Chuanxin Mo1†Wenshuai Chen2†Zhuolin Lv1
Jun Zhao1Yuchan Li3Qiannan Shi1Kaimi Dong1Zhen Wei1
Zeru Yu1Xueling Wang1Chao Xin1Zhen Ni2
Ma Yu1*
Hua Chen1*- 1College of Life Sciences and Agri-forestry, Southwest University of Science and Technology, Mianyang, Sichuan, China
- 2Institute of Plateau Biology of Xizang Autonomous Region, Lhasa, Xizang, China
- 3Jining Polytechnic, Jining, Shandong, China
The root biomass and saikosaponins yield of Bupleurum chinense DC. are crucial factors determining its economic value. This study developed triploid F1 hybrid materials of the B. chinense by crossing a diploid (2n=2x1 = 12, x1 = 6) maternal parent with a tetraploid (2n=4x2 = 20, x2 = 5) paternal parent. The resulting hybrids exhibited a stable intergenomic karyotype (2n=x1+2x2 = 16) and significant heterosis. Two-year field trials confirmed strong over-dominance in root architecture, with root dry weight exceeding that of the diploid and tetraploid parents by 37.3% and 166.5%, respectively. The yield of the bioactive compounds saikosaponin A (13.42 mg) and D (13.20 mg) increased by an average of 60.7% and 57.5%, respectively, compared to the diploid parent, highlighting substantial potential for pharmaceutical development. Based on the transcriptome comparison in the seedling and maturity stage of the root, the remarkable heterosis might be supported by unique genomic architecture and sophisticated transcriptional reprogramming. The intergenomic imbalance, likely provide a stable foundation for heterosis by facilitating functional compartmentalization and synergistic interaction between the parental subgenomes. Transcriptome analysis revealed that the heterotic traits were arranged by a complex relationship of gene networks: root morphology was optimized through the additive and transgressive expression of hormone signaling genes, while the enhanced synthesis of saikosaponins was driven by the synergistic expression of key biosynthetic genes. This research provides a novel strategy for exceed conventional plant breeding by demonstrating engineered genomic asymmetry, specifically intergenomic triploid and could help to unlock superior and stable heterosis.
1 Introduction
Radix Bupleuri, the dry root of plants in the Bupleurum genus, has been extensively employed in traditional Chinese medicine for more than 2000 years (Ashour and Wink, 2011; Teng et al., 2023; Zeng et al., 2023). Species in this genus are widely distributed across East and Central Asia, North America, North Africa, and Europe (Pan, 2006; Zeng et al., 2023). According to the Chinese Pharmacopoeia (2025), the official sources of Radix Bupleuri are the dried roots of Bupleurum chinense and B. scorzonerifolium, which are recognized for their therapeutic effects, including antipyretic, analgesic, hepatoprotective, and antidepressant properties (Teng et al., 2023).
Due to increasing market demand, natural populations of B. chinense and B. scorzonerifolium have experienced significant depletion in recent years. Therefore, cultivation and breeding efforts have become essential to ensure sustainable supply. Current breeding strategies for B. chinense include the introduction of cultivated varieties from other regions, domestication of local wild germplasm, and selection of superior individuals to develop locally adapted varieties (Zheng et al., 2010; Xu et al., 2018). Hybridization breeding is widely employed in agriculture because of its ability to generate offspring exhibiting heterosis (hybrid vigor), where hybrids outperform their parents in yield and growth (Schnable and Springer, 2013; Kim and Zhang, 2018). For instance, heterosis has contributed to yield increases of 3.5–15% in wheat (Whitford et al., 2013), 55% in rice (Yuan, 1994), and up to 200% in canola (Shehzad et al., 2015). Similar effects have been observed in medicinal plants, including Panax ginseng (Seo et al., 2019), licorice (Glycyrrhiza spp.) (Lee et al., 2017), Chrysanthemum (Deng et al., 2012), and Chinese bellflower (Platycodon grandiflorus) (Wei et al., 2011). In the genus of Bupleurum L., an interspecific F1 hybrid between B. chinense and B. marginatum exhibited significant heterosis in root yield and growth adaptability (Xu, 2019). Interspecific hybrid materials are generally not adopted in commercial medicinal‐material production because their safety, efficacy, and quality stability require rigorous risk assessment. Despite the demonstrated potential of hybridization breeding, polyploid‐based hybrid breeding, particularly triploid formation through asymmetric ploidy crosses, has not yet been systematically explored in B. chinense, leaving its effects on growth performance and secondary metabolism largely unknown.
Hybridization leads to the combination of divergent alleles, often resulting in structural rearrangements and epigenetic reprogramming. These molecular events induce heritable changes in gene expression and phenotype, contributing to species diversification and evolution (Jackson, 2017; Nieto Feliner et al., 2020). A commonly observed phenomenon in hybrids is “transcriptome shock,” wherein the merging of distinct parental genomes triggers non-additive gene expression patterns, with F1 expression levels deviating from the mid-parent value (Rapp et al., 2009). Among non-additively expressed genes, transgressive expression refers to expression levels that exceed the parental range, while expression-level dominance describes gene expression in the F1 that matches one parent but diverges significantly from the other (Wu et al., 2016; Díaz-Valenzuela et al., 2023). Transcriptome shock also entails extensive rewiring of regulatory networks and differential partitioning of parental homeolog expression (Quan et al., 2022).
Saikosaponins, the principal bioactive constituents of B. chinense, are pentacyclic triterpenoid saponins synthesized through a multistep biosynthetic pathway (Yu et al., 2020; Sui et al., 2011). The biosynthesis begins with the formation of the triterpenoid backbone, primarily via the cytosolic mevalonate (MVA) pathway, although the plastidial methylerythritol phosphate (MEP) pathway also contributes. The MVA pathway starts from acetyl-CoA and proceeds through a series of enzymatic reactions to generate isopentenyl diphosphate (IPP), a key precursor. IPP is subsequently converted to squalene, which is epoxidized to form 2,3-oxidosqualene. This intermediate is cyclized by oxidosqualene cyclases (OSCs) to yield various triterpenoid skeletons (Xue et al., 2012; Wang et al., 2022). These scaffolds undergo further structural diversification through modifications catalyzed by cytochrome P450 monooxygenases (P450s) and UDP-glycosyltransferases (UGTs), which introduce functional groups and glycosyl residues, respectively, thereby enhancing the bioactivity and chemical diversity of saikosaponins (Zhang et al., 2019; Li et al., 2023). Although heterologous biosynthesis offers a promising platform for the production of complex natural products, it requires comprehensive elucidation of metabolic intermediates and pathway enzymes (Park et al., 2020). The objectives of the present study were (a) to compare root morphological traits and saikosaponin profiles between F1 hybrids and their parents; (b) to investigate the genetic and transcriptomic differences using RNA sequencing; and (c) to identify heterosis- regulating root development and saikosaponin biosynthesis.
2 Materials and methods
2.1 Plant materials and hybridization
This study utilized an F1 generation derived from a cross between two commercial varieties of B. chinense (Chuanbeichai #1 and Chuanchai #2), along with both parents (Figure 1). Chuanbeichai #1 (hereafter CBC) is a diploid (2n = 2x1) cultivar cultivated primarily in northern China (Hebei, Shanxi, and Gansu provinces), is characterized by a high dry root weight and elevated saikosaponin content, but exhibits moderate tolerance to soil waterlogging. The male parent of Chuanchai #2 (hereafter: CC2) is an atuotetraploid (2n = 4x2) line introduced from Rongxian county in Sichuan province and was officially registered in 2023. This cultivar shows strong adaptation to the rainy weather of summer in Sichuan province but produces a relatively low dry root yield. Because pistil maturation in CBC lagged behind stamen development, uniformly developed inflorescences were selected and pruned to prevent self-pollination. After the stamens are removed manually, CC2 at full flowering was used as the pollen donor, and F1 hybrids were obtained by controlled pollination. Seeds for F1 and both cultivars were kindly provided by Dr. Yu from the College of Life Sciences and Agri-forestry, Southwest University of Science and Technology.
Figure 1. Comparative analysis of root biomass and karyotypes in F1 hybrids and parental lines. (A) Root architecture phenotypes at seedling stage, from left to right are CBC,F1, and CC2; (B) Root architecture phenotypes at maturity stage, from left to right are CBC,F1, and CC2; (C–E) Fluorescence in situ hybridization patterns of CBC (C), F1 hybrid (D), and CC2 (E) using 5S rDNA probes (cyan pseudo-color) and 18S rDNA probes (red pseudo-color).
2.2 Fluorescence in situ hybridization
Roots of B. chinense (1.5–2.0 mm in length, containing active root apical meristems) were harvested from 7-day-old in vitro cultured seedlings. To enrich for metaphase cells, the cultures were exposed to nitrous oxide gas at 1 MPa for two hours to synchronize mitosis. Following treatment, approximately 5 mm root tips were excised, minced, and subjected to enzymatic digestion in a solution of 1% pectolyase and 2% cellulase Onozuka at 37°C for one hour. The resulting cell suspension was centrifuged, and the pellet was resuspended in 90% acetic acid. The suspension was placed onto clean glass slides within a humidified chamber to prevent drying.
Chromosomes were stained with 4′,6-diamidino-2-phenylindole (DAPI), and visualized under a fluorescence microscope (Zeiss LSM880, Germany). For fluorescence in situ hybridization (FISH), probes specific to 5S and 18S rDNA repeat sequences were used to detect multiple chromosomal loci. The probes were denatured in hybridization buffer (50% formamide, 10% dextran sulfate, 2×SSC) at 80°C for 10 minutes, hybridized overnight at 42°C, and stringently washed with 2×SSC/0.1% Tween-20 at 55°C. Methodological details are available in Zhang et al. (2022). Fluorescent signals were visualized using confocal microscopy (Leica Microsystems, Wetzlar, Germany). Karyotype construction was based on measurements of relative chromosome length and centromeric index, with a minimum of 15 well-spread metaphase cells analyzed per sample.
2.3 Field experiment and phenotypic evaluation
Parental lines and F1 hybrids were cultivated at the Longshan Research Farm, Southwest University of Science and Technology, Mianyang, China (31°32′N, 104°42′E) across two growing seasons—sown on March 12th, 2021, and March 28th, 2022, respectively. The experiment employed a randomized complete block design (RCBD) with three replicates. Each plot (1.5 × 8.0 m) contained two rows with 25 cm row spacing and 3–5 cm plant spacing. Standard agronomic practices were maintained throughout the growing season to ensure uniform growth (Luo, 2024). Plant height was measured in October each year. For each plot, fresh root weight, root length, and taproot diameter were determined from 10 randomly selected plants after harvesting. Roots were oven-dried (Thermo Fisher Scientific, USA) at 120°C for 30 minutes followed by 60°C for 72 hours to determine dry root weight.
On July 18, 2024, parents and F1 materials were grown in a controlled-environment greenhouse at Southwest University of Science and Technology using a completely randomized design with three replicates. Greenhouse conditions were maintained at 24°C and 40–60% relative humidity. Plants were individually planted in round pot (10 cm diameter × 11 cm heigh), with 10 plants per replicate. Root length and fresh weight were recorded prior to bolting, with drying procedures identical to the field experiment.
2.4 Stress treatment
To validate candidate genes, seeds of the CC2 line were sown in a mixed substrate of peat moss and vermiculite (1:1, v/v) and cultivated under controlled greenhouse conditions. At the five-true-leaf stage, seedlings were subjected to irrigation treatments as follows: (1) 100 g/L polyethylene glycol (PEG-6000) to simulate drought stress; (2) 100 μmol/L methyl jasmonate (MeJA) to simulate MeJA stress, according to established protocols (Sui et al., 2011; Yang et al., 2022). Control seedlings received distilled water. Root tissues were harvested at 0, 1, 2, 4, and 8 hours post-treatment initiation, flash-frozen in liquid nitrogen, and stored at −80°C for subsequent RNA extraction, saikosaponins quantification and gene expression validation. For each time point, seedlings from every three pots were pooled to constitute one biological replicate, yielding three biological replicates per treatment group.
2.5 Determination of saikosaponin content
Dried roots were ground to a fine powder through a 60-mesh sieve (Wanshi, China). For each sample, 0.5 g of root powder was extracted in 25 mL of 5% ammonia-methanol solution using ultrasonication for 30 min and then lyophilized. The dried extract was redissolved in 10 mL methanol. Saikosaponins A and D were quantified using a HPLC system (Waters, USA) equipped with an ASB-vensil C18 column (4.6 × 250 mm, 5 μm). Reference standards were obtained from the National Institutes for Food and Drug Control (Beijing, China), and analysis was performed following the method described by Xu (2019).
2.6 RNA-seq library preparation and sequencing
Fresh root samples of CBC, F1, and CC2 were collected in October 2021 for maturity and in October 2024 for seedlings, with three biological replicates for each sample. The roots were immediately frozen in liquid nitrogen and kept at -80°C for RNA extraction. Then, the samples were sent to Novogene Bioinformatics Technology Corporation (Beijing, China) for RNA-sequencing. The library construction and sequencing process have been reported by Yu et al. (2021).
2.7 RNA-seq data analysis and gene function annotation
The raw sequencing data were filtered by FastQC (Andrews, 2023) and Trimmomatic (Bolger et al., 2014). Clean reads were mapped to the CC2 genome (https://ym-lab.vip.cpolar.cn) using HISAT2 for further analysis (Kim et al., 2019). The reads count of each gene and original expression matrix were generated by featureCounts software (Liao et al., 2014). The matrix was then normalized by edgeR (Robinson et al., 2010) package in R (v. 4.2.2). The principal component analysis (PCA) and Pearson correlation coefficient analysis were conducted using the PCAtools in R. DESeq2 (Love et al., 2014) was used to identify differential expression genes (DEGs) among the two parents and F1 generation. Significant DEGs were filtered using a false-discovery rate (FDR) threshold of < 0.01 and log2 (FoldChange) > 1. Venn diagram analysis between samples was conducted using the TBtools (Chen et al., 2020). The software ClusterProfiler (Yu et al., 2012) was applied to perform KEGG enrichment analysis. The KEGG (Kanehisa et al., 2023) background was acquired from eggNOG-mapper v2 (Cantalapiedra et al., 2021). KEGG pathway analysis was applied to the “enricher()” function for enrichment. A hypergeometric test with a P-value threshold of 0.05 was applied.
2.8 Co-expression network analysis and hub gene identification
Weighted Gene Co-expression Network Analysis (WGCNA) was conducted with the R package WGCNA. The built-in function “goodSamplesGenes()” (with minFraction = 1/2) and the R package genefilter (with var.cutoff = 0.5) were applied to filter the normalized expression matrix. The filtered genes were used to construct a weighted co-expression network. The ME (module eigengene, the first principal component of a module) value was computed for each module to assess the relationship with the trait. The key genes were identified among the genes in the module most strongly associated with the trait, based on their connectivity (high kME values) and the number of network connections (more edges). The network of key genes was visualized using cytoscape 3.10 (Kohl et al., 2011). The top score hub genes were computed and sorted with the MCC method implemented in the Cytoscape plugin cytoHubba (Chin et al., 2014).
2.9 Candidate genes validation
Ten DEGs associated with root morphotype and saikosaponin biosynthesis were selected for qRT-PCR validation. Gene-specific primers were designed using Primer3Plus (Untergasser et al., 2012) with the following parameters: amplicon size = 80–150 bp, primer length = 18–25 nt, Tm = 58–62°C. Exon-spanning primers were employed to prevent genomic DNA amplification. Primer sequences are listed in Supplementary Table S1.
qRT-PCR amplification was performed using cDNA templates from: (1) stress-treated CC2 seedlings and (2) three parental genotypes used for RNA sequencing. Reactions utilized the TransStart Top Green qPCR SuperMix Kit (TransGen Biotech, China) on a LightCycler 96 system (Roche Diagnostics, Switzerland). The BcADF5 gene (Yu et al., 2019) served as the endogenous reference. Gene expression levels were quantified using the 2−ΔΔCt method (Livak and Schmittgen, 2001). Three biological replicates were analyzed, each with three technical replicates.
2.10 Statistical analysis
Analysis of variance (ANOVA) for all traits was performed using SPSS software v21.0 (SPSS Inc., Chicago, USA).
3 Results
3.1 Karyotype analysis
The root-related traits of F1 hybrids were better than those of both parental lines during both seedling (Figure 1A) and maturity (Figure 1B) stages. Through karyotype analysis, the parental line CBC (2n=2x1 = 12, x1 = 6) was found to possess 12 somatic chromosomes (Figure 1C), while CC2 (2n=4x2 = 20, x2 = 5) contained 20 chromosomes (Figure 1D). The F1 hybrids (2n=x1+2x2 = 16) exhibited 16 chromosomes (Figure 1E), half of the combined parental chromosome number. FISH with 18S rDNA probes revealed that both parental lines CBC and CC2 exhibited two fluorescent signals localized on homologous chromosomes, which were conserved in F1 hybrids. The 5S rDNA probe signals followed a similar inheritance pattern to chromosome numbers: CBC displayed two signals, CC2 showed four signals, and F1 hybrids carried three signals.
3.2 Phenotypic evaluation
At the seedling stage, the F1 hybrids displayed significant over-dominance (P < 0.05) in lateral root number, root diameter, root fresh weight, root dry weight, content of saikosaponin A and D, and the yields of saikosaponin A and D (Figure 2; Supplementary Table S2). For instance, F1 hybrid developed 29 lateral roots, exceeding CBC and CC2 by 1.2-fold and 1.5-fold, respectively; root dry biomass (0.17 g) exceeded CBC and CC2 by 89% and 325%, respectively; while saikosaponin A yield ranked F1 (0.61 mg) > CBC (0.34 mg) > CC2 (0.10 mg), with saikosaponin D following an identical trend.
Figure 2. Statistical and correlation analysis of root-related traits and content of saikosaponin in F1 and parental lines. (A) Lateral root number; (B) Root diameter; (C, D) Fresh root weight and dry root weight; (E, F) saikosaponin A and D content; (G, H) Average yield of saikosaponin (A, D) per plant; The bar represents the standard error (SE). Different letters represent significant difference (P < 0.05).
Two-year field trails confirmed F1 hybrids over-dominance in root architecture (lateral root number, diameter, fresh/dry weight) and saikosaponin yields (Figures 2A-H; Supplementary Table S2). Mean root dry weight reached 2.50 g in F1, exceeding CBC (1.83 g) by 37.3% and CC2 (0.94 g) by 166.0%. Saikosaponin yields averaged 13.42 mg/plant for saikosaponin A (60.7% higher than CBC; 111.2% higher than CC2) and 13.20 mg/plant for saikosaponin D (57.5% higher than CBC; 103.3% higher than CC2). Although Saikosaponins contents in F1 were lower than in CC2, they surpassed CBC by 16.3% (saikosaponin A) and 14.3% (saikosaponin D).
Correlation analysis (Supplementary Figure S2) revealed that lateral root number negatively correlated with root diameter (rn-rd= -0.52), dry weight (rn-dw = -0.48), and fresh weight (rn-fw = -0.43). Extreme collinearity occurred between dry and fresh weights (rdw-fw = 0.99), followed by root diameter and dry/fresh weight (r = 0.97). Saikosaponin A and D metrics showed positive associations with all root size parameters.
3.3 Transcriptome profile analysis
Transcriptome sequencing of 18 biological samples (2 stages × 3 genotypes × 3 replicates) generated 54.18 Gb clean data. All libraries exceeded quality thresholds (Q30 > 94.29%, mapping rate > 80.0%), with high inter-replicate correlations (r>0.95) (Figure 3A; Supplementary Table S3). PCA showed dominant stage and genotype separation (PC1 = 85.33% variance; PC2 = 9.26% variance) (Figure 3B).
Figure 3. Bioinformatics analysis of transcriptome differential genes between F1 and parental lines. (A) Dendrogram heat map; (B) Principal component analysis; (C) Venn diagram of differential genes in seedling stage; (D) Venn diagram of differential genes in maturity stage; (E) Double volcano map between F1 and parental lines at seedling stage; (F) Double volcano map between F1 and parental lines at maturity stage.
Differential expression analysis revealed genotype-distinct transcriptional profile, with seedling-stage (Figure 3C; Supplementary Figure S3A) comparisons identifying 7,772 DEGs between CBC and CC2 (2,771 up-/5,001 down-regulated), 3,516 DEGs between F1 and CBC (2,457 up-/1,059 down-regulated), and 5,255 DEGs between F1 and CC2 (2,114 up-/3,141 down-regulated). At the maturity stage (Figure 3D; Supplementary Figure S3B), DEG counts decreased to 6,586 for CBC vs CC2 (2,771 up-/3,815 down-regulated), 2,956 for F1 vs CBC (1,787 up-/1,169 down-regulated), and 3,364 for F1 vs CC2 (1,611 up-/1,753 down-regulated). Volcano plots detailed F1-specific expression patterns: 19 genes were upregulated in F1 versus both parents at the seedling stage (Figure 3E), increasing to 24 genes in maturity (Figure 3F).
Transcriptomic shock classification (Rapp et al., 2009) showed stage-dependent divergence (Table 1): maternal-dominant expression prevailed at both stages (seeding with 2,168 DEGs; maturity with 1,623 DEGs), followed by paternal-dominance. While additive patterns decreased 73.9% from seedling to maturity stages. Over-dominant genes declined by 58.3%, from 242 DEGs at seedling stage to 101 DEGs at the maturity stage.
Table 1. Possible additive and nonadditive gene expression patterns in a F1 hybrid relative to its parents.
The KEGG pathway analysis identified 17 significantly enriched metabolic pathways (P < 0.05) at the seedling stage and 21 at the maturity stage, with 11 core pathways conserved across both developmental stages including phenylpropanoid biosynthesis, sesquiterpenoid and triterpenoid biosynthesis, zeatin biosynthesis, and brassinosteroid biosynthesis (Supplementary Figure S4). Phenylpropanoid biosynthesis exhibited the highest DEG enrichment, displaying 115 and 79 DEGs in CBC vs. CC2 comparisons at seedling and maturity stages respectively, while F1 vs. CBC comparisons revealed 47 (seedling) and 32 (maturity) DEGs. Notably, F1 vs. CC2 comparisons showed 79 DEGs at the seedling stage but an absence of DEGs at maturity.
3.4 WGCNA and hub gene identification
WGCNA identified 31 trait-associated modules from 39,996 seedling-stage genes (Figure 4A; Supplementary Figure S5A). Lateral root number, root diameter, biomass, and saikosaponin-related traits exhibited significant positive correlations (P < 0.05) with the MEbrown module (5,821 genes) and MEgreenyellow module (670 genes) and the correlation coefficient were above 0.73, while showing negative correlations (P < 0.01) with the MEpink module and the correlation coefficient was less than -0.7. Root length demonstrated exclusive positive correlations with MEwhite and MEtan modules (P < 0.05). Analysis of 39,332 maturity-stage genes revealed 28 modules (Figure 4B; Supplementary Figure S5B). Root architecture traits and saikosaponin yields correlated positively with MEpink, MEmagenta, and MEMidnightblue modules (r > 0.83, P < 0.05), but negatively with MEdarkred module (r = -0.87, P < 0.01). Lateral root number only showed positive associations with MEred (1,772 genes) and MEgreen modules (1,790 genes).
Figure 4. Weighted gene co-expression network analysis (WGCNA) of root genes at seedling and maturity stages. (A) Correlation analysis of gene modules and phenotypic traits at seedling stage; (B) Correlation analysis of gene modules and phenotypic traits at maturity stage. The modules on the left are color-coded, with the number of genes under the module name. Heatmap showing correlations between co-expression modules (vertical axis) and traits (horizontal axis). P-values are shown inside parentheses; (C) Venn diagram of genes related to saikosaponins content (SSA, SSD, Total_SSA and Total_SSD; module cor > 0.5) in the co-expressed gene modules at seedling and maturity; (D) Venn diagram of genes related to root morphology (RL, LRN, RD, RFW and RDW; module cor > 0.5) in the co-expressed gene modules at seedling and maturity; (E, F) Co-expression network of Saikosaponins and root morphology related modules; The size of nodes represents the Degree value of genes; The color of nodes represents the ClosenessCentrality value; and the color of the outer ring of nodes represents gene function or family classification.
Comparative analysis of phenotype-associated modules across developmental stages identified 5,717 conserved genes in saikosaponin-related modules (Figure 4C) and 9,195 in root morphology (Figure 4D) modules, prompting construction of dedicated co-expression networks. The saikosaponin biosynthesis network (Figure 4E) exhibited significant enrichment in specialized metabolic pathways including phenylpropanoid biosynthesis (ko00940), terpenoid backbone biosynthesis (ko00900), and saikosaponin-modifying enzymes—particularly cytochrome P450s and UGTs. Among its top 10 hub genes, functional annotation revealed two P450s (Hap1_chr2G6951, Hap1_chr3G2362), three UGTs (Hap1_chr1G1890, Hap1_chr1G999, Hap1_chr3G3900), three phenylpropanoid pathway genes (Hap1_chr2G565, Hap1_chr4G1377, Hap1_chr3G6767), and two transcription factors (WRKY family member Hap1_chr1G9955 and PHD-finger protein Hap1_chr1G2044). While, the root architecture network (Figure 4F) showed pronounced enrichment in plant hormone signal transduction (ko04075), with core hubs dominated by transcription factor families: MYB proteins, AP2/ERF regulators, and GRAS members. Its top 10 hub genes comprised four hormone signaling components, three AP2 factors, one GRAS transcription factor (Hap1_chr3G5679), and one GARP family regulator.
3.5 Key genes associated with root morphology
Integrated WGCNA and differential gene analysis revealed 147 root morphology-associated DEGs enriched in plant hormone signaling pathways (Figure 5; Supplementary Figure S6), distributed as follows: 51 DEGs in the auxin (Aux) pathway, 24 in the cytokinin (CTK) pathway, 11 in the gibberellin (GA) pathway, 15 in the abscisic acid (ABA) pathway, 9 in the ethylene (ETH) pathway, 14 in the brassinosteroid (BR) pathway, 15 in the jasmonic acid (JA) pathway, and 8 in the salicylic acid (SA) pathway.
Figure 5. The core differentially expressed genes (Log2Foldchange > 1) from WGCNA analysis associated with root related traits in F1 and parental lines. In the heat map, the three columns on the left are seedling samples (CBC, F1, CC2 in order), and the three columns on the right are maturitysamples (CBC, F1, CC2 in order), and the heat map data of seedling and maturitysamples are normalized in “row” respectively.
Within the auxin signaling pathway, the AUX/IAA gene family comprised 18 differentially expressed genes (DEGs), the GH3 family contained 8 DEGs, and the SAUR family included 21 DEGs. Notably, BcIAA13 (Hap1_chr1G8281) exhibited significant downregulation in parental comparisons (CBC vs. CC2) across both developmental stages, while demonstrating additive expression in F1 hybrids (Table 1).
Within the cytokinin signaling pathway, eight GARP-family ARR-B genes and seven ARR-A transcriptional regulators were identified. The gibberellin pathway contained eight GRAS-family DELLA genes, while ethylene signaling featured eight AP2-family AP2-ERF genes. Notably, two ethylene-responsive ERF genes (Hap1_chr1G4638 and Hap1_chr2G1753) exhibited transgressive upregulation (Table 1) in F1 hybrids across both developmental stages—a pattern mirrored in brassinosteroid signaling by TCH4 enzymes Hap1_chr1G9738 and Hap1_chr2G8015. In jasmonic acid pathways, ten bHLH-family MYC2 genes were detected, including Hap1_chr2G8944 with significantly elevated expression in CC2 (Padj < 0.05; 4-fold > F1, 7-fold > CBC). This gene displayed additive expression in seedlings and maternal-biased low expression in maturity (Table 1). Salicylic acid pathways contained four bZIP-family TGA genes, while abscisic acid signaling featured five bZIP-family ABF regulators.
3.6 Genes associated with saikosaponins synthesis
Integrated WGCNA-KEGG analysis revealed 179 DEGs enriched in saikosaponin biosynthesis, spanning three pathway stages. (Figure 6A). Terpenoid backbone biosynthesis (ko00900) contained 21 enzyme-encoding DEGs, including three HMGR genes (Hap1_chr4G869, Hap1_chr4G1155, Hap1_chr4G5502) - key rate-limiting enzymes in the MVA pathway - and a single-copy DXR gene (Hap1_chr4G5524) from the MEP pathway that showed significant inter-parental expression divergence (P < 0.01). Sesquiterpenoid/triterpenoid biosynthesis (ko00909) featured eight DEGs: one squalene synthase (SS, Hap1_chr2G7650), three squalene epoxidases (SQE, Hap1_chr4G6682, Hap1_chr5G7293, Hap1_chr1G9139), and four saponin synthases (β-AS, Hap1_chr1G5523, Hap1_chr1G5343, Hap1_chr1G5366, Hap1_chr3G323). Post-modification stages comprised 148 DEGs, including 102 DEGs encoding CYP450s enzymes and 46 DEGs encoding UGTs (UDP-glycosyltransferases). The P450 genes were classified into three subfamilies: CYP71 (59 genes), CYP72 (30 genes), and CYP85 (13 genes). While, UGT genes were categorized into four subfamilies: UGT71 (four genes), UGT73 (22 genes), UGT74 (nine genes), and UGT91 (11 genes).
Figure 6. The core differentially expressed genes (Log2Foldchange > 1) from WGCNA analysis associated with saikosaponins content in F1 and parental lines. (A) In the heat map, the three columns on the left are seedling samples (CBC, F1, CC2 in order), and the three columns on the right are maturity samples (CBC, F1, CC2 in order), and the heat map data of seedling and maturity samples are normalized in “row” respectively; (B, C) Evolutionary relationship of differential expressed genes in hub Cytochromes P450 (P450s) and UDP-glucuronosyltransferases (UGTs) gene families, respectively. The number on the branch of the evolutionary tree represents the bootstrap value.
Phylogenetic relationships between DEGs encoding P450s and UGTs and functionally characterized catalytic genes were analyzed (Figure 6B). In the P450s phylogenetic tree, Hap1_chr3G7452 clustered most closely with KU878849.1, which encodes a C-28 oxidase. Three tandemly duplicated genes (Hap1_chr5G3928, Hap1_chr5G3929, and Hap1_chr5G3926) formed a clade with AHF45909.1 (100% support), a known C-16α hydroxylase. Gene Hap1_chr1G5524 segregated on a distinct branch with Bupleurum-derived JF803813, a gene of unknown function. In the UGTs phylogeny (Figure 6C), Hap1_chr1G9522 showed closest affinity to C-3 glycosyltransferase MH819286.1, whereas Hap1_chr3G5569 was phylogenetically proximate to C-28 glycosyltransferase VhUGT74M1 (DQ915168.1. Additionally, Hap1_chr4G4244 and Hap1_chr3G8747 demonstrated the strongest homology to putative glycosyltransferases BcUGT5 (JF803821.1) and BcUGT3 (JF803819.1), respectively.
3.7 Candidate genes function verification
Candidate hub genes identified from transcriptome analysis were selected for further functional verification in B. chinense. These included two auxin response factors (BcIAA13.1, BcSAUR24.1), two cytochrome P450 genes (BcCYP716Y1.1, BcCYP716A83.1), and two glycosyltransferase genes (BcUGT73.1, BcUGT74.1).
Under drought stress, six candidate genes exhibited dynamic expression patterns across all measured time points (Figures 7A-F). BcIAA13.1 and BcSAUR24.1 were sustainedly up-regulated. In contrast, BcCYP716A83.1 and BcUGT74.1 expression peaked at 2 h. Gene BcCYP716Y1.1 reached its maximum at 2 h, while BcUGT73.1 showed a delayed peak at 4 h. Under MeJA treatment, the expression profiles differed: BcIAA13.1 and BcSAUR24.1 peaked at 4 h and 2 h, respectively. Genes BcCYP716A83.1 and BcUGT74.1 were initially down-regulated but subsequently stimulated, both reaching their highest expression at 8 h. Gene BcCYP716Y1.1 was consistently up-regulated, peaking at 8 h, and BcUGT73.1 expression peaked at 2 h before decreasing. The HPLC analysis results showed (Figures 7G, H) that the contents of saikosaponin A and D in B. chinense were significantly higher under drought and MeJA treatment compared to the control (CK). The effect of MeJA on the accumulation of saikosaponin D was significantly greater than that of drought stress (Figure 7H).
Figure 7. Candidate genes response and Saikosaponins accumulation under abiotic stress. (A, B) Auxin early response genes: BcIAA13.1, BcSAUR24.1; (C, D) Cytochromes P450: BcCYP716Y1.1, BcCYP716A83.1; (E, F) UDP-glucuronosyltransferases: BcUGT73.1, BcUGT74.1; (G, H) Saikosaponin A and D contents in Abiotic stress. Different letters represent statistically significant expression differences (P < 0.05).
To further validate the transcriptome results, qRT-PCR was performed on seedling-stage samples of the F1 and parental lines using the six genes with differential expression identified under abiotic stress (Supplementary Figure S7). The expression trends obtained by qRT-PCR were highly consistent with the RNA-seq data, confirming the reliability of the transcriptome-derived differential expression profiles. In detail, BcIAA13.1, BcCYP716Y1.1, BcUGT73.1 and BcUGT74.1 displayed additive expression patterns, BcCYP716A83.1 followed the parental expression pattern, and BcSAUR24.1 exhibited transgressive up-regulation, all in agreement with the RNA-seq analysis (Table 1).
4 Discussion
This study developed novel triploid B. chinense germplasm through intraspecific hybridization, as summarized by a schematic model of heterosis mechanisms (Supplementary Figure S8): diploid CBC (Chuanbeichai #1, 2n=2x1 = 12, x1 = 6) was crossed as female parent with tetraploid CC2 (Chuanchai #2, 2n=4x2 = 20, x2 = 5) as male parent, yielding an F1 triploid population (2n=x1+2x2 = 16). The hybrids exhibited significant heterosis, with dry root yield reaching 1.37-fold that of CBC and total saikosaponin (A+D) content increasing to 1.59-fold, highlighting their pharmaceutical potential. Triploid plants which harbor three sets of chromosomes and exhibit significant advantages in biomass, metabolism, and stress tolerance are common in plants, particularly in in crops harvested for vegetative organs and seedless fruit crops (Wang et al., 2016; Ramakrishnan et al., 2025). For instance, watermelon ‘Vertigo’ (2n=3x=33) achieved yields up to 45.95 t/ha (Cushman et al., 2003). Cavendish banana (2n=3x=33) combined high productivity with strong resistance to Fusarium wilt (Li et al., 2024). Triploid dominance was also found in medicinal plants. Triploid Siraitia grosvenorii (2n=3x=42) enhanced yield and processing quality (Sreekumari et al., 1999) and produced fruits with 36.28% higher mogroside V content than diploids (Wei et al., 2022). Triploid Eucommia ulmoides (2n=3x=51) surpassed the diploids in growth rate, photosynthetic efficiency, and secondary metabolite accumulation (Li et al., 2019). Salvia miltiorrhiza triploids show 215.33% greater fresh root weight with tanshinone content exceeding Chinese Pharmacopoeia standards (Li and Chen, 2012). Erigeron breviscapus triploids produced 214% more biomass while maintaining 13% higher scutellarin content than diploids (Wu et al., 2011).
FISH revealed intergenomic triploid F1 hybrids derived from the CC2 (x=5) × CBC (x=6) cross, characterized by unequal basic chromosome numbers between parental genomes. This architecture differs fundamentally from conventional triploids and standard aneuploidy involving numerical variation within a single chromosome set (Zamariola et al., 2014). Typical triploids exhibit chromosome numbers exactly triple the haploid base (e.g., cassava 2n=3x=54, x=18; banana/watermelon 2n=3x=33, x=11) (Ramakrishnan et al., 2025). The observed intergenomic imbalance might also provide a stable genomic foundation for sustained heterosis. Chromosomal rearrangement between subgenomes could facilitate transcriptional reprogramming, enable modular partitioning of key biological pathways—such as root development and specialized metabolite biosynthesis—into complementary subgenomic units (Chen, 2013; Deb et al., 2023). This functional compartmentalization strategy might optimize existing parental pathways rather than relying on de novo evolution and show a sub-genome recombination case for optimizing perennial medicinal crops breeding.
Heterosis in B. chinense F1 hybrids was observed at both seedling and maturity stages, as revealed by integrated phenotypic and transcriptomic analyses. Since the root is the medicinal organ of this species, root biomass and the concentration of bioactive compounds determine its therapeutic quality. Two-year field trials and greenhouse experiments demonstrated significant heterotic advantages in root architecture traits including lateral root density, diameter, and biomass yield. Transcriptional upregulation underpins this heterosis, with parental-dominance expression patterns (> 70%) predominating among DEGs, and approximately 3% of DEGs in the F1 hybrids exhibited transgressive upregulation at both seedling and maturity stages (Table 1). This expression pattern diverged from rapeseed studies where transgressive regulation dominated F1 hybrids (Shalby et al., 2021) but aligned with rice research reporting prevalent paternal-dominant expression patterns. The triploid genomic constitution may further enhance metabolic capacity through gene dosage effects and allelic complementation between divergent parental genomes, whereby favorable alleles jointly optimize regulatory robustness and pathway efficiency (Chen, 2013; Schnable and Springer, 2013). In current study, significant heterosis manifested at both developmental stages, with greater DEG abundance observed during rapid seedling development. These findings suggested advantageous traits in mature plants might originate from developmental priming during early growth phases.
Root morphology-associated DEGs were enriched in plant hormone signaling transduction pathways. Auxin, as a key growth regulator, played a critical role in lateral root development. Within the auxin signal transduction pathway, the SCFTIR1/AFB-Aux/IAA-ARF module regulated the downstream genes expression. In this study, five genes in Aux/IAA family expressed relatively higher in parents than in the F1 generation. Through comparative transcriptome (Yu et al., 2021), B. chinense exhibited lower expression of the auxin inhibitor BcIAA13 (Hap1_chr1G8281) but greater lateral root density than B. scorzonerifolium. This inverse correlation aligned with root suppression in rice Osiaa13 gain-of-function mutants (Kitomi et al., 2012). The remaining four Aux/IAA genes were reported to inhibit lateral root initiation in different species. For instance, an auxin-resistant Arabidopsis mutant, Atiaa28-1, exhibited severe defects in lateral root formation, and reduced maturity size and decreased apical dominance (Rogg et al., 2001). In addition, Segmental duplication might also influence root development across plant species (Fan et al., 2018). Transgenic apple plants overexpressing MdIAA27T exhibited enhanced tolerance to phosphorus deficiency, characterized by the development of longer and denser adventitious roots (Zhao et al., 2022). Sl-IAA27 expression was up-regulated by the AM fungus and had a positive impact on AM colonization (Guillotin et al., 2017). Ectopic expression of the auxin-responsive gene TrIAA27 in Arabidopsis has been shown to promote biomass accumulation and improve tolerance to both drought and salt stress (Iqbal et al., 2024). Two tandem BcIAA27 duplication pairs were identified on chromosome 4 in this study. BcIAA27.1 (Hap1_chr4G1934) and BcIAA27.2 (Hap1_chr4G1944) shared 100% sequence identity with about 120 kb segmental duplication; while, BcIAA27.3 (Hap1_chr4G5053) and BcIAA27.4 (Hap1_chr4G5064) exhibit >98% similarity with about 95 kb duplication. In the downstream of the auxin transduction pathway, sixteen tandemly arranged SAUR genes on chromosome 4 were found, among which only BcSAUR24.1 (Hap1_chr4G363) and BcSAUR50.1 (Hap1_chr4G379) exhibited differential expression in the F1 generation. Together, the additive expression of BcIAA13 in the hybrid might stabilize auxin responsiveness and prevent excessive inhibition of lateral root formation, while the tandemly duplicated BcIAA27 and BcSAUR genes might provide a genomic basis for enhanced environmental adaptability—demonstrating how hormone-related transcriptomic reprogramming modulates both biomass heterosis and adaptive trait optimization in B. chinense.
In B. chinense, more than 100 types of saikosaponins were reported, among which saikosaponin A and saikosaponin D were the primary medicinal components. On the pentacyclic triterpenoid saponins biosynthesis pathway, the post-modification genes P450s and UGTs have been extensively studied in different species (Li et al., 2023). The protein of P450s could participate in the hydroxylation and oxidation at the C-11, C-16, C-21, C-23, and C-28 positions of triterpenoid saponins. Phylogenetic analysis revealed that differentially expressed P450s were mainly grouped into subfamilies CYP72, CYP71, and CYP85, respectively (Figure 6B). In the Bupleurum genus, only CYP716Y1 (AHF45909.1) classified into CYP85, has been cloned from B. falcatum (Moses et al., 2014), which catalyzed the C-16α hydroxylation of triterpenes. In the current study, a tandem of BcCYP716Y1.1 (Hap1_chr5G3928) and BcCYP716Y1.2 (Hap1_chr5G3929) were clustered in this group, and showed a significant inter-parental expression difference (P < 0.01), suggesting its potential role as a key driver of metabolic innovation in saikosaponin profiles. Zhang et al. (2022) identified 266 P450s genes from the genome of B. chinense, the sequences of these genes were unpublished, it was unable to determine the homology of these genes with those in our experiment. Transcriptome analysis by Sui et al. (2011) identified a candidate P450 oxidation gene, BcCYP716A41 (JF803813.1), which shares high sequence identity (>98%) with Hap1_chr1G5524, suggesting that they might represent the same gene. In addition, BcCYP716A83.1 (Hap1_chr3G7452) might possess a similar function to CaCYP716A83 (KU878849.1), as they were clustered within the same phylogenetic clade. Further investigation is warranted to elucidate the functions of the 30 differentially expressed P450 genes in the CYP72 subfamily and the 53 differentially expressed P450 genes in the CYP71 subfamily. As for UGTs genes, glycosylation mostly happened at C-3, C-23 and C-28 positions of saikosaponins. Phylogenetic analysis indicated that Hap1_chr4G4244 shared high identity (>84%) with BcUGT5 (JF803821.1), and Hap1_chr3G8747 shared high identity (>89%) with BcUGT3 (JF803819.1) (Sui et al., 2021), suggesting that these sequences might represent the same gene (Figure 6B). In ginsenosides, the glycosyltransferase gene PzOAGT3 (MH819286) transferred glucuronic acid at C-3 positions of oleanolic acid to form oleanolic acid 3-O-β-glucuronide (Tang et al., 2019). This gene showed a high sequence similarity (>77%) to BcUGT73.1 (Hap1_chr1G9522), suggesting potential functional similarities (Figure 6C). The gene UGT74M1 (DQ915168) from Saponaria vaccaria forming a glucose ester at C-28 position during monodesmoside biosynthesis (Meesapyodsuk et al., 2007). This gene was closely associated with BcUGT74.1 (Hap1_chr3G5569), indicating a potential role in catalyzing glycosylation at the C-28 position (Figure 6C). Collectively, members of CYP71/72 and UGT73/74 subfamilies emerged as dominant players in the saikosaponin modification stage, with BcCYP716Y1.1 exemplifying a parental-expression-divergent gene that may underpin novel metabolite profiles in the hybrid. Such specialized metabolic rewiring, in concert with hormone-regulatory networks, likely contributes to the optimization of both medicinal quality and ecological adaptability in B. chinense.
In addition, the mechanisms enabling hybridization between parents with unequal basic chromosome numbers, the stability of heterosis-associated transcriptional reprogramming during further domestication, and the feasibility of large-scale seed production require further investigation. Future studies combining multi-environment trials and functional genomics approaches will be essential to evaluate the stability and application potential of triploid breeding in B. chinense.
5 Conclusions
This study reported the karyotype profiles of two dominant B. chinense cultivars (CBC: 2n=2x1 = 12, x1 = 6; CC2: 2n=4x2 = 20, x2 = 5) and their triploid F1 hybrids (2n=x1+2x2 = 16), revealing a unique ploidy-driven genomic architecture. Alongside demonstrating root biomass heterosis, the F1 hybrids achieved a breakthrough in saikosaponin biosynthesis efficiency. The remarkable heterosis might be underpinned by unique genomic architecture and sophisticated transcriptional reprogramming. Intergenomic imbalance established a stable foundation for heterosis by facilitating functional compartmentalization and synergistic interaction between the parental subgenomes. Transcriptome analysis revealed that the heterotic traits were arranged by a complex relationship of gene networks: root morphology was optimized through the additive and transgressive expression of hormone signaling genes, while the enhanced synthesis of saikosaponins is driven by the synergistic and divergent expression of key biosynthetic genes. This research might exceed conventional plant breeding by demonstrating engineered genomic asymmetry, specifically intergenomic triploid and could help to unlock superior and stable heterosis.
Data availability statement
The RNA-seq data that support the findings of this study have been deposited into CNSA with accession number CNP0008950.
Author contributions
CM: Investigation, Writing – original draft, Data curation, Visualization. WC: Writing – review & editing. ZL: Writing – original draft, Visualization, Validation, Data curation. JZ: Supervision, Writing – review & editing. YL: Data curation, Investigation, Writing – review & editing. QS: Methodology, Formal analysis, Writing – original draft. KD: Writing – original draft, Formal analysis, Data curation. ZW: Writing – original draft, Validation. ZY: Writing – original draft, Visualization. XW: Writing – original draft, Visualization. CX: Writing – review & editing, Supervision. ZN: Writing – review & editing. MY: Writing – review & editing, Supervision, Funding acquisition. HC: Project administration, Funding acquisition, Writing – review & editing, Conceptualization.
Funding
The author(s) declared that financial support was received for this work and/or its publication. This research was funded by the China Agriculture Research System (CARS-21), the CAMS Innovation Fund for Medical Sciences (2021-I2M-1-032), the Science and Technology Department of Sichuan Province (2023YFH0044, 2023YFH0018), and the Regional Innovation Cooperation Project of the Science and Technology Department of Sichuan Province (2026YFHZ0163).
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Generative AI statement
The author(s) declared that generative AI was not used in the creation of this manuscript.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpls.2026.1736464/full#supplementary-material
References
Andrews, S. (2023).FastQC: A quality control tool for high throughput sequence data. Available online at: http://www.bioinformatics.babraham.ac.uk/projects/fastqc (Accessed January 27, 2026).
Ashour, M. L. and Wink, M. (2011). Genus Bupleurum: a review of its phytochemistry, pharmacology and modes of action. J. Pharm. Pharmacol. 63, 305–321. doi: 10.1111/j.2042-7158.2010.01170.x
Bolger, A. M., Lohse, M., and Usadel, B. (2014). Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinf. (Oxford England) 30, 2114–2120. doi: 10.1093/bioinformatics/btu170
Cantalapiedra, C. P., Hernández-Plaza, A., Letunic, I., Bork, P., and Huerta-Cepas, J. (2021). eggNOG-mapper v2: Functional Annotation, Orthology Assignments, and Domain Prediction at the Metagenomic Scale. Mol. Biol. Evol. 38, 5825–5829. doi: 10.1093/molbev/msab293
Chen, Z. J. (2013). Genomic and epigenetic insights into the molecular bases of heterosis. Nat. Rev. Genet. 14, 471–482. doi: 10.1038/nrg3503
Chen, C., Chen, H., Zhang, Y., Thomas, H. R., Frank, M. H., He, Y., et al. (2020). TBtools: an integrative toolkit developed for interactive analyses of big biological data. Mol. Plant 13, 1194–1202. doi: 10.1016/j.molp.2020.06.009
Chin, C. H., Chen, S. H., Wu, H. H., Ho, C. W., Ko, M. T., and Lin, C. Y. (2014). cytoHubba: identifying hub objects and sub-networks from complex interactome. BMC Syst. Biol. 8 Suppl 4, S11. doi: 10.1186/1752-0509-8-S4-S11
Cushman, K. E., Horgan, T. E., Snyder, R. G., Hudson, P. M., Coker, C. H., and Ely, M. (2003). Evaluation of elongated and oval triploid (seedless) watermelon genotypes. Annual Report, 398, 339–345.
Deb, S. K., Edger, P. P., Pires, J. C., and McKain, M. R. (2023). Patterns, mechanisms, and consequences of homoeologous exchange in allopolyploid angiosperms: A genomic and epigenomic perspective. New Phytol. 238, 2284–2304. doi: 10.1111/nph.18927
Deng, Y., Jiang, J., Chen, S., Teng, N., Song, A., Guan, Z., et al. (2012). Combination of multiple resistance traits from wild relative species in Chrysanthemum via trigeneric hybridization. PLoS One 7, e44337. doi: 10.1371/journal.pone.0044337
Díaz-Valenzuela, E., Hernández-Ríos, D., and Cibrián-Jaramillo, A. (2023). The role of non-additive gene action on gene expression variation in plant domestication. EvoDevo 14, 3. doi: 10.1186/s13227-022-00206-4
Fan, M., Ma, W., Liu, C., Zhang, C., Wu, S., Chen, M., et al. (2018). Evolution and expression characteristics of receptor-like cytoplasmic protein kinases in maize, rice and Arabidopsis. Int. J. Mol. Sci. 19, 3680. doi: 10.3390/ijms19113680
Guillotin, B., Etemadi, M., Audran, C., Bouzayen, M., Bécard, G., and Combier, J. P. (2017). Sl-IAA27 regulates strigolactone biosynthesis and mycorrhization in tomato (var. MicroTom). New Phytol. 213, 1124–1132. doi: 10.1111/nph.14246
Iqbal, M. Z., Liang, Y., Anwar, M., Fatima, A., Hassan, M. J., Ali, A., et al. (2024). Overexpression of auxin/indole-3-acetic acid gene TrIAA27 enhances biomass, drought, and salt tolerance in Arabidopsis thaliana. Plants (Basel Switzerland) 13, 2684. doi: 10.3390/plants13192684
Jackson, S. A. (2017). Epigenomics: dissecting hybridization and polyploidization. Genome Biol. 18, 117. doi: 10.1186/s13059-017-1254-7
Kanehisa, M., Furumichi, M., Sato, Y., Kawashima, M., and Ishiguro-Watanabe, M. (2023). KEGG for taxonomy-based analysis of pathways and genomes. Nucleic Acids Res. 51, D587–D592. doi: 10.1093/nar/gkac963
Kim, D., Paggi, J. M., Park, C., Bennett, C., and Salzberg, S. L. (2019). Graph-based genome alignment and genotyping with HISAT2 and HISAT-genotype. Nat. Biotechnol. 37, 907–915. doi: 10.1038/s41587-019-0201-4
Kim, Y. J. and Zhang, D. (2018). Molecular control of male fertility for crop hybrid breeding. Trends Plant Sci. 23, 53–65. doi: 10.1016/j.tplants.2017.10.001
Kitomi, Y., Inahashi, H., Takehisa, H., Sato, Y., and Inukai, Y. (2012). OsIAA13-mediated auxin signaling is involved in lateral root initiation in rice. Plant Sci. 190, 116–122. doi: 10.1016/j.plantsci.2012.04.005
Kohl, M., Wiese, S., and Warscheid, B. (2011). Cytoscape: software for visualization and analysis of biological networks. Methods Mol. Biol. (Clifton N.J.) 696, 291–303. doi: 10.1007/978-1-60761-987-1_18
Lee, S.-H., Kim, Y. B., Lee, J.-M., Lee, J.-H., Park, C.-B., Bang, J.-W., et al. (2017). Cultivation and Characteristics of Licorice F1 (Glycyrrhiza glabra x G. uralensis) Lines. Hortic. Sci. Technol. 35, 276–286. doi: 10.12972/kjhst.20170030
Li, X. L. and Chen, L. (2012). Cultivation of triploid Salvia miltiorrhiza and its sustainable utilization research. Chin. Traditional Herbal Drugs 43, 375–379. Available online at: https://www.cnki.net/ (Accessed January 27, 2026).
Li, Y., Wang, J., Li, L., Song, W., Li, M., Hua, X., et al. (2023). Natural products of pentacyclic triterpenoids: from discovery to heterologous biosynthesis. Natural product Rep. 40, 1303–1353. doi: 10.1039/d2np00063f
Li, Y., Yang, J., Song, L., Qi, Q., Du, K., Han, Q., et al. (2019). Study of variation in the growth, photosynthesis, and content of secondary metabolites in Eucommia triploids. Trees 33, 817–826. doi: 10.1007/s00468-019-01818-5
Li, X., Yu, S., Cheng, Z., Chang, X., Yun, Y., Jiang, M., et al. (2024). Origin and evolution of the triploid cultivated banana genome. Nat. Genet. 56, 136–142. doi: 10.1038/s41588-023-01589-3
Liao, Y., Smyth, G. K., and Shi, W. (2014). featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinf. (Oxford England) 30, 923–930. doi: 10.1093/bioinformatics/btt656
Livak, K. J. and Schmittgen, T. D. (2001). Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods (San Diego Calif.) 25, 402–408. doi: 10.1006/meth.2001.1262
Love, M. I., Huber, W., and Anders, S. (2014). Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550. doi: 10.1186/s13059-014-0550-8
Luo, Y. (2024). Research on growth regulation and root rot prevention of the Bupleurum Variety Chuanbeichai No.1 (Master’s thesis, Southwest University of Science and Technology). doi: 10.27415/d.cnki.gxngc.2024.001643. (Accessed January 27, 2026)
Meesapyodsuk, D., Balsevich, J., Reed, D. W., and Covello, P. S. (2007). Saponin biosynthesis in Saponaria vaccaria. cDNAs encoding beta-amyrin synthase and a triterpene carboxylic acid glucosyltransferase. Plant Physiol. 143, 959–969. doi: 10.1104/pp.106.088484
Moses, T., Pollier, J., Almagro, L., Buyst, D., Van Montagu, M., Pedreño, M. A., et al. (2014). Combinatorial biosynthesis of sapogenins and saponins in Saccharomyces cerevisiae using a C-16α hydroxylase from Bupleurum falcatum. Proc. Natl. Acad. Sci. United States America 111, 1634–1639. doi: 10.1073/pnas.1323369111
Nieto Feliner, G., Casacuberta, J., and Wendel, J. F. (2020). Genomics of evolutionary novelty in hybrids and polyploids. Front. Genet. 11. doi: 10.3389/fgene.2020.00792
Pan, S.-L. (2006). Main Chemical Constituents of Bupleurum Genus. 71–95. doi: 10.1201/9781420009071-7
Park, D., Swayambhu, G., and Pfeifer, B. A. (2020). Heterologous biosynthesis as a platform for producing new generation natural products. Curr. Opin. Biotechnol. 66, 123–130. doi: 10.1016/j.copbio.2020.06.014
Quan, C., Chen, G., Li, S., Jia, Z., Yu, P., Tu, J., et al. (2022). Transcriptome shock in interspecific F1 allotriploid hybrids between Brassica species. J. Exp. Bot. 73, 2336–2353. doi: 10.1093/jxb/erac047
Ramakrishnan, R., Banadka, A., Dubey, S., Al-Khayri, J. M., and Nagella, P. (2025). Advances in triploid plant production: Techniques, benefits, and applications. Plant Cell Tissue Organ Culture (PCTOC) 160, 70. doi: 10.1007/s11240-025-03011-8
Rapp, R. A., Udall, J. A., and Wendel, J. F. (2009). Genomic expression dominance in allopolyploids. BMC Biol. 7, 18. doi: 10.1186/1741-7007-7-18
Robinson, M. D., McCarthy, D. J., and Smyth, G. K. (2010). edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinf. (Oxford England) 26, 139–140. doi: 10.1093/bioinformatics/btp616
Rogg, L. E., Lasswell, J., and Bartel, B. (2001). A gain-of-function mutation in IAA28 suppresses lateral root development. Plant Cell 13, 465–480. doi: 10.1105/tpc.13.3.465
Schnable, P. S. and Springer, N. M. (2013). Progress toward understanding heterosis in crop plants. Annu. Rev. Plant Biol. 64, 71–88. doi: 10.1146/annurev-arplant-042110-103827
Seo, H. W., Suh, J. H., Kyung, J. S., Jang, K. H., and So, S. H. (2019). Subacute Oral Toxicity and Bacterial Mutagenicity Study of a Mixture of Korean Red Ginseng (Panax ginseng C.A. Meyer) and Salvia plebeia R. Br. Extracts. Toxicological Res. 35, 215–224. doi: 10.5487/TR.2019.35.3.215
Shalby, N., Mohamed, I. A. A., Xiong, J., Hu, K., Yang, Y., Nishawy, E., et al. (2021). Overdominance at the Gene Expression Level Plays a Critical Role in the Hybrid Root Growth of Brassica napus. Int. J. Mol. Sci. 22, 9246. doi: 10.3390/ijms22179246
Shehzad, A., Ashraf, M. F., Sultan, S., Ali, M., and Sadaqat, H. A. (2015). Heterosis Studies for Some Morphological, Seed Yield and Quality Traits in Rapeseed (Brassica napus L.). J. Biology Agric. Healthcare 5, 39–47. Available online at: https://www.iiste.org/Journals/index.php/JBAH/article/view/27568/28284 (Accessed January 27, 2026).
Sreekumari, M. T., Jos, J. S., and Nair, S. G. (1999). ‘Sree Harsha’: A superior triploid hybrid in cassava. Euphytica 106, 1–6. doi: 10.1023/A:1003473118487
Sui, C., Han, W. J., Zhu, C. R., and Wei, J. H. (2021). Recent progress in Saikosaponin biosynthesis in Bupleurum. Curr. Pharm. Biotechnol. 22, 329–340. doi: 10.2174/1389201021999200918101248
Sui, C., Zhang, J., Wei, J., Chen, S., Li, Y., Xu, J., et al. (2011). Transcriptome analysis of Bupleurum chinense focusing on genes involved in the biosynthesis of saikosaponins. BMC Genomics 12, 539. doi: 10.1186/1471-2164-12-539
Tang, Q. Y., Chen, G., Song, W. L., Fan, W., Wei, K. H., He, S. M., et al. (2019). Transcriptome analysis of Panax zingiberensis identifies genes encoding oleanolic acid glucuronosyltransferase involved in the biosynthesis of oleanane-type ginsenosides. Planta 249, 393–406. doi: 10.1007/s00425-018-2995-6
Teng, L., Guo, X., Ma, Y., Xu, L., Wei, J., and Xiao, P. (2023). A comprehensive review on traditional and modern research of the genus Bupleurum (Bupleurum L., Apiaceae) in recent 10 years. J. ethnopharmacology 306, 116129. doi: 10.1016/j.jep.2022.116129
Untergasser, A., Cutcutache, I., Koressaar, T., Ye, J., Faircloth, B. C., Remm, M., et al. (2012). Primer3–new capabilities and interfaces. Nucleic Acids Res. 40, e115. doi: 10.1093/nar/gks596
Wang, X., Cheng, Z.-M., Zhi, S., and Xu, F. (2016). Breeding triploid plants: A review. Czech J. Genet. Plant Breed. 52, 41–54. doi: 10.17221/151/2015-CJGPB
Wang, J., Guo, Y., Yin, X., Wang, X., Qi, X., and Xue, Z. (2022). Diverse triterpene skeletons are derived from the expansion and divergent evolution of 2,3-oxidosqualene cyclases in plants. Crit. Rev. Biochem. Mol. Biol. 57, 113–132. doi: 10.1080/10409238.2021.1979458
Wei, R., Tu, D., Huang, X., Luo, Z., Huang, X., Cui, N., et al. (2022). Genome-scale transcriptomic insights into the gene co-expression network of seed abortion in triploid Siraitia grosvenorii. BMC Plant Biol. 22, 173. doi: 10.1186/s12870-022-03562-4
Wei, J. H., Yang, C. M., Sui, C., Huang, L. Q., Shi, F. H., Chu, Q. L., et al. (2011). New Chinese Bellflower Cultivars ‘Zhonggeng 1’, ‘Zhonggeng 2’ and ‘Zhonggeng 3’ developed by using the male sterile line. Acta Hortic. Sin. 38, 1217–1218. Available online at: https://www.cnki.net/ (Accessed January 27, 2026).
Whitford, R., Fleury, D., Reif, J. C., Garcia, M., Okada, T., Korzun, V., et al. (2013). Hybrid breeding in wheat: technologies to improve hybrid wheat seed production. J. Exp. Bot. 64, 5411–5428. doi: 10.1093/jxb/ert333
Wu, H. Z., Mei, L., Zheng, S. X., Yang, S. C., and Li, X. B. (2011). Cultivation of triploid Erigeron breviscapus and observation of its biological characteristics. Chin. Traditional Herbal Drugs 42, 991–995. Available online at: https://www.cnki.net/ (Accessed January 27, 2026).
Wu, Y., Sun, Y., Wang, X., Lin, X., Sun, S., Shen, K., et al. (2016). Transcriptome shock in an interspecific F1 triploid hybrid of Oryza revealed by RNA sequencing. J. Integr. Plant Biol. 58, 150–164. doi: 10.1111/jipb.12357
Xu, D. M. (2019). Heterosis Analysis of Major Traits in F1 Hybrids of Bupleurum Chuanbeichai No.1 × B. marginatum (Southwest University of Science and Technology). Available online at: https://www.cnki.net/ (Accessed January 27, 2026).
Xu, D., Yu, M., Shen, J., Zhang, Z., Li, Y., Tang, Z., et al. (2018). Properly breed region to Chuanbeichai no.1 of new Bepleurum chinense in Sichuan area. China J. Chin. Materia Med. 41, 834–838. doi: 10.13863/j.issn1001-4454.2018.04.012
Xue, Z., Duan, L., Liu, D., Guo, J., Ge, S., Dicks, J., et al. (2012). Divergent evolution of oxidosqualene cyclases in plants. New Phytol. 193, 1022–1038. doi: 10.1111/j.1469-8137.2011.03997.x
Yang, L., Qiao, L., Su, X., Ji, B., and Dong, C. (2022). Drought stress stimulates the terpenoid backbone and triterpenoid biosynthesis pathway to promote the synthesis of saikosaponin in Bupleurum chinense DC. Roots. Molecules (Basel Switzerland) 27, 5470. doi: 10.3390/molecules27175470
Yu, M., Chen, H., Liu, Q., Huang, J., Semagn, K., Liu, D., et al. (2021). Analysis of unigenes involved in lateral root development in Bupleurum chinense and B. scorzonerifolium. Planta 253, 128. doi: 10.1007/s00425-021-03644-x
Yu, M., Chen, H., Liu, S. H., Li, Y. C., Sui, C., Hou, D. B., et al. (2020). Differential Expression of Genes Involved in Saikosaponin Biosynthesis Between Bupleurum chinense DC. and Bupleurum scorzonerifolium Willd. Front. Genet. 11. doi: 10.3389/fgene.2020.583245
Yu, M., Liu, D., Li, Y. C., Sui, C., Chen, G. D., Tang, Z. K., et al. (2019). Validation of reference genes for expression analysis in three Bupleurum species. Biotechnol. Biotechnol. Equip. 33, 154–161. doi: 10.1080/13102818.2018.1557556
Yu, G., Wang, L. G., Han, Y., and He, Q. Y. (2012). clusterProfiler: an R package for comparing biological themes among gene clusters. Omics 16, 284–287. doi: 10.1089/omi.2011.0118
Yuan, L. P. (1994). Increasing yield potential in rice by exploitation of heterosis (Agricultural and Food Sciences). Available online at: https://api.semanticscholar.org/CorpusID:221963033.
Zamariola, L., Tiang, C. L., De Storme, N., Pawlowski, W., and Geelen, D. (2014). Chromosome segregation in plant meiosis. Front. Plant Sci. 5. doi: 10.3389/fpls.2014.00279
Zeng, C., Zhao, J., Chen, H., Xin, C., Wang, B., Yu, M., et al. (2023). Traditional use, germplasm identification, phytochemistry, pharmacology of Bupleuri Radix: a review. Medicinal Plant Biol. 2, 18. doi: 10.48130/MPB-2023-0018
Zhang, Q., Li, M., Chen, X., Liu, G., Zhang, Z., Tan, Q., et al. (2022). Chromosome-level genome assembly of Bupleurum chinense DC provides insights into the saikosaponin biosynthesis. Front. Genet. 13. doi: 10.3389/fgene.2022.878431
Zhang, X., Yu, Y., Jiang, S., Yu, H., Xiang, Y., Liu, D., et al. (2019). Oleanane-Type Saponins Biosynthesis in Panax notoginseng via Transformation of β-Amyrin Synthase Gene from Panax japonicus. J. Agric. Food Chem. 67, 1982–1989. doi: 10.1021/acs.jafc.8b07183
Zhao, S., Zhao, X., Xu, X., Han, Z., and Qiu, C. (2022). Transcription factor IAA27 positively regulates P uptake through promoted adventitious root development in apple plants. Int. J. Mol. Sci. 23, 14029. doi: 10.3390/ijms232214029
Keywords: Bupleurum chinense, heterosis, karyotype, polyploid breeding, saikosaponins
Citation: Mo C, Chen W, Lv Z, Zhao J, Li Y, Shi Q, Dong K, Wei Z, Yu Z, Wang X, Xin C, Ni Z, Yu M and Chen H (2026) Heterosis unveiled in root-related traits and saikosaponins content between triploid F1 hybrids and parental Bupleurum chinense DC.. Front. Plant Sci. 17:1736464. doi: 10.3389/fpls.2026.1736464
Received: 31 October 2025; Accepted: 14 January 2026; Revised: 15 December 2025;
Published: 12 February 2026.
Edited by:
Kioumars Ghamkhar, AgResearch Ltd, New ZealandReviewed by:
Wei Sun, China Academy of Chinese Medical Sciences, ChinaNarendra Singh, Indian Agricultural Research Institute (Jharkhand), India
Copyright © 2026 Mo, Chen, Lv, Zhao, Li, Shi, Dong, Wei, Yu, Wang, Xin, Ni, Yu and Chen. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Ma Yu, bWF5dTAwNzNAc3d1c3QuZWR1LmNu; Hua Chen, aGNoZW5Ac3d1c3QuZWR1LmNu
†These authors have contributed equally to this work