|Home | About | Journals | Submit | Contact Us | Français|
The plastid genomes of different plant species exhibit significant variation, thereby providing valuable markers for exploring evolutionary relationships and population genetics. Glycine soja (wild soybean) is recognized as the wild ancestor of cultivated soybean (G. max), representing a valuable genetic resource for soybean breeding programmes. In the present study, the complete plastid genome of G. soja was sequenced using Illumina paired-end sequencing and then compared it for the first time with previously reported plastid genome sequences from nine other Glycine species. The G. soja plastid genome was 152,224 bp in length and possessed a typical quadripartite structure, consisting of a pair of inverted repeats (IRa/IRb; 25,574 bp) separated by small (178,963 bp) and large (83,181 bp) single-copy regions, with a 51-kb inversion in the large single-copy region. The genome encoded 134 genes, including 87 protein-coding genes, eight ribosomal RNA genes, and 39 transfer RNA genes, and possessed 204 randomly distributed microsatellites, including 15 forward, 25 tandem, and 34 palindromic repeats. Whole-plastid genome comparisons revealed an overall high degree of sequence similarity between G. max and G. gracilis and some divergence in the intergenic spacers of other species. Greater numbers of indels and SNP substitutions were observed compared with G. cyrtoloba. The sequence of the accD gene from G. soja was highly divergent from those of the other species except for G. max and G. gracilis. Phylogenomic analyses of the complete plastid genomes and 76 shared genes yielded an identical topology and indicated that G. soja is closely related to G. max and G. gracilis. The complete G. soja genome sequenced in the present study is a valuable resource for investigating the population and evolutionary genetics of Glycine species and can be used to identify related species.
The chloroplast (cp) is a key organelle in photosynthesis and in the biosynthesis of fatty acids, starches, amino acids, and pigments [1, 2]. In angiosperms, plastomes are typically circular and highly conserved, ranging from 115 to 165 kb in length and comprising a small single-copy region (SSC; 16–27 kb) and a large-single-copy region (LSC; 80–90 kb), separated by a pair of inverted repeats (IRs) [3, 4]. Most plastomes also contain 110–130 genes encoding up to 80 unique proteins and approximately 4 rRNAs and 30 tRNAs. Most of the protein-coding genes are associated with photosynthesis or other biochemical processes in plant cells, such as synthesis of amino acids, sugars, vitamins, lipids, pigments, and starches, storage, nitrogen metabolism, sulphate reduction, and immune responses [5, 6]. In contrast to mitochondrial and nuclear genomes, the plastomes of plants are highly conserved in regard to gene structure, organization, and content . However, gene duplications, mutations, rearrangements, and losses have been observed in some angiosperm lineages . Rearrangements of plastid gene order are generally observed in taxa with plastomes that exhibit at least one of the following qualities: variable IR region size, loss of one IR region, a high frequency of small dispersed repeats, complete or near-complete lack of photosynthesis, or biparental cp inheritance . In addition, plastome inversions have been reported in a number of angiosperm families, including Asteraceae , Campanulaceae , Onagraceae , Leguminosae , and Geraniaceae [13, 14]. The plastomes of several members of the Papilionoideae also exhibit significant variation and rearrangement, including the loss of an IR region  and inversion of a 50-kb portion of the LSC [16, 17]. These features, as well as the loss of introns from the rps12 and clpP genes [18, 19] and transfer of rpl22 to the nucleus [20, 21], have been well documented, and their occurrence has been mapped onto the phylogeny of Leguminosae .
The genus Glycine comprises at least 28 species, which are separated into two subgenera, Glycine and Soja. The annuals include cultivated soybean, G. max, and the wild soybean, G. soja, that are native to eastern Asia, whereas most of the other species are perennials that are native to Australia. Researchers previously classified Glycine species into various groups (A-I) on the basis of fertility of artificially produced hybrids and the degree to which meiotic chromosomes pair , and Ratnaparkhe et al. (2011)  further reviewed the nine genome groups using isozymes and sequences of two nuclear loci (H3D and nrDNA ITS).
Plastid data from various Glycine species (annual and perennial) have been used in studies of phylogenetic and genetic diversity [24–28], including the investigation of neopolyploidy [29, 30]. For example, Doyle et al. (1990b)  identified three major clades within the perennial subgenus, showing varying degrees of agreement with nuclear phylogenies. However, additional research revealed incongruence between the plastid and nuclear phylogenies of the various genome groups . The most noticeable incongruity was the placement of G. falcata, which is the sole species in the F-genome group. According to the H3D gene-based phylogeny, G. falcata is sister to all other perennial species, whereas chloroplast DNA fragment- based phylogenies strongly supported the placement of G. falcata in the A-genome clade [16, 30, 32].
The advent of high-throughput sequencing technology has facilitated rapid progress in the field of genomics, especially in cp genetics. Since the first plastome was sequenced in 1986 , over 800 complete plastid genome sequences have been made available through the National Center for Biotechnology Information (NCBI) organelle genome database, including 300 from crop and tree genomes . To date, complete plastomes have been reported for nine Glycine species [35–37]. In the present study, the complete plastome of G. soja was sequenced (GenBank accession number: KY241814) with the aim of elucidating global patterns of structural variation in the G. soja plastome and comparing it for the first time with the available plastomes of nine other Glycine species (G. max, G. gracilis, G. canescens, G. cyrtoloba, G. dolichocarpa, G. falcata, G. stenophita, G. syndetika, and G. tomentella).
The G. soja (accession KLG90379), seeds were received from the National Gene Bank of the Rural Development Administration of the Republic of Korea. Plants were cultivated in greenhouse at the Kyungpook National University, Republic of Korea. Plastid DNA was extracted from young leaves using the protocol described by Hu et al. , and the resulting DNA was sequenced using the Illumina HiSeq-2000 platform (San Diego, CA, USA) at Macrogen (Seoul, Korea). The G. soja plastome was then assembled de novo using a bioinformatics pipeline (http://phyzen.com). More specifically, a 400-bp paired-end library was produced according to the Illumina PE standard protocol, which resulted in 28,110,596 bp of sequence data, with a 101-bp average read length. Raw reads with Phred scores of 20 or lower were removed from the total PE reads using the CLC-quality trim tool, and de novo assembly of the trimmed reads was accomplished using CLC Genomics Workbench v7.0 (CLC Bio, Aarhus, Denmark) with a minimum overlap of 200 to 600 bp. The resulting contigs were compared against the G. max plastome using BLASTN with an E-value cutoff of 1e-5, and five contigs were identified and temporarily arranged based on their mapping position in the reference genome. After initial assembly, primers were designed (S1 Table) based on the terminal sequences of adjacent contigs, and PCR amplification and subsequent DNA sequencing were employed to fill in the gaps. PCR amplification was performed in 20-μl reactions that contained 1× reaction buffer, 0.4 μl dNTPs (10 mM), 0.1 μl Taq (Solg h-Taq DNA Polymerase), 1 μl (10 pm/μl) primers, and 1 μl (10 ng/μl) DNA, under the following conditions: initial denaturation at 95°C for 5 min; 35 cycles of 95°C for 30 s, 60°C for 20 s, and 72°C for 30 s; and a final extension step of 72°C for 5 min. After incorporating the additional sequencing results, the complete plastome was used as a reference to map the remaining unmapped short reads to improve the sequence coverage of the assembled genome.
The G. soja plastome was annotated using DOGMA  and checked manually, and codon positions were adjusted based on comparison with homologs in the plastome of G. max. The transfer RNA sequences of the G. soja plastome were verified using tRNAscan-SE version 1.21 , with the default settings, and structural features were illustrated using OGDRAW . To examine deviations in synonymous codon usage by avoiding the influence of the amino acid composition, the relative synonymous codon usage (RSCU) was determined using MEGA 6 . Finally, the divergence of the new G. soja plastome from both perennial and annual Glycine species was assessed with mVISTA  in Shuffle-LAGAN mode, employing the new G. soja genome as a reference.
Repeat sequences, including direct, reverse, and palindromic repeats, were identified within the plastome using REPuter , with the following settings: Hamming distance of 3, ≥90% sequence identity, and minimum repeat size of 30 bp. Additionally, SSRs were detected using Phobos version 3.3.12 , with the search parameters set to ≥10 repeat units for mononucleotide repeats, ≥8 repeat units for dinucleotide repeats, ≥4 repeat units for trinucleotide and tetranucleotide repeats, and ≥3 repeat units for pentanucleotide and hexanucleotide repeats. Tandem repeats were identified using Tandem Repeats Finder version 4.07 b , with default settings.
The average pairwise sequence divergence of 76 shared genes and the complete plastomes of 11 Glycine species were analysed using data from G. soja new (KY241814), G. soja old (NC022868), G. max, G. gracilis, G. canescens, G. cyrtoloba, G. dolichocarpa, G. falcata, G. stenophita, G. syndetika, and G. tomentella. Missing and ambiguous gene annotations were confirmed through comparative sequence analysis, after assembling a multiple sequence alignment and comparing gene order. The complete genome dataset was aligned using MAFFT version 7.222 , with default parameters, and Kimura’s two-parameter (K2P) model was selected to calculate pairwise sequence divergence . A sliding window analysis was conducted to determine the nucleotide diversity (Pi) of the cp genome using DnaSP (DNA Sequences Polymorphism version 5.10.01) software . The step size was set to 200 bp, with a window length of 800 bp. Similarly, Indel polymorphisms among the complete genomes were identified using DnaSP 5.10.01 , and a custom Python script (https://www.biostars.org/p/119214/) was employed to identify single-nucleotide polymorphisms. To resolve the phylogenetic position of G. soja within the genus Glycine, ten published Glycine species plastomes were downloaded from the NCBI database for phylogenetic analysis. Multiple alignment of the complete plastomes were constructed based on the conserved structure and gene order of the plastid genomes , and four methods were employed to construct phylogenetic trees: Bayesian inference (BI), implemented using MrBayes 3.1.2 ; maximum parsimony (MP), implemented using PAUP 4.0 ; and both maximum likelihood (ML) and joining-joining (NJ), implemented using MEGA 6 , employing previously described settings [52, 53]. In a second phylogenetic analysis, 76 shared cp genes from eleven Glycine species and two outgroup species (Phaseolus vulgaris and Vigna radiata) were aligned using ClustalX with default settings, followed by manual adjustment to preserve reading frames. Finally, the same four phylogenetic inference methods were employed to infer trees from the 76 concatenated genes, using the same settings [52, 53].
A total of 2,611,513 reads with an average read length of 101 bp were obtained, and these reads provided 1514.9× coverage of the plastome. The consensus sequence for a specific position was generated by assembling reads that were mapped with at least 934 reads per position and was used to construct the complete sequence of the G. soja plastome. The assembled G. soja plastome of was typical of angiosperms, with a pair of IR regions (25,574 bp), an LSC of 83,181 bp, and an SSC of 178,963 bp (Fig 1); a total size of 152,224 bp; and a GC content of 35.4% (Table 1). In addition, approximately 33.23% of the genome was non-coding, whereas protein-coding, rRNA, and tRNA genes constituted 52.06, 5.94, and 1.92% of the plastome, respectively (Table 2), similar to the values observed in other legume genomes. As observed in other angiosperm plastomes, the GC content was unequally distributed in the G. soja plastome; it was high in the IR regions (41.8%), moderate in the LSC region (32.8%), and low in the SSC region (28.73%; Table 1). The high GC content of the IR regions is due to the presence of eight ribosomal RNA (rRNA) sequences in these regions, as reported previously [54, 55].
The total coding DNA sequences (CDSs) were 79,250 bp in length and encoded 87 genes, including 26,416 codons (Table 3). The codon-usage frequency of the G. soja plastome was determined based on tRNA and protein-coding gene sequences (Table 4). Leucine (10.6%) and cysteine (1.2%) were the most and least frequently encoded amino acids, respectively, and isoleucine, serine, glycine, arginine, and alanine constituted 9.0%, 7.7%, 6.5%, 5.8%, and 5.0% of the CDSs, respectively, as reported previously [54, 56].
Among these codons, the most and least frequently used were AAA (n = 1,181), which encodes lysine, and ATC and ATT (n = 1, n = 1), which both encode methionine. The AT contents of the 1st, 2nd, and 3rd codon positions of CDSs were 55.7%, 62.9%, and 72.4%, respectively (Table 3). The high AT content observed at the 3rd codon position is similar to that reported for the plastomes of other terrestrial plants [54, 57, 58]. In addition, 46.36% and 57.65% of the preferred synonymous codons (RSCU > 1) ended with A or U and C or G, respectively, and 44.30% and 55.20% of the non-preferred synonymous codons (RSCU < 1) ended with C or G and A or U, respectively. However, there was no bias in start codon usage (AUG or UGG; RSCU = 1; Table 4).
The G. soja genome map (Fig 1) was representative of known Glycine plastomes in general, and no structural rearrangement was detected among these plastomes. The length of the G. soja plastome was 152,224 bp, which is similar to that of G. max (152,217 bp) , but smaller than those of G. dolichocarpa, G. falcata, G. sydetika, and G. tomentella (Table 1). Among the sequenced Glycine plastomes, that of G. max is smallest, and that of G. dolichocarpa is largest (Table 1). Furthermore, a total of 134 genes were identified in the G. soja plastome, of which 110 were unique, including 87 protein-coding genes, 39 tRNA genes, and eight rRNA genes (Fig 1, Table 5). Similar to other legumes, the plastome of G. soja lacked the rpl22 gene, probably due to an ancient transfer to the nuclear genome . The duplicated IR regions of the G. soja plastome resulted in complete duplication of the rpl2, rpl23, ycf2, ycf15, ndhB, and rps7 genes as well as duplication of exons 1 and 2 of rps12, all four rRNA genes, and seven tRNA genes. The LSC region included 61 protein-coding and 24 tRNA genes, whereas the SSC region included only 12 protein-coding genes and one tRNA gene. The protein-coding genes included nine genes encoding large ribosomal proteins (rpl2, 14, 16, 20, 22, 23, 32, 33, and 36), 12 genes encoding small ribosomal proteins (rps2, 3, 4, 7, 8, 11, 12, 14, 15, 16, 18, and 19), five genes encoding photosystem I components (psaA, B, C, I, and J), 16 genes related to photosystem II (Table 5), and six genes encoding ATP synthase and electron transport chain components (atpA, B, E, F, H, and I; Table 5).
Among the coding genes, rps12 was unequally divided, with its 5′ exon being located in the LSC region and one copy of the 3′ exon and intron being located in each of the IR regions, as in other angiosperms. The ycf1 gene was located at the IRa/SSC boundary, leading to incomplete duplication of the gene within the IR regions. We also identified 12 intron-containing genes, including nine that contained a single intron and three (ycf3, clpP, and rps12) that contained two introns (Table 6). This is in contrast to the situation in Cicer arietinum, Medicago truncatula, Trifolium subterraneum, Pisum sativum, and Lathyrus sativus, all of which have lost an intron from both clpP and rps12 . The largest intron was found in trnK-UUU (2583 bp) and included the entire matK gene, whereas trnL-UAA contained the smallest intron (508 bp). Introns play an important role in the regulation of gene expression, and recent research has shown that introns can improve exogenous gene expression when located at specific positions. Therefore, introns can be a valuable tool for improving transformational efficiency . Furthermore, intron sequences in legume chloroplast DNA have become important tools in phylogenetic analyses . In addition, even though ycf1 and ycf2 [62, 63], rpl23 , and accD [65, 66] are often absent in plants , they have been reported to occur the plastomes of various Glycine species . atpB-atpE pairs were observed to overlap with each other by ~1 bp. However, psbC-psbD exhibited a 53-bp overlap in G. soja plastomes, similar to what is observed in G. max  and G. falcata , Arabidopsis arenosa (17-bp overlap) , Gossypium (53-bp overlap) , and Camellia (52-bp overlap) . Previously, Addachi et al. (2012)  reported the importance of the partial overlap of psbC and psbD cistrons. They demonstrated that the translation of the psbC cistron largely depends on the translation of the preceding psbD cistron, indicating a contribution form independent psbC translation. Similar results were reported in tobacco, where ndhC and ndhK cistrons overlap, and ndhK translation is strictly dependent on the upstream termination codon .
Repeat analysis of the G. soja plastome identified 34 palindromic repeats, 15 forward repeats, and 25 tandem repeats (Fig 2A). Among these repeats, 12 of the forward repeats were 30–44 bp in length, while all 25 tandem repeats were 15–29 bp in length (Fig 2A–2D). Similarly, 27 of the palindromic repeats were 30–44 bp in length, and three repeats were 45–59 bp in length (Fig 2D). Overall, 74 repeats were identified in the G. soja plastome, which is a similar number to the 75, 75, 76, 83, 81, 83, 88, 80, and 80 repeat sequences found in the plastomes of G. max, G. gracilis, G. canescens, G. cyrtoloba, G. dolichocarpa, G. falcata, G. stenophita, G. syndetika, and G. tomentella, respectively (Fig 2A). Therefore, G. soja is more similar to G. max and G. gracilis in terms of repeats. Approximately 29.4% of these repeats were distributed in protein-coding regions. Previous reports suggest that repeat sequences, which contribute to genome rearrangements, can be very helpful in phylogenetic studies [58, 73]. In addition, analyses of various plastomes have shown that repeat sequences induce indels and substitutions , and both sequence variation and genome rearrangement occur as a result of slipped-strand mispairing and improper recombination of such repeat sequences [73, 75, 76]. Furthermore, the presence of repeat sequences indicates that loci are hotspots for genome reconfiguration [58, 77], and repeats can be used to develop genetic markers for phylogenetic and population studies .
Simple sequence repeats (SSRs), or microsatellites, are repeating sequences, typically of 1–6 bp in length, that are distributed throughout the genome. In the present study, we identified perfect SSRs in the plastome of G. soja and in those of nine other Glycine species (Fig 3A). Certain parameters were set because SSRs of 10 bp or longer are prone to slipped-strand mispairing, which is believed to be the main mechanism of the formation of SSR polymorphisms [78–80].
A total of 204 perfect microsatellites were identified in the G. soja plastome (Fig 3A), which is a similar number to the 206, 206, 210, 204, 216, 223, 213, 214, and 211 perfect microsatellites identified in the plastomes of G. max, G. gracilis, G. canescens, G. cyrtoloba, G. dolichocarpa, G. falcata, G. stenophita, G. syndetika, and G. tomentella, respectively (Fig 3A). The majority of the SSRs possessed dinucleotide repeat motifs, varying in number from 66 in G. soja to 76 in G. falcata and G. dolichocarpa, while trinucleotide SSRs were the second most common, ranging in number from 69 in G. syndetika to 74 in G. stenophita. Using our search criterion, two pentanucleotide SSRs were identified in G. soja, G. max, and G. stenophita, and two hexanucleotide SSRs were identified in G. gracilis and G. dolichocarpa (Fig 3A). In G. soja, the majority of the mononucleotide SSRs were A (98.1%) and C (1.81%) motifs, and the majority dinucleotide SSRs were A/T (71.64%) and A/G (23.940%) motifs (Fig 3B, Table 7). In addition, 61.7% of the SSRs were located in non-coding regions, whereas 2.9% and 0.49% were located in rRNA and tRNA genes, respectively (Fig 3C). Further analysis indicated that 64.7% of the SSRs were located in the LSC region, whereas 20.58% and 14.7% were located in the IR and SSC regions, respectively (Fig 3D). These results are similar to previous reports that SSRs are unevenly distributed in plastomes, and the findings might provide more information for selecting effective molecular markers for detecting intra- and interspecific polymorphisms [81–84]. Furthermore, most of the mono- and dinucleotide repeats consisted of A and T, which may have contributed to the bias in base composition, as in the plastomes of other species . Our findings are comparable to previous reports that SSRs in plastomes are generally composed of polythymine (polyT) or polyadenine (polyA) repeats and infrequently contain tandem cytosine (C) or guanine (G) repeats , thereby contributing to AT richness [55, 56, 86].
Ten complete Glycine plastomes were compared with the G. soja plastome. Analysis of genes with known functions indicated that G. soja shared 76 protein-coding genes with nine Glycine species. In addition, the gene content and organization of the G. soja plastome were similar to those of other Glyine species plastomes , but different from the usual gene order of angiosperm plastomes, due to a large inversion (~51 kb) that reversed the order of the genes between trnK and accD (Fig 1). This 51-kb inversion was previously reported in other members of the legume family, especially members of subfamily Papilionaoideae [16, 24, 87], and other inversions have been reported in the plastomes of other species, including a 5.6-kb inversion in Milletia , a 78-kb inversion in various closely related legumes, including Phaseolus and Vigna [17, 89], and a 36-kb inversion within the 51-kb inversion found in Lupinus and other genisotoids . This change in gene order has been ascribed to the contraction and expansion of IR regions, leaving the gene order as described in papilionoids, retaining the 51-kb inversion, but alerting the genes bordering the IR region [89, 91].
Furthermore, the IR region overlaps the ycf1 gene by 478 bp, as observed in legumes exhibiting the same inverted repeat as G. soja. This feature has been shown to distinguish the plastomes of legumes from those of other angiosperms, in which the IR region and ycf1 typically overlap by 1,000 bp . Moreover, as found in the plastomes of other legumes, the plastome of G. soja possessed variation and was missing two cp genes, rpl22 and infA, , both of which have been replaced by cp-targeting nuclear copies [59, 92]. Absence of the rps16 gene from the plastome has also been reported in other legume lineages, excluding Glycine, and the mitochondrial copy is dually targeted to both the cp and mitochondria [19, 93]. Furthermore, loss of the introns in rps12 and clpP has been detected in the plastomes of various species , including those of Glycine species [35, 67].
Pairwise alignment of the new G. soja plastome with the old G. soja plastome and those of nine other genomes showed a high degree of synteny. The annotation of the new G. soja plastome was used as a reference for plotting the overall sequence identity of the plastomes of the other ten Glycine species in mVISTA (Fig 4). In the results, relatively lower sequence identity was observed between the plastomes of the seven other perennial species, especially in the rpoC1, atpF, accD, clpP, rpl2, ndhA, ndhF, rps8, rps19, and ycf1 genes (Fig 4). In addition, the LSC and SSC regions were less similar than the two IR regions in all Glycine species, and the non-coding regions were more divergent than the coding regions. Highly divergent regions included the matK-rbcL, ycf3-psaA, trnC-rpoB, rpl20-clpP, rps16-trnQ, trnfM-trnM, psbM-petN, atpI-atpH, petA-psbJ, and ycf1-rps15 spacers, as reported previously [54, 55]. Our results also confirmed similar differences among various coding regions in the analysed species, as suggested by Kumar et al. . On the other hand, G. soja exhibited high sequence identity with annual Glycine species (S1 Fig), which suggest that they are highly conserved. However, the variation in similarity levels revealed various coding and non-coding regions where the G. soja exhibits divergence from these annual Glycine species (S1 Fig). Similarly, we detected 10 relatively highly variable regions, including 4 gene regions and 6 intergenic regions of the cp genomes, that might be undergoing more rapid nucleotide substitution at species and cultivar levels (S2 Fig) (atpB-rbcL, trnT-trnL, trnS(GGA)-trnG(UCC), psbD-trnT, rps16, rpl33-rpl18, rpl16-rps3, ndhB, ycf1 and ycf15). These regions can be used as potential molecular markers for application in phylogenetic analyses of Glycine. Furthermore, various researchers have determined coding and non-coding regions of particularly high variability as potential molecular markers for Glycine species, such as trnS(GGA)-trnG(UCC), rpl16-rps3, trnT-trnL and atpB-rbcL [95–97]. Similarly, it has been reported that non-coding regions in cp DNA show greater variability in nucleotide regions than coding regions, and these regions have become a major source of variability for phylogenetic studies in various species, including studies within Glycine species [98–100]. Furthermore, comparison of the plastomes of G. soja and related species revealed 72 SNPs and 26 indels in relation to G. max and G. gracilis, respectively (S2 Table). These results confirmed that the highly conserved plastome can include interspecific mutations that may be useful for analysing both genetic diversity and phylogenetic relationships.
Similarly, we calculated the average pairwise sequence divergence among the plastomes of the ten Glycine species (S3 Table). The plastome of G. soja exhibited an average sequence divergence of 0.0096, whereas that of G. cyrtoloba possessed the highest average sequence divergence (0.00567), and those of G. soja and G. max displayed the lowest average sequence divergence (0.00010 and 0.00020, respectively). Furthermore, the nine most divergent genes among these genomes were accD, matK, ycf1, rps16, rpl20, psbM, psbN, petL, and petN. The accD gene exhibited the greatest average sequence divergence (0.07825), followed by ycf1 (0.0241), rps16 (0.0201), and matK (0.0194; Fig 5), most of which were located in the LSC region, and the accD gene of G. soja was highly divergent from those of nine other Glycine species (S3 Fig). The highest nucleotide diversity (Pi) (0.0916) and total number of mutations (Eta) (119 bp) in comparison with the G. soja accD gene was observed in G. cyrtoloba among the plastomes of the nine Glycine species, whereas the lowest were observed in G. syndetika (S4 Table). The length of the accD gene was 1,299 bp (433 aa) in G. soja, G. max, and G. gracilis and 1527 bp (523 aa) in the seven other Glycine species (S3 Fig). Similar differences in gene length within small cpDNA regions have been observed in a variety of other angiosperms . In legume species, both ycf4 and accD exhibit extensive length variation. The expansion of the accD gene is partly explained by the presence of numerous tandemly repeated sequences . This accD gene encodes a subunit of acetyl-CoA carboxylase, which is related to fatty acid synthesis within the plastid. Previous gene knockout experiments have shown that the function of accD is vital, and this gene is expected to be indispensable . However, various studies have identified widespread pseudogenization or absence of accD in a variety of relatively distant lineages, including the Ericaceae, Campanulaceae, Geraniaceae, Acoraceae, Poaceae, and Fabaceae [10, 21, 102–106], which implies that deletion or pseudogenization events occur independently.
Variations in the size of angiosperm plastomes are mostly the result of expansion or contraction of the IR regions [79, 107–109]. In the present study, a detailed comparison of the four junctions (JLA, JLB, JSA, and JSB) between the two IR regions (IRa and IRb) and the two single-copy regions (LSC and SSC) of the 10 Glycine species was performed (Fig 6). Despite the similar lengths of the IR regions of G. soja and the other nine Glycine species, some expansion and contraction were observed, with the IR regions ranging from 25,432 bp in G. stenophita to 25,591 bp in G. dolichocarpa. The genes that marked the beginnings and ends of the IR regions were only partially duplicated, including 68 bp of rpp19 in G. soja, G. max, and G. gracilis and 65 bp of rpp19 in G. dolichocarpa, G. falcata, G. stenophita, and G. tomentella. In G. canescens, G. syndetika, and G. cyrtoloba, this distance was 61 bp in IR region from JLB. Similarly, the hypothetical cp gene ycf1 was partially duplicated, with 478 bp of this sequence being duplicated in G. soja, G. max, and G. gracilis; 463 bp in G. falcata, G. stenophita, and G. tomentella; and 442 bp in G. canescens and G. cyrtoloba. JLA was located between rpl2 and psbA, and the distance between rpl2 and JLA was 122 bp in all of the species except for G. cyrtoloba, where rpl2 is located 188 bp from the JLA border. Additionally, the distance between psbA and the JLA in the G. soja plastome was 314 bp, which was similar to that in the G. max and G. gracilis plastomes. Furthermore, the ndhF gene traversed the SSC and IRa regions, with 1 bp being located in the IR region of G. soja, 37 bp being located in the IR region of G. dolichocarpa, and 19 bp being located in the IR region of G. falcata and G. stenophita (Table 7).
Plastid genomes have been useful in phylogenetic, evolutionary, and molecular studies. During the last decade, many analyses based on the comparison of plastid protein-coding genes [110, 111] and complete genome sequences  have addressed phylogenetic questions at deep nodes and enhanced our understanding of enigmatic evolutionary relationships among angiosperms. The genus Glycine includes 28 species, separated into two subgenera (Soja and Glycine), the former of which includes both cultivated soybean (G. max) and its wild annual progenitor (G. soja), which are distributed in East Asia, including Japan, Korea, China, Russia, and Taiwan. G. max and G. soja are both diploid (2n = 40) and interfertile and are thought to share highly similar genetic variation, although G. soja is much more variable than G. max [25, 113]. Polymorphisms in the cpDNA of G. max and G. soja have been used in numerous studies to assess maternal lineages and cytoplasmic diversity [114–119]. Continued efforts have expanded our ability to differentiate and understand the genomic structure and phylogenetic relationships of Glycine species [28, 120, 121]. The phylogeny and taxonomy of Glycine species in the Soja subgenus have been extensively investigated based on DNA variation, including nucleotide variation in nuclear ribosomal DNA (rDNA), intergenic spacer (ITS) regions , cpDNA restriction sites [24, 29], the histone gene H3-D , A-199a , and cpDNA intergenic spacer regions . However, the complete genome sequence provides more detailed insight [52, 55, 124]. In the present study, the phylogenetic position of G. soja within its genus was established using the complete plastomes (S5 Table) and shared genes of 10 Glycine species and various methods of phylogenetic analysis. Phylogenetic analysis indicated that the complete plastome and the 76 shared genes contained the same phylogenetic signal. In both datasets, G. soja formed a clade with G. max and G. gracilis, with high BI and bootstrap support values (Fig 7, S4 Fig). Moreover, the tree topology confirmed previously reported relationships based on SSR and plastome data [114, 125]. These results of the present study are in general agreement with the results of Gao and Gao (2016) , who reported that G. gracilis is intermediate between the two species and is more closely related to G. max than G. soja. Furthermore, the results of the present study suggest that there is no conflict between the complete genome and the 76 shared gene datasets.
In the present study, the complete plastome sequence of G. soja (152,224 bp) was determined. The gene order and structure of the G. soja plastome were found to be highly conserved with the plastomes of other Glycine species. The present study also revealed the distribution and location of repeat sequences and SSRs as well as the sequence divergence among the plastomes and shared genes between G. soja and nine of its congeners. No major structural rearrangement was observed in relation to annual Glycine species. However, in the perennial species, accD was found to be the most divergent gene, while relatively lower identity was observed in some other regions, especially in the rpoC1, atpF, accD, and clpP genes. Furthermore, phylogenetic analyses based on complete plastomes and shared genes yielded trees with the same topology, at least in regard to the placement of G. soja. Thus, the present study provides a valuable analysis of the complete plastome of G. soja and related species, which may facilitate species identification and both biological and phylogenetic studies.
VISTA-based identity plot showing the sequence identity among the ten Glycine species, using G. soja (new) as a reference. Vertical scale indicates the percentage of identity, ranging from 70% to 100%. Horizontal axis indicates the coordinates within the chloroplast genome. Arrows indicate the annotated genes and their transcriptional direction. A thick black line indicates the inverted repeat (IR) regions.
(Window length: 800 bp, step size: 200 bp). X-axis, position of the midpoint of a window; Y- axis, nucleotide diversity of each window.
The data from the 76 shared genes were analysed with four different methods: joining-joining (NJ), maximum parsimony (MP), maximum likelihood (ML) and Bayesian inference (BI)). The numbers above the branches are the bootstrap values from the NJ, MP, and ML methods and the posterior probabilities of BI. A red dot represents the position of G. soja (KY241814).
This work was supported by Korea Institute of Planning and Evaluation for Technology in Food, Agriculture, Forestry and Fisheries (IPET) through Agriculture, Food and Rural Affairs Research Center Support Program, funded by Ministry of Agriculture, Food and Rural Affairs (MAFRA)(716001-7).
GenBank accession number: KY241814.