Cold stratification was used to examine the dormancy-breaking capacity of freshly harvested Pyrus betulaefolia seeds. Non-stratified seeds demonstrated strong dormancy (hereinafter referred to as “dormant seeds”) ( Figure 1A ). In contrast, seeds subjected to 2-week cold stratification showed significantly enhanced germination potential, reaching 93.33% after 5 days of incubation at 21°C. Maximal dormancy release was achieved following 5 weeks of cold stratification, resulting in 100% germination (hereinafter referred to as “nondormant seeds”), suggesting the experimental seeds possessed dormancy while maintaining high viability ( Figure 1A ). To elucidate the mechanisms underlying seed dormancy in Pyrus betulaefolia, germination assays were conducted using isolated embryos dissected from dormant seeds (hereinafter referred to as “coatless embryos”). Comparative analysis revealed that while excised embryos from dormant seeds exhibited delayed germination compared to non-dormant seeds, they ultimately achieved 100% germination within 6 days of incubation ( Figures 1B, C ). In stark contrast, intact dormant seeds showed a germination rate of 8.70%. These findings provided compelling evidence that the seed coat served as the primary constraint inhibiting germination in Pyrus betulaefolia seeds.

Seed coat-imposed dormancy in seeds involves three potential mechanisms, including physical impermeability to water, mechanical constraint, and endogenous inhibitors. To investigate the role of water impermeability, we measured water uptake kinetics in dormant seeds. As defined by Baskin and Baskin (2007), the seeds we used exhibited triphasic water uptake kinetics (Phase I: rapid 0–8 h absorption; Phase II: 8 h-3 d plateau; Phase III: post-3 d surge coupled with germination), confirming unimpeded water penetration through the seed coat ( Figure 1D ). The synchronized onset of Phase III and radicle emergence further verified that dormancy arrest occurs post-hydration, implicating physiological constraints. Using a seed coat bedding assay to physically separate seed coats from embryos ( Figures 1E, F ), germination rate of isolated embryos was 100% efficiency, while those cultured with dormant seed coats showed significantly suppressed germination rate (13.33% after 5 days of incubation). Together, these results demonstrated that dormancy of Pyrus betulaefolia seed was primarily mediated by inhibitors within the seed coats, which are gradually eliminated or degraded during cold stratification.

Given the well-established role of ABA in promoting seed dormancy and suppressing germination, the effects of fluridone, an ABA biosynthesis inhibitor, on dormant seeds in Pyrus betulaefolia were investigated. Treatment with 10 μM fluridone significantly promoted germination, achieving an 82.22% germination rate after 13 days under 16h/8h (light/dark) photoperiod at 22 ± 2°C ( Figures 2A, B ). To further examine the role of ABA in seed coat-mediated dormancy,seed coat bedding assays were conducted with or without 100 μM ABA supplementation ( Figure 2C ). Isolated embryos exhibited 100% germination, and non-dormant seed coats had no inhibitory effect. However, ABA completely suppressed germination in both isolated embryos and embryos co-cultured with non-dormant seed coats ( Figure 2C ), suggesting that dormant seed coats function similarly to exogenous ABA in preventing embryo germination. Besides, application of 10 μM fluridone in the seed coat bedding assay significantly attenuated the inhibitory effect of dormant seed coats, restoring embryo germination capacity ( Figure 2C ). These findings provide compelling evidence that dormant seed coats inhibit germination through endogenous ABA accumulation, and that this inhibition can be reversed by blocking ABA biosynthesis.

Expression patterns of key ABA metabolic and signaling genes in coatless embryos treated with either exogenous ABA or dormant seed coats were analyzed ( Figure 3 ; Supplementary Figure 1 ). Exogenous ABA treatment significantly increased transcript levels of the ABA biosynthesis gene PbeNCED-3 in both isolated embryos and embryos co-cultured with nondormant seed coats ( Figure 3A ). Similarly, dormant seed coats markedly enhanced the expression of ABA biosynthesis genes PbeNCED-1 and PbeNCED-3 ( Figure 3B ). Furthermore, we observed that exogenous ABA upregulated ABA signaling transcription factors PbeABI3-1, PbeABI4-1, PbeABI5-1, PbeABI5-2, and PbeABI5-5 ( Figure 3A ), while dormant seed coats specifically induced PbeABI3-1, PbeABI5-1 and PbeABI5-5 expression ( Figure 3B ). These transcriptional profiles implied that dormant seed coats modulated ABA homeostasis through coordinated regulation of NCED genes and selective activation of ABI3/ABI5-family signaling components, thereby maintaining seed dormancy.

Investigation of GA metabolic gene expression revealed that both exogenous ABA treatment and dormant seed coats significantly downregulated GA biosynthetic genes (PbeGA20ox-1, PbeGA20ox-2, and PbeGA3ox-1) while concurrently upregulating GA catabolic genes (PbeGA2ox-2 to PbeGA2ox-4, PbeGA2ox-6, and PbeGA2ox-9 to PbeGA3ox-12) in coatless embryos, with PbeGA2ox-3 and PbeGA2ox-4 showing strong induction (4-18-fold and 2-6-fold increases, respectively) under both treatments ( Figure 4 ). Notably, qRT-PCR analysis demonstrated stable expression of DELLA protein family genes (PbeGAI-Like-1 to PbeGAI-Like-3 and PbeRGL1-Like-1 to PbeRGL1-Like-3) regardless of treatment ( Supplementary Figure 2 ). Additionally, coatless embryos were highly sensitive to the GA biosynthesis inhibitor PAC, while exogenous GA_{4 + 7} could partially overcome dormancy, though both treatments exhibited similar early germination patterns that exceeded controls ( Figure 5 ). These findings demonstrated that seed coat-imposed dormancy involved ABA-mediated suppression of GA biosynthesis coupled with enhanced GA catabolism.

Based on the observed co-expression patterns of PbeABI5-1, PbeABI5-5, PbeGA2ox-3 and PbeGA2ox-4 during seed germination, the potential transcriptional regulation of GA2ox genes by ABI5 transcription factors was investigated. Bioinformatics analysis identified three ABRE motifs in the 2-kb promoter regions of both PbeGA2ox-3 and PbeGA2ox-4 ( Figure 6A ). Yeast one-hybrid assays revealed that PbeABI5-1 and PbeABI5-5 bound to multiple regions of the PbeGA2ox-3 promoter, while only PbeABI5-5 interacted with the PbeGA2ox-4 promoter ( Figures 6B, D ). Transient expression assays in Nicotiana benthamiana leaves demonstrated that PbeABI5-5 significantly activated transcription from both the PbeGA2ox-3 and PbeGA2ox-4 promoters, as shown by both qualitative and quantitative LUC reporter assays ( Figures 6E, F ). These findings establish that PbeABI5-5 directly binds to and activates transcription of GA catabolic genes, providing a molecular mechanism for the coordinated regulation of ABA signaling and GA metabolism during seed germination.

Seeds of Pyrus betulaefolia have a dormant stage, in which embryos with morphological dormancy are not yet fully developed, and they usually germinate after being stored for a period of time at room temperature (Zhang et al., 2022). Our study shows that the germination rate of Pyrus betulaefolia dormant seeds is low, and even if stored at room temperature for a period of time, the germination rate did not increase ( Figure 1 ). Considering that the germination rate of coatless embryos reached 100% after 6 days of incubation, together with rapid germination of coatless embryos, fully developed embryos and permeable seed coats, we confirmed that Pyrus betulaefolia seeds exhibit physiological dormancy. Following Baskin’s classification (Baskin and Baskin, 2007), we identified non-deep physiological dormancy based on three key traits: (1) coatless embryos germinated like non-dormant seeds ( Figure 1 ), (2) cold stratification (5 weeks) fully released dormancy ( Figure 1 ), and (3) GA_{4 + 7} treatment significantly enhanced germination ( Figure 5 ). Subsequent investigations could focus on performing detailed comparative analyses of dormancy depth throughout the Rosaceae family, with particular emphasis on identifying phylogenetic patterns and ecological correlates that may explain the evolutionary development and maintenance of these dormancy strategies.

In the mature seed of most angiosperms, the embryo is encased by two covering layers (coats): the living endosperm tissue and the outer layer of dead tissue (Topham et al., 2017). Experiments involving the removal of external testa layer while leaving the endosperm layer have shown that the endosperm provides most if not all of the germination repressive activity (Bethke et al., 2007). Previous reports have shown that seed coats surrounding the embryo contain compounds possessing germination-inhibitory activities, including ABA (Shimizu et al., 2022). Indeed, gene expression studies in isolated Arabidopsis seed coat tissue have shown that the endosperm expresses genes involved in ABA metabolism, and it has been demonstrated that the seed coat contains ABA (Barrero et al., 2010). While our data confirmed physiological dormancy in Pyrus betulaefolia ( Figure 1 and Figure 2 ), the role of seed coat extends beyond a passive barrier. Crucially, dormant seed coats actively synthesize ABA to maintain dormancy, as demonstrated in Arabidopsis by Lee et al. (2010), seed coat bedding assay reveals that imbibed dormant coats release ABA to suppress even non-dormant embryos. This mechanism seems conserved across species - in capeweed (Arctotheca calendula), seed coat-derived ABA imposes shallow dormancy (Chapman, 2000), while in recalcitrant Avicennia marina seeds, the pericarp (homologous to seed coats) retains high ABA levels that constrain germination post-dehydration (Farrant et al., 1993).

The germination rate was significantly higher than that of the control group when the seeds and embryos were treated with fluridone ( Figure 2 ). Consistently, in Arabidopsis, treating imbibed dormant wild-type seeds with fluridone will trigger their germination (Debeaujon and Koornneef, 2020; Millar et al., 2006), suggesting that repression of germination is an active process requiring de novo and constitutive production of the phytohormone ABA upon seed imbibition. Further supported by the observation that dormant and nondormant dry seeds contain similarly high amounts of ABA, which played an essential and well established developmental role during seed maturation, and upon seed imbibition, the high ABA levels present in both dormant and nondormant seeds drop drastically by about 10-fold within the first 12 h, and what indeed distinguishes dormant seeds from nondormant seeds is the capacity of dormant seeds to synthesize ABA de novo and constitutively once the initial drop of ABA levels has occurred upon seed imbibition (Ali-Rachedi et al., 2004; Piskurewicz et al., 2008), indicating that de novo ABA produced in the Pyrus betulaefolia seed coats and released toward the embryo to repress germination. Furthermore, the reported observations should not be meant to imply that the ABA derived from the seed coats is the only cause of dormancy. The strength of dormancy in a given seed, i.e., its capacity to restrain from germinating upon imbibition, is most likely related to the seed’s capacity to synthesize ABA as a whole: in both the endosperm and the embryo. Rather, we wish to propose that the endosperm is an essential contributor to sustain dormancy, i.e., to prevent germination over time upon imbibition.

That ABA synthesized in embryos also contributes to overall seed dormancy is most likely reflected by the observation that dormant embryos green and germinate slower than nondormant embryos upon dissection, and coatless embryos dissected from isolated dormant seeds 24 h after imbibition contained 2.5-fold more ABA than coatless embryos isolated from nondormant seeds (Lee et al., 2010). In agreement with this view, we found that the dormant seed coats promoted the expression of the ABA biosynthesis genes PbeNCED-1 and PbeNCED-3 in embryos ( Figure 3 ). It should be pointed out that dormancy-promoting function of the seed coat is not strictly restricted to the endosperm. However, PbeABI3-1 was significantly up-regulated under exogenous ABA or dormant seed coat treatment, which was similar to the expression patterns of PbeABI5-1 or PbeABI5-5 ( Figure 3 , 4 ). Studies have also shown that ABI3 has a domain interacting with ABI5, and ABI3 can also positively regulate the expression of ABI5 (Lopez-Molina et al., 2001; Nakamura et al., 2001; Lopez-Molina et al., 2002), indicating the relationship between PbeABI3 and PbeABI5 need to be further studied. Even with a larger relative increase, the absolute amount of PbeABI5-1 transcript induced by exogenous ABA or dormant seed coat treatment might still be considerably less than the basal and induced levels of PbeABI5-5 ( Figure 3 , 4 ). Crucially, our functional assays clearly demonstrated that only the PbeABI5-5 protein possesses the specific ability to bind the identified cis-element in the PbeGA2ox promoter and activate its expression ( Figure 6 ). While both are ABA-responsive bZIP transcription factors, they appear to have evolved distinct target gene specificities. In contrast, the expression of the ABA catabolic genes PbeCYP707A-1 to PbeCYP707A-5 were also significantly upregulated in the coatless embryos both treated with exogenous ABA and dormant seed coats ( Figure 3 ). Considering the similar feedback regulation in Prunus persica and Arabidopsis (Wang and Deng, 2014; Shu et al., 2016a), we hypothesized that the upregulation could be due to feedback regulation by the higher ABA content in these embryos. Notably, although gene expression analysis revealed potential modulation of ABA/GA homeostasis during dormancy release, we emphasize that the absence of ABA and GA quantification constitutes a limitation. While established models indicate NCED upregulation typically elevates ABA levels and GA2ox induction promotes GA catabolism (Hedden, 2020; Aloryi et al., 2025), future direct hormone measurements will allow us to rule out potential decoupling between gene expression patterns and actual hormone concentrations caused by post-transcriptional regulation or compensatory metabolic pathways. The relative contributions of embryonic versus seed coat-derived ABA also deserves careful consideration. While qRT-PCR data showed rapid upregulation of de novo ABA biosynthesis genes in isolated embryos ( Figure 3 ), potential priming effects from coat-translocated ABA prior to tissue isolation cannot be excluded. Future research will prioritize direct quantification of ABA levels to definitively resolve the contributions of seed coat-derived versus embryo-synthesized ABA in this system.

While studies have established the antagonistic relationship between abscisic acid and gibberellins in seed dormancy and germination (Lee et al., 2015; Hoang et al., 2014), the precise molecular mechanisms underlying their interaction remain incompletely understood. Crucially, seed coat-specific adaptations diverge across species: in Arabidopsis, endosperm-derived ABA transporters regulate embryonic ABI5 activation (Shimizu et al., 2022); in legumes, ABI5 directly promotes seed coat sclerification by repressing GA catabolism genes (Zinsmeister et al., 2016). AtABI4 has been identified as a crucial transcription factor in post-germination stages, which transcription and protein stability were enhanced by ABA, subsequently upregulating AtGA2ox7 expression to reduce active GA content; Conversely, GA signaling suppresses AtABI4 expression and accelerates its protein degradation, thereby inhibiting AtNCED6 transcription and ABA biosynthesis (Shu et al., 2016b). Cantoro et al. identified direct binding of SbABI4 and SbABI5 to the SbGA2ox3 promoter in vitro using electrophoretic mobility shift assay, suggesting these transcription factors may trigger GA catabolism to suppress germination in dormant grains (Cantoro et al., 2013).

Our research reveals another key player in this hormonal crosstalk - PbeABI5-5. Based on the induction of PbeGA2ox-3 and PbeGA2ox-4 genes expression by ABA ( Figure 4 ), we demonstrated that PbeABI5-5 directly bound to and activates the promoters of the two genes ( Figure 5 ). Through yeast one-hybrid and tobacco transient expression assays, we demonstrate that PbeABI5-5 directly binds to and activates the promoters of PbeGA2ox-3 and PbeGA2ox-4 genes ( Figure 6 ). These findings support our proposed model ( Figure 7 ), where in Pyrus betulaefolia seed coats enhance embryonic ABA biosynthesis. Elevated ABA levels upregulate PbeABI5-5 expression, which in turn activates PbeGA2ox genes through direct promoter binding. The resulting increase in GA catabolism reduces bioactive GA levels, maintaining embryonic dormancy by inhibiting germination. This regulatory cascade represents a novel mechanism through which ABA-dominated seed coat signaling influences embryonic GA metabolism to sustain dormancy.In conclusion, our study establishes that seed coat-derived ABA sustains dormancy in Pyrus betulaefolia by activating PbeABI5-5, which transcriptionally upregulates PbeGA2ox genes to enhance GA catabolism and suppress embryonic germination. This ABA-GA crosstalk mechanism, mediated through direct PbeABI5-GA2ox promoter binding, reveals a novel regulatory layer wherein seed coat signaling controls embryonic GA metabolism to maintain physiological dormancy.

Freshly harvested Pyrus betulaefolia seeds were collected from Xinjiang Uygur Autonomous Region, China. The seeds were rinsed with sterile water and air-dried in a shaded environment before storage in self-sealing bags at room temperature for subsequent experiments.

For germination assay, seeds were surface-sterilized with 5% NaClO for 5 min, followed by five washes with sterile water. After a 4-h imbibition in sterile water, seeds were placed on two layers of moist filter paper in Petri dishes. Each dish was treated with 2 mL of sterile water (control), ABA, fluridone, GA, or PAC, depending on the experimental conditions. The plates were incubated in a growth chamber (22 ± 2°C, 60–70% relative humidity, 16-h light/8-h dark cycle). Germination was monitored twice daily, and sterile water or treatment solutions were replenished as needed to maintain moisture. Seeds with a 2-mm radicle protrusion were scored as germinated. The germination rate was calculated as: Germination rate (%) = (Number of germinated seeds/Total seeds) × 100.

For cold stratification, sterilized seeds were mixed with moist sand and stored at 4°C. After stratification, seeds were transferred to germination conditions as described above.

The seed coat bedding assay was adapted from Lee et al (Lee et al., 2010). Briefly, imbibed seeds were dissected into embryos and seed coats using fine forceps. Ten coatless embryos were placed on filter paper overlying 30 seed coats, supplemented with 2 mL of sterile water, ABA, or fluridone, and incubated under standard conditions ( Supplementary Figure 3 ).

Water imbibition rates were determined by weighing seeds before and after incubation. Surface moisture was removed prior to initial weighing. The imbibition rate was calculated as: Water imbibition rate (%) = [(Post-incubation weight − Initial weight)/Initial weight] × 100.

The reference genome used was Pyrus x bretschneideri genome sequence. Total RNA was extracted using the HiPure Plant RNA Mini Kit (Magen, China) according to the manufacturer’s instructions. First-strand cDNA was synthesized from 2 µg of total RNA using the PrimeScript™ RT Reagent Kit with gDNA Eraser (Takara, Japan) following the manufacturer’s protocol. Quantitative real-time PCR (qRT-PCR) was performed using SYBR Premix Ex Taq™ (Takara, Japan), on a CFX96 Real-Time PCR System (Bio-Rad, USA). Gene-specific primers (designed with Primer3, http://bioinfo.ut.ee/primer3-0.4.0/) are listed in Supplementary Table S1 . Prior to analysis, primer efficiency was validated using a standard curve (80–120% efficiency required). Each reaction was run in three technical replicates, and gene expression levels were calculated using the 2^-ΔΔCt method, with PpActin (JN684184) as the internal reference (Liu et al., 2012).

The 2000bp region upstream from the start codon of PbeGA2ox genes were extracted from the Pyrus betulaefolia genome as a promoter. Then, the sequences of promoters were uploaded to the Plantcare (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) to predict ABRE motifs.

Yeast one-hybrid (Y1H) assays were performed using the Matchmaker™ Gold Yeast One-Hybrid Library Screening System (Clontech, USA) following the manufacturer’s protocol. The primers used for constructing bait and prey are listed in Supplementary Table S2 . Different promoter fragments of PbeGA2ox containing the ABRE sequence were cloned into the pAbAi vector. The recombinant plasmid was linearized and integrated into the genome of the Y1H Gold yeast strain. Subsequently, yeast cells were co-transformed with either PbeABI5-AD or the empty AD vector. Protein-DNA interactions were assessed based on yeast growth on SD/-Leu medium supplemented with 100–200 ng/mL aureobasidin A.

The dual-luciferase assay was performed in Nicotiana benthamiana leaves following the method of Niu et al (Niu et al., 2016). The full-length coding sequences of PbeABI5-1 and PbeABI5-5 were cloned into the pGreenII 0029 62-SK vector, while the promoter regions of PbeGA2ox-3 and PbeGA2ox-4 were inserted into pGreenII 0800-LUC. All constructs were transformed into Agrobacterium tumefaciens strain GV3101 using the freeze-thaw method. For infiltration, Agrobacterium cultures (OD₆₀₀ = 1.0) carrying the effector and reporter constructs were mixed at a 10:1 ratio and incubated at room temperature for 2 h before infiltration into leaves. After 48-72 h post-infiltration, leaf samples were collected for observation. For qualitative detection, luciferase activity was visualized using the NightSHADE LB 985 imaging system (Berthold Technologies) after infiltration of 0.2 mM luciferin (Promega) 30 min prior to imaging. For quantitative analysis, firefly and Renilla luciferase activities were measured using the Dual-Luciferase^® Reporter Assay System (Promega) on a Modulus™ Luminometer (Promega). Reporter gene expression was calculated as the ratio of LUC_firefly: REN_Renilla. Each experiment included three independent biological replicates, with six technical replicates per biological replicate.

All experiments were arranged in a completely randomized design. Data were analyzed with SPSS Statistics v.18.0 (IBM, USA). Statistical significance was determined at p < 0.05 using appropriate tests. Data visualization was performed using GraphPad Prism 7.00. Values are presented as mean ± standard error (SE) from at least three independent biological replicates.