Complete Plastome of Three Korean Asarum (Aristolochiaceae): Confirmation Tripartite Structure within Korean Asarum and Comparative Analyses

The genus Asarum (Aristolochiaceae) is a well-known resource of medicinal and ornamental plants. However, the taxonomy of Korean Asarum is ambiguous due to their considerable morphological variations. Previously, a unique plastome structure has been reported from this genus. Therefore, we investigated the structural change in the plastomes within three Korean Asarum species and inferred their phylogenetic relationships. The plastome sizes of Asarum species assembled here range from 190,168 to 193,356 bp, which are longer than a typical plastome size (160 kb). This is due to the incorporation and duplication of the small single copy into the inverted repeat, which resulted in a unique tripartite structure. We first verified this unique structure using the Illumina Miseq and Oxford Nanopore MinION platforms. We also investigated the phylogeny of 26 Aristolochiaceae species based on 79 plastid protein-coding genes, which supports the monophyly of Korean Asarum species. Although the 79 plastid protein-coding gene data set showed some limitations in supporting the previous classification, it exhibits its effectiveness in delineating some sections and species. Thus, it can serve as an effective tool for resolving species-level phylogeny in Aristolochiaceae. Last, we evaluated variable sites and simple sequence repeats in the plastome as potential molecular markers for species delimitation.


Introduction
The genus Asarum L. (Aristolochiaceae), which consists of about 100 species, is mainly distributed in temperate regions of the Northern Hemisphere [1][2][3][4][5]. Most of the species (~80 species) occur in East Asia, including China, Japan, and Korea, while approximately 15 species and a single species are distributed in North America and Europe, respectively [1,6,7]. Asarum species are well known as oriental folk medicinal herbs and have horticultural potential as well [8][9][10][11][12][13]. In particular, the dried roots and rhizomes of Asarum species (common name "Seshin" in Korea, "Xixin" in China, "Saishin" in Japan, and "Wild ginger" in England) are widely used as a drug in traditional medicinal practices worldwide [14,15]. Asarum sieboldii Miq. and A. heterotropoides F. Schmidt are valued as remedies for aphthous stomatitis, toothache, and gingivitis in Korea and China [10,16,17]. Taxonomically, these medicinal herbs are placed in the section Asiasarum in subgenus Heterotropa of genus Asarum [18].
The section Asiasarum, distributed in northeastern Asia (China, Japan, Korea, and eastern Russia), is defined by perennial rhizome, deciduous leaves, connate style at the base, longitudinally ridged inner surface of calyx tube, and half-inferior to superior ovary [19][20][21][22]. In Korea, these taxa are highly variable in morphology, resulting in taxonomic confusion divergent plastomes have been reported in some American Asarum species, which were suggested as contigs of the sequence, but not a completely circularized map [42]. The structural change was supported based on the complete plastome map of A. sieboldii, a member of the section Asiasarum [43]. However, it remains unclear whether changes in the IRs boundary would be present throughout the section. These rearrangements in the plastome might provide important evidence for inferring the phylogenetic relationship and evolutionary history of their lineage. Additionally, the variable regions of the plastome have been developed as DNA barcode systems to identify each taxon. Owing to the improvements in sequencing technologies, the whole plastome sequence is often used as a super-barcode for the identification of plants these days [44].
In this study, we investigated the plastomes of four Korean Asarum species (all members of section Asiasarum) suggested by Oh [5] to generate robust data for resolving taxonomic/phylogenetic problems. We sequenced and assembled the whole plastomes of three Asarum species (A. heterotropoides, A. misandrum B. U. Oh and J. G. Kim, and A. maculatum Nakai) and one related species (Aristolochia contorta Bunge). Additionally, to confirm the rearrangement events and structural changes in the plastome of Korean Asarum species, we used long reads produced by Oxford Nanopore Technology (ONT) MinION together with short reads generated by Illumina MiSeq for A. maculatum. Moreover, a comparative analysis of the plastomes was conducted using previously published data to establish a molecular basis for the development of novel DNA barcode markers and to infer the phylogenetic relationships of Aristolochiaceae.

Plastome Features and Gene Content
A total of four complete plastomes (three Asarum and one Aristolochia species) were newly sequenced and successfully assembled using high-quality Illumina MiSeq and ONT MinION data. In detail, we acquired 1.35 to 1.70 Gbp of Illumina MiSeq short reads per species and 2.22 Gbp of ONT MinION long reads for one species, A. maculatum. For Illumina MiSeq data, the number of reads after quality trimming (>q20) was about 4,624,218 to 6,918,313, and the average coverage for the plastome ranged from 549 to 2019 (Table S1). For ONT MinION data, the number of reads after quality trimming (> q7) was 523,864, and the average coverage for the plastome was 291 (Table S1). The complete plastome size of the three Korean Asarum species (A. heterotropoides, A. misandrum, and A. maculatum) ranged  (Table S1). For ONT MinION data, the number of reads after quality trimming (> q7) was 523,864, and the average coverage for the plastome was 291 (Table S1). The complete plastome size of the three Korean Asarum species (A. heterotropoides, A. misandrum, and A. maculatum) ranged from 190,168 to 193,163 bp, and the GC ratio ranged from 36.22% to 36.78% ( Figure 1; Table 1).  (1) functionally bipartite.  These plastomes had tripartite structures consisting of a single copy (SC) region ranging from 93,336 to 96,357 bp and a pair of IRs ranging from 48,402 to 48,416 bp. The complete plastome of Aristolochia contorta was 160,556 bp in length and was smaller than that of Asarum species. The GC ratio of Aristolochia contorta was 38.49%, being slightly higher than those of Asarum species. The genome structure of Aristolochia contorta had a  (19,882 bp), and a pair of IRs (25,459 bp) ( Figure 1; Table 1).

Comparative Plastome Structure and Polymorphism
To understand synteny (gene order) and polymorphism in the plastomes in Aristolochiaceae occurring in Korea, we compared the SC/IR boundaries in six Korean Aristolochiaceae species, including the plastome of A. sieboldii of section Asiasarum previously published [43]. The comparative analysis revealed that the plastomes of Asarum species were different from that of Aristolochia in terms of structures (Figures 1 and 2). A disruption in the SSC region was found in the plastomes of Asarum compared to the plastome of Aristolochia ( Figure 1). This involved the incorporation of the entire SSC into the IRs, implying that the plastome of Asarum species is functionally bipartite. In contrast, the plastome of Aristolochia species has a quadripartite structure typical of angiosperms ( Figure 1). We compared the duplicated genes between Asarum and Aristolochia species. A total of 16 genes, detected on the SSC of Aristolochia, were duplicated in the IRs of Asarum (see above). For ndhF, this is located on the SSC/IRB boundary in Aristolochia species, but extends on the boundary of two IRs, with 1 bp of the 3 portion in each other IR, in Asarum ( Figure 2). The ycf1 gene, placed at the SSC/IRA boundary in Aristolochia, is nested interior to the IR in Asarum ( Figure 1). In the SC of Asarum, the region from trnH-GUG to trnE-UUC is inversed, when compared with the LSC of Aristolochia ( Figure 1). The comparison of the plastomes from the same Asarum species also showed these patterns. When two plastome sequences of A. sieboldii [MG551543 (193,356 bp) vs. MW034667 (167,293 bp)] and A. heterotropoides [MN132860 (190,168 bp) vs. MK577409 (159,944 bp)] were compared, two distinct variations were identified. The first one was the reversed orientation of some LSC area which corresponds to the region from trnH-GUG to trnE-UUC (blue arrow in Figure S2), and the second one was the incorporation of SSC into the IR regions (red and orange arrows in Figure S2). This synteny and rearrangements in the plastome of Asarum were verified using the combination of two sequencing datasets (ONT MinION and Illumina MiSeq data) from the identical sample of A. maculatum. All boundaries in the plastome of Asarum were well assembled by those reads ( Figure S1). Comparison of the plastome polymorphism in Asarum species revealed that the plastome sequences are fairly conserved across the four species except for a few regions with variations ( Figure 3). The sequences of exons and untranslated regions are nearly identical throughout the four Asarum species, whereas most of the detected variations were found  Comparison of the plastome polymorphism in Asarum species revealed that the plastome sequences are fairly conserved across the four species except for a few regions with variations ( Figure 3). The sequences of exons and untranslated regions are nearly identical throughout the four Asarum species, whereas most of the detected variations were found in the non-coding sequence ( Figure 3). Within Asarum, the largest size of the IR was seen in A. heterotropoides despite its smallest overall plastome size (Table 1), resulting from nucleotide indel in the non-coding sequence. Distance from the boundary to genes was compared within these species. The distances between exon 1 of rpl16 and the SC/IR boundary are identical within Korean Asarum (973 bp), but the location of psbM with respect to the SC/IRA boundary varies across species. Among the selected taxa, the location of the gene in A. sieboldii is the farthest from the boundary (2523 bp), whereas in A. heterotropoides, it is the closest from the boundary (2473 bp) ( Figure 2). We also examined nucleotide polymorphism (π, nucleotide diversity) among the four Korean Asarum species (Figure 4). The average π was 0.0009, and the value of π ranged from 0 to 0.0297. Overall, nucleotide sequences of the IRs are more conserved (average π = 0.0003) than those of SC (average π = 0.0016). The most highly variable regions (π > 0.020) include one protein-coding gene (rpoA: π = 0.015) and seven intergenic spacers (IGS) (ycf3 intron: π = 0.023, ndhC-trnV-UAC: π = 0.020, trnG-UCC-trnS-GCU: π = 0.015, trnH-GUG-trnT-GGU: π = 0.014, psbE-petL: π = 0.012, atpF-atpH: π = 0.011, psbA-trnH-GUG: π = 0.010).

Repeat Sequences in the Plastome of Asarum Species
The simple sequence repeats (SSRs) in the plastomes of the four Korean Asarum species were analyzed. The total number of SSRs found in the plastome of these species ranges from 953 to 1039 (Table S2). The mono-, di-, tri-, tetra-, penta-, and hexa-nucleotide SSRs were detected among them, and the number of each SSR in the plastome varies across the four Asarum species (Table 3 and S2). In all the species, the most abundant type of SSR is the penta-nucleotide repeat (494 to 547 in number), whereas the least abundant type is the tetra-nucleotide repeat (6 to 7 in number). Except for the mononucleotide SSR with G or C base located in the IGS of trnQ-UUG-rps16, all mono-repeats are composed of A or T base in all the four species. All dinucleotide SSRs are AT or TA base. We further investigated the long sequence repeats (LSRs) shared by the Korean Asarum species (Table S3). A total of 49 LSRs were identified, consisting of an average of 7 palindromic, 20 forward, 16 reverse, and 4 complement repeats. Of these, the longest unit size (120 bp) of the repeat was found in atpH-atpF of A. sieboldii, whereas the shortest unit size (42 bp) of the repeat was detected in trnH-GUG-trnT-GGU of A. maculatum. Most of the repeats (86-92%) are less than 90 bp and nearly a quarter of the repeats (about 16-24%) are situated in or at the border of genic regions. Among the palindromic repeats within the coding region, there is one palindromic repeat on ycf3 of A. maculatum, and no palindromic repeat on ycf3 of the other three Asarum species.

Phylogenetic Inference
The Maximum likelihood (ML) tree was constructed to explore the phylogenetic relationships among 26 species of Aristolochiaceae using 79 protein-coding gene sequences. This analysis included the four Korean Asarum species (newly sequenced species and previously reported A. sieboldii). The overall topology of the phylogeny computed from ML and Bayesian inference (BI) analyses were identical (BI tree not shown). In the phylogeny, all the Asarum species examined here were strongly resolved as a monophyletic group (bootstrap value [BV] of ML/posterior probability [PP] of BI = 100/1) in Aristolochiaceae ( Figure 5). Asarum canadense L., a member of section Asarum of subgenus Asarum, was sister to all the remaining Asarum species, agreeing with the previous analysis [42]. However, subgeneric classification was not supported by our analysis. For example, A. epigynum Hayata belongs to subgenus Geotaenium, but it formed a clade with A. shuttleworthii Britten and Baker fil., a member of section Hexastylis of subgenus Heterotropa. All the remaining Asarum species are members of subgenus Heterotropa, but its monophyly was not supported since A. shuttleworthii was not included ( Figure 5). Members of section Heterotropa formed two clades, while members of section Asiasarum formed a monophyletic clade supported by high BV/PP (100/1) ( Figure 5). In a clade of section Asiasarum, two clades were formed. One clade includes three Korean Asarum species (A. maculatum, A. misandrum, and A. sieboldii) and A. sieboldii from unknown locality [42], while another one contains Chinese accessions of A. sieboldii and A. heterotropoides and Korean A. heterotropoides although the latter one received weak support (BV = 74) ( Figure 5). Saruma henryi Oliv., a monotypic species in this genus, was positioned as the sister to the genus Asarum ( Figure 5). In addition, 10 species of Aristolochia formed another strongly supported monophyletic group and were divided into two subclades (1) subgenus Siphisia (BV/PP = 100/1) and (2) subgenus Aristolochia (BV/PP = 100/1) ( Figure 5).

Structural Changes in the Plastome of Korean Asarum Species
In this study, we demonstrate the tripartite plastome structure in Korean Asarum us- Additional phylogenetic analysis with reassembled plastome sequences resolved similar relationships among Asarum species, except the clustering of A. maculatum and A. misandrum ( Figure S3).

Structural Changes in the Plastome of Korean Asarum Species
In this study, we demonstrate the tripartite plastome structure in Korean Asarum using a hybrid strategy of ONT MinION in combination with Illumina MiSeq. Prior to sequencing the genome with the ONT MinION platform, we tried to validate the IRA/IRB boundary with a conventional PCR, but the region was not amplified. It was inferred that the sequence of the IRA/IRB boundary region formed a hairpin or cruciform structure which made the amplification of this region by PCR difficult. Therefore, we verified the tripartite structure identified in all Korean Asarum species using ONT read mapping ( Figure S1). A recently developed hybrid pipeline including a base error correction tool on the ONT long reads along with Illumina short reads showed higher accuracy than the assembly done with short reads only [45][46][47][48]. Our sequencing results showed that the structural variations were caused by the incorporation and duplication of the SSC into the IR regions. In addition, the complete plastomes of three Asarum and one Aristolochia species were sequenced ( Figure 1; Table 1), and the organization of genes was characterized ( Table 2). Structural variations across taxa in the family, found in this study, might be applicable for phylogenetic studies.
Angiosperm plastomes generally have very little variation at the structural and genic levels, including length, GC ratio, gene order (synteny), and gene content [49,50]. However, the analysis of complete plastomes of Asarum in this study revealed that the structure is syntenically different from those of other plastomes reported from the Magnoliids, including their closely related genera, such as Aristolochia and Saruma in Aristolochiaceae ( Figure 1). This result supports previous genomic studies [42,43]. The plastome structure of the Korean Asarum section Asiasarum is functionally bipartite because of the incorporation of the SSC into the IRs (Figure 1). For such incorporation of the SSC into the IRs, palindromic sequences in the 3 portion of ndhF, was inferred as a factor for the structural change [42]. Palindromic sequences could take the shape of a cruciform and are associated with chromosome break that leads to rearrangement in genomes [51]. However, it seems that such structural changes occur independently according to section [42]. Three species belonging to each section of subgenus Heterotropa, A. minus F. Maek. (section Hexastylis), A. megacalyx (F. Maek.) T. Sugaw. (section Heterotropa), and A. delavayi Franch. (section Longistylis), had a functionally tripartite structure with expanded inverted repeats, while A. canadense (section Asarum of subgenus Asarum) and A. sieboldii (section Asiasarum) were characterized by tripartite structures (= functionally bipartite) [42]. Additionally, there were differences among species in the placement of ndhF, which is placed on the IRA/IRB boundary. In the case of A. canadense, 12 bp of the 3 portion of ndhF expanded in another IR, whereas only 1 bp of the gene intruded the IR in A. sieboldii [42]. Although Lim et al. [43] have reported such structural changes around ndhF in A. sieboldii, it is not clear whether these structural changes are shared within the section. In this regard, our observation that only 1 bp of the 3 portion of ndhF intruded the IR was conserved in Korean Asarum is notable ( Figure 2); this could be an important genomic event that distinguishes this section from others. We confirmed this feature by reassembling short reads from Chinese accessions of A. sieboldii and A. heterotropoides and three species of section Heterotropa: A. costatum (F. Maek.) T. Sugaw., A. minamitanianum Hatus., and A. sakawanum Makino. The IRA/IRB by ndhF intrusion was only observed in A. sieboldii and A. heterotrpoides, but not in members of section Heterotropa ( Figure S4). Therefore, it seems that this functionally bipartite structure could be unique to section Asiasarum.
In comparison with the Aristolochia plastome, there is another positional shift on the SC (=LSC in Aristolochia)/IR boundaries of the plastomes within Korean Asarum; this is the incorporation of a part of the SC region (rps3, rpl22, rps19, and partial rpl16) into the IRs (Figure 1). This implies that the SC/IRs boundary is also unstable in Korean Asarum. In addition, we rechecked the inversion of a large portion including all genes between trnT-GUU and trnH-GUG in the SC region in the plastomes in this section (Figure 1). A mechanism for the rearrangement of the SC in Asarum was described, and 12 bp inverted repeats with low complexity (AATATAAATAAT) flanking trnT-GUU and trnH-GUG were suggested as factors for the rearrangement [40]. Genomic rearrangement by the IRs has been reported across many lineages, such as Monsonia (Geraniaceae), Trachelium (Campanulaceae), and the tribe Desmodieae (Fabaceae) [52][53][54]. The discussion on such events could help understand the evolution of the plastome in Korean Asarum.

Inferring the Molecular Phylogeny of Aristolochiaceae
The genus Asarum is well known for taxonomic difficulty due to morphological similarities, high variability of diagnostic characteristics, and natural hybridization [6,7,[22][23][24]28,30]. In previous studies, the taxonomic assessment of Asarum has been conducted based on flower, leaf, cataphyll, and fruit morphology with a few molecular markers, which may in part complicate the problems [1,18,23,24,30,31]. The whole plastome or plastid proteincoding genes have shown considerable values for reconstructing phylogenetic relationships among the complex taxa at various taxonomic levels during the past decade [55][56][57]. Therefore, we utilized 79 plastid protein-coding gene sequences from 26 taxa in Aristolochiaceae to infer phylogenetic relationships among the four Korean Asarum species and the related species ( Figure 5). The phylogenetic relationships among 26 taxa were partially congruent with the previously reported relationships for Aristolochiaceae [7,58,59]. Our data showed that members of the family were clustered into two clades, Aristolochia and Asarum/Saruma. However, within the genus Asarum, classification at subgeneric and sectional levels was not supported. For example, A. epigynum, a member of section Geotaenium of subgenus Geotaenium, formed a clade with A. shuttleworthii (section Heterostylis of genus Heterotropa) ( Figure 5). Since we included single species for most of sections, the utility of the plastome in delineating classification of Asarum should be tested with more samples, as well as nuclear data. Nonetheless, members of section Heterotropa failed to form a monophyletic group. Since section Heterotropa includes many species (about 80 species), its monophyly needs to be investigated with thorough sampling of taxa and nuclear genomic regions. Interestingly, two clades formed by species of section Heterotropa correspond to their geographic distributions: one clade is composed of A. costatum, A. sakwanum, and A. megacalyx which occur in Japan, while the remaining species distribute in southern China and Taiwan. The section Asiasarum formed a monophyletic group which is sister to the Japanese Heterotropa clade in the ML phylogeny ( Figure 5). Although the branch length was very short, this result is congruent with the previous sister relationships between section Asiasarum and Heterotropa based on nuclear and plastid data sets [30]. In the clade of section Asiasarum, Korean species did not form a clade due to low sequence divergence. However, three accessions of A. heterotropoides were clustered together with Chinese accessions of A. sieboldii collected from eastern China close to North Korea ( Figure 5), indicating either misidentification of A. sieboldii or introgression of the plastome of A. heterotropoides into A. sieboldii. Korean accession of A. sieboldii was included in the clade of other Korean Asarum species along with another A. sieboldii which of locality is unknown [42]. This result reflects the taxonomic confusion in A. sieboldii which has many synonyms. Further analyses with more accessions could resolve the entity of A. sieboldii occurring in the Korean peninsula. Two endemic species of A. maculatum and A. misandrum are closely related based on the nearly complete plastome sequences ( Figure S3). These two species exhibit morphological differentiation, for example, A. misandrum has recurved calyx lobes and a glabrous abaxial surface, while A. maculatum has calyx, which is not curved, and whitish blotches on the adaxial leaf surface [5]. Thus, this result might suggest that these two endemic species have shared a maternal parent sometime in their evolutionary history. However, ITS phylogeny showed that A. misandrum is separated from the remaining Korean Asarum species [31], thus, more samples with nuclear data should be investigated further to resolve phyloge-netic relationships among Korean Asarum species. Although 79 plastid protein-coding genes showed some limitations in supporting the previous classification, it exhibits its effectiveness in delineating some sections and species, therefore, it can serve as an effective tool for resolving species-level phylogeny in Aristolochiaceae.

Molecular Identification
The precise identification of Asarum species is essential to use them as a medicinal resource, but it is often restricted by considerably morphological variation within species. Although DNA barcoding was implemented based on universal barcode markers, such as ITS, rbcL, and matK, for accurate usage of Korean Asarum as a medicinal herb, the discriminatory power at the species level was very low [60]. In such a case, the discovery of novel barcodes with high discriminatory power or the use of the whole plastome as a super-barcode seems to be necessary. Therefore, we identified eight hyper-variable sites located in the SC region through comparison of taxa ( Figure 4); these comprised, one from rpoA, six from the IGS (atpF-atpH, psbA-trnH-GUG, ndhC-trnV-UAC, psbE-petL, trnG-UCC-trnS-GCU, and trnH-GUG-trnT-GGU), and one from the intron of ycf3. None of these regions were employed in previous phylogenetic studies [1,30,31,59]. Of these variable regions, trnG-UCC-trnS-GCU and ndhC-trnV-UAC are also highly variable in the genus Aristolochia. Given this fact, these regions could be estimated as candidates for DNA barcodes on the genus Asarum as well as on the family Aristolochiaceae. Various SSR regions of four Korean Asarum species were detected (Table 3), and some of these loci could be used as markers to infer the historical migration route and genetic diversity/structure within a species. Numerous population genetics studies employed plastid SSR as well as nuclear SSR [61][62][63]. Our SSR data show notable differences in type (tetra-nucleotide) between A. heterotropoides and remaining species (Table 3 and Table S2), supporting its isolation from the remaining Korean Asarum species in the ML tree ( Figure 5). Palindromic repeat was only observed in A. maculatum (Table S1). These sequence variations in the CP genome are potential markers for identification of species.

Sampling, DNA Library Preparation, and Sequencing
Fresh young leaves of three Asarum species (A. misandrum, A. maculatum, and A. heterotropoides) and Aristolochia contorta that grow in Korea were sampled under the permission of the local government (Table S5). Each species was identified based on the key morphological characters described previously [5]. The voucher specimens for all the four species were deposited in the herbarium of the National Institute of Biological Resources (KB), Incheon, Korea. Total genomic DNA was extracted using the DNeasy Plant Mini Kit (Qiagen Co., Hilden, Germany) following the manufacturer's protocol. The purity of the extracted DNA was measured with a NanoDrop ND1000 (Thermo Fisher Scientific, Waltham, MA, USA), and the DNA was also checked visually through 1% agarose gel electrophoresis. After DNA fragmentation with an LE220-plus Focused-ultrasonicator (Covaris

Plastome Assembly and Annotation
For the plastome assembly, we followed the overall assembly pipeline of the dnaLCW method as described by Kim et al. [64]. In brief, high-quality trimmed reads with phred scores greater than 20 were obtained using the CLC quality trim tool, and then de novo assembly was conducted using the QIAGEN CLC Assembly Cell version 4.2.1 (QIAGEN Digital Insights, Aarhus, Denmark). The putative contigs of the plastome were selected and then mapped to reference sequences of Asarum sieboldii (MG551543) and Aristolochia manshuriensis (MN132862). Subsequently, the contigs were merged into a single draft sequence by connecting the overlapping terminal sequences. The draft plastome sequences were curated by remapping PE reads. The sequencing data of A. maculatum were obtained by ONT MinION base-calling using the Albacore version 2.0 (available online: https: //github.com/Albacore/albacore) with default options. Trimming of the adapter and chimeric sequences was performed using Porechop (available online: https://github.com/ rrwick/Porechop) with default options. Trimmed reads were de novo assembled using the SMARTdenovo program (available online: https://github.com/ruanjue/smartdenovo). A single contig was selected by comparing the plastome sequence assembled from Illumina MiSeq data. Gene annotation was performed using GeSeq (available online: http:// chlorobox.mpimp-golm.mpg.de/geseq.html) [65], with tRNAscan-SE and BLAT search of default options. Circular maps were drawn using OGDRAW (available online: http: //ogdraw.mpimp-golm.mpg.de/) [66]. Finally, the sequences of the plastomes of Asarum and Aristolochia species were deposited in GenBank (MN132858-MN132861).

Genome Structure and Comparative Analysis
We compared the overall genome structure, genome size, gene content, and repeats across all six Korean Aristolochiaceae species, and the previously reported plastomes of A. sieboldii (MG551543) and Aristolochia manshuriensis (MN132862) [43,67]. The GC content was compared using the Geneious Prime 2021.0.3 (Biomatters Ltd., Auckland, New Zealand). The Mauve version 2.4.0 software was used for aligning sequences and determining the rearrangements in the plastomes in Aristolochiaceae [68]. We also examined the sequence divergence among the six Korean Aristolochiaceae species through a sliding window analysis by computing π among the plastomes using the DnaSP version 6.0 [69]. For the sequence divergence analysis, we applied a window size of 600 bp with a 200 bp step size. Additionally, the whole plastomes of the four Asarum species were aligned using the MAFFT version 7.450 (available online: http://mafft.cbrc.jp/alignment/server/) and visualized using the Shuffle-LAGAN mode in mVISTA (http://genome.lbl.gov/vista/mvista/submit.shtml). For the mVISTA plot, we used the annotated plastome of A. sieboldii as a reference. Some misaligned regions were manually curated using BioEdit (available online: http: //www.mbio.ncsu.edu/bioedit/bioedit.html). Thereafter, polymorphic regions showing SNPs and indels were investigated at the interspecific levels.
We found repeat elements with two approaches. In the first approach, a web-based simple sequence repeats finder, MISA-web (available online: https://webblast.ipk-gatersleben. de/misa/), was employed to identify SSRs with thresholds of 10 repeat units for mono-, 5 repeat units for di-, 4 repeat units for tri-, and 3 repeat units for tetra-, penta-, and hexa-nucleotide SSRs. In the second approach, we investigated the size and type of LSRs in the plastome of Korean Asarum using Vmatch version 2.3.0 (available online: http://www.vmatch.de). The parameters were set as follows: a minimal repeat size of 40 bp, a minimal repetition number of two, and the type of LSRs (forward, reverse, palindromic, complementary).

Phylogenetic Analysis
Phylogenetic relationships were inferred using 26 species (31 accessions) in Aristolochiaceae as follows: four plastomes newly sequenced in this study (A. heterotropoides, A. maculatum, A. misandrum, Aristolochia contorta) and 27 previously reported CP genomes, including 16 Asarum species, 10 Aristolochia species, and Saruma henryi (Table S4). Two Piper species (P. cenocladum [DQ887677] and P. kadsura [KT223569]) were added as the outgroup. Due to structural variations and incomplete sequences in plastomes of this group, we used 79 protein-coding gene sequences. Protein-coding gene sequences were retrieved using annotation software Chloë (web application: https://chloe.plantenergy.edu.au/) and sorted alphabetically. After removing duplicated gene sequences, sorted sequences were concatenated and aligned using MAFFT and manually edited using the Geneious alignment viewer. Gaps in the sequences were treated as missing. We inferred the phylogeny using two approaches, the ML analysis and BI. We constructed the ML phylogeny using IQ-TREE version 1.6.11 with the TVM+F+R2 model [70] and 1000 bootstrap replicates for evaluating the node support. The nucleotide substitution model used in the phylogenetic analysis was chosen based on the Akaike information criteria implemented in the jModel-Test version 2.1.10 [71]. BI phylogeny was reconstructed using MrBayes version 3.2.6 [72] under the following settings: a simple nucleotide model (lset nucmodel = 4by4, lset nst = 6), a proportion of invariable sites (lset rates = invgamma). The Markov chain Monte Carlo (MCMC) algorithm was applied for 5,000,000 generations after 25% burn-in, and sampling of trees every 1000 generations. The consensus trees were finally edited using the Figtree version 1.4.3 (available online: http://tree.bio.ed.ac.uk/software/figtree/).
In addition, we reassembled the plastome sequences of five Asarum species using A. sieboldii (MG551543) as a reference to better resolve the relationships among Korean Asarum species. Short reads generated by either Illumina HiSeq or the Ion Torrent Personal Genome Machine (PGM) were downloaded from NCBR Short Read Archive (SRA) and were assembled with Geneious Prime. The plastomes from six and three species belonging to section Asiasarum and Heterotropa, respectively, were aligned using MAFFT, and the AT-rich regions were excluded due to ambiguity. The ML phylogeny was constructed as stated above, except employing a model of TVM+F+I.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/plants10102056/s1, Figure S1: In silico validation of plastome junctions by Illumina MiSeq and Oxford Nanopore MinION read, Figure S2: Comparison of the plastome structure between fully and partially assembled plastomes of the same species, Figure S3: A ML phylogeny of Asarum species based on 186,718 bp of the assembled plastome with exclusion of the AT-rich regions, Figure S4: In silico validation of plastome junctions in the plastome sequences reassembled with sequences from NCBI SRA, Table S1: Sequencing and mapping results for newly sequenced five Korean Aristolochiaeae species, Table S2: Information of SSR across four Korean Asarum species, Table S3: Information of long sequence repeats across four Korean Asarum species, Table S4: Sequence information of species used in phylogenetic analysis, Table S5: Voucher information the Korean Asarum and Aristolochia species examined in this study.