DNA was extracted from four species of Polygonum, such as P. aviculare, P. bistorta, P. orientale and P. perfoliatum. Fresh Polygonum leaves were collected from Chengdu (Sichuan Province, China). The specimens of Polygonum have been kept in CDCM (Traditional Chinese medicine herbarium of Chengdu University of Traditional Chinese Medicine) (Supplementary Table S1). The cetyltrimethylammonium bromide method has been used to obtain the whole genome DNA from fresh leaves. The ND-2000 spectrometer was used to quantify the DNA. A shotgun library (250 bp) was built following the constructer guidelines. The X Ten Platform (Illumina, San Diego, CA, United States) was used to sequence through the double terminal sequencing technique with pair-end 150. The total raw data from the mensuration DNA was about 3.5G, and about 12 million paired-ends scrutinizes were finished (Supplementary Table S2).

The raw reads were trimmed using Skewer v0.22 (skewer -q 20 -Q 30–l 100 -t 32) (Jiang et al., 2014). BLAST was used to predict the chloroplast-like reads by cleaning the reads with the sequences of the reference P. cuspidatum (MW411186.1) (Deng et al., 2015; Ye, Lin, Zhou, He, Yan and Cheng 2021). Generally, SOAPdenovo-2. 04 (SOAPdenovo-127mer all -s config. txt -o out -K 51 -R) was used to accumulate the sequences using chloroplast reads (Gogniashvili et al., 2015). Then, those accumulated assembled sequences were extended with the help of SSPACE-3. 0 (SSPACE_standard_v3.0.pl -l library.txt -s out.config -x 1 -T 4 -b sspace.out) and GapCloser-1.12 (GapCloser -a scaffols.fa -b library.txt -o Fanal.fa)were used to fill the gaps (Boetzer et al., 2011; Acemel et al., 2016). To authenticate the accuracy of the connection splicing, a random primer was designed to check the connections of the sequence by polymerase chain reaction. The PCR primer information and amplification conditions are shown in the Supplementary Table S3. The Sanger sequencing results were compared with the assembled chloroplast genome sequence to verify the accuracy of genome linkage.

CpGAVAS (Liu et al., 2012) was used for sequence annotation. The annotation results were checked by DOGMA (http://dogma.ccbb.utexas.edu/) and BLAST (Wyman et al., 2004). In addition, the tRNA genes were classified using tRNAscanSEv1. 21 (Chan and Lowe 2019). The OGDRAWv1. 2 (Lohse et al., 2007) and MEGA5. 2 (Tamura et al., 2011) were used to plot the structural features of the chloroplast genome and define the relative utilization of synonymous codons. MEGA5. 2 (Tamura, Peterson, Peterson, Stecher, Nei and Kumar 2011) was adopted to analyze the relative synonymous codon usage (RSCU). The assembled chloroplast genome sequences of the four Polygonum species were deposited in NCBI under the Genbank accession number MZ748474–MZ748477.

Tandem Repeats Finder (Benson 1999) and REPuter (Kurtz et al., 2001) have been used to find the tandem, forward, and palindromic repeats from the four Polygonum chloroplast genomes. The Misa.pl (Beier et al., 2017) was used to recognize the SSRs and the finding parameters of mononucleotides transfer to eight repeatable elements, dinucleotides, and trinucleotides four repeatable elements, tetranucleotides, pentanucleotides, and hexanucleotides transfer to three repeatable elements. Primer3 (Untergasser et al., 2012) was used to design the SSR primers.

Genome structures among six Polygonum chloroplast genomes, Including four polygonum species in this study and two polygonum species published by the NCBI (P. cuspidatum and P. chinense), were completed by mVISTA software (Shuffle-LAGAN mode) (Frazer et al., 2004) using the genome of P. cuspidatum as the reference. Pi values and sequence polymorphisms of six Polygonum species were analyzed using DNAsp v. 6.12.03 (Rozas et al., 2017). The step size was set to 200 bp, and the window length was set to 800 bp.

A total of 19 chloroplasts sequences (Supplementary Table S4) were used to build the phylogenetic trees. Each of the 67 protein-coding genes shared by all the genomes was compared individually and then linked end to end to form a supergene from each species. The sequences alignment was carried out using the MAFFT v7.309. The best model was determined using the modeltest-ng-0.1.6 software with default parameters; ML analysis was performed using RAxMLNG v0.9.084 based on Linux edition using default parameters. The parameters were GTR + FU + IU + G4m, noname = 1–51,039. Chrysanthemum x morifolium has been situated likewise those out-groups.

The genome sizes of the four Polygonum chloroplasts are 163,461 bp (P. aviculare), 159,476 bp (P. bistorta), 159,015 bp (P. orientale), and 160,680 bp (P. perfoliatum), respectively. The whole GC content are 37.43% (P. aviculare), 37.37% (P. bistorta), 38.21% (P. orientale), and 37.96% (P. perfoliatum), individually. The LSC region, SSC region, and a couple of inverted repeat regions (IRA/IRB) are alike in Polygonum chloroplast genomes than other plants (Guo et al., 2018). The length of the LSC region is 83,583–88,021 bp and the GC content is 35.48–36.59% in Polygonum chloroplast genomes. The distribution of length in the SSC regions is 12,928–13,306 bp and the GC content is about 32.46–33.19%. The GC content in those IR regions is 41.27–41.45% and the length is 31,067–31,184 bp (Table 1). The GC content is a significant marker to identify the genetic relationship of species; moreover, the Polygonum has comparable cpDNA GC content. The phenomenon is also common in other plants (Guo, Guo, Zhao, Xu, Li, Zhang, Shen, Wu and Hou 2018; Liang et al., 2019) that the GC content in those IR regions is more than that of other regions (LSC, SSC). The high GC content of the IR areas usually points to the rRNA and tRNA genes (He et al., 2016; Shen et al., 2017). Also, the whole chloroplast genome sequences of four Polygonum species can be checked in the National Center for Biotechnology Information (NCBI) database afterward annotation and the GenBank accession number can be found in Table1.

The annotation results (GB files) of four Polygonum chloroplast genomes which were measured in this study were submitted to the OGDraw software, and the physical map of the Polygonum chloroplast genomes were drawn. The results can be found in Figure 1. In total 112 genes were found in the chloroplast genomes, such as four rRNA genes, 30 tRNA genes, and 78 protein-coding genes (Figure 1; Table 2). The main genes of the four Polygonum chloroplasts can be coarsely divided into three classes, named chloroplast self-replication-related genes, photosynthesis-related genes, and other genes (Saski et al., 2005).

There are 16 genes with introns among the 112 genes of four Polygonum chloroplast genomes (Table 3), with 5 tRNA genes and 11 functional genes. The tRNA genes include trnI-GAU, trnL-UAA, trnV-UAC, trnK-UUU, and trnA-UGG. The 11 functional genes include ndhB, ndhA, atpF, petB, petD, rpoC1, ycf3, clpP, rps12, rpl16, and rpl2. The 5’ termination of the rps12 gene is in the LSC region of the chloroplast genome, and the 3’ end is in the IR region of the chloroplast genome. Like the other angiospermous chloroplasts trans-splicing phenomenon also occurs in the rps12 gene of the Polygonum chloroplast genome. Three of the 17 intron-containing genes cover two introns (rps12, ycf3, and clpP), and the other genes only contain one intron, of which trnK-UUU covers the main intron (2,510 bp), and this intron covers the entire gene of matK.

Relative synonymous codon usage (RSCU) is a synonymous codon correlative effect, which values the 64 vital synonymous codons (Wu et al., 2007). RSCU is calculated as the ratio of the actual observed value to the average usage of the synonymous codons. The value of RSCU can be divided into three types: greater than 1, less than 1, and equal to 1. If the value of RSCU is greater than 1, it indicates that the codon is used more frequently than other codons. If the value of RSCU is less than 1, it means that other synonymous codons of this codon are used more frequently than this codon. If the value of RSCU is equal to 1, it indicates that there is no bias in the use of a codon. According to the statistical study of four Polygonum chloroplast genomes, the extent of CDS is from 80,286 to 83,403 bp, and for which accounts for about 50% of the total chloroplast genome length. The number of codons is in between 26,762–27,801. As for the statistical study of RSCU, there was some certain bias in the use of other amino acids, except for Trp and Met (Tables 4).

In this research, we also studied the repeated sequence of four Polygonum chloroplast genomes, with tandem repeats (T), forward repeats (F), reverse repeats (R), and palindromic repeats (P). The results of the repeated study of four Polygonum chloroplast genomes are in Figure 2. There are 99 repeated sequences in P. aviculare, including 49 tandem repeats, 22 forward repeats, 26 palindromic repeats, and 2 reverse repeats; 86 repeated sequences in P. bistorta, including 36 tandem repeats, 23 forward repeats, 22 palindromic repeats, and 5 reverse repeats; 72 repeated sequences in P. orientale, including 22 tandem repeats, 22 forward repeats, 22 palindromic repeats, and 6 reverse repeats; 76 repeated sequences in P. perfoliatum, including 27 tandem repeats, 20 forward repeats, 22 palindromic repeats, and 7 reverse repeats. Among all the categories of repeated sequences, the sequences of length 20–50 bp are the most (Figure 2).

Simple sequence repeat (SSR), also known as microsatellite sequence, is a repeat sequence composed of one to six bases as repeat units in series, which is of great significance to the study of plant populations. SSRs with a length of over 10 bp are inclined on slipped-strand mispairing, which is approved to be the principal mutational mechanism of SSR polymorphisms (Asaf et al., 2017). Additionally, SSRs which are changeable at the intraspecific position in the chloroplast genome are often used as the genetic marker in the investigation of population genetics and evolution (Yang et al., 2016; Zl et al., 2016). There are 228 SSRs in P. aviculare, including, three trinucleotides, 57 dinucleotides, 159 mononucleotides, 6 tetranucleotides, and 3 pentanucleotides; 225 repeated sequences in P. bistorta, including 7 tetranucleotides, 8 trinucleotides, 161 mononucleotides and 49 dinucleotides; 181 repeated sequences in P. orientale, including 136 mononucleotides, 38 dinucleotides, 3 trinucleotides and 4 tetranucleotides; 204 repeated sequences in P. perfoliatum, including 152 mononucleotides, 44 dinucleotides, 4 trinucleotides, and 4 tetranucleotides (Figure 3). Based on these SSR results, we designed 10 pairs of primers as molecular markers. P. orientale, P. cuspidatum, and P. aviculare were used for molecular marker amplification. The results showed that four pairs of primers could distinguish well between species (Supplementary Figure S1). Although the four of them can be distinguished at the genus level, the PCR amplification products may be chloroplast DNA or nuclear DNA, which requires further study.

The chloroplast genome sequence of P. cuspidatum is used as a reference sequence to draw an analogy among the genomic sequences of six Polygonum chloroplast genomes (P. cupidatum, P. chinense, P. aviculare, P. bistorta, P. orientale, and P. orientale). Figure 4 displayed regions in common. The figure displayed that, aiming at the Polygonum chloroplast genomes; the rate of changes for the LSC and SSC regions is visibly greater than that of the IR region. To further clarify the variation in the coding regions, the Pi (nucleotide diversity) was also calculated (Figure 5). Six divergent loci (psbI-trnS-GCU, rpoB-trnC-GCA, trnE-UUC-trnT-GGU, trnT-GGU-psbD, trnT-UGU-trnL-UAA, and rpl32-trnL-UAG) had a Pi value greater than 0.12. All of these six divergent loci were intergenic regions and were present in the LSC region, except for rpl32-trnL-UAG, which occurred in the SSC region, with none being detected in the IR region. These highly variable regions may also resolve the interspecific relationships of Polygonum in the Polygonaceae phylogeny. With the forward progress of the chloroplast genome because of the wide usage of the chloroplast DNA fragments, some conundrums in plant species authentication, phylogenetics, and other related researche studies can be solved. Obviously, the chloroplast DNA fragments have higher resolution loci than general DNA fragments which have low distinguishability and mutation rates in some proximal related groups. Meanwhile, these high-efficiency DNA fragments from the chloroplast genome can promote the development of species identification and population diversity by comparing the differences between the chloroplast genome sequences of different groups.

IR Contraction and Expansion in Polygonum Chloroplast Genome

In this research, the analysis of IR-LSC and IR-SSC border structure and location of four Polygonum species was finished (Figure 6). The results found out that the SSC/IRa assembly was located in the ndhF region in the four species of Polygonum chloroplast genome, and spread a length of 60–63 bp into the IRa region in the four species. Also, the rps19 gene was located in the IRa region and the length is about (P. orientale, 21 bp; P. perfoliatum, 40 bp) into the LSC region. The ycf1 gene of four Polygonum chloroplast genomes completely exists in the IRb region, with a terminal 225 bp (P. aviculare), 275 bp (P. bistorta), 261 bp (P. orientale), and 257 bp (P. perfoliatum) from the SSC/IRa border. For now, the trnH gene existed in the LSC region, and it had a length of 2, 1, 3, and 5 bp from the LSC/IRb border in the chloroplast genome of P. aviculare, P. bistorta, P. orientale, and P. perfoliatum, individually.

Until now, the expansion mechanism of the IR region has been debated; moreover, the double-strand break repair (DCBR) theory is written off as the prime mechanism for the expansion of the IR region (Machour and Ayoub 2020). There is little probability of the IR region, which has a large shrink. Furthermore, it is believed that the DCBR model is not only the core mechanism of IR region expansion but also the mechanism of IR region contraction.

Phylogenetic analysis was accomplished on an alliance of concatenated nucleotide sequences of all genes from 19 angiosperm species. We used ML to construct a phylogenetic tree ground based on these gene data, while C. morifolium was agreed as the outgroup. The P. bistorta, P. orientale, P. perfoliatum, and P. chinense are divided into a branch, namely, Polygonum. And in the branch of Polygonum, it is further divided into two small sub-branches. The upper sub-branch includes P. orientale, P. perfoliatum, and P. chinense. Moreover, P. orientale and P. perfoliatum clustered on the small branch. The P. aviculare is closer to F. sachalinensis, F. aubertii, and Polygonum cuspidatum, which is far away from the branch of Polygonum. Therefore, we put P. aviculare and P. cuspidatum into the branch of Fallopia (Figure 7).