The I. pes-caprae L. plants used for this study were naturally growing on the beach located in Changle, Fuzhou (Latitude 25°, 54’, 33˝ N, Longitude 119°, 40’, 42˝ E), Fujian, China. The leaf materials were used for DNA extraction using the modified CTAB method (Li et al., 2013).

The genome DNA samples were subject to SMART laboratory construction and then sequenced on the PacBio Sequel II sequencing platform. The CCS (Circular Consensus Sequence) reads corresponding to the chloroplast genome were extracted by mapping all the reads to the chloroplast genomes of all the Convolvulaceae species with Bowtie 2 (Langmead and Salzberg, 2012). Subsequently, the resulting CCS reads were considered to be derived from the chloroplast genome and were used for chloroplast genome assembly using Canu (V2.2) software (Nurk et al., 2020). The complete chloroplast genome sequences were annotated using the program PGA (Qu et al., 2019). Both tRNAs and rRNAs were identified by BLASTN and BLASTP by searching against the references composed of all the available chloroplast genomes of Convolvulaceae species. The annotation results were checked manually, and the codon positions were adjusted by comparing them to a previously homologous gene from various chloroplast genomes. The circular chloroplast genome map was drawn by the program Chloroplot (Zheng et al., 2020). The assembly and annotation of the chloroplast genome were submitted to NCBI (Accession No: MZ557416).

Repeat elements in chloroplast genomes of I. pes-caprae were investigated using two different programs. The program MicroSAtellite identification tool (Katoh et al., 2002), (Beier et al., 2017) was used to identify the SSR repeat, setting the parameters with thresholds of 10, 5, 4, 3, 3, and 3 repeat units for mono-, di-, tri-, tetra-, penta-, and hexa-nucleotides, respectively. The program REPuter was used to detect the repeat sequences within the chloroplast genome (Kurtz et al., 2001). Four types of repeats, including forward repeats, reverse repeats, complement repeats, and palindromic repeats, were investigated in this analysis.

The chloroplast genome sequences of 26 Convolvulaccae species, including 14 Cuscuta species and 12 Ipomoea species, were used for phylogenetic analysis using three model species as outgroups (Arabidopsis thaliana, Amborella trichopoda, and Oryza sativa). The taxonomical details of the investigated species and the accession numbers of their chloroplast genome are listed in Supplementary Table 1 . To accurately identify the phylogenetic position of I. pes-caprae, two phylogenetic trees were constructed based on the complete chloroplast genome sequences and the protein sequences, respectively. The alignment was conducted by MAFFT (Katoh et al., 2002), and the phylogenetic trees were constructed by MEGAX using Maximum Likelihood methods (Kumar et al., 2018).

To investigate the sequence divergence among Ipomoea species chloroplast genomes, The chloroplast genome of I. pre-caprae generated in this study, together with 11 released Ipomoea chloroplast genomes retrieved from NCBI, were used for comparative analysis. The sequences were aligned using the mVISTA program with Shuffle-LAGAN mode (https://pgrc.ipk-gatersleben.de/misa/) (Frazer et al., 2004).

To identify the SNPs and hypervariable regions within the chloroplast genome of I. pes-caprae in comparison with other Ipomoea species, the chloroplast genomes of other Ipomoea were aligned to the chloroplast genome of I. pes-caprae using MAFFT (Katoh and Standley, 2013). The nucleotide diversity (Pi) along the chloroplast genome was calculated using DnaSP version 5 software (Librado, 2009) with sliding window analysis. The window length was set to 800 base pairs, and the step size was set to 50 base pairs.

Using the PacBio Sequel II sequencing platform, 23252902 whole-genome long reads of I. pes-caprae were produced for genome assembly (unpublished project), and the mean read length is around 14 kb. Converted raw reads to CCS reads of the whole genome, the reads number from 23252902 to 1467275 ( Supplementary Table 2 ). The chloroplast genome size of 161667 bp of I. pes-caprae was derived from the Canu program assembly. The I. pes-caprae complete chloroplast genome had a typical circular structure and with typical quadripartite organization consisting of the four conserved constitute regions, a pair of 30670 bp inverted repeats (IRs 30670 bp), an 88210 bp long single-copy regions (LSC), and a 12117 bp short single copy region (SSC) ( Figure 1 ; Supplementary Table 3 ).

A number of 136 functional genes were identified in the chloroplast genome of I. pes-caprae, including 86 protein-coding genes, 42 tRNA genes, and 8 rRNA genes (16S, 23S, 5S, and 4.5S) ( Figure 1 ; Table 1 ). According to their functions, the 136 I. pes-caprae chloroplast genes were assigned into three categories: Most protein-coding genes are related to photosynthesis. There are 45 genes in this category, including genes encoding subunits of ATP synthase, subunits of photosystem II, subunits of cytochrome b/f complex, subunits of NADH-dehydrogenase, assembly/stability of photosystem I, subunits of photosystem I, subunits of photosystem II, cytochrome c synthesis (ccsA), photosystem biogenesis factor 1 (pbf1), and subunit of rubisco (rbcL). The second category consists of 86 genes associated with chloroplast transcription and translation, including protein-coding genes encoding the large subunits of the ribosome, the small subunits of ribosomal proteins, DNA-dependent RNA polymerase, and translational initiation factor, and two types of non-coding genes, i.e., rRNA genes, tRNA genes. And the rest of the four genes were classified as other genes because of their unique or unknown function, including matK with function in RNA processing and 3 conserved open reading frames (ycf1 and ycf2×2) encoding proteins of unknown function ( Table 1 ; Supplementary File 3 ).

The sequence repeats widely exist in eukaryotic genomes, and the simple sequence repeat (SSR) is the most abundant and typical repeat type. SSRs in chloroplast genomes exhibit high copy numbers, which play an essential role in genome rearrangement and recombination and are important molecular markers in plant phylogenetic and evolutionary studies (Kuang et al., 2011; Yang et al., 2011; Jiao et al., 2012; Xue et al., 2012). From the perspective of evolution, the differences in repetitive sequences among species resulted from natural selection (Huang et al., 2021). 35 SSRs were identified in the chloroplast genome of I. pes-caprae ( Supplementary Table 4 ): 32 are A/T single nucleotide repeats, and one is the TA dinucleotide repeat. Notably, there was no di, tri-, tetra-, penta-, or hexanucleotide repeat detected in the chloroplast genome of I. pes-caprae. The longer sequence repeats within the chloroplast genome were identified using Reputer 2.0 software. As a result, 50 repeats consisting of 26 forward and 24 palindromic repeats were obtained, while no complement or reverse repeats were detected ( Supplementary Table 5 ). The largest repeat unit with a size of 30670 bp is the inverted repeats of chloroplast, which is essential for the chloroplast structure organization. And the lengths of the rest of the repeat units ranged from 119 and 242 bp. The repeat information of the chloroplast genome of I. pes-caprae is valuable for developing genetic markers for phylogenetic and population studies (Nie et al., 2012).

In evolution, species are affected by natural selection pressure and genetic drift, resulting in differences in the use frequency of most genetic codons. Thus, the different genomes might have specific codon preferences (Hershberg and Petrov, 2008). In the complete chloroplast genome of I. pes-caprae, There are 53889 codons within the protein-coding genes. The I. pes-caprae chloroplast genome encoded all 20 amino acids, and 61 types of amino-acid codons were observed ( Supplementary Table 6 ). The UU-started codons are found to be more frequent than the others ( Supplementary Table 7 ). Among the 20 amino acids, Leucine was the most abundant (number of codons encoding Leucine = 5708, 10.59%), Serine was the second abundant (number of codons encoding Serine = 4961, 9.21%), while the rarest one is Tryptophan (713 codons, approximately 2.12%). Thirty-three codons were observed to be used more frequently than the expected usage at equilibrium (RSCU (Relative Synonymous Codon Usage) > 1), and 27 codons showed the codon usage bias (RSCU < 1). However, the frequency of use for the AUG (Methionine, start codon), UGG (Tryptophan), ACC (Threonine), and GGU (Glycine) showed no bias (RSCU = 1). In aspects of amino acids, most of them have codon preferences, except for Methionine and Tryptophan ( Figure 2 ). And most of the 20 amino acids with at least two codons, and the acids Arginine, Leucine, and Serine have six codons. To investigate whether the codon usage preferences are conserved among the species in the Ipomoea genus, the protein-coding genes of the other 11 Ipomoea species, including I. batatas (53822 amino acids), I. hederifolia (53791), I. involucrata (53506), I. minutiflora (53729), I. murucoides (53327), I. nil (53965), I. obscura (53749), I. purpurea (54015), I. tricolor (53589), I. trifida (53711), and I. triloba (53916) were investigated. As shown in Supplementary File 1 , the codon usage preferences of all the 11 investigated species showed a similar tendency as that of I. pes-caprae ( Supplementary File 1 ), indicating the conservation of codon bias in the genus of Ipomoea, which might be because that the ancestors of those species underwent the shared evolutionary history.

The ENc plots are usually used to indicate the factors that affect the codon bias. To understand the relative importance of natural mutation and evolutionary selection in producing codon usage patterns, ENc (Effective Number of Codons) values of four Ipomoea species (I. pes-caprae, I. involucrata, I. murucoide, and I. tricolor) were estimated and plotted against the GC3s values ( Figure 3 ). From the ENc plots, it is clear that the protein-coding genes of the four species showed similar codon bias patterns. Most of the genes were distributed on both sides of the standard curve, and more than half of the genes were below the curve, suggesting that the selection pressures predominantly influence codon bias in the chloroplast genome of the Ipomoea species. Photosynthesis-related genes are distributed most discretely, suggesting that some other factors might influence the codon bias, or maybe these genes are more conserved than self-replication and other group genes.

Expansion and contraction of IR regions are common events that frequently happen in the evolutionary history of land plants (Raubeson et al., 2007; Kode et al., 2010; Yao, 2015). LSC/IR and IR/SSC junctions are sometimes regarded as an index of chloroplast genome evolution. The map illustrating the LSC/IRb (JLB), IRb/SSC (JSB), SSC/IRa (JSA), and IRa/LSC (JLA) boundaries/junctions of the 12 Ipomoea chloroplast genomes was constructed according to heir chloroplast genome sequences and annotation information ( Figure 4 ). It is shown that the chloroplast genome organizations are highly conserved across the 12 Ipomoea species with only minor variations. The chloroplast genomes of 12 species have similar sizes ranging from 160072 bp (I. murucoides) to 161897 bp (I. nil). However, the sizes of LSC, SSC, and IR are varied significantly. For example, The lengths of IRs ranged from 8265 bp (I. tricolor) to 31061 bp (I. triloba), and the size of IR of I. pes-caprae was 30670 bp. The LSC/IR and SSC/IR boundaries in all 12 Ipomoea species were distributed with different genes, and only a few chloroplast genomes with consensus genes. The RS lines (the boundary lines between IRb/IRb and SSC) were mainly located between ycf1 and trnN.

On the contrary, the RL lines (the boundary lines between IRa/IRb and LSC) were more variable than RS lines. Compared to species of other genera, the LSC regions of I. involucrata, I. murucoides, I. obscura, and I. tricolor have the expansion of ycf2 to the IR LSC region, decreasing IR length. On the other hand, the SSC Regions of I. pes-caprae, I. nil, I. purpurea, and I. triloba were concentrated compared to those of other species, with the loss of ycf1 in the SSC regions. As previous studies suggested, the expansion and contraction detected in the IR regions might be a primary mechanism in the length variation of three regions (LSC, SSC, and IR) of the chloroplast genomes in Ipomoea species (Sun et al., 2019).

In order to understand the evolutionary relationships of I. pes-caprae with the Ipomoea genus, the chloroplast genomes of 28 Convolvulaccae species and three outgroup species (A. thaliana, A. trichopoda, and O. sativa) were used for phylogenetic analysis. The complete chloroplast genomes were downloaded from NCBI (National Center for Biotechnology Information database) and used for constructing the phylogenetic trees. We first constructed the phylogenetic trees based on the complete chloroplast genome sequences ( Figure S1 ). The Convolvulaccae species were classified into the Ipomoea clade and the Cuscuta clade. The phylogenetic tree generated based on completed chloroplast sequences showed consistent relationships with taxonomical classification, suggesting confidence in the phylogenetic analysis based on the chloroplast genome. However, one of the outgroups species, A. trichopoda, the base angiosperm, was not in the basal position in the phylogenetic tree. One reason might be because of the sequence variation of the non-coding sequences. To this end, the sequences of orthologous single-copy genes were used to construct a maximum likelihood tree. As shown in Figure 5 , the Basal Angiosperms, Monnocotes, and Diocts outgroups species showed a clear taxonomical relationship, and the 12 ipomoea species and 14 Cuscuta species were clustered into two clades. Our results showed that the chloroplast genomes could be used for constructing the phylogenetic tree reflecting the evolutionary relationships of land plants, and the orthologous single-copy genes sequence-based phylogenetic relationships based would be much more confident than the complete chloroplast genome sequences-based phylogenetic relationships.

Cuscuta spp plants are annual parasitic herbs, and most of them are leafless, lack chlorophyll, and have a wide range of hosts (Albert et al., 2008). By comparing the protein-coding genes between Cuscuta and Ipomoea species, it’s evident that many genes were lost in Cuscuta species ( Supplementary Table 8 ). In Cuscuta species, only C. exaltata maintains a few ndh (Subunits of NADH-dehydrogenase) genes. The other 13 Cuscuta species missed all of the ndh genes related to the photosynthesis pathway ( Figure 6 ; Supplementary File 2 ). The result of the comparative analysis showed that C. boldinghii, C. erosa, and C. strobilacea have the least number of genes, which are 31, 33, and 33, respectively, and C. exaltata has the most number of genes, which is 67. Except for C. exaltata, other Cuscuta species don’t have complete photosynthesis-related genes, such as the genes coding subunits of cytochrome b/f complex, subunits of photosystem I, and subunits of photosystem II. Therefore, it is concluded that the loss of genes involved in photosynthesis in Cuscuta chloroplast genomes happened gradually, and Cuscuta species lost the photosynthesis ability to various extents, which is consistent with their nutritional performance.

We also analyzed their chloroplast genomes differences to explore further the genetic relationship among the I. involucrata, I. murucoides, and I. tricolor. The mVISTA program was used for global alignment to exhibit the variation of the chloroplast genomes in different regions using the genome sequence and annotation of I. pes-caprae chloroplast as the reference ( Figure 7 ). The gene organization was highly conserved across the four chloroplast genomes, with few variation regions, consistent with previous studies (Nguyen et al., 2021). The results also exhibited that the divergences in LSC and SSC regions were higher than in IR regions. Besides, the sequences in the coding regions tended to be more conserved, whereas most of the variations detected were found in conserved non-coding sequences (NCS). The sequences of exons had almost 100% similarity throughout the four taxa. Among the coding genes, the highly disparate sequences are the regions harbouring rpl2, ycf2, ndhK, ndhD, and ycf1 genes.

To visualize the overall sequence divergence of the 12 Ipomoea species, the pi values of the chloroplast genome sequence were calculated with a slide window length of 600 bp and a step size of 100 bp. The sliding window pi plots showed that the average pi value of the LSC (Pi = 0.008101) and SSC (Pi = 0.054394) regions was much higher than that in the IR (Pi = 0.001942) regions, which showed that LSC and SSC regions contained the most of the variation ( Figure 8 ). The plots also showed that the SSC region is hypervariable. Further investigation revealed that most of the SSC region of I. pes-caprae was much shorter than other Ipomoea species, indicating an invasion of IRs to this region. These regions with higher pi values are more variables that might experience rapid nucleotide substitution and could be used to develop molecular markers for identification and phylogenetic analysis (Lyu et al., 2020).