Background

Caragana Fabr. comprises over 100 species and belongs to the family of Fabaceae. These plants are mainly distributed in arid and semi-arid regions of Asia and Europe. Of these species, 66 were found in China, 32 of which are endemic [1]. Caragana plants are renowned for their drought, infertile conditions, cold and heat tolerance [2]. They are widely cultivated due to their ability to adapt to dry conditions [3]. Similar to other Fabaceae family plants, these plants can convert atmospheric nitrogen into usable nutrients via nodules on their roots, playing a role in rejuvenating infertile soils, combating dust storms, and hindering desertification [4]. The distribution of various Caragana plants in China has been extensively studied (Table 1).

Table 1 Introduction to 9 species of the genus Caragana
Full size table

Additionally, previous studies have shown that this genus includes more than 10 plants with excellent pharmacological properties. These plants have been utilized to treat various diseases such as fever, inflammation, wound infections, headaches, rheumatoid arthritis, and cancer [4, 14, 15]. The C. tibetica studied in this article is mentioned in both Mongolian medicine and Tibetan medicine as a potential treatment for rheumatoid arthritis, wounds, hypertension, and anemia [4].

The current research indicates a limited availability of data on plants in Caragana, with only 14 chloroplast (cp) genomes reported. However, the evolutionary analyses using nuclear ITS (Internal Transcribed Spacer) and plastid marker sequence data (matK, trnL-F, and psbA-trnH) for studying the phylogenetic relationships of Caragana plants lack resolution, leaving unanswered questions about the classification of certain medicinal plants like Caragana changduensis Liou f., Caragana frutex (L.) C. Koch, and Caragana polourensis Franch. [1, 16,17,18,19]. These plants, representing different species within Caragana, exhibit varying morphological features in terms of flowers, leaves, stems, and other aspects. Additionally, their habitat preferences contribute to morphological adaptations and variations, as they inhabit diverse ecological environments. Thus, relying solely on morphology for identifying different Caragana species may introduce errors and uncertainties. Consequently, finding an accurate and convenient method for plant identification in Caragana is crucial.

According to reports, researchers have gained a deeper understanding of chloroplasts, including their origins, structures, evolution, and genetic engineering [20, 21]. Chloroplasts contain their genetic system [22], and most plants have chloroplasts existing in the form of covalently closed circular DNA [23]. The rapid development of sequencing technologies has led to the discovery of more efficient molecular markers within the chloroplast genome, which are advantageous for accurate species identification. The chloroplast genome is an ideal choice for molecular identification, phylogenetic analysis, and species conservation research according to a previous study [24]. Unlike the nuclear genome, the chloroplast genome is particularly valuable for plant phylogenetic studies due to its unique features: it is typically inherited from only one parent, had a simpler structure, and contains multiple copies of each gene [23, 25]. The plastid chromosome, which is circular and has a length of 120 ~ 160kb [25], consists of four regions containing two inverted repeat regions (IRs). These regions separate the large single copy region (LSC) and the small single copy region (SSC) [26]. Due to their high level of conservation and relative small size, plastid structure and gene content is easy to obtain completely and worth studying in species identification, population genetics, and phylogenetics [27, 28]. Currently, various plants, such as Desmodieae, Picea, and Epimedium [29,30,31], utilize the chloroplast genome to study their phylogenetic relationships. Reports have indicated the occurrence of inverse repeated loss of the evolutionary branch (IRLC) in Fabaceae [32,33,34,35,36], including the 8 species of Caragana with IRLC that have been reported [1, 16, 17, 37]. Therefore, it will be possible to study Caragana as a lineage representing a broad spectrum of IRLC with the improvement in the chloroplast genome data of Caragana.

In this study, we compared the complete chloroplast genomes of C. tibetica and C. turkestanica to those of other species within the Caragana genus. Additionally, we analyzed the structural characteristics and phylogenetic relationships of these chloroplast genomes with other species within the Fabaceae family. The results of this study have advanced the knowledge of chloroplast genome data within the genus Caragana, providing valuable insights for species identification, systematic evolutionary studies, and germplasm conservation and utilization.

Results

Characteristics of Caragana chloroplast genomes

Draw gene maps using OGDRAW for C. tibetica (Fig. 1A) and C. turkestanica (Fig. 1B) based on the annotation results of their chloroplast genomes. The chloroplast genomes sizes of C. tibetica and C. turkestanica were found to be 128,433 and 129,453bp, respectively. With the loss of the IR region in the two plants, they do not have the typical quadripartite structure found in most flowering plants’ chloroplast genomes, and their lengths were accordingly shorter. Nevertheless, the cp genome structures, gene contents and direction were strongly comparable (Fig. 1A and B).

Fig. 1
figure 1

The diagram illustrates the chloroplast gene maps of C. tibetica (A) and C. turkestanica (B). The genes located on the outer circle are transcribed counterclockwise, whereas those on the inner circle are transcribed clockwise. Different functional gene groups are represented by different color codes. In addition, changes in GC content are represented by light gray in the inner circle, while changes in AT content are represented by dark gray

Full size image

The annotation results of the chloroplast genomes revealed that C. tibetica had a total of 110 specific genes in chloroplast genome, while C. turkestanica shared totally 111 specific genes, comprising 76 protein coding genes, 31 (30) tRNA genes and 4 rRNA genes (Table 2). The GC contents of the two species were very similar, with values of 34.30% and 34.71%, respectively. Seven cp genomes in the Caragana species (C. arborescens, C. opulens, C. jubata, C. rosea, C. microphylla, C. kozlowii, C. korshinskii) with missing IR regions were compared with C. tibetica (128,433 bp) and C. turkestanica (129,453 bp). The results revealed that the total sequence lengths ranged from 128,132 to 133,122 base pairs. The deletion of the IR region resulted in the shortest chloroplast genome length of 128,132 bp in C. jubata and the longest in C. rosea (133,122 bp). In addition, the number of genes in C. turkestanica, C. arborescens, and C. opulens was one more than that in other species (tRNA encoded by trnN-GUU gene). At the same time, the number of protein-encoding genes and ribosomal RNA genes was consistent among the nine plants.

Table 2 Summary of complete chloroplast genomes for nine Caragana species
Full size table

From the standpoint of gene contents, the nine plants had the highest number of protein-encoding genes, which accounted for approximately half of the entire genome length. Following the most abundant genes were tRNA genes, which were shorter in length than other genes. In general, the sequence length and gene content of chloroplast genomes in the nine Caragana species were roughly consistent. We also analyzed the difference in GC content among the three types of genes. The GC content of rRNA genes was highest, exceeding 50% and were consistently so. GC content of tRNA genes were next in amount. The lowest GC content was observed for protein-coding genes, which was approximately 37%. Moreover, the average GC content of the nine species were around 34%, which suggested that the sequence of Caragana species was relatively conserved during the process of evolution.

The genes encoded by the chloroplast genomes of C. tibetica and C. turkestanica can be divided into three categories, similar to other species. There were 57 genes related to self-replication, including ribosomal RNA, transfer RNA, and three subunits (large, small, and DNA-dependent RNA polymerase) that encode chloroplast RNA polymerase; 44 photosynthesis-related genes; the remaining genes were categorized as other genes and unknown genes. In the chloroplast genomes of C. tibetica and C. turkestanica, 17 genes with introns were detected. Thereinto, C. tibetica had two genes (clpP and ycf3) with two introns, while C. turkestanica only had one gene (ycf3) with two introns, and the remaining 15 genes (rpl16, rpl2, rps12, rpoC1, trnA-UGC, trnG-UCC, trnI-GAU, trnK-UUU, trnL-CAA, trnV-UAC, ndhA, ndhB, petB, petD, atpF) had only one intron (Fig. 2, Table 3).

Fig. 2
figure 2

Gene contents of C. tibetica and C. turkestanica chloroplast genomes. The color of each gene is unique, and the horizontal axis indicates that each box is proportional to the size (bp) of the gene

Full size image
Table 3 Genes in the chloroplast genome of Caragana species
Full size table

Analyses of repeats and simple sequence repeat (SSR)

Repetitive units played a critical role in evolution of the genome by facilitating genetic variation and diversity. Through size evolution and structural rearrangements, they promoted genomic mutations and diversity, providing the opportunity for organisms to adapt to new environments and develop new functionalities [38,39,40]. In the present study, we identified the repetitive sequences present in the cp genomes of C. tibetica and C. turkestanica, and analyzed content of the two plants. In the genomes of C. tibetica and C. turkestanica, a total of 119 (length range: 30–337 bp) and 128 (length range: 30–249 bp) repetitive sequences were identified, respectively, consisting of forward (F), palindromic (P), reverse (R), and complementary (C) repeats (Additional file 1: Table S1). In C. tibetica and C. turkestanica, repetitive analysis detected 88 and 84 forward repeats, 30 and 36 palindromic repeats, 1 and 7 reverse repeats, and 0 and 1 complementary repeats, respectively (Fig. 3A). Among all types of repeats, the frequency of occurrence was highest for sequences with a length ranging from 30 to 49 base pairs (bp). In C. tibetica, there were 44 forward repeats, 27 palindromic, and 1 reverse repeats with lengths ranging from 30 to 49 base pairs (Fig. 3B-D). Similarly, in C. turkestanica, 53 forward repeats, 27 palindromic, 7 reverse repeats and 1 complementary repeats were 30–49 bp in length (Fig. 3B-D).

Fig. 3
figure 3

Repeat sequences analysis of 9 Caragana cp genomes. A The total number of four types of repeat sequences in 9 Caragana species; B The frequency of forward repeats by length; C The frequency of palindrome repeats by length; D The frequency of reverse repeats by length

Full size image

In addition, 129, 229, 80, 259, 178, 380, and 127 repeats were found in the reported C. arborescens, C. opulens, C. jubata, C. rosea, C. microphylla, C. kozlowii, and C. korshinskii cp genomes, respectively (Additional file 1: Table S1, Fig. 3A). This finding suggested that C. tibetica and C. turkestanica have a higher degree of similarity in repeat frequency with C. arborescens and C. korshinskii.

Simple Sequence Repeat (SSR) loci exhibit extensive and highly variable polymorphism within the genome. As a result, SSRs are considered as effective molecular markers for investigating genetic variations and individual genetic relationships within the genome [41,42,43]. In this study, we identified intact SSRs in the chloroplast genomes of C. tibetica and C. turkestanica together with seven additional Caragana species (Fig. 4A-C). Based on the propensity of SSRs with 10 bp or longer to undergo slippage and mismatch in the DNA chain, specific parameters have been set to address this phenomenon, which is considered the primary mechanism for SSR polymorphism [43]. We detected a total of 27 types in the two Caragana plants, using software MISA (Fig. 4A). Among these, 239 and 277 SSRs loci were detected in C. tibetica and C. turkestanica. Similarly, we found 277 SSRs in C. arborescens, 265 SSRs in C. opulens, 281 SSRs in C. jubata, 261 SSRs in C. rosea, 275 SSRs in C. microphylla, 287 SSRs in C. kozlowii, and 279 SSRs in C. korshinskii (Additional file 1: Table S2). The SSRs in these chloroplast genomes were mainly composed of mononucleotide and trinucleotide repeats motifs. The mononucleotide repeats (A/T and C/G) varied from 150 (62.76%) in C. tibetica to 168 (63.64%) in C. opulens, while varying from 68 (28.45%) in C. tibetica to 86 (31.16%) in C. microphylla for trinucleotide repeats (AT/AT and AG/CT) (Fig. 4B, Additional file 1: Table S2). Among them, A/T repeat sequences were the most abundant SSRs. There were 146 and 158 SSRs containing A or T, respectively, in the sequenced species, while only 1 contains C or G.

Fig. 4
figure 4

Statistics of SSRs detected in the plastome of nine Caragana species. A Number of SSRs determined in different repetition types; B The amount of different SSR types found in nine Caragana species genomes; and C The number of SSRs were found in coding (CDS), and intronic regions, intergenic (IGS), Respectively

Full size image

In addition, there were two pentanucleotide repeats in C. jubata, with one present in C. opulens, C. microphylla, C. kozlowii and C. korshinskii, one hexanucleotide repeats was found in C. jubata, C. rosea and C. kozlowii using our identification criteria (Fig. 4B, Additional file 1: Table S2). Furthermore, we analyzed the distribution of SSRs in the coding and non-coding regions. The results showed that the number of SSRs in protein-coding regions was significantly lower compared to the non-coding regions (Fig. 4C). We have discovered that the clpP gene in C. tibetica contains the longest simple repeat sequence, which was a mononucleotide repeat sequence with a length of 49 bp, whereas, the longest SSR was found on the ycf1 gene in C. turkestanica, and it was a single nucleotide repeat sequence with a length of 46 bp ( Additional file 1: Table S3).

Codon usage bias analysis

During the process of biological evolution, there is a widespread codon usage bias observed in plastids. By analyzing codon usage bias, it is possible to uncover phylogenetic relationships between organisms and molecular evolution of genes, thereby providing potential insights into the origins, mutation patterns, and evolution of species [44, 45]. We have compiled the codon usage information for the protein-coding sequences of nine species (Additional file 1: Table S4). C. tibetica and C. turkestanica presented the 63 RSCU, and composed of 21,965 and 22,035 codons. There were 21,998 (C. arborescens), 22,035 (C. opulens), 21,998 (C. jubata), 41,710 (C. rosea), 40,657 (C. microphylla), 41,030 (C. kozlowii) and 40,552 (C. korshinskii) codons, respectively. Leucine was the amino acid with the highest quantity (2,333 codons in C. tibetica, 2,347 codons in C. turkestanica, 2,327 codons in C. arborescens, 2,347 codons in C. opulens, 2,331 codons in C. jubata, 4,584 codons in C. rosea, 4,195 codons in C. microphylla, 4,270 codons in C. kozlowii and 4,116 codons in C. korshinskii), while the least prevalent were cysteine (259 codons in C. tibetica, 261 codons in C. turkestanica, 257 codons in C. arborescens, 261 codons in C. opulens, 258 codons in C. jubata) and Tryptophan (656 codons in C. rosea, 670 codons in C. microphylla, 630 codons in C. kozlowii and 639 codons in C. korshinskii.

Meanwhile, we also calculated the Relative Synonymous Codon Usage (RSCU) values to assess the codon usage preference in nine Caragana species. Using a threshold of 1, codons with RSCU values over 1 were considered as optimal codons. 30 preferred and 32 non-preferred (RSCU < 1.00) codon usages were detected in five species, which were C. tibetica (Additional file 2: Fig. S1), C. turkestanica (Additional file 2: Fig. S2), C. arborescens (Additional file 2: Fig. S3), C. opulens (Additional file 2: Fig. S4), and C. jubata (Additional file 2: Fig. S5), 28 preferred and 33 non-preferred in C. rosea (Additional file 2: Fig. S6) and C. korshinskii (Additional file 2: Fig. S7), 29 preferred and 32 non-preferred in C. microphylla (Additional file 2: Fig. S8), 31 preferred and 30 non-preferred in C. kozlowii (Additional file 1: Table S4, Additional file 2: Fig. S9). Furthermore, the RSCU values for most A/U-ending codons were > 1, while C/G-ending codons were < 1 (Additional file 1: Table S2).

Sequence divergence analysis

To reveal the conservative character and divergence in Caragana species, we used mVISTA to compare the plastid sequences of C. tibetica and C. turkestanica studied in this paper and other seven species of Caragana plants that have been reported. The annotated chloroplast genome sequence of C. jubata served as the reference sequence (Fig. 5). The results revealed a high degree of similarity in the nine plastid genome sequences. However, the sequences were found to exhibit differences in the intergenic spacer (IGS) regions of certain genes, such as matK-rbcL, psbD-psbM, atpA-psbI, and etc. In addition, most of the protein-coding gene sequences were highly conserved, except for a few genes (rpoC2, accD, ycf2, and ycf1). Furthermore, compared to the non-coding regions, the coding regions were more conserved. This suggested that the rapidly evolving regions in Caragana genus were located in IGS.

Fig. 5
figure 5

The chloroplast genome of nine Caragana species were compared by mVISTA. The gray arrow in the figure indicates the direction of gene translation; The x-axis represents the coordinates in the chloroplast genome; The y-axis represents the percentage between 50 and 100%; Blue indicates protein coding (exon); Light green indicates untranslated region (UTR); Orange indicates conserved non-coding sequences (CNSs)

Full size image

We utilized the Mauve software to analyze the chloroplast DNA rearrangements in nine species of the genus Caragana. The alignment results revealed a high degree of consistency in the types, numbers, and arrangements of coding genes, including CDS, tRNA, and rRNA, across these plants, with no structural inversions or gene rearrangement events observed (Fig. 6).

Fig. 6
figure 6

Comparative genomic analysis of chloroplast genome of nine Caragana species. Note: Color bands represent genes, and different colors represent different blocks. The squares with the same color between different genes represent homologous regions. In each block, the similarity map of the genome sequence was drawn by Mauve software, and the height of the similarity map corresponded to the average conservative level of the genome sequence region. The two rows below the color band represent the gene. The upper side is on the positive chain, and the lower side is on the complementary chain. The white squares represent CDS, the thin lines in the white squares represent introns, and the green and red squares represent tRNA and rRNA

Full size image

Then, we utilized DNAsp6 to detect nucleotide diversity, and identified highly mutated regions in the chloroplast genomes of nine Caragana species. The pi values range from 0 to 0.11847, with an average of approximately 0.01257 (Fig. 7), indicating significant differences among the sequences. We have identified five regions that were most likely to be variable, including rps2-atpI (π = 0.11847), accD-psaI-ycf4 (π = 0.05819), cemA-petA (π = 0.04949), psbN-psbH (π = 0.04144) and rpoA-rps11 (π = 0.04065). Among these regions, the rps2-atpI region had the highest π value (0.11847).

Fig. 7
figure 7

Nucleotide variability (π) values of nine Caragana plants. The linear gene graph spectrum of Caragana species is given below

Full size image

Phylogenetic analysis

To investigate the phylogenetic positions of 9 species of genus Caragana in the family Fabaceae, we conducted phylogenetic analysis using Maximum Likelihood (ML) and Bayesian Inference (BI) methods. Except for the genus Caragana, the remaining 8 genus included Wisteria (1), Glycyrrhiza (2), Astragalus (1), Calophaca (1), Cicer (1), Medicago (3), Trifolium (3), and Lathyrus (4). The number in parentheses represents the number of species in the corresponding taxa.

The phylogenetic trees obtained from two methods showed similar topology, and the different datasets generally yielded consistent phylogenetic trees with strong support values. The phylogenetic analysis revealed that all samples were divided into three major branches. Four pairs of species showed closer relationships: C. tibetica and C. jubata, C. rosea and C. opulens, C. microphylla and C. korshinskii, and C. turkestanica and C. arborescens (Fig. 8). Notably, the close relationship between the genus Astragalus and genus Caragana (bootstrap: 100%) belonging to the Subtribe Astragaliinae was worth mentioning.

Fig. 8
figure 8

A phylogenetic tree based on the chloroplast genomes of 24 Fabaceae family and one outgroup Arabidopsis thaliana was constructed using BI and ML methods. The numbers following the nodes represent bootstrap values. GenBank accession numbers were provided after each species. C. tibetica and C. turkestanica were highlighted in red

Full size image

Discussion

This study presented the assembly and annotation of two complete chloroplast genomes. By analyzing these genomes, we gained in-depth insights into the chloroplast genomes of the Caragana genus and conducted comparative studies on seven previously reported Caragana species.

Reports indicate that researcher, by analyzing characteristics such as morphological features, chromosomes, and pollen morphology of 72 species within the Caragana genus and employing cladistic methods, successfully classified these species into 12 series and 5 groups [46]. However, this classification method based on morphology could be influenced by environmental changes and convergent evolution, leading to potential discrepancies in the classification outcomes. With the advancement of molecular marker technology, researcher further utilized DNA sequences from the rbcL gene, trnS-trnG introns and spacer regions, and the ITS region to study the phylogenetic relationships between 12 Caragana species and 48 other legume plants [47]. Although molecular markers have improved the accuracy of classifications in certain aspects, their limitations still exist, preventing the resolution of some evolutionary disputes. Given the chloroplast genome’s low mutation rate, ease of sequencing, and high sequence conservation, it has become a crucial tool for analyzing genetic differences among closely related species, helping to overcome the limitations of traditional methods and enhance classification accuracy [48].

In contrast to the majority of angiosperms, both C. tibetica and C. turkestanica, along with previously reported Caragana species, showed a notable absence of the inverted repeat (IR) region. This absence leaded to an unclear demarcation between the large single-copy (LSC) and small single-copy (SSC) regions. For example, C. rosea, C. microphylla, C. intermedia, C. jubata, C. erinacea, C. opulens, and C. bicolor [1, 16, 17, 19]. Comparative analysis of chloroplast genomes within the Caragana genus was conducted. The lengths of the chloroplast genomes varied to some degree among the nine Caragana plants, with the longest being 133,122 bp in C. rosea and the shortest being only 128,132 bp in C. jubata. C. tibetica and C. turkestanica encoded 110 and 111 genes respectively, including 76/76 protein-coding genes, 4/4 rRNA genes, and 30/31 tRNA genes. In this study, we identified a unique gene, trnN-GUU, in the genome of C. turkestanica, which was missing in the genome of C. tibetica. The trnN-GUU gene encoded a tRNA that transports valine, representing a unique sequence within C. turkestanica’s genome. Through codon usage bias analysis, we discovered that the RSCU of valine-related codons encoded by trnN-GUU in C. turkestanica genome exceeds 1, indicating that the use of valine in relation to trnN-GUU may be elevated during protein synthesis. In contrast, no homologous sequence for the trnN-GUU gene was identified in C. tibetica genome, suggesting that the two subspecies may have different evolutionary histories and paths of adaptive evolution. The presence or absence of trnN-GUU in the genomes of C. turkestanica and C. tibetica, respectively, may reflect their divergent evolutionary branches. Future research should investigate the functional role of the trnN-GUU gene in the adaptive evolution of C. turkestanica, as well as its evolutionary importance in adaptation to various environments. Additionally, 239 and 277 SSRs were found to be randomly distributed in their genomes. Furthermore, approximately 88/84 forward repeats, 30/36 palindromic repeats, 1/7 reverse repeats, and 0/1 complementary repeats were identified in both cp genomes. We have identified 17 genes containing introns in both chloroplast genomes, with 15 genes containing 1 intron each in C. tibetica (16 in C. turkestanica), while the genes clpP and ycf3 contained 2 introns in C. tibetica, C. turkestanica only had one gene ycf3 contain 2 introns. Introns play a significant role in gene expression regulation, and recent studies have indicated that different introns can enhance the expression of exogenous genes at specific locations [49,50,51]. In transgenic mice, the addition of an intron from the rabbit-globin gene was found to enhance the expression of the human growth hormone gene, leading to increased levels of the hormone [52]. Similarly, research indicates that deleting the introns of the Drosophila alcohol dehydrogenase (Adh) gene leads to reduced expression levels of the Adh gene, whether measured by enzymatic activity or RNA levels. This accentuates the important role of introns in the regulation of gene expression [53]. It was found that ycf2 [54], rpl23, and accD were frequently missing in some plants, but they were detected in chloroplast genome of Caragana plants reported in this paper [55,56,57]. Previous research has indicated that specific Caragana species’ chloroplast genome, including C. kozlowii, C. korshinskii, C. microphylla, and C. rosea, exhibit a loss of genes such as rps16, infA, rpl22, and ycf15 [1]. Similarly, this study also did not find the rps16, infA, rpl22, and ycf15 loci in C. tibetica and C. turkestanica. Among these genes, infA is an unusually unstable flowering plant chloroplast gene, whereas rpl22 encodes the ribosomal protein CL22 and has been eliminated from the chloroplast genome, relocating to the nucleus [58, 59]. Nonetheless, recent studies have discovered that the infA gene is present in C. jubata, C. erinacea, C. opulens, and C. bicolor [4]. These findings strongly indicated that the infA gene might not be lost in some Caragana species. Due to the scarcity of available chloroplast genomic data for Caragana, it is imperative to conduct extensive experimental research in order to validate these conclusions.

A total of 119 and 128 repeats were detected in the chloroplast genomes of C. tibetica and C. turkestanica, respectively. These repeats include forward, palindromic, reverse, and complementary sequences. These repeats serve as crucial genetic markers and are closely associated with species emergence and development [60]. Repeat sequences are highly valuable in phylogenetic investigations and also contribute to genome rearrangements [61, 62]. Furthermore, the analysis of various cp genomes has established the vital role of repeat sequences in indel and substitution events [63]. Moreover, no rearrangements have been observed in the plastids of C. tibetica and C. turkestanica. Previous studies have reported the absence of the IR region in several Caragana species, such as C. microphylla, C. bicolor, and C. jubata [1, 16]. Similarly, the cp genomes of C. tibetica and C. turkestanica were found to lack the IR region in this study. Additionally, the G/C content of the chloroplast DNA (cpDNA) is crucial for determining inter-specific affinity [16]. The DNA G/C content of the two Caragana species discussed in this paper is highly alike. Additionally, SSRs are regarded as vital molecular markers for analyzing genetic variation within populations, and they are widely employed in assessing phylogenetic relationships, evolution, and genetic diversity [64]. A total of 239 to 277 SSRs were found in the chloroplast genomes of two Caragana plants, exhibiting a significant bias towards A/T. The majority of SSR types consist of single nucleotide repeats, and the highest number of SSRs was observed in the non-coding regions (IGSs) of the cp genomes. These SSRs are valuable starting points for the development of genetic markers in Caragana species, and their utilization is applicable in phylogenetic and ecological research.

Codon usage preferences are known to reflect the species of origin and the mutational model. Analyzing patterns of codon usage bias in chloroplast genomes can provide insights into plant phylogenetic relationships, gene expression mechanisms, and molecular evolution [44]. Leucine (Leu) is the most abundant amino acid in C. tibetica, C. turkestanica, and other Caragana species. Furthermore, our research found that the majority of synonymous codons preferred for RSCU values end with A/U, thereby contributing to a higher AT content in genes. Based on our analysis, we hypothesize that natural selection and gene mutation may be responsible for this codon usage pattern. Nevertheless, it is important to note that codon preference and utilization patterns provide only a partial reflection of the evolutionary relationship between species, and further research is necessary.

The plastid genome contains numerous variable nucleotides, which can be utilized as valuable DNA barcodes for determining the relationships between species or genera [65, 66]. We simultaneously identified five intergenic spacer regions (IGSs) with relatively high differentiation values: rps2-atpI, accD-psaI-ycf4, cemA-petA, psbN-psbH, and rpoA-rps11. These variable regions have the potential to function as DNA barcodes for studying phylogenetic relationships, species identification, and population genetics research. Next, we compared the sequence variations of the nine assembled Caragana plants. Comparative analysis of the chloroplast genomes confirmed that the coding regions were more conserved than the non-coding regions, which is consistent with findings from other Caragana species [16].

Recently, the chloroplast genome has become a preferred option for studying the phylogenetic relationships of diverse plant species. For instance, the utilization of the complete chloroplast genomes from three Lycoris plants in phylogenetic analysis revealed a strongly supported relationship between Lycoris plants and the Narcissus genus [67]. Phylogenetic analysis of 23 Swertia plant species revealed that Swertia is paraphyletic rather than monophyletic [68]. The phylogenetic position of C. tibetica and C. turkestanica within Fabaceae was determined by constructing a comprehensive genomic dataset composed of 66 genes shared among 24 representatives from nine genera. The phylogenetic analysis indicated that Caragana species constituted a distinct clade, with C. microphylla and C. korshinskii showing a closer genetic relationship, which aligns with earlier findings [1, 16]. The phylogenetic relationships inferred from the chloroplast genome offer novel insights and perspectives to advance our understanding of plant evolution.

For the first time, we assembled and analyzed the complete chloroplast genomes of C. tibetica and C. turkestanica, and compared them to other members of the Caragana genus. Our findings revealed that the sizes of their genomes, gene compositions, gene arrangements, GC content, and codon usage were similar to previously documented chloroplast genomes within the Caragana genus. Additionally, this study identified the positions and distributions of repetitive sequences in both species, and determined the sequence variability and nucleotide variation sites. The results of this study provide guidance for future studies on the phylogenetic evolution and species identification of the Caragana genus, as well as the development of new molecular markers. Ultimately, this discovery contributes to the augmentation of the chloroplast genome database for the Caragana genus.

Conclusions

Our study presents the initial assembly and annotation of C. tibetica and C. turkestanica and compares them with seven other Caragana species. Due to the absence of a pair of IRs, the plastomes of C. tibetica and C. turkestanica were shorter, with sizes ranging from 128,433 bp to 129,453 bp. The long repeats, SSR loci, and certain genes found in the IGS region (matK-rbcL, psbD-psbM, atpA-psbI) as well as five regions exhibiting high variability (rps2-atpI, accD-psaI-ycf4, cemA-petA, psbN-psbH, and rpoA-rps11) identified in our study will promote future research efforts. These efforts include the development of new molecular markers and investigations of population genetics and phylogenetic analysis. By analyzing the sequence and structural information of the chloroplast genomes of two Caragana plants comprehensively, we have determined their genetic evolutionary position and relationships with other Caragana species. This information establishes a foundation for extensive and detailed research on the identification, genetic diversity, and phylogenetics of Caragana species. Furthermore, our study has significantly contributed to the enrichment of the chloroplast genome database for Caragana plants.

Methods

Plant material, DNA extraction and sequencing

The fresh and young leaves of C. tibetica, and C. turkestanica were gathered from eastern part of Qinghai Province (N36°43′24.80′′, E101°44′54.11′′), China. During outdoor sampling, the leaf tissues were temporarily stored in a low-temperature insulated box. After returning to the laboratory, the samples were immediately placed in a -80℃ ultra-low temperature freezer for storage. We used a modified cetyltrimethylammonium bromide (CTAB) [69] method to extract DNA from fresh tissue samples of Caragana plants. Ultrasound was used to fragment the DNA fragments, and the fragment size was selected by agarose gel electrophoresis. The selected fragments were amplified by PCR to form a sequencing library, and the qualified library was sequenced using the Illumina NovaSeq platform, with 150 bp pair-end reads.

Gene annotation, genome assembly and sequence analyses

Before assembly, we conducted a rigorous preprocessing of the raw data. The raw data were filtered using the Trimmomatic v 0.39 [70] tool to remove low-quality data. After that, we used SPAdes v3.10.1 (http://cab.spbu.ru/software/spades/) [71] to assemble chloroplast genome sequences to obtain their SEED sequences, K-mer analysis was conducted on the seed sequence to obtain Contigs. We employed SSPACE v2 [72] for the assembly of contig sequences into scaffolds. Subsequently, Gapfiller v2.1.1 [73] was utilized to resolve any gaps within these scaffolds, thereby enhancing the coherence and completeness of the pseudo-genome assembly. Then, based on the structure of the chloroplast genome, the corrected pseudo genome sequences were reordered and aligned, resulting in two complete circular chloroplast genome sequences. After assembly, quality control of the final sequence was performed using the reference sequence for C. kozlowii in this study. The annotation information for the CDS, rRNA, and tRNA sequences in the chloroplast genome were gained using Blast v2.2.25 (https://blast.ncbi.nlm.nih.gov/Blast.cgi), hmmer v3.1b2 [74] (http://www.hmmer.org/), and ARAGORN v1.2.38 [75] (http://130.235.244.92/ARAGORN/) software, respectively. The chloroplast genome maps for C. tibetica and C. turkestanica were plotted by the online tool Chloroplot in OGDRAW [76]. Finally, the annotated chloroplast genome sequences of the genus Caragana were submitted to GenBank using the online submission tool BankIt, and the accession numbers OQ942026 and OQ942027 were obtained, respectively.

Repeats, simple sequence repeats and codon usage bias analysis

Vmatch [77] was chosen to search for forward, reverse, palindromic, and complementary repeats in the chloroplast genome sequences of Caragana species. Additionally, simple sequence repeats (SSRs) in this study were identified by MISA [78] with the following parameter settings: mono-nucleotide set as 8, di-nucleotide set as 5, tri-nucleotide set as 3, tetra-nucleotide set as 3, penta-nucleotide set as 3, and hexa-nucleotide set as 3, respectively. The program CodonW1.4.2 [79] was applied to calculate relative synonymous codon usage (RSCU) values of protein-coding genes under default settings.

Sequence divergence and comparative genome analysis

The chloroplast genome sequences of Caragana species were aligned using MEGA7 [80], and DnaSP6 [81] was used to calculate nucleotide diversity (π) values with the following parameter settings: window length of 600 bp and step size of 200 bp, which are commonly used in the literature. Comparison and visualization of complete Caragana chloroplast genomes using mVISTA [82] program (Shuffle-LAGAN mode). C. jubata plastome was labeled as reference. Utilizing Mauve software [83], an analysis was conducted on the chloroplast DNA rearrangements of nine species within the genus Caragana, aiming to identify changes in gene order, potential large-scale sequence rearrangements, and local tandem duplication events. This step employed the software’s recommended default parameters.

Phylogenetic analysis

To determine the phylogenetic positions of Caragana species in this study, we downloaded plastid genome sequences of 24 legume species which belong to the IRLC, and one outgroup (Arabidopsis thaliana) from NCBI. Phylogenetic trees were constructed using PhyloSuite v1.2.2 [84]. First, we extracted 66 common protein-coding gene sequences from each of 25 chloroplast genomes. These protein gene sequences were then aligned in batches using the MAFFT [85] with auto strategy and normal alignment mode. This study constructed phylogenetic trees based on the chloroplast genome using two methods: Bayesian inference (BI) and maximum likelihood (ML). In the Bayesian Inference (BI) analysis, we employed Modelfinder [86] to determine the most suitable model, opting for the GTR + F + I + G4 nucleotide substitution model. Subsequently, phylogenetic trees were inferred using Bayesian inference [87], leveraging a partition model to enhance the accuracy of our findings. It runs in parallel with one execution, performing a total of 1,000,000 generations. During the analysis, the initial 25% of sampled data is discarded as a burn-in period. To ensure the convergence of the Markov Chain Monte Carlo (MCMC) algorithm, these samples are excluded from the final analysis. Additionally, the average standard deviation of split frequencies is set to a threshold greater than 0.01 [88]. In an evolutionary tree constructed using BI, a value of 1 represents the highest probability or maximum posterior probability for support, indicating that a particular structure or branch is highly reliable given the data and model. For ML, the IQ-TREE [89] tool was used, with automatic partition selection and 1,000 ultrafast bootstrap [90]replicates performed to assess the confidence of each branch. And we set the threshold for Bootstrap values at 70% as the criterion for evaluating support. Therefore, when the Bootstrap value of a particular clade reaches or exceeds 70%, it is considered to have strong support.