Plastomes Provide Insights into Differences between Morphology and Molecular Phylogeny: Ostericum and Angelica (Apiaceae) as an Example

: Traditional classiﬁcation based on morphological characters suggests that the genus Ostericum is closely related to Angelica , but molecular phylogenetic studies suggest that the genus Ostericum is related to Pternopetalum rather than Angelica . In this study, the plastomes of nine Oster-icum species and ﬁve Angelica species were used to conduct bioinformatic and comparative analyses. The plastomes of Ostericum and Angelica exhibited signiﬁcant differences in genome size, gene numbers, IR junctions, nucleotide diversity, divergent regions


Introduction
The division of species and the evolutionary relationship between them are the basis of biology.Over the last 50 years, phylogeny has become increasingly more based on molecular data, increasingly favoring homologous sequences rather than morphological characters [1].Molecular phylogeny had been widely used to evaluate traditional plant taxonomy (e.g., Angiosperm Phylogeny Group III and IV (APG III, 2009; APG IV, 2016)) [2,3].The differences between morphology and molecular phylogeny come under observation in many plants.For example, the phylogeny based on molecular data of Hedyosmum (Chloranthaceae) indicated differences between morphology and molecular phylogeny [4], and similar circumstances have been found among Restionaceae, Anarthriaceae, and Centrolepidaceae [2,5].In addition, this phenomenon is very common in Apiaceae, especially in Apioideae.The Apioideae were divided into 8 tribes and 10 subtribes based on fruit morphology, but a total of 41 major clades within Apioideae have been identified based on molecular phylogenetic studies [6][7][8].
Apioideae is the largest and best-known subfamily of Apiaceae and includes many familiar edible and medicinal plants in China (e.g., Daucus carota var.sativa Hoffm., Coriandrum sativum L., Peucedanum praeruptorum Dunn, Bupleurum chinense DC.) [7,9,10].The species of Apioideae are difficult to attribute to known genera and species.Most traditional classifications of Apiaceae have relied almost exclusively on fruit characters [11].However, major classifications of Apioideae produced some differences between morphology and molecular phylogeny [8,[12][13][14].A typical example is that Angelica L. is consistent with Ostericum Hoffm. in the morphological characteristics (e.g., fruits (Figure 1), flowers, and leaves) but distantly related to Ostericum in molecular phylogenetic studies [8,15,16].Angelica is a large and taxonomically complex genus of Apiaceae that comprises more than approximately 100 species around the world [9,10,17,18].The genus Angelica with related genera, including Archangelica Wolf, Coelopleurum Ledeb., Conioselinum Fisch.ex Hoffm., Czernaevia Turcz., Glehnia F. Schmidt ex Miq., Levisticum Hill, and Ostericum, is composed of a large and controversial group called Angelicinae Drude in Flora Reipublicae Popularis Sinicae [9] or Angelica sensu lato in some other publications [19].This group is widely distributed in northern temperate regions and exhibits high diversities and variations in morphology, resulting in problematic generic limits [9,10].In addition, there are many economically important plants in Angelica and Ostericum.Among them, A. dahurica (Fisch.ex Hoffm.)Benth.et Hook.f. ex Franch.e, A. decursiva (Miquel) Franchet & Savatier, and O. citriodorum (Hance) Yuan et Shan are commonly used as traditional Chinese medicines [9,10]; A. amurensis Schischk.Is used as food or fodder for animals; and O. grosseserratum (Maxim.)Kitagawa and O. maximowiczii (Fr.Schmidt ex Maxim.)Kitagawa can also be used to extract aromatic oils [9,10].An in-depth study of the two genera will undoubtedly provide new basic information on the origin and species evolution and improve the resource utilization of these genera.
Ostericum was first described by Hoffmann, with O. palustre (Besser) Besser as the type species.The genus is mainly distributed in the north temperate zone, containing approximately 13 species worldwide, with 11 species distributed in China [9,10,17,20].Since the establishment of the genus Ostericum, the taxonomic position of this genus has been controversial.Maximowicz (1873) moved Ostericum into Angelica as an infrageneric section [21].Drude (1898), based on fruit morphology, incorporated Ostericum and eight other genera into Angelica, and Ostericum became a subgenus named subgen.Ostericum Maxim [11].In contrast, other botanists suggested that Ostericum should be independent of Angelica.Kitagawa suggested that Ostericum should be treated as a relatively independent genus based on fruit morphological studies in 1935 [22].Through chemical composition analysis, Harborne held the same opinion [23].Additionally, after study of the fruit anatomy and pollen ultrastructure of Angelica L. (S.L.), Qin et al. (1995) and Sheh et al. (1997) [24][25][26] regarded Ostericum as an independent genus and the controversy was also embodied in the flora of different regions.In the Umbelliferae of Japan and North American Flora, Ostericum was moved into Angelica [19,27].In Flora Reipublicae Popularis Sinicae and Flora of China, Ostericum was treated as an independent genus but was still considered to be closely related to Angelica [9,10].
With the development of DNA sequencing technology, several molecular markers, such as nuclear DNA (nrDNA), including internal transcribed spacer (ITS) sequences and external transcribed spacer (ETS), and the plastid fragments (rpl16 and rps16), have been used to infer the phylogeny of Apioideae [6][7][8][28][29][30][31]. The same is true for the genus Ostericum and Angelica, as molecular phylogenetic analysis showed that Ostericum and Angelica had distant relationships based on ITS, ETS, and plastid fragments [8,15,16 [16].These studies' main target species was the genus Angelica, which contains few species of Ostericum, and does not contain the type species of Ostericum (O.palustre).
The plastid DNA has many features, such as monolepsis, small subfractions, multiple replications, and moderate nucleotide substitution rates.Because of these specific features, and with an increasing number of plastomes available from the National Center for Biotechnology Information (NCBI), plastomes have been widely used for the reconstruction of phylogenetic relationships at different taxonomic ranks in angiosperms [32][33][34][35][36].Moreover, plastomes have been exploited in molecular identification, comparative genomics, genome evolution, and population genetics [37][38][39][40].
To date, many complete plastomes of Apiaceae species have been published, including some species of Angelica and Ostericum [41 -43].In this study, we sequenced and annotated the plastomes of five species of Angelica and eight species of Ostericum, combined with the plastome of O. palustre (obtained from NCBI).We aimed to (1) explore the possible causes of differences in molecular phylogenetic and morphological studies between Angelica and Ostericum; (2) perform comparative analysis of plastomes to provide insights into differences between morphology and molecular phylogeny; and (3) reconstruct phylogenetic relationships of Ostericum based on complete plastomes.The scientific names and abbreviations of the species involved in this study are listed in Supplementary Material Table S1.

Taxon Sampling, DNA Extraction, and Sequencing
Fresh leaves were collected from wild plants, and they were desiccated and stored in silica gel.We extracted the total genomic DNA from the stored dry leaves using a modified cetyltrimethylammonium bromide method [44].The herbarium specimens of these species were stored in the Herbarium, College of Life Sciences, Sichuan University (SZ).The specimen voucher details are shown in Table S2.These plants are not key protected plants and the collection of plant material complied with institutional or national guidelines and was conducted following local legislation.
For nrITS, PCR amplification of the complete ITS region used the primers of ITS4 (5 -TCC TCCGCT TAT TGA TAT GC-3 ) and ITS5(5 -GGA AGTAAA AGT CGT AAC AAG G-3 ) [45].PCR amplification proceeded in a 30 µL volume reaction, containing 3 µL of plant total DNA, 1.5 µL of each forward primer and reverse primer, 10 µL of ddH 2 O, and 15 µL of 2 × Taq MasterMix (CWBIO, Beijing, China).PCR amplification of the nrITS region was performed under the setting of initial denaturation for 4 min at 94 • C, followed by 30 cycles of 45 s at 94 • C, 45 s at 53 • C, and 60 s at 72 • C, and then the final extension of 10 min at 72 • C. All PCR products were sent to Sangon (Shanghai, China) for sequencing (single-pass sequencing for ITS4 primer) after being examined using a 1.5% (w/v) agarose TAE gel.The DNA sequences of nrITS were applied for phylogenetic analysis and detailed information is shown in Table S2.For the plastome, these total genomic DNA were sequenced using the Illumina Novaseq-PE1500 platform at Novogene (Beijing, China), with paired-end reads of 2 × 150 bp.

Plastome Assembly and Annotation
The clean data were assembled using NOVOPlasty 2.7.1 [46] with the default K-mer value of 39 and rbcL of A. sylvestris Linnaeus (GenBank accession No. DQ133798.1)was used as seed input for Angelica species and rbcL of O. sieboldii (Miq.)Nakai (GenBank accession No.: D44579.1) was used as seed input for Ostericum species.Preliminary genome annotation was conducted using PGA [47] and manual modifications for uncertain genes, and uncertain start and stop codons were corrected based on comparison with other related plastomes using Geneious R11 [48].To agree on the standard and reduce error, all the plastomes obtained from NCBI (National Center for Biotechnology Information) were reannotated.Protein-coding gene extraction was performed by PhyloSuite [49].The sequenced species' annotated genome sequences were submitted to GenBank, and their corresponding accession numbers are listed in Table S2.Circular gene maps of the annotated genomes were constructed using the online program Chloroplot (https: //irscope.shinyapps.io/chloroplot/;accessed on 19 May 2022) [50].

Plastome Comparative Analyses
The junctions between single-copy regions (LSC region and SSC region) and inverted repeat regions (IRA region and IRB region) among these species (we sequenced five species of Angelica, eight species of Ostericum, and the plastome sequences of O. palustre and two species of Pternopetalum Franch., which were obtained from NCBI) were compared using Geneious R11 and then visualized manually.
The plastome simple sequence repeats (SSRs) of these species were generated using Perl script MISA [51] with the same settings: 10 repeats for mononucleotide, 5 repeats for dinucleotide, 4 repeats for trinucleotide, 3 repeats for tetranucleotide, 3 repeats for pentanucleotide, and 3 repeats for hexanucleotide.
To compare the plastomes' difference in sequences at the level of the genome, the whole plastomes' alignment of five species of Angelica, nine species of Ostericum, and P. davidii Franch.were generated and visualized using the mVISTA program with the Shuffle-LAGAN model [52], with P. vulgare (Dunn) Hand.-Mazz.as the reference.We trimmed the sequences at the ycf1 gene promoter in IRA region, keeping all sequences of LSC region, IRB region and SSC region.
To detect the nucleotide diversity of plastomes, single nucleotide polymorphism (SNP) analyses were generated using DnaSP v5 [53].The parameters were set as follows: the window length was 600 bp and the step size was 200 bp.The plastome sequences were aligned using MAFFT v7.402 [54] and calibrated manually in Geneious R11.To minimize the impact of sequences on the plastomes, the alignments of five Angelica species and nine Ostericum species were trimmed from the trnH-GUG gene terminator to the ycf1 gene promoter manually.

Phylogenetic Analyses
To infer the phylogenetic relationships between Angelica and Ostericum, 80 proteincoding genes from 97 complete plastomes and 112 ITS sequences were used to reconstruct phylogenetic trees based on the Bayesian inference (BI) and maximum likelihood (ML) methods.Chamaesium H. Wolff was chosen as the outgroup based on previous studies [14].The protein-coding genes (CDS) were extracted from plastomes using the PhyloSuite program [49] and with manual checks.The best-fit model was chosen using Modeltest 3.7 [55].The best-fit model for ITS was GTR+G (BI and ML) and for cpDNA (CDS) was GTR+G (ML) and GTR+G+I (BI), respectively.Maximum likelihood (ML) analyses were undertaken using RAxML v8.2.4 [56] with 1000 bootstrap replicates.Bayesian inference (BI) analyses were performed in MrBayes version 3.2 [57] using a Markov chain Monte Carlo (MCMC) method.We set the number of generations to 10 million starting from a random tree and sampling one tree every 1000 generations.All runs were inspected to check that the average standard deviation of split frequencies was <0.1.The first 20% of the obtained trees were discarded as burn-in and the remaining trees were used to calculate the 50% majority-rule consensus topology and posterior probability (PP) values.The tree display and annotation were performed using the iTOL (https://itol.embl.de/;accessed on 19 May 2022) online tool [58].In addition, we used the MEGA6 [59] program to detect the characteristics of the CDS data sets and ITS data sets for phylogenetic analysis.

The Plastome Features of Angelica and Ostericum
All complete plastomes of Angelica and Ostericum had a single and typical quadripartite structure (Figure 2) that was divided into four regions: two inverted repeat regions (IRs), a large single-copy region (LSC), and a small single-copy region (SSC).The size of the plastomes of Angelica ranged from 146,765 (A.biserrata (Shan et Yuan The genome total GC content was 37.5% in Angelica and 37.4-37.7% in Ostericum.The plastomes of the Angelica and Ostericum shared 114 unique genes, including 80 protein-coding genes (PCGs), 30 transfer RNA genes (tRNAs), and 4 ribosomal RNA genes (rRNAs).The total number of genes in Angelica was 129 while in Ostericum, it was 134.The total number of protein-coding genes (PCGs) in Angelica was 84 while in Ostericum, it was 87.The total number of transfer RNA genes (tRNAs) in Angelica was 36 while in Ostericum, it was 37. Compared to Angelica, Ostericum had four more duplicated genes: trnI-CAU, ycf2, rpl23, and rpl2 (Tables 1 and 2).The amount of genes, PCGs, tRNAs, and rRNAs of Pternopetalum was identical to Ostericum.Additionally, the size of Pternopetalum plastomes, LSC length, SSC length, and IR length were similar to Ostericum (Table 1).

Analyses of Inverted Repeat Contraction and Expansion
To assess the expansion and contraction of the IR regions in Angelica, Ostericum, and Pternopetalum (which is related to Ostericum in the phylogenetic trees), we illustrated the junctions of IR/LSC and IR/SSC (Figure 3).The junctions between the single-copy regions and IR regions are designated as J LA (LSC/IRA), J LB (LSC/IRB), J SA (SSC/IRA), and J SB (SSC/IRB).Among all the species, the junctions of J SA and J SB exhibited high conservation in these three genera, but J LA and J LB exhibited significant differences between Angelica and Ostericum, which contrasted with J LA and J LB of Ostericum being the same as Pternopetalum.The positions of the junctions J SB and J SA are relatively consistent among genera: J SB between the ψycf1 gene and ndhF gene, and J SA occurs in the ycf1 gene.The specific description of the junctions among these genera are as follows.
In Angelica, junction J LB occurs in the ycf2 gene, with ycf2 435-669 bp located in the IRB region.Junction J LA occurs between the trnL-CAA gene and trnH-GUG gene.The trnL-CAA gene located in the IRA region with 775-1009 bp away from J LA , and the trnH-GUG gene located in the LSC region with 461-1042 bp away from J LA .Junction J SB is between the ψycf1 gene and ndhF gene except for A. dahurica, in which J SB occurs within the 69-bp end of the ndhF gene.Junction J SA occurs in the ycf1 gene, with 1609-1901 bp of the ycf1 gene duplicated in the IRA region.
In Ostericum, the IR regions expand outwards by ~8000 bp such that the ycf2, rpl23, rpl2, and trnI-CAU genes are all contained within the IRs.Thus, junction J LB occurs in the rps19 gene, resulting in the duplication of part of this gene named ψrps19 (81 bp) in the IRA region, and there is 3-bp noncoding sequence between J LA and the trnH-GUG gene.Junction J SB is between the ψycf1 gene and ndhF gene, and there are 9-4906 bp of noncoding sequences between J SB and the ndhF gene.Junction J SA occurs in the ycf1 gene, with 782-1939 bp of the ycf1 gene duplicated in the IRA region.
In Pternopetalum, junctions J LA and J LB are completely consistent with Ostericum.In P. davidii, junction J SB is between the ψycf1 gene and ndhF gene, 87 bp of noncoding sequences away from J SB , and junction J SA occurs in ycf1, with 2399 bp of the ycf1 gene duplicated in the IRA region.In P. vulgare, junction J SB occurs within the 6-bp end of the ndhF gene, and junction J SA occurs in ycf1, with 1848 bp of the ycf1 gene duplicated in the IRA region.LSC (large single-copy), SSC (small single-copy), and IR (inverted repeat) regions.

Single Sequence Repeat (SSR) Analyses
We detected six types of SSRs (mononucleotide, dinucleotide, trinucleotide, tetranucleotide, pentanucleotide, and hexanucleotide) in the Angelica and Ostericum plastomes (Figure 4).There was no discernible difference in the number of each SSR type, the total number of SSRs, and the distribution of SSRs in the plastomes between Angelica and Ostericum (Table S3, Figure S1).We identified 59-86 SSRs.In these species, O. citriodorum had the smallest number of SSRs (59) while A. biserrate had the largest (86).The number of mononucleotide SSR was the largest, followed by dinucleotide, tetranucleotide, and trinucleotide repeats.Pentanucleotide and hexanucleotide repeats were very rare, especially hexanucleotide repeats (Figure 4).Most SSRs were located in the LSC region, followed by the SSC region, and then the IRA/IRB region (Figure S1).However, beyond this, we found that the repeat units of SSR types have similarities within the genus while appearing to be different between the two genera (Figure 4, Table S4).The type of SSR repeat unit was similar within the genus but different between Angelica and Ostericum, mainly showing trinucleotide, tetranucleotide, and pentanucleotide repeats (Figure 4C-E, Table S4).In the trinucleotide, the repeat units in Angelica were almost all ATT and ATA while in Ostericum, they were almost all TAT, AAT, and TTC.For the pentanucleotide repeat units, the Angelica species shared TTTTA, which was not detected in Ostericum, while most Ostericum species shared AAATA or TACTA, which were not detected in Angelica.The repeat units of tetranucleotides are special: AGGT, ATTT, CAAT, and CTAC were detected in both Angelica and Ostericum.However, the Angelica species shared TCCT, TCTT, TTTA, AAAT, and AATA within the genus and these repeat units were not detected in Ostericum.On the other hand, the Ostericum species shared GAAA, TAAA, ACTA, and ATTA, which were not detected in Angelica.

Sequence Diversity Analyses
The mVISTA online program with the Shuffle-LAGAN mode was employed to analyze the comprehensive sequence discrepancy of Angelica and Ostericum using P. vulgare as a reference.The sequence identity results revealed that the whole plastome was conserved within the genus, but there were significant differences between Angelica and Ostericum.Plastomes among Ostericum and Pternopetalum were more similar and conservative (Figure 5).The IR regions were more conserved compared to the LSC and SSC regions and coding regions had more sequence conservation than noncoding regions for all genomes.In addition, the nucleotide diversity (Pi) of plastomes in Angelica and Ostericum was calculated to estimate the sequence divergence level of different regions (Figure 6, Table S5).The nucleotide diversity results suggest that sequences with high Pi values are predominantly in intergenic spacers; however, the ycf1 gene regions were an exception in both Angelica and Ostericum.In Angelica and Ostericum, the Pi values ranged from 0.00 to 0.025 and 0.00 to 0.23912, respectively.The corresponding averages of the Angelica and Ostericum plastome sequences were 0.0031 and 0.0059, respectively.We marked the regions with the top 10 Pi values in Angelica (Figure 6A).The top 10 Pi values are located in four regions, including the trnE-UCC-trnT-GGU, petA-psbL, ndhF-rpl32, and ycf1 gene regions.Moreover, the petA-psbL region had the highest Pi values (0.025).In Ostericum (Figure 6B), the majority of the top 10 Pi values were located in ψycf1-ndhF and the highest Pi value was 0.23912.This difference is too large.This is because O. maximowiczii and O. scaberulum (Franch.)Yuan et Shan have longer sequences (~4000 bp) inserted between the ycf1 pseudogene and ndhF gene.For comparison with Angelica, we chose another four regions with high Pi values, excluding the ψycf1-ndhF region.The four regions were the atpI-atpH, ndhC-trnV-UAC, psbH-petB, and ycf1 gene regions, respectively.Among them, the psbH-petB region had the highest Pi values (0.01894), excluding the ψycf1-ndhF region.
Figure 6.The nucleotide diversity of (A) five Angelica species and (B) nine Ostericum species.The regions with higher Pi values are marked.LSC (large single-copy), SSC (small single-copy), and IR (inverted repeat) regions.

Phylogenetic Analyses
To investigate the phylogenetic relationships between Angelica and Ostericum, we used 80 different protein-coding genes from 97 complete plastome sequences (14 newly sequenced in this study and 83 obtained from NCBI) to reconstruct phylogenetic trees based on the maximum likelihood (ML) and Bayesian inference (BI) methods (Figures 7 and S2).We also used 112 ITS sequences (we sequenced 5 sequences of Angelica and 17 sequences of Ostericum, listed in Table S2) to reconstruct phylogenetic trees between Angelica and Ostericum (Figures 7 and S3).The alignments of the CDS data sets for phylogenetic analysis showed a length of 62,264 bp, with 11,726 variable sites (18.83%) and 6525 parsimonyinformative characters (PICs; 10.48%).The alignments of the ITS data sets for phylogenetic analysis showed a length of 670 bp, with 411 variable sites (61.34%) and 357 parsimonyinformative characters (PICs; 53.28%).
Both the ML and BI analyses produced congruent tree topology, indicating that Angelica and Ostericum are not monophyletic.For Angelica, the majority of members are distributed in the Selineae tribe (plastid trees: BS = 100, PP = 1.00;ITS trees: BS = 60, PP = 0.72) while the rest (A.ternate Regel et Schmalh., A. multicaulis Pimenov, and A. paeoniifolia Shan et Yuan) are located in the Hymenidium (Sinodielsia) clade.However, Ostericum is located in the Acronema clade and is closely related to Pternopetalum (plastid trees: BS = 100, PP = 1.00;ITS trees: BS = 100, PP = 1.00).The relationships between Ostericum and Angelica were far from the phylogenetic trees.In addition, it is worth noting that O. grosseserratum (GenBank number: KT8524844) is a sister to A. tianmuensis (plastid trees: BS = 100, PP = 1.00) in the phylogenetic trees but the O. grosseserratum we sequenced in this study is located in Ostericum and is a sister to O. atropurpureum (plastid trees: BS = 100, PP = 1.00).In addition, O. huadongense was placed in Ostericum (plastid trees: BS = 100, PP = 1.00;ITS trees: BS = 100, PP = 1.00) rather than Angelica in this study.Thus, there is no doubt that O. huadongense is a member of Ostericum.O. muliense (R.  Phylogenetic relationships between Angelica and Ostericum and related groups are inferred from Bayesian inference (BI) based on the protein-coding genes of plastid genomes (left) and nrITS (right).Bayesian posterior probabilities (BI PPs) are presented at the nodes.The complete plastid genome sequences and ITS sequences obtained from NCBI show the GenBank number adjacent to the species names.Angelica is highlighted in blue, Ostericum is highlighted in red, and Pternopetalum is highlighted in green.The illustration of the fruit in the transverse section from top to bottom is A. dahurica, O. scaberulum, and P. davidii.

Comparative Analyses of Plastomes
Generally, plastomes are highly conserved in genome structure, GC content, and gene order [60,61].However, previous studies have confirmed that expansion and contraction of IRs in plastomes has often occurred in Apiaceae [62,63].In this study, we conducted comparative analyses of these plastomes of Angelica and Ostericum, and all plastomes presented a typical quadripartite structure, including the LSC region, SSC region, and two IR regions.For all these plastomes, the SSC region contained the same 12 protein-coding genes (the protein-coding gene ycf1 counted in the SSC region) and this is the same in Apioideae [37][38][39]64,65].Moreover, the gene numbers distributed in the LSC and IR regions were different between Angelica and Ostericum.The LSC region had 64 protein-coding genes in Angelica while the LSC region had 61 protein-coding genes in Ostericum, and the IR region had 4 protein-coding genes in Angelica while the IR region had 7 proteincoding genes in Ostericum.However, the plastomes' characters of Ostericum were consistent with Pternopetalum.These results indicate that Ostericum is more similar to the genus Pternopetalum rather than the genus Angelica in genome structure, gene numbers, and gene order.
The contraction and expansion of IR regions is important for genome size variations [66][67][68].The IRs of Ostericum and Pternopetalum were ~8000 bp longer than those of Angelica, causing the number of IR genes in Ostericum and Pternopetalum to be greater than that in Angelica.The expansion of IR regions in Apiaceae has been reported in previous studies, showing that junctions J SB and J SA have similar gene positions in all Apiaceae plastomes, and that IRs' expansion and contraction mainly occurred in J LA and J LB [43,62,63].In this study, Ostericum was consistent with Pternopetalum in all junctions.However, in the junction of IR/LSC (J LA and J LB ), contraction and expansion of IRs between Angelica and Ostericum was found.In addition, we observed that junctions J SA and J SB generated slight expansion and contraction.The IRB region expanded to the SSC region, as reflected in the junction J SB with the ndhF gene in A. dahurica and P. vulgare, and this phenomenon appeared in A. apaensis [43].Meanwhile, compared to other species in Apiaceae, the size of the SSC region, ranging from 17,436 to 23,685 bp in Ostericum, is unique [37,39,43,62,63,69].The contraction and expansion of IRs suggests that Ostericum is more similar to the genus Pternopetalum rather than the genus Angelica.
Highly divergent regions were detected in plastomes, with Pi values indicating substitutions in the respective regions, and the highly divergent regions displayed high Pi values [38].The mVISTA analyses and nucleotide diversity (Pi) results reveal that the IR regions were more conserved than the LSC and SSC regions [38,70,71].Because intergenic regions are under weaker selection pressure and possess a higher evolutionary rate than genes, intergenic regions have more substitutions than gene regions [72,73].Thus, intergenic regions are suitable for the study of the classification and evolution of low taxonomic levels, and plastid barcoding markers have been applied in some plants (e.g., Salvia subg.Perovskia) [72][73][74].In this project, for Angelica, we found four regions with high Pi values (top 10 Pi values).For Ostericum, we also chose four regions with high Pi values, excluding the ψycf1-ndhF region.Among these regions, only the ycf1 gene region was detected in both genera while the other regions did not overlap.Different highly divergent regions were selected for Angelica and Ostericum, which might be more suitable for the development of potential molecular markers and species identification for the two genera.Moreover, we observed that Ostericum had the highest Pi value and a higher average value, suggesting that Ostericum has more substitutions than Angelica.The different highly divergent regions observed in Angelica and Ostericum suggest that Ostericum may not be closely related to Angelica.
Due to the high level of polymorphism, simple sequence repeats (SSRs) have been recognized as one of the main sources of molecular markers and have been widely used in plant population genetics and evolutionary studies [34,75,76].Furthermore, next-generation sequencing can be used to select SSRs and produce SSR markers more conveniently [77].
In our study, the most abundant SSRs were mononucleotides, followed by dinucleotide, tetranucleotide, trinucleotide, pentanucleotide, and hexanucleotide repeats (in decreasing abundance) in both genera.This result was common in Allium [78], Liliaceae [69], and Apiaceae [64].The SSRs were mainly distributed in the LSC region, followed by the SSC region, and the same results have appeared in previous research studies [37,79].The type of SSR repeat unit was similar within the genus but different between Angelica and Ostericum, mainly showing trinucleotide, tetranucleotide, and pentanucleotide sequences (Figure 4C-E, Table S4).Therefore, these SSRs may be promising SSR markers for the identification, classification, and genetic divergence of Angelica and Ostericum.The differences between SSR types indicated a distant relationship between Angelica and Ostericum.

Phylogenetic Position and Intergeneric Relationship of Ostericum
In the phylogenetic trees, our results exhibited significantly improved support and resolution.The relationships between Angelica and Ostericum established by 70 protein-coding genes were compatible with previous studies [8,15,16,80].The species of Angelica were not grouped into a monophyletic group and the majority of members were distributed in the Selineae tribe while a few species were located in the Hymenidium (Sinodielsia) clade [8,81].However, Ostericum occurred in the Acronema clade [8] and formed a monophyletic group with P. neurophyllum and Ostericum was found to be a sister to Pternopetalum.These phylogenetic studies confirm that Ostericum should be treated as an independent genus and that Angelica has a distant relationship with Ostericum.
Focusing on Ostericum, the sequence of O. grosseserratum obtained from GeneBank (GeneBank number: KT8524844) clustered with A. tianmuensis but that newly sequenced in this study occurred in Ostericum.From the genome size and molecular phylogenetic position, we found that KT8524844 has high consistency with A. tianmuensis.Thus, we consider that KT8524844 is probably the sequence of A. tianmuensis.In addition, O. huadongense occurred in Angelica and was found to be a sister to Czernaevia laevigata Turcz. in a previous study [16].In our study, O. huadongense belonged to Ostericum and was found to be a sister to O. sieboldii.Therefore, O. huadongense is undoubtedly a member of Ostericum.
The position of the species P. neurophyllum is special, nested in the Ostericum clade and sister to O. palustre, the type-species of Ostericum.We examined the herbarium specimens of P. neurophyllum in iPlant (http://www.iplant.cn/;accessed on 28 May 2022) and found that P. neurophyllum is similar to O. citriodorum in morphology.These results imply that this species should be transferred into Ostericum.However, morphological data on P. neurophyllum is currently lacking, and further research is needed on P. neurophyllum.In addition, our phylogenetic analyses resolved the taxonomic controversy of O. muliense.O. muliense was first described by Yuan and Shan in 1985 as O. maximowiczii var.alpinum Yuan et Shan from Sichuan, SW China [82] and was accepted as O. maximowiczii var.alpinum in Flora of China [83].However, Pimenov thought the taxon was a separate species in the combination of characters differing from O. maximowiczii by investigation of the type [17,20].Our phylogenetic analyses uncovered that O. scaberulum is a sister to O. maximowiczi and then clustered with O. muliense in plastid trees.Though the phylogenetic position of O. muliense within Ostericum in ITS-based tree is uncertain (low BS and PP), this result still supports the conclusion of Pimenov that O. muliense is a separate species.
The taxonomic position of Ostericum is controversial in its taxonomic history.Ostericum has been treated as a member of Angelica [11,21].Based on chemical studies, flavonoids are ubiquitous in Ostericum [23,92], and karyotypes and pollen ultrastructural studies suggest that Ostericum is a relatively independent genus from Angelica [25,26,93,94].However, they still think that Ostericum is closely related to Angelica based on some similar morphological characteristics, especially fruit morphological characteristics (e.g., the lateral wings of mericarp are separate from each other, elliptic fruit, base slightly round or slightly heart-shaped, dorsally compressed, and vittae obvious) (Figures 1 and 7).On the contrary, the fruits of Pternopetalum are very different from the genus Ostericum (Figure 7) [9,10].Though fruit anatomy and morphology suggested that Ostericum has convex and thickened outer walls and that the exocarp consists of one layer of cells, which is different from all other members of Angelica s.l.[16,24].Combined with the conspicuous calyx teeth of Ostericum [9,10], we can only conclude that Ostericum is an independent genus from Angelica, and we still do not understand whether Ostericum is related to Angelica.However, our phylogenetic study revealed that Ostericum is closely related to Pternopetalum rather than Angelica, which contrasts with the fruit morphological features' distribution.This may be caused by convergent morphology and incomplete lineage sorting (ILS) [86,[95][96][97].Convergent morphology means that distinct lineages independently evolve similar morphological traits [97].This may be the result of adaptation to shared environments.For Ostericum and Angelica, they share the same environment [9,10].Convergence is widely used to interpret the phenomenon of morphological similarities between distantly related lineages.In addition, a recent study has shown that incomplete lineage sorting (ILS) makes ancestral genetic polymorphisms persist during rapid speciation events, and ILS is likely to have affected complex morphological traits in extant species [86].Identification of the real reason for the phenomenon between Angelica and Ostericum requires more research.

Conclusions
In this study, we sequenced and annotated the plastomes of five species of Angelica and eight species of Ostericum (O.sieboldii with two populations).The plastome of O. palustre obtained from NCBI was added for analysis.We found that the plastomes of Angelica and Ostericum exhibited high conservation within the genus but presented significant differences between the two genera in genome size, gene numbers, IR junctions, nucleotide diversity, divergent regions, and the repeat units of SSR types.In contrast, Ostericum was more similar to Pternopetalum than Angelica in comparative analyses of the plastomes.These results of the plastome comparisons were consistent with the phylogenetic analyses.The phylogenetic analyses indicated that Angelica had a distant relationship with Ostericum: Angelica was mainly located in the Selineae tribe while Ostericum occurred in the Acronema clade and was found to be a sister to Pternopetalum.Our results robustly support the taxonomic treatment that separated Ostericum from Angelica as an independent genus and suggest that Ostericum is distantly related to Angelica.Furthermore, our results suggest that O. muliense is a separate species that differs from O. maximowiczii and imply that P. neurophyllum may be a member of Ostericum.There are differences in the comparative morphological analysis and molecular phylogeny between Ostericum and Angelica, and plastomes provide insights into these differences.The cause of the differences may be convergent morphology and incomplete lineage sorting (ILS).Our study provides abundant genetic resources for future molecular phylogeny, evolution, and population genetic studies of Ostericum and Angelica.

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d14090776/s1, Figure S1: Number of SSR types and SSRs' distribution; Figure S2: Phylogenetic relationships between Angelica and Ostericum and related groups are inferred from maximum likelihood (ML) based on the protein-coding genes of plastid genomes; Figure S3: Phylogenetic relationships between Angelica and Ostericum and related groups are inferred from maximum likelihood (ML) based on nrITS; Table S1: Scientific names and abbreviations of the species; Table S2: Voucher details and GenBank accession numbers of taxa sequenced in this study; Table S3: Number of SSR types and SSRs' distribution; Table S4: Repeat units of SSR types; Table S5: Nucleotide diversity analyses of Ostericum and Angelica.

Figure 2 .
Figure 2. Plastid genome map of (A) Angelica and (B) Ostericum type-species.The species name and specific information regarding the genome (length, GC content, and the number of genes) are depicted in the center of the plot.The lengths of the corresponding single short copy (SSC), inverted repeat (IRa and IRb), and large single-copy (LSC) regions are shown.Represented with arrows, the transcription directions for the inner and outer genes are listed clockwise and anticlockwise, respectively.The optional shaded area stretching from the inner sphere toward the outer circle marks the IR regions.Genes are color-coded by their functional classification.

Figure 5 .
Figure 5. Sequence alignment of all the plastid genome sequences sequenced (P.vulgare as the reference).The y-axis represents the percent similarity between 50% and 100%.Different colors represent different genetic regions.

Figure 7 .
Figure 7. Phylogenetic relationships between Angelica and Ostericum and related groups are inferred from Bayesian inference (BI) based on the protein-coding genes of plastid genomes (left) and nrITS (right).Bayesian posterior probabilities (BI PPs) are presented at the nodes.The complete plastid genome sequences and ITS sequences obtained from NCBI show the GenBank number adjacent to the species names.Angelica is highlighted in blue, Ostericum is highlighted in red, and Pternopetalum is highlighted in green.The illustration of the fruit in the transverse section from top to bottom is A. dahurica, O. scaberulum, and P. davidii.

Table 1 .
Features of the plastid genomes of Angelica, Ostericum, and Pternopetalum species.

Table 2 .
List of genes and pseudogenes in Angelica and Ostericum species.