Bioinformatic analysis and functional characterization of ZjABCG15, a plasma membrane-localized ABCG transporter associated with reproductive development in Ziziphus jujuba Mill. ‘Zhongqiusucui’ Authors: Yutong Li, Juan Chen, Fengxia Shao, Xinxin Lian, Kuo Yang, Sen Wang, Yafeng Wen Published: 2026 ABSTRACT BackgroundEmbryo abortion is a major factor limiting hybrid breeding efficiency in jujube (Ziziphus jujuba Mill.). To investigate whether the ABCG transporter gene ZjABCG15 is associated with reproductive development and ovule/seed abortion-related phenotypes, we performed bioinformatic analysis, expression profiling, and heterologous functional validation using the jujube cultivar ‘Zhongqiusucui’.ResultsThe cDNA sequence of ZjABCG15 is 2,019 bp and encodes a 672-amino-acid protein. Bioinformatic analysis suggested that ZjABCG15 is a plasma membrane-associated ABCG transporter. Subcellular localization further confirmed its plasma membrane localization. Quantitative real-time PCR (qRT-PCR) showed that ZjABCG15 was expressed in multiple jujube tissues, with relatively high transcript levels in leaves and kernels. For functional validation, overexpression and RNA interference (RNAi) constructs of ZjABCG15 were introduced into Arabidopsis thaliana. Compared with the wild type, the average seed yield per plant increased by 44.2% in the overexpression line, but decreased by 12.65% in the RNAi line. In addition, RNAi plants showed shorter siliques, fewer seeds, and increased frequencies of seed abortion-related phenotypes, whereas overexpression plants produced more siliques and plump seeds.ConclusionsThese results suggest that ZjABCG15 is a plasma membrane-localized ABCG transporter gene associated with ovule-to-seed development in jujube. Heterologous functional analysis in Arabidopsis supports its conserved positive role in reproductive development, whereas its direct role and regulatory mechanism in jujube still require further validation.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12870-026-08890-3. Background Ziziphus jujuba Mill., belonging to the genus Ziziphus of the family Rhamnaceae, is an important woody grain tree species in China. Jujube has a long cultivation history and abundant germplasm resources in China. However, the breeding of new cultivars still mainly depends on the selection of superior individual plants, and the utilization of germplasm innovation remains to be improved. Hybrid breeding, as an important approach for germplasm creation, has not yet achieved substantial progress in jujube. The key limiting factor is the widespread occurrence of embryo abortion, which prevents the formation of viable and fully developed seeds, thereby making it difficult to obtain hybrid progeny and seriously restricting the progress of hybrid breeding and the efficient utilization of germplasm resources [1, 2]. At present, studies on embryo abortion in jujube have mainly focused on cytological, embryological, physiological, and biochemical aspects, whereas systematic investigations of its underlying mechanisms at the molecular level remain relatively limited. Therefore, the identification of key regulatory genes associated with embryo abortion is of great significance for elucidating its molecular mechanism and promoting jujube breeding. The ABCG transporter family provides a new perspective for understanding abnormal reproductive development in plants. The ABCG subfamily is the largest group within the ABC transporter superfamily and is widely distributed in higher plants, playing important roles in transmembrane transport. These proteins are capable of transporting a wide range of substrates, including lipids, cutin and wax precursors, phytohormones, and secondary metabolites, and are involved in various biological processes such as plant growth, development, and stress responses [3–5]. Because many of their transported substrates are associated with cell surface structure formation, tissue differentiation, and the maintenance of the reproductive microenvironment, ABCG proteins are not only closely related to pollen wall and anther development, but may also participate in ovule development, integument differentiation, seed coat formation, and early seed development. Therefore, they are of considerable importance in the study of plant reproductive development. In model plants, the roles of ABCG transporters in reproductive development have been well documented. For example, in Arabidopsis thaliana, AtABCG26 is involved in the transport of components required for pollen exine formation, and loss of its function leads to abnormal pollen development and a marked reduction in fertility [6–9]. Likewise, ZmMS2 in maize, as a homologous gene, also affects normal microspore exine formation when mutated [10]. In addition, several ABCG members in Arabidopsis participate in the transport of pollen wall precursors to the pollen surface and play important roles in pollen exine formation and tryphine deposition [11]. These studies indicate that ABCG transporters play important and relatively conserved roles in male reproductive development in plants [12–14]. It should be emphasized, however, that what these proteins mediate is not simply “male fertility” itself, but rather the transmembrane transport of lipids and other related hydrophobic substances. Such substances also serve as essential structural and signaling components for ovule development, integument differentiation, seed coat formation, and early seed establishment. Therefore, although current studies have focused mainly on anther and pollen development, the intrinsic molecular functions of ABCG transporters suggest that they may also participate in female reproductive development and reproductive defects associated with ovule/seed abortion. In recent years, with the deepening of studies in model plants, the roles of ABCG transporters in reproductive development have attracted increasing attention, providing new insights into the molecular mechanisms underlying reproductive defects and ovule/seed abortion in jujube. Previous studies have shown that ZjABCG15 plays an important role in pollen wall development in jujube, but its function in ovule and seed development remains unclear. Our previous work showed that ZjABCG15 was highly expressed during the early stages of ovule development in jujube (SW, OE, and FrI), but was significantly down-regulated during the critical transition stage before ovule/seed abortion (FrI vs. FrII) and remained at low levels during the abortion stages (FrII and FrIII). Here, SW denotes the sepal flattening stage, whereas OE, FrI, FrII, and FrIII represent sequential ovule/fruit developmental stages defined in our previous transcriptomic dataset [13]. This expression pattern suggests that the function of ZjABCG15 may not be limited to male reproductive development, but may also be involved in normal ovule development and early seed formation. Its persistently low expression during the critical stage associated with ovule/seed abortion further indicates that abnormal expression of this gene may be closely associated with impaired ovule or seed development. In summary, current research on the molecular mechanisms of embryo abortion in jujube is still limited, and the relevant regulatory genes, key biological processes, and metabolic networks remain largely unclear. Given the transmembrane transport function of the protein family to which ZjABCG15 belongs, the existing evidence for its role in reproductive development, and its specific expression pattern during ovule development in jujube, we hypothesized that ZjABCG15 may participate in the continuous process of ovule-to-seed development and may be closely related to reproductive defects leading to ovule/seed abortion, rather than to embryo-specific defects. Considering the conserved roles of ABCG transporters in plant reproductive development, Arabidopsis was used here as a preliminary heterologous system to assess whether ZjABCG15 has a conserved reproductive-development-related function. To test this hypothesis, the jujube cultivar ‘ZhongqiuSucui’ was used as the experimental material in this study. Materials and methods Plant materials Jujube The ovule/seed-abortion-prone cultivar Ziziphus jujuba Mill. ‘Zhongqiusucui’, a main fresh-eating jujube variety cultivated in Hunan Province, was used as the experimental material. Eight-year-old trees with robust growth, relatively uniform vigor, and no pest or disease damage were selected. The experimental site was located at the Jujube Experimental Base of Central South University of Forestry and Technology, Qidong County, Hunan Province, China (112.12° E, 26.78° N). This region has a subtropical climate, characterized by cold winters, hot and humid summers, abundant but unstable thermal conditions, and ample precipitation with uneven seasonal distribution. During the 2024 growing season, samples of flowers, fruits, leaves, roots, and stems were collected from jujube trees at the base. All samples were immediately frozen in liquid nitrogen and subsequently stored at -80 °C until further use. Tobacco Nicotiana benthamiana, maintained in our laboratory, was used in this study. Four- to six-week-old tobacco plants were grown in an artificial climate chamber using a substrate composed of peat, vermiculite, and perlite mixed at a volume ratio of 3:1:1. The growth conditions were set at a 16 h light/8 h dark photoperiod, a constant temperature of 22 °C, a light intensity of 120–150 µmol·m⁻²·s⁻¹, and a relative humidity of 60%–70%. During cultivation, plants were watered according to substrate moisture status to keep the substrate moist without excessive water accumulation, generally once every 2–3 d. In addition, potassium dihydrogen phosphate solution was applied once per week according to plant growth status. The growth environment was kept relatively stable throughout the cultivation period to meet the requirements of subsequent transient expression, subcellular localization, and genetic transformation experiments. Arabidopsis Arabidopsis thaliana ecotype Col-0 was used as the recipient material for Agrobacterium-mediated transformation, and the corresponding overexpression and RNAi transgenic lines were generated in this study. Before sowing, seeds were subjected to low-temperature treatment at 4 °C for 2–3 days to break dormancy and improve germination uniformity. The seeds were then surface-sterilized and sown on MS medium for germination. After the seedlings had grown to an appropriate stage, they were transplanted into the growth substrate and further cultivated in a growth chamber. The substrate consisted of peat, vermiculite, and perlite mixed at a volume ratio of 3:1:1. Growth conditions were maintained at 22–24 °C under a 16 h light/8 h dark photoperiod, with a light intensity of 100–120 µmol·m⁻²·s⁻¹ and a relative humidity of 60%–70%. During cultivation, plants were watered as needed according to their growth status to maintain substrate moisture, generally once every 2–3 d. After successful transplantation and establishment, potassium dihydrogen phosphate solution was applied once per week to ensure normal plant growth. All materials were grown in the same substrate and managed under identical conditions, with a relatively stable cultivation environment maintained throughout the experiment to support subsequent phenotypic observation, gene function verification, and related physiological and biochemical analyses. Bioinformatic analysis of the ZjABCG15 gene SnapGene was used to visualize sequencing results. The nucleotide sequence of ZjABCG15 was first confirmed using BLASTN/BLASTP on the NCBI website (https://blast.ncbi.nlm.nih.gov/Blast.cgi). Unless otherwise specified, default parameters were used for all bioinformatic analyses. For BLAST searches, only sequences with an E-value of 0 were considered significant. The open reading frame (ORF) was subsequently identified using the NCBI ORF Finder tool (https://www.ncbi.nlm.nih.gov/orffinder/), and the corresponding amino acid sequence was deduced. Physicochemical properties, including molecular weight and theoretical isoelectric point, were predicted using ProtParam (http://web.expasy.org/protparam/). Signal peptide and transmembrane domain predictions were performed using SignalP 6.0 (https://services.healthtech.dtu.dk/service.php?SignalP-5.0) and TMHMM 2.0 (https://services.healthtech.dtu.dk/service.php?TMHMM-2.0), respectively. Amino acid sequence alignment and conserved domain analysis were conducted using DNAMAN (version 7) with default alignment parameters. The secondary structure of the protein was predicted using SOPMA (https://npsa-prabi.ibcp.fr/cgi-bin/npsa_automat.pl?page=npsa_sopma.html) with the following parameters: window width = 17, similarity threshold = 8, and number of states = 4. The three-dimensional structure was analyzed using Phyre2 (http://www.sbg.bio.ic.ac.uk/~phyre2) in mode with default settings. To investigate the phylogenetic relationship of ZjABCG15, its deduced amino acid sequence was submitted to the NCBI BLASTP database to retrieve homologous protein sequences from different species. After representative homologous sequences were selected, multiple sequence alignment was performed using DNAMAN 7.0 with default settings, and the alignment results were imported into MEGA 12.0, where the phylogenetic tree was constructed using the Construct/Test Neighbor-Joining Tree function under the Phylogeny menu with default settings. Bootstrap analysis with 1,000 replicates was performed to evaluate the support for each branch. Bootstrap support values were displayed at the corresponding nodes of the phylogenetic tree. To predict the potential interacting proteins of ZjABCG15, a protein-protein interaction network analysis was conducted using the STRING database. Based on the target protein or its homologous protein information, protein interaction relationships were retrieved under the corresponding species background, and candidate interacting proteins were screened according to interaction scores and known functional annotations for subsequent analysis. Gene expression pattern analysis Gene-specific primers for quantitative real-time PCR (qRT-PCR) were designed using Primer5 software according to the established principles for qPCR primer design. All primers were synthesized by Sangon Biotech (Shanghai) Co., Ltd (Supplementary Table S1). Total RNA was extracted from jujube leaves, roots, young shoots, floral organs at different developmental stages, and various fruit tissues (previously flash-frozen in liquid nitrogen and stored at − 80 °C) using the Plant RNA Extraction Kit (TIANGEN Biotech, Beijing, China). Reverse transcription was performed using the PrimeScript™ RT Reagent Kit (Clontech Laboratories, Inc., Mountain View, CA, USA) according to the manufacturer’s instructions. The jujube ZjEF1 gene was used as an internal reference. Primers were designed within the 3′ untranslated region (3′-UTR) of ZjABCG15 to ensure amplification specificity. Semi-quantitative RT-PCR was conducted to assess the specificity of gene expression, and qRT-PCR was employed to quantify the expression levels of ZjABCG15 in different tissues. For qRT-PCR analysis, each sample was examined using three biological replicates, and each biological replicate was analyzed with three technical replicates. Relative gene expression levels were calculated using the 2^-ΔΔCt method. These analyses provided foundational guidance for subsequent functional characterization of the gene. Subcellular localization The ZjABCG15 coding sequence without the stop codon was recombined into the pBI121-GFP vector. The recombinant plasmid was introduced into Agrobacterium tumefaciens strain GV3101 using the freeze-thaw method. Four- to six-week-old N. benthamiana plants were used for the agroinfiltration experiments. N. benthamiana plants were placed in darkness for 3 h prior to infiltration. The abaxial side of fully expanded leaves was infiltrated with the Agrobacterium suspension using a needleless syringe. After infiltration, the plants were kept in darkness for 16 h and then transferred to normal light conditions for 48 h. Subcellular localization of the ZjABCG15-GFP fusion protein was observed using a laser scanning confocal microscope. GFP fluorescence was detected at 488 nm laser excitation and 500–530 nm emission filter settings. Construction of overexpression and RNAi vectors and genetic transformation of arabidopsis Appropriate restriction enzyme sites were selected, and the cloned sense fragment of ZjABCG15 was inserted into the plant expression vector pBI121 to construct the overexpression vector. For construction of the RNAi suppression vector, a 300–400 bp specific fragment from the 3′ end of ZjABCG15 was selected and inserted into the pBI121 vector in an inverted-repeat configuration separated by a spacer sequence. The 3′-end fragment was chosen because this region is generally less conserved and more specific than homologous regions in related genes, thereby reducing the risk of off-target silencing during RNAi. The recombinant vectors were finally verified by PCR amplification and sequencing. Genetic transformation and screening of transgenic plants: T0 seeds obtained after Agrobacterium-mediated transformation were surface-sterilized and sown on MS medium containing 50 mg/L hygromycin for selection. After stratification at 4 °C for 2 d and growth at 22 °C under light for 1 week, normally growing resistant seedlings were regarded as putative transgenic plants. These positive seedlings were then transplanted into soil and grown under greenhouse conditions. When the leaves had fully expanded, genomic DNA was extracted from the putative transgenic plants, and PCR amplification was performed using hygromycin phosphotransferase (HPT) primers to identify positive T1 transgenic lines. The confirmed transgenic plants were subsequently used for phenotypic observation and comparative analysis with wild-type plants, with particular attention to growth status and reproductive traits. Molecular identification and phenotypic analysis of transgenic plants. Putative transgenic plants were identified by PCR amplification using hygromycin phosphotransferase (HPT) primers after hygromycin selection. Positive transgenic lines were transplanted into soil and further grown under greenhouse conditions. After molecular confirmation, genetically stable homozygous T3 generation lines were used for subsequent phenotypic observation and quantitative analysis. Phenotypic traits, including plant growth status, silique development, seed set, and seed morphology, were compared between transgenic and wild-type plants. For seed-yield analysis, ten individuals from each line with comparable growth status were selected as biological replicates. Upon full maturity and natural drying of siliques, whole plants were harvested individually, total seeds per plant were counted, and the statistical significance of differences among groups was assessed by one-way analysis of variance (ANOVA) (Supplementary Table S2). Membrane-based yeast two-hybrid assay To verify the predicted association between ZjABCG15 and LRR1, a membrane-based yeast two-hybrid assay was performed using the split-ubiquitin system. The bait construct pBT3-N-ABCG15 and the prey construct pPR3-N-LRR1 were co-transformed into the yeast strain NMY51. The combinations pTSU2-APP + NubG-Fe65 and pTSU2-APP + pPR3-N were used as the positive and negative controls, respectively. pBT3-N-ABCG15 + pOst1-NubI was used for bait functionality verification, and pBT3-N-ABCG15 + pPR3-N was used for self-activation testing. Transformed yeast cells were first selected on DDO medium (-Trp/-Leu). Colonies were then resuspended in sterile water and spotted onto DDO and QDO/X (-Trp/-Leu/-His/-Ade + X-α-gal) media, followed by incubation at 30 °C for 3–5 d. Growth and blue coloration on the selective media were used as indicators for evaluating protein-protein interaction. Results and analysis Bioinformatic prediction and analysis of the protein encoded by ZjABCG15 The nucleotide sequence of ZjABCG15 was analyzed using the NCBI BLAST online tool. The gene (GenBank accession No. LOC107431645) is located on chromosome 11 and contains nine exons (Supplementary Fig. S1A). Its coding sequence (CDS) is 2,019 bp in length and encodes a 672-amino-acid polypeptide (Fig. 1, Supplementary Fig. S1D). Putative transmembrane helices and signal-peptide features are shown in Supplementary Fig. S1B, C. Physicochemical analysis indicated that ZjABCG15 has a predicted molecular mass of approximately 75.6 kD, a theoretical isoelectric point (pI) of 8.83, and slight overall hydrophobicity, supporting its potential membrane-associated nature. Phylogenetic analysis showed that ZjABCG15 clustered closely with ABCG15 orthologs from other Ziziphus accessions and related plant species, indicating a conserved evolutionary relationship. TMHMM predicted six distinct transmembrane helices, and no canonical N-terminal signal peptide was detected, suggesting that ZjABCG15 is a typical integral membrane protein rather than a secretory protein. Detailed phylogenetic relationships, secondary and tertiary structure predictions, and the STRING-based protein interaction network are provided in Supplementary Fig. S2. Fig. 1Amino acid sequence alignment of ZjABCG15 homologous proteins Amino acid sequence alignment of ZjABCG15 homologous proteins Subcellular localization of the ZjABCG15 protein To determine the functional localization of ZjABCG15 within plant cells, subcellular localization analysis was performed. The ZjABCG15-GFP fusion vector was transiently transformed into epidermal cells of N. benthamiana leaves, and the green fluorescent protein (GFP) signal of ZjABCG15 was observed using laser scanning confocal microscopy. The results showed that the ZjABCG15-GFP fusion protein exhibited clear membrane localization in tobacco epidermal cells. The GFP signal formed a continuous ring-shaped distribution along the cell periphery, which coincided with the cell boundaries visualized by differential interference contrast (DIC) microscopy. When co-expressed with the plasma membrane marker mCherry-PIP2A, the merged channel displayed continuous ring-shaped orange signals, confirming that ZjABCG15 is a plasma membrane-localized protein (Fig. 2). This finding is consistent with the functional characteristics of ABCG family transporters in lipid secretion. Fig. 2Subcellular localization of ZjABCG15 in N. benthamiana epidermal cells. (The original uncropped images are shown in Supplementary Fig. S3.) Subcellular localization of ZjABCG15 in N. benthamiana epidermal cells. (The original uncropped images are shown in Supplementary Fig. S3.) Representative images showing ZjABCG15 membrane localization are presented in Fig. 2, including DIC, GFP fluorescence, mCherry fluorescence, and merged channels. At least three independent experiments were performed for each combination, all of which yielded similar results. RT-qPCR analysis of ZjABCG15 expression The expression patterns of ZjABCG15 in different jujube tissues were further analyzed by reverse transcription-quantitative real-time PCR (RT-qPCR) (Fig. 3E). Tissues including flowers, pericarp, pulp, leaves, roots, stems, seed coat, and kernel were examined. The results revealed that ZjABCG15 exhibited distinct tissue-specific expression, with significantly higher transcript levels in leaves and kernels than in other tissues (P < 0.01), followed by flowers at the sepal flattening stage. In contrast, extremely low expression levels were detected in roots, stems, pericarp, pulp, and seed coat. Combined with its relatively high expression in kernels, ZjABCG15 is presumed to be associated with the transport and accumulation of related substances during seed development; however, whether it is directly involved in the allocation of photoassimilates remains to be further verified. Fig. 3Molecular cloning, transgenic plant identification, and tissue-specific expression analysis of the ZjABCG15 gene. A Electrophoresis detection of Myc-ZjABCG15 cloning. (The original uncropped images are shown in Supplementary Fig. S4.) (B) Electrophoresis detection of RNAi-ZjABCG15 cloning. (The original uncropped images are shown in Supplementary Fig. S5) (C) Identification of ZjABCG15-Myc overexpressing lines by HPT PCR. (The original uncropped images are shown in Supplementary Fig. S6.) (D) Identification of RNAi-ZjABCG15 lines by HPT PCR. (The original uncropped images are shown in Supplementary Fig. S7.) (E) Relative expression levels of ZjABCG15 in different jujube tissues. Data are presented as mean ± SD of three biological replicates, with three technical replicates for each biological replicate. Relative expression levels were calculated using the 2^-ΔΔCt method. Different letters indicate significant differences among tissues (P < 0.01) Molecular cloning, transgenic plant identification, and tissue-specific expression analysis of the ZjABCG15 gene. A Electrophoresis detection of Myc-ZjABCG15 cloning. (The original uncropped images are shown in Supplementary Fig. S4.) (B) Electrophoresis detection of RNAi-ZjABCG15 cloning. (The original uncropped images are shown in Supplementary Fig. S5) (C) Identification of ZjABCG15-Myc overexpressing lines by HPT PCR. (The original uncropped images are shown in Supplementary Fig. S6.) (D) Identification of RNAi-ZjABCG15 lines by HPT PCR. (The original uncropped images are shown in Supplementary Fig. S7.) (E) Relative expression levels of ZjABCG15 in different jujube tissues. Data are presented as mean ± SD of three biological replicates, with three technical replicates for each biological replicate. Relative expression levels were calculated using the 2^-ΔΔCt method. Different letters indicate significant differences among tissues (P < 0.01) Functional validation of ZjABCG15 via genetic transformation To investigate the function of ZjABCG15 in plant reproductive development, ZjABCG15-overexpressing (Myc) and RNA interference (RNAi) knockdown lines were generated via genetic transformation, and their growth and developmental phenotypes were systematically compared with those of wild-type A. thaliana (WT). The results (Fig. 4) showed that, compared with WT, the RNAi lines exhibited pronounced abnormalities at multiple developmental stages: overall plant vigor was weakened, plant height was significantly reduced, and branch number was markedly decreased; leaf development was impaired, with a significant reduction in leaf number; reproductive capacity was compromised, as reflected by fewer siliques and significantly shortened silique length; and seed development was severely defective, with some seeds displaying yellowing and shrinkage typical of abortion, and both transverse and longitudinal diameters of non-aborted seeds were significantly smaller than those of WT (p < 0.05). In contrast, the overexpression lines displayed enhanced overall vigor, with more robust plants and a significant increase in branch number; leaf development was normal; both silique and seed numbers were the highest among all lines, with seeds densely and uniformly distributed within siliques, appearing plump and light green, and no obvious seed abortion was observed; the transverse and longitudinal diameters of seeds were comparable to those of WT, with no significant differences. Fig. 4Phenotypic observation and statistical analysis of flowers and seeds in wild-type, RNAi, and overexpressing A. thaliana. A Phenotypic observation of wild‑type and transgenic A. thaliana. (The boxed area and arrow indicate aborted seeds.) (B) Quantitative analysis of wild‑type and transgenic A. thaliana. C Percentage of aborted and normal seeds in wild‑type and transgenic A. thaliana. (Scale bars: 1500 μm for whole-plant phenotypes, 1000 μm for leaf and flower phenotypes, 500 μm, 1500 μm, and 2000 μm for siliques, and 2000 μm for seeds.) Phenotypic observation and statistical analysis of flowers and seeds in wild-type, RNAi, and overexpressing A. thaliana. A Phenotypic observation of wild‑type and transgenic A. thaliana. (The boxed area and arrow indicate aborted seeds.) (B) Quantitative analysis of wild‑type and transgenic A. thaliana. C Percentage of aborted and normal seeds in wild‑type and transgenic A. thaliana. (Scale bars: 1500 μm for whole-plant phenotypes, 1000 μm for leaf and flower phenotypes, 500 μm, 1500 μm, and 2000 μm for siliques, and 2000 μm for seeds.) These findings suggest that the introduced overexpression and RNAi constructs of ZjABCG15 are associated with changes in vegetative growth and reproductive development in A. thaliana. Together, these results support a positive role of ZjABCG15 in plant reproductive development in the heterologous Arabidopsis system. Gene silencing leads to inhibited plant growth, retarded leaf development, and impaired seed formation, whereas overexpression promotes enhanced growth and reproductive fitness. Together, these results indicate that ZjABCG15 plays a positive role in plant reproductive development in a heterologous Arabidopsis system. Quantitative analysis of seed yield in transgenic Arabidopsis lines To systematically elucidate the regulatory role of ZjABCG15 in plant reproductive development, seed yield was statistically analyzed in genetically stable homozygous T3 generation lines of wild-type (WT), overexpressing (Myc), and RNA interference (RNAi) A. thaliana plants, following the comprehensive phenotypic observations described above. All plants were grown under strictly uniform conditions (22 °C, 16 h light/8 h dark photoperiod). Ten individuals of each line (WT, RNAi and Myc) with consistent growth status—comparable plant height and rosette leaf number at the same post-germination stage—were selected as biological replicates. Upon full maturity and natural drying of siliques, whole plants were harvested individually. Total seeds per plant were accurately counted using manual counting combined with microscopic imaging validation. The mean seed yield per plant was calculated for each group, and statistical significance of differences among groups was assessed by one-way analysis of variance (ANOVA). Statistical analysis showed that the average seed yield per plant in the wild type (WT) was 2161.00 ± 103.86 seeds. Compared with WT, the average seed yield per plant in the ZjABCG15 overexpressing line (Myc-ZjABCG15) increased significantly to 3115.60 ± 112.86 seeds, representing a 44.2% increase (p < 0.05), indicating that high expression of ZjABCG15 can significantly promote seed formation and accumulation. In contrast, the average seed yield per plant in the RNAi suppression line decreased to 1887.60 ± 67.53 seeds, which was 12.65% lower than that of WT (p < 0.05), suggesting that suppression of this gene markedly weakens the seed-setting capacity of the plant (Supplementary Table S3). Further analysis of yield-related traits revealed that the average number of siliques per plant in WT was 18.20 ± 1.81, whereas it increased to 22.80 ± 1.93 in the overexpression line, representing a 25.27% increase; in the RNAi line, however, it decreased to 12.50 ± 1.96, which was 31.32% lower than that of WT. Regarding the number of seeds per silique, the WT averaged 46.00 ± 1.63 seeds, while the overexpression line had 46.60 ± 1.65 seeds, showing a slight increase of 1.30% relative to WT; by contrast, the RNAi line had only 34.60 ± 3.66 seeds per silique, a 24.78% decrease compared with WT. Meanwhile, the average silique length and width in WT were 29.79 ± 0.26 mm and 1.68 ± 0.03 mm, respectively. In the overexpression line, the silique length was 28.89 ± 0.21 mm and the width was 1.85 ± 0.03 mm, indicating slightly shorter but wider siliques. In the RNAi line, silique length decreased to 24.95 ± 0.18 mm, representing a 16.25% reduction relative to WT, while silique width was 1.80 ± 0.03 mm. Seed morphology measurements further showed that the average seed length and width in WT were 1.34 ± 0.07 mm and 0.70 ± 0.06 mm, respectively. In the overexpression line, seed length and width were 1.40 ± 0.11 mm and 0.75 ± 0.09 mm, representing increases of 4.10% and 7.61%, respectively, compared with WT. In contrast, in the RNAi line, seed length and width were only 1.12 ± 0.11 mm and 0.56 ± 0.09 mm, respectively, representing decreases of 16.78% and 19.68% relative to WT. In addition, obviously shriveled or aborted seeds were observed in the RNAi line, with an average length and width of only 0.64 ± 0.10 mm and 0.23 ± 0.08 mm, respectively, which were markedly smaller than those of normal seeds.These yield data are highly consistent with the previously observed phenotypic characteristics: overexpression lines produced more siliques and plump seeds, whereas RNAi lines exhibited fewer siliques, higher seed abortion rates, and smaller seed size. Thus, the quantitative yield analysis further corroborates the reliability of the preceding phenotypic observations. In summary, the quantitative data showed that changes in ZjABCG15 expression were positively associated with seed yield in the heterologous Arabidopsis system. ZjABCG15 affected both seed-related traits and the total number of seeds per plant, further supporting its involvement in reproductive development. These findings provide a quantitative basis for future investigation of ZjABCG15 function in jujube seed development. ZjABCG15 interacts with LRR1 in a membrane-based yeast two-hybrid assay To further investigate the potential regulatory mechanism of ZjABCG15, candidate interacting proteins were analyzed, and LRR1 was identified as a potential interacting partner. A membrane-based yeast two-hybrid assay was then performed to test whether ZjABCG15 directly interacts with LRR1. All co-transformation combinations grew on DDO medium, indicating successful transformation of the yeast cells. The positive control grew and turned blue on QDO/X medium, whereas the negative control did not grow, confirming that the assay system functioned properly. The bait functionality verification group also grew on QDO/X medium, indicating that the bait fusion protein was correctly expressed and that the split-ubiquitin system was functional. Importantly, the interaction group pBT3-N-ABCG15 + pPR3-N-LRR1 also grew and turned blue on QDO/X medium, demonstrating that ZjABCG15 interacts with LRR1 in the yeast system (Fig. 5). Fig. 5Membrane-based yeast two-hybrid validation of the interaction between ZjABCG15 and LRR1 Membrane-based yeast two-hybrid validation of the interaction between ZjABCG15 and LRR1 Discussion Embryo abortion is one of the major constraints limiting hybrid breeding and germplasm innovation in jujube, and its molecular regulatory mechanisms remain largely unclear. Previous studies have shown that members of the ABCG transporter family are widely involved in the transmembrane transport of substances related to plant reproductive development, particularly in processes such as pollen exine formation, cuticle deposition, and surface structure establishment [5, 15, 16]. However, compared with male reproductive development, the functions of ABCG family members in ovule, seed, and embryo development remain relatively underexplored. In the present study, expression analysis and heterologous functional validation were performed for ZjABCG15 from jujube. The results suggest that ZjABCG15 may be involved in ovule-to-seed development and normal seed formation, thereby providing a promising candidate gene for understanding reproductive defects associated with ovule/seed abortion in jujube. Tissue-specific expression analysis showed that ZjABCG15 was highly expressed in leaves and kernels. Combined with our previous transcriptomic data, this gene was found to be strongly expressed during the early stages of ovule development in jujube(SW, OE, and FrI; where SW represents the sepal flattening stage, and OE and FrI represent early ovule developmental stages), but significantly down-regulated during the critical period before and during ovule/seed abortion [17]. This expression pattern suggests that ZjABCG15 is likely closely associated with normal ovule development, seed formation, and the accumulation of storage materials. The kernel is an important site for embryo development and reserve accumulation, and high expression of related genes in this tissue generally implies a role in the transport and allocation of nutrients or structural precursors during development. Therefore, the relatively high expression of ZjABCG15 in kernels, together with its altered expression during the critical stage associated with ovule/seed abortion, supports a functional link between this gene and normal ovule-to-seed development. Heterologous transformation experiments in Arabidopsis thaliana provided further evidence for the biological role of ZjABCG15. RNAi lines exhibited reduced plant height, fewer branches, shorter siliques, decreased seed number, and abnormal seeds showing yellowing, shrinkage, and collapse, whereas overexpressing lines displayed stronger growth vigor and improved seed-setting performance. These results indicate that the expression level of ZjABCG15 is positively correlated with both vegetative growth and reproductive output, especially with silique development and normal seed formation. Notably, the most pronounced abnormalities in the RNAi lines were concentrated at the ovule/seed developmental stage, suggesting that the function of ZjABCG15 may not be limited to early organ formation, but may instead be more directly involved in the supply, transport, or homeostasis of key substances during reproductive development leading to ovule/seed abortion-related phenotypes. Based on the known functions of the ABCG family, ZjABCG15 is likely involved in the transmembrane transport of lipids or other hydrophobic molecules, thereby affecting surface structure formation, storage accumulation, or developmental signal maintenance during ovule-to-seed development. Plant ovule and seed development depend not only on the accumulation of carbohydrates and proteins, but also on lipids, which play essential roles in membrane system construction, energy storage, and signal transduction. If such transport processes are disrupted, abnormal cell differentiation, impaired structural formation, or metabolic imbalance may occur during reproductive development, ultimately leading to poor seed development or even abortion. Therefore, the abnormal siliques and seeds observed in the RNAi lines in this study may result from inhibition of key transport processes mediated by ZjABCG15. This hypothesis is also consistent with reports that ABCG family members are involved in reproductive structure development and seed-related processes in other plant species [13, 18–20]. Although these candidate genes have mainly been reported in studies of pollen development or reproductive organ differentiation, and there is currently no direct evidence demonstrating a clear regulatory relationship between them and ovule/seed abortion in jujube, the biological processes in which they are involved—such as cell differentiation, surface structure formation, and maintenance of developmental signaling—are also relevant to ovule and seed development. Therefore, these results should be regarded as clues for understanding the possible regulatory network involving ZjABCG15 in reproductive development, rather than as direct evidence. Further studies, including protein interaction validation, genetic interaction analysis, and substrate identification, will be required to clarify the mode of action of ZjABCG15 and its associated molecular network. It should be noted that the current evidence for ZjABCG15 function is mainly derived from bioinformatic analysis, expression profiling, and heterologous transformation in Arabidopsis thaliana. Although this system may reflect a conserved role in plant reproductive development, it does not by itself demonstrate that ZjABCG15 plays the same mechanistic role in reproductive defects associated with ovule/seed abortion in jujube. It should also be noted that direct quantitative validation of ZjABCG15 transcript levels in the Arabidopsis overexpression and RNAi lines was not shown in the present study. Therefore, the observed transgenic phenotypes should be interpreted as heterologous functional evidence associated with the introduced constructs, rather than as direct proof of expression-level effects. Although this system can effectively reflect the conserved role of the gene in plant reproductive development, its precise function and regulatory mechanism in jujube still need to be confirmed through in situ functional analyses. Although heterologous transformation confirms its conserved function, the direct mechanism by which it leads to reproductive defects associated with ovule/seed abortion in jujube (e.g., the substrate it transports, and how it affects embryonic development) still needs final confirmation through homologous transformation experiments in jujube itself (e.g., stable genetic transformation or transient silencing) [21]. In addition, the actual transport substrate(s) of ZjABCG15, its upstream regulatory factors, and the specific stage(s) at which it functions during ovule-to-seed development in jujube remain unclear. These issues will be the focus of future studies. Results from heterologous functional validation in Arabidopsis support a conserved positive role of ZjABCG15 in reproductive development, particularly in seed development and seed-setting capacity, while its direct function and regulatory mechanism in jujube reproductive defects associated with ovule/seed abortion remain to be confirmed. These findings not only provide a new candidate gene for elucidating the molecular mechanisms underlying reproductive defects leading to ovule/seed abortion in jujube, but also lay a foundation for future studies on seed-setting improvement and molecular breeding in this species. Consistent with the result of the membrane-based yeast two-hybrid assay (Fig. 5), ZjABCG15 was shown to interact with LRR1 in the yeast system, providing preliminary support for its possible involvement in a broader regulatory network. Although the biological significance of this interaction in jujube reproductive development remains to be clarified, this finding offers an experimental basis for future studies on the molecular partners and regulatory mechanism of ZjABCG15. Supplementary Information Supplementary Material 1: Fig. S1. Basic characteristics and sequence analysis of the ZjABCG15 protein. Fig. S2. Advanced structural and functional prediction of the ZjABCG15 protein. Fig. S3. Subcellular localization of ZjABCG15 in N. benthamiana epidermal cells. Fig. S4. ZjABCG15 overexpression cloning electrophoresis images. Fig. S5. ZjABCG15-RNAi cloning electrophoresis images. Fig. S6. Identification of ZjABCG15-Myc overexpressing lines by HPT PCR. Fig. S7. Identification of RNAi ZjABCG15 lines by HPT PCR. Table S1. Primers used in this study. Table S2. Raw data of seed yield per plant in different lines. Table S3. Statistical summary of seed yield in wild-type and transgenic Arabidopsis lines. Supplementary Material 1: Fig. S1. Basic characteristics and sequence analysis of the ZjABCG15 protein. Fig. S2. Advanced structural and functional prediction of the ZjABCG15 protein. Fig. S3. Subcellular localization of ZjABCG15 in N. benthamiana epidermal cells. Fig. S4. ZjABCG15 overexpression cloning electrophoresis images. Fig. S5. ZjABCG15-RNAi cloning electrophoresis images. Fig. S6. Identification of ZjABCG15-Myc overexpressing lines by HPT PCR. Fig. S7. Identification of RNAi ZjABCG15 lines by HPT PCR. Table S1. Primers used in this study. Table S2. Raw data of seed yield per plant in different lines. Table S3. Statistical summary of seed yield in wild-type and transgenic Arabidopsis lines.