chinensis Bunge, S. flavovirens Z. H. Chen, D. D. Ma & W. Y. Xie, S. giraldii R. H. Shan & S. L. Liou, S. lamelligera Hance, S. orthacantha S. Moore and S. rubriflora F. Schmidt [2, 17, 20, 21]. However, it seemed that the samples of S. chinensis and S. orthacantha used in previous study were likely mixed up due to their same voucher information. Additionally, few Sanicula species had been involved in molecular studies [2, 22] and fewer effective markers were discovered to deal with their inter- and intra- specific relationships.

The aim of this study was to 1) determine the whole plastome sequence of 13 Chinense Sanicula taxa, including the four newly sequenced taxa; 2) compare the global structural patterns of available Chinese Sanicula cp genomes; 3) examine variations in the SSRs and repeat sequences among 13 Sanicula cp genomes; 4) to reconstruct the phylogeny of Chinese Sanicula taxa, and improve the understanding of the relationship and evolution in Chinese Sanicula taxa.

Results

Chloroplast genome structures of four taxa in Chinenese Sanicula L. and one species in Eryngium L

All five new Sanicula cp genomes (Table 1) were similar to other species of Sanicula or other genera in Apiaceae [17]. The size of four new cp genome in Sanicula ranged from 155,396 bp in S. orthacantha var. brevispina to 155,757 bp in S. caerulescens, exhibiting a typical quadripartite structure with two single copy regions (LSC and SSC) which were separated by a pair of inverted repeats (IRa and IRb) (Fig. 1). The length of the large single-copy (LSC) ranged from 85,818 bp (S. orthacantha var. brevispina) to 86,209 bp (S. caerulescens), the small single-copy (SSC) ranged from 17,089 bp (S. hacquetiodes) to 17,106 bp (S. tienmuensis), and IR regions ranged from 26,225 bp (S. caerulescens) to 26,332 bp (S. hacquetiodes) (Table 1). The overall GC content ranged from 38.16% (S. caerulescens) to 38.21% (S. hacquetiodes).

Table 1 Summary of chloroplast genome features in this study, including four new chloroplast genomes of the Sanicula taxa and one newly in Eryngium, a showing the new chloroplast genome reported in this study
Full size table
Fig. 1
figure 1

A circular gene map of four newly sequenced Sanicula chloroplast genomes. Genes shown outside are transcribed clockwise, and inside the circle are transcribed counterclockwise. Genes are color-coded to distinguish different functional groups. The dark grey and the light grey plots in the inner circle correspond to the GC content and AT content, respectively

Full size image

All four newly sequenced Sanicula cp genomes here encoded 103 unique genes, including 79 unique protein-coding genes (PCGs), 20 unique tRNA genes and four unique rRNA genes, and 23 of these were duplicated, with a total of 126 genes (Table 1; * showing the new chloroplast genomes reported in this study). 13 genes contain one (atpF, ndhA, ndhB, petB, rpl16, rpl2, rpoC1, rps16, trnA-UGC, trnI-GAU) or two (clpP1, rps12, ycf3) introns, and two of these were tRNA genes (Table 2, Fig. 1). The cp genome contained coding regions ranging from 55.92% to 56.07% and non-coding regions ranging from 43.93% to 44.08%, including both intergenic spacers and introns (Table 2). They were divided into four categories, consisting of photosynthesis, self-replication, other genes, and function unknown genes (Table 2). The length of Eryngium foetidum L. was 155,270 bp, consisting of a LSC region of 85,874 bp, an SSC region of 17,074 bp, and a pair of inverted repeats region of 26,161 bp (Fig. 2). The overall GC content was 38.13%. It contained 127 genes, including 86 PCGs, 33 tRNA genes and 8 rRNA genes (Table 1; (a) showing the new chloroplast genomes reported in this study), and divided into four categories, consisting of photosynthesis, self-replication, other genes, and function unknown genes (Table 2).

Table 2 List of annotated genes in the chloroplast genomes of four newly sequenced Sanicula taxa and one sample of Eryngium foetidum
Full size table
Fig. 2
figure 2

A circular gene map of one newly sequenced chloroplast genomes of Eryngium foetidum. Genes shown outside are transcribed clockwise, and inside the circle are transcribed counterclockwise. Genes are color-coded to distinguish different functional groups. The dark grey and the light grey plots in the inner circle correspond to the GC content and AT content, respectively

Full size image

Inverted repeats expansion, contraction, and interspecific comparison

In total, we analyzed and compared 15 cp genomes’ IR/LSC and IR/SSC boundary structures (including four Sanicula and one Eryngium samples from GenBank, four newly sequenced chloroplast genomes of Sanicula and one newly of Eryngium; Fig. 3, * showing the new chloroplast genomes reported in this study). The IRb/LSC boundary was located within the rps19 gene (with the 5′ end of the rps19 located in the IRb region while 3′ end located in the LSC), except in S. flavovirens sample (NC_061752), with an expansion length of 55 or 58 bp. The IRa/SSC boundary was in the ycf1 gene (the 5′ end of the ycf1 located in the IRa region while the 3′ end located in the SSC), with spanned 1122–1872 bp in the IRa region. The IRb/SSC boundary obviously varied: three samples were located within ndhF, with expanded 1–34 bp to the IRb region, while other 12 samples with 5 or 6 bp away from the IRb/SSC boundary.

Fig. 3
figure 3

Comparison of the SC/IR junctions among the 15 chloroplast genomes, including 13 Sanicula and two Eryngium chloroplast genomes. JLA indicates LSC/IRa boundary; JSA indicates SSC/IRa boundary; JSBindicates SSC/IRb boundary; JLB indicates LSC/IRb boundary 

Full size image

The mVISTA result showed that the non-coding regions were more variable than the coding regions, the LSC and SSC regions had higher level of sequence divergence than the two IR regions, and intergenic spacers (IGS) regions were the most divergent regions (Fig. 4). The highly divergent regions among the 13 chloroplast genomes occurred in 19 of the intergenic spacers, 17 in the LSC regions, including atpH-atpI, ndhC-trnM, petB-petD, petD-rpoA, petN-psbM, psaJ-rpl33, rbcL-accD, rpoB-trnC, rps16-trnQ, trnE-psbD, trnF-ndhJ, trnH-psbA, trnS-psbZ, trnS-trnR, trnT-trnF, ycf3-trnS, ycf4-cemA; and one in the boundary between IRa and SSC region: trnN-ndhF; one in IR regions: trnV-rps12. Apart from these regions, one coding region ycf1 also showed high sequence variation (Fig. 4).

Fig. 4
figure 4

Plots of percent sequence identity of the chloroplast genomes of 12 Sanicula taxa with S. orthacantha var. stonifera (NCBI accession no. MT561028) as a reference

Full size image

The value of nucleotide diversity (Pi) ranged from 0 to 0.01658, with average value of 0.003326 among the whole chloroplast (Fig. 5; Additional File 1: Table S1). The IR region were observed to have lower Pi value than LSC and SSC regions. The LSC region showed the highest nucleotide diversity (Pi = 0.01658), while the lowest Pi is in the IR regions (Pi = 0). 12 hypervariable sites with Pi more than 0.01 in LSC regions were screened (Fig. 5), namely cemA-petA (Pi = 0.01009), ndhJ-ndhK (Pi = 0.01124), petA-psbJ (Pi = 0.01059), petD-rpoA (Pi = 0.01436), petE-psbL (Pi = 0.01145), petN-psbM (Pi = 0.01265), psbZ-trnG (Pi = 0.01103), rpoB-trnC (Pi = 0.01658), trnH-psbA (Pi = 0.0145), trnR-atpA (Pi = 0.01175), trnS-trnR (Pi = 0.01551), ycf3-trnS (Pi = 0.01047). Two hypervariable sites, rps15-ycf1 (Pi = 0.01128) and ycf1 (Pi = 0.01389), with high Pi value more than 0.01 in SSC regions were also screened in Fig. 5.

Fig. 5
figure 5

The nucleotide diversity of the whole chloroplast genomes of the 13 Sanicula taxa. LSC indicates large single copy region, IR indicates inverted repeat region, SSC indicates small single copy region

Full size image

Repeat structures and simple sequence repeats

The characteristics of Simple sequence repeats (SSRs) in four newly sequenced Sanicula cp genomes (S. caerulescens, S. hacquetiodes, S. orthacantha var. brevispina and S. tienmuensis) were analyzed, and the patterns of SSRs distribution were shown in Additional File 2: Table S2; Fig. 6A, B. A total of 40, 39, 38 and 35 SSRs loci were detected in these four newly sequenced Sanicula cp genome, respectively. The most abundant SSRs were A or T nucleotide repeats, which accounted for 28.95% to 35% of the total. SSRs were mainly distributed in LSC regions (76.47%–78.13%), and were significantly lower in the SSC (11.11%–13.89%) and IR (8.33%–11.76%) regions. Furthermore, they were only having mono- and di-nucleotide repeats. Among them, mono-nucleotide repeats were the most common SSR, accounting for 77.5%, 71.79%, 73.68% and 74.29% respectively, followed by 22.5%, 28.21%, 26.32%, and 25.71% in di-nucleotide repeats.

Fig. 6
figure 6

Statistics of repeats in four newly sequenced Sanicula taxa samples. A. Number of SSRs distributed in LSC, SSC and IR regions. B. Number of SSRs types. C. Number of four types SDRs. D. Number of different lengths of SDRs

Full size image

The REPuter screening discovered 42 to 68 dispersed repeats of 30 bp or longer among the four newly sequenced Sanicula cp genomes examined (Additional File 3: Table S3; Fig. 6C, D). The number of categories and the total number in repeats of S. tienmuensis (4; 68) were higher than S. caerulescens (3; 42), S. hacquetiodes (2; 44), and S. orthacantha var. brevispina (3; 46) (Fig. 6C). Only one, two and one reverse repeats were found in S. caerulescens, S. tienmuensis and S. orthacantha var. brevispina, respectively, while no reverse repeats was discovered in S. hacquetiodes. The complement repeat accounted for one only in S. tienmuensis (Fig. 6C). Among these four newly sequenced Sanicula cp genomes, the number of repeats with length between 30–40 bp exceeded those with lengths of  41–50 bp, 51–60 bp, 61–70 bp and over 70 bp (Fig. 6D).

Statistics of codon usage

According to the codon usage analysis, the total sequence sizes of the PCGs were 67,857–67,863 bp in the four newly sequenced Sanicula taxa genomes; 22,673–22,690 codons were encoded (Additional File 4: Table S4). Leucine encoded with the maximum number of codons ranged from 2382 to 2390, followed by isoleucine, with the number of codons ranged from 1909 to 1918. Cysteine was the least with 237–239. The relative synonymous codon usage (RSCU) values varied slightly among the four newly sequenced Sanicula genomes (Fig. 7). Thirty-two codons were used frequently with RSCU ≥ 1 and 34 codons used less frequently with RSCU < 1. AUG showed a preference in all the four cp genomes. The frequency of use for the codon UGG, encoding the tryptophan (Trp), showed no bias (RSCU = 1).

Fig. 7
figure 7

Codon content of 20 amino acids and stop codons in Sanicula caerulescens (a), S. hacquetiodes (b), S. orthacantha var. brevispina (c) and S. tienmuensis (d)

Full size image

Phylogenetic analysis

A total of three datasets, including the whole cp genomes sequences (Additional File 5: Fig. S1A), concatenation of 126 unique IGS regions (Additional File 5: Fig. S1B), concatenation of the unique 79 unique PCGs regions (Fig. 8) were constructed to investigate the phylogenetic relationships among 13 Sanicula taxa, with Eryngium planum L. and E. foetidum as outgroup taxa. By using the maximum likelihood (ML) method, three phylogenetic trees were built based on the three respective datasets, which exhibited highly concordant between one another. Therefore, only the ML topology of concatenation of the 79 unique PCGs regions among 13 Sanicula taxa, which also by using the Bayesian inference (BI) and Maximum Parsimony (MP) analyses, were shown here with the ML/MP/BI support [bootstrap support (bs) / bs / posterior probability (pp)] values added at each node with only slight differences (Fig. 8).

Fig. 8
figure 8

Phylogenetic tree based on complete cp genomes resulting from ML, MP and BI analysis of 13 Sanicula samples and two Eryngium species as references based on concatenation of 79 unique coding genes. The bootstrap support values and posterior probability values are displayed on the branches in the order ML/MP/BI, and values less than 50/50/0.5 are not shown

Full size image

Our analyses confirmed that the genus Sanicula was monophyletic with strongly supported. Two major clades (clade I and II) were resolved within the monophyletic genus. The clade I comprised two fully supported subclades (A and B). Subclade A was consistent with Sect. Pseudopetagnia Wolff. Subclade B contained two species belonging to two different sections in the genus Sanicula, namely sect. Sanicla DC. and Sect. Tuberculatae Drude. Clade II divided into two subclades (C and D) with fully supply supported. Subclade C included three samples representing S. flavovirens and S. chinensis, respectively. However, they belonged to two different sections, i.e., Sect. Tuberculatae and Sect. Sanicla. Subclade D contained two samples representing S. rubriflora in Sect. Tuberculatae. These results showed that Sect. Tuberculatae and Sect. Sanicla were not natural monophyletic sections.

A total of 20 highly divergent regions (atpH-atpI, ndhC-trnM, petB-petD, petD- rpoA, petN-psbM, psaJ-rpl33, rbcL-accD, rpoB-trnC, rps16-trnQ, trnE-psbD, trnF-ndhJ, trnH-psbA, trnN-ndhF, trnS-psbZ, trnS-trnR, trnT-trnF, trnV-rps12, ycf3-trnS, ycf4-cemA and ycf1) and concatenation of 20 highly divergent regions were also evaluated for phylogenetic analysis in our study (Additional Files 6, 7, 8, 9, 10, 11: Figs. S2–S7). Three molecular fragments, including trnE-psbD (Additional File 8: Fig. S4B), trnS-trnR (Additional File 9: Fig. S5C) and the concatenation regions (Additional File 11: Fig. S7), yielded similar topological results. However, compared to the three topological trees constructed by the whole cp genomes sequences (Additional File 5: Fig. S1A), concatenation of 126 unique IGS regions (Additional File 5: Fig. S1B), concatenation of 79 unique PCGs regions (Fig. 8), the supporting values were different observed from the nodes based on different sequences dataset. For example, the nodes in clades I derived from the dataset of trnE-psbD and trnS-trnR both showed strong supports (bs = 89.9% and bs = 85%; Additional Files 8, 9: Figs. S4B, S5C) lower than those from concatenation of 79 unique PCGs regions, whole cp genomes and concatenation of 126 unique IGS regions (bs = 100%, 100%, 99.4%; Fig. 8; Additional File 5: Fig. S1A, B). Additionally, the concatenation of 20 highly divergent regions had nodes strong supports in clades I and II (Additional File 11: Fig. S7). These results indicated a well resolution of the whole complete cp genomes, concatenations of PCGs regions and IGS regions as well as the concatenations of highly divergent regions compared to the single divergent region, which may serve as a reliable proof to reconstruct the phylogenetic relationship in Sanicula.

Discussion

The chloroplast genomic features, sequence variation and the potential molecular markers in Sanicula

The genus of Sanicula L. could be easily distinguished by basal leaves orbicular, rounded-cordate or cordate-pentagonal, usually palmately lobed; flowers polygamous, umbels in racemous, cymous or corymbose inflorescences from other genera of Subfam. Saniculoideae in Apiaceae. However, to understand the taxonomy and phylogenetic relationships in Sanicula had been particularly difficult based on its varied morphological characters in rhizomes, leaves, inflorescences and fruits. As a result, the previously reported chloroplast genomes of certain Sanicula species, which have been associated with ambiguous or incorrect information and potential misidentifications, were not included in our analysis, including S. astrantiifolia, S. chinensis, S. giraldii, S. lamelligera, S. orthacantha. In this study, 13 Sanicula genomes (including four newly sequenced, five re-annotated, and four previously reported) representing nine species, one variety and two Eryngium species were used to clarify the phylogenetic relationship.

The structure, gene orders and GC content were highly conserved and nearly similar in the samples of Sanicula analyzed here, and were also identical to other cp genomes in other genera of Apiaceae and other angiosperms [2, 17, 20, 21, 23,24,25,26]. The size of the 13 cp genomes varied from 155,335 (S. flavovirens; NC_061752 and OP703176) to 155,764 bp (S. lamelligera; OP703174) (Table 1). The Sanicula cp genomes sequenced here all contained total 126 genes (including 103 unique genes) with the total GC content being 38.16% or 38.25% (Table 1). However, some species were found to contain different numbers of genes in different samples, for examples, S. flavovirens (NC_061752), S. orthacantha var. stolonifera (MT561028), S. rubriflora (MT528260) and S. rubriflora (NC_060324) were reported to contain 129, 133, 133, 130 genes, respectively, whereas all annotated here with 126 genes. To eliminate the influences of references used and annotation software, the 13 samples were re-annotated using Plastid Genome Annotator (PGA) and Geneious Prime 2020.0.5 with Heteromorpha arborescens (NC_053554), and their tRNA genes were verified by tRNA-SE. Unexpectedly, we examined all the 13 sequences re-annotated only with 126 genes and did not find any gene loss in this study (Table 2).

The variation of length in cp genomes usually hinted the IR region expansions, which were useful in evolutionary studies in some taxa [23,24,25, 31,32,33]. However, our findings indicated that there were only minor variations observed in the cp genomes of Sanicula examined, with no significant expansions or contractions. Among 13 Sanicula cp genome, the length of the IR region varied, with S. rubriflora (26,340 bp; MT528260) exhibiting the longest IR length, while S. flavovirens (26,217 bp; NC_061752) had the shortest. Only the ndhF gene, with an expansion length of 34 bp for S. rubriflora expanded to the IRb region, for remaining 12 Sanicula samples were entirely located within the SSC region. And the rps19 gene with contractions length of 27 bp away from IRb region only in S. flavovirens (NC_061752). These results were also similar to the expansion in the cp genome of other species in Apiaceae among IR regions [18, 34].

Genome composition, including the factors such as gene sequence length, tRNA abundance, GC distribution position, and other related features, along with natural selection, were the two major factors affecting codon usage bias [27, 28, 35,36,37]. The total number of 63 codons present across the Sanicula cp genomes encoding 20 amino acids and codon usage was biased towards A or U at the third codon position, which was in consistent with other Apiaceae taxa [2, 23, 29, 30].

Many works proved that the variation of SSRs in cp genomes were widely used in population genetic studies, species identification and evolutionary relationship [26, 34, 38]. In this study, the characteristics of SSRs and short dispersed repeats (SDRs) were also similar among these Sanicula cp genomes. Our results suggested that the mononucleotide (A/T) account for the most abundant repeat type, and the IR regions contained less SDRs and SSRs than LSC and SSC regions, which were consistent with the analyses in other Apiaceae taxa [2, 18, 34]. Therefore, this indicated that the LSC and SSC regions possessed high level of nucleotide variability, which could be used as potential polymorphic molecular markers for identification, phylogeny, and evolutionary study in Sanicula.

Our analysis of nucleotide diversity revealed that 19 IGS within the non-coding regions (17 in the LSC regions, including atpH-atpI, ndhC-trnM, petB-petD, petD-rpoA, petN-psbM, psaJ-rpl33, rbcL-accD, rpoB-trnC, rps16-trnQ, trnE-psbD, trnF-ndhJ, trnH-psbA, trnS-psbZ, trnS-trnR, trnT-trnF, ycf3-trnS, ycf4-cemA; and one in the boundary between IRa and SSC region: trnN-ndhF; one in IR regions: trnV-rps12.), as well as one coding region (ycf1), exhibited high levels of divergence in Sanicula (Fig. 5). This finding was consistent with the diverse patterns typically observed in angiosperms, where nucleotide diversity tended to be higher in non-coding regions compared to coding regions [39]. However, among these variable sequences, the only one chloroplast marker, rps16-trnQ, which was applied in phylogenetic utility for subfam. Saniculoideae of Apiaceae [8, 13]. In this study, 18 hypervariable regions, including atpH-atpI, ndhC-trnM, petB-petD, petD-rpoA, petN-psbM, psaJ-rpl33, rbcL-accD, rpoB-trnC, rps16-trnQ, trnF-ndhJ, trnH-psbA, trnS-psbZ, trnT-trnF, ycf3-trnS, ycf4-cemA, trnN-ndhF, trnV-rps12 and ycf1, had contributed to a certain degree of confusion in the topology (Additional Files 6, 7, 8, 9, 10: Figs. S2–S6). However, the phylogenetic analysis based on trnE-psbD (Additional File 8: Fig. S4B), trnS-trnR (Additional File 9: Fig. S5C) and the concatenation regions (Additional File 11: Fig. S7), resulted in similar topological trees as the whole cp genomes sequences (Additional File 5: Fig. S1A), concatenation of 126 unique IGS regions (Additional File 5: Fig. S1B), concatenation of 79 unique PCGs regions (Fig. 8) could well differentiate the two clades in Sanicula. Therefore, the two new highly variable chloroplast marker, trnE-psbD and trnS-trnR, might be the promising potential molecule makers in phylogeny reconstruction.

Phylogenetic analysis

Based on the conservatism and heritance, cp genomes were effective in inferring the phylogenetic relationships at various taxonomic levels [40]. In our phylogenetic analysis using the 79 unique coding genes, the monophyly and infrageneric classification in Chinese Sanicula were investigated. The status of S. chinenesis, S. hacquetiodes, and S. giraldii was also re-evaluated.

The systematics of Chinese Sanicula had been discussed based on molecular phylogenetic research [2, 8, 13, 14, 17, 21, 41] and morphological study [6, 41]. The phylogenetic trees obtained here were found to be consistent with those reported by Vargas et al. [41] using nuclear ribosomal DNA internal transcribed spacer (ITS). For instance, S. lamelligera and S. orthacantha formed a monophyletic lineage within Sect. Pseudopetagnia, as proposed by Wolff [12]. However, the other two sections, including Sect. Tuberculatae and Sect. Sanicla, defined by Drude [42] and de Candolle [43], and relationship among species within these sections suggested by Shan & Constance [6] were not supported.

The samples in clade I had involucellate bracteoles small and shorter than umbellets, fertile flowers 1 to 3 per umbellule, fruits characterized by tuberculate, prickly, lamellate, squamosa [3, 6]. Two monophyletic subclades, including subclade A and B, were resolved in this clade. Subclade A contained six taxa belonged to Sect. Pseudopetagnia, which characterized by involucellate bracteoles small and shorter than umbellets, fertile flowers only one per umbellule and fruits squamosa, lamellate or with straightly spiculate spicules. Subclade B included two species, S. giraldii of Sect. Sanicla and S. hacquetiodes of Sect. Tuberculatae, with fully supported nested within clade I. Furthermore, Shan & Constance [6] noted that the noteworthy relationship of S. hacquetiodes with Sect. Pseudopetagnia based on the similar morphological characters, including the presence of generally one fertile flowers and tendency towards a subracemose inflorescence structure. However, two species in subclade B lack of a consistent morphological synapomophy except for involucellate bracteoles small and shorter than umbellets.

Species in clade II could be easily distinguished from taxa in clade I by involucellate bracteoles often longer than umbellets in flowering and fertile flowers often 3 or more per umbellule [3, 6]. In this study, clade II showed that Sanicula chinensis of Sect. Sanicla was nested within Sect. Tuberculatae (including S. flavovirens [5] and S. rubriflora), which formed the sister group of S. flavovirens in well supported. In accordance with previous publications [2], it was suggested that S. chinensis and S. orthacantha formed a strong-supported sister group to S. lamelliegera. However, upon conducting a critical examination, the sample of S. chinensis (MK208987) used in the referenced paper [2] might be a misidentification. Thus, it was advisable to exercise caution when utilizing the sequence in future, as the reliability of the results obtained from this data remained debatable. Within clade II, two subclades, namely subclades C and D, were fully supported. Morphologically, species belonging to subclade D exhibited considerably longer involucellate bracteoles length compared to those in subclade C. Subclade C encompassed two species, formerly assigned to two sections (Sect. Tuberculatae and Sect. Sanicla), which could be easily distinguished by flower characteristics. Subclade D contained two samples of S. rubriflora, a species that was previously classified within Sect. Tuberculatae. It was suggested by Shan & Constance [6] that S. rubriflora may potentially represent an ancestral species within the genus Sanicula. Notably, S. rubriflora was more closely related to S. flavovirens in having numerous staminate flowers with pedicels and base tuberculate fruits with stout uncinate prickles above, rather than to S. chinensis, which had bits of staminate flowers with deciduous pedicel and fruit only covered with uncinate prickles. Thus, to better address the issue of inconsistent classification of chloroplast (cp) genomes and morphology more effectively, it was crucial to obtain additional samples from other species in Sanicula. Particularly, the ones within Sect. Tuberculatae and Sect. Sanicla should be included to validate their placement within the Chinese Sanicula.

Taxonomic inconsistencies in the delimitation of taxa continue to pose a challenge within the genus Sanicula. For instance, S. orthacantha var. brevispina was treated as a synonym of S. orthacantha var. orthacantha by Shan & Constance [6] and Hiroe [7], while had been reinstated as a distinct variety by Liou [11], Fu [44], Wang [45], Sheh & Phillippe [3] and Pimenov [4]. Additionally, S. orthacantha var. stolonifera was only recognized by Sheh & Phillippe (2005) along with its publication. In previous study [10], we found that S. orthacantha var. orthacantha definitely differed from S. orthacantha var. brevispina only by short rhizome, oblique rootstock bearing elongated, fibrous roots, sometimes fleshy stoloniferous (vs. slender, elongate and lignified nodes stoloniferous), and S. orthacantha var. stolonifera was a synonym of S. orthacantha var. brevispina. In this study, the results strongly supported the clustering of S. orthacantha var. brevispina with S. orthacantha var. stolonifera, while weakly supporting the relationship between S. orthacantha var. orthacantha and S. orthacantha var. brevispina. Thus, our findings provided substantial support for the treatment proposed by Li et al. [10].

Conclusion

This study reports four newly sequenced complete cp genomes of Sanicula taxa, i.e. S. caerulescens, S. hacquetiodes, S. orthacantha var. brevispina, S. tienmuensis, following the analysis of SSRs, codon usage, IR boundaries, sequence divergence estimates with other nine Chinese Sanicula samples. Insight into the interspecific relationships in the 11 Chinese Sanicula taxa (including 13 samples) verifies, in some degree, the traditional system based on morphology analysis. These results will help to understand the relationship and evolution clearly in Sanicula at the molecular level and benefit their identification, utilization, and protection as herbal medicinal genus.

Methods

Plant materials, DNA extraction and sequencing of the chloroplast genomes

Eight species and one variety of Sanicula L. and one species of Eryngium L. were collected from field observation in China (Table 3). Fresh and healthy leaf tissues were collected in field and stored in silica gel. Voucher specimens were deposited in the herbarium of Institute of Botany, Jiangsu Province and Chinese Academy of Sciences (NAS), and their deposition numbers were listed in the Additional file 12: Table S5. In addition, four complete chloroplast genomes of Sanicula species (Table 1) and one of Eryngium species (Table 1) that publicly available in NCBI GenBank were downloaded with annotations.

Table 3 Collecting information for the nine taxa of Sanicula L. and one species of Eryngium L. sequenced in the study
Full size table

Total genomic DNA was extracted from silica-dried leaf tissues following a modified CTAB method [46]. DNA integrity was examined by electrophoresis in 1% (w/v) agarose gel, and concentration was measured using a NanoDrop spectrophotometer 2000 (Thermo Scientific; Waltham, MA, USA), then accurate quantifications were completed by Qubit 2.0. High-quality DNA libraries constructed and sequenced at Novogene Bioinformatics Technology Co., Ltd. (https://www.novogene.com/, accessed on March 2011 Tianjin, China). The strategy of Nova-PE150 was selected for high-throughput sequencing, with an insert size of 350 bp.

Complete chloroplast genomes assembly and annotation

The clean data of sequencing were directly assembled using the GetOrganelle pipeline [47,48,49]. Bandage v.5.6.0 [50] was used to visualize and manually correct the assembly results. The annotation of the chloroplast genomes was performed in PGA program [51]. Manual correction of start/stop codons and intron/exon boundaries was performed in Geneious Prime 2020.0.5 [52]. All genome maps were drawn by Organellar Genome DRAW v.1.3.1 [53]. The annotated chloroplast genomes were deposited in GenBank (Table 1).

Genome comparison, codon usage analyses and simple sequence repeat analysis

We applied MAFFT v7.490 [54] to align the total 13 cp genomes sequences (Table 1) for examining the divergence regions among Sanicula species. The aligned sequences were performed in Shuffle-LAGAN model via mVISTA program (http://genome.lbl.gov/vista/ mvista/submit.shtml) with the annotated cp genome sequence of S. orthacantha var. stolonifera (GenBank accession no. MT561028) as a reference genome. DnaSP v6 [55] was applied to examine the sequence divergence hotspots with conducting a sliding window analysis to calculate pi values among the cp genomes, with windows size of 600 bp and step size of 200 bp.

IRscope software was used for the 13 cp genome sequences to visualize their IR/SC boundaries. CodonW [56] was implemented to analysis the codon usage bias for all PCGs. SSRs were identified by Web-based simple sequence repeats finder MISA-web (https://www.webblast.ipk-gatersleben.de/misa/), with minimum numbers of 10 repeat units for mono-, 6 repeat units for di-, 5 repeat units for tri-, tetra-, penta-, and hexa-nucleotide SSRs. The maximum length of a sequence between two SSRs was set as 10. REPuter was implemented to detect the SDRs [57], including forward, reverse, complement and palindromic, with the following parameters: a maximal repeat size of 5000, a minimal repeat size of 30, and hamming distance of 3.

Phylogenetic analysis

A total of 24 datasets, including the 13 complete cp genome sequences of Sanicula, concatenation of 126 unique IGS regions, concatenation of 79 unique PCGs regions, 20 highly divergent regions (atpH-atpI, ndhC-trnM, petB-petD, petD-rpoA, petN-psbM, psaJ-rpl33, rbcL-accD, rpoB-trnC, rps16-trnQ, trnE-psbD, trnF-ndhJ, trnH-psbA, trnN-ndhF, trnS-psbZ, trnS-trnR, trnT-trnF, trnV-rps12, ycf3-trnS, ycf4-cemA, and ycf1) and concatenation of 20 highly divergent regions, with two Eryngium species (including one newly reported taxon), i.e. E. planum L. and E. foetidum, selected as outgroup taxa, were used for phylogenetic analysis. Additionally, phylogenetic analyses were performed using BI, ML and MP based on concatenation of 79 unique PCGs included in the final alignment. BI and MP analyses were conducted on the CIPRES Science Gateway website [58]. BI analyses were run with MrBayes on XSEDE version 3.2.7a [59]. Models were selected among model analyzed by MrBayes using Bayesian model choice criteria (nst = mixed, rates = gamma). MP analyses were run with PAUP on XSEDE version 4.a168 [31] using the heuristic search option with 1000 random sequence additions. ML phylogenetic analyses were performed in the IQ-tree program [32, 33] with auto substitution model and 1000 bootstrap replicates for evaluating the node support. FigTree v 1.4 (http://tree.bio.ed.ac.uk/ software/ fgtree/) was used to visualize the resulting trees.