Comparative chloroplast genomics of Limeum (Limeaceae) species: New insights into genome evolution and phylogenomic implications
Article information
Abstract
Limeum is the sole genus of Limeaceae, comprising approximately 20 species distributed across Africa, Southwest Asia, and India. However, genomic resources for the family are limited, and phylogenetic relationships within the genus remain poorly understood. In this study, we sequenced, assembled, and analyzed the complete chloroplast genomes of six Limeum species to investigate their genomic characteristics and phylogenetic relationships. The chloroplast genomes ranged from 156,354 bp to 160,377 bp in length and exhibited the typical quadripartite structure. Each genome contained 131 genes, consisting of 86 protein-coding genes, 37 tRNA genes, and eight rRNA genes. Comparative analyses identified 455 tandem repeats and 616 simple sequence repeats, revealing 16 highly variable regions that may serve as potential molecular markers for future studies. Phylogenetic analyses based on complete chloroplast genome sequences strongly supported the monophyly of Limeum and robustly resolved interspecific relationships within the genus. These results provide the first comprehensive plastome dataset for Limeum and a valuable genomic resource for species identification, genetic diversity assessment, phylogenetic reconstruction, and evolution studies of Limeum and the family Limeaceae.
INTRODUCTION
Limeum L. is the sole genus of Limeaceae comprising approximately 20 psammophytic species (Hernández-Ledesma et al., 2015; Sukhorukov and Kushunina, 2020) distributed mainly in Africa, with two species extending into Southwest Asia and India (Endress and Bittrich, 1993; Hernández- Ledesma et al., 2015). Species of Limeum are annual or perennial herbs, subshrubs, or shrubby lianas characterized by the alternate leaves that are generally glabrous or bear simple or glandular trichomes, a sepaloid green perianth with white margins, petals that are often reduced or absent, flowers with 5–7 stamens arranged in dense terminal or axillary inflorescences, and fruits breaking into two mericarps. These characters readily distinguish Limeum from morphologically similar members of Caryophyllales.
The systematic placement of Limeum has long been controversial. The genus has historically been assigned to Aizoaceae (Oliver, 1871; Schellenberg, 1913; Gonçalves, 1970), Phytolaccaceae (Heimerl, 1889), and most commonly Molluginaceae (Ecklon and Zeyher, 1836; Friedrich, 1956; Endress and Bittrich, 1993). Molecular phylogenetic analyses demonstrated that Limeum is distinct from Molluginaceae and represents an isolated lineage within Caryophyllales (Cuénoud et al., 2002), leading to the recognition of Limeaceae as a separate family (Shipunov in Hoogland and Reveal 2005). Subsequent studies have consistently supported this treatment (The Angiosperm Phylogeny Group, 2009; Christin et al., 2011; The Angiosperm Phylogeny Group et al., 2016; Zuntini et al., 2024). Recent phylogenomic analyses further indicate that Limeum is sister to the ‘globular inclusion’ clade comprising Aizoaceae s.str., Molluginaceae s.str., and Phytolaccaceae s. str. (Brockington et al., 2009, 2013).
Although the familial placement of Limeaceae is now well-established, relationships within the genus remain poorly understood due to limited taxon sampling and sparse genomic data. To date, only nearly 60 DNA sequences of Limeum are available in GenBank, most of which are short DNA fragments (Table 1). Partial plastid sequences of L. aethiopicum Burm.f. (MH286320.1) and L. africanum L. (MK397924.1) were included in a broad-scale study of Caryophyllales (Yao et al., 2019), but no investigation has focused specifically on plastome evolution or species-level relationships in Limeum. Consequently, the extent of plastome variation and its phylogenetic utility within the genus remains largely unknown.
Molecular markers used in previous studies of Limeum.
Complete plastid genomes have become an important resource of data for plant systematics and evolutionary studies because they provide abundant phylogenetically informative characters while retaining a relatively conserved genomic structure (Downie & Palmer, 1992; Soejima and Wen, 2006; Jansen and Ruhlman, 2012; Ruhfel et al., 2014; Gitzendanner et al., 2018; Wu et al., 2020; Zhao et al., 2021). In addition to nucleotide sequence variation, structural features such as expansions and contractions of inverted repeats (IR), gene loss, and genome rearrangements may contribute valuable evidence for evolutionary inference (Daniell et al., 2016; Li and Zheng, 2018). Comparative plastome analyses have therefore become increasingly useful for resolving phylogenetic relationships and identifying molecular markers at different taxonomic levels (Li et al., 2021b; Zhao et al., 2021; Chen et al., 2022; Fu et al., 2025).
With rapid advances in next-generation sequencing, sequencing complete plastid genomes of representative species is essential for establishing a robust foundation for species identification, molecular systematics, phylogenomics, and population genetic studies. Moreover, as many species of Limeum are adapted to sandy habitats, plastome-level analyses and development of molecular markers may help elucidate their adaptive evolution and support genetic conservation, particularly under scenarios of global climate change.
To date, no comprehensive chloroplast-based phylogeny of Limeum has been published, resulting in the absence of a stable evolutionary framework and limiting comparative genomic analyses. To address this gap, we sequenced and assembled the complete chloroplast genomes of six species, L. arabicum Friedrich, L. dinteri G. Schellenb., L. myosotis H. Walter, L. obovatum Vicary, L. sulcatum (Klotzsch.) Hutch., and L. viscosum (J. Gay) Fenzl. Specifically, we aimed to (1) characterize plastome structure and genome features within Limeum; (2) identify highly variable regions with potential utility for species identification and future phylogenetic analyses; and (3) evaluate phylogenetic relationships within the genus using complete plastome sequences. The resulting data provide a genomic framework for future systematic and evolutionary studies of Limeum and contribute to a better understanding of plastome evolution in Limeaceae.
MATERIALS AND METHODS
Plant materials, DNA extraction, and sequencing
In this study, six species of Limeum were sampled and newly sequenced from Oman, Namibia, Angola, Saudi Arabia, and Senegal. Among them, L. sulcatum was collected in the field from Rehoboth, Hardap Region, Namibia, whereas the remaining five species (L. arabicum, L. dinteri, L. myosotis, L. obovatum, L. viscosum) were obtained from herbarium specimens collected between 1976 and 2017. Voucher information for all samples is provided in Table 2.
Voucher information for the samples newly sequenced in this study.
Notably, L. viscosum exhibits remarkable morphological variation and has a broad geographic distribution. In this study, L. viscosum refers specifically to the taxon whose sample originated from the same region as the type specimens (K000232116 & K00023117, as Gaudinia viscosa J. Gay). According to that, this taxon is treated here as L. viscosum sensu stricto.
Total genomic DNA was extracted from silica gel-dried leaves using a modified CTAB protocol (Doyle and Doyle, 1987). For herbarium specimens, DNA was extracted from approximately 20 mg of leaf tissue using the Tiangen DNA secure Plant Kit (DP320, TIANGEN Biotech, Beijing, China) following the manufacturer’s instructions. DNA quality and concentration were evaluated using a NanoDrop 2000 spectrophotometer, and only samples with concentrations exceeding 30 ng/μL were used for library preparation. DNA was fragmented to approximately 300 bp, and sequencing libraries were prepared following the standard Illumina protocol. High-throughput sequencing was conducted on the Illumina platform, generating a total of 348,560,432 reads, with an average yield of approximately 10 Gb per sample.
Genome assembly and annotation
Raw reads were filtered using Fastp (v0.23.4) (Chen et al., 2018a) to remove adapter sequences and low-quality reads. Clean reads were then assembled de novo into complete chloroplast genomes using GetOrganelle v1.7.7.1 (Jin et al., 2020) with parameters set to -F embplant_pt, -R 15, and -K 21, 45, 75, 105, 127. Genome annotation was conducted using Plastid Genome Annotator (Qu et al., 2019), with Lophiocarpus tenuissimus Hook.f. (NC_040951) serving as the reference genome.
Annotations were manually inspected and corrected in Geneious Prime 2025.0.2 (Kearse et al., 2012) to correct potential mis-annotations and recover missing genes. Genomic features, including gene content, GC content, gene and intron length, were analyzed in Geneious Prime 2025.0.2 (Kearse et al., 2012). Circular genome maps were visualized using Organellar Genome Draw (OGDraw) (https://chlorobox.mpimpgolm.mpg.de/OGDraw.html).
Simple and long repeat sequence analysis
Simple sequence repeats (SSRs) were identified using MISA v2.1 (http://pgrc.ipk-gatersleben.de/misa/) with minimum repeat thresholds: set to 10 for mononucleotide, five for dinucleotide, and four for trinucleotide motifs; and three for tetra-, penta-, and hexanucleotide motifs (Beier et al., 2017). Long sequence repeats, including forward, palindromic, complement, and reverse repeats, were detected using the online REPuter program (Kurtz et al., 2001) with the minimum repeat size of 30 bp and the Hamming distance of three.
Comparative analysis of chloroplast genomes
Expansion and contraction of IR regions among the six plastomes of Limeum were analyzed using CPJSdraw (Li et al., 2023). Sequence similarity among chloroplast genomes was evaluated using mVISTA program (Frazer et al., 2004) in Shuffle-LAGAN mode with default settings, using L. arabicum as the reference genome.
Codon usage analysis
Codon usage bias was evaluated using Cpstools 3.0 (Huang et al., 2024) by calculating relative synonymous codon usage (RSCU) values (Sharp and Li, 1987) for all protein-coding genes. RSCU values > 1 indicate preferential codon usage, values < 1 indicate underrepresented codons, and RSCU = 1 indicates no codon usage bias. The results were visualized using customized Python scripts to illustrate codon usage patterns among species.
Nucleotide divergence analysis
Highly variable regions among the six chloroplast genomes were identified by calculating nucleotide diversity (π) for both coding and non-coding regions. Sequence alignments and π calculations were performed using Cpstools 3.0 (Huang et al., 2024), and the results were visualized using custom in-house Python-based scripts to facilitate comparative interpretation.
Hypervariable regions were identified using the criterion π > μ + k × σ, where μ represents the mean nucleotide diversity across all regions and σ represents the standard deviation, and k is a threshold coefficient. Here, k was set to 1.5, which selects regions with π values exceeding approximately 93.3% of the data distribution and was applied consistently to both coding regions and intergenic spacers to facilitate comparative analyses.
Phylogenetic analysis
Phylogenetic analyses were conducted using six complete plastome sequences of Limeum generated in this study, together with two plastome sequences retrieved from GenBank (L. aethiopicum, MH286320; L. africanum, MK397924) (Yao et al., 2019). Although the latter sequences are incomplete, they were included to maximize species representation within the genus.
To test the monophyly of Limeaceae, a total of eight Limeum accessions and 10 representatives from six families (Aizoaceae, Amaranthaceae, Kewaceae, Macarthuriaceae, Nyctaginaceae, and Phytolaccaceae) were included as ingroup. Four taxa (Agrostemma githago L., Cerastium davuricum Fisch. ex Spreng., Honckenya peploides (L.) Ehrh., and Silene jeniseensis Willd.) were selected as outgroups for tree rooting.
Two data sets were constructed for phylogenetic inference: (1) complete plastome sequences with one copy of the IR region removed, and (2) the 16 hypervariable regions identified in this study. Sequence alignments were performed using MAFFT 7.490 (Katoh et al., 2019) under default parameters and subsequently refined with TrimAL v1.4 (Capella-Gutiérrez et al., 2009).
Phylogenetic relationships were reconstructed using both maximum Likelihood (ML) and Bayesian Inference (BI) methods. Node support was assessed using bootstrap support (BS) and posterior probability (PP) values.
ML analyses were conducted using RAxML 8.2.12 under the GTR+GAMMA model (Stamatakis et al., 2008; Stamatakis, 2014) with 1000 rapid bootstrap replicates. BI analyses were performed with MrBayes v.3.2.7a (Ronquist et al., 2012) under the GTR + G + I substitution model, which was selected as the best-fitting model by jModeltest 2 (Darriba et al., 2012) based on the Akaike Information Criterion. Markov Chain Monte Carlo analyses were run for 20,000,000 generations, sampling every 1,000 generations. The first 25% of sampled trees were discarded as burn-in. The resulting phylogenetic trees were visualized and edited using FigTree v1.4.2 (Rambaut, 2014) and TreeGraph 2 (Stöver and Müller, 2010).
RESULTS
Chloroplast genome features of the six species of Limeum
The complete chloroplast genomes of the six species of Limeum ranged in size from 156,354 bp in L. arabicum to 160,377 bp in L. sulcatum (Table 3). All plastomes exhibited the typical quadripartite circular structure, consisting of a large single-copy (LSC) region (87,710–91,839 bp), a small singlecopy (SSC) region (19,200–20,060 bp), and a pair of IR regions (24,658–24,779 bp) (Fig. 1, Table 3). The overall GC content varied from 35.7% to 36.5%. Among the four regions, the IR regions exhibited the highest GC content (43.2%), compared with the LSC (32.9–34.1%) and SSC regions (29.3–30.2%).
Plastome size and gene content of six Limeum species.
Gene map of the chloroplast genomes of six Limeum species. Genes located inside and outside the circle are transcribed in the clockwise and counterclockwise directions, respectively. Genes belonging to different functional categories are shown in different colors.
Each chloroplast genome encoded 131 genes, including 86 protein-coding genes, 37 tRNA genes, and eight rRNA genes. This genus could be categorized into three functional groups: photosynthesis-related genes, self-replication genes, and genes with other functions. Six protein-coding genes (ndhB, rpl2, rpl23, rps7, rps19, and ycf2), seven tRNAs (trnA-UGC, trnI-CAU, trnI-GAU, trnL-CAA, trnN-GUU, trnR-ACG, and trnV-GAC), and all four rRNA genes (rrn4.5, rrn5, rrn16, and rrn23) were duplicated within the IR regions (Fig. 1).
A total of 17 genes contained introns (Table 4). Among them, 14 genes contained a single intron (atpF, ndhA, ndhB, petB, petD, rpl16, rps16, rpoC1, trnA-UGC, trnG-UCC, trnIGAU, trnK-UUU, trnL-UAA, and trnV-UAC), whereas three genes (clpP, rps12, and ycf3) contained two introns. The rps12 gene was trans-spliced, with its 5′ exon located in the LSC region and 3′ exons duplicated in the IR regions. Most protein-coding genes used the canonical AUG start codon, whereas a few genes employed alternative start codons, including ACG (psbL) and GUG (rps19).
Functional classifications of plastid genes in Limeum chloroplast genomes.
Repetitive sequences and SSR
A total of 455 long repeats were identified across the six plastomes, including forward (F), reverse (R), complement (C), and palindromic (P) repeats (Fig. 2A). Among the six species, L. sulcatum possessed the highest number of long repeats (100), comprising 51 forward, 24 reverse, two complement, and 23 palindromic repeats. The numbers of long repeats identified in L. dinteri, L. myosotis, L. viscosum, L. obovatum, and L. arabicum were 94, 79, 76, 60, and 46, respectively.
Analysis of long repeat sequences in the chloroplast genomes of six Limeum species. A. Number of repeat types detected in each plastome. B. Frequency of different repeat types. C. Distribution of repeat lengths.
Forward repeats were the most abundant (50.77%), followed by palindromic repeats (34.07%), whereas reverse and complement repeats accounted for 12.09% and 3.08% of the total, respectively (Fig. 2B). Repeat lengths ranged from 30 to 80 bp, with the highest frequencies observed at 30, 31, 33, 38, and 74 bp (Fig. 2C).
In total, 616 SSRs were detected among the six plastomes of Limeum. The numbers of SSRs identified in L. arabicum, L. dinteri, L. myosotis, L. obovatum, L. sulcatum, and L. viscosum were 89, 117, 107, 90, 101, and 112, respectively. Mononucleotide repeats represented the most abundant SSR type, ranging from 51 in L. sulcatum to 74 in L. dinteri (Fig. 3A). Among these, A/T mononucleotide repeats predominated, with counts of 58, 73, 63, 55, 51, and 62, respectively.
Analysis of simple sequence repeats (SSRs) in the chloroplast genomes of six newly sequenced Limeum species. A. Number of different SSR types. B. Number of different SSR motifs.
In total, 22 SSR motif types were identified. Motifs shared among all species included mononucleotides (A/T), dinucleotides (AT/AT, AG/CT), trinucleotides (AAT/ATT, and tetranucleotides (AAGG/CCTT, AAAT/ATTT, AAAG/CTTT, AATT/AATT, ACCT/AGGT) (Fig. 3B).
Comparative chloroplast genome structure
Comparison of IR/SC (single copy) boundary regions among the six Limeum species revealed a highly conserved overall plastome structure, with only minor variations at the junctions involving ndhF, rpl22, rps19, trnN, trnH, and ycf1 (Fig. 4). The lengths of the IR regions ranged from 24,658 to 24,779 bp.
Comparison of the boundaries among the large single-copy (LSC), inverted repeat (IR), and small single-copy (SSC) regions in the chloroplast genomes of the six Limeum species. Genes are annotated in different colors, and the distances between genes and the corresponding junctions are indicated. JLB, JSB, JSA, and JLA represent the junctions between LSC/IRb, IRb/SSC, SSC/IRa, and IRa/LSC, respectively.
Among the six species, only the SSC/IRa junction (JSA) extends into the ycf1 gene. The rpl22 gene was entirely located within the LSC region, 12–29 bp from the LSC/IRb boundary (JLB). The rps19 gene was positioned downstream of the JLB and completely duplicated within the IR region. The ndhF gene spanned the IRb/SSC boundary (JSB).
The mVISTA analysis (Fig. 5) showed that sequence divergence was generally higher in the LSC and SSC regions than in the IR regions. In addition, non-coding regions exhibited greater variability than coding regions, whereas rRNA genes remained highly conserved.
Sequence alignment of six Limeum chloroplast genomes using mVISTA, with Limeum arabicum as the reference. The X-axis represents the coordinates of the chloroplast genome, and the Y-axis indicates sequence identity (50–100%). Gene transcriptional directions are shown by grey arrows.
Codon usage analysis
After excluding duplicated genes, short sequences (<300 bp), and genes lacking a canonical ATG start codon, 53 protein-coding genes were retained for RSCU analysis. The results revealed that 30 codons had RSCU values > 1, indicating preferential usage, whereas 32 codons had RSCU < 1 (Fig. 6). Methionine (Met) and tryptophan (Trp) exhibited no codon usage bias (RSCU = 1).
Codon usage analysis of six Limeum chloroplast genomes. A. Codon frequencies of the 20 amino acids and stop codons in protein-coding genes. B. Heat map showing codon usage bias among the six chloroplast genomes. Color intensity corresponds to the Euclidean distance.
Most preferred codons terminated with A or T, except for TTG (Leu). Among all codons, TTA (Leu) exhibited the highest usage bias (RSCU ≥ 1.98), whereas CTG was the least preferred codon (RSCU ≤ 0.37).
Nucleotide diversity analysis
Nucleotide diversity analysis revealed variable levels of sequence divergence among coding genes and intergenic spacers across the six Limeum plastomes. Protein-coding genes showed π values ranging from 0 to 0.03407, with five hypervariable loci (π > 0.019): clpP, ndhF, psaI, rpl22, and ycf1. These loci were distributed across the LSC (three) and SSC (two) regions. The average nucleotide diversity of coding regions was 0.0065 (Fig. 7B).
Sliding-window analysis of nucleotide diversity across six Limeum chloroplast genomes. A. Nucleotide diversity (π) of intergenic spacers. B. Nucleotide diversity (π) of protein-coding genes.
Intergenic spacers exhibited substantially higher levels of variation, with π values ranging from 0 to 0.14425. 11 hypervariable intergenic regions (π > 0.07) were identified, most located in the LSC region (atpA-atpF_2, psaA-ycf3_3, rpl33-rps18, rps4-trnT-UGU, rps16_1-trnQ-UUG, trnE-UUCtrnT- GGU, trnH-GUG-psbA, trnK-UUU_1-rps16_2, trnRUCU- atpA, and trnS-GCU-trnG-UCC_1), and one in the SSC (rpl32-trnL-UAG) (Fig. 7A). The average π value of noncoding regions (0.0268) was markedly higher than that of coding sequences.
Phylogenetic analysis
Phylogenetic relationships were reconstructed using both ML and BI analyses based on two separate datasets: complete chloroplast genome sequences and the 16 hypervariable regions. All sampled Limeum species formed a strongly supported monophyletic group (ML BS = 100%, BI PP = 1.00).
Analyses based on complete plastome sequences recovered two strongly supported subclades (Fig. 8). One comprised L. dinteri, L. myosotis, and L. viscosum, whereas the other included L. arabicum, L. obovatum, and L. sulcatum clade (100%, 1.00). Within the latter clade, the sister relationship between L. sulcatum and the clade formed by L. arabicum plus L. obovatum received relatively weak ML support (59%, 1.00).
Best-scoring maximum likelihood (ML) phylogenetic tree of Limeum inferred from the complete chloroplast genome sequences. Support values (bootstrap support [BS] ≥ 50 or posterior probability [PP] ≥ 0.99) are shown in the order maximum likelihood bootstrap support (MLBS)/Bayesian inference posterior probability (BIPP). An asterisk (*) indicates BS = 100% and PP = 1.00.
Phylogenetic reconstruction based on the 16 hypervariable regions also supported the monophyly of Limeum (Fig. 9). However, relationships within the genus were only partially resolved, forming a polytomy with three major lineages. The first lineage (92%, 1.00) included L. arabicum and L. obovatum, which together formed a clade sister to the clade comprising L. africanum and L. aethiopicum. The second lineage consisted of L. dinteri, L. viscosum sensu stricto, and L. myosotis, whereas the third lineage was represented exclusively by L. sulcatum.
Best-scoring maximum likelihood (ML) phylogenetic tree of Limeum inferred from 16 hypervariable regions. Support values (bootstrap support [BS] ≥ 50 or posterior probability [PP] ≥ 0.99) are shown in the order maximum likelihood bootstrap support (MLBS)/Bayesian inference posterior probability (BIPP). An asterisk (*) indicates BS = 100% and PP = 1.00.
DISCUSSION
In this study, we sequenced, assembled, and annotated the complete chloroplast genome of six species of Limeum (Fig. 1, Tables 3, 4). While two Limeum plastomes were previously available, they were sequenced primarily as representatives of Limeaceae for higher-order phylogeny reconstruction. Our study expands plastome sampling to six species and provides the first comparative genomic framework for Limeum, allowing assessment of plastome evolution, sequence divergence, and phylogenetically informative regions across the genus. The plastomes ranged from 156,354 bp (L. arabicum) to 160,377 bp (L. sulcatum), with length differences primarily resulting from indels in non-coding regions. Consistent with observations in most angiosperms (Ruhlman and Jansen, 2014), Limeum plastomes were highly conserved in genome structure, gene content, order, GC contents, and intron number. Their overall structural organization was comparable to that in other plastomes of the genera with a different position within Caryophyllales, such as Alternanthera sessilis (L.) DC., Amaranthaceae (Wang et al., 2024), Bougainvillea glabra Choisy, Nyctaginaceae (Ni et al., 2019), Monococcus echinophorus F.Muell., Petiveriaceae (Yao et al., 2019) and Talinum paniculatum (Jacq.) Gaertn., Talinaceae (Liu et al., 2018). However, several notable features were detected in Limeum. In contrast to most Caryophyllales plastomes, Limeum possesses two complete copies of rps19 located within the IR regions, indicating a lineage-specific expansion of the IR boundary. Because IR expansion and contraction are among the major mechanisms driving plastome size variation (Jansen and Ruhlman, 2012; Wang et al., 2017), this feature may represent a synapomorphic genomic character of Limeum and could provide additional phylogenetic information for future studies within Limeaceae and related lineages.
Several non-standard start codons were identified, a phenomenon widely reported in vascular plants (Wolf et al., 2003; Raubeson et al., 2007; Gichira et al., 2017; Alzahrani et al., 2020). For example, psbL initiated with ACG, consistent with previous findings in diverse lineages, including gymnosperms, monocots, and early-diverging angiosperms (Raubeson et al., 2007; Chen et al., 2011; Kim and Kim, 2013; Yao et al., 2015). Similarly, initiation of rps19 with GUG has been documented in multiple taxa (Kim and Kim, 2013). GC content patterns were similar to those in most angiosperms, with the IR region showing the highest GC content due to the presence of GC-rich rRNA genes (Dong et al., 2018; Mower et al., 2019).
Repetitive elements play important roles in plastome evolution, including regulation of gene expression, facilitation of recombination, and maintenance of genome organization (Aranda-Olmedo et al., 2002; Tobes and Pareja, 2005; Rocco et al., 2010; Zhang et al., 2024). Across the six plastomes, 455 long repeats were identified, predominantly forward (50.77%) and palindromic repeats (34.07%).
SSRs, or microsatellites, constitute an important class of repetitive elements in chloroplast genomes. Owing to their high polymorphism, abundance, and codominant inheritance, SSRs are widely applied as effective molecular markers in genetic diversity, population structure, and phylogeographic studies (Powell et al., 1995; Ni et al., 2016). In the present study, a total of 616 SSRs were detected across six plastomes of Limeum, with mononucleotide repeats being the most abundant (362 loci). These mononucleotide SSRs were primarily composed of A/T motifs and exhibited notable variation in abundance among species, ranging from 51 in L. sulcatum to 73 in L. dinteri (Fig. 3). The predominant AT composition is consistent with the generally AT-rich nature of plastid genomes and aligns with similar findings in other Caryophyllales taxa (Liu et al., 2024a). The interspecific variation in SSR number and distribution observed here suggests a high level of polymorphism, highlighting their potential value as informative genetic markers. The observed variation in SSR abundance among species suggests that microsatellite evolution remains active within Limeum. Such variation may reflect lineage-specific mutation processes and demographic histories, making these loci particularly useful for investigating population differentiation and genetic connectivity in widespread desert-adapted species. These SSRs provide an important genomic resource for future studies on population genetics, species identification, and phylogeographic inference in Limeum.
Although plastome structure in Limeum was largely conserved, we detected a modest IR expansion that resulted in the duplication of rps19. This feature, which is uncommon in Caryophyllales plastomes, may represent a genus-level synapomorphy for Limeum. Previous studies have suggested that IR expansion can result from gene conversion and double-strand break repair processes, which contribute to shifts in IR boundaries (Wang et al., 2008). Whether this duplication originated from a single ancestral IR expansion event or from multiple independent boundary shifts remains unclear and will require broader sampling across Limeaceae and closely related Caryophyllales lineages. Sequence variation was unevenly distributed, with the LSC and SSC regions exhibiting greater divergence than the IR regions. Non-coding regions also showed higher variability than coding regions (Fig. 5), a pattern widely observed in other angiosperms (Liu et al., 2022; Peng et al., 2022).
Codon usage analysis revealed that leucine was the most frequently encoded amino acid in Limeum, consistent with patterns reported in other plant lineages (Liu et al., 2020, Liu et al., 2024b; Shang et al., 2023). Among the 64 identified codons, those ending in A or T were preferentially used over codons ending in G or C, which is in line with the AT-rich nucleotide composition commonly observed in chloroplast genomes (Nair et al., 2012; Wang et al., 2022). Specifically, 30 codons exhibited RSCU values greater than 1, indicating preferential usage, while 32 codons showed RSCU values below 1 and two had RSCU equal to 1 (Fig. 6). This overall codon usage pattern suggests a translational bias toward AT-rich codons, further reflecting compositional constraints of the Limeum plastome (Ma et al., 2015; Parvathy et al., 2022).
Nucleotide diversity is a widely used indicator of genetic variation within and between species (Li et al., 2021a), and is valuable for detecting highly variable regions suitable for DNA barcoding (Tatarinova et al., 2016). Identifying loci with high polymorphism can provide effective molecular markers to support species identification, phylogenetic resolution, and the establishment of germplasm resources (Menezes et al., 2018; Singh et al., 2020). In the present study, 16 hypervariable regions were detected across the Limeum plastomes (Fig. 7), including five coding genes (clpP, ndhF, psaI, rpl22, and ycf1) and 11 intergenic regions (atpA-atpF_2, psaA-ycf3_3, rpl32- trnL-UAG, rpl33-rps18, rps4-trnT-UGU, rps16_1-trnQ-UUG, trnE-UUC-trnT-GGU, trnH-GUG-psbA, trnK-UUU_1-rps16_2, trnR-UCU-atpA, and trnS-GCU-trnG-UCC_1). Among these, ycf1 is one of the largest coding genes in the chloroplast genome and is widely recognized as a core hypervariable locus in angiosperms (Dong et al., 2015). The same is true in Caryophyllales. For instance, Franck et al. (2012) verified that ycf1 has a higher level of sequence variation than traditional chloroplast markers in Cactaceae, and recommended its application for phylogenetic research in Caryophyllales. Drew and Sytsma (2011) further demonstrated that ycf1 possesses high phylogenetic utility at multiple taxonomic scales within Lamiaceae. Notably, most hypervariable loci were located in intergenic spacers, which is consistent with the reduced functional constraints typically associated with non-coding regions. The coexistence of highly variable coding loci (e.g., ycf1 and ndhF) and intergenic spacers provides a valuable marker set suitable for resolving both deep and shallow evolutionary relationships within the genus.
Chloroplast genomes contain abundant phylogenetically informative characters and have been widely applied to resolve evolutionary relationships across taxonomic scales (Yan et al., 2019; Yao et al., 2019; Li et al., 2021b; Cui et al., 2024). For each dataset, both ML and BI analyses yielded consistent topologies (Figs. 8, 9), supporting the monophyly of Limeum. Within this genus, L. arabicum and L. obovatum formed a fully supported clade (100%, 1.00). This relationship is congruent with morphological observations, as both species possess pseudopposite inflorescences, consistent with their previously recognized affinity (Friedrich, 1956). Limeum viscosum and L. dinteri also clustered together with high support values (100%, 1.00), with L. myosotis recovered as their successive sister lineage. The concordance between molecular and morphological evidence supports the taxonomic value of inflorescence architecture in Limeum and suggests that several traditionally recognized species groups may reflect natural evolutionary lineages. Future studies incorporating the remaining species of the genus will be necessary to evaluate the monophyly of currently recognized infrageneric groups.
The incongruent placement of L. sulcatum between datasets likely reflects differences in evolutionary rate and phylogenetic signal. Complete plastomes integrate information from numerous conserved loci and therefore tend to recover deeper evolutionary relationships, whereas hypervariable regions evolve more rapidly and may better capture recent divergence events. In addition, substitution saturation and homoplastic changes in rapidly evolving regions can obscure true evolutionary history and generate alternative topologies (Rodríguez-Ezpeleta et al., 2007; Philippe et al., 2011).
Complete chloroplast genomes have proven widely recognized as effective “super-barcodes” for species identification and phylogenetic inference (Chen et al., 2018b; Zhang et al., 2019), as they provide more comprehensive genetic information than traditional DNA barcode loci (Wu et al., 2021; Wang et al., 2023). In this study, the complete plastomes of six Limeum species successfully resolved interspecific relationships and clearly discriminated among closely related taxa, highlighting their utility as super-barcodes within the genus. Although only six of the approximately 20 currently recognized species were sampled, this study establishes the first plastome-based genomic framework for Limeum. The newly generated plastome resources, together with the identified hypervariable loci, provide valuable tools for future taxonomic, phylogenetic, and population-level investigations. Expanded species sampling and integration of nuclear genomic data will further clarify evolutionary relationships within the genus and contribute to a more comprehensive understanding of diversification in Limeaceae.
Notes
ACKNOWLEDGMENTS
We are grateful to the Molecular Biology Experiment Center (Germplasm Bank of Wild Species in Southwest China) for providing molecular laboratory facilities. We thank Ruizhu Bai and Shuhan Li for their assistance in data analysis. We also acknowledge the staff of the herbaria K, M, and MW for a chance to obtain the leaves used in the phylogenetic analysis. The study of APS was conducted under the state assignment of Lomonosov Moscow State University.
CONFLICTS OF INTEREST
The authors declare that there are no conflicts of interest.