De Novo assembly and phylogenetic analysis of the complete mitochondrial genome of Eleutherococcus senticosus and related araliaceous species

Abstract

Eleutherococcus senticosus, a unique herb with significant medicinal and ecological importance, has remained largely unexplored regarding its mitochondrial genome. This research details the complete assembly and annotation of the E. senticosus mitogenome, achieved using a hybrid sequencing strategy that integrates data from Illumina short-read and Nanopore sequencing long-read sequencing data. The mitogenome is 548,869 bp in length and encompasses 78 annotated genes, including 38 essential protein-coding genes (PCGs), 24 transfer RNAs, 13 variable genes, and three ribosomal RNAs. An in-depth investigation of repetitive sequences identified simple sequence repeats, tandem repeats, and dispersed repeats, which are implicated in homologous recombination and genomic variability. Prediction of RNA-editing sites uncovered 587 modifications, primarily involving C-to-U transitions that lead to changes in amino acid sequences within PCGs. Furthermore, the chloroplast genome, measuring 156,802 bp, was examined to uncover structural variations and assess mitochondrial-plastid DNA integration events. Phylogenetic analyses utilizing mitochondrial PCGs elucidated the evolutionary position of E. senticosus within the Araliaceae family, demonstrating a close genetic relationship to Panax species while showcasing unique genomic adaptations. These results provide critical insights into the genetic framework, evolutionary processes, and biological functions of E. senticosus, delivering essential genomic data for further studies on phylogenetics, conservation, and molecular breeding of this valuable medicinal plant.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-06809-y.

Introduction

Belonging to the Araliaceae family, Eleutherococcus senticosus (commonly known as Siberian ginseng) has long been valued for its medicinal properties in traditional medicine across East Asian countries such as China, Korea, and Japan [1, 2]. This plant has gained considerable scientific interest due to its pharmacologically active compounds, which include triterpenoid saponins, eleutherosides, flavonoids, and polysaccharides [3, 4]. These bioactive compounds exhibit a broad range of biological activities, including anti-inflammatory, immunomodulatory, antioxidant, and anticancer effects [5–8]. Among these, eleutheroside D is particularly noteworthy for its neuroprotective and adaptogenic properties, while polysaccharides from Acanthopanax have been shown to enhance immune function [9]. Bioactive saponins, especially triterpenoid saponins, are recognized for their role in modulating immune responses and protecting against inflammation-related diseases. Despite the extensive research into its chemical composition and pharmacological effects, there is a significant lack of genomic studies on E. senticosus. While comparative genomic studies on related species within the Araliaceae family, such as Panax notoginseng and Panax. ginseng, have provided insights into the biosynthesis pathways of secondary metabolites, the genomic structure of E. senticosus remains largely unexplored [10–12].

Mitochondria are indispensable organelles in eukaryotic cells, crucial for facilitating energy metabolism and cellular signaling pathways [13–16]. Unlike other cellular organelles, mitochondria possess their own semi-autonomous genome. In plants, research on mitochondrial genomes (referred to as mitogenomes) is of paramount importance for species identification, as well as for studying evolutionary relationships and population genetics [17, 18]. Moreover, mitogenomes play a key role in breeding strategies aimed at enhancing stress resistance traits [19]. These genomes are also essential for investigating cytoplasmic male sterility, a phenomenon vital to hybrid seed production [20–22].Compared to chloroplast and plastid genomes, plant mitogenomes exhibit remarkable structural complexity, with a sparse distribution of highly conserved genes interspersed among abundant non-coding DNA (ncDNA) regions [16]. While plant mitogenomes often exhibit varying degrees of gene loss [23], their overall size is predominantly influenced by the abundance of repetitive sequences, AT-rich regions, large introns, and other non-coding elements, rather than the quantity of genes [24]. Furthermore, the structural diversity of plant mitogenomes includes a variety of configurations, such as circular, linear, branched, and even multi-chromosomal forms [25–28]. This structural variation is influenced, in part, by homologous recombination events facilitated by repetitive sequences, resulting in dynamic alterations to genome organization and structure [29–31].

This study presents a detailed reconstruction and annotation of the mitochondrial and chloroplast genomes (referred to as the cpgenome) of E. senticosus, highlighting the investigation of repeat sequences and codon usage preferences. A comprehensive analysis was conducted to predict RNA-editing sites within mitochondrial protein-coding genes (PCGs), offering insights into potential post-transcriptional modifications. Additionally, the cpgenome was reconstructed, and mitochondrial plastid sequences (MTPTs) were examined to explore inter-organellar gene transfer. Phylogenetic analysis based on mitochondrial PCGs from 26 species was performed, shedding light on the evolutionary relationships between E. senticosus and its close relatives in the Apiales order. Homologous collinear blocks were identified and compared across four species, advancing our understanding of genomic synteny and structural evolution. This research provides a foundation for future investigations into the biological features of E. senticosus and offers essential resources to study genetic evolution and diversity in plant mitogenomes.

Materials and methods

Collection of plant materials, library preparation, and sequencing

Tender leaves of E. senticosus were obtained from Longdong University, located in Gansu Province, China (coordinates: 35°44′10″N, 107°40′03″E). Following collection, the samples were forwarded to the university’s plant taxonomists and experienced herbal practitioners for confirmation of species identification. After verification, the leaves were carefully washed with DEPC-treated water and stored at − 80 °C for subsequent analyses. High-quality genomic DNA was extracted using the CTAB protocol.The quality and concentration of DNA were assessed using 0.75% agarose gel electrophoresis, a Qubit 3.0 Fluorometer (Life Technologies, Carlsbad, CA, USA), and a NanoDrop One spectrophotometer (Thermo Fisher Scientific). For Illumina second-generation sequencing, paired-end libraries with an average insert size of 300 bp were generated using the Nextera DNA Flex Library Prep Kit (Illumina, San Diego, CA, USA) and sequenced on the NovaSeq 6000 platform. For third-generation sequencing, libraries for Oxford Nanopore Technologies (ONT) were prepared using the SQK-LSK110 ligation kit and sequenced on the PromethION system, a high-throughput sequencing device. To ensure high data quality, Illumina short reads were processed with fastp (v0.22.0) to eliminate adapter sequences and low-quality reads [32]. Similarly, ONT long reads underwent quality filtering and adapter removal using Porechop_ABI (v0.5.0) [33] and NanoFilt (v2.8.0) [34].

Methodological approach for DNA barcoding in species identification

A detailed comparison was conducted with specimens from the National Plant Specimen Resource Center of China Digital Herbarium to ensure accurate identification of the samples. The specimen of E. senticosus stored in the Herbarium of the Institute of Botany, Chinese Academy of Sciences, with the specimen number PE02031576 (Fig. 1A), was primarily used as a reference for comparison with the collected samples. To achieve accurate species identification, genomic DNA was used to amplify the rbcL and matK barcode regions [35]. The specific primers and PCR conditions employed are outlined in Table 1. The primer pairs for rbcL amplification, rbcL-F and rbcL-R, and for matK, the KIM-3 F and Kew-5R primer sets, were based on methods from [35]. Additionally, two other matK primer pairs, M1 (M1-F, M1-R) and M2 (M2-F, M2-R), were designed in-house for further refinement. The PCR products were then sent to Sangon Biotech in Shanghai for Sanger sequencing. Once the sequences were obtained, they were compared with the Barcode of Life Data Systems (BOLD) database (https://www.boldsystems.org) to verify species identification. Identification was determined through alignment scores, sequence similarity, and E-values derived from the comparison results.

Fig. 1.

Morphological, Genetic, and Codon Usage Overview of E. senticosus. (A) Morphological characteristics of E. senticosus. The left panel shows the specimen we collected, while the right panel is from the National Plant Specimen Resource Center of China Digital Herbarium, with the specimen number PE02031576. The physical specimen is stored at the Herbarium of the Institute of Botany, Chinese Academy of Sciences. (B) Circular map of the E. senticosus mitochondrial genome, showing gene orientations and GC content. (C) Codon usage patterns in the E. senticosus mitochondrial genome, illustrating the frequency of specific codons

Table 1.

Primer sequences and PCR conditions for target gene amplification

Gene	Primers	Product length (bp)	Sequence 5’-3’	PCR procedure	Refs
rbcL	rbcL-F	599	ATGTCACCACAAACAGAGACTAAAGC	95℃ 3 min; 94℃ 15 s, 53℃ 30s, 72℃ 1 min (35 cycles); 72℃ 5 min	[35]
	rbcL-R		GTAAAATCAAGTCCACCGCG
matK	KIM-3 F	972	CGTACAGTACTTTTGTGTTTACGAG
	Kew-5R	942	GTTCTAGCACAAGAAAGTCG
	Kew-xF		TAATTTACGATCAATTCATTC
	M1-F	925	TTGAGGCGCAACAAGAATTTGTAT		/
	M1-R		CGGTCCAAACCGCCTTACT
	M2-F	1195	TTTGAGGCGCAACAAGAATTTGTAT
	M2-R		CCTCTGCGAAGCGGAAGAT

Assembly and annotation of organelle genomes

The mitogenome of E. senticosus was assembled using a hybrid strategy that effectively leveraged the long-read sequencing from ONT and short-read data from Illumina. Initially, high-quality Nanopore reads were aligned to the mitochondrial sequences of E. senticosus (PP894749), a genome previously assembled using PacBio sequencing data, which has not yet been published. Minimap2 (v2.26) was employed to map the Nanopore reads and isolate the mitochondrial-specific sequences from PP894749 [36]. Subsequently, Canu (v2.2) was employed to correct and trim these long reads to enhance their specificity for mitochondrial assembly [37]. For further refinement, Illumina short reads were aligned to the processed mitochondrial reads using Bowtie 2 (v2.5.1), ensuring accurate selection of mitochondrial data [38]. The hybrid genome assembly was conducted with Unicycler (v0.4.8) under its standard configuration, integrating both long and short reads seamlessly [39]. Contigs with anomalously high coverage, likely originating from the chloroplast, were excluded manually. The assembly’s graphical layout, presented in GFA format, was examined using Bandage (v0.8.1) [40]. For annotation, the complete mitochondrial sequence of E. senticosus was annotated using the GESeq online tool, referencing the mitogenomes of Panax notoginseng (MZ826158.1), Panax ginseng (MZ389476.1), Panax quinquefolius (MZ826160.1), and Panax vietnamensis (MZ826164.1) [41]. To enhance annotation accuracy and standardization, the dataset was further processed using the Plant Mitochondrial Genome Annotator (PMGA), employing its built-in Dataset1 as an additional reference [42]. Discrepancies between automatic annotations were manually curated and resolved using Apollo (v2.5.0) [43]. Visualization of the final mitogenome was performed using OGDRAW (v1.3.1) to generate high-quality graphical representations [44].

Comprehensive analysis of repetitive elements and homologous recombination in mitogenome

Eukaryotic genomes contain a variety of repetitive sequences, which are broadly categorized into simple sequence repeats (SSRs), tandem repeats, and dispersed repeats, based on their structural characteristics and spatial organization within the genome [45–49]. In this research, SSRs were identified using the misa.pl tool (v2.1), with thresholds set for mono-, di-, tri-, tetra-, penta-, and hexanucleotide repeats at 10, 5, 4, 3, 3, and 3 repeat units, respectively [50]. An allowable maximum distance of 1000 bp between two SSRs was applied. Tandem repeats were identified using the Tandem Repeats Finder (TRF v4.09) under specific parameter settings (‘2 7 7 80 10 50 500 -f -d -m’) to ensure precise detection [51, 52]. Dispersed repeats, consisting of non-contiguous repetitive sequences, were identified using the REPuter tool [53] with specific parameters: a Hamming distance of 3, up to 500 repeats detected, and a minimum repeat length of 30 nucleotides. Their distribution across the mitogenome was comprehensively mapped and visualized using Circos (v0.69-8), offering detailed insights into their genomic arrangement [54].

To investigate the role of repetitive sequences in facilitating homologous recombination, we conducted a detailed analysis of repeat pairs along with their 100 bp flanking sequences. Two potential recombination conformations were constructed: a reference conformation representing the original genomic arrangement and a recombinant configuration illustrating the hypothesized rearrangement resulting from recombination events. Nanopore sequencing reads were aligned to both conformations to determine the involvement of specific repeat pairs in mediating recombination. This approach allowed us to assess whether these repetitive elements were actively participating in the genomic rearrangement processes, providing deeper insights into their functional significance in genome dynamics and structural variation. Experimental validation was conducted using Polymerase Chain Reaction (PCR) to amplify the flanking regions of repeat sequences. The PCR reaction mixture was prepared to a final volume of 25 µl, consisting of 1 µl of genomic DNA as the template, 1 µl each of 8 µM forward and reverse primers, 13 µl of 2× Taq PCR Master Mix, and 10 µl of nuclease-free distilled water. The amplification process was initiated with an initial denaturation step at 96 °C for 3 min to ensure complete DNA denaturation. This was followed by 32 cycles of denaturation at 94 °C for 30 s to separate the DNA strands, annealing at 62 °C for 30 s to allow primers to bind to the target sequences, and extension at 72 °C for 1 min to synthesize new DNA strands. A final extension step at 72 °C for 10 min ensured the completion of any remaining DNA synthesis. The resulting PCR products were analyzed for their integrity and size using agarose gel electrophoresis and subsequently sequenced using the Sanger method to confirm the presence of homologous recombination events.

Comprehensive assembly of E. senticosus cpgenome and identification of mitochondrial-plastid DNA transfer

The cpgenome of E. senticosus was assembled utilizing Unicycler (v0.4.8), combining both clean Illumina short reads and ONT long data to ensure high-quality assembly. The genome annotation process was carried out using the CPGAVAS2 web server [55], which facilitated the identification and labeling of genomic features. Visualization of the E. senticosus cpgenome was achieved with OGDRAW (v1.3.1), allowing for the creation of detailed genome maps [44].

Plant mitogenomes often exhibit structural variation due to the presence of dispersed repeats, which promote homologous recombination, generating multiple conformations and adding complexity to the genome. To explore the extent of mitochondrial-plastid DNA transfer in E. senticosus, BLASTn (v2.13.0) was used to identify mitochondrial plastid sequences (MTPTs) within the mitogenome [56]. This transfer of plastid-derived DNA to the mitogenome is a common occurrence in plants. For the analysis, the mitogenome was set as the target sequence, and the plastome served as the query, using BLASTn with an e-value threshold of 1e-6. The identified MTPTs were visualized using TBtools (v2.010) and further annotated through the GESeq platform [57].

RNA-Editing site prediction and codon usage optimization analysis

RNA-editing is a prevalent post-transcriptional process in plant mitogenomes, significantly influencing gene expression and protein functionality. This modification plays a crucial role in fine-tuning genetic information, contributing to the adaptability and regulation of mitochondrial functions in plants [58]. RNA-editing sites were predicted using the PREPACT3 online tool (http://www.prepact.de/prepact-main.php, accessed on 6 Oct 2024). This tool utilizes protein sequences from 25 reference species (Arabidopsis thaliana, Beta vulgaris, Brassica napus, Chaetosphaeridium globosum, Chara vulgaris, Citrullus lanatus, Cucurbita pepo, Isoetes engelmannii, Lotus japonicus, Marchantia polymorpha, Millettia pinnata, Naegleria gruberi, Nicotiana tabacum, Oryza sativa, Physcomitrella patens, Reclinomonas americana, Selaginella moellendorffii, Silene latifolia, Vitis vinifera, Adiantum capillus veneris, Anthoceros formosae, Atropa belladonna, Hevea brasiliensis, and Pisum sativum) within the mitochondrial database, applying a BLASTX cutoff value of 0.001, as described in previous studies [59].

Codon optimization, referring to the preferential use of specific codons to enhance translational accuracy and efficiency, is extensively observed across both prokaryotic and eukaryotic systems [60, 61]. Factors influencing codon bias include evolutionary pressures such as natural selection, species-specific differences, and genetic drift [62]. In addition, PCGs from the mitogenome of E. senticosus were analyzed using TBtools (v2.010). To evaluate codon preferences, the relative synonymous codon usage (RSCU) values were computed with CodonW (v1.4.4) [63].

Phylogenetic and collinear analysis of E. senticosus mitogenome

To construct the phylogenetic framework, mitogenome data from 26 species closely related to E. senticosus across four taxonomic orders were analyzed, with one species from the Solanales order serving as the outgroup. The mitogenome sequences were sourced from GenBank, followed by processing through Phylosuite (v1.2.3) for extraction and format standardization. Although E. senticosus had its genome previously sequenced using the PacBio platform (GenBank accession number PP894749), the data has not yet been published in a formal article. To perform multiple sequence alignment (MSA), the MAFFT (v7.313) tool was utilized [64]. PartitionFinder2 was then applied to identify the optimal evolutionary models for the concatenated dataset [65]. Phylogenetic trees were constructed using the maximum likelihood approach implemented in IQ-TREE2 (v2.1.4) [66], and the visual enhancement of these trees was accomplished with iTOL (v6) for improved presentation and interpretation [67].

In addition, three mitogenomes from the Araliaceae family were selected for the collinearity assessment. BLASTn was employed to identify collinear blocks, applying an e-value threshold of 1e-6. Conserved blocks were defined as homologous sequences exceeding 500 bp in length. The synteny among these genomes was illustrated using the NGenomeSyn tool (v1.41) [68].

Results

Genetic identification of collected specimens

The five pairs of PCR primers successfully achieved specific amplification on the target genes matK and rbcL (Figure S1), with fragment sizes aligning with expectations (Table S1). The database match for the M1 amplification product showed that our collected samples are highly homologous to E. senticosus (Score = 879, Similarity = 99.89, Figure S2). Although the rbcL amplification product also ranked E. senticosus as the top match, it displayed high homology as well (Score = 566, Similarity = 99.65, Figure S3).

Genomic features and structure of the E. senticosus mitogenome

The raw sequencing reads of E. senticosus is now publicly accessible in NGDC under the accession number PRJCA030727 (https://ngdc.cncb.ac.cn). Sequencing efforts included approximately 22 Gb of ONT long-read data with an N50 of 25,116 bp and 32 Gb of paired-end Illumina data (150 bp reads). Second-generation sequencing provided a total of 1,163,027 clean reads, while third-generation sequencing yielded 48,510,823 clean reads. The resulting mitogenome assembly is a complete circular DNA structure spanning 548,869 bp (Fig. 1B, Figure S4).

The mitogenome of E. senticosus, as annotated using the PMGA, comprises a total of 65 unique genes, including 23 core protein-coding genes, 17 variable genes, 4 plastid-derived genes, 3 ribosomal RNA (rRNA) genes, and 24 transfer RNA (tRNA) genes (Table 2, Figure S5).The 23 core protein-coding genes are involved in key mitochondrial functions: five ATP synthase genes (atp1, atp4, atp6, atp8, atp9), nine NADH dehydrogenase genes (nad1, nad2, nad3, nad4, nad4L, nad5, nad6, nad7, nad9), one gene for ubiquinol cytochrome c reductase (cob), four genes associated with cytochrome c biogenesis (ccmB, ccmC, ccmFC, ccmFN), three cytochrome c oxidase genes (cox1, cox2, cox3), and one maturase gene (matR).

Table 2.

Functional classification of mitochondrial-encoded genes in E. senticosus. The genes are grouped into core genes (including ATP synthase, NADH dehydrogenase, ubiquinol cytochrome c reductase, cytochrome c biogenesis, cytochrome c oxidase, and maturases), variable genes (ribosomal protein genes of LSU and SSU, succinate dehydrogenase, and membrane transport genes), plastid-derived genes, ribosomal RNA genes, and transfer RNA genes

Group of genes		Name of genes
Core genes	ATP synthase	atp1, atp4, atp6, atp8, atp9
	NADH dehydrogenase	nad1, nad2, nad3, nad4, nad4L, nad5, nad6, nad7, nad9
	Ubichinol cytochrome creductase	cob
	Cytochrome c biogenesis	ccmB, ccmC, ccmFC, ccmFN
	Cytochrome c oxidase	cox1, cox2, cox3
	Maturases	matR
Variable genes	Large subunit of ribosome (LSU)	rbl2, rbl5, rpl10, rpl16 *
	Small subunit of ribosome (SSU)	rps1, rps3, rps4, rps7, rps8, rps10, rps11, rps12 * , rps13 * , rps14 *
	Succinate dehydrogenase	sdh3, sdh4 (x2)
	Membrane transport	mttB (tatB)#
Plastid-Derived genes		rps12 * , rps13 * , rps14 * , rpl16 *
rRNA genes	Ribosome RNA	rrn5, rrn18, rrn26
tRNA genes	Transfer RNA	trnM-CAU (x5), trnW-CCA (x2), trnQ-UUG (x2), trnP-UGG (x2), trnD-GUC, trnS-UGA, trnY-GUA, trnN-GUU, trnC-GCA, trnK-UUU, trnP-CGG, trnG-GCC, trnS-GCU, trnF-GAA, trnV-GAC, trnH-GUG, trnE-UUC

Note: Numbers in parentheses indicate gene copy counts. * These genes are variably present in angiosperm mitochondrial genomes and/or exhibit plastid origin, pseudogenization, or dual genomic location. # This gene is not commonly found in land plant mitochondria; its annotation remains putative and may reflect horizontal gene transfer or prokaryotic origin

The 17 variable genes include 14 ribosomal protein genes—three from the large subunit (rpl2, rpl5, rpl10) and eleven from the small subunit (rps1, rps3, rps4, rps7, rps8, rps10, rps11, rps12, rps13, rps14, and rpl16). Additionally, two succinate dehydrogenase genes (sdh3 and sdh4) and a putative membrane transport gene (mttB) are annotated. Notably, rps12, rps13, rps14, and rpl16 are variably retained across angiosperms and are considered non-canonical or plastid-derived in many species, suggesting possible horizontal gene transfer or intracellular relocation events. Likewise, mttB exhibits sequence similarity to the bacterial tatB gene, and its presence in plant mitochondria is rare, indicating it may represent a horizontally acquired or misannotated locus. In addition, the genome encodes three rRNA genes (rrn5, rrn18, rrn26) and 24 tRNA genes corresponding to 18 amino acids, with copy number variation detected for trnM-CAU (×5), trnW-CCA (×2), trnO-UUG (×2), and trnP-UGG (×2).

Codon usage analysis

The mitochondrial PCGs of E. senticosus exhibit distinctive codon usage patterns, shedding light on the evolutionary forces and biological processes influencing codon selection. The mitogenome encodes 61 codons to produce 20 amino acids (Fig. 1C), with all amino acids—except methionine (Met)—represented by multiple synonymous codons. This codon usage preference is thought to be shaped by various factors, including natural selection, mutational tendencies, and genetic drift. Codons with a Relative Synonymous Codon Usage (RSCU) value greater than 1 are regarded as preferred. Notable examples include GCU (alanine), CGU (arginine), GGU (glycine), AUA (isoleucine), CUA (leucine), UGG (tryptophan), and GUA (valine), which demonstrate higher frequencies (1.53–1.79), implying potential selective benefits (Table S2). Conversely, codons such as GCG (alanine), CGG (arginine), UUA (leucine), and GUG (valine) are less favored, with RSCU values below 0.5. Furthermore, adenine (A) and uracil (U) are commonly found at the third codon position, reflecting a broader nucleotide preference. On the other hand, stop codons display no discernible usage bias, indicating minimal selective constraints on termination signals within this mitogenome.

Repeat elements analysis

A comprehensive survey of simple sequence repeats (SSRs) in the mitogenome of E. senticosus identified a total of 124 repeats, with lengths ranging from 10 to 84 base pairs (Table S3). As shown in Fig. 2A, these SSRs were classified by nucleotide type—mononucleotide, dinucleotide, trinucleotide, tetranucleotide, pentanucleotide, hexanucleotide, and compound repeats—and further grouped by repeat unit numbers, ranging from three to more than seven. Tetranucleotide repeats (p4) were the most abundant, with 45 instances, followed by compound repeats (c) with 28 occurrences and dinucleotide repeats (p2) with 24 occurrences. Mononucleotide repeats (p1) and trinucleotide repeats (p3) were less frequent, with 11 and 9 instances, respectively. Pentanucleotide (p5) and hexanucleotide (p6) repeats were rare, observed in 5 and 2 cases, respectively. This distribution highlights the predominance of tetranucleotide and compound SSRs in the E. senticosus mitogenome, suggesting their potential contributions to genomic organization and evolutionary adaptation.

Fig. 2.

Analysis of repeat elements in E. senticosus mitogenome. Statistics are shown for the number of (A) SSRs and (B) both tandem and dispersed repeats. (C) Distribution of repeat elements is visualized, with SSRs in the outermost circle, followed by tandem repeats, and dispersed repeats represented by connecting lines in the innermost circle. Pink lines indicate forward repeats, and purple lines mark palindromic repeats

Additionally, Dispersed repeats were the most prevalent, with a total of 396 identified, including 216 forward repeats and 180 palindromic repeats, ranging in length from 30 bp to 11,396 bp (Fig. 2B, Table S4). In contrast, tandem repeats were less abundant, reflecting the diverse distribution and varying roles of these repeat types within the genome (Table S5). The distribution of these repeat elements, including SSRs, tandem repeats, and dispersed repeats, is visually mapped in Fig. 2C, providing an in-depth overview of their locations throughout the mitogenome. This detailed mapping contributes valuable insights into the organization and potential functional impacts of repeat elements within the E. senticosus mitogenome.

Repeated sequence-mediated homologous recombination

Through analysis of ONT read alignment to potential recombinant conformations, three repeat pairs (r28, r160, and r166) were identified as candidates for homologous recombination events. To confirm these recombination products, primers were designed on 100 bp sequences flanking each repeat sequence. Using these primers (F1&R1 and F2&R2), PCR confirmed the presence of both the native genomic sequence and recombined sequences (indicated by F1&R2 and F2&R1 amplification). As shown in Figs. 3A-C, the distinct PCR bands validate the structural diversity induced by recombination. Figure 3D provides a schematic representation of the recombination mechanism facilitated by forward repeats, illustrating the generation of diverse mitochondrial DNA configurations.

Fig. 3.

Structural diversity of mitochondrial DNA driven by repeat-mediated recombination events. (A-C) PCR results showcase the multiple configurations of mitochondrial DNA, with each panel displaying five lanes: (1) molecular marker, (2) and (3) two major conformations, and (4) and (5) two minor conformations, corresponding to recombination within three repeat sequences (r28, r160, r166). The PCR bands clearly demonstrate the structural differences among these conformations. (D) Diagrammatic illustration of the recombination mechanism mediated by forward repeats, highlighting the generation of diverse mitochondrial DNA conformations

Mitochondrial-plastid DNA transfer and genome characterization

The cpgenome of E. senticosus was successfully assembled and annotated, revealing a total length of 156,802 bp (Fig. 4A), which is nearly identical to the assembly reported in (156,768 bp) [69, 70]. Key functional genes identified in the genome include ycf2, rpoC2, psbC, and ndhC, which play vital roles in chloroplast functionality and the photosynthesis process (Fig. 4B). The gene ycf2 is essential for the synthesis of a key chloroplast protein, while rpoC2 is involved in RNA polymerase activity for transcription [71]. psbC is part of the photosystem II complex, critical for light-dependent reactions of photosynthesis, and ndhC is involved in the NAD(P)H dehydrogenase complex, which plays a role in electron transport and energy production within the chloroplast [72]. Further investigation into intracellular sequence transfer revealed the presence of MTPTs, which are DNA fragments transferred from the cpgenome into the mitogenome [73, 74]. A total of 27 MTPTs, spanning 9,293 bp, were found to be shared between the cpgenomes and mitogenomes, representing 1.69% of the mitogenome and 5.93% of the cpgenome of E. senticosus (Table S6). The MTPT lengths ranged from 30 to 1,544 bp, labeled MTPT1 through MTPT27. Some MTPTs contained partial rRNA genes (e.g., rrn16S in cpgenome and rrn18 in mitogenome), specifically in MTPT1, MTPT2, MTPT6, and MTPT7. Additionally, seven MTPTs contained complete tRNA genes, including trn-GAC (MTPT1, MTPT2), trnW-CCA (MTPT5), trnP-UGG (MTPT5), trnN-GUU (MTPT15), and trnM-CAU (MTPT17, MTPT23, MTPT24). Furthermore, 10 MTPTs included partial plastidial PCGs, such as nadC, nadK, nadJ in MTPT3, atpA in MTPT4, petG in MTPT5, and psbC in MTPT8. The results suggest that while tRNA genes were fully transferred and potentially retain functional roles, the transferred PCGs appear to be incomplete and may lack functional significance. These findings underscore the dynamic interplay between organellar genomes and provide essential data for understanding genome evolution and adaptation in E. senticosus.

Fig. 4.

Structural and comparative analysis of E. senticosus cpgenome and MTPT integration. (A) Comprehensive map of cpgenome, with genes color-coded according to functional categories, illustrating its structural organization. (B) Visualization of MTPTs in the cpgenome, highlighting regions of sequence transfer between the cpgenome and mitogenome, which may provide insights into inter-organellar gene transfer. (C) Comparative analysis of LSC, SSC, and IR boundaries among cpgenome from four Araliaceae species, demonstrating differences in genomic structure and boundary positioning across the family

Structurally, the mitogenome comprises the large single-copy (LSC) region (86,178 bp), the small single-copy (SSC) region (18,153 bp), and two inverted repeat (IR) regions, each measuring 25,930 bp (Fig. 4C). These structural regions contribute to the overall integrity and function of the cpgenome. The LSC region is essential for encoding genes involved in photosynthesis, while the SSC region plays a role in the regulation of chloroplast processes. The IR regions are involved in the maintenance of genomic stability and the prevention of large-scale deletions, which is crucial for chloroplast functionality. This comprehensive assembly provided a foundation for further comparative analyses.

The comparative analysis of cpgenome structures among E. senticosus and its related species (P. ginseng, P. vietnamensis, and P. notoginseng) reveals notable differences that emphasize the unique genomic features of E. senticosus (Fig. 4C). Overall, the cpgenome of E. senticosus is slightly larger compared to the Panax species, with the genomes of Panax species ranging in size from similar to slightly smaller sizes. The LSC region in E. senticosus is notably longer than in the Panax species, while the SSC and IR regions remain highly conserved across these species. Key differences are evident at the junctions between the LSC, SSC, and IR regions, particularly in the arrangement and length of crucial genes such as ycf1, ndhF, and rps19 (Table S6). Specifically, E. senticosus displays unique shifts at the JLB (Junction of the LSC region and the Border of the chloroplast genome) and JSA (Junction of the SSC region and the Adjacent IR region) junctions, with the ycf1 gene being longer than its counterparts in Panax species. Significant differences are also observed in the spacer regions between genes such as rps19 and ndhF, as well as between ycf1 and trnH, which further highlight structural variations that may lead to differences in function or regulation. These structural discrepancies suggest potential evolutionary divergence within these genomes.

Variation and RNA-editing site analysis in mitochondrial protein-coding genes

Nucleotide and amino acid sequence alignments of eight protein-coding genes (atp4, ccmB, cox3, nad4L, nad4, nad7, matR, and rps4) were performed for four mitogenomes, including E. senticosus (PP894749), P. ginseng (MZ389476), P. vietnamensis (MZ8296164), and the newly sequenced E. senticosus genome. The alignments, shown in Fig. 5 (Panels A-H), reveal extensive conservation across all species, with most regions exhibiting high sequence similarity. However, specific nucleotide and amino acid differences, highlighted by red boxes, indicate divergence in certain regions. Interestingly, while PP894749 is also derived from E. senticosus, noticeable differences were observed compared to our newly sequenced E. senticosus genome. These variations may stem from differences in sequencing platforms, which could introduce technical discrepancies, or may reflect biological factors such as differences in strain lineage or adaptation to distinct selective pressures. For example, certain genes such as matR and rps4 show sequence mismatches and indels that could be attributed to these factors. Despite these variations, the overall high conservation across the four mitogenomes underscores the evolutionary stability of core mitochondrial genes within the Araliaceae family. These findings suggest that while E. senticosus maintains a largely conserved mitogenomic structure, subtle genetic differences may arise due to environmental or evolutionary pressures, as well as potential technical artifacts from sequencing methodologies. These insights provide a basis for further exploration of intraspecific diversity and the functional implications of the observed variations.

Fig. 5.

Alignments of the nucleotide and amino acid sequences of PCGs from four mitogenomes. Multiple sequence alignments were conducted using MAFFT software with default parameters. Panels (A-H) correspond to the alignment results of the atp4, ccmB, cox3, nad4L, nad4, nad7, matR, cox2, and rps4 genes, respectively. The mitogenomes represented by MZ389476 and MZ8296164 correspond to P. ginseng and P. vietnamensis

RNA editing is a key post-transcriptional process in plant mitochondria, involving specific nucleotide changes that can impact protein functionality [66]. This research focused on predicting RNA-editing sites in mitochondrial PCGs to assess their possible roles in regulating gene expression and protein properties. A total of 587 RNA-editing events were detected across the mitochondrial PCGs, predominantly characterized by C-to-U transitions (Fig. 6A, Table S7). Among the 33 PCGs analyzed, nad4 exhibited the most RNA-editing sites, with 42 modifications, followed by ccmFn and ccmB, which had 35 and 34 sites, respectively. Conversely, genes such as rps10, rps1, atp8, and rps14 displayed fewer than five editing events each. A significant number of these edits resulted in amino acid substitutions, often converting hydrophobic residues to hydrophilic ones, which could potentially impact protein structure and functionality (Fig. 6B, Table S8). These alterations may play a critical role in mitochondrial adaptation by influencing enzymatic activity or protein stability under varying physiological conditions.

Fig. 6.

RNA-Editing Sites and Phylogenetic Relationships in Mitochondrial PCGs of E. senticosus. (A) Distribution of predicted RNA-editing sites across mitochondrial PCGs in E. senticosus, with the number of RNA-editing sites indicated for each gene. (B) Analysis of amino acid substitutions resulting from RNA-editing events, detailing the types and frequencies of observed changes. (C) Phylogenetic tree of 26 species, including E. senticosus, showing their evolutionary relationships within the Apiales order, with E. senticosus highlighted. (D) Heatmap of RNA-editing site distribution across mitochondrial PCGs in various species, illustrating the frequency of gene edits

Evolutionary relationships among closely related species

The phylogenetic tree shown in Fig. 6C illustrates the evolutionary relationships among species from different plant orders, specifically Solanales, Fabales, Sapindales, and Apiales. The analysis uses the mitochondrial PCGs of 27 species as shown in Table S9, with Solanum aethiopicum, a species from the order Solanales, designated as the outgroup. This outgroup choice aids in rooting the tree and providing a reference point for tracing evolutionary divergences within the depicted taxa.

Within each order, individual species branch according to their genetic similarities, with bootstrap values displayed at nodes to indicate the statistical confidence of each bifurcation, where higher values (close to 100) suggest stronger support for the evolutionary relationships shown. Notably, within Apiales, E. senticosus (Siberian ginseng) and its related Araliaceous species (family Araliaceae), such as various Panax species (Panax quinquefolius, Panax ginseng, Panax vietnamensis, and Panax notoginseng), show a close clustering. This grouping indicates significant genetic similarity and a recent common ancestor, underscoring the tight evolutionary link within Araliaceae. This relationship not only reflects the phylogenetic proximity of Araliaceous species but also highlights the evolutionary and genetic characteristics of E. senticosus in relation to its Panax relatives.

This dot plot in Fig. 6D visualizes the presence and abundance of various mitochondrial genes across different plant species, particularly within the context of their orders. The species are grouped by their evolutionary lineage, with E. senticosus and related Araliaceous species highlighted in the middle section, especially in Apiales. In particular, E. senticosus along with its close relatives like Panax species (P. quinquefolius, P. ginseng, P. vietnamensis, and P. notoginseng), display notable gene presence in specific mitochondrial genes, such as those in the nad family (e.g., nad1, nad2, nad4). These genes are crucial components of the mitochondrial electron transport chain, which is fundamental to cellular respiration, suggesting that these Araliaceous species might have similar mitochondrial functionality and evolutionary adaptation patterns. Comparatively, other species such as C. cyminum and D. carota show variation in gene counts across several genes, which may reflect different mitochondrial adaptations or evolutionary paths within Apiales. This comparative gene presence data provides insights into the genetic and functional diversity across these plant species, highlighting how certain mitochondrial genes are conserved within specific lineages, like the Araliaceae family, potentially linking them to similar physiological traits or adaptations. The variation across species in gene abundance can inform future studies on mitochondrial function, evolutionary biology, and the medicinal attributes of these plants, particularly for E. senticosus and its close relatives.

Synteny analysis of plant mitogenomes

Plant mitogenomes are highly conserved in terms of the number, type, and sequence of functional genes, yet these genes often differ significantly in location and order across species. By comparing genome sequences of closely related species, large syntenic (co-linear) blocks can be identified, revealing extensive homology. These conserved regions provide valuable information for assessing evolutionary relationships and genetic kinship among species, as well as aiding in the discovery of novel genes and enhancing genome annotations [75]. The comparative analysis between PP89474794 and E. senticosus reveals a strong dot plot alignment in Fig. 7A, indicating a high degree of sequence similarity and conserved synteny between the two genomes. This result aligns with expectations given their evolutionary proximity. Supporting this observation, Fig. 7B showcases the BLAST analysis, where PP89474794 exhibits the most extensive alignment and higher homology with E. senticosus compared to the other four species.

Fig. 7.

Comparative genomic analysis of E. senticosus and related species. Panel (A) displays dot plots comparing the genomic sequence of E. senticosus against PP894794. Panel (B) shows the BLAST alignment of E. senticosus against four other mitogenomes, with a detailed comparison at the gene level. Panel (C) illustrates a synteny plot comparing the mitogenome of E. senticosus with P. ginseng, highlighting conserved regions and structural similarities. Panel (D) presents dot plots comparing the genomic sequence of E. senticosus against P. ginseng

In addition to sequence-level similarity, structural comparison of the annotated mitochondrial genomes further reinforces their close relationship. The total genome size of PP894749 (547,838 bp) is nearly identical to that of E. senticosus (548,869 bp), with only ~ 1 kb difference, suggesting minimal divergence at the genome scale (Figure S5). Both genomes encode a conserved set of core protein-coding genes, including those for ATP synthase, NADH dehydrogenase (Complex I), cytochrome c oxidase (Complex IV), ribosomal proteins, and maturases. A comparative inspection of tRNA content reveals that the two mitogenomes share a largely overlapping set, including trnM-CAU, trnD-GUC, trnE-UUC, trnY-GUA, trnS-UGA, trnN-GUU, trnQ-UUG, trnW-CCA, and trnG-GCC. However, subtle differences exist; for instance, E. senticosus uniquely contains trnK-UUU and trnP-CGG, while PP894749 features duplicate copies of trnI-CAU and trnP-UGG. rRNA genes also display positional shifts—both genomes contain rrn5, rrn18, and rrn26, yet their genomic context varies, with rrn26 flanking different NADH subunit loci. Furthermore, genes such as ccmFC, ccmFN, cox1, and the nad gene cluster appear in rearranged orders between the two species, suggesting local structural rearrangements despite overall synteny conservation.

To further elucidate sequence-level discrepancies, pairwise alignments of representative protein-coding genes were conducted (Fig. 5). Although the two assemblies exhibit high overall sequence identity, several non-synonymous substitutions are present in PP894749 but absent from the current assembly. These include base changes in conserved coding regions of nad4, cox2, and rps3, which lead to amino acid replacements such as L→P, F→S, or V→I. The observed differences, confirmed at both the nucleotide and amino acid levels, may result from sequencing errors or annotation inconsistencies in PP894749. These findings collectively indicate that while the two assemblies are closely related, the newly assembled E. senticosus mitogenome offers improved structural accuracy and sequence fidelity.

The mauve collinearity analysis comparing E. senticosus with P. ginseng revealed numerous homologous syntenic blocks, indicating conserved regions among the genomes (Fig. 7C). However, the dot plot analysis highlighted sparse collinear regions between the mitogenomes of E. senticosus and P. ginseng, reflecting low overall synteny (Fig. 7D). Similar observations were made through a multiple synteny plot (Figure S6), which demonstrated that the co-linear blocks were rearranged differently across the five closely related mitogenomes. These findings suggest that E. senticosus has undergone extensive genomic rearrangements (Table S10). As a result, the mitogenome of E. senticosus exhibits a highly variable structure, emphasizing its low conservation in genomic organization.

Discussion

The comprehensive assembly, annotation, and analysis of plant mitogenomes and cpgenomes are fundamental to advancing our understanding of plant biology and genomics [15, 76]. As key organelles responsible for energy production and photosynthesis, their genomes are indispensable for investigating plant physiology, evolutionary history, and adaptive responses to environmental challenges. However, the structural complexity and prevalence of repetitive sequences in plant mitogenomes make their assembly particularly difficult [23, 25]. In this study, the complete mitogenome and cpgenome of E. senticosus were successfully assembled simultaneously using a hybrid sequencing strategy that combined Illumina short reads with ONT long reads. These high-quality assemblies serve as a foundational resource for exploring the molecular mechanisms underpinning the biology of E. senticosus. Moreover, they provide valuable reference genomes that facilitate comparative genomic studies within the Apiales order and support future efforts in the molecular breeding and genetic improvement of this species.

The combination of DNA barcoding techniques, including the amplification of rbcL and matK regions, demonstrates a robust methodological approach for accurate species identification, as successfully applied to E. senticosus. The dual-marker approach, leveraging rbcL and matK, capitalizes on their complementary strengths—rbcL for its universality and matK for its high discriminatory power. The use of multiple primer sets, including in-house-designed primers, further enhanced the specificity and efficiency of amplification, as evidenced by the successful PCR results with expected fragment sizes. By utilizing specimens from the National Plant Specimen Resource Center, the study ensured reliable taxonomic comparison and verification. This integrated methodological combination not only validates the identity of E. senticosus but also exemplifies a scalable framework for other plant species, highlighting its potential utility in taxonomic studies, conservation efforts, and biodiversity assessments [77, 78]. Such a methodological combination ensures high accuracy and reliability, particularly for species with complex genomic structures or closely related taxonomic groups.

The comparative analysis of cpgenome structures among E. senticosus and its related species (P. quinquefolius, P. ginseng, P. vietnamensis, and P. notoginseng) highlights significant differences that underscore the distinct genomic characteristics of E. senticosus. These variations are likely driven by contrasting ecological niches and evolutionary pressures experienced by these species [2, 19]. E. senticosus thrives in the temperate regions of East Asia, including northern China, Korea, and Japan, where it is exposed to cooler climates and unique environmental challenges. In contrast, Panax species are predominantly found in subtropical and tropical regions with relatively stable and warm conditions [79]. The specific adaptive pressures in E. senticosus ‘s habitat, such as seasonal temperature fluctuations and varying soil nutrient availability, may have contributed to the observed structural rearrangements and expansions in its cpgenome.

To support the genomic findings, the alignment and sequence conservation of key protein-coding genes within the mitogenome of E. senticosus and its closely related species were analyzed. These genes display notable nucleotide substitutions and indels, particularly in the coding regions, which may result in alterations to protein function or stability. For instance, the atp1 and ccmB genes exhibit unique sequence modifications in E. senticosus, indicating potential adaptations associated with energy metabolism and stress response mechanisms. Similarly, variations in nad genes, which play a critical role in the electron transport chain, highlight the influence of ecological and evolutionary pressures on mitochondrial function. These differences in sequence conservation and structural organization emphasize the genomic divergence of E. senticosus, reflecting its distinct adaptive strategies and evolutionary trajectory within the Araliaceae family.

Phylogenetic analysis of the complete mitogenomes of E. senticosus and its closely related Panax species also provides a strong basis for advancing genetic and evolutionary studies. While E. senticosus exhibits phylogenetic proximity to Panax species, it retains distinct genomic features that highlight its evolutionary divergence. These variations, coupled with the above distinct cpgenome characteristic, are likely the result of geographic isolation, ecological adaptation, and genetic drift, which together have shaped its speciation and genomic architecture [80, 81]. This analysis offers valuable insights into the genetic diversity and evolutionary pathways within the Araliaceae family. The comparative genomic approach not only enhances our understanding of the unique genetic attributes of E. senticosus but also elucidates the broader evolutionary dynamics within this family. Moreover, these findings significantly contribute to the field of plant genomics by revealing the complex interplay of genetic and environmental factors driving diversification in Araliaceous species. The knowledge gained from this study lays a foundation for further exploration of genome evolution, adaptive mechanisms, and the development of targeted strategies for conservation, cultivation, and potential biotechnological applications.

In summary, the genomic differences between E. senticosus and its Panax relatives highlight the interplay of evolutionary dynamics, ecological adaptability, and habitat-driven selection pressures. These results offer significant insights into the evolutionary mechanisms driving genetic diversity within the Araliaceae family. By highlighting the genetic relationships and adaptations among species, this study lays a foundation for future research into their evolutionary trajectories and ecological specialization.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Abbreviations

MTPTs

Mitochondrial plastid DNAs

tRNAs

Transfer RNAs

RSCU

Relative synonymous codon usage

Leu

Leucine

APG

Angiosperm Phylogeny Group

PCGs

Protein coding genes

GC

Guanine-cytosine

rRNA

Ribosomal RNA

Val

Valine

MSA

Multiple sequence alignment

Arg

Arginine

Gly

Glycine

mt

Mitochondrial contig

Ser

Serine

SSRs

Simple sequence repeats

bp

Base pairs

IR

Inverted repeat

LSC

Large single-copy

SSC

Small single-copy

NADH

Nicotinamide adenine dinucleotide hydride

ATP

Adenosine triphosphate

ccm

Cytochrome c biogenesis genes

N50

Median read length of the longest contigs

DEPC

Diethyl pyrocarbonate

CTAB

Cetyltrimethylammonium bromide

PCR

Polymerase chain reaction

RNA

Ribonucleic acid

Author contributions

Z. Wang contributed to the conception, design, data acquisition, analysis, and drafting of the manuscript. H. Wang provided substantial input in data interpretation and offered critical revisions to the manuscript. X. Gong assisted in data interpretation and manuscript revision. Y. Guo supported experimental procedures and manuscript preparation. X. Ou was responsible for sample collection and identification. M. Tang supervised the project, overseeing study design, data interpretation, and manuscript revision, and provided final approval of the version to be published.

Funding

This research was supported through funding from several prestigious sources, including the City Social Development Project of Zhenjiang (SH2024080) and (SH2024038), the Zhenjiang Science and Technology Plan Projects (Project No. SH2023078), and the Medical Education Collaborative Innovation Fund of Jiangsu University (Grant No. JDYY2023009). These contributions were essential for the successful completion of this work.

Data availability

The raw sequencing data from both Illumina and Nanopore platforms generated in this study have been deposited in the Genome Sequence Archive at the National Genomics Data Center (NGDC), China National Center for Bioinformation (CNCB), Beijing Institute of Genomics, Chinese Academy of Sciences, under accession number CRA019353 (https://ngdc.cncb.ac.cn/gsa) [82]. Additionally, the data reported in this paper have also been deposited in GenBase at NGDC/CNCB under accession number C_AA107960.1, publicly accessible at https://ngdc.cncb.ac.cn/genbase [83].

Declarations

Ethics approval and consent to participate

This study was conducted in full compliance with institutional, national, and international ethical standards. Plant materials were sourced from publicly available resources, including authorized collections and databases, ensuring adherence to relevant legal and regulatory frameworks. No specific permits were required for the collection and analysis of Eleutherococcus senticosus samples. All experimental procedures were approved by the ethical committee of Jiangsu University, ensuring rigorous oversight of the research methodology. The study did not involve endangered or protected species, and all sampling followed sustainable and ethical practices.

Consent for publication

All authors have reviewed and approved the final version of the manuscript and consent to its publication. No individual or identifiable personal data is included in this study, and as such, specific consent for publication from individuals is not applicable. The authors affirm that the work is original and has not been submitted or published elsewhere.

Competing interests

The authors declare no competing interests.